Taxonomy, phylogeny and biogeography of francolins (‘Francolinus’ spp.) Aves: Order Galliformes Family: Phasianidae To Thesis presented for the degree of w n Tshifhiwa Gift Mandiwana-Neudani DOCTOR OF PHILOSOPHY ap e Faculty of Science C DST/NRF Centre of Excellence at the of Percy FitzPatrick Institute of African Ornithology ty Department of Biological Sciences Private Bag X3 Rondebosch, 7701 South Africa U ni ve r si UNIVERSITY OF CAPE TOWN Principal supervisor: Professor Timothy M. Crowe, DST/NRF Centre of Excellence at the Percy FitzPatrick Institute of African Ornithology, Department of Biological Sciences, University of Cape Town, South Africa. E-mail: timothy.crowe@uct.ac.za Co-supervisor: Professor Rauri C.K. Bowie, Museum of Vertebrate Zoology and Department of Integrative Biology, University of California, Berkeley, USA. E-mail: bowie@berkeley.edu October 2013 n of C ap e To w The copyright of this thesis vests in the author. No quotation from it or information derived from it is to be published without full acknowledgement of the source. The thesis is to be used for private study or noncommercial research purposes only. U ni ve rs ity Published by the University of Cape Town (UCT) in terms of the non-exclusive license granted to UCT by the author. TABLE OF CONTENTS Page(s) Dedication i Declaration ii Acknowledgements iii-viii Abstract ix-xi Chapter 1 General introduction: A review of the taxonomy, phylogeny and biogeography of francolins (‘Francolinus’ spp.)…………………………………..1-36 Chapter 2 Phylogenetics of evolutionarily enigmatic terrestrial gamebirds (Aves: Galliformes) with special regard to ‘Francolinus’ spp...............................................................................37-80 Chapter 3 A study of the gross morphological and histological syringeal features of francolins and spurfowls in a phylogenetic context………………………………………………81-103 Chapter 4 Taxonomic and phylogenetic utility of variation in advertising calls of francolins and spurfowls (Galliformes: Phasianidae)……………………………………………………...104-151 Chapter 5 Taxonomy and phylogeny of ‘true’ francolins………..................152-214 Chapter 6 Taxonomy and phylogeny of spurfowls…………………………215-273 Chapter 7 Historical biogeography of francolins and spurfowls……………274-319 Chapter 8 Summary and synthesis…………………………………………..320-327 Chapter 9 References………………………………………………………..328-363 DEDICATION I dedicate this thesis to my parents who shared the same belief with Robert M. Maciver who once said “…When you educate a man you educate an individual; when you educate a woman you educate a whole family…” Without them, this work would never have begun. To Enos, whose love and support made this journey possible, and to my beloved Uhone whose flexibility and adaptability afforded the completion of this work… i DECLARATION I hereby declare that the work presented in this thesis is my own, unless otherwise stated. Apart from the guidance received from my supervisors, assistance from all institutions and individuals in this thesis is acknowledged. This thesis has not been previously submitted for the degree at this or any other university and I therefore present it for examination for the degree of PhD. Signature Removed 10/03/2014 ………………………………………. ……………………….............. Tshifhiwa G. Mandiwana-Neudani Date ii ACKNOWLEDGEMENTS Many people and institutions contributed to the success of this work in one way or the other. First, there is of course Mrs B.P. Hall whose pioneering monograph on the evolution of francolins published in 1963 laid key foundations for this research. Second, I thank my supervisors, Professors Timothy M. Crowe and Rauri C.K. Bowie, for helping to develop and extend her research on these fascinating birds. Their financial support, mentorship and intellectual contributions were essential to the completion of this academic journey. I am indebted to the directors of institutions and heads and curators of collections housed at the: American Museum of Natural History (Professor Joel L. Cracraft, Mr Paul R. Sweet, Mr Peter Capainolo - Division of Vertebrate Zoology, Ornithology); the Natural History Museum at Tring (Dr Robert Prys-Jones and Mark Adams – Department of Zoology); Ditsong National Museum of Natural History (formerly Transvaal Museum - Northern Flagship Institution) (former director Professor Francis Thackeray, Ms Tamar Cassidy); Iziko Museums of Cape Town (Natural History) (Dr Hamish Robertson, Ms Denise Hamerton – Terrestrial Vertebrates, Birds); and the Field Museum of Natural History (Drs Shannon Hackett, John Bates and David E. Willard - Division of Birds) for opening the doors of their collections allowing me to personally examine the material under their care and their willingness to allow me to subsample toe pads for DNA analysis. I am also indebted to the University of Uganda, the United States National Museum of Natural History, the Zoological Museum of the University of Copenhagen, the Museum of Comparative Zoology (Harvard University), the Los Angeles County Natural History Museum and the Louisiana State University Natural History Museum for providing the specimens. Professor Rebecca Kimball iii provided key additional DNA sequences for a range of galliforms and wingshooters provided key tissues. The first sequences of DNA data were generated at the Molecular Ecology and Evolution Programme (MEEP) laboratory at the University of Pretoria with the help of the students who were using the laboratory at the time. Professor Paulette Bloomer, the Director of the MEEP laboratory is thanked for generously affording me a space in the laboratory and teaching me the basic techniques required for DNA sequencing. Approximately 75% of the DNA samples sequenced in this project were subsampled from long-dead museum bird study skins, and this had implications with regard to the technical expertise required to successfully sequence these samples. Dr Robert G. Moyle and Ms Julie Feinstein (Division of Vertebrate Zoology, American Museum of Natural History) played an important role in passing on their skills of sequencing ancient DNA samples to me. I am very thankful to my friend and former classmate Jack Henning, Professor Lorenzo Prendini (Division of Invertebrate Zoology, American Museum of Natural History) and his wife Dr Elizabeth Scott who generously accommodated me during my visits to the AMNH. The former manager of the University of Cape Town (UCT) Department of Biological Sciences Systematics laboratory Dr Tracey Nowell and Dr Pia Eldenäs, laboratory engineer of the molecular laboratory (Swedish Museum of Natural History) are thanked for their help with laboratory techniques. Professor Mari Källersjö and Dr J. Steven Farris (molecular laboratory, Swedish Museum of Natural History) are thanked for making the collaboration possible through the South African – Swedish Bilateral Partnership. Recordings of vocalizations were sourced from various institutions including the Transvaal Museum Sound Library, British Library Sound Archive and Macaulay iv Library of Natural Sounds. Michael Mills of Birding Africa is thanked for providing us with the Gibbon’s (1995) and Chappuis’ (2000) collection and blood sample of Swierstra’s Spurfowl. Callan Cohen (Birding Africa & Percy FitzPatrick Institute) is thanked for passing on a recording of Jackson’s Spurfowl from Brian Finch of WINGS birding. The analyses of calls would not have been possible without the help of Mr Jean-pierre Richard of the Éthologie Animale et Humaine’ research group, Université de Rennes, who taught me to use his software package ‘ANA’ (Jean-Pierre Richard, Unpublished version) and to interpret sonograms. Dr Martine Hausberger, the director of Éthologie Animale et Humaine’ research group (Université de Rennes) and Dr Laurence Henry (Université de Rennes) are thanked for their useful comments on the vocalization chapter of this thesis and also for handling all the logistics relating to my trips to France and making my stay in Rennes enjoyable. I am indebted to the wonderful contribution made by Dr Cecilia Kopuchian (División Ornitología, Museo Argentino de Ciencias Naturales, Argentina) on the chapter involving the anatomy of syringes. I owe this to Dr Pablo A. Goloboff (Instituto Miguel Lillo, San Miguel de Tucumán, Argentina) who linked me with Dr Kopuchian. She provided me with protocols to stain gross morphology of galliform syringes and useful literature on syringeal-systematics work. She helped with the interpretation of the observations made and also by making invaluable comments on the syringeal draft manuscript. I am very grateful for the contribution made by Ms Morea Peterson (Department of Human Biology, University of Cape Town) for her superb technical assistance during the preparation of histological sections. Ms Barbara Moore and Barbara Young (Department of Human Biology, UCT) also provided assistance with histological procedures. I also would like to thank Professor Graham Louw v (Department of Human Biology, UCT) who helped with the interpretation of histological sections and provided very useful comments on the histological component of the chapter which focused on syringes of francolins and spurfowls. Various gamebird hunters and Owen Davies (Percy FitzPatrick Institute of African Ornithology, University of Cape Town) are also thanked for help with the collection of whole specimens from which I extracted syringes and for providing hearts for DNA-based research. I also would like to thank Ms Tamar Cassidy of Ditsong National Museum of Natural History for providing me with the whole alcohol-preserved specimens from which I extracted the syringes. Ms Anna Crowe, my fellow colleagues at the FitzPatrick Institute, Graeme Oatley and Lisa Nupen are thanked for preparing sub-samples of hearts of francolins and spurfowls on various farms in the Eastern, Western Cape and Mpumalanga Province, and Owen Davies is thanked for his technical involvement in the syringeal work. J. Salvador Arias (Instituto Miguel Lillo, S.M. de Tucumán, Argentina) played an important role in introducing me to the historical biogeographical analyses using his spatial analysis of vicariance approach which was implemented in the software Vicariance Inference Program (VIP). Professor Terry Hedderson (Department of Biological Sciences, UCT) also helped with ancestral trait reconstruction analyses which were performed on Mesquite and the interpretation of results thereof. Dr Fenton Cotterill (AEON, UCT) provided illuminating discussions about the natural history of key francolin species and the effects of topography and vegetation on francolin biogeography. Dr Jérôme Fuchs, a post-doctoral fellow at the Percy FitzPatrick Institute, UCT and Museum of Vertebrate Zoology, University of California, Berkeley is thanked for the help he provided with some phylogenetic inference methods. vi Professor William Bond (Department of Biological Sciences, UCT) advised me on aspects that relate to the vegetation map. I am also thankful to all my fellow colleagues in the Department of Biological Sciences for consistently giving me support through this journey. Ms Sandy Smuts (Department of Biological Sciences Senior Administrator) is thanked for her administrative support and Mr Gonzalo Aquilar (Department of Biological Sciences) for the technical support. The administrative staff at the Percy FitzPatrick Institute is thanked for their help in various ways. Ms Margaret Koopman (Niven librarian) is thanked for her fast response and dedication in helping with literature. Mr Chris Tobler (Principal Technical officer) is thanked for helping with computational technicalities. Ms Hilary Buchanan and Tania Jansen are thanked for their quick response and help with all the trips undertaken during this study and Mr Lionel Mansfield (former administrative staff member) is thanked for ordering laboratory consumables when needed. The financial support for this project was provided by the South African Department of Science and Technology (DST)/National Research Foundation (NRF) Centre of Excellence at the Percy FitzPatrick Institute of African Ornithology. My work on vocalizations including visits to Rennes, France was funded through the South African (NRF) - France (CNRS) Bilateral Programme while the syringeal component was partly supported by CONICET PIP 112-200801-0074 through my collaboration with Dr Kopuchian. The use of illustrations of francolins and spurfowls from Handbook of the birds of the world: Vol. II. (del Hoyo et al. 1994) was allowed by Lynx edicions, Barcelona. I cannot forget to thank my friend Dr Tshifhiwa Nangammbi (Department of Zoology, University of Venda) for always making sure that I stayed positive and vii acknowledging my effort at times when I could not see the light. Enos, your love and support of all the time kept me alive. Uhone, you surprised me, you did not act like a 10-year old when you bowed to our earnest request that you needed to be away from me so that I can complete this thesis. My sincere thanks to my sister-in-law Melta Neudane who heartfully made sure that my house was always in order in my absence and presence. My parents played an immense role in shaping my future as I advanced in my career. Their parenting role generated a sense of work and effort without which this body of work would never have been completed. viii ABSTRACT MANDIWANA-NEUDANI, T.G. 2013. Taxonomy, phylogeny and biogeography of francolins (‘Francolinus’ spp.) Aves: Order Galliformes, Family Phasianidae. PhD thesis, DST/NRF Centre of Excellence at the Percy FitzPatrick Institute of African Ornithology, Department of Biological Sciences, University of Cape Town, Private Bag X3, Rondebosch, 7701, South Africa. Francolins (Francolinus spp.) are small to medium-sized, sedentary, Old World, partridge/quail-like, terrestrial gamebirds (Order Galliformes) that occupy diverse habitats ranging from dry/open/scrubby lowland and montane grasslands, bushveld and savanna/woodland to mesic montane/lowland forests and forest edges. Some francolins have complex distribution patterns and also are morphologically, ecologically and behaviourally diverse. At the start of this research, Francolinus Stephens, 1819 was considered a monophyletic galliform genus comprising 41 species (36 African and five Asiatic) divided among eight putatively monophyletic species groups and four taxonomically enigmatic species. However, different taxonomic revisions, especially post Hall’s (1963) classic monograph, challenged the monophyletic status of the genus and that of some of its designated species groups differed markedly in the number of recognized subspecies. Furthermore, there was debate concerning the geographical origin of the genus: Asia versus Africa. Some of the early molecular research on a few exemplar francolin species based on partial mitochondrial Cytochrome-b DNA sequences and Restriction Fragment Length Polymorphisms (RFLPs) also challenged the monophyly of the genus and that of some of Hall’s (1963) species groups. These findings suggested that francolins may form at least two distantly related lineages called ‘patryse’ (partridges) and ‘fisante’ (pheasants) by Afrikaans-speaking people. Patryse, or ‘true’ francolins, had been divided into as many as five genera (Francolinus, Ortygornis, Dendroperdix, Peliperdix, Scleroptila) and fisante, or spurfowls, all grouped into a single genus, Pternistis. Research in this thesis is based on: mitochondrial and nuclear DNA sequences (5554 base pairs), organismal and vocal characters of francolins and spurfowls. Galliform terminal taxa rooted variably on Anseriformes and Megapodes (Chapter 2), Gallus gallus and Bambusicola thoracica (francolins) and Coturnix coturnix, Alectoris chukar (spurfowls) (Chapters 3, 4, 5, 6, 7). The galliform phylogeny (Chapter 2), morphology of syringes (Chapter 3) and the ix vocalizations of many taxa (Chapter 4) confirm this phylogenetic dichotomy between francolins and spurfowls and the Bayesian reconstruction of molecular divergence date reveal they last shared a common ancestor at c. 33.6 mya. In addition, the taxonomic and phylogenetic outcome of francolins (Chapter 5) and spurfowls (Chapter 6) suggest the need to recognize two additional genera of francolins: Ortygornis for the Asian Grey Francolin ‘Francolinus’ pondicerianus, Swamp Francolin, ‘Francolinus’ gularis (two of four of Hall’s taxonomically enigmatic species), and the African Crested Francolin ‘Dendroperdix’ sephaena; and Afrocolinus gen. nov., for the African Latham’s Francolin ‘Francolinus’ lathami (another of Hall’s taxonomically enigmatic species), which links the basal francolins, Francolinus and Ortygornis spp. with relatively terminal genera, Peliperdix and Scleroptila. The ‘true’ francolins, those related to the nominate species, Black Francolin Francolinus francolinus, are monophyletic (having originated at c. 7.6 mya), and sister to junglefowls (Gallus spp.) and Bamboo Partridges (Bambusicola spp.), and are relatively small and have at most one (generally short) tarsal spur positioned about half-way down the tarsus, give more musical and whistling calls, generally roost on the ground, and have striped and barred rufous dorsal plumage resembling that of quail. Spurfowls which originated at c. 8.7 mya are monophyletic, but are distantly related to francolins. They are sister to the Bush Quail (Perdicula asiatica) and the Sand Partridge (Ammoperdix heyi) and a range of Palaearctic and African partridges and quails and are relatively large, often have two (generally long) tarsal spurs (the lower of which is positioned about two-thirds down the tarsus), emit raucous grating or cackling advertisement calls (given mainly at dawn/dusk). They have been observed perching in trees and have dark dorsal plumage vermiculated with white or buff. Thus, both francolins and spurfowls appear to have Asian, not African origins. Four of eight of Hall’s (1963) groups were recovered and the phylogenetic hypotheses for all four of Hall’s taxonomically enigmatic species are offered. The putative taxonomic ‘link’ species between francolins and spurfowls, the Crested Francolin Dendroperdix (now Ortygornis) sephaena, is shown to consistently group with two of Hall’s unplaced Asian species O. pondicerianus and O. gularis. Another important discovery was that one of the unplaced species, the African Nahan’s Francolin Francolinus (now Ptilopachus) nahani is not a francolin, but is sister x to the African Stone Partridge Ptilopachus petrosus (Chapter 2), which in turn, are sister to the New World quails (Odontophoridae). xi CHAPTER 1 A review of the taxonomy, phylogeny and biogeography of francolins Quail-like – ‘True’ Francolins – “patryse” Francolinus francolinus Ortygornis ‘Dendroperdix’ sephaena Peliperdix coqui Scleroptila shelleyi 1 Partridge-like Francolins – Spurfowls – “fisante” Pternistis hartlaubi Pternistis afer Pternistis erckelii Pternistis capensis Pternistis squamatus 2 Nahan’s ‘Francolin’ ‘Acentrortyx’ ‘Francolinus’ nahani 3 Introduction – the status quo Taxonomy - What are francolins? Francolins (Table 1.1), at the initiation of this research, were classified as small to medium-sized, sedentary, Old World, partridge/quail-like gamebirds adapted to varied, but primarily tropical/sub-tropical, habitats ranging from dry, lowland grassland to montane forests (Hall 1963, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). At the outset of this research, taxonomically, francolins were invariably placed in the Order Galliformes and family Phasianidae. In some more finely partitioned classifications, within the Phasianidae, they were placed in the sub-family Phasianinae (including Phasianus and related species) and, with other Old World partridge- and quail-like gamebirds (e.g. Perdix and Coturnix spp.), in the tribe Perdicini (Chapin 1932, Peters 1934, Wolters 1975-82, Crowe et al. 1986, Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994, Madge and McGowan 2002). Although relatively stable at higher taxonomic levels, there remains controversy with respect to the number and limits of subspecies, species and genera of francolins. A brief respite The distributions of 36 francolin species are confined to Africa. Those of five further species extend outside of Africa north to the Caspian Sea and east to central Asia and southern China (Johnsgard 1988). Based on the findings of B.P. Hall’s (1963) monograph, francolins were (and generally still are) lumped into a single perdicine genus, Francolinus, comprising 41 species (del Hoyo et al. 1994). Francolinus Stephens, 1819 is currently the most specious genus in the Galliformes (Morony et al. 4 1975, del Hoyo et al. 1994) and one of the largest genera in the class Aves (Bock and Farrand 1980). However, other taxonomic treatments have partitioned francolins into as many as eight genera (Francolinus, Pternistis, Peliperdix, Scleroptila, Dendroperdix, Ortygornis, Acentrortyx, and Chaetopus) (Roberts 1924, Chapin 1932, Peters 1934, Mackworth-Praed and Grant 1952, 1962, 1970, Hall 1963, Wolters 1975-82, Hockey et al. 2005). Nevertheless, virtually immediately after lumping francolins into a single genus, Hall (1963) hypothesized that all but four Francolinus spp. could be assigned to eight putatively monophyletic groups (Table 1.1). Morphology and behaviour Francolins, like many perdicine gamebirds, are characterized by having 14 tail feathers that are presumably moulted centrifugally (Johnsgard 1988). Some more partridge-like species of francolins are: relatively large, often have two (generally long) tarsal spurs (positioned about two-thirds down the tarsus), emit raucous grating or cackling advertisement calls (given mainly at dawn/dusk). They have been observed perching in trees or other elevated structures and have dark dorsal plumage vermiculated and/or striped with white or buff (Crowe et al. 1986, Milstein and Wolff 1987, Crowe et al. 1992). Other more quail-like francolins are relatively small, have at most one (generally short) tarsal spur (positioned about half-way down the tarsus), have relatively musical and whistling calls, generally roost on the ground, and have striped and barred rufous dorsal plumage resembling that of quails (Coturnix spp.) (Crowe et al. 1986, Milstein and Wolff 1987). This has led the Afrikaans-speaking people to group francolins into two supra-specific groupings; fisante (or pheasants) for the former and patryse (or partridges) for the latter (Milstein and Wolff 1987). 5 However, the above distinction was difficult to apply to the Crested Francolin, Dendroperdix sephaena, since it has attributes of both fisante and patryse. It is relatively small, has quail-like back plumage like patryse, but emits a raucous call, has well-spurred, red tarsi and roosts in trees like many fisante. Thus, it is a ‘linking form’ that persuaded Hall to lump all 41 species into a single genus Francolinus. The tortuous taxonomic history of francolins The genus Francolinus was first described by Stephens in 1819. Since that time, various taxonomic treatments and revisions have attempted to partition the francolins into two or more genera, a variable numbers of species and a plethora of subspecies (Table 1.2). However, Hall’s (1963) monograph still remains the standard reference as far as the classification of francolins is concerned (Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994). Table 1.3 attempts to summarize the taxonomic chaos while Table 1.4 outlines the types of characters that have been used in various treatments. Alternative taxonomic treatments Roberts (1924), in his checklist of birds of South Africa, recognized six genera of francolins (Table 1.2): 1. Pternistis Wagler, 1832 -- the species swainsonii, afer, humboldtii and castaneiventer were recognized, the last of which was divided into three subspecies castaneiventer, notatus and krebsi. 2. Chaetopus Swainson, 1837 -- capensis, adspersus and natalensis. 3. Peliperdix Bonaparte, 1856 -- hartlaubi. 6 4. Dendroperdix Roberts, 1922 -- rovuma (recognized by most avian taxonomists as a subspecies) and sephaena. 5. Scleroptila Blyth, 1849 -- afra (= afra), levaillantii, shelleyi and gariepensis (= levaillantoides). 6. Ortygornis Reichenbach, 1853 -- coqui. Roberts, in his first edition (1940) of ‘The Birds of South Africa’ followed the above-mentioned checklist, recognizing the same genera as in Roberts (1924), with just one change in which he erected the genus Chapinortyx Roberts, 1928 to distinguish the species hartlaubi. The same species as those in the 1924 checklist were retained, but additional subspecies were delineated. Within Ortygornis coqui, three subspecies were recognized (coqui, campbelli, vernayi); six within Dendroperdix sephaena (the nominate sephaena, zuluensis, zambesiae, chobiensis, mababiensis, thompsoni); and six within Scleroptila levaillantoides (levaillantoides, gariepensis, ludwigi, pallidior, langi, kalaharica). In addition to the above species, Scleroptila jugularis was delineated within which he recognized one subspecies cunenensis; four subspecies were recognized for Pternistis swainsonii (swainsonii, damarensis, chobiensis, gilli) and two subspecies for Pternistis afer (afer, cunenensis). A further subspecies lehmanni, was recognized for Pternistis castaneiventer. In the 2nd edition of ‘Roberts Birds of Southern Africa’ (McLachlan and Liversidge 1957), the number of genera recognized was reduced to two (Francolinus and Pternistis). Francolinus was assigned to all partridges and fisante (including hartlaubi plus the three subspecies hartlaubi, bradfieldi, crypticus) that do not belong to 7 Hall’s Bare-throated Group, whereas Pternistis was assigned strictly to the Barethroated taxa (Table 1.2). Even though the 1st and the 2nd editions covered the same geographic area, the delineation of taxa continued to vary with the lumping and splitting of previously recognized taxa and the erection of new taxa. This resulted in great inconsistency and instability in the number of taxa recognized. The following taxa were typically recognized: subspecies coqui, vernayi, with the addition of hoeschianus and the removal of campbelli within F. coqui; subspecies mababiensis and chobiensis were synonymized with F. sephaena; only five subspecies were recognized within F. levaillantoides (levaillantoides, pallidior, langi, wattii and cunenensis which was considered a subspecies of S. jugularis) leaving out gariepensis, ludwigi, kalaharica; subspecies gilli was synonymized with swainsonii; and notatus, swynnertoni and humboldti were atributed to P. afer. The subspecies lehmanni was removed from P. castaneiventer. The 3rd and 4th (McLachlan and Liversidge 1970, 1978 respectively) editions were similar to the 2nd edition with regard to the taxa which were recognized. Subspecies lundazi was recognized as an additional subspecies of P. swainsonii, and gilli was re-considered, whereas chobiensis was synonymized with swainsonii. Subspecies lehmanni was re-considered as a valid subspecies of P. afer. The 5th and the 6th editions (Maclean 1985, 1993 respectively) covered the southern African region and recognized the same set of species, the only major difference being that the 5th edition used the genus Francolinus for both patryse and fisante and that there was no mention of subspecies with the exception that it is 8 indicated that rovuma was considered a subspecies of D. sephaena. The 6th edition made use of the genera Francolinus and Pternistis as in the previous editions and had the same composition of species as in the 4th edition. The most recent 7th edition of Roberts’ Birds of Southern Africa (Hockey et al. 2005), recognized four genera: 1. Peliperdix -- coqui within which they recognized the subspecies coqui, which included campbelii, hoeschianus, lynesi, kasaicus, ruahdae, stuhlmanni, vernayi and angolensis. 2. Dendroperdix -- sephaena comprised of sephaena which included the subspecies zuluensis, rovuma and zambesiaie (synonyms of which are thompsoni, chobiensis and mababiensis). 3. Scleroptila -- afra, levaillantii, shelleyi within which they recognized subspecies shelleyi (synonym sequestris) and canidorsalis; levaillantoides within which they recognized the subspecies levaillantoides (synonyms ludwigi, langi and kalaharica), jugularis (synonyms cunenensis and stresemanni) and pallidior (synonym wattii). 4. Pternistis -- hartlaubi, adspersus (subspecies - kalahari and mesicus), capensis, natalensis, afer (subspecies - afer, swynnertoni, castaneiventer) and swainsonii (subspecies - swainsonii and lundazi). Chapin (1932) divided francolins between two genera Pternistis and Francolinus: 1. Pternistis -- included the three species afer, rufopictus and swainsonii 9 2. Francolinus -- this genus was assigned to all the other francolins with the exception of nahani. Francolinus nahani was described by (Dubois, 1905) and moved to Peliperdix by van Someren (1926). During the same year, Chapin (1926) suggested nahani be placed in a monotypic genus, Acentrortyx, as he argued that this perdicine was not congeneric with Francolinus. Peters (1934) recognized 39 species and split the francolins between two genera: Francolinus and Pternistis. However, he retained the use of the genus Francolinus in the broader sense, pending a detailed revision of all the species. Peters’ deliniation of Francolinus included members of Hall’s (1963) Spotted, Montane, Scaly, Vermiculated, Red-winged, Striated and Red-tailed Groups. The genus Pternistis was restricted to members of Hall’s Bare-throated Group. He recognized ‘schlegelii’ as a subspecies of Francolinus coqui. Another Peters’ ‘peculiarity’ is that he elevated Francolinus castaneicollis ‘atrifrons’ to full species status, whereas other authors considered this taxon a subspecies of castaneicollis. Mackworth-Praed and Grant (1952, 1962, 1970) also recognized two genera: 1. Pternistis -- rufopictus; leucoscepus (subspecies leucoscepus, infuscatus, muhamedben-abdullah, kilimensis); swainsonii (subspecies swainsonii, damarensis, chobiensis, gilli), and afer (subspecies afer, cranchii, humboldtii, benguellensis, intercedens, castaneiventer, krebsi, swynnertoni, notatus, lehmanni, cunenensis, loangwae, harterti, nyanzae, leucoparaeus, bohmi, melanogaster, itigi). 2. Francolinus -- sephaena (subspecies sephaena, grantii, zambesiae); streptophora; and rovuma (which they elevated to species, subspecies rovuma and spilogaster); 10 squamatus (subspecies squamatus, schuetti, doni, maranensis, usambarae, uzungwensis, chyuluensis); griseostriatus and ahantensis (subspecies ahantensis, hopkinsoni); nobilis (subspecies nobilis, chapini); castaneicollis (subspecies castaneicollis, gofanus, ogoensis, atrifrons, kaffanus); jacksoni (subspecies jacksoni, gurae, pollenorum); erckelii (subspecies erckelii, pentoni); swierstrai; camerunensis; hildebrandti (subspecies hildebrandti, altumi, johnstoni); harwoodi; bicalcaratus (subspecies bicalcaratus, thornei, adamauae, ogilvie-granti, ayesha); clappertoni (subspecies clappertoni, gedgii, heuglini, cavei, sharpii, konigseggi, nigrosquamatus); icterorhynchus (subspecies icterorhynchus, emini); adspersus; capensis; natalensis (subspecies natalensis, neavei); hartlaubi (subspecies hartlaubi, bradfieldi, crypticus); coqui (subspecies coqui, angolensis, vernayi, hoeschianus, hubbardi, ruahdae, thikae, maharao, kasaicus, spinetorum); schlegelii; albogularis (subspecies albogularis, buckleyi, dewittei); finschi; levaillantoides (subspecies levaillantoides, gariepensis, jugularis, pallidior, ludwigi, langi, cunenensis, kalaharica, wattii); levaillantii (subspecies levaillantii, kikuyuensis, crawshayi, benguellensis, clayi); shelleyi (subspecies shelleyi, whytei, theresae, elgonensis, uluensis, macarthuri); afer (= africanus = afra) and the unplaced nahani and lathami (subspecies lathami, schubotzi). Mackworth-Praed and Grant (1952, 1962, 1970) demonstrate the taxonomic problems created by their variable use of the name ‘afer’ at both the specific and subspecific level: ‘afer’ was used in their 1952 publication to refer to Grey-winged Francolin, currently deliniated as Scleroptila afra (they described this species’ distribution as ranging from the Cape to Ethiopia, but with interruptions) and ‘cranchii’ for Red-necked Spurfowl currently Pternistis afer. The Red-necked Spurfowl is widely 11 distributed in the lowlands of central and east Africa, from Gabon, western Angola, northwest Namibia, and Zimbabwe to South Africa. However, in 1962 they used ‘afer’ for both the Grey-winged Francolin and Red-necked Spurfowl. Currently, the specific name ‘afer’ is used for Red-necked Spurfowl Pternistis afer whereas ‘africanus/afra’ is used for the Grey-winged Francolin Scleroptila afra. Furthermore, within the Grey-winged Francolin (presently considered a South African and Lesotho endemic – Clancey 1986, Sinclair and Ryan 2003), MackworthPraed and Grant (1952, 1962, 1970) included several taxa which occur in sub-Saharan Africa. These were uluensis, lorti, gutturalis, psilolaema, ellenbecki, archeri, friedmanni, marcathuri and stantoni. Other authors (Table 1.2) assigned uluensis and macarthuri to Shelley’s Francolin F. shelleyi, and gutturalis, archeri and friedmanni to the Orange River francolin F. levaillantoides. While they considered psilolaema a subspecies of Grey-winged Francolin, other authors (Table 1.2) recognized this taxon to be a full species of the Red-winged Group with ellenbecki being one of its subspecies. Wolters (1975-82) proposed the first cladistic system of genera and subgenera for francolins in which he recognized Hall’s (1963) 41 species. He recognized six genera: 1. Francolinus -- this genus included Hall’s Spotted taxa francolinus, pintadeanus, pictus and her unplaced gularis, with francolinus, pintadeanus, pictus (except gularis) lumped into a subgenus, Francolinus gularis was not assigned to any subgenus. 2. Pternistis -- Hall’s Bare-throated leucoscepus, rufopictus, swainsonii and afer (all placed in the subgenus Pternistis) as well as the Montane species erckelii, 12 ochropectus, castaneicollis, jacksoni, camerunensis and swierstrai (not placed in any subgenus), the Scaly squamatus, griseostriatus and ahantensis (no subgenus), adspersus and capensis (without subgenus). Wolters also assigned nahani to genus Pternistis, placing it in subgenus Acentrortyx following Chapin (1926). Also, placed in the genus Pternistis were the Vermiculated natalensis, hildebrandti, harwoodi, hartlaubi, bicalcaratus, icterorhynchus and clappertoni, all within the subgenus Chaetopus. 3. Scleroptila -- Hall’s Red-winged afra (= africanus), finschi, shelleyi, psilolaema, levaillantii and levaillantoides. 4. Dendroperdix -- Hall’s Striated sephaena and streptophora. 5. Peliperdix -- Hall’s Red-tailed albogularis, schlegelii, coqui and the unplaced lathami. 6. Ortygornis -- this genus was assigned to the unplaced pondicerianus. Crowe and Crowe (1985) questioned, but did not refute, the monophyletic status of the genus Francolinus as defined in Hall (1963). However, they retained this genus for all the species and suggested a system of subgenera as follows: subgenus ‘Francolinus’ included the Spotted species francolinus, pictus, pintadeanus; ‘Ortygornis’ was assigned to the unplaced pondicerianus; ‘Scleroptila’ for all the Redwinged species finschi, levaillantii, afra, psilolaema, shelleyi, and levaillantoides that also incorporated one species from Hall’s Striated Group streptophora; ‘Peliperdix’ for the unplaced lathami; ‘Dendroperdix’ for the Striated sephaena; ‘Pternistis’ for the Bare-throated afer, swainsonii, leucoscepus and rufopictus; and the subgenus ‘Acentrortyx’ for nahani. Within the subgenus Peliperdix, they assigned only lathami. 13 No subgenera were assigned to coqui, albogularis, schlegelii, gularis, griseostriatus, ahantensis, squamatus, adspersus, hildebrandti, natalensis, hartlaubi, capensis, castaneicollis, jacksoni, erckelii, ochropectus, nobilis, camerunensis, swierstrai, harwoodi, icterorhynchus, clappertoni and bicalcaratus. Crowe and Crowe (1985) also modified Hall’s (1963) classification, by suggesting that F. streptophora should be moved from the Striated to the Red-winged Group based on the cladistic analysis of organismal characters. They further maintained that the two Montane Group species F. camerunensis and F. swierstrai form a second superspecies (= a set of species with ranges that do not overlap) within the Montane Group. However, they supported Hall’s suspicion that lathami has affinities with members of the Red-tailed Group whereas nahani was thought to have affinities with members of the Scaly Group. Milstein and Wolff (1987) adopted Wolters’ (1975-82) classification system in their work on the southern African francolins; they formally recognized the genus ‘Pternistis’ which included the Bare-throated species swainsonii and afer, and the Vermiculated adspersus, capensis and natalensis. They assigned the genus ‘Scleroptila’ to the Red-winged afra, shelleyi, levaillantoides, and levaillantii whereas ‘Peliperdix’ and ‘Dendroperdix’ were assigned to a Red-tailed species coqui and the Striated species sephaena, respectively. Despite all this taxonomic debate, Crowe et al. (1986) and Sibley and Monroe (1990) recognized only one genus ‘Francolinus’ with 41 species within it. They 14 recognized subspecies francolinus, arabistanicus, henrici, asiae, melanonotus within F. francolinus; pallidus, pictus, watsoni within F. pictus; phayrei, pintadeanus within F. pintadeanus; pondicerianus, mecranensis, interpositus within F. pondicerianus. Francolinus gularis was recognized as a monotypic species. Francolinus lathami consisted of two subspecies, lathami and schubotzi. Four subspecies were recognized within F. coqui (spinetorum, maharao, hubbardi, coqui) with thikae being included in maharao; buckleyi in spinetorum; angolensis, campbelli, ruahdae, lynesi, vernayi, hoeschianus and kasaicus were all included in the nominate coqui. Three subspecies albogularis, buckleyi and dewittei were recognized within F. albogularis. Francolinus schlegelii, F. streptophora, F. afra, F. finschi and F. nahani were recognized as monotypic species. Two subspecies, kikuyuensis and levaillantii were recognized within F. levaillantii; subspecies shelleyi (which included uluensis, canidorsalis, sequestris) and whytei were recognized within F. shelleyi; psilolaema (which included ellenbecki) and elgonensis (included theresae) within F. psilolaema; gutturalis, lorti (which included archeri), jugularis (which included cunenensis), levaillantoides (which included pallidior, kalaharica) within F. levaillantoides; sephaena, rovuma, spilogaster, zambesiae and grantii were recognized within F. sephaena. Francolinus squamatus, F. ahantensis, F. griseostriatus, F. hartlaubi, F, icterorhynchus, F. clappertoni, F. harwoodi, F. adspersus, F. capensis, F. natalensis, F. hildebrandti, F. leucoscepus, F. rufopictus were all recognized as monotypic species. The two subspecies ayesha and bicalcaratus were recognized for F. bicalcaratus. Francolinus swierstrai, F. camerunensis, F. nobilis, F. jacksoni, F. ochropectus, F. erckelii were recognized as monotypic species. Two subspecies were recognized within F. castaneicollis, castaneicollis (which included bottegi, ogoensis, kaffanus, 15 gofanus) and atrifrons; lundazi and swainsonii (included damarensis, chobiensis, gilli) within F. swainsonii; harterti, afer, cranchii (intercedens), leucoparaeus, melanogaster (included loangwae), swynnertoni and castaneiventer (included notatus, lehmanni) were recognized within F. afer. Crowe et al. (1992) again challenged the status of monophyly of Francolinus and proposed a system of genera and subgenera. Genus ‘Francolinus’ included the Spotted pictus, francolinus, pintadeanus and were all placed in the subgenus Francolinus; the unplaced pondicerianus was then placed in the subgenus Ortygornis and gularis in the subgenus Limnocolinus (nov.). Genus ‘Peliperdix’ included the Redtailed coqui, albogularis and schlegelii, and the unplaced lathami, were all placed in the subgenus Peliperdix; the Striated species sephaena was assigned to subgenus Dendroperdix, but placed in genus Peliperdix. Genus ‘Scleroptila’ included the Striated species streptophora and the Red-winged finschi, levaillantii, afra, psilolaema, shelleyi, and levaillantoides. The genus ‘Pternistis’ included nahani which was placed in the subgenus Acentrortyx; the Vermiculated hartlaubi was placed in subgenus Chapinortyx; the other Vermiculated species such as bicalcaratus, icterorhynchus, clappertoni and harwoodi were placed in subgenus Chaetopus; the balance of the Vermiculated Group, adspersus, capensis, natalensis, hildebrandti were placed in subgenus Notocolinus (nov.); the Scaly squamatus, ahantensis, and griseostriatus were placed in subgenus Squamatocolinus; the Bare-throated leucoscepus, rufopictus, afer, swainsonii in subgenus Pternistis; the Montane jacksoni, nobilis, camerunensis, swierstrai, castaneicollis, erckelii, ochropectus were all placed in subgenus Oreocolinus (nov.). 16 Despite the newly proposed taxonomic classification, del Hoyo et al. (1994) followed Hall’s (1963) classification system and adopted the genus Francolinus (sensu lato) for all the species, but did not comment on its monophyly. del Hoyo’s species and subspecies delineations were similar to those in Crowe et al. (1986) with the exception that no subspecies were attributed to castaneiventer. Dickinson (2003) did not deviate much from Crowe et al. (1986) with regard to genera, species and subspecies that were recognized. He attributed 41 species to the genus genus Francolinus. Within F. francolinus, subspecies francolinus (which included billypayni), arabistanicus, bogdanovi (included festinus), henrici, asiae (including parkerae), melanonotus were recognized; pallidus, pictus, watsoni were delineated within F. pictus; phayrei, pintadeanus within F. pintadeanus; pondicerianus (included ceylonensis), mecranensis, interpositus (included prepositus) within F. pondicerianus. Francolinus gularis (which included ridibundus) and F. lathami were recognized as monotypic species. Four subspecies were recognized within F. coqui (spinetorum, maharao, hubbardi, coqui) with thikae being included in maharao; angolensis, campbelli, ruahdae, lynesi, vernayi, bourqoii, hoeschianus and kasaicus were all included in the nominate coqui. The taxa albogularis, buckleyi, dewittei (included meinertzhageni) were recognized within F. albogularis; F. schlegelii (included confusus); F. streptophora; F. afra (which included proximus); F. finsch; F. griseostriatus and F. nahani were recognized as monotypic species. Three subspecies kikuyuensis (which included benguellensis, clayi), crawshayi, and levaillantii were recognized within F. levaillantii; subspecies shelleyi (which included uluensis, canidorsalis, sequestris) and whytei within F. shelleyi; psilolaema 17 (which included ellenbecki) and elgonensis (included theresae) within F. psilolaema; gutturalis (included eritreae), archeri (included friedmanni, stantoni), lorti, jugularis (included cunenensis, stresemanni), pallidior (included wattii) and levaillantoides (included kalaharicus) within F. levaillantoides; sephaena (included zuluensis), rovuma, spilogaster (included somaliensis), zambesiae (included thompsoni) and grantii were recognized within F. sephaena; squamatus (included whytei), schuetti (included tetraoninus), zappeyi, maranensis (included chyuluensis), usambarae, uzungwensis and doni within F. squamatus; hopkinsoni and ahantensis were recognized within F. ahantensis; F. hartlaubi was represented by three subspecies (hartlaubi, bradfieldi, crypticus); altumi, hildebrandti (included helleri, fischeri), johnstoni (included grotei) within F. hildebrandti; natalensis (included thamnobium) and neavei within F. natalensis; ayesha, bicalcaratus (which included adamauae), thornei and ogilviegranti (included molunduensis) within F. bicalcaratus; clappertoni, sharpii, heuglini, konigseggi, gedgii (included cavei) and nigrosquamatus (included testis) within F. clappertoni; F. harwoodi; F. capensis; F. swierstrai; F. camerunensi; F. nobilis (included chapini); F. jacksoni (included pollenorum); F. ochropectus; F. erckelii; F. rufopictus and F. leucoscepus were recognized as monotypic species. Two subspecies were recognized within F. adspersus, adspersus (included kalahari) and mesicus; castaneicollis (included bottegi, gofanus), ogoensis, kaffanus (included patrizii) and atrifrons within F. castaneicollis; lundazi and swainsonii (included damarensis, gilli) within F. swainsonii; harterti, afer (included chio, palliditectus), cranchii (boehmi, punctulatus, benguellensis, intercedens, nyanzae, itigi), leucoparaeus, melanogaster (included loangwae, aylwinae, tertius) swynnertoni and castaneiventer (included notatus) were recognized within F. afer. 18 Phylogenetics Hall (1963) presented the first phylogenetic hypothesis for francolins (Fig. 1.1). Bloomer and Crowe (1998) provided the first DNA-based evidence (Fig. 1.2) to suggest a lack of monophyly for Francolinus as circumscribed by Hall, but offered no taxonomic recommendations. However, their molecular phylogeny provided support for splitting the francolins into what is today referred as two distinct clades, the ‘francolins’ and ‘spurfowls’. Highlights of some key taxonomic disagreements:  It is clear that taxonomic confusion exists, not just among taxonomists, but also between taxonomists and gamebird enthusiasts and hunters. Mackworth-Praed and Grant (1952, 1962, 1970) in all their reviews, split francolins into what they called ‘francolins’ and ‘spurfowls’. They assigned the genus Pternistis to the spurfowl taxa belonging to Hall’s (1963) Bare-throated Group, and the genus Francolinus to all African francolins including the balance of the spurfowl groups identified in Hall’s monograph (Table 1.1). The results of Bloomer and Crowe (1998) led to the split of francolins into two major lineages (Fig. 1.2) which they called the quail-francolin (or partridges)’ and partridge-francolin (or pheasants) clades. The common names francolin and spurfowl will be used hereafter, since Bloomer and Crowe (1998) found them to be evolutionarily distantly-related from each other.  The evolutionary affinities of the Crested Francolin Ortygornis sephaena have been debated given its possession of features that are found in both francolins 19 and spurfowls. Hall (1963) considered this species to be a 'linking form' and as a consequence placed all 41 species into a single genus Francolinus.  Confusion has surrounded the number of genera recognized among francolins and spurfowls. The earlier revisions (Roberts 1924, 1940) included seven genera (Pternistis, Chaetopus, Peliperdix, Chapinortyx, Dendroperdix, Scleroptila, Ortygornis), which were reduced to either two (Francolinus and Pternistis) in Chapin (1932), Peters (1934), Mackworth-Praed and Grant (1952), McLachlan and Liversidge (1957), Mackworth-Praed and Grant (1962), Mackworth-Praed and Grant (1970), McLachlan and Liversidge (1970) or just one genus Francolinus (Crowe and Crowe 1985, Maclean 1985, Crowe et al. 1986, Maclean 1993) in several later revisions (Table 1.3). More recently, Crowe et al. (1992) suggested that some earlier genera were indeed valid (Francolinus, Dendroperdix, Peliperdix, Scleroptila and Pternistis). It should be noted that based on Crowe et al. (2006) and the results emanating from this study, the number of genera recognized within ‘francolins’ might see an increase, but are expected to remain stable beyond the completion of this study.  The lack of stability and consistency in the subspecies recognized and several revisions of the species complexes, even when similar evidence and geographic areas were covered, has clearly demonstrated the profound difficulties with respect to using morphological characters to delineate taxa.  Francolinus nahani has remained something of an enigma among francolins/spurfowl with a great deal of uncertainty pertaining to its phylogenetic placement. Van Someren (1926) proposed the genus Peliperdix for 20 this species, and Chapin (1926) had during the same year proposed to move nahani to the genus Acentrortyx. Chapin argued that this partridge was not congeneric with the type species of the genus of francolins Francolinus francolinus. However, subsequent revisions placed this species back in the genus Francolinus (Peters 1934, Hall 1963, Mackworth-Praed and Grant 1970). Wolters (1975-82) and Crowe et al. (1992) placed nahani within the genus Pternistis in subgenus Acentrortyx. Most recently nahani has been moved from Francolinus to ‘Ptilopachus’, a clade of two taxa, the other being the Stone Partridge Ptilopachus petrosus which intriguingly forms the sister clade to New World quails Odontophoridae (Crowe et al. 2006, Cohen et al. 2012). The genus Francolinus - African or Asian in origin? The geographical origin of the genus ‘Francolinus’ has remained controversial (Hall 1963, Crowe and Crowe 1985). Two hypotheses are postulated. The first based on the suggestion that the genus Francolinus shares its closest affinities with other Palaearctic and Asian genera. Hall (1963) strongly argued for the genus to be of Asian origin with its age being traced back to the Oligocene +-25-35. She further alluded to reduced competition as being the driving factor for the speciation of francolins in Africa. Although Crowe and Crowe (1985) concurred with Hall that the ancestor was quail-like, they hypothesized an African origin for the genus, with Asia being colonized by a nomadic or migratory ancestor that diversified and became sedentary. The reported age of the divergence francolins from Old World quail Coturnix spp. was traced to the Pliocene (approx. 4.5-5 mya) using Shields and Wilson’s (1987) calibration of 2% divergence per million years (Crowe 1992). Bloomer and Crowe (1998) recovered a 21 divergence date between the quail versus the partridge-like lineages at 5 and 8 mya, whereas the transversion rate from suggested a divergence time of 3-6 mya using the using Shields and Wilson’s (1987) conventional calibration of 2% per million years implemented on mitochondrial Cytochrome-b data. They believe that the minimum divergence estimates appear to be corroborated by the availability of well-differentiated francolin fossils (Crested Francolin Ortygornis sephaena; Crowe 1992). However, as hinted at (although not actually calculated) in Crowe et al. (2006), the divergence of francolins and spurfowl may be considerably older. Conclusions The taxonomic review detailed above demonstrates that the confusion over the number of genera and terminal taxa of francolins and spurfowls is profound and that what is needed are characters from a range of sources that can provide solid and objective explanations for every taxonomic decision that is taken. This is especially important given the enormous morphological variation observed within certain species complexes, which has generated considerable debate with respect to the delimitation of species and subspecies. Defensible and robust analytical methods and approaches also need to be adopted to enable objective conclusions to be drawn. The major aims of this study On the basis of the confusion/disagreement/hypotheses outlined above regarding: (1) debate on the monophyly of the genus ‘Francolinus’, (2) the monophyly and evolutionary relationships of, and within the different species-groups suggested by Hall (1963), (3) the instability/inconsistency of the number of terminal taxa (species, 22 subspecies and genera) recognized, and (4) the dispute on where the Francolinus clade first diversified – Africa or Asia - this study aimed at achieving the following: 1. To undertake a general review of the taxonomy of ‘francolins’ (taking views largely – but not exclusively – of Mackworth-Praed and Grant (1952, 1962, 1970) and Hall 1963 as alternative hypotheses) [Chapter 1]. 2. To contrast alternative hypotheses concerning the monophyly of the genus ‘Francolinus’ (e.g. Hall 1963 versus Bloomer and Crowe 1998) and to further explore DNA-based, vocalization and behavioural evidence in order to test the phylogenetic affinities of the Stone Partridge Ptilopachus petrosus and Nahan’s Francolin Francolinus nahani as suggested in Crowe et al. (2006) [Chapter 2]. 3. To investigate if there are any syringeal features (anatomical) that can be used to further our understanding of the phylogenetic relationships among francolins and spurfowls [Chapter 3]. 4. To test the validity of the francolin-spurfowl taxonomic dichotomy based on vocal characters [Chapter 4]. 5. To test Hall’s (1963) hypotheses on the monophyly of the suggested eight species groups within her circumscription of the the genus Francolinus, and thereby provide a modern systematic revision of the terminal taxa and genera of francolins and spurfowls [Chapter 5 & 6]. 6. To describe and explore patterns of distribution of francolins and spurfowls, using Hall’s (1963) monograph as a null hypothesis [Chapter 7]. Tables and Figures 23 Table 1.1. List of francolin and spurfowl species and their designated species groups recognized by Hall (1963). English and specific names are as proposed in the IOC list (Gill and Donsker 2013) and the generic classification follows Crowe et al. (2006) and Mandiwana-Neudani, this thesis. Genus Francolinus Stephens, 1819, Peliperdix Bonaparte, 1856, Scleroptila Blyth, 1849, Dendroperdix Roberts, 1922, Ortygornis Reichenbach, 1852, Pternistis Wagler, 1832, Chaetopus Swainson, 1837, Afrocolinus gen. nov. Mandiwana-Neudani, this thesis. Taxon authority citation with a year in parentheses indicates that either the specific or the subspecific epithet was originally published in another genus by the first author, but moved to the present genus by the second revising author. Species group English name Scientific name 1. Spotted Black Francolin Painted Francolin Chinese Francolin Francolinus francolinus (Linnaeus, 1766) F. pictus (Jardine & Selby, 1828) F. pintadeanus (Scopoli, 1786) 2. Bare-throated Red-necked Spurfowl Swainson’s Spurfowl Yellow-necked Spurfowl Grey-breasted Spurfowl Pternistis afer (Müller, 1776) P. swainsonii (Smith, 1836) P. leucoscepus (Gray, 1867) P. rufopictus Reichenow, 1887 3. Montane Erckel’s Spurfowl Djibouti Spurfowl Chestnut-naped Spurfowl Jackson’s Spurfowl Handsome Spurfowl Mount Cameroon Spurfowl Swierstra’s Spurfowl P. erckelii (Rüppell, 1835) P. ochropectus (Dorst & Jouanin, 1952) P. castaneicollis Salvadori, 1888 P. jacksoni O. Grant, 1891 P. nobilis Reichenow, 1908 P. camerunensis Alexander, 1909 P. swierstrai (Roberts, 1929) 4. Scaly Ahanta Spurfowl Scaly Spurfowl Grey-striped Spurfowl P. ahantensis Temminck, 1854 P. squamatus Cassin, 1857 P. griseostriatus O. Grant, 1890 5. Vermiculated Double-spurred Spurfowl Heuglin’s Spurfowl Clapperton’s Spurfowl P. bicalcaratus (Linnaeus, 1766) P. icterorhynchus Heuglin, 1863 P. clappertoni (Children & Vigors, 1826) P. hildebrandti Cabanis, 1878 P. natalensis Smith, 1833 Hildebrandt’s Spurfowl Natal Spurfowl 24 Species group English name Scientific name Hartlaub’s Spurfowl Harwood’s Spurfowl Red-billed Spurfowl Cape Spurfowl P. hartlaubi Bocage, 1869 P. harwoodi Blundell & Lovat, 1899 P. adspersus Waterhouse, 1838 P. capensis (Gmelin, 1789) 6. Striated Crested Francolin Ring-necked Francolin Ortygornis sephaena (Smith, 1836) Scleroptila streptophora O. Grant, 1891 7. Red-winged Shelley’s Francolin Grey-winged Francolin Orange River Francolin Red-winged Francolin Finsch’s Francolin Moorland Francolin Scleroptila shelleyi O. Grant, 1890 S. afra (Latham, 1790) S. levaillantoides Smith, 1836 S. levaillantii (Valenciennes, 1825) S. finschi Bocage, 1881 S. psilolaema Gray, 1867 8. Red-tailed Coqui Francolin White-throated Francolin Schlegel’s Francolin Peliperdix coqui (Smith, 1836) P. albogularis Hartlaub, 1854 P. schlegelii Heuglin, 1863 Latham’s Francolin Swamp Francolin Grey Francolin Nahan’s Francolin Afrocolinus lathami Hartlaub, 1854 Ortygornis gularis (Temminck, 1815) O. pondicerianus (Gmelin, 1789) Francolinus ‘Ptilopachus’ nahani (Dubois, 1905) Unplaced species 25 Table 1.2. Summary of the taxonomic designations from selected revisions pertaining to francolins, detailed according to putative super-species group, genera, species and subspecies. Fran - Genus Francolinus Stephens, 1819, Peli - Peliperdix Bonaparte, 1856, Scler Scleroptila Blyth, 1849, Dend - Dendroperdix Roberts, 1922, Orty - Ortygornis Reichenbach, 1852, Pter - Pternistis Wagler 1832, Chae - Chaetopus Swainson, 1837, Afrocolinus gen. nov. (this thesis). n/a denotes taxa not investigated (i.e. either falling outside the geographic region covered or taxa not being of interest), + denotes taxa recognized, - denotes taxa not recognized, ~ denotes taxa lumped/synonymized with others. Putative genera - ~8, Putative species - ~41, Putative subspecies - ~164. Taxon authority citation with a year in parentheses indicates that either the specific or the subspecific epithet was originally published in another genus by the first author, but moved to the present genus by the second revising author. Putative species and subspecies Roberts (1924) Hockey et al. (2005) Peters (1934) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Spotted Group n/a n/a Fran n/a Fran Fran Fran Species francolinus subspecies francolinus (Linnaeus, 1766) arabistanicus Zarudny & Härms, 1913 bogdanovi Zarudny, 1906 henrici Bonaparte, 1856 asiae Bonaparte, 1856 melanotus Hume, 1888 n/a n/a + n/a + + + n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + + + + + n/a n/a n/a n/a n/a + + + + + + n/a n/a n/a n/a n/a pictus pictus (Jardine & Selby, 1828) pallidus (Gray, 1831) watsoni Legge, 1880 n/a n/a n/a n/a n/a n/a + + - n/a n/a n/a + + + + - + n/a n/a pintadeanus pintadeanus (Scopoli, 1786) phayrei (Blyth, 1843) n/a n/a n/a n/a + + n/a n/a + + + + + n/a 26 Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Striated Group Dend Dend Fran Fran Fran Dend Orty/Scler sephaena sephaena (Smith, 1836) spilogaster Salvadori, 1888 somaliensis Grant & Praed, 1934 schoanus Heuglin, 1873 jubaensis Zedlitz, 1913 grantii Hartlaub, 1866 rovuma Gray, 1867 zambesiae Praed, 1920 chobiensis Roberts, 1932 thompsoni Roberts, 1924 zuluensis Roberts, 1924 mababiensis Roberts, 1932 + n/a n/a n/a n/a n/a + + + + + - + n/a n/a n/a n/a n/a + + ~ ~ ~ ~ + + + + + + + + ~ ~ ~ + + + + + - + + + + + - + + + + - + + ~ ~ ~ + + ~ ~ ~ ~ ~ streptophora streptophora O. Grant, 1891 n/a n/a + + + + + Orty Peli Fran Fran Fran Peli Peli coqui coqui (Smith, 1836) spinetorum Bates, 1928 buckleyi Peters, 1934 maharao Sclater, 1927 ruahdae Someren, 1926 hubbardi O. Grant, 1895 angolensis Rothschild, 1902 lynesi Sclater, 1932 vernayi (Roberts, 1932) campbelli (Roberts, 1928) thikae Grant & Praed, 1934 kasaicus White, 1945 hoeschianus Stresemann, 1937 stuhlmanni Reichenow, 1889 + n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + n/a n/a n/a ~ ~ ~ ~ ~ n/a ~ ~ ~ + + + + + + + + + + ~ + + + + + + + + + + + - + + + + + + + + + + - + + - + + ~ + + + ~ ~ + ~ ~ + ~ + schlegelii schlegelii Heuglin, 1863 n/a n/a + + + + + albogularis albogularis Hartlaub, 1854 buckleyi O. Grant, 1892 dewittei Chapin, 1937 meinertzhageni White, 1944 gambagae (Praed, 1920) n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + + + + + + - + + + + - + + + + - + + + ~ ~ Red-tailed Group 27 Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Red-winged Group Scler Scler Fran psilolaema psilolaema Gray, 1867 ellenbecki Erlanger, 1905 elgonensis O. Grant, 1891 theresae Meinertzhagen, 1937 n/a n/a n/a n/a n/a n/a n/a n/a + + + - shelleyi shelleyi O. Grant, 1890 uluensis O. Grant, 1892 whytei Neumann, 1908 macarthuri Someren, 1938 trothae Reichenow, 1901 sequestris Clancey, 1960 canidorsalis (Lawson, 1963) + n/a n/a n/a n/a - + n/a n/a n/a ~ ~ afra afra (Latham, 1790) + Fran Fran Scler Scler + + + + + ~ ~ ~ + + + + + + ~ + + + + + - + + + + - + + + ~ - + - + + + ~ ~ ~ ~ + + + + + + levaillantoides levaillantoides Smith, 1836 kalaharica Roberts, 1932 pallidior Neumann, 1908 langi Roberts, 1932 wattii Macdonald, 1953 jugularis Büttikorfer, 1889 cunenensis Roberts, 1932 stresemanni Hoesch & Niethammer, 1940 gutturalis (Rüppell, 1835) lorti Sharpe, 1897 archeri Sclater, 1927 ludwigi Neumann, 1920 + n/a n/a n/a n/a n/a n/a + ~ + ~ ~ + ~ + ~ + + + ~ + + + + + + + + + + + - + + - + ~ + ~ ~ + ~ n/a n/a n/a n/a + ~ n/a n/a n/a ~ + + + + + + + - + + + - + + + - ~ + ~ ~ ~ levaillantii levaillantii (Valenciennes, 1825) kikuyuensis O. Grant, 1897 crawshayi O. Grant, 1896 benguellensis Neumann, 1908 clayi White, 1944 + n/a n/a n/a n/a + n/a - + + + + - + + + + + + + + - + - + + + ~ ~ finschi finschi Bocage, 1881 n/a n/a + + + + + 28 Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Vermiculated Group Chae/Peli Pter Fran Fran Fran Pter Pter hartlaubi hartlaubi Bocage, 1869 crypticus Stresemann, 1939 bradfieldi (Roberts, 1928) ovambensis (Roberts, 1928) + n/a n/a n/a + - + + + + + + - + ~ ~ - + - + ~ ~ ~ adspersus adspersus Waterhouse, 1838 kalahari de Schauensee, 1931 mesicus Clancey, 1996 + - + ~ + + + - + - + - + - + ~ ~ capensis capensis (Gmelin, 1789) + + + + + + + natalensis natalensis Smith, 1833 neavei Praed, 1920 + - + - + + + + + + + - + ~ hildebrandti hildebrandti Cabanis, 1878 altumi Fischer & Reichenow, 1884 helleri Mearns, 1915 fischeri Reichenow, 1887 johnstoni Shelley, 1894 grotei Reichenow, 1919 n/a n/a + + + + + n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + + + + + + + - + + - + + - ~ ~ + ~ ~ bicalcaratus bicalcaratus (Linnaeus, 1766) ogilvie-grantii Bannerman, 1922 ayesha Hartert, 1917 adamauae Neumann, 1915 thornei Ogilvie-Grant, 1902 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + + + + + + + + + + + + + + + - + ~ ~ + ~ icterorhynchus icterorhynchus Heuglin, 1863 dybowski Oustalet, 1892 ugandensis Neumann, 1907 emini Neumann, 1907 n/a n/a n/a n/a n/a n/a n/a n/a + + + ~ + + + + - + + - + ~ ~ ~ clappertoni clappertoni (Children & Vigors, 1826) sharpii Ogilvie-Grant, 1892 n/a n/a n/a n/a + + + + + + + + + + 29 Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Vermiculated Group Chae/Peli heuglini Neumann, 1907 gedgii Ogilvie-Grant, 1891 nigrosquamatus Neumann, 1902 konigseggi Madarasz, 1914 testis Neumann, 1928 cavei Macdonald, 1940 n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a harwoodi harwoodi Blundell & Lovat, 1899 n/a Fran Fran Pter Pter + + + ~ + - + + + + + + + + + - - ~ ~ ~ ~ ~ ~ n/a + + + + + n/a n/a Fran Fran Fran Pter Pter erckelii erckelii (Rüppell, 1835) pentoni Praed,1920 n/a n/a n/a n/a + + + + + + + - + ~ nobilis nobilis Reichenow, 1908 chapini Grant & Praed, 1934 n/a n/a n/a n/a + - + + + + + - + ~ camerunensis camerunensis Alexander, 1909 n/a n/a + + + + + swierstrai swierstrai (Roberts, 1929) n/a n/a + + + + + castaneicollis castaneicollis Salvadori, 1888 bottegi Salvadori, 1898 gofanus Neumann, 1904 ogoensis Praed, 1920 kaffanus Grant & Praed 1934 atrifrons (Conover, 1930) n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a + + + + + + + + + + + + + + + + + ~ ~ ~ ~ ochropectus ochropectus (Dorst & Jouanin, 1952) n/a n/a - - + + + jacksoni jacksoni O. Grant, 1891 n/a pollenorum Meinertzhagen, 1937n/a gurae Bowen, 1931 n/a n/a n/a n/a + + + + - + + - + - + ~ ~ Montane Group Pter Fran 30 + Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Scaly Group Fran Fran Pter Pter + + + + + + + + + + + + + + + + ~ + ~ n/a n/a n/a + + + + - + + - - ~ ~ ~ n/a n/a n/a n/a + - + - - ~ ~ ahantensis ahantensis Temminck, 1854 hopkinsoni Bannerman, 1934 n/a n/a n/a n/a + - + + + + + - + ~ griseostriatus griseostriatus O. Grant, 1890 n/a n/a + + + + + Pter Pter Pter Pter Fran Pter Pter + n/a n/a n/a n/a n/a n/a n/a n/a + + + + + + + + + + + + + + + + + + + + + - + + - + ~ ~ ~ ~ ~ + ~ ~ n/a ~ + + + + + + + + + + - ~ + ~ + n/a n/a ~ ~ ~ ~ ~ + ~ ~ ~ ~ - + + + + + + + - + + + + + - + + + - ~ ~ ~ ~ ~ ~ ~ ~ squamatus squamatus Cassin, 1857 maranensis Mearns, 1910 schuetti Cabanis, 1880 usambarae Conover, 1928 uzungwensis Bangs & Loveridge, 1931 doni Benson, 1939 zappeyi Mearns, 1911 tetraoninus Blundell & Lovat, 1899 chyuluensis Someren, 1939 Bare-throated Group n/a n/a Fran n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a afer afer (Müller, 1776) + harterti Reichenow, 1909 n/a nyanzae Conover, 1929 n/a böhmi Reichenow, 1885 n/a intercedens Reichenow, 1909 n/a itigi (Bowen, 1930) n/a cranchii (Leach & Koenig, 1818) n/a punctulatus (Gray, 1834) n/a benguellensis Bocage, 1893 n/a leucoparaeus (Fischer & Reichenow, 1884) n/a humboldtii (Peters, 1854 + swynnertoni Sclater, 1921 n/a castaneiventer Gunning & Roberts, 1911 + melanogaster Neumann, 1898 n/a loangwae Grant & Praed, 1934 n/a lehmanni Roberts, 1931 notatus Roberts, 1924 + krebsi Neumann, 1920 + cunenensis Roberts, 1932 n/a cooperi Roberts, 1947 - 31 Putative species and subspecies Roberts (1924) Hockey Peters et al. (1934) (2005) MackHall Wolters worth(1963) (1975Praed 82) & Grant (1952,62,70) MandiwanaNeudani (this study) Genera Bare-throated Group Pter Pter swainsonii swainsonii (Smith, 1836) lundazi White, 1947 gilli Roberts, 1932 damarensis Roberts, 1932 chobiensis Roberts, 1932 + n/a n/a n/a n/a + + ~ ~ ~ rufopictus rufopictus Reichenow, 1887 n/a leucoscepus leucoscepus (Gray, 1867) infuscatus Cabanis, 1868 holtemülleri Erlanger, 1904 keniensis Mearns, 1911 kilimensis Mearns, 1911 tokora Stoneham, 1930 muh.-ben-abdul. Erlanger, 1904 n/a n/a n/a n/a n/a n/a n/a Pter Pter Fran Pter Pter + ~ ~ ~ + + + + + ~ ~ ~ - + - + ~ ~ ~ ~ n/a + + + + + n/a n/a n/a n/a n/a n/a n/a + + + + ~ + + + + + + + ~ - + - + + ~ ~ ~ ~ ~ Unplaced taxa pondicerianus Hall (1963:167) speculates affinities with sephaena and/or coqui pondicerianus (Gmelin, 1789) mecranensis Zarudny & Härms, 1913 interpositus Hartert, 1917 ceylonensis Whistler, 1941 n/a n/a + n/a + + + n/a n/a n/a n/a n/a n/a + + - n/a n/a n/a + + + - n/a n/a n/a gularis Hall (1963:167-168) “comparable in size and general proportions to the largest member of F. francolinus” isolated from other francolins” …. “divergence over a long period” gularis (Temminck, 1815) n/a n/a + n/a + + + nahani Hall (1963:166-167) speculates affinities with squamatus and other ‘Scaly’ francolins nahani (Dubois, 1905) n/a n/a + + + + - lathami Hall (1963:165-166) speculates affinities with coqui and other ‘Red-tailed’ francolins lathami Hartlaub, 1854 schubotzi Reichenow, 1912 n/a n/a n/a n/a + + + + + + + - A new genus Afrocolinus gen. nov. is recognized for lathami Hartlaub, 1854 in this study. 32 + + Table 1.3. Summary of the number of putative genera, species and subspecies outlined in Table 1.2. Taxonomic category Author/s Roberts (1924) Peters (1934) Mackworth-Praed & Grant (1952, 62, 70) Hall (1963) Wolters (1975-82 Crowe et al. (1986) del Hoyo et al. (1994) Dickinson (2003) Hockey et al. (2005) Mandiwana-Neudani (this thesis) Geographic area covered Putative genera Putative species Putative subspecies South Africa Asia & Africa 6 2 11 39 20 119 Africa Asia & Africa Asia & Africa Asia & Africa Asia & Africa Asia & Africa Southern Africa Asia & Africa 2 1 5 1 1 1 4 8 34 41 41 41 41 41 11 40 113 101 56 50 50 85 17 54 33 Table 1.4. Summary of characters used by authors to justify their positions on the classification of francolins. Authors Characters used Roberts (1924)/Hockey et al. (2005) Plumage, calls, size, bare skin on throat, distribution range†, breeding information Chapin (1932) Habitat type, distribution range†, morphology Peters (1934) Breeding information, distribution range† Mackworth-Praed & Grant (1952, 1962, 1970) Morphology, geographic distribution†, habits, nest and eggs, breeding information, calls Hall (1963) Plumage, size, sexual dimorphism, extent of bare skin, colour of bill and legs, spurs, field habits, vocalizations, chicks, eggs Wolters (1975-82) Distribution range† Crowe and Crowe (1985) Skeletal and integument characters, vocalizations, ethological and ecological information Crowe et al. (1986) Morphology, field characters, voice, habitat and food preferences, breeding habits Milstein and Wolff (1987) Calls of adults and chicks, incubation periods, spur development, natal down, hybridization, general behaviour (whether a particular species is a squatter or runner) Sibley and Monroe (1990) Habitat preferences, hybridization information, distribution range† Crowe et al. (1992) Mitochondrial DNA Restriction Fragment Length Polymorphisms, morpho-behavioural characters del Hoyo et al. (1994) Feeding habit, habitat preferences, breeding, distribution range†, morphology Bloomer and Crowe (1998) Mitochondrial Cytochrome b sequence characters, morphobehavioural characters Dickinson (2003) Primarily results from published molecular DNA phylogenies, also follow Peters (1934) † indicates that the distribution ranges of taxa were not presented as point localities instead as roughly defined ranges. 34 Figure 1.1. The re-drawn hypothetical cladogram of francolins according to Hall (1963). The partridge-quail francolin dichotomy follows Milstein and Wolff (1987). The acronyms used on the tree abbreviate the following: SPG stands for Spotted Group, BTG - Bare-throated Group, MTG - Montane Group, SCG - Scaly Group, VMG - Vermiculated Group, RWG - Redwinged Group, STG - Striated Group and RTG abbreviates Red-tailed Group. AS stands for Asia and AF - Africa. 35 Figure 1.2. The francolin phylogeny based on combined analysis of 200 Cytochrome-b and 25 morphobehavioral characters (adopted from Bloomer and Crowe 1998). Bootstrap support values are indicated at nodes. 36 CHAPTER 2 Phylogenetics of evolutionarily enigmatic terrestrial gamebirds (Aves: Galliformes) with special regard to ‘Francolinus’ spp. Part of the information presented in this chapter is derived from Crowe et al. (2006) and Cohen et al. (2012) (both of which I am a co-author), where I contributed in a primary capacity with regard to content and molecular DNA sequence data on francolins (Francolinus, Ortygornis, Afrocolinus, Peliperdix and Scleroptila spp.) and spurfowls (Pternistis). I was also involved at all levels with respect to the preparation of these manuscripts. Abstract The ultimate goal of the cladistic analysis of any putative taxonomic group is to demonstrate (or refute) its monophyly decisively. If the former proves to be the case, the next task is to identify well-resolved and supported monophyletic assemblages within it. Until Crowe et al. (2006), these goals had not been achieved for the terrestrial gamebirds (Aves: Galliformes), a geographically widespread assemblage that harbours some of the most economically important bird species and a spectrum of biologically and geographically diverse taxa that have been used as models for studying broad-scale physiological, ecological, evolutionary and biogeographical phenomena. However, some authors, e.g., Eo et al. (2009), Shen et al. (2010), still maintain that this goal 37 remains unattained. In particular, the lack of a well-resolved hypothesis concerning phylogenetic relationships within the crown Galliformes, and especially phasianine galliforms (pheasants, partridges, ‘francolins’, quails, grouse, turkeys and peafowls), hinders interpretation of their morphological and ecological evolution and their utility as biogeographical indicators needed for conservation and management initiatives. In this chapter, four mitochondrial and three nuclear markers were sequenced for a selection of galliform species (and not genera, are terminals as in previous studies). Further evidence in terms of the large number of francolin and spurfowl species analyzed, mitochondrial and nuclear DNA characters (and new evidence particularly from vocalizations and behaviour of particularly F. nahani) was explored in the context of the phylogenetic affinities suggested by Crowe et al. (2006) with the goal of producing a robust phylogenetic hypothesis for Africa's diverse Galliform lineages, and in particular francolins sensu lato. Results from this study rejects the traditional classification of galliform families, and instead supports the recognition of only five families, which are the: Megapodiidae, Cracidae, Numididae, Odontophoridae and Phasianidae. Members of the two traditionally recognized families, the Meleagrididae (turkeys) and Tetraonidae (grouse and allies), were nested within the Phasianidae. Regarding the focal taxa, Nahan’s Francolin F. nahani is not a francolin but a partridge sister to the Stone Partridge P. petrosus, and they diverged c. 9.6 mya. Their relationship is strongly supported and contradicting all other published treatments of the galliformes. The two form a basal clade relative to ‘true’ francolins and spurfowls suggesting that they represent a relictual group sister to the New World quails (Odontophoridae), and are only distantly related to the other Old World galliforms. This study also confirms earlier studies in 38 demonstrating that francolins and spurfowls are not each other’s closest relatives as suggested by Hall (1963). The genus Francolinus comprises at least two distantly related lineages (excluding F. nahani), the partridge-like spurfowls, which are related to quails (e.g. Coturnix, Excalfactoria, Margaroperdix, Perdicula and Ammoperdix spp.) and certain Old World partridges (e.g. Tetraogallus and Alectoris spp.), and the quaillike francolins (Dendroperdix, Peliperdix and Scleroptila spp.), which are related to the ‘true’ Asian francolins (Francolinus spp.), chickens (Gallus spp.), and other Old World partridges (Bambusicola spp.). The estimated divergence date between the Gallus/Bambusicola/francolins and the Coturnix/Alectoris/spurfowls clade is recovered at c. 33.6 mya. 39 Introduction Prior to 2006, there was no consensus on the phylogenetic relationships of, or within, the Galliformes, except that the Galliformes were monophyletic (a view initially challenged by Prager and Wilson 1976, Jolles et al. 1976, 1979). The sister-group relationship between Galliformes and Anseriformes (ducks, geese and screamers) in the avian tree of life is very strongly supported (Hackett et al. 2008), although relationships among the lineages within Galliformes still remain controversial (Wang et al. 2013). The number of supra-generic monophyletic assemblages within Galliformes has varied greatly. Adopting an extreme ‘splitters’ viewpoint (Verheyen 1956, Johnsgard 1973, 1986, 1988, 1999), eight groupings with associated core biogeographical affinities have been identified: Megapodiidae (Australasia - megapodes, scrubfowl, brush-turkeys), Cracidae (Neotropics - curassows, guans and chachalacas), Numididae (Afrotropics – guineafowls), Phasiani(n/d)ae (Afro/Asiotropical - pheasants, junglefowls - chickens, peafowls, and peacock- and argus-pheasants), Perdicinae (Palaearctic and Afro/Asiotropical - partridges, francolins and Old World quails), Meleagridi(n/d)ae (Nearctic – turkeys), Tetraoni(n/d)ae (Holarctic – grouse) and Odontophori(n/d)ae (Neotropical and Nearctic - New World quails). With respect to these groupings (reviewed by Johnsgard 1973, 1986, 1988, 1999 and Sibley and Ahlquist 1985, 1990), there is general agreement that the megapodes and cracids are cladistically basal (and probably sister-taxa – Wetmore 1960, Sibley and Ahlquist 1990), followed by the guineafowls (Cracraft 1981, Crowe 1988) or New World quails (Sibley and Ahlquist 1990, Kornegay et al. 1993) as sister to the remaining phasianine assemblages. Within phasianine galliforms, it is traditionally accepted that the turkeys and grouse form monophyletic clades, with Sibley and 40 Ahlquist (1990) speculating that they are each other's closest relatives. This leaves the Phasiani(n/d)ae and Perdicinae for which several authors have suggested that their constituent taxa are polyphyletic (e.g. Bloomer and Crowe 1998). The Afrotropics harbour 49 species of galliform gamebirds, occurring in virtually all habitats across the continent largely south of the Sahara (Crowe et al. 1986, del Hoyo et al. 1994). Crowe et al. (2006) demonstrated decisively (see also Milstein and Wolf 1987, Crowe et al. 1992, Bloomer and Crowe 1998), that Africa’s largest gamebird genus (currently 36 spp.), Francolinus Stephens, 1819 (sensu Hall 1963) comprises at least two distantly related African radiations. The partridge-like spurfowls (Pternistis, 24 spp.) are related to quail (e.g. Coturnix, Excalfactoria, Margaroperdix, Perdicula and Ammoperdix spp.) and certain Old World partridges (e.g. Tetraogallus and Alectoris spp.). The quail-like francolins (Dendroperdix, Peliperdix and Scleroptila, 12 spp.) are related to ‘true’ Asian francolins (Francolinus, five spp.), junglefowls (Gallus) and other Old World partridges (Bambusicola). The remaining African galliforms comprise the Old World quails (Coturnix and Excalfactoria, three spp.), the endemic guineafowls (Numididae, six spp.) and four species with putative IndoMalaysian affinities: the Congo Peafowl Afropavo congensis, the Stone Partridge Ptilopachus petrosus, and the Udzungwa and Rubeho Forest Partridges Xenoperdix udzungwensis and X. obscurata (Dinesen et al. 1994, Johnsgard 1988, Madge and McGowan 2002, Bowie and Fjeldså 2005, Crowe et al. 2006). All of these African taxa were thought to have their nearest phylogenetic relatives elsewhere in the Old World (Sibley and Ahlquist 1990). Thus, it was surprising when Crowe et al. (2006) suggested that the Stone Partridge Ptilopachus petrosus and Nahan’s Francolin Francolinus 41 nahani were sister-species, and that this clade (the Ptilopachinae, Bowie et al. 2013) was sister to the New World Quails (Odontophoridae). Based on an analysis of 158 ingroup taxa representing 65 genera and all putative suprageneric galliform taxa rooted on representatives of the Anseriformes, Crowe et al. (2006) (including myself as a junior author) offered a novel, generally well-resolved and well-supported, phylogenetic hypothesis for the Galliformes (Fig. 2.1). Characters analyzed included 102 morpho-behavioural attributes (Dyke et al. 2003) and 4452 nucleic acid base pairs (bp) from four mitochondrial markers (Cytochrome-b, ND2, 12S and control region) and a single nuclear marker, ovomucoid intron G. At the generic level, parsimony-based cladistic analysis of the concatenated character data set yielded a single, completely resolved, supra-generic cladogram, often with high values of nodal jackknife support (Fig. 2.1), which suggested the need for a revised classification for the phasianine galliforms. A brief review of the major galliform results of Crowe et al. (2006) The megapodes are monophyletic (clade 1 in Fig. 3.1) and cladistically basal within a monophyletic Galliformes. The next clade comprises the monophyletic cracids (clade 2), followed by the relatively basal guineafowls (clade 3), and not the New World quails (clade 4), both of which are monophyletic and in turn sister to the balance of the phasianids. The New World quails are basal within the remaining phasianids, but include an African sister clade formed by the monotypic stone partridge Ptilopachus petrosus and Hall’s (1963) previously phylogenetically enigmatic Nahan’s ‘Francolin’ (‘Francolinus/Acentrortyx’ nahani) (clade 4). Within the remainder of phasianids, the grouse and ‘true’ (i.e. wattled) pheasants were the only demonstrably monophyletic 42 supra-generic traditionally recognized clades; Phasianids including the remainder of the Phasianinae and Perdicinae (especially the latter) are polyphyletic; basal among these phasianids is a novel clade comprised of Afro/Asian ‘partridges’ (Xenoperdix, Rollulus and Arborophila spp. - clade 5), with the remaining Phasianinae and Perdicinae taxa sundered phylogenetically into five clades which are: the Old World quails (Coturnix and Excalfactoria spp.) including the monotypic Madagascar ‘partridge’ Margaroperdix madagarensis), some Old World partridges (e.g. Alectoris, Tetraogallus, Ammoperdix and Perdicula spp.), and some ‘francolins’ (Pternistis spp. which are commonly known as spurfowls) (clade 6); the junglefowls (Gallus spp.), Bambusicola partridges and ‘true’ (= quail-like) francolins (Francolinus, Dendroperdix, Peliperdix and Scleroptila spp.) (clade 7); the Afro-Asian peafowls sensu lato including the Argus Pheasant (Argusianus and Rheinardia spp.) the Congo Peacock (Afropavo congensis), and the peacock-pheasants (Polyplectron spp.) (clade 8); the turkeys (Meleagris spp.) and ‘the’ partridges (Perdix spp.) as sister-species (with relatively weak jackknife support) and sister to the grouse (e.g. Tetrao, Bonasa and Tympanuchus spp.) (clade 9); and the ’true’ pheasants (Phasianus, Catreus, Syrmaticus, Chrysolophus, Lophura and Crossoptilon spp.) that exclude the peacock-pheasants and junglefowls, but that include (with no jackknife nodal support) the somewhat partridge-like taxa Ithaginis, Lophophorus, Pucrasia and Tragopan spp. (clade 10). A brief review of Galliform studies published since Crowe et al. (2006) Since the publication of Crowe et al. (2006) several further manuscripts have been published that have bearing on its conclusions. They, and their key findings, are as follows: Cox et al. (2007) analyzed eight nuclear and three mitochondrial markers for 43 16 ingroup galliform taxa (rooted on megapodes), and produced a maximum likelihood cladogram that was consistent with the finding that the guineafowls and not the New World Quails are basal phasianoids, i.e. confirming Crowe et al. (2006). Analysis of 20 ingroup galliforms (rooted on Anseriformes) for 25 retroposed elements (large insertions of nuclear DNA commonly called ‘jumping genes’) by Kriegs et al. (2007), recovered a cladogram that was consistent with that in Crowe (2006) except in the placement of ‘the’ partridge, Perdix perdix. Rather than being sister to the turkey (Meleagris gallopavo), it was placed as sister to the Chrysolophus pheasants, leaving the turkey as sister to grouse. However, the methods and criteria used to generate this cladogram were not stated. Subsequent correspondence with the authors revealed that parsimony as implemented in PAUP*10b (Swofford 2002) and that the IRREV.UP option had been used to generate their preferred tree. This strategy assumes an ‘all-plesiomorphic’ outgroup and prevented the retroposed elements from being lost (reversing to absent) once they appear in a clade. Although this may be a reasonable assumption for retroposed elements, an unconstrained analysis with TNT (Goloboff et al. 2008) using an all-zero outgroup did not resolve the placement of either the partridge or turkey beyond grouping them in a polytomy with grouse and pheasants. Kimball and Braun’s (2008) model-based analysis of four nuclear introns and two mitochondrial coding regions for 41 ingroup taxa rooted on megapodes also recovered a cladogram largely congruent with that in Crowe et al. (2006) except that the turkey was placed as sister to grouse, and the partridges were placed as sister to the true pheasants, while the pavonines emerged paraphyletic. Eo et al. (2009) conducted a supertree analysis (Gatesy et al. 2002, Bininda-Edmonds et al. 2004) of available GenBank data with the resulting topology also being highly congruent with that of 44 Crowe et al. (2006). Model-based phylogenetic analysis (Shen et al. 2010) of complete mitochondrial genomes of 34 galliform taxa placed turkeys with grouse, the partridges with true pheasants and the pavonines as polyphyletic. This dataset also placed the New World quails as basal to guineafowls. Finally, a paper recently published by Wang et al. (2013) represents the most comphrehensively sampled (in terms of number of characters) molecular phylogeny to date. Their results refuted the traditional consideration of seven families and strongly supported the recognition of five major families within galliformes in the evolutionary sequence: Megapodiidae, Cracidae, Numididae, Odontophoridae and Phasianidae. This is similar to the findings in Crowe et al. (2006). The study also strongly supported the hypothesis that the deepest divergence within extant galliforms is between the Megapodes and all the other galliforms species, with the next divergence corresponding to that between Cracids and Phasianoidea. Similarly, as with Crowe et al. (2006), the turkeys (traditionally classified in Meleagrididae) and the grouse and ptarmigan (Tetraonidae) were nested within Phasianidae, hence this finding rejects the hypothesis that the turkeys and grouse form independent families. Ksepka (2009) and a range of other fossil-related publications Ksepka’s (2009) parsimony analysis of an expanded and revised matrix of 120 organismal characters combined with sequences from four mitochondrial markers (control region, 12S rDNA, CYTB, and ND2) and a nuclear ovomucoid intron G for 56 ingroup species rooted on Anseriforms and a Tinamus sp. (Tinamiformes), produced a cladogram largely congruent with the one presented in Crowe et al. (2006). However, the major point of criticism presented by Ksepka (2009) is the phylogenetic placement 45 of a key Eocene fossil, Gallinuloides wyomingensis (+/-54 mya old) from North America. He maintains that, contra Crowe et al. (2006), it is better placed at the stem (and not in the crown) of the galliform cladogram. Furthermore, there are several other fossil-based studies (Mlikovsky 1989, Stidham 2008, Elanowski and Stidham 2011, Mourer-Chauvire et al. 2011) that have challenged the divergence times ascribed to clades in Crowe et al. (2006). This chapter In this chapter, the focal taxa are francolins sensu lato. They are small to medium-sized, sedentary, Old World, partridge/quail-like gamebirds which occur in varied habitats, from dry, lowland grassland to montane forests (Hall 1963, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). Taxonomically, francolins were invariably placed in the family Phasianidae. In some more finely partitioned classifications, within the Phasianidae, they are placed in the sub-family Phasianinae (including Phasianus and related species), and together with other Old World partridgeand quail-like gamebirds (e.g. Perdix and Coturnix spp.), in the tribe Perdicini (Chapin 1932, Peters 1934, Wolters 1975-82, Crowe et al. 1986, Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994, Madge and McGowan 2002). Controversy has centered on the status of monophyly of the genus Francolinus sensu Hall (1963) and debate over the geographic origin of the genus ‘Francolinus’, for which two contrasting hypotheses have been postulated (Hall 1963, Crowe and Crowe 1985). Based on the notion that the genus Francolinus shares its closest affinities with other Palaearctic and Asian genera, Hall (1963) strongly argued for the genus to be of Asian origin with its age being traced back to the Oligocene +/- 25-35. Although Crowe and Crowe (1985) 46 concurred with Hall that the ancestor was quail-like, they hypothesized an African origin for the genus, with Asia being colonized by a nomadic or migratory ancestor that diversified and became sedentary. The enigmatic Stone Partridge occurs on rocky outcrops in the arid habitats of the northern savanna belt including the Sahel south of the Sahara, from Gambia to Ethiopia, and south to Cameroon and northern Kenya (Crowe et al. 1986). It was described initially as a Tetrao by Gmelin (1789), but was subsequently placed into a monotypic genus Ptilopachus by Swainson (1837). Nahan’s Francolin in contrast, is a highly-localized species associated with core areas of primary forests of the eastern equatorial lowlands of the Democratic Republic of the Congo and Uganda (Crowe et al. 1986, Sande et al. 2009). This taxon was first placed by Dubois (1905) in the genus Francolinus, but subsequently moved by Chapin (1926) into a monotypic genus, Acentrortyx. Hall (1963) placed it back into Francolinus because she doubted the value of characters used to split it from Francolinus. Furthermore, she linked it tentatively to members of her putatively monophyletic ‘Scaly Group’ of spurfowls (Ahanta Francolin Pternistis ahantensis, Scaly Francolin P. squamatus and Grey-striped Francolin P. griseostriatus; all previously placed into Francolinus by Hall) on the basis of bare-part colouration and plumage characteristics. In a morphometric analysis based on osteological features, Crowe and Crowe (1985) also placed F. nahani near members of Hall’s Scaly Group, closest to P. ahantensis. In a further reworking of the francolins, Crowe et al. (1992) placed F. nahani in a resurrected monotypic subgenus, Acentrortyx, within the African spurfowl genus Pternistis, although speculating that it might represent a phylogenetically relictual taxon, unrelated to other African galliforms. 47 In this chapter, further evidence in terms of the large number of francolin and spurfowl species analyzed, mitochondrial and nuclear DNA (and new evidence from vocalizations and behaviour of particularly F. nahani) is explored in the context of the phylogenetic affinities suggested by Crowe et al. (2006) with the goal of producing a robust phylogenetic hypothesis for Africa's diverse Galliform lineages, and in particular francolins sensu lato. Materials and methods Collection of data Taxon sampling The taxon sampling was based on that of Crowe et al. (2006), with a number of important changes. In order to increase the confidence that no taxa had been overlooked, all additional African ‘francolin’ species (e.g. several additional Pternistis spp.) as well as additional species of Asian and New World galliforms were sequenced for this study and further sequences were obtained from GenBank (Table 2.1). A sample of Ptilopachus petrosus was obtained from Ghana, and three samples of F. nahani were obtained from Budongo Forest, Uganda. Molecular approach Four mitochondrial (mtDNA) markers and three nuclear (nucDNA) markers, which occur on distinct chromosomes and thus provide independent estimates of phylogeny, were used in this study. The mitochondrial markers (Cytochrome-b – CYTB, NADH Dehydrogenase Subunit 2 - ND2, 12S Ribosomal DNA - 12S, and Control Region CR), and nuclear markers (Ovomucoid intron G – OVOG, Transforming Growth Factor 48 Beta 2 intron 5 – TGFB, and GAPDH intron 11 - GAPDH) were investigated since these markers have helped to resolve the phylogenetic status of other galliform genera and species (Armstrong et al. 2001, Dimcheff et al. 2000, 2002, Crowe et al. 2006, Hackett et al. 2008). Laboratory techniques Total genomic DNA was extracted from blood, heart and liver tissue using the DNeasy animal tissue protocol provided with the DNeasy tissue kit (Qiagen). The initial CYTB primers amplified 1337 base pairs (Table 3.2). Due to the length of this region, an internal primer (Table 2.2) was also used in sequencing this region. The initial CYTB primer pair did not amplify Grey-striped Spurfowl Pternistis griseostriatus and Yellownecked Spurfowl P. leucoscepus, thus further galliform specific primers were also used (Table 2.2). Double stranded DNA templates were amplified by polymerase chain reaction (PCR) using 0.75 units of BIOTAQTM DNA polymerase (Bioline) in 30 µl reactions. Reactions also contained 1 x NH4 buffer, 2.5 mM MgCl2, each dNTP at 0.1 mM, each primer at 0.3 µM, and 3 µl of extracted DNA was used as template. The thermal profile used comprised an initial denaturation step at 94°C for two minutes, followed by 30 cycles of 94°C for one minute, 52°C for one minute and 72°C for two minutes, with a final extension step of 72°C for seven minutes. PCR-amplified products were cleaned from solution or gel using the GFX TM PCR DNA and gel band purification kit (Amersham Biosciences) prior to cyclesequencing with the ABI PRISM Big DyeTM Terminator v3.1 cycle-sequencing Ready Reaction Kit (Applied Biosystems). Sequencing products were resolved on an ABI 49 PRISM 3100 Genetic Analyser. Sequences were assembled and checked for incorrect base calling and the presence of stop codons using SeqMan II (LaserGene systems software, DNAstar, Inc.). Consensus sequences were aligned using Clustal and adjusted manually in MegAlign (LaserGene systems software, DNAstar, Inc.). Analyses of data Phylogenetic analyses Three methods of phylogenetic analysis with different optimality criteria were employed to generate phylogenetic hypotheses: Bayesian inference (BI), maximum likelihood (ML) and parsimony. In all analyses, indels and ambiguous character states were considered as 'missing' and all characters were treated as non-additive. Parsimony-based phylogenetic analyses were conducted using TNT (Tree analysis using New Technology - Goloboff et al. 2008). In TNT, the search strategy employed was the ‘traditional’ search option. When multiple, equally parsimonious cladograms persisted, a strict consensus cladogram was constructed. The extent to which each non-terminal node is supported by different character partitions was determined by using the ‘jackknife’ resampling strategy with: 1000 replicates, TBR branch-swapping, five random additions of taxa per replicate with the deletion of 36% of the characters per jackknife replicate (Farris et al. 1996, Källersjö et al. 1998). Since gene regions can evolve under different models of evolution, it has been argued that a partioned, mixed-model approach should be used when concatenating these different datasets in a model-based phylogenetic analysis (Ronquist and Huelsenbeck 2003, Nylander et al. 2004). Mixed-model Bayesian analyses were undertaken in MrBayes v3.1.2 (Huelsenbeck and Ronquist 2001, Ronquist and 50 Huelsenbeck 2003). Substitution models for each locus were determined in PAUP*4b10 (Swofford 2002) with Modeltest 3.06 (Posada and Crandall 1998), using the Akaike Information Criterion (Akaike 1973, Posada and Buckley 2004). Mixed-model analyses allowed different parameters (base frequencies, rate matrix or transition/transversion ratio, shape parameter, proportion of invariable sites) to vary between the partitions (gene regions and codon positions) (Nylander et al. 2004). Four Metropolis-coupled MCMC chains (one cold and three heated) were run for 10 million generations with trees sampled every 100 generations. A Dirichlet distribution was assumed for estimation of the base frequency parameters and an uninformative (flat) prior was used for the topology. The ‘burn-in’ period (discarded cycles before the chains had reached stationarity) varied per analysis but was typically 500 000 generations (5000 trees); posterior probabilities (PP) were estimated from the remaining generations. Each Bayesian analysis was run twice (random starting point for each run). The loglikelihood values and posterior probabilities were checked using Tracer v1.4.1 (Rambaut and Drummond 2007) to confirm that the chains had reached stationarity. The potential scale reduction factor was confirmed to approach 1.0 (for all parameters) and the average deviation of split frequencies converged towards zero. Mixed-model maximum likelihood analyses were performed using the Randomised Axelerated Maximum Likelihood algorithm for High Performance Computing (RAxML) v7.0.4 (Stamatakis 2006, Stamatakis et al. 2008) as implemented on the CIPRES portal. Mixed-model RAxML analyses make use of a GTR++ model partitioned by gene or codon postion. The following analyses were run: mixed-model mtDNA (one model for each codon position, and also as a single data partition); a mixed-model analysis of the nuclear DNA genes, partitioned by each of the four gene 51 regions, and a mixed-model analysis of the combined mtDNA and nuclear DNA datasets. Support at nodes was assessed with 100 non-parametric bootstrap (BS) pseudoreplicates. The use of different methodological approaches (optimality criteria) facilitated the identification of method-based incongruence. Divergence date estimation The previous and most comprehensive dating analysis of Galliformes was conducted by Crowe et al. (2006), which made use of the Eocene fossils Gallinuloides wyomingensis (Green River Formation) and Amitabha urbsinterdictensis (Bridger Formation) as calibration points. Further preparation and re-examination of Amitabha resulted in this fossil being removed from Galliformes and it is now placed in the Rallidae (Ksepka 2009). Further, Ksepka (2009) argues that Gallinuloides is best placed at the stem and not within the crown of the galliform phylogeny, as previously suggested by Crowe et al. (2006). If Ksepka (2009) is correct, this would suggest that previous estimates of divergence dates among galliform lineages have been overestimated. As a consequence, a new dating analysis was conducted using BEAST v. 1.6.2 (Drummond and Rambaut 2007), omitting both Gallinuloides and Amitabha and instead calibrated via the use of three additional fossil calibration points: (1) a fossil of a Crested Francolin Dendroperdix sephaena at 4.5-5.0 mya as a minimum date for the age of the true Francolins (Crowe 1992), (2) a basal date for the Tetraoninae (Grouse and allies) of 27-29 mya (Crowe and Short 1992), and (3) a basal date for Polyplectron (Peacock-Pheasants) of 34-36 mya (Olson 1974, modified by T.M. Crowe unpubl. data). An uncorrelated lognormal clock was used with the same data partitions and nucleotide substitution models as described for the Bayesian analyses above. The 52 analysis was run for 80 million generations with tree sampling taking place every 2000 generations. Convergence was determined as described above for the Bayesian phylogenetic analyses. Field observations of behaviour and vocalizations of P. petrosus and F. nahani Behavioural observations and vocalizations were recorded in the field: F. nahani was observed in the Budongo (1.714°N, 31.543°E) and Mabira (0.399°N, 33.049°E) forests, Uganda, in 1999, 2002, 2008 and 2009 by Callan Cohen; P. petrosus was observed near Mora (11.083°N, 14.114°E) and Benoue National Park (8.116°N, 13.679°E) in Northern Cameroon in 2002, 2004 and 2010, and near Bandiagara (14.359°N, 3.584°E), central Mali, in 2006, also by Callan Cohen. Sound recordings were made using a stronglydirectional Sennheiser ME-67 microphone with a K6 power module. The recordings were made onto various media including a Fostex FR-LE-2 solid-state recorder, a Sony RH1 minidisc recorder in uncompressed format, and an Edirol R-09HR. These were supplemented by further vocalizations of F. nahani from Brian Finch (unpubl. data) and from Chappuis (2000). Vocal analyses Calls of P. petrosus and F. nahani were compared aurally to all available African galliform species on Gibbon (1995) and Chappuis (2000), supplemented by additional calls from the British Library Sound Archive and Macaulay Library of Natural Sounds. In addition, sonograms were made from typical advertisement calls (heard most often at dawn and dusk) for P. petrosus and F. nahani and compared with those of putative sister taxa (spurfowls - Pternistis spp. and francolins - Scleroptila spp.) and other 53 African galliforms. Sonograms were generated in Raven Lite (Version 1.0, Cornell Laboratory of Ornithology). Results Phylogenetic analyses The parsimony tree adopted from Crowe et al. (2006) is presented in this chapter (Fig. 2.1) since it yielded a single, well-resolved, supra-generic cladogram, often with high values of nodal jackknife (JK) support (Fig. 2.1). A combination of all seven sequenced markers resulted in 5554 base pairs, involving 84 taxa, from which the Bayesian inference (BI) and maximum likelihood (ML) analyses were generated (Fig. 2.2 and 2.3). In comparing the BI (Fig. 2.2) and ML (Fig. 2.3) topologies generated with that of Crowe et al.'s (2006) parsimony tree (Fig. 2.1: based on combined mtDNA, nucDNA and organismal characters) all analyses recovered five major lineages of galliforms and not seven families as traditionally circumscribed. The sequence of divergence of these families is such that the Megapodiidae diverged first, followed in sequence by the Cracidae, Numididae, Odontophoridae and Phasianidae. Members of the two traditionally recognized families, the Meleagrididae (turkeys) and Tetraonidae (grouse and allies), are nested within the Phasianidae with 1.0 posterior probability (PP) and 86% ML bootstrap (BS) support, contrary to the findings in parsimony analysis where the turkeys (Meleagris spp.) and ‘the’ partridges (Perdix spp.) are sister (with relatively weak 71% JK), and in turn sister to the grouse (e.g. Tetrao, Bonasa and Tympanuchus spp.) (clade 9) with a lack of JK support. 54 The nodal support among the five prominent families (Megapodiidae, Cracidae, Numididae, Odontophoridae and Phasianidae) is 1.0 PP, whereas ML analyses recovered 100% BS support for the node between Cracids and Guineafowls, Guineafowls and Odontophorids and 91% between Odontophorids and Phasianids with no support between the basal Megapodes and Cracids. The parsimony topology recovered 100% JK support between Megapodes and Cracids, Guineafowls and Odontophorids, 98% between Cracids and Guineafowls and 91% between Odontophorids and Phasianids. The phylogenetically enigmatic phasianines remain Ithaginis, Tragopan, Pucrasia, Meleagris, and Perdix spp. (Table 2.3). With regard to the francolins and spurfowls, the parsimony, BI and ML topologies reject monophyly of the genus Francolinus decisively (Fig. 2.1, 2.2, 2.3) in support of francolins sensu lato comprising two distantly related species assemblages. Spurfowls (Pternistis spp.) form a strongly supported monophyletic lineage with 1.0 PP and 100% BS but interestingly a lack of JK support in parsimony. Lack of support in parsimony could be attributed to among other things, fewer spurfowl species having been included in the analysis, specific differences underlying the analytical principles for the three phylogenetic inference methods, and analyses based on genetic markers representing loci evolving at different rates. On the other hand, francolins (Francolinus, Dendroperdix, Peliperdix, Scleroptila spp.) are monophyletic with a PP of 0.99 and BS of 78% respectively, but also with a lack of JK support in parsimony. This could be for the same reasons as highlighted above. However, there is consensus among all three phylogenetic methods in revealing that the closest evolutionary relatives of spurfowls are Old World quails (Coturnix spp.), the monotypic Madagascar ‘partridge’ Margaroperdix sp., Old World partridges such as Alectoris spp., Tetraogallus spp., 55 Ammoperdix spp. and Perdicula spp. In contrast, francolins are most closely related to the chickens (Gallus spp.) and Bamboo partridges (Bambusicola spp.) One other major revelation regards what was traditionally considered a francolin that is, Nahan’s Francolin Francolinus nahani. All three inference methods strongly refute this in support of Nahan’s Francolin being sister to the Stone Partridge Ptilopachus petrosus with 1.0 PP and 100% BS (Fig. 2.4, see also Cohen et al. 2012). Also these results demonstrate strong support that the ‘duo’ form a sister relationship with representatives of the New World Odontophorid quails, Callipepla, Colinus, Oreortyx and Cyrtonyx spp. with 1.0 PP and 100% ML BS (Fig. 2.2, 2.3, respectively), 100% JK (Fig. 2.4) and 98 JK (Fig. 2.1). Analyses of data partitions Mitochondrial versus nuclear DNA analyses The combined mt- and nucDNA BI topologies are similar to that of the mt- and nucDNA ML topologies, and as such only the mt- and nucDNA ML trees with branchlengths are presented. The combined mtDNA ML (Fig. 2.5) and BI analyses recovered similar topologies to that of the combined total evidence BI (Fig. 2.2) and ML (Fig. 2.2) DNA analyses, supporting the five major galliform families in the same sequence of evolution, although with slighly less support at several nodes. In both mtDNA BI and ML (Fig. 2.5) analyses, there is further support for the split of francolins into francolins and spurfowls, and strong support for the phylogenetic placement of Nahan’s Francolin and P. petrosus as sister species. Further, both spurfowl and francolin taxa were recovered as distinct monophyletic assemblages in the BI and ML mtDNA analyses; these results are consistent among each of the data partitions (Table 2.4). 56 Both the combined nucDNA BI and ML (Fig. 2.6) analyses appear to have been influenced by missing data for some taxa. For instance, one Cracid species (Crax spp.) is nested in the Megapodes and the placement of Rollulus rouloul is uncertain. Another difference is that even though the PPs are generally high at the deeper nodes in the BI analyses, there is poor nodal support (less than 50% BP) in the ML tree coupled with very short branch-lengths. The branch between Numididae and Odontophoridade is unresolved. However, despite some uncertanty at the base of the phylogeny for nucDNA in the BI and ML analyses, the phylogenetic resolution of the focal taxa, that is francolins and spurfowls and also Nahan’s Francolin largely remain the same with the exception that the francolins are recovered as monophyletic in both the nucDNA BI and ML (Fig. 2.6) trees. Otherwise, the francolin-spurfowl dichotomy is maintained with spurfowls being monophyletic (1.0 PP, 100% BS) and the relationship between Nahan’s Francolin and P. petrosus is strongly supported (1.0 PP, 100% BS). Further, the sister relationship between Nahan’s Francolin and P. petrosus and the New World quails is maintained with 1.0 PP and 99% BS. Divergence times The molecular divergence dates (Fig. 2.7) of the deeper nodes point to the oldest split being between the Cracids/Megapodes and the rest of the galliform species and could have happened at around 65.0 mya (HPD 53.5-79.8). The Megapodes and Cracids diverged from each other at c. 56.8 mya (HPD 42.9-72.4), the Numididae at 46.5 mya (HPD 42.1-51.8) and the Odontophorids and Phasianids diverged at c. 44.5 mya (HPD 40.6-49.2). 57 The estimated date for the timing of the split between P. petrosus and F. nahani was 9.6 mya (95% HPD 5.8-14.1), and 37.4 mya (95% HPD 31.8-43.1) for the divergence between this clade and the New World Quails. The estimated divergence date between the Gallus/Bambusicola/francolins and the Coturnix/Alectoris/spurfowls clade is recovered at c. 33.6 mya (HPD 29.8-37.1) with the split between Gallus/Bambusicola and francolins at around 10.7 mya (HPD 8.4-13.7) and Coturnix/Alectoris and spurfowls at c. 23.1 mya (HPD 20.7-30.0). Interestingly, the time to most recent common ancestor of extant spurfowls is around 8.7 mya (HPD 7.410.4) and that for francolins slightly younger at c. 7.6 mya (HPD 7.0-8.3). Whether this could be correlated to habitats in which spurfowls and francolins thrive is a difficult question to answer. Vocal and behavioural comparison between F. nahani and P. petrosus The calls of P. petrosus (Fig. 2.8a) and F. nahani (Fig. 2.8b) are strikingly similar and differ from those of other francolins and spurfowls both sonographically and aurally: the exemplars presented here are from the widely available Chappuis (2000). They consist of a long series of whistles that increase in volume and are often joined by additional birds calling near the end of the sequence. The structure of the whistle begins with a short lead in tone between 1-1.5 kHz, followed by a double-peaked whistle with high and low frequency values of 1.5 and 2.5 kHz respectively, and associated harmonics (Fig. 2.8c-e). Interspecific variation based on our additional recordings is limited and influenced largely by the number of group members calling simultaneously. These calls differ qualitatively to a large degree from any other African galliform. Thus, it is not possible to identify homologous call units to enable direct comparison among 58 Galliform lineages. No other African galliform examined has a similar whistle structure. In particular, these calls strongly contrast with typical spurfowl calls of the putative relatives of F. nahani, which consist of slurred, almost grating, raucous calls that do not very much in frequency (Fig. 2.8c-e; from Chappuis 2000). Behavioural observations (substantiated by photographs and extensive field observations) indicate that both P. petrosus (Fig. 2.9a) and F. nahani (Fig. 2.9b) hold their tails in a distinctive, bantam-like cocked position. Discussion Phylogenetic relationships In general the phylogenetic results for this study, including the family status within galliformes, are similar to those reported by Crowe et al. (2006), with the exception that here species, and not genera, are terminals. With respect to the focal taxa, the supported sister relationship between F. nahani and P. petrosus contradicts all other published treatments of the Galliformes (e.g. Hall 1963, Crowe et al. 1985, 1986, 1992, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). Furthermore, the basal position of this clade relative to ‘true’ francolins and spurfowls suggests that they represent a relictual group sister to the New World quails (Odontophoridae), and are only distantly related to the other Old World galliforms. Intriguingly, both species occupy habitats - dense primary forest understorey and rocky outcrops - that have been suggested by Kingdon (1989) as having a higher than expected proportion of relictual species. Further, the present study confirms the results of Crowe and Bloomer (1998) and Crowe et al. (2006) in suggesting that francolins and spurfowls are not each other’s 59 closest relatives as suggested by Hall (1963). The genus Francolinus comprises at least two distantly related lineages (excluding F. nahani, see above). The partridge-like spurfowls, which are related to quails (e.g. Coturnix, Excalfactoria, Margaroperdix, Perdicula and Ammoperdix spp.) and certain Old World partridges (e.g. Tetraogallus and Alectoris spp.), and the quail-like francolins (Dendroperdix, Peliperdix and Scleroptila spp.), which are related to the ‘true’ Asian francolins (Francolinus spp.), chickens (Gallus spp.), and other Old World partridges (Bambusicola spp.). Morphological, behavioural and vocal similarities between F. nahani and P. petrosus Morphological similarities shared by F. nahani and P. petrosus, include: small size, red bare skin around the eye, lack of spurs and the lack of sexual dimorphism (Hall 1963, Johnsgard 1988, Madge and McGowan 2002). Although it is well known that Ptilopachus has a long, vaulted and regularly cocked tail (Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002), the same condition in F. nahani is less well known, because of its rarity and dense forest habitat (Stevenson and Fanshawe 2002). Hence, most bird artists have depicted the shape of the bird as that of a typical francolin or spurfowl (see illustrations in Crowe et al. 1986, del Hoyo et al. 1994, Sinclair and Ryan 2003). Only one relatively recent publication (Stevenson and Fanshawe 2002) has depicted the posture of this species correctly. This posture is illustrated (Fig. 2.9) based on two photographs of F. nahani in natural habitat (in Budongo and Mabira Forests, Uganda) and one P. petrosus taken in Cameroon. Dendroperdix sephaena is the only other African galliform known to cock its tail (Madge and McGowan 2002), but it is not closely related to these species (Crowe et al. 2006). 60 The biology of F. nahani is very poorly known (Crowe et al. 1986, Sande et al. 2009), and its voice has only been described relatively recently (Chappuis 2000, Stevenson and Fanshawe 2002), thus hampering the correct taxonomic placement of this species. The calls of both F. nahani and P. petrosus are a series of whistles increasing in volume, and are strikingly similar (see Fig. 2.8). Chappuis (2000), in a booklet accompanying his CD set, noted this similarity, as did Brian Finch, who worked on the voice section of Stevenson and Fanshawe (2002). Furthermore, our field observations attest that both species live in small, family groups and have interactive calling. Given the long divergence time between these two species (5-13 mya), it is interesting that the nature of the calls have been so well conserved. The group duetting may indicate a strong social cohesiveness function and the calls could be subject to stabilizing selection in this regard (Payne 1971). Another matter to consider is the exact nature of the habitat of these species. Whereas Ptilopachus is found in the arid zone, it does inhabit dense bush growth among large boulders, a challenging environment for the broadcast of sounds, with many obstacles, similar to the dense forest understorey inhabited by F. nahani. Indeed, given the likely Miocene divergence between these species, it is most likely that their common ancestor inhabited forest habitats (Fjeldså and Bowie 2008). The open savannas and arid land lineages of mammals only seem to have radiated later, in the Plio-Pleistocene when dry habitat became much more widespread in Africa (e.g. deMenocal 2004). The plumage of these similar birds seems to have been very well conserved, and besides aspects of colouration that presumably relate to camouflage (F. nahani is darker above, whereas P. petrosus is somewhat paler), there has been remarkably little divergence. 61 Historical biogeography As expected the omission of the fossil Gallinuloides wyomingensis did result in the recovery of younger divergence times. For example, Crowe et al. (2006) estimated that the stem Ptilopachus plus Odontophoridae clade to have diverged at 55.5 mya (95% HPD 50.1-65.9) whereas our analyses based on three ingroup fossils recovered this node at 37.4 (95% HPD 31.7-43.0). Overall, these results are in agreement with the view of Ksepka (2009), that although stem galliformes likely existed in the Cretaceous (i.e. pre 65 mya), the divergence of crown-group lineages remains inconclusive. Our estimated divergence between Ptilopachus and the New World quails occurred around the middle of the Eocene period (55.8-33.9 mya). The Eocene was a remarkable period in earth history, with high temperatures and precipitation, in essentially an ice-free world (Eberle and Greenwood 2012, Harrington et al. 2012). Connections existed between Africa and Europe, and Europe and North America via Greenland, although by about 40 mya, the time of our inferred Ptilopachus-New World quail split, it seems unlikely that this landbridge was still open (Scotese 2001). However, Eocene and Oligocene fossils have been discovered from France that are most similar to New World quails (Crowe and Short 1992, Crowe et al. 2006, see MourerChauvire 1992 for another view) suggesting that Europe likely played an important part in the biotic exchange between African and North American lineages. Should the ‘Greenland landbridge’ have been closed, an alternate connection may have been via Asia to North America along the Bering Strait. Given the relatively sedentary habitats of both Ptilopachus and New World quails, it seems highly unlikely that direct dispersal between African and the New World occurred, as for example inferred for some lineages of birds such as Thrushes (Turdus spp., Voelker et al. 2009). In summary, 62 although difficult to infer without additional fossil evidence, it seems likely that one of the above mentioned landbridges played an important role in shaping the biogeographic origins of both the African and New World members of the Odontophoridae. At 5-13 mya the super-African rainforest was likely still expansive (Fjeldså and Bowie 2008) which may suggest that P. petrosus either secondary invaded its present arid and rocky habitat, or that it was more geographically restricted in the past, and that F. nahani likely occupies the ancestral habitat of these taxa. Interestingly many of the extant New World Quails are more open habitat associated (de Hoyo et al. 1994), with closer habitat affinities to P. petrosus than F. nahani. The relative divergence time between Gallus/Bambusicola/francolins and the Coturnix/Alectoris/spurfowls clade is recovered at c. 33.6 mya whereas the split between Gallus/Bambusicola and francolins and Coturnix/Alectoris and spurfowls was recovered at around 10.7 and 23.1 mya respectively. Thus the diversification of, francolins in Africa happened much more recently. Further, the divergence times among the lineages in each of the major clades seem to have occurred consistently over roughly the same period and given that the spurfowls evolved at around 8.7 mya and the francolins c. 7.6 mya, this indicates that spurfowls are a little older than the francolins. Most spurfowl species occur in forests and along forest edges, with only a few species occurring in distinctive habitats such as the rocky hills preferred by P. hartlaubi and semi-wooded grasslands preferred by members of the Bare-throated species complex. Francolins on the other hand occur in different types of grasslands including, montane, wooded and scrubby grassland, with the exception of one species, Francolinus lathami, which is a forest specialist. 63 Taxonomic recommendations On the basis of the close genetic relationship between F. nahani and P. petrosus, as well as their shared behavioural and vocal characters, we recommend that F. nahani be moved to the genus Ptilopachus Swainson (on the basis of priority). We recommend the placement of Ptilopachus in the Odontophoridae to emphasize its sister relationship to this New World family of galliform birds. Ptilopachus should be placed first in the sequence of genera of the Odontophoridae. With regard to the spurfowls and francolins, this study supports assignment of the genus Pternistis to all spurfowl taxa. However, there seems to be a dire need to revise the taxonomic status of francolins in particular, the distinct forest species F. lathami, one of Hall’s phylogenetically enigmatic species, as well as the other two enigmatic species, the Asian F. gularis and F. pondicerianus, which emerged as sister species to the enigmatic African Dendroperdix sephaena. This will form the focus of a later chapter in this thesis. Conclusions After several studies focusing on improving our understanding of the galliform phylogeny, it is frustrating that relationships within the Phasianidae, in particular, the position of the turkeys and grouse and allies, remains unresolved. However, this study clearly demonstrates the support for the five families considered in modern galliform classification thereby rejecting the recognition of the seven traditionally recognized families (including Meleagrididae and Tetraonidae) in the order outlined in Crowe et al. (2006) and Wang et al. (2013). As Wang et al. 2013 concluded “additional data collection will be necessary to resolve the remaining uncertainties within Galliformes”, 64 they also consider the challenges that remain to be resolve within the galliform phylogeny, to be aggravated by for example, the use of group names such as ”pheasants” and “partridges” which do not imply common descent. It would be safe to conclude that the traditionally known francolins represent at least two distantly related lineages with F. nahani being pulled to the base of the galliform tree to go with P. petrosus. Thus, F. nahani is not a francolin as traditionally considered but a partridge related to the New World quails. 65 Tables and Figures Table 2.1. GenBank accession numbers of the samples analyzed in this study. Genus Species Sample no. Locality CYTB ND2 CR 12S OVOG TGFB GAPDH Acryllium vulturinum – – AF536742 AF536745 – AF536739 DQ832070 – – Afropavo congensis – – AF013760 DQ768253 DQ834507 – AF170991 – – Alectoris chukar – – L083781 DQ768273 DQ834525 – AF170987 FR694121 FR694070 Alectoris Alectoris Alectura Arborophila graeca rufa lathami javanica – – Z487724 – DQ834524 – – – – – – – – Z487754 NC007227 – AY274051 DQ834523 DQ834465 – AY274004 AF170988 DQ832069 – EU737326 – – – – AM236890 DG093804 – DQ832097 DQ832074 – – Arborophila torqueola – – AM236889 – DQ834475 – – – – Bambusicola thoracica – – EU165706 AF222538 DQ834513 EU165706 AF170978 – – Bonasa Callipepla Callipepla umbellus californica gambelii – – AF230167 AF222541 DQ834476 U83740 – – – – – AB120131 AF028773 DQ834473 – – Submitted Submitted – – L083821 AF028761 DQ834472 – – – – Catreus wallichii – – AF028792 DQ768254 DQ834499 – AF170980 – – Chrysolophus amherstiae – – AB120130 – – DQ832102 – – – Chrysolophus pictus – – AF028793 DQ768255 DQ834497 – – – – Colinus cristatus – – – – – – – EU737357 – Colinus virginianus – – EU372675 AF222545 DQ834469 AF222576 – – Submitted Coturnix coturnix – – L083771 X57246 DQ834529 X57245 – Submitted EU737363 Coturnix japonica – – NC003408 NC003408 – NC003408 – – – Crax alector – – AY141921 – – – – Submitted – Crax rubra – – AY956378 AY274050 AY145307 AY274003 – – – Crossoptilon crossoptilon – – AF028794 DQ768256 DQ834500 – AF170981 – Cyrtonyx montezumae – – AF068192 AF028779 DQ834467 – AF170976 – – Dendroperdix sephaena TMC9 Marico, SA FR694140 DQ768274 DQ834515 FR691559 DQ832083 FR694111 FR694102 Falcipennis canadensis – – AF170992 AF222548 DQ834478 AF222577 AF170986 – – Francolinus francolinus FR691376 – – – – Francolinus gularis Francolinus lathami Francolinus pictus AMNH 776813 India Francolinus pondicerianus Gallus gallus Gallus AF013762 AMNH DOT8023 India U90649 India U906497 – – – – – – Cameroon AM236893 DQ768257 FR691377 FR691546 DQ832082 FR694113 FR694080 FR694142 – – – – – – AMNH DOT8050 India FR691632 DQ768279 – FR691547 DQ832081 FR694114 FR694081 – – L083761 AB086102 DQ834510 NC001323 AF170979 FR694110 FR694078 varius – – AB044988 AF222551 – – – – – Guttera pucherani – – AM236882 – – – – – – Ithaginis cruentus – – AF068193 DQ768258 DQ834487 – DQ832076 – – Leipoa ocellata – – AM236879 AF394619 – AF222586 – – – Lophophorus impejanus – – AF028796 DQ768259 DQ834486 DQ832098 DQ832075 – – Lophura nycthemera – – L083801 DQ768261 DQ834498 – – – – Margaroperdix madagarensis – – U906407 – DQ834528 – – – – Megapodius eremita – – AF082065 AY274052 – AY274005 – – – Meleagris gallopavo – – L083811 AF222556 DQ834485 U83741 AF170984 – – Numida meleagris – – L083831 NC006382 DQ834466 AF222587 AF170975 EU737410 FR694071 Oreortyx pictus – – AF252860 AF028782 DQ834468 – AF170977 Submitted Submitted Ortalis vetula – – L083841 AF394614 – – AF170974 – – Pauxi pauxi – – AF068190 AY140750 AF165439 AF165449 AF170973 – – Pavo cristatus – – L083791 AF394612 DQ834508 AY722396 AF170990 – – Peliperdix coqui PFIAO 45 Settlers, SA AM236895 DQ768278 FR691379 FR691549 DQ832084 FR694115 FR694082 Perdix perdix – – AF028791 AF222560 DQ834484 AF222590 AF170982 – – Phasianus colchicus – – AY368060 AF222561 DQ834495 U837426 – – – Polyplectron bicalcaratum – – AF534564 DQ768263 DQ834503 – AF331959 – – Polyplectron emphanum – – AF330062 DQ768265 DQ834504 – AF331955 – AM236893 66 – Genus Species Sample no. Locality CYTB ND2 CR 12S OVOG TGFB GAPDH Pternistis adspersus PFIAO 206A – AM236910 DQ768276 DQ834535 DQ832113 DQ832095 FR694122 FR694087 Pternistis afer AM236908 DQ768281 DQ834533 DQ832111 DQ832092 FR694123 FR694088 Pternistis bicalcaratus U906377 FR691578 FR691370 FR691551 FR691690 FR694103 FR694089 Pternistis camerunensis TMC 42 Mount Cameroon FR694142 FR691577 FR691382 FR691552 FR691694 FR694124 FR694090 Pternistis capensis PFIAO 229 Kakamas, SA AM236909 DQ768282 DQ834534 DQ832112 DQ832093 FR694125 FR694091 Pternistis castaneicollis GB – AM236903 – – – – – – Pternistis clappertoni AMNH 541305 Takoukout, Cameroon FR691602 FR691576 FR691383 FR716655 FR691693 FR694126 FR694092 Pternistis erckelii AMNH DOT11039 Ethiopia U906387 – – – – – – Pternistis griseostriatus Ndalla Tanda AM236905 DQ768284 FR691384 FR691554 DQ832089 FR694128 FR694094 Pternistis hartlaubi Namibia U906397 FR691572 FR691555 FR691692 FR694129 FR694095 Pternistis hildebrandti – U906317 – – – – – – Pternistis icterorhynchus Fanadji FR691601 – – – – – – Pternistis jacksoni East slope, Mt Kenya FR691594 – – – – – – Kenya AM236906 FR691387 FR691556 DQ832090 FR694131 FR694097 FR694097 Marico River, SA AM236911 DQ834536 FR691557 DQ832094 FR694132 FR694098 FR694098 West Ruwenzori FR691592 – – – – – – Djibouti FR691590 – – – – – – PFIAO 108 TM 14682 AMNH 541411 TMC 121 GB AMNH 156922 AMNH26192 Watervalboven, SA Gold Coast, Hinterland Pternistis leucoscepus PFIAO 109 Pternistis natalensis TMC 120 Pternistis nobilis AMNH1759 Pternistis ochropectus FNHM 1971-1072 Pternistis rufopictus AMNH 202503 Gagayo, Muranza FR691588 – – – – – – Pternistis squamatus AMNH 541409 Nr York Pass, Sierra Leone AM236904 DQ768286 DQ834531 DQ832109 DQ832088 FR694133 FR694099 Pternistis swainsonii TMC 40 Marico River, SA AM236907 DQ768287 DQ834532 DQ832110 DQ832091 FR694134 FR694100 Pternistis swierstrai TMC 67 Angola FR691593 – – – – – – Ptilopachus nahani – Budongo forest, Uganda AM236885 DQ768288 FR691374 FR691545 DQ832071 FR694107 FR694075 Ptilopachus petrosus – Ghana AM236886 DQ768289 FR691375 FR691544 DQ832072 FR694108 FR694076 Pucrasia macrolopha – – AF028800 DQ768269 DQ834490 – AF170983 – – Rollulus rouloul – – AM236888 – – – – Submitted – Scleroptila afra PFIAO 59 Eastern Cape, SA AM236897 AF222550 DQ834517 AF222581 DQ832086 FR694116 FR694083 Scleroptila finschi AMNH 308887 Angola AM236896 DQ768290 – – – – – Scleroptila levaillantii TM 78622 Sterkspruit, SA AM236913 DQ768291 DQ834516 DQ832106 DQ832085 FR694117 FR694084 Scleroptila levaillantoides TMC 12 Petrus steyn, SA AM236900 DQ768292 DQ834519 DQ832108 – FR694118 FR694085 Scleroptila psilolaema Kenya FR691614 – – – – – – Scleroptila shelleyi PFIAO 47 Ayton farm, SA AM236898 DQ768295 DQ834518 DQ832107 DQ832087 FR694119 FR694101 Scleroptila streptophora TMC11 Cameroon FR691617 FR691573 FR691380 FR691550 – FR694120 FR694086 Syrmaticus ellioti – – – DQ768270 – – – – – Syrmaticus humiae – – AF534706 – DQ834491 DQ832099 DQ832077 – – Tetrao urogallus – – AB120132 AF222565 DQ834480 AF222594 – – – Tragopan temminckii – – AF229838 AF222566 DQ834488 AF222595 – – – Tympanuchus Xenoperdix phasianellus udzungwensis – – AF068191 AF222569 DQ834483 AF222598 AF170985 – – – – AM236887 DG093800 DQ834474 DQ832096 DQ832073 – – BM 80 1 1 1066 67 Table 2.2. DNA markers sequenced and primers used for PCR amplifications and sequencing of preserved tissue. Primer name Primer sequence (5’to 3’) Reference All Galliformes (General primers) Cytochrome b L14578 MH15364 ML15347 H15915 cta gga atc atc cta gcc cta ga act cta cta ggg ttt ggc c atc aca aac cta ttc tc aac gca gtc atc tcc ggt tta caa gac J.G. Groth (pers. comm.) P. Beresford (pers. comm.) P. Beresford (pers. comm.) Edwards & Wilson (1990) Control region PHDL PH-H521 PH-L400 PHDH agg act acg gct tga aaa gc tta tgt gct tga ccg agg aac cag att tat tga tcg tcc acc tca cg cat ctt ggc atc ttc agt gcc Fumihito et al. (1995) E.A. Scott (pers. comm.) E.A. Scott (pers. comm.) Fumihito et al. (1995) 12S rRNA L1267 H2294 aaa gca tgg cac tga ag(atc) tg gtg cac ctt ccg gta cac ttac c Moum et al. (1994) O. Haddrath (S. Pereira pers. comm.) NADH dehydrogenase subunit 2 (ND2) L5216 gcc cat acc ccr aaa atg H6313 ctc tta ttt aag gct ttg aag gc Sorenson et al. (1999) Sorenson et al. (1999) Ovomucoid G OVO-G Forward OVO-G Reverse caa gac ata cgg caa caa rtg ggc tta aag tga gag tcc crt t Armstrong et al. (2001) Armstrong et al. (2001) GAPDH intron-11 GapdL890 GapdH950 acc ttt aat gcg ggt gct ggc att gc cat caa gtc cac aac acg gtt gct gta Friesen et al. (1997) Friesen et al. (1997) Transforming Growth Factor Beta2 intron-5 TGFb2-5F ttg tta ccc tcc tac aga ctt gag tc TGFb2-6R gac gca ggc agc aat tat cc Primmer et al. (2002) Primmer et al. (2002) Cytochrome b Spurfowl-specific primers L14851 (General) Pt-H195 Pt-H194 MH15145 cct act tag gat cat tcg ccc t ttt cgr cat gtg tgg gta cgg ag cat gtr tgg gct acg gag g aag aat gag gcg cca ttt gc Kornegay et al. (1993) R. Moyle & T. Mandiwana-Neudani R. Bowie P. Beresford Pt-L143 Pt-H361 gcc tca tta ccc aaa tcc tca c gtg gct att agt gtg agg ag R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani Pt-L330 Pt-H645 tat act atg gct cct acc tgt ac ggg tgg aat ggg att ttg tca gag R. Bowie R. Moyle & T. Mandiwana-Neudani Pt-L633 Pt-H901 ggc tca aac aac cca cta ggc agg aag ggg att agg agt agg at R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani 68 Primer name Primer sequence (5’to 3’) Reference Cytochrome b Spurfowl-specific primers L2-2312 H15696 cat tcc acg aat cag gct c aat agg aag tat cat tcg ggt ttg atg R. Bowie Edwards et al. (1991) Pt-L851alt Pt-H1050 cct att tgc cta cgc cat cct ac gat gct gtt tgg ccg atg R. Bowie R. Bowie Pt-L961 Pt-L961alt HB20 (General) cga acc ata aca ttc cca c ctc atc cta ctc cta atc ccc ttg gtt cac aag acc aat gtt R. Moyle & T. Mandiwana-Neudani R. Bowie J. Feinstein (pers. comm.) 69 Table 2.3. Cladistic placement of phylogenetically enigmatic phasianine galliforms in concatenated analyses. Source of data Parsimony Concatenated organismal, F. nahani Meleagris Ithaginis Tragopan Pucrasia Perdix sister to NWQ sister to sister to sister to sister to sister to Tetraonini Pucrasia Tetraophasis + Ithaginis true pheasants mt & nucDNA Lophophorus Bayesian inference mt & nucDNA sister to P. petrosus sister to sister to sister to sister to sister to Tetraonini large Phasianine Lophophorus Perdix + true pheasants assemblage mtDNA sister to P. petrosus true pheasants sister to sister to sister to sister to sister to Tetraonini large Phasianine Lophophorus Perdix + true pheasants assemblage nucDNA Maximum likelihood mt & nucDNA mtDNA nucDNA sister to P. petrosus sister to P. petrosus sister to P. petrosus sister to P. petrosus sister to sister to Tetraonini true pheasants _ polytomous with polytomous with large Phasianine assemblage of assemblage of assemblage Phasianine spp. Phasianine spp. _ sister to sister to sister to sister to Tetraonini large Phasianine true true assemblage pheaseants pheasants sister to sister to sister to sister to sister to Tetraonini large Phasianine true true true pheasants pheasants pheasants sister to assemblage paraphyletic with _ paraphyletic with paraphyletic with large Phasianine Assemblage large Phasianine assemblage Lophophorus 70 large Phasianine assemblage Table 2.4. Support for the relationship of Ptilopachus petrosus and Francolinus nahani from different data partitions, + indicates supported branch; U - unresolved (Adopted from Cohen et al. 2012). Bayesian Inference Parsimony Maximum Likelihood Clade mtDNA nucDNA mtDNA nucDNA mtDNA nucDNA CYTB CR ND2 12S OVOG TGFB2 GAPDH Odontophoridae sister to P. petrosus & F. nahani 1.0 1.0 100 100 + + + + + + + + + P. petrosus sister to F. nahani 1.0 1.0 100 100 + + + + + + + + U 71 Figure 2.1. Summary of the single strict consensus parsimony cladogram recovered for the Galliformes by Crowe et al. (2006). Anseriforms, Megapodiidae, Cracidae, Guineafowls Numididae, New World quails sensu stricto Odontophorines, Ptilopachus spp., XEN/ROLL/ARB (Xenoperdix + Rollulus + Arborophila spp.), COT/MAD/ALEC/TETG (Coturnix + Excalfactoria + Margaroperdix + Alectoris + Tetraogallus spp.), Perdicula + Ammoperdix spp., Spurfowls Pternistis spp., Gallus + Bambusicola spp., Francolins Francolinus, Ortygornis, Afrocolinus, Peliperdix, Scleroptila spp., Rhenardia + Argusianus spp., Pavo + Afropavo spp., Polyplectron spp., Perdix spp., Turkey Meleagris gallopavo, Grouse Tetraonini spp., Lophophorus spp., Pucrasia macrolopha, Tragopan spp., Ithaginis cruentus, Wattled – Phesants (Phasianus, Lophura, Syrmaticus, Catreus, Crossoptilon Chrysolophus spp. Numbers at nodes are jackknife support values. See text for meaning of circled numbers (1-10) on clades. 72 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 .96 .97 1.0 .96 .99 1.0 1.0 .97 1.0 .99 .99 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 .99 .98 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 .98 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Figure 2.2. The 50% Bayesian majority-rule consensus tree generated from the combined four mitochondrial (Cytochrome-b – CYTB, NADH Dehydrogenase Subunit 2 - ND2, 12S Ribosomal DNA 12S, Control Region - CR) and three nuclear (Ovomucoid intron G – OVOG, Transforming Growth Factor Beta 2 intron 5 – TGFB, GAPDH intron 11 - GAPDH) DNA markers. Numbers associated with nodes represent posterior probability values (≥ 0.95 are shown). 73 100 100 Alectura lathami Leipoa accelata Megapodiidae Megapodius sp. Ortalis sp. 100 Crax sp. Cracidae 100 Pauxi pauxi Acryllium vulturinum 100 Numididae Guttera pucherani Numida meleagris Afropavo congensis 100 Pavo cristatus Bambusicola thoracica 95 Gallus gallus 99 Gallus varius Peliperdix coqui F Scleroptila streptophora 97 99 r Scleroptila psilolaema 100 Scleroptila finschi a 100 Scleroptila levaillantoides n Scleroptila shelleyi 74 78 Scleroptila afra 74 c Scleroptila levaillantii o Francolinus pictus 89 Francolinus francolinus l Francolinus lathami i Francolinus pondicerianus 100 99 Francolinus gularis n Dendroperdix sephaena Margaroperdix madagarensis s 100 100 Coturnix japonica Coturnix coturnix Pternistis leucoscepus 89 Pternistis swainsonii 94 Pternistis rufopictus 77 Pternistis afer S P Pternistis hartlaubi p h Pternistis squamatus 100 Pternistis icterorhynchus u 100 99 a 100 Pternistis bicalcaratus r s Pternistis griseostriatus 100 Pternistis camerunensis 100 f i Pternistis nobilis o a Pternistis castaneicollis 90 Pternistis ochropectus 79 w n Pternistis erckelii 100 74 l Pternistis swierstrai i Pternistis clappertoni s d Pternistis jacksoni Pternistis adspersus a 97 Pternistis natalensis 99 e Pternistis hildebrandti Pternistis capensis Alectoris rufa 100 95 Alectoris chukar Alectoris graeca Polyplectron emphanum 100 Polyplectron bicalcaratum Bonasa umbellus 100 Tympanuchus phasianellus 100 Falcipennis canadensis 86 100 96 Tetrao urogallus Meleagris gallopavo Pucrasia macrolopha Chrysolophus sp. Crossoptilon crossoptilon 77 88 80 Lophura nycthemera Catreus wallichii 100 85 Phasianus colchicus Syrmaticus sp. 100 Perdix perdix 91 Lophophorus impejanus Tragopan temminckii Ithaginis cruentus Xenoperdix udzungwensis 99 Rollulus rouloul 81 Arborophila javanica 100 Arborophila torqueola Cyrtonyx montezumae Colinus sp. 100 100 Callipepla californica 100 95 Callipepla gambelii 100 Odontophoridae Oreortyx pictus Ptilopachus nahani 100 Ptilopachus petrosus Figure 2.3. Maximum likelihood tree obtained from combined four mitochondrial (Cytochrome-b – CYTB, NADH Dehydrogenase Subunit 2 - ND2, 12S Ribosomal DNA - 12S, Control Region CR) and three nuclear (Ovomucoid intron G – OVOG, Transforming Growth Factor Beta 2 intron 5 – TGFB, GAPDH intron 11 - GAPDH) DNA markers. Numbers associated with nodes are bootstrap support values (only ≥ 70% are shown). 74 Ptilopachus petrosus Francolinus nahani Odontophoridae Other Phasianines Numididae Megapodiidae Figure 2.4. Parsimony cladogram for Ptilopachus petrosus and Francolinus nahani. Numbers at nodes are parsimony jackknife/maximum likelihood bootstrap/Bayesian probability values. 75 Pauxi pauxi Cracidae Crax rubra Ortalis sp. ∞ Acryllium vulturinum Numididae Numida meleagris * ∞ Guttera pucherani Cyrtonyx montezumae Oreortyx pictus *∞ Colinus virgianianus * * ∞ Odontophcallfornica * ∞ * Callipepla * Callipepla callfornica ∞ oridae ∞ Ptilopachus ∞ nahani *∞ Ptilopachus petrosus Ithaginis cruentus Pucrasia macrolopha Syrmaticus sp. Phasianus colchicus Lophura nycthemera * ∞ Catreus wallichii * ** *∞ ∞ ∞ Crossoptilon crossoptilon Chrysolophus sp. Perdix perdix Tragopan temminckii Lophophorus impejanus * Bonasa umbellus ∞ Tympanuchus phasianellus * * ∞ Tetrao urogallus ∞* *∞ Falcipennis canadensis ∞ Meleagris gallopavo Pavo cristatus * Afropavo congensis ∞ *∞ Gallus varius * gallus * ∞ Gallus * ∞ Bambusicola thoracica ∞ Francolinus lathami F * Francolinus pictus ∞ *∞ Francolinus r francolinus Francolinus pondicerianus a * * ∞* Francolinus gularis n Dendroperdix sephaena ∞ Peliperdix coqui c Scleroptila streptophora * o ∞ Scleroptila levaillantii * * Scleroptila shelleyi ∞ l * ∞ Scleroptila levaillantoides ∞ *∞ Scleroptila afra i ∞ * Scleroptila psilolaema ∞ n Scleroptila finschi ∞ s Polyplectron bicalcaratum * Polyplectron emphanum ∞ Pternistis capensis adspersus *∞ Pternistis Pternistis hildebrandti *∞ Pternistis natalensis S Pternistis griseostriatus Pternistis hatlaubi p * ∞ ∞ Pternistis squamatus ∞ Pternistis icterorhynchus u ∞ * Pternistis bicalcaratus r ∞ Pternistis swierstrai f Pternistis ochropectus o erckelii *∞∞* Pternistis castaneicollis w *∞ ∞* Pternistis Pternistis nobilis * Pternistis camerunensis l ∞ Pternistis clappertoni s Pternistis jacksoni Pternistis swainsonii * Pternistis rufopictus ∞ * Pternistis afer ∞∞ Pternistis leucoscepus Margaroperdix madagarensis *∞ * Coturnix coturnix ∞ * Coturnix japonica *∞ ∞Alectoris graeca * Alectoris chukar ∞ * ∞ Alectoris rufa Arborophila javanica * Arborophila torqueola ∞ * Rollulus rouloul ∞ * ∞ Xenoperdix udzungwensis Leipoa ocellata Alectura lathami Megapodiidae * ∞ ∞ * ∞ Megapodius sp. Figure 2.5. Maximum likelihood tree obtained from combined four mitochondrial (Cytochrome-b – CYTB, NADH Dehydrogenase Subunit 2 - ND2, 12S Ribosomal DNA - 12S, Control Region - CR) DNA markers. Symbols above branches (*) represent bootstrap support values (only ≥ 70% are presented) and symbols below branches (∞) represent the Bayesian posterior probabilities (only ≥ 0.95 are presented) extracted from the 50% majority-rule consensus tree. 76 P h a s i a n i d a e Crax sp. * * ∞ ∞ Al. lathami Megapodius sp. Megapodiidae Pauxi pauxi Cracidae Ortalis sp. Numida meleagris Numididae *∞ Acryllium vulturinum Ptilopachus petrosus * Ptilopachus nahani ∞ Colinus sp. Odontophoridae * * Callipepla callfornica ∞ ∞ Oreortyx pictus ∞ Cyrtonyx montezumae Coturnix coturnix Pternistis hatlaubi Pternistis capensis S *∞ Pternistis adspersus p Pternistis natalensis u * * Pternistis griseostriatus ∞ r ∞ Pternistis squamatus f ∞* Pternistis camerunensis o Pternistis bicalcaratus w ∞ Pternistis leucoscepus l * * ∞ * Pternistis swainsonii s * ∞ ∞ Pternistis clappertoni Pternistis afer Alectoris chukar *∞ * ∞ Alectoris rufa Xenoperdix udzungwensis F Scleroptila shelleyi r *∞ * Scleroptila levaillantoides a afra * ∞ Scleroptila n ∞ Scleroptila streptophora c Scleroptila levaillantii Phasianidae o Peliperdix coqui l Dendroperdix sephaena * i ∞ * Francolinus pondicerianus *∞ ∞ n Francolinus lathami s Bambusicola thoracica * Gallus gallus ∞ Ithaginis cruentus Perdix perdix Pucrasia macrolopha Syrmaticus sp. *∞ Crossoptilon crossoptilon Catreus wallichii ∞* Tympanuchus phasianellus * ∞ Falcipennis canadensis Lophophorus impejanus * Meleagris gallopavo ∞ Polyplectron bicalcaratum * Polyplectron emphanum ∞ Pavo cristatus * Afropavo congensis ∞ Rollulus rouloul * ∞ * ∞ * ∞ Figure 2.6. Maximum likelihood tree obtained from combined three nuclear (Ovomucoid intron G – OVOG, Transforming Growth Factor Beta 2 intron 5 – TGFB, GAPDH intron 11 - GAPDH) DNA markers. Symbols above branches (*) represent boostrap support values (only ≥ 70% are presented) and symbols below branches (∞) represent the Bayesian posterior probabilities (only ≥ 0.95 are presented as extracted from the 50% majority-rule consensus tree. 77 37.4 9.3 46.5 34.2 B 44.5 34.2 30.1 32.1 40.9 37.6 23.1 65.0 35.2 8.7 25.2 33.6 36.3 7.2 7.6 A 31.6 C 19.5 13.7 56.8 11.9 5.4 Numida meleagris Numididae Guttera pucherani Acryllium vulturinum Cyrtonyx montezumae Callipepla californica Callipepla gambelii OdontopColinus sp. Oreortyx pictus horidae Ptilopachus nahani Ptilopachus petrosus Arborophila javanica Arborophila torqueola Rollulus rouloul Xenoperdix udzungwensis Ithaginis cruentus Tympanuchus phasianellus Tetrao urogallus Falcipennis canadensis Bonasa umbellus Meleagris gallopavo Tragopan temminckii Pucrasia macrolopha Perdix perdix Syrmaticus sp. Phasianus colchicus Crossoptilon crossoptilon Catreus wallichii Lophura nycthemera Chrysolophus sp. Lophophorus impejanus Afropavo congensis Pavo cristatus Alectoris chukar Alectoris graeca Alectoris rufa Pternistis capensis Pternistis hildebrandti Pternistis natalensis Pternistis adspersus P Pternistis clappertoni S Pternistis jacksoni h p Pternistis griseostriatus a Pternistis hartlaubi u Pternistis bicalcaratus s Pternistis icterorhynchus r i Pternistis squamatus f a Pternistis camerunensis o Pternistis nobilis n Pternistis erkelii w Pternistis ochropectus i l Pternistis castaneicollis d Pternistis swierstrai s a Pternistis afer Pternistis rufopictus e Pternistis leucoscepus Pternistis swainsonii Coturnix coturnix Coturnix japonica Margaroperdix madagarensis Francolinus lathami F Francolinus francolinus r Francolinus pictus Francolinus gularis a Francolinus pondicerianus n Dendroperdix sephaena Peliperdix coqui c Scleroptila finschi o Scleroptila psilolaema Scleroptila afra l Scleroptila levaillantoides i Scleroptila shelleyi Scleroptila levaillantii n Scleroptila streptophora s Gallus gallus Gallus varius Bambusicola thoracica Polyplectron bicalcaratum Polyplectron emphanum Alectura lathami Leipoa ocellata Megapodiidae Megapodius sp. Megapodiidae Crax sp. Pauxi pauxi Cracidae Ortalis sp. Figure 2.7. The Bayesian reconstruction of divergence dates with uncorrelated lognormal molecular clock. Numbers (only at critical nodes) at nodes are divergence dates estimated with three calibration points, (A) a fossil of Crested Francolin Dendroperdix sephaena at 4.5-5.0 mya (Crowe 1992), (B) a basal date for the Tetraoninae (Grouse and allies) of 27-29 mya (Crowe and Short 1992), and (C) a basal date for Polyplectron (Peacock-Pheasants) of 34-36 mya (Olson 1974, modified by T. M. Crowe unpubl. data). 78 a b c d e Figure 2.8. Sonograms of the call of (a) Stone Partridge Ptilopachus petrosus, (b) Nahan’s Francolin Francolinus nahani, (c) Scaly Francolin Pternistis squamatus, (d) Ahanta Francolin Pternistis ahantensis and (e) Red-necked Spurfowl Pternistis afer. Frequency (kHz) on the vertical axis with time (seconds) on the horizontal axis. 79 ,,) (b) Figure 2.9. Line drawing to show the posture of (a) Ptilopachus petrosus and (b) Francolinus nahani, after photographs by Callan Cohen, Ron Hoff and Nik Borro. 80 CHAPTER 3 A study of gross morphological and histological syringeal features of true francolins (Galliformes: Francolinus, Ortygornis, Afrocolinus, Peliperdix and Scleroptila spp.) and spurfowls (Pternistis spp.) in a phylogenetic context Abstract Modern taxonomies of francolins recognize 41 congeneric species, forming the largest genus of terrestrial gamebirds (Galliformes). Recent molecular, ecological and behavioural studies challenge this view, suggesting that they comprise two unrelated, monophyletic groups. There are 'true' francolins (Francolinus, Ortygornis, Afrocolinus, Peliperdix and Scleroptila spp.) that are relatively small, ground-roosting birds (with the exception of Ortygornis sephaena which roosts in trees), and spurfowls (Pternistis spp.) that are large birds that can roost in trees. This study explores gross morphological and histological syringeal anatomy of francolins, spurfowls and their respective sister taxa to test whether differences are concordant with a molecular-based hypothesis. Differences found were the presence of a shield- versus diamond-shaped tympanum among francolins and spurfowls, respectively. The first bronchial half rings are mineralized among francolins except in O. sephaena, whereas almost no mineral deposition was observed among spurfowls. 81 Histologically, francolins have a small, rounded pessulus (except in O. sephaena, which has a rounded, larger pessulus) in contrast to the larger pessulus observed among spurfowls, which is rounded and triangular in Pternistis capensis and P. natalensis. Both gross and histological similarities within, and differences between, francolin and spurfowl syringes support this division. However, O. sephaena shows intermediate features between francolins and spurfowls. 82 Introduction Syringeal characters in taxonomy and systematic studies The syrinx or avian voice box is located in the base of the neck, at the junction of the trachea and bronchi (Ames 1971, Lewis 1983, Seller 1983, Gill 1990). It can be formed from either the tracheal or bronchial tissues, or both. Myers (1917) categorized three types of syrinx based on its location relative to the trachea and the bronchi: ‘syrinx trachealis’ if it is found at the lower end of the trachea, ‘syrinx bronchialis’ in the case where the syrinx is located below the bifurcation, and ‘syrinx trachea-bronchialis’ when the syrinx is located at a position that includes both the lower end of the trachea and the upper parts of the bronchi. The syringeal structure of songbirds (Passeriformes) has been widely compared with that of non-songbirds (Frank et al. 2006), with passerine birds known to produce complex vocalizations as opposed to the relatively simple vocalizations given by many non-passerine birds. Whether the complexity of the vocalizations is dependent on the number or the complexity of the syringeal components is a question that requires detailed functional investigation of each part. This is because certain non-passerine species, e.g. cockatiels Nymphicus hollandicus, which are known to have only three pairs of syringeal muscles and two pairs of tracheal muscles (Larsen and Goller 2002), can mimic many types of sound (Tsukahara et al. 2008). The presence or absence of certain components of the syringes or their musculature, contribute largely to voice production, and has been found to play a significant role in the classification of birds. This is supported by evidence in studies of the syringes of the Red jungle-fowl (Chicken) Gallus gallus, the male Mallard Duck Anas platyrhynchos and the Greater 83 Sage-Grouse Centrocercus urophasianus (Myers 1917, Frank et al. 2006, Krakauer et al. 2009, respectively). The syrinx is anatomically complex and interspecifically diverse even in species that lack special structures. Syringeal morphology has been found to be informative in many systematic studies on passerine birds, e.g. Tyrannidae (Lanyon 1986, Prum and Lanyon 1989, Mobley and Prum 1995), Pipridae (Prum 1992) and Furnariidae (Zimmer et al. 2008), as well as in non-passerines, e.g. Anatidae (Delacour and Mayr 1945, Humphrey 1955, Johnsgard 1961, Livezey 1986), Charadriiformes (Brown and Ward 1990), Falconidae (Griffiths 1994a, 1994b) and Psitacidae (Gaban-Lima and Höfling 2006). As an example of the utility of syringeal characters in non-passerines, Livezey (1986) in his phylogenetic analysis of Anseriformes studied tracheal characters and found several important synapomorphies with which to deliniate clades. Because the demonstrated utility of syringeal characters in studies of phylogenetic relationships among non-passeriform taxa and in the spirit of using multiple lines of evidence to test hypotheses of monophyly (Templeton 1989, Farais et al. 2000, Pruett and Winker 2010), this study explored the gross anatomy and histology of francolin and spurfowl syringes, and the concordance of this feature with phylogenetic relationships supported in previous studies on this group (Crowe and Crowe 1985, Milstein and Wolff 1987, Crowe et al. 1992, Bloomer and Crowe 1998, Crowe et al. 2006, Cohen et al. 2012, Chapter 2). This is, to my knowledge, the first truly comparative paper of syringeal morphology in francolins and spurfowls. Data on a number of lesser-known species in a group that tends to be dominated by data from a few domesticated or managed northtemperate species is presented 84 Taxonomy, distribution, ecology and morphology Based on traditional morphological research, the 41 currently recognized species of francolins are placed within a single genus Francolinus Stephens, 1819 within the order Galliformes and family Phasianidae (Hall 1963). They are distributed throughout subSaharan Africa (with an isolated population of one species Pternistis bicalcaratus occurring in Morocco), the Middle East and Asia (Johnsgard 1988, Madge and McGowan 2002), and are adapted to a variety of habitats, comprised primarily of tropical and unforested vegetation types (McGowan 1994). All francolins have 14 tail feathers and most species are sexually monomorphic in plumage, with males having single- or double-spurred tarsi (Hall 1963, Johnsgard 1988). However, relatively recent morphological, eco-ethological and molecular studies of francolins (Crowe and Crowe 1985, Milstein and Wolff 1987, Crowe et al. 1992, Bloomer and Crowe 1998, Crowe et al. 2006, Chapter 2) have suggested that they form two distantly related lineages: the ‘true’ francolins and spurfowls (Fig. 3.1). ‘True’ francolins (allocated to five genera: Francolinus, Ortygornis (details for taxonomic designation are in Chapter 5), Afrocolinus gen. nov. (details for taxonomic designation are in Chapter 5), Peliperdix and Scleroptila) are relatively small, ground-roosting birds with striped and barred rufous dorsal plumage resembling that of quails Coturnix spp., with only males possessing relatively small tarsal spurs. Spurfowls are placed within a single genus Pternistis and are generally larger, often roost in trees, and have dark dorsal plumage usually vermiculated with white or buff, and both sexes usually have spurred (often two) tarsi that are much longer in males. Furthermore, spurfowls generally emit atonal, raucous, grating advertisement calls (given at dawn and dusk), whereas francolins have more tonal, often whistling, calls (Milstein and Wolff 1987). 85 This distinction between the two assemblages becomes blurred owing to O. sephaena, the Creasted Francolin which, like other francolins, has quail-like plumage but, like spurfowls has long tarsal spurs, roosts in trees and has an advertisement call with both grating and tonal elements (Milstein and Wolff 1987). It was on the basis of this ‘linking form’ that Hall (1963) decided to place all 41 species into a single genus, Francolinus, which DNA-based data no longer supports (Fig. 3.1). In this context, the aim of this study was to examine the gross morphological and histological anatomical structure of the syringes of selected francolins, spurfowls and their putative sister taxa, as well as to determine whether there are any syringeal characters that could be taxonomically and phylogenetically informative in investigating the proposed diphyletic status of the ‘true’ francolins and spurfowls. Materials and methods Syringeal sampling (gross morphology and anatomy) Syringes were examined from species that are representative of francolins and spurfowls and those that are from their closest extant relatives (Table 3.1): Common Quail Coturnix coturnix and Chukar Partridge Alectoris chukar (sister to the spurfowls), G. gallus (sister to the francolins), and Helmeted Guineafowl Numida meleagris, a distant relative to chickens, quails, partridges, francolins and spurfowls (Chapter 2). It is appropriate at this point to highlight some of the differences between the syringes of male and female birds. As explained by Frank et al. (2007), the syrinx of male and female birds may have morphological differences that in turn could alter the properties of the voices of each sex. However, Appel (1929) did not find differences between the syringes of males and females in his research on the chicken Gallus gallus, 86 a galliform he studied intensively, even in ovariectomised females. Syringes from immature birds were avoided in order to ensure that differences found between syringes were not because of age differences between individuals. As Hogg (1982) described for Galliformes, mineralisation begins well before maturity is reached and virtually achieves its final extent during the growing period, so all adult birds are considered to have a fully developed syrinx. Of course, it would be ideal to confirm our observations by analysing more individuals of each of the species and in particular a developmental sequence of one or more taxa, but this comparative material is at present not available. Thus, the difficulty of acquiring multiple syringeal samples (represented by male and female individuals of a particular species) and the inability to physically sex some of the specimens were major constraints. Analyses All syringes examined for gross morphological and anatomical purposes were dissected from frozen whole bird specimens with the exception of three syringes that were dissected from alcohol-preserved whole specimens provided by the Ditsong National Museum of Natural History, South Africa. The syringes were immersed in 70% ethanol until such time that they were processed. Gross syrynx morphology The double-staining protocol in Cannell (1988) for the clearing and staining procedures of the syringes was followed. Ossified tissues stained red with Alizarin red, cartilaginous tissues stained blue with Alcian blue, and muscles stained brownish in Lugol solution. Cleared and stained syringes were examined using a Leica S8APO 87 stereomicroscope and the LAS EZ version 1.7.0 software. Syringes of large birds such as Gallus gallus, Numida meleagris, Pternistis capensis, P. swainsonii and P. natalensis were stained for a few additional days relative to those excised from the smaller francolins. Histology For the tissue-sectioning procedure, the syringes were transferred from 70% ethanol to 10% buffered formalin overnight and then taken through a series of 70%, 90% and absolute ethanol washes before being cleared in xylene (1 h in each solution). They were then embedded in paraffin wax and longitudinal sections of 5 μm thickness were cut. Two stains were used, haematoxylin and eosin (Bancroft and Gamble 2002) for more general biological examination, and a more specialised multiple stain, orceinpicroindigocarmine (Steven et al. 2000). The orcein-picroindigocarmine stain is appropriate for staining bird syringes since it has the potential to differentiate tissues and their components. Tissue sections were stained with orcein, indigocarmine and picric acid following the Steven et al. (2000) protocol. The sections were deparaffinised and placed in distilled water, stained with orcein solution for 45 min at room temperature and rinsed in distilled water. Sections were then differentiated in 96% ethyl alcohol (two times) for 30 s each, rinsed in distilled water and stained with picroindigocarmine solution for 30 min. Slides were drained and differentiated in 70% ethyl alcohol for 2 min and finally mounted with Entellan and examined using a Nikon Stereoscopic Zoom microscope (NIS-Elements version 2.10). The sectioning of tissues was more difficult for the larger birds examined and this was a contributing factor in the distortion of those sections. Finally, even though 88 the staining of whole syringes and the sectioning of samples of larger syringes was challenging, most of the stained syringes and sections were intact and clear with the exceptions of Figure 3.5d and 3.6f. Results Syringeal gross morphology The basic structure of the syringeal morphology (Fig. 3.2) of a typical francolin/spurfowl is characterised by cartilaginous rings (which stained blue with Alcian blue), a very distinctive mineralized tympanum (which stained red with Alizarin red) and muscles (which stained brownish in Lugol solution). All the species analysed have a typical ‘tracheobronchialis’ type of syrinx (Fig. 3.3-3.5), which means that both the tracheal and the bronchial tissues are involved in the formation of the syrinx and thus possibly in shaping the structure of their vocalizations. Two significant gross morphological differences between true francolins and spurfowls were observed. Firstly, unlike their putative sister taxon G. gallus, which has a mineralized triangular-shaped tympanum (Fig 3.4a), francolins (Fig. 3.4b–f) and N. meleagris (Fig. 3.3) have a shield-like calcified tympanum clearly visible from the ventral perspective (Fig. 3.3). Numida meleagris, G. gallus and francolins show extensive mineral deposition in the tympanum in both ventral and dorsal views but Coqui Francolin Peliperdix coqui and Red-winged Francolin Scleroptila levaillantii (Fig. 3.4c and d, respectively) show reduced mineral deposition on the dorsal side compared with the ventral side. In spurfowls, as in their putative sister taxa A. chukar and C. coturnix (Fig. 3.5a and b, respectively), the tympanum is diamond-shaped with the degree of calcification 89 not as extensive as in francolins and relatives (Fig. 3.5c–e). Secondly, on the ventral side, the first bronchial half rings are well mineralized in G. gallus (Fig. 3.4a) and francolins (Fig. 3.4c, d and f), but not so in O. sephaena (Fig. 3.4b), spurfowls (Fig. 3.5c–e) and their relatives, A. chukar and C. coturnix (Fig. 3.5a and b). A further observation is that some species such as S. levaillantoides, G. gallus and N. meleagris show mineral deposition in some tracheal rings. Histology The histological structure of the syringes of francolins, spurfowls and their near relatives are generally similar, with only a few small differences observed (Fig. 3.6). The pessulus, which is present in all the species examined (Fig. 3.6a–h), varies considerably in size and shape. It is generally small and rounded in the two true francolins, i.e. S. levaillantii and Grey-winged Francolin S. afra, and N. meleagris (Fig. 3.6a, b and h, respectively) (Table 3.3). The pessulus is large and also rounded in O. sephaena (Fig. 3.6c) and in spurfowls even though it is almost triangular in Natal Spurfowl Pternistis natalensis (Fig. 3.6f) and C. coturnix (Fig. 3.6g). It is markedly large and triangular in G. gallus (Fig. 3.6d, see also Myers 1917). A distinctive interclavicular air sac, which is bound by the internal tympaniform membrane, is found in the two francolins (S. levaillantii and S. afra), Cape Spurfowl Pternistis capensis, C. coturnix and N. meleagris (Fig. 3.6a, b, e, g and h, respectively) and is absent in O. sephaena and G. gallus (Fig. 3.6c and d, respectively), while its presence could not be determined in P. natalensis (Fig. 3.6f). It was observed that the inner wall of the interclavicular air sac and sometimes the medial bronchial wall are lined with a layer of connective tissue that differs in thickness. Connective tissue fills 90 spaces and provides support to organs. The francolins S. levaillantii and S. afra (Fig. 3.6a and b) and their relative G. gallus (Fig 3.6d) have thin connective tissue lining the walls of the membranes that tends to be moderate in P. natalensis, C. coturnix and N. meleagris (Fig 3.6f, g and h, respectively). Numida meleagris has connective lining that runs along the medial bronchial walls. Pternistis capensis (Fig 3.6e), as in O. sephaena (Fig 3.6c), has a large amount of connective tissue that is restricted to the far corners of the internal tympaniform membrane. This tissue pushes against the internal tympaniform membrane forcing it to expand and hence results in the narrowing of the surrounding airway. The external tympaniform membrane, which is found in all the species examined, differs remarkably in length. This structure is extremely long in G. gallus (Fig. 3.6d) and much shorter in the other species (Fig. 3.6a–c and e–h). Discussion Although a limited number of species were sampled, there are some marked structural differences in the syrinx that distinguish francolins from spurfowls. The gross morphology of the syrinx is generally consistent and supports the split of francolins and spurfowls into two independent clades. The shape of the tympanum places O. sephaena decisively with other francolins. However, O. sephaena, as in the two spurfowls P. capensis and P. natalensis and their evolutionary relatives C. coturnix and A. chukar, shows no mineral deposition in their first bronchial half rings with P. swainsonii showing short first bronchial half rings with very little mineralization. Thus, the aspect of mineral deposition is not fully consistent among francolins and spurfowls. Our observation on the degree of mineralisation in tracheal rings of S. levaillantoides, G. Gallus and N. meleagris could not at this stage translate to any coherent conclusion 91 apart from the fact that a similar observation was made by Hogg (1982), thought this could have to do with conferring rigidity in the tracheal rings, which is an adaptation to vocalization. With regard to the histology of the syringes, the features that separate francolins from spurfowls are based on the size of the pessulus. There is variation in the shape of the pessulus, the presence/absence of the defined interclavicular air sac and the amount of connective tissue such that differences are observed even between the two spurfowl species. However, the amount of connective tissue puts species with tonal and whistling calls together (S. levaillantii and S. afra), separated from those that have atonal, raucous and grating calls (P. capensis, P. natalensis and O. sephaena). This feature could be related to differences in sounds between them, given that whistle-like sounds could be generated by shearing forces as a column of air is forced through a narrow aperture (see Gaunt et al. 1982), and the connective tissue could exert pressure to modify the lumen of the syrinx, which could be considered the column through which air passes, generating the sound. At first glance, this explanation would be contrary to what we would expect, given that spurfowls have thick connective tissue that would narrow their syrinx, but they have raucous calls. However, we know the syrinx is only one component in a vocal system (Gaunt et al. 1982) and other factors may be modulating the final sound we hear. Conclusions A number of gross morphological and histological features were identified as differentiating francolins from spurfowls (though with O. sephaena being a possible exception). What emerged as an area for future research was investigation into the role 92 of each part of the syrinx in shaping francolin and spurfowl vocalizations, as well as the need to include other galliform species in addition to the sampled francolin and spurfowl taxa. This will require a reasonable number of individuals to be sampled per species, such that any presence or absence of intra- or inter-specific variation can be determined. Finally, it could be concluded that the outcome of this work points to the distinction between francolin and spurfowl assemblages with O. sephaena (as indicated in Hall 1963) still presenting some difficulties owing to its possession of characteristic features that are typical of both francolins and spurfowls. 93 Tables and Figures Table 3.1. List of species for which syrinxes were analyzed. Gross morphology Histology Common name Scientific name Sample name1 Sex2 Age Scientific name Sample Name1 Sex2 Age Grey-winged Francolin Orange River Francolin Red-winged Francolin Crested Francolin Coqui Francolin Cape Spurfowl Swainson's Spurfowl Natal Spurfowl Common Quail Chukar Partridge Helmeted Guineafowl Chicken/Red Jungle Fowl Scleroptila afra S. levaillantoides S. levaillantii Ortygornis sephaena Peliperdix coqui Pternistis capensis P. swainsonii P. natalensis Coturnix coturnix Alectoris chukar Numida meleagris TMC67 TMC65 TMC60 TM78245 TM75627 TMC70 TMC48 TM60042 TMC50 TM74489 TMC45 M F F M F M M ? F ? M Adult Adult Adult Adult Adult Adult Adult Adult Adult ? Adult S. afra TMC01 F Adult S. levaillantii O. sephaena TMC15 TMC13 F M Adult Adult P. capensis TMC07 M Adult P. natalensis C. coturnix TMC20 TMC14 M F Adult Adult N. meleagris TMC17 F Adult Gallus gallus TMC30 ? Adult G. gallus TMC02 ? Adult 1 TM, Transvaal Museum; TMC, Timothy M Crowe (Percy FitzPatrick Institute of African Ornithology, (University of Cape Town). 2 ? = Unknown. Table 3.2. Comparison of gross morphological features of the syringes. Taxon Tympanum shape Bronchial half ring 1 mineralization Tracheal ring mineralization S. afra S. levaillantoides S. levaillantii O. sephaena P. coqui P. capensis P. swainsonii P. natalensis C. coturnix A. chukar N. meleagris G. gallus Shield-like Shield-like Shield-like Shield-like Shield-like Diamond Diamond Diamond Diamond Diamond Shield-like Triangular Mineralized Mineralized Mineralized Non-mineralized Mineralized Non-mineralized Mineralized Non-mineralized Non-mineralized Non-mineralized Mineralized Mineralized Non-mineralized Mineralized Non-mineralized Non-mineralized Non-mineralized Non-mineralized Non-mineralized Non-mineralized Non-mineralized Non-mineralized Mineralized Mineralized 94 Table 3.3. Comparison of the histological features of syringes. Taxon Pessulus shape & size Interclavicular air sac Amount of connective tissue S. afra S. levaillantii O. sephaena P. capensis P. natalensis C. coturnix N. meleagris G. gallus Rounded, small Rounded, small Rounded, larger Rounded, larger Triangular, larger Triangular, larger Rounded, small Triangular, larger Bound by internal tympaniform membrane Bound by internal tympaniform membrane Absent Bound by internal tympaniform membrane ? Bound by internal tympaniform membrane Bound by internal tympaniform membrane Absent Thin Thin Thick Thick Moderate Thin Thin Thin External tympaniform membrane Shorter Shorter Shorter Shorter Shorter Shorter Shorter Long ? = Presence or absence of feature cannot be determined from distorted tissue section. 95 Figure 3.1. Strict consensus cladogram for some 'true' francolins (Ortygornis, Peliperdix and Scleroptila spp.) and spurfowls (Pternistis spp.) modified from Crowe et al. (2006). Numbers at nodes are jackknife support values. 96 Figure 3.2. Example of a ventral view of the syrinx of a typical Coqui Francolin Peliperdix coqui with illustrations of the various parts. Labels: tracheo-lateralis muscle (tlm), tracheal cartilaginous ring (tcr), membrana-trachealis (mt), sterno-trachealis muscle (stm), tympanum (t), external tympaniform membrane (etm), bronchial half ring 1 (bhr1), membrane-interanularis (mi), bronchus cartilaginous ring (bcr). Red colour mineralized cartilage, blue - cartilage, transparent light brown - muscle. Figure 3.3. Ventral (V) and dorsal (D) views o f the syrinx of Helmeted Guineafowl Numida meleagris. Red colour - mineralized cartilage, blue - cartilage, transparent light brown - muscle. 97 Figure 3.4. Ventral (V) and dorsal (D) views of the syringes of: a. Gallus gallus, b. Ortygornis sephaena, c. Peliperdix coqui, d. Scleroptila levaillantii, e. Orange River Francolin S. levaillantoides and f. S. afra. Red colour - mineralized cartilage, blue cartilage, transparent light brown - muscle. 98 Figure 3.4 (concl.). Ventral (V) and dorsal (D) views of the syringes of: a. Gallus domesticus, b. Ortygornis sephaena, c. Peliperdix coqui, d. Scleroptila levaillantii, e. Orange River Francolin S. levaillantoides and f. S. afra. Red colour - mineralized cartilage, blue - cartilage, transparent light brown - muscle. 99 Figure 3.5. Ventral (V) and dorsal (D) views of the syringes of: a. Alectoris chukar, b. Coturnix coturnix, c. Pternistis natalensis, d. P. capensis, e. Swainson’s Spurfowl P. swainsonii. Red colour - mineralized cartilage, blue - cartilage, transparent light brown muscle. 100 Figure 3.5 (concl.). Ventral (V) and dorsal (D) views of the syringes of: a. Alectoris chukar, b. Coturnix coturnix, c. Pternistis natalensis, d. P. capensis, e. Swainson’s Spurfowl P. swainsonii. Red colour - mineralized cartilage, blue - cartilage, transparent light brown - muscle. 101 Figure 3.6. Histological structure of the syringes of francolins (a. S. levaillantii, b. S. afra and c. O. sephaena), d. Chicken G. gallus, spurfowls (e. P. capensis and f. P. natalensis), g. Common Quail C. coturnix and h. Helmeted Guineafowl N. meleagris. References: tracheo-lateralis muscle (tlm), lumen of trachea (lt), interclavicular air sac (ias), connective tissue (ct), pessulus (p), internal tympaniform membrane (itm), external tympaniform membrane (etm), medial bronchial wall (mbw), bronchiodesmus (bd), bronchial lumen (bl), sterno-trachealis muscle (stm). 102 Figure 3.6 (concl.). Histological structure of the syringes of francolins (a. S. levaillantii, b. S. afra and c. O. sephaena), d. Chicken G. gallus, spurfowls (e. P. capensis and f. P. natalensis), g. Common Quail C. coturnix and h. helmeted guineafowl N. meleagris. Labels: tracheo-lateralis muscle (tlm), lumen of trachea (lt), interclavicular air sac (ias), connective tissue (ct), pessulus (p), internal tympaniform membrane (itm), external tympaniform membrane (etm), medial bronchial wall (mbw), bronchiodesmus (bd), bronchial lumen (bl), sterno-trachealis muscle (stm). 103 CHAPTER 4 Taxonomic and phylogenetic utility of variation in advertising calls of francolins and spurfowls (Galliformes: Phasianidae) Abstract Systematists have not often made use of avian vocalizations to assess the taxonomic rank of birds, or to infer their phylogenetic relationships. The likely reasons for this stem from the perceived inability to distinguish genetic and ecological components of variation in vocalizations, the difficulty in detecting homology across taxa, as well as the diverse selection pressures acting on vocal characters which may make such characters particularly prone to convergent evolution. In this study, we scored and analysed DNA and vocal characters of two delineated assemblages of gamebirds, francolins and spurfowls. Our phylogenetic results suggest that short strophes evolved from longer strophes among taxa within the genera Scleroptila and Peliperdix. More generally, our results corroborate the francolin-spurfowl dichotomy, with francolin calls generally being long and tonal, containing a series of discrete elements that have detectable harmonics. In contrast, most spurfowls render short, atonal calls with elements that generally have no harmonics, although they may contain discrete elements. Phylogenetically, Ortygornis sephaena is placed decisively with ‘true’ francolins and its closest relatives are the two phylogenetically enigmatic Asian 104 francolins, the Grey Francolin Ortygornis pondicerianus and Swamp Francolin O. gularis. 105 Introduction Vocalizations are a means of communication and are often of special significance in the social behaviour of animals, and birds in particular. Many birds rely on acoustic signals to convey information to each other over long distances, in dense cover, and at night (Marler 1969, Gill 1990). Systematists have not often made use of vocalizations to assess the taxonomic rank of birds, or to infer their phylogenetic relationships. Possible reasons for this may stem from the perceived inability to distinguish genetic and ecological components of vocalizations (McCracken and Sheldon 1997), as well as the difficulty in detecting homology across taxa (Lanyon 1969). Further, McCracken and Sheldon (1997) maintain that the diverse selection pressures acting on vocal characters may make such characters particularly prone to convergent evolution. For example, species that live in dense vegetation tend to have vocalizations with lower frequencies and narrower frequency ranges than those that inhabit open habitats (Morton 1975). This is because longer wavelengths propagate energy more efficiently through vegetation than shorter wavelengths, which attenuate due to the scattering effects of the vegetation. In addition, several bird species from diverse clades have repertoires of many distinct song types, whereas others have just one simple and stereotyped song or call (Price and Lanyon 2002, Lei et al. 2005), and some species learn the songs of other species, often making it difficult to distinguish homologous components from learned songs (Lei et al. 2005). Other confounding factors include the possibility that parts of the song may change in response to different seasons (Lei et al. 2005, Aubrecht and Holzer 2000, Frank et al. 2007), as well as morphological constraints of the syrinx (McCracken and Sheldon 1997). Finally, bird song usually plays a strong role in mate recognition, and is thus under strong sexual 106 selection, and thereby adapted to the environment in which signalling is taking place (Price and Lanyon 2002). Despite the above caveats, over the past two decades, authors have started to place greater emphasis on exploring the systematic utility of vocal character variation (Enggist-Düblin and Birkhead 1992, Alström 2001, Price and Lanyon 2002, Päckert et al. 2003, Lei et al. 2005, Farnsworth and Lovette 2008). Non-singing gamebirds such as francolins belonging to the genera Francolinus Stephens, 1819, Ortygornis Reichenbach, 1852 (see Chapter 5 for details in taxonomic designations), Afrocolinus gen. nov. (Mandiwana-Neudani, this thesis) (see Chapter 5 for details in taxonomic designations), Peliperdix Bonaparte, 1856 and Scleroptila Blyth, 1849, as well as spurfowls Pternistis Wagler, 1832, are known to have a few stereotyped call types making up their repertoires (van Niekerk 2010). Furthermore, francolins and spurfowls appear to not be able to learn the calls of other related species (Milstein and Wolff 1987). Here we explore the taxonomic and phylogenetic utility of francolin and spurfowl ‘advertisement’ calls, that is those calls usually heard at dawn and dusk, primarily during the breeding season. Systematics Genus Francolinus Stephens, 1819 Hall’s (1963) monograph of ‘francolins’ sensu lato placed 41 traditionally recognized species into a single genus Francolinus, with 37 species partitioned into eight putative monophyletic groups of related species and four unplaced species (Hall 1963, Table 1.1). Seven groups have representatives in Africa south of the Sahara, and one group is confined to the Middle East and Asia. The current phylogenetic treatment divides francolins into two distantly related groups of phasianine Galliforms: ‘francolins’ 107 (shared among genera Francolinus, Ortygornis, Afrocolinus, Scleroptila, Peliperdix) and ‘spurfowls’, all classified in one genus Pternistis (Milstein and Wolff 1987, Crowe et al. 1992, Crowe et al. 2006, Mandiwana-Neudani, this thesis). Among attributes that distinguish francolins from spurfowls, is the difference in the acoustic properties of their advertising calls. Aurally, most spurfowl species have raucous grating or cackling calls whereas francolins tend to have more musical and whistling calls (Crowe et al. 1986, Milstein and Wolff 1987, Crowe et al. 1992). Furthermore, spurfowls can perch in trees or other elevated structures and have scaly or striped underpart plumage, whereas francolins roost on the ground and generally have barred, blotched or uniform underparts and quail-like back plumage. However, the above distinction does not apply well to the Crested Francolin O. sephaena because it has attributes of both francolins and spurfowls. It is relatively small, but emits a somewhat raucous call and roosts in trees like many spurfowls. Thus, it is the ‘linking form’ that persuaded Hall (1963) to include all 41 species into a single genus Francolinus. In this study, auditory, visual and quantitative descriptions of the advertising calls of francolins and spurfowls are analysed in order to assess their phylogenetic utility, investigate the validity of the francolin-spurfowl dichotomy and to assess variation between putative related species. Materials and methods Collection of data Call sampling The calls were sourced from various institutional sound libraries and published collections of recordings (Table 4.1), although many were unfortunately of poor quality. 108 Sonograms posted online at Xeno-canto [http://www.xeno-canto.org/] were also examined, but were also mostly of poor quality. Frank et al. (2007) highlighted that the syrinx of male and female birds may differ morphologically, which could in turn determine the properties of their voices. However, given the difficulty in gathering multiple vocalizations (for male and female individuals of a particular species), lack of precise knowledge about the habitat from which the sourced call recordings were collected, and often lack of knowledge as to whether the recorder actually saw the bird in close proximity or not, we decided to analyse only the best quality calls and as a consequence we were not always able to account for sex and/or the specific habitat the bird was recorded in for all taxa. Phylogenetic sampling Molecular characters Twenty terminal taxa of francolins (including two outgroups - Gallus gallus and Bambusicola thoracica) and 24 spurfowl taxa (with two outgroups - Coturnix coturnix and Alectoris chukar) were sequenced for the mitochondrial Cytochrome-b gene (CYTB - 1143 base pairs- bp; Table 4.2). In contrast to the earlier work of Crowe et al. (1992) and Bloomer and Crowe (1998), which focused on relatively few species, almost all traditionally recognized species were included in our sampling. Total genomic DNA was extracted from tissue using the animal tissue protocol provided with the DNeasy tissue kit (Qiagen). Primers used in sequencing CYTB gene for fresh tissues and from museum toe-pads are detailed in Table 4.3. Double stranded DNA templates were amplified by polymerase chain reaction (PCR) using 0.75 units of BIOTAQTM DNA polymerase (Bioline) in 30 µl reactions. Reactions also contained 1 x NH4 buffer, 2.5 109 mM MgCl2, each dNTP at 0.1 mM and each primer at 0.3 µM, as well as 3 µl of template DNA. The thermal profile used comprised an initial denaturation step at 94°C for two minutes, followed by 30 cycles of 94°C for one minute, 52°C for one minute and 72°C for two minutes, with a final extension step of 72°C for seven minutes. PCR-amplified products were cleaned from solution or gel using the GFXTM PCR DNA and gel band purification kit (Amersham Biosciences) prior to cyclesequencing with the ABI PRISM Big DyeTM Terminator v3.1 cycle-sequencing Ready Reaction Kit (Applied Biosystems). Sequencing products were resolved on an ABI PRISM 3100 Genetic Analyser. Vocal characters Twenty terminal taxa of francolins (including two outgroups - Gallus gallus and Bambusicola thoracica) and 24 spurfowl taxa (with two outgroups - Coturnix coturnix and Alectoris chukar) were sampled and character matrices of vocal characters for the francolins (Table 4.4.) and the spurfowls (Table 4.5) were generated. All calls (in wave format) were analysed using a Linux-based software package ‘ANA’ (Jean-Pierre Richard, Unpublished version). Multiple recordings of the calls of most species were analysed to examine intraspecific variation and also to assess the quality of the calls for each species. The terminology used hereafter is: strophe, element and pause (following Alström et al. 2008) (Fig. 4.1). A ‘strophe’ is a call that may be repeated multiple times in sequence and is comprised of elements that are defined as parts within a strophe separated by a pause of duration of at least 0.01s. A ‘pause’ is an interval between the elements of a strophe. Strophe duration was measured from the start of the first element 110 to the end of the last element on oscillograms and not on sonograms, in order to maximise accuracy. Sonograms with different acoustic properties were produced. The strophe variables: strophe length; number of elements in a strophe; number of introductory elements in a strophe; nature of the strophe in terms of whether it contains a harmonic or trill; raucousness of advertisement; presence and absence of a warbling ending to the strophe; whether the strophe is a simple musical advertisement, short complex musical advertisement call, or a long complex musical call; whether the inter-element pause exists and is distinct or indistinct; the presence or absence of a cackle-trill ending; the presence or absence of ‘ka-waak/ko-rak' component; whether the pitch of the strophe ascends or descends as it progresses; and whether the strophe is antiphonal or not; were scored from the francolins and spurfowls vocal recordings of suitable quality (Table 4.4, 4.5). Analyses of data The generated nucleotide sequences were edited and assembled in the Staden Package (Staden et al. 2003) and aligned using MAFFT (Katoh et al. 2009). All CYTB gene sequences were checked for stop codons and insertions or deletions. Two phylogenetic inference methods with different optimality criteria were employed to generate phylogenetic hypotheses: Parsimony was used to construct phylogenetic trees for both CYTB and the vocal data matrices, and a maximum likelihood (ML) approach was used in analysing the CYTB data set. The francolin data matrices were rooted on Gallus gallus and Bambusicola thoracica and the spurfowl matrices were rooted with Coturnix coturnix and Alectoris chukar (see Crowe et al. 2006). Parsimony-based phylogenetic 111 analyses were conducted using PAUP* ver. 4.0b10 (Swofford 2002) under a full heuristic search with all characters unordered and with equal weight, starting tree(s) obtained via stepwise addition; tree-bisection-reconnection branch-swapping, and 1000 random addition replicates (Maddison 1991). When multiple, equally parsimonious cladograms were recovered, a strict consensus cladogram was constructed. The extent to which each non-terminal node is supported by different character partitions was determined by using the bootstrap (BS) (Felsenstein 1985) resampling with 1000 pseudoreplicates, and five random addition replicates of taxa per bootstrap pseudoreplicate. Mixed-model maximum likelihood analyses were performed using the Randomised Axelerated Maximum Likelihood algorithm for High Performance Computing (RAxML) v7.0.4 (Stamatakis 2006, Stamatakis et al. 2008) implemented on the CIPRES portal. Mixed-model RAxML analyses make use of a GTR++ model partitioned by gene or codon position. The following analyses were run: mixed-model mtDNA (one model for each codon position, and also as a single data partition). Support at nodes was assessed with 100 non-parametric bootstrap (BS) pseudoreplicates. PAUP* was used to determine three measures of phylogenetic signal by estimating: the consistency index, the retention index and the scaled retention index. This enabled us to directly compare the phylogenetic signal inherent to the mtDNA and vocal character matrices. Results Calls 112 Both structural and temporal characteristics distinguish the calls of francolins and spurfowls, as well as among assemblages of taxa within them. Structural characteristics involve: strophe duration, the number of elements per strophe, and use of harmonics, trills and cackles. Temporal attributes include the sequence of elements and the relative length of the pauses between them. Francolins Francolinus: The three members of the Spotted Francolin complex, F. francolinus, F. pictus and F. pintadeanus are characterised by having similarly structured strophes which are 2.0 s, 2.11 s and 1.62 s long, respectively. These strophes are comprised of distinctive elements: seven elements in F. francolinus, and five in both F. pintadeanus and F. pictus (Appendix 4.1). The strophe of F. francolinus is relatively tonal with the second element being higher-pitched. The first three elements of the strophe of F. pintadeanus are tonal with the second and third elements ending in trills. The third element is remarkably protracted whereas the fourth and fifth elements are fully trilled. Francolinus pictus has a less tonal strophe with very fuzzy trilling elements. These taxa also differ in that F. francolinus and F. pictus both have one introductory element followed by a series of further elements, whereas F. pintadeanus has two introductory elements. The pause between the introductory element/s and the subsequent elements is relatively long in F. pictus (0.58 s), 0.43 s in F. francolinus and 0.24 s in F. pintadeanus. Another difference is seen in the shape of the harmonics, which is rising in F. pintadeanus, but comprises a mainly overslurred fundamental 113 frequency in F. francolinus. The tonal part in the first element of F. pictus is also overslurred. Ortygornis: Ortygornis sephaena has a complicated strophe, which sounds like a ‘hybrid’ between a francolin and spurfowl strophe to the ear (Appendix 4.1). The advertisement strophe of O. sephaena is a high-pitched, squealing cooperative duet led by a male and followed by a female (Little and Crowe 2000, van Niekerk 2010), and is thought to be used in claiming a territory (van Niekerk 2010). Its strophe sounds similar to that of the two Asian francolins, O. pondicerianus and O. gularis. The global structure of the sonograms of these species looks similar despite the difference in the duration of the strophes. Ortygornis pondicerianus and O. gularis are among the four species that Hall (1963) did not place decisively in one of her eight species groups. The duration of the strophes of O. sephaena is much shorter (0.55 s) compared to that of the Kenyan O. grantii (0.74 s). The strophes of O. sephaena and O. grantii have four elements, one introductory element followed by three complex elements. There is interspecific variation that relates to the second element, which has two parts that are fused in O. grantii but encompasses a distinctive pause in O. sephaena. The first component of the second element has harmonics descending in shape in O. sephaena and almost linear in O. grantii, and trilling parts. The third element has complex variation, with the two parts either fused or separated. The fourth element is quite faint and has an overslurred harmonic-like structure. Thus, O. sephaena, O. grantii, O. gularis and O. pondicerianus have one introductory element followed by a pause, and thereafter two to three additional complex elements. The third element is almost broken into two components with descending harmonics. Ortygornis gularis has strophe duration of 0.70 s, which, 114 approaches that of O. grantii (0.74 s), with O. pondicerianus having the shortest duration (0.46 s) compared to O. sephaena and O. gularis. There is a much longer pause (0.32 s) between the first and the last sets of elements in O. gularis when compared to the other taxa: 0.14 s in O. grantii, 0.12 in O. sephaena, and 0.09 s in O. pondicerianus. Peliperdix: All the Peliperdix spp. (also including A. lathami) have strophes with an introductory element followed by a pause and then another five to seven elements that trail away in volume, and have tonal and trilling parts. The global structure of the sonograms remains similar despite differences in strophe duration, and the number and structure of elements. The strophes of Peliperdix spp. (Appendix 4.1) are relatively high pitched, harsh and tinny (Crowe et al. 1986) to the ear. Afrocolinus lathami being an exclusively forest species, has a low frequency ‘cooing’ strophe which suits the type of habitat it thrives in, but has the basic elemental structure similar to that of the Peliperdix spp. The major difference among the strophes of Peliperdix spp. is in their duration. The strophe of Peliperdix albogularis is short (1.0 s) consisting of six tonal elements with relatively stable harmonics with some parts of the elements trilled. The strophe of P. schlegelii is a little longer (1.89 s) and consists of seven elements, with element E1 more tonal and has clearly defined rising harmonics, element E2 is fully trilled, and the last four element (E 3-7) start with an overslurred component followed by trilling parts. The strophe of P. coqui (1.95 s) has eight elements, which can be split into two parts. The first part has stable harmonics followed by the second component which comprises trilling and descending harmonics. The strophe of A. lathami (1.89 s) is made up of eight very distinctive elements with E1 being more tonal followed by E 2-8 which begin with tonal parts and end in trills. There is also a remarkable difference in the duration of 115 the pause between E1 and the rest of the elements in the strophes. The inter-element pause (which narrows as the strophe progresses) is 0.10 s long in P. albogularis, 0.37 s in P. schlegelii, 0.11 s in A. lathami, and somewhat longer than that in P. coqui (0.25 s). There is a general decreasing trend in the duration of the inter-element pause in strophes within the group. Thus, the strophes of the Peliperdix species represent a stepped cline moving from the South African P. coqui to the Central African P. schlegelii, to the West African P. albogularis. Scleroptila: Unlike the Red-tailed Peliperdix francolins, which have harsh short duration strophes, members of the Red-winged Group have whistling strophes. Four species (S. levaillantoides, S. finschi, S. shelleyi and S. gutturalis) have strophes similar in structure (four elements) that may be rendered “I’ll drink YER-BEER” (Newman 2002) and differ primarily in strophe and pause duration (Appendix 4.1). The strophe of S. levaillantoides (0.53 s) is the shortest, with those of S. finschi (0.71 s), S. gutturalis (0.79 s), and S. shelleyi, (0.89 s), respectively increasing in duration. The remaining three species have strophes of much longer duration, (S. afra - 2.0 s, S. levaillantii 2.11 s, and S. streptophora - 2.22 s) and seven to ten elements (Appendix 4.1). The typical advertisement strophe of S. levaillantoides, S. gutturalis, S. finschi and S. shelleyi (also reported by Madge and McGowan 2002, Hockey et al. 2005 excluding S. gutturalis) are similar in that they have four complex elements with the first two elements introducing the strophes. Strophes of Scleroptila finschi have harmonics of complex shapes. The element following the two introductory elements is very complex and does not have markedly definable third and fourth elements as in S. levaillantoides and S. shelleyi. The first two elements of S. levaillantoides have rising harmonics, which descend in the second component of E2, E3 is trilled, whereas E4 116 begins with trilling parts and ends with descending harmonics. The overall strophe structure of S. shelleyi is similar with some minor variation in the structure of E3 and E4. Scleroptila streptophora, S. levaillantii and S. afra have longer strophes with several elements. Scleroptila streptophora has a dove-like cooing strophe, which is very tonal and has one introductory element followed by six additional elements. All the elements have well-defined harmonics, which are stable in the first three elements and descend in the fourth and fifth elements. The fourth through to the seventh elements make a warbling sound. Scleroptila levaillantii has a strophe (with overslurred harmonics) which is introduced by seven elements followed by three elements (E 8-10) which make a warbling sound, whereas the strophe of S. afra (with overslurred harmonics) is introduced by six elements followed by two elements (E 7 and 8), but includes no warbles. There is a difference in the duration of the pause between the introductory elements and the subsequent elements, which is quite long in S. streptophora (0.46 s), and shorter in the other taxa: 0.27 s in S. shelleyi, 0.21 s in S. afra, 0.18 s in S. finschi, 0.12 s in S. levaillantii, and 0.10 s in S. levaillantoides. Spurfowls - Pternistis species All four Bare-throated species have a basic atonal ‘ka-waak’ advertisement strophe (0.75 s for P. swainsonii, 0.36 s – P. afer, 0.57 s – P. rufopictus and P. leucoscepus) (Appendix 4.1) sounding higher-pitched and more guttural in P. swainsonii and lower-pitched and more nasal to the ear in P. afer. Furthermore, the strophes of P. leucoscepus and P. rufopictus are both high-pitched, with an element of screeching and more protracted trilling second elements. 117 The strophes of P. swainsonii, P. afer, P. leucoscepus and P. rufopictus are differentiated into two elements, which can be characterised as trills and are separated by a fuzzy pause (except in P. swainsonii). The first element is more tonal in P. swainsonii and trills in the other species, whereas E2 is completely trilled in P. swainsonii, P. leucoscepus and P. rufopictus (exceptionally starting with a slower trill and ending with a much faster trill). The inter-element pause is well-differentiated in P. swainsonii (0.16 s long) in contrast to the fuzzy pause in P. afer (0.04 s), P. rufopictus (0.08 s) and P. leucoscepus (0.09 s). There is a decreasing trend in the duration of elements and the pause between them from the northeast African P. leucoscepus through to the east African P. rufopictus to the southern African P. afer. Pternistis swainsonii has a much longer strophe. Like the Bare-throated spurfowls, the northern Vermiculated species P. icterorhynchus, P. bicalcaratus, P. clappertoni and P. harwoodi, all give basic ‘kawaak’ strophes that differ mainly in the duration over which they are rendered (0.28 s, 0.33 s, 0.64 s and 0.61 s, respectively; Appendix 4.1). The strophes of these species have two trilling elements, which are separated by a fuzzy pause that is 0.06 s, 0.04 s, 0.14 s and 0.08 s long in P. icterorhynchus, P. bicalcaratus, P. clappertoni, and P. harwoodi, respectively. Among the four species, P. clappertoni is the only species in which its strophe shows a much faster and loud trilling part at the beginning of the second element. The southern Vermiculated species, P. adspersus, P. capensis, P. hartlaubi, P. hildebrandti and P. natalensis give strophes which are extremely divergent from one another, with the only similarity being that P. capensis and P. adspersus have strophes ending in ‘cackle-trills’. The overall structure of the sonograms, the duration of the strophes (2.10 s, 4.43 s, 1.90 s, 0.34 s and 0.62 s, respectively), number of elements 118 (some with components) and the duration of the inter-element pause differ remarkably. Pternistis adspersus gives a strophe that is introduced by four elements followed by two ‘cackle-trill’ elements. Pternistis capensis renders the longest strophe within the group which is more tonal consisting of seven elements (the first five elements with increasing volume followed by a low-pitched ‘cackle-trill’). The first two elements have two components, followed by two three-component elements, one four-component element and ends in two five-component ‘cackle-trilled’ elements. The most complicated strophe is that of Hartlaub’s Spurfowl P. hartlaubi, which has a complex high-pitched antiphonal strophe often given by the male (Komen 1987; Maclean 1993). This strophe has tonal and less well-defined trilling parts. Pternistis hildebrandti in contrast, gives a trilling strophe with six elements that have no clearly defined pause. The advertisement strophe of P. natalensis is a high-pitched, fourelement strophe characterized by some tonal, slow and fast trilling parts. Among the members of the Scaly Group, the advertisement strophe of P. squamatus is a high-pitched nasal ‘ke-rak’ which is repeated several times and increases in volume as it progresses (Appendix 4.1). Pternistis ahantensis has a much shorter strophe (0.44 s) compared to P. squamatus and P. griseostriatus (1.09 s and 0.81 s, respectively). The Montane species, P. camerunensis has a relatively tonal strophe with two elements which are almost fused Appendix 4.1), with the first element being higherpitched relative to the second element. The same bird also gives a slow two-element strophe with both elements descending in pitch and having a clearly defined pause in between. Interestingly, the strophes from two recorded individuals of P. camerunensis differed in their duration (0.37 s and 0.67 s, respectively), as well as the duration of the 119 pause between the two elements (0.03 s and 0.18 s, respectively). A second montane species, P. nobilis, has an atonal, two-element strophe (0.60 s, which is reminiscent of the ‘ka-waak’ strophes of members of the Bare-throated species group. The first element has trilling and tonal parts followed by a pause (0.10 s) before a second protracted trilling element. Pternistis ochropectus also gives a two-element tonal strophe (0.48 s) with elements descending in pitch as it ends. The two elements were separated by a distinctive pause (0.21 s). What is remarkable about P. ochropectus is that based on the recordings analysed, the same individual sometimes switches the sequence of the elements around starting with the lower-pitched second element followed by a higher-pitched first element. Three species, P. erckelii, P. swierstrai and P. castaneicollis, render strophes, which are much longer (3.80 s, 4.24 s and 2.67 s long, respectively) with a series of elements that build up in intensity and complexity as the strophes progress. The strophe of P. erckelii is made up of 17 elements that increase in intensity from the start of E 1-4, with E 5-14 beginning to subside and the last three elements (E 15-17) forming a ‘cackle-trill’ that further subsides in intensity as it ends. The duration of the interelement pause ranges from 0.05 s to 0.21 s. The strophe of P. swierstrai consists of seven two-component elements that increase in intensity (from E 1-4) and decrease in volume from E 5-7. The last three elements (E 5-7) form a ‘cackle-trill’, which starts to break away in the fourth element becoming more apparent in the fifth element and so forth. The duration of the inter-element pause increases from 0.14 s to 0.33 s. The strophe of P. castaneicollis is characterized by six elements with the first element having no components and followed by three two-component elements, one three-component element, and ends in one four-component element. The last two 120 elements (E 5-6) again form a ‘cackle-trill’ as in P. erckelii and P. swierstrai. The strophe of P. castaneicollis has variable intensity from the beginning to the end. The first element starts at a low volume and then builds up in volume (E 2-4) and ends fainter with E 5-6 comprising a 'cackle-trill'. There is an increasing trend in the duration of the pause between elements (0.18 s - 0.21 s). It was reported in Madge and McGowan (2002) that P. jacksoni gives very loud, high-pitched series of cackles that sound similar to those of P. squamatus. The available recording of P. jacksoni was unfortunately of poor quality such that no sonogram could be produced. However, to the ear, it appears to give a loud series of cackles. Phylogeny Molecular and vocal characters - francolins The parsimony analysis of the CYTB data set with 1143 characters, contained 248 parsimony informative characters (21.7%); a further 110 (9.6%) variable characters were parsimony uninformative. Five trees of 965 steps were recovered. The vocal character matrix, with nine characters, contained nine parsimony informative characters, and recovered 28 equally parsimonious trees of 225 steps. The CYTB parsimony and ML (not shown) trees (Fig. 4.2) recovered three of Hall’s hypothesised monophyletic groups (Spotted - Francolinus, Red-tailed Peliperdix, Red-winged - Scleroptila). Further, the inclusion of Dendroperdix streptophora into the Red-winged Group of francolins (Scleroptila spp.) in the CYTB trees confirms the findings of Crowe and Crowe (1985) and Crowe et al. (1992). In the parsimony derived vocal tree (Fig. 4.3), the Red-tailed species group together, but with 121 the inclusion of A. lathami; however, overall the resulting phylogeny is largely unresolved. Molecular and vocal characters - spurfowls The parsimony analysis for the CYTB data set with 1143 characters, contained 231 parsimony informative characters (20.2%); a further 155 (13.6%) variable characters were parsimony uninformative. Four trees of 926 steps were recovered. The vocal character matrix, with eight characters, contained seven parsimony informative characters, one parsimony uninformative character, and recovered 17 equally parsimonious trees of 16 steps. The CYTB parsimony and ML (not shown) trees (Fig. 4.4), recovered some phylogenetic structure compared to the vocal tree (Fig. 4.5), which largely forms a polytomy. The CYTB tree recovered at least one of Hall’s hypothesised monophyletic groups, the Bare-throated Group (82% BS value – Fig. 4.4), though P. hartlaubi is recovered within the group. Within the Bare-throated Group, P. rufopictus and P. afer are recovered as sister taxa. The Vermiculated Group is split geographically into a monophyletic clade comprising the southern Vermiculated taxa, whereas the northern Vermiculated taxa are polyphyletic (P. bicalcaratus, P. clappertoni, P. icterorhynchus, P. harwoodi). Montane species are primarily distributed towards the base of the Pternistis tree, with the exception of P. jacksoni which is sister to the montane Angolan endemic P. swierstrai, although with poor support. The East African species (P. erckelii, P. ochropectus, P. castaneicollis) form a clade, with the montane West and East African taxa P. camerunensis and P. nobilis forming a sister group basal to the three East African montane species. The Namibian arid savanna endemic P. 122 hartlaubi is recovered as the basal taxon to all spurfowl species in the CYTB parsimony tree. Phylogenetic signal In comparing three common parsimony metrics for the francolin CYTB and vocal data sets, the vocal matrix has less homoplasy but considerably fewer characters: CYTB consistency index (CI) of 0.516 compared to 0.818 in the vocal matrix, retention index (RI) of 0.458 compared to 0.897 in vocal matrix, and a rescaled retention index (RC) of 0.236 compared to 0.734 for the vocal matrix. Similarly for the spurfowl CYTB data matrix, the CI is 0.521 compared to 0.625 for the vocal matrix, with a RI of 0.457 versus 0.870, and a RC of 0.238 versus 0.543. In summary, there is less homoplasy in the vocal matrices for both francolins and spurfowls than in the CYTB data sets, but because of considerably fewer characters there is less phylogenetic resolution. Discussion The quantitative and visual structural descriptions of the advertisement calls support the phylogenetic distinction between francolins and spurfowls. Francolins are found to generally give long, tonal strophes (either whistling, or harsh and tinny calls) that comprise a series of distinctive elements that usually have harmonics. Spurfowls are characterised by having generally short (with a few exceptions), atonal and grating strophes, with very few elements and no harmonics. Within the Red-winged Group (Scleroptila spp.), both the CYTB and the vocal trees recover the separation of the core-Scleroptila clade from S. streptophora and S. levaillantii. This is generally in accordance with the sonographic presentations (except 123 S. afra), which break Scleroptila into two quite distinct subgroups: (1) species with short, four-element strophes (S. finschi, S. levaillantoides S. gutturalis and S. shelleyi) and those with long strophes consisting of many more elements (S. streptophora, S. levaillantii and S. afra), often ending with a warble. It could be that the similarities in the advertisement calls among Scleroptila species may be a consequence of sharing common descent, as in contrast to the parsimony CYTB phylogeny, the ML-based topology (not shown) places S. afra at the base of a clade comprising S. finschi, S. levaillantoides, S. gutturalis and S. shelleyi. Even though the Red-winged Scleroptila species occur in varied habitats at different latitudes, their primary preference is grassland (Johnsgard 1988, Madge and McGowan 2002), which collectively comprises the open lowland grasslands in which S. levaillantoides, S. gutturalis, S. finschi thrive, and the open hilly (S. shelleyi) and highland grasslands preferred by S. afra, S. psilolaema, S. levaillantii, and S. streptophora. Within the Red-tailed group, Peliperdix species, the CYTB topology places P. albogularis and P. schlegelii together, both taxa with fast, short strophes. Basal to these two taxa is P. coqui that has a slower and longer strophe. Although the topology based on vocal characters is poorly resolved, it interestingly recovers the enigmatic Afrocolinus lathami to be nested inside Peliperdix, rather than as a highly distinct lineage, sister to Francolinus revealed in the CYTB topology (although with limited support). The grouping together of A. lathami and Peliperdix taxa in the vocal tree could be attributed to it sharing a series of elements (almost eight) forming its strophe that fade as the strophe progresses, as in other Peliperdix species, particularly P. coqui. The three Red-tailed Peliperdix species inhabit open grassland and woodland savannas. Afrocolinus lathami remains the only African francolin to be restricted to dense forested 124 habitats (del Hoyo et al. 1994, Madge and McGowan 2002), hence it has a low frequency ‘cooing’ strophe (versus the high-pitched, broadband strophe in Peliperdix spp.), which suits the forested habitat it thrives in. Frequency was not a character we could reliably score, and hence it was not included in the phylogenetic analyses. If it were included, it may well place A. lathami outside the Peliperdix clade. The CYTB topology suggests that shorter strophes evolved from longer strophes among both Scleroptila and Peliperdix species. The placement of Ortygornis (Dendroperdix) sephaena has long been a puzzle because of its aberrant life-history and vocal characteristics, that is, roosting in trees and emitting a somewhat raucous call more similar to spurfowls than francolins. The CYTB and vocal topologies suggest O. sephaena to be sister to the African O. grantii, with the DNA data suggesting this clade to be sister to the Asian O. gularis and O. pondicerianus. Ortygornis sephaena may thus represent a ‘linking form’ between Asian and African francolins. The global structure of the sonograms of O. sephaena, O. grantii, O. gularis and O. pondicerianus look similar despite the difference in the duration of the strophes. These species all have one introductory element followed by a pause, and thereafter two to three additional complex elements. The third element is almost broken into two components in O. gularis and O. pondicerianus. Aurally, the strophe of O. sephaena and O. grantii sound similar to that of the two Asian francolins, O. pondicerianus and O. gularis. Despite the Spotted Francolinus species emerging as monophyletic, they have remarkably variable strophes but share the feature of having similar introductory elements. Differences between species of spurfowls are mainly in strophe duration, the number of elements per strophe, the duration of the pause between the introductory element and the subsequent elements. The Bare-throated species (P. swainsonii, P. afer, 125 P. rufopictus and P. leucoscepus) and perhaps the northern Vermiculated taxa (P. bicalcaratus, P. icterorhynchus, P. clappertoni and P. harwoodi) have similar sonographic features despite being well-separated in the CYTB topology. The Barethroated species also inhabit similar habitats to members of the Vermiculated Group (P. hartlaubi, P. capensis, P. natalensis, P. adspersus, P. hildebrandti, P. clappertoni, P. icterorhynchus, P. bicalcaratus, P. harwoodi), which generally occupy dense bushy thickets and shrubby grasslands (del Hoyo et al. 1994, Madge and McGowan 2002). The vocal topology recovered for spurfowls is almost completely unresolved and hence is of limited phylogenetic utility. There is strong divergence among the strophes of members of the Montane species complex with just the two northeast African species (P. erckelii and P. castaneicollis) and perhaps P. swierstrai (Angolan endemic) being brought together by having multiple elements (with components) and a ‘cackle-trill’ that ends their strophes. The relationship between the strophes of P. camerunensis, P. nobilis and P. ochropectus is uncertain except that they all have two elements that descend in pitch. Pternistis nobilis, however, gives a two-element strophe that is reminiscent of those of the Barethroated species, suggesting that the ‘ka-waak’ strophe might be the plesiomorphic state within spurfowls. The vocal relationship within the Scaly Group (P. squamatus, P. griseostriatus, P. ahantensis) is uncertain and their strophes are highly divergent. Conclusions Comparing the outcome of DNA and vocal characters in a phylogenetic context was not an easy task due to the vocal phylogenetic topology being largely unresolved relative to that recovered from the DNA characters, particularly among spurfowls. However, on 126 the basis of the level of phylogenetic signal within each data set (DNA and vocal characters), vocal characters exhibit less homoplasy. Hence, if a sufficient number of characters can be scored relative to the number of taxa being studied, such characters should be of considerable use in reconstructing phylogenies. Spurfowls, unlike francolins are very divergent vocally, morphologically, as well as with respect to habitat preference. There is however, some phylogenetic concordance between DNA and vocal characters of francolins. Perhaps the most useful phylogenetic outcome of this research are the marked differences between the calls of francolins and spurfowls and the decisive placement of the ‘linking form’ Ortygornis sephaena with francolins, in particular with two of Hall’s (1963) unplaced taxa, O. pondicerianus and O. gularis. 127 Tables and Figures Table 4.1. List of species for which calls were analysed. BLSA abbreviates British Library Sound Archive, MLNS - Macaulay Library of Natural Sounds. _____________________________________________________________________________ Taxon name ID. No. Supplier Recorder Locality AFRICAN SPURFOWLS Bare-throated Group Pternistis afer P. swainsonii P. leucoscepus P. rufopictus CD2.5 - 198 CD2.5 - 199 100205 03052 R1C1 Michael Mills M. Mills MLNS BLSA G. Gibbon G. Gibbon R. Linda Cornell Natal Namibia Ethiopia Tanzania Montane Group P. camerunensis P. ochropectus P. nobilis P. erckelii P. swierstrai P. castaneicollis P. jacksoni CD5 - 95 cc26744 CD5 - 96 68756 BD10 Unknown 100242 Unknown Michael Mills BLSA Michael Mills BLSA Michael Mills MLNS C. Cohen C. Chappuis G. Welch C. Chappuis S. Smith M. Mills R. Linda B. Finch Mt. Cameroon Djibouti Uganda Unknown Angola Ethiopia Kenya Vermiculated Group P. bicalcaratus P. icterorhynchus P. clappertoni P. harwoodi P. hildebrandti P. capensis P. natalensis P. adspersus P. hartlaubi 216-231 cc1238 BD20 cc1239 BD21 24718 BD1 CD5 – 92 CD2 - 195 CD2 - 196 61790 BD24 CD2 - 197 Michael Mills BLSA BLSA BLSA Michael Mills Michael Mills Michael Mills BLSA Michael Mills C. Chappuis C. Bourguignon C. Chappuis H. Shirihai C. Chappuis G. Gibbon G. Gibbon D. Watts G. Gibbon Ivory Coast Zaire Nigeria Ethiopia Malawi South Africa South Africa Namibia Namibia Scaly Group P. ahantensis P. griseostriatus P. squamatus cc24693BD32 103772BD49 CD5 - 89 BLSA BLSA Michael Mills C. Chappuis I. Sinclair C. Chappuis Senegal Angola Gabon MLNS MLNS BLSA B. King P. Holt C. Carter Sri Lanka India Burma Michael Mills Michael Mills Michael Mills C. Chappuis G. Gibbon G. Gibbon Gabon South Africa South Africa ASIATIC FRANCOLINS Spotted Group Francolinus pictus F. francolinus F. pintadeanus 551 86293 113300 BD42 AFRICAN FRANCOLINS Red-winged Group Scleroptila finschi S. shelleyi S. afra CD5 - 86 CD2.5 - 191 CD2.4 - 190 128 ___________________________________________________________________________ Taxon name ID. No. Supplier Recorder Locality S. levaillantoides S. gutturalis S. levaillantii CD2.5 - 193 48414 R1C1 CD2.5 - 192 Michael Mills BLSA Michael Mills G. Gibbon D. Pearson G. Gibbon Namibia Somalia South Africa Red-tailed Group Peliperdix coqui P. schlegelii P. albogularis CD2 - 188 cc1214 BD22 cc1212 BD20 M. Mills BLSA BLSA G. Gibbon J. Brunel C. Chappuis South Africa Chad Senegal Striated Group Ortygornis sephaena O. grantii S. streptophora 1095 CD5 - 88 CD5 - 84 M. Hausberger Michael Mills Michael Mills M. Hausberger C. Chappuis C. Chappuis South Africa Kenya Cameroon C. Chappuis P. Holt P. Holt Ivory Coast Nepal Pakistan Unplaced taxa (excluding Ptilopachus nahani) Afrocolinus lathami CD5 - 80 Michael Mills Ortygornis gularis 65642 BD20 BLSA O. pondicerianus 41433 R1C10 BLSA 129 4.2. Francolin and spurfowl taxa for which DNA sequences were generated. Acronyms. AMNH abbreviates American Museum of Natural History, TM Transvaal Museum, BM - British Museum - Natural History Museum at Tring, FMNH – Field Museum of Natural History, PFIAO - Percy FitzPatrick Institute of African Ornithology, TMC - Timothy M. Crowe, University of Cape Town, South Africa, ‘-’ – Unknown. Genera are as recorded on specimen label. Taxa name Sample no. Origin Date collected GenBank no. AMNH DOT8023 India AMNH 776813 India GenBank China - AF013762 FR694142 NC011817 TMC 9 Marico River, South Africa BM 1902 1 20 300 Hulul, Ethiopia TMC 11 Cameroon 2004 1902 2005 FR694140 FR694144 PFIAO 47 PFIAO 59 TMC 12 AMNH 541174 TM 78622 AMNH 308887 Ayton Farm, South Africa Eastern Cape, South Africa Petrus Steyn, South Africa Ethiopia Sterkspruit, South Africa Angola 2002 2002 2005 1941 AM236898 AM236897 FR691612 FR691613 U90642 FR691607 PFIAO 45 BM 1949 30 19 BM 1929 3 13 1 Settlers, South Africa Mboro, Bahr-El-Ghazel, Chad1949 Farafeni, Gambia 1929 AM236895 FR694149 FR694145 Francolins Spotted Group F. francolinus F. pictus F. pintadeanus Striated Group Francolinus sephaena F. sephaena grantii F. streptophora FR691617 Red-winged Group F. shelleyi F. afra F. levaillantoides levaillantoides F. l. gutturalis F. levaillantii levaillantii F. finschi Red-tailed Group F. coqui F. schlegelii F. albogularis Ungrouped species F. lathami F. pondicerianus F. gularis GenBank Cameroon AMNH DOT8050 India GenBank India - AM236893 FR691632 U90649 PFIAO 108, TMC 40 AMNH 202503 PFIAO 109 2004 2004 2004 AM236908 AM236907 FR691588 AM236906 Spurfowls Bare-throated Group Francolinus afer F. swainsonii F. rufopictus F. leucoscepus Tudor East, Watervalboven Marico River Gagayo, Muranza Kenya 130 Taxa name Sample no. Origin Date collected GenBank no. AMNH 541471 FMNH 1971-1072 GenBank AMNH261929 AMNH1759 TMC 42 AMNH 419126 Badaltino, Shoa Djbouti East slope, Mt. Kenya West Ruwenzori Mount Cameroon Angola - FR691589 FR691590 AM236903 FR691594 FR691592 FR691591 FR691593 PFIAO 117 AMNH 541411 Ndalla Tanda - AM236905 F. bicalcaratus F. clappertoni F. icterorhynchus F. hildebrandti F. natalensis F. hartlaubi F. capensis F. adspersus F. harwoodi TM 14682 TMC 68 AMNH 156922 GenBank TMC 120 TMC 121 PFIAO 229 PFIAO 206A BM 1927.11.5.18 Gold Coast, Hinterland Cameroon Fanadji Marico River, South Africa Namibia Kakamas, South Africa - 1901 2005 2004 2006 1927 FR691624 FR691602 FR691601 FR691595 AM236911 FR691618 AM236909 FR691623 FR691600 Outgroups Gallus gallus Bambusicola thoracica Alectoris chukar Coturnix coturnix - - - L083761 EU165706 L083781 L083771 Montane Group F. erckelii F. ochropectus F. castaneicollis F. jacksoni F. nobilis F. camerunensis F. swierstrai Scaly Group F. squamatus F. griseostriatus AM236904 Vermiculated Group 131 Table 4.3. DNA markers sequenced and primers used for PCR amplifications and sequencing of preserved tissues. Primer name Primer sequence (5’to 3’) Reference Fresh tissues Francolins & Spurfowls (General primers) Cytochrome b L14578 MH15364 ML15347 H15915 cta gga atc atc cta gcc cta ga act cta cta ggg ttt ggc c atc aca aac cta ttc tc aac gca gtc atc tcc ggt tta caa gac J.G. Groth (pers. comm.) P. Beresford (pers. comm.) P. Beresford (pers. comm.) Edwards & Wilson (1990) Toe-pads Cytochrome b Spurfowl-specific primers L14851 (General) cct act tag gat cat tcg ccc t Pt-H195 ttt cgr cat gtg tgg gta cgg ag Pt-H194 cat gtr tgg gct acg gag g MH15145 aag aat gag gcg cca ttt gc Kornegay et al. (1993) R. Moyle & T. Mandiwana-Neudani R. Bowie P. Beresford Pt-L143 Pt-H361 gcc tca tta ccc aaa tcc tca c gtg gct att agt gtg agg ag R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani Pt-L330 Pt-H645 tat act atg gct cct acc tgt ac ggg tgg aat ggg att ttg tca gag R. Bowie R. Moyle & T. Mandiwana-Neudani Pt-L633 Pt-H901 ggc tca aac aac cca cta ggc agg aag ggg att agg agt agg at R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani L2-2312 H15696 cat tcc acg aat cag gct c R. Bowie aat agg aag tat cat tcg ggt ttg atg Edwards et al. (1991) Pt-L851alt Pt-H1050 cct att tgc cta cgc cat cct ac gat gct gtt tgg ccg atg R. Bowie R. Bowie Pt-L961 Pt-L961alt HB20 (General) cga acc ata aca ttc cca c ctc atc cta ctc cta atc ccc ttg gtt cac aag acc aat gtt R. Moyle & T. Mandiwana-Neudani R. Bowie J. Feinstein (pers. comm.) 132 Primer name Primer sequence (5’to 3’) Reference Cytochrome b Francolin-specific primers L14851 (General) cct act tag gat cat tcg ccc t Franc-H1 cag cag aca cyt cyc tyg cct tc MH15145 aag aat gag gcg cca ttt gc Kornegay et al. (1993) R. Bowie P. Beresford Franc-L1 Franc-H2 Franc-L2 Franc-H3 tgc ctc aca acc caa atc ctc ac agg agr agr att act cct gtg ttt cag g gcc tca ttc tty ttc aty tgy atc ttc c ggr tgg aat ggg att ttg tca gag R. Bowie R. Bowie R. Bowie R. Bowie Franc-L3 Franc-H4 tcatcyractcygacaaaatccc gar rgg gat tag rag gag gat R. Bowie R. Bowie Franc-L4 Franc-H5 tat tcg cct ayg cya tcc twc gct c gta gga rag kga tgc tat ttg gcc R. Bowie R. Bowie Franc-L5 HB20 (General) ctc atc ctc ctc cta atc cc ttg gtt cac aag acc aat gtt R. Bowie J. Feinstein (pers. comm.) 133 Table 4.4. Vocal character states scored and used for the phylogenetic analysis of francolins. 1. Strophe length: <1.0 s = 1; =1.0 s <2 s = 2; ≥2.0 s = 3 2. No. of elements in strophe: <5 = 1; ≥5 = 2 3. Harmonics: Absent = 0; Indistinctive = 1; Predominately distinct = 2; Distinct mixed with distinctive trills = 3 4. Raucous advertisement: Absent = 0; Type 1 = 1; Type 2 = 2 5. No. of introductory elements: one = 1; two = 2; More than two = 3 6. Strophe warbling ending: Absent = 0; Present = 1 7. Simple musical advertisement: Absent = 0; Slow = 1; Fast = 2; Very fast = 3 8. Short complex musical advertisement call: Absent = 0; Slow I'll drink-yer-beer = 1; Fast I'll drink-yer-beer = 2 9. Long complex musical call: Absent = 0; Type 1 = 1; Type 2 = 2 A matrix of vocal state scores used for phylogenetic analyses of the francolins. Characters Taxon 1 2 3 4 5 6 7 8 9 Gallus gallus 1 1 2 1 1 0 0 0 0 Bambusicola thoracica Francolinus francolinus 1 1 2 1 1 0 0 0 0 3 2 2 1 1 0 0 0 0 Francolinus pintadeanus 2 2 3 1 2 0 0 0 0 Francolinus pictus 3 2 1 1 1 0 0 0 0 Afrocolinus lathami 2 2 0 0 1 0 1 0 1 Ortygornis pondicerianus 1 1 2 1 1 0 0 0 0 Ortygornis gularis 1 1 2 1 1 0 0 0 0 Ortygornis grantii 1 1 3 2 1 0 0 0 0 Ortygornis sephaena 1 1 3 2 1 0 0 0 0 Peliperdix albogularis 1 2 2 0 1 0 3 0 1 Peliperdix coqui 2 2 2 0 1 0 1 0 1 Peliperdix schlegelii 2 2 2 0 1 0 2 0 1 Scleroptila afra 3 2 2 0 3 1 0 0 2 Scleroptila finschi 1 1 2 0 2 0 0 1 0 Scleroptila levaillantii 3 2 2 0 3 1 0 0 2 Scleroptila levaillantoides 1 1 2 0 2 0 0 2 0 Scleroptila gutturalis 1 1 2 0 2 0 0 1 0 Scleroptila shelleyi 1 1 2 0 2 0 0 1 0 Scleroptila streptophora 3 2 2 0 1 0 0 0 2 134 Table 4.5. Vocal characters and states scored and used for phylogenetic analysis of spurfowls. 1. Strophe length: <1 s = 1; ˃1 s =2 2. No. of elements in strophe: 2-3 = 1; ≥4 = 2 3. Strophe type: Less to no trill = 1; Predominately trilled = 2; Predominately harmonic = 3 4. Inter-element pause: Absent/indistinct = 1; Distinctive = 2 5. Cackle-trill ending: Absent = 0; Present = 1 6. Ka-waak/Ko-rak component: Absent = 0; Present = 1 7. Strophe pitch: Stable = 1; Descends as strophe ends = 2; Ascends as strophe ends = 3 8. Strophe antiphonal?: No = 0; Yes = 1 A matrix of vocal character used to generate the vocal character phylogeny. Characters Taxon 1 2 3 4 5 6 7 8 Coturnix coturnix 1 1 1 1 0 0 ? 0 Alectoris chukar 1 1 1 1 0 0 ? 0 Pternistis hartlaubi 2 2 1 2 0 0 1 1 Pternistis camerunensis 1 1 3 2 0 0 1 0 Pternistis nobilis 1 1 2 2 0 1 1 0 Pternistis erckelii 2 2 2 2 1 0 2 0 Pternistis swierstrai 2 2 2 2 1 0 2 0 Pternistis castaneicollis 2 2 2 2 1 0 2 0 Pternistis jacksoni ? ? ? ? 1 0 ? ? Pternistis ochropectus 1 1 3 2 0 0 1 0 Pternistis squamatus Pternistis ahantensis 2 1 1 1 3 1 2 2 0 0 0 0 3 1 0 0 Pternistis griseostriatus 1 ? 1 1 0 0 1 0 Pternistis bicalcaratus 1 1 2 2 0 1 1 0 Pternistis icterorhynchus 1 1 2 1 0 1 1 0 Pternistis clappertoni 1 1 2 1 0 1 1 0 Pternistis harwoodi 1 1 2 1 0 1 1 0 Pternistis hildebrandti 1 1 2 1 0 0 1 0 Pternistis natalensis 1 2 2 2 0 0 1 0 Pternistis adspersus 2 2 2 2 1 0 2 0 Pternistis capensis 2 2 3 2 1 0 3 0 Pternistis leucoscepus 1 1 2 1 0 1 1 0 Pternistis rufopictus 1 1 2 1 0 1 1 0 Pternistis afer 1 1 2 1 0 1 1 0 Pternistis swainsonii 1 1 2 2 0 1 1 0 135 Figure 4.1. Example of a sonogram of a typical francolin, Coqui Francolin Peliperdix coqui, illustrating and defining the variables studied: E - element, E1 - element number 1, P - pause, H - harmonic, H1 - harmonic number 1 and FF - fundamental frequency. 136 Gallus gallus Bambusicola thoracica Scleroptila streptophora Scleroptila shelleyi Scleroptila levaillantoides 98 82 77 R W G Scleroptila afra Scleroptila guturalis 99 100 Scleroptila finschi Scleroptila levaillantii 83 92 Peliperdix schlegelii 95 Peliperdix albogularis 86 99 R T G Peliperdix coqui Ortygornis sephaena 100 100 80 98 Ortygornis grantii Ortygornis spp. Ortygornis pondicerianus 100 100 Ortygornis gularis Francolinus pintadeanus Francolinus pictus 82 S P G Francolinus francolinus Afrocolinus lathami Figure 4.2. A parsimony tree (1 of 5 most parsimonious trees) of francolins obtained from mitochondrial Cytochrome-b characters. Numbers above branches represent parsimony boostrap support values and those below branches are maximum likelihood bootstrap support values (only ≥ 70% are presented). RWG stands for Red-winged Group, RTG – Red-tailed Group, SPG – Spotted Group. 137 Gallus gallus Bambusicola thoracica Francolinus francolinus S P G Francolinus pintadeanus Francolinus pictus Afrocolinus lathami Peliperdix coqui Peliperdix schlegelii R T G Peliperdix albogularis Ortygornis pondicerianus Ortygornis gularis Ortygornis Spp. Ortygornis grantii 81 Ortygornis sephaena Scleroptila afra 90 Scleroptila levaillantii Scleroptila streptophora Scleroptila finschi R W G Scleroptila levaillantoides 78 Scleroptila guturalis Scleroptila shelleyi Figure 4.3. The strict consensus parsimony tree of francolins obtained from vocal characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). RWG stands for Red-winged Group, RTG – Red-tailed Group, SPG – Spotted Group. 138 Alectoris chukar Coturnix coturnix Pternistis adspersus Pternistis capensis 87 95 Pternistis hildebrandti 99 98 S V t Pternistis natalensis Pternistis afer 81 81 B T G Pternistis rufopictus Pternistis leucoscepus 82 Pternistis swainsonii 100 97 Pternistis clappertoni Pternistis harwoodi Pternistis jacksoni Pternistis swierstrai Pternistis squamatus Pternistis griseostriatus M T t S C t Pternistis bicalcaratus 100 Pternistis icterorhynchus Pternistis castaneicollis 89 92 88 Pternistis erckelii 93 90 89 90 100 Pternistis ochropectus M T t Pternistis camerunensis Pternistis nobilis Pternistis hartlaubi Figure 4.4. A parsimony tree (1 of 4 most parsimonious trees) of spurfowls obtained from mitochondrial Cytochrome-b characters. Numbers above branches represent parsimony boostrap support values and numbers below branches represent maximum likelihood boostrap support values (only ≥ 70% are presented). MTt stands for Montane taxa, SCt – Scaly taxa, SVt – Southern Vermiculated taxa and BTG – Bare-throated Group. 139 Coturnix coturnix Alectoris chukar Pternistis hartlaubi Pternistis camerunensis Pternistis erckelii Pternistis swierstrai Pternistis castaneicollis Pternistis jacksoni Pternistis adspersus Pternistis capensis Pternistis squamatus Pternistis ochropectus Pternistis ahantensis Pternistis natalensis Pternistis nobilis Pternistis bicalcaratus Pternistis swainsonii Pternistis icterorhynchus Pternistis clappertoni Pternistis harwoodi Pternistis leucoscepus Pternistis rufopictus Pternistis afer Pternistis hildebrandti Pternistis ochropectus Figure 4.5. The strict consensus parsimony tree of spurfowls obtained from vocal characters. No nodes received bootstrap (i.e. > 50%) support. 140 Appendix 1 Francolinus francolinus F. pictus F. pintadeanus i Sonograms of the Francolinus species. 141 Ortygornis sephaena O. grantii O. pondicerianus O. gularis Sonograms of Ortygornis species. 142 Peliperdix albogularis P. schlegelii P. coqui Afrocolinus lathami Sonograms of Peliperdix species and Afrocolinus lathami. 143 Scleroptila finschi S. levaillantoides S. gutturalis S. shelleyi Sonograms of Scleroptila species. 144 Scleroptila streptophora S. levaillantii S. afra Sonograms of Scleroptila species. 145 Pternistis swainsonii P. afer P. rufopictus P. leucoscepus i Sonograms of the Bare-throated spurfowls. 146 Pternistis icterorhynchus P. bicalcaratus P. clappertoni P. harwoodi Fi Sonograms of the Vermiculated spurfowls. 147 Pternistis hildebrandti P. hartlaubi P. adspersus P. natalensis P. capensis Fi Sonograms of the Vermiculated spurfowls. 148 Pternistis squamatus P. ahantensis P. griseostriatus Fi Sonograms of the Scaly spurfowls. 149 Pternistis camerunensis (1) P. camerunensis (2) P. nobilis P. ochropectus Fi Sonograms of the Montane spurfowls. 150 Pternistis erckelii P. swierstrai P. castaneicollis Fi Sonograms of the Montane spurfowls. 151 CHAPTER 5 Taxonomy and phylogeny of ‘true’ francolins Abstract The development of a plethora of species and subspecies concepts has had major effects on the delineation of terminal taxa across the Class Aves. The taxonomy and phylogeny of small, quail-like, Afro-Asian phasianine birds now known as ‘true’ francolins, but traditionally placed in a single genus Francolinus Stephens, 1819 with a range of other taxa are revised. The number of taxa that have been recognized as species and, to a larger extent as subspecies among ‘true francolins’ (Francolinus, Dendroperdix, Peliperdix and Scleroptila spp.) (sensu Crowe et al. 2006) has never been stable. This is due to a lack of objective, evolutionarily relevant species, subspecies and generic circumscription. This study aimed to establish a classification system of francolins which takes into account the evolutionary relationships among taxa, and which, in turn, could generally bring stability to the number of taxa recognized as valid species, subspecies and genera based on congruent multiple lines of evidence. The model-based Maximum likelihood and parsimony analyses of separate and combined DNA and organismal characters resulted in some putative subspecies being elevated to the species level, others placed into more inclusive entities, and two new genera being established. Most of the phylogenetic hypotheses presented by Hall (1963) were rejected. The genus Dendroperdix is replaced by Ortygornis. The ‘true’ francolins are divided among five genera in the following pectinate phylogenetic sequence: Francolinus / Ortygornis / 152 Afrocolinus gen. nov. / Peliperdix / Scleroptila. A multi-faceted character approach seems to be a fitting strategy with which to delineate intra-generic relationships in francolins. 153 Introduction Despite the challenges and uncertainties that the galliform phylogeny still presents (see Chapter 2), we do know that ‘francolins’ sensu lato do not share common ancestry as traditionally circumscribed. The genus Francolinus Stephens, 1819 (sensu Hall 1963), is split into two distantly related assemblages: ‘francolins’, represented by taxa classified in the genera Francolinus Stephens, 1819, Dendroperdix Roberts, 1922, Peliperdix Bonaparte, 1856 and Scleroptila Blyth, 1849, and ‘spurfowls’, with all members being assigned to the genus Pternistis Wagler, 1832. Traditionally, 41 species (Table 1.1.) are assigned to the genus Francolinus, 36 occur in sub-Saharan Africa and five in Asia, thus making it the most specious genus in the Galliformes (Morony et al. 1975, del Hoyo et al. 1994) and one of the largest genera in the class Aves (Bock and Farrand 1980). Francolins are placed in the sub-family Phasianinae, and together with other Old World partridge- and quail-like gamebirds (e.g. Perdix and Coturnix spp.) in the tribe Perdicini (Chapin 1932, Peters 1934, Wolters 197582, Crowe et al. 1986, Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994, Madge and McGowan 2002). The traditional composition of the genus Francolinus has been disputed for over 50 years, with continuing uncertainty regarding the delineation and relationships of taxa within and between constituent species. Species of francolins and their distribution The focus in this chapter centres on the ‘true’ francolins (Francolinus, Dendroperdix, Peliperdix and Scleroptila spp.), that is those taxa assigned to Hall’s (1963) species groups (see Table 1.1) with the addition of four species that Hall failed to assign to any of her groups. The species groups are: the Spotted (represented by three Asian species), 154 Striated, Red-tailed and the Red-winged Group (all restricted to Africa). The four species she was unable to assign to a group are two African species, Latham’s Francolin Francolinus lathami and Nahan’s Francolin Ptilopachus ‘Francolinus’ nahani, and two Asian species, Swamp Francolin Francolinus gularis and Grey Francolin Francolinus pondicerianus (Table 1.1). Generally, francolins are small to medium sized, quail-like resident birds that can run or fly for short distances when faced with a threat. They are mainly diagnosed by having 14 tail feathers that moult centrifugally. Most species are sexually monomorphic in plumage, with males of most species having single spurred tarsi, with females in only a few cases having relatively smaller spurs (Johnsgard 1988). They represent a morphologically, ecologically and behaviourally diverse group, and have complex distribution patterns (Snow 1978). They occur in diverse habitats of a tropical to subtropical nature, with one African species F. lathami occurring in forested habitat (excluding P. nahani) (Hall 1963, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). Francolins occur at varying altitudes and most of the species assemblages show allopatric distributional patterns among their constituent taxa (Hall 1963). Species group diversity, distribution and morphology Spotted Group – Genus Francolinus Stephens, 1819 This Asian group is represented by three recognized species (Table 1.1), the Black Francolin Francolinus francolinus (Linnaeus, 1766), Painted Francolin F. pictus (Jardine & Selby, 1828) and Chinese Francolin F. pintadeanus (Scopoli, 1786). The nominate species of the genus, F. francolinus, with its putative subspecies (Table 1.2) 155 occurs from Cyprus to Manipur in northeastern India (Forcina et al. 2012). This species inhabits lowland cultivations, river deltas and lake edges with scrub and reeds (Forcina et al. 2012). Francolinus pictus represents the southern counterpart of F. francolinus, is distributed from Sri Lanka to north India, where it favours semi-dry undulating grasslands with scrub or cultivation (Forcina et al. 2012). Francolinus pintadeanus represents the eastern counterpart of F. francolinus (Forcina et al. 2012), inhabiting subtropical or tropical lowland forest from northeastern India, across Myanmar to south eastern China, extending across western and southern Thailand (Forcina et al. 2012). Striated Group – Genus Dendroperdix Roberts, 1922 This group comprises two species, the polytypic Crested Francolin Dendroperdix sephaena (Smith, 1836), and the monotypic Ring-necked Francolin Dendroperdix streptophora O. Grant, 1891. The distributional range of D. sephaena extends from Somalia through eastern Africa to KwaZulu-Natal in South Africa, and extends westwards across Namibia and southern Angola, where the species inhabits mostly acacia savanna and steppe habitats (Fig. 5.1; Mackworth-Praed and Grant 1952, 1962, 1970, Hall 1963). Several subspecies (Table 1.2) are recognized: in D. sephaena there are sephaena (Smith, 1836), spilogaster Salvadori, 1888, somaliensis Grant & Praed, 1934, schoanus Heuglin, 1873, jubaensis Zedlitz, 1913, grantii Hartlaub, 1866, rovuma Gray, 1867, zambesiae Praed, 1920, chobiensis Roberts ,1932, thompsoni Roberts, 1924, zuluensis Roberts, 1924 and mababiensis Roberts, 1932. Generally, D. sephaena is a rich reddish-brown bird with broad white shaft streaks. It has a chestnut collar that is interrupted with white. The patterned feathers of the lower neck are rich dark brown or blackish with broad white centres. This species is 156 slightly dimorphic in that the females unlike the males are slightly vermiculated. The rest of the belly is buff with triangular chestnut markings on the breast, narrow greyish barring on the lower breast and it has an unpatterned belly. This species can be categorized into two distinct types, the "coastal type", which has fine drop-shaped light chestnut streaks on the belly and the "inland type" with no streaks on the belly (Fig. 5.1). Dendroperdix streptophora has a markedly disjunct distribution occurring in savanna-grassland in Cameroon in the west, and northern Uganda and northwestern Kenya and Tanganyika in the east (Fig. 5.2). It differs from most of the other Redwinged francolins in having an unpatterned crown, throat colour buff rimmed with rufous, the absence of a gorget, barred breast, blotched and mottled belly, brown (not red) primaries, unpatterned side of the head, and a tiny spur bump. Red-winged Group – Genus Scleroptila Blyth, 1849 Members of this quail-like plumaged group, Shelley’s Francolin Scleroptila shelleyi O. Grant, 1890, Grey-winged Francolin S. afra (Latham, 1790), Orange River Francolin S. levaillantoides Smith, 1836, Red-winged Francolin S. levaillantii (Valenciennes, 1825), Finsch’s Francolin S. finschi Bocage, 1881, Moorland Francolin S. psilolaema Gray, 1867 are relatively homogenous in their overall morphology (Hall 1963, Snow 1978) and are distinguished by having red or rufous on their primaries, which is reduced in the Grey-winged Francolin S. afra. The distributional range of the group encompasses most of eastern and southern Africa extending from Ethiopia and Eritrea to the Cape, and westwards to Angola and the Congo (Fig. 5.2, 5.3, 5.4). Some species are allopatric (Snow 1978), and others have complex distributions, with species occupying diverse habitats at different altitudes. 157 Taxonomically, this is Hall’s (1963) most contentious group. Several of the species in this complex are polytypic with multiple subspecies as follows: four in S. psilolaema (psilolaema Gray, 1867, ellenbecki Erlanger, 1905, elgonensis O. Grant, 1891, theresae Meinertzhagen, 1937); seven within S. shelleyi (shelleyi O. Grant, 1890, uluensis O. Grant, 1892, whytei Neumann, 1908, macarthuri Someren, 1938, trothae Reichenow, 1901, sequestris Clancey, 1960, canidorsalis (Lawson, 1963)); 12 in S. levaillantoides (levaillantoides Smith, 1836, kalaharica Roberts, 1932, pallidior Neumann, 1908, langi Roberts, 1932, wattii Macdonald, 1953, jugularis Büttikorfer, 1889, cunenensis Roberts, 1932, stresemanni Hoesch & Niethammer, 1940, gutturalis (Rüppell, 1835), lorti Sharpe, 1897, archeri Sclater, 1927, ludwigi Neumann, 1920); five in S. levaillantii (levaillantii (Valenciennes, 1825), kikuyuensis O. Grant, 1897, crawshayi O. Grant, 1896, benguellensis Neumann, 1908, clayi White, 1944). Scleroptila finschi Bocage, 1881 and S. afra (Latham, 1790) are monotypic species. Red-tailed Group – Genus Peliperdix Bonaparte, 1856 The Red-tailed Group is represented by small francolins generally with a wing length of less than 150 mm (Mackworth-Praed and Grant 1952, 1962, 1970). Their distribution range stretches from Senegal to Sudan and from central Kenya west to the central Congo and Angola, and south to the Transvaal and Natal in South Africa (Fig. 5.5). There is an isolated subspecies in the Ethiopian Rift. The three traditionally recognized species are Coqui Francolin Peliperdix coqui (Smith, 1836), White-throated Francolin P. albogularis Hartlaub, 1854, and Schlegel’s Francolin P. schlegelii Heuglin, 1863 (Table 1.1), and all are allopatric (Hall 1963, Snow 1978). Members of this group occupy wooded grasslands. All the above taxa generally have a quail-like patterning on 158 the back which is consistent and well-defined, though the basic colour and the colour of the crown varies from brown to rufous in different taxa. Peliperdix coqui is one of the two African francolins (including F. lathami) that exhibits truly marked sexual dimorphism in plumage (Hall 1963). The male birds differ from the females in having the sides of the head and throat ochre or light buff without a black eye stripe or necklace. Females have the sides of the head similar in colour to the males, but the throat is whiter and they have a black necklace and a black eye stripe that continues as a black line down the sides of the head. The breast and lower neck of the male is barred with black, whereas the barring on the breast of the female is overlaid by a pinkish wash and there is no barring on the lower neck. Peliperdix coqui and P. albogularis are the two species in this group that have recognized subspecies (Table 1.2): 14 in P. coqui (coqui (Smith, 1836), spinetorum Bates, 1928, buckleyi Peters, 1934, maharao Sclater, 1927, ruahdae Someren, 1926, hubbardi O. Grant, 1895, angolensis Rothschild, 1902, lynesi Sclater, 1932, vernayi (Roberts, 1932), campbelli (Roberts, 1928), thikae Grant & Praed, 1934, kasaicus White, 1945, hoeschianus Stresemann, 1937, stuhlmanni Reichenow, 1889), and five for P. albogularis (albogularis Hartlaub, 1854, buckleyi O. Grant, 1892, dewittei Chapin, 1937, meinertzhageni White, 1944, gambagae (Praed, 1920)). The unplaced species of Hall (1963) The species Ptilopachus ’Francolinus’ nahani, one of Hall’s enigmatic species (others F. lathami, F. gularis and F. pondicerianus) (Table 1.1), is excluded from this chapter on francolins since this species was found not to be a ‘francolin’, but a ‘partridge’, with close phylogenetic affinity to the African Stone Partridge Ptilopachus 159 petrosus, and both these species in turn are the closest extant relatives of the New World quails (see Chapter 2, Cohen et al. 2012, Bowie et al. 2013). Latham’s Francolin F. lathami is a small francolin with a disjunct distribution and is restricted to forest habitat. Its distribution extends from Sierra Leone to western Uganda and southern Sudan, with two further outlying pockets of occupancy in the Congo (Fig. 5.6). Two subspecies, lathami Hartlaub, 1854 and schubotzi Reichenow, 1912, are recognized (Table 1.2). Francolinus pondicerianus lives on dry plains in semi-desert areas (Forcina et al. 2013) in the Gulf of Oman and on the plains of India. Francolinus gularis is patchily distributed between south-western Nepal and extreme north-eastern India (del Hoyo et al. 1994, Forcina et al. 2012), where it is confined to marshes and reeds (Hall 1963) on the plains of the Ganges River. What do we know about their taxonomy and phylogeny? Chapter 1 outlines the taxonomic disagreement and confusion that various authors have had with delineating taxa at the subspecific level (Table 1.2), as well as with attempts to assign francolin species to various genera. The classic monograph of Hall (1963) on francolins remains one of the most significant works on any group of African birds. It has significantly enhanced our knowledge and understanding of the taxonomic status of francolins, their evolutionary relationships, current distribution ranges including how species might have spread to their current habitats. What ignited further investigation of systematic relationships among francolins was Hall’s (1963) conclusion that the genus Francolinus was monophyletic, and that the 37 species she recognized could be divided among eight putatively monophyletic groups comprising ecologically similar, but largely allo- or parapatric species (Bloomer and 160 Crowe 1998). The reason Hall (1963) did not recognize distinct genera was because of the difficulties presented by the Crested Francolin Dendroperdix sephaena and an atypical Ptilopachus nahani. However, she indicated that should generic subdivision be desirable she would have two major groups, one consisting of members that she referred to as belonging to the Spotted Group (Asian francolin group), Bare-throated (African spurfowls), Montane (African spurfowls), Scaly (African spurfowls), Vermiculated Group (African spurfowls) and one of the unplaced Asian francolin F. gularis. The second group would be comprised of the Red-winged (francolins), Striated (francolins) and Red-tailed groups (francolins), and two unplaced species F. pondicerianus (Asian francolin) and F. lathami (African francolin). Wolters (1975-82) like Roberts (1924) (Table 1.2) recognized the genera: Francolinus, Scleroptila, Dendroperdix, Peliperdix and Ortygornis. Contrary to Roberts, Peters (1934), and Mackworth-Praed and Grant (1952, 1963, 1970), Wolters assigned the genus Pternistis (Table 1.2) to all taxa which are known today as ‘spurfowls’. In short, Wolters’s system of genera represents what is used today (Table 1.2) with the exception of the genus Ortygonis, which he assigned to F. pondicerianus. The number of traditionally recognized francolin species ranged widely over time with 17 excluding P. nahani (in Peters 1924 - who covered Africa and Asia), 11 excluding Scleroptila psilolaema and P. nahani (Mackworth-Praed and Grant 1952, 1962, 1970 - Africa only), 17 excluding P. nahani (Wolters 1975-82 - Africa and Asia), and 17 excluding P. nahani (Hall 1963 - Africa and Asia). Phylogenetic difficulties have been articulated by Crowe and Crowe (1985), Crowe et al. (1986), Crowe et al. (1992), Bloomer and Crowe (1998), Crowe et al. (2006) and most recently by Forcina et al. (2012), with regard to the monophyletic status of the 161 genus Francolinus, and in particular, the status of the different putative monophyletic groups recognized by Hall (1963). It is important to note that different studies used different types of characters (Table 1.4), focussed on different geographic areas of francolin distribution (Table 1.3), and made use of different methods of phylogenetic analysis. Groups that various authors have recovered as monophyletic are the Spotted Group (Crowe and Crowe et al. 1985, Crowe et al. 1992, Forcina et al. 2012), Red-tailed. Group (Crowe and Crowe et al. 1985, Crowe et al. 1992), and the Red-winged Group (Crowe and Crowe et al. 1985, Crowe et al. 1992, Bloomer and Crowe 1998). No study has ever recovered the monophyly of the Striated Group, with Crowe et al. (1986) placing the Ring-necked Francolin Dendroperdix streptophora in the Red-winged Group (Scleroptila). In most analyses (Crowe and Crowe 1985, Crowe et al. 1992, Bloomer and Crowe 1998, Chapter 2), F. pondicerianus and F. gularis group with D. sephaena and as such it would have been ideal for Forcina et al. (2012) to have included some African francolins to strengthen their findings on the Asian francolins and purported delineation of the Spotted Group as being monophyletic. On the basis of the chaotic taxonomic status and the phylogenetic uncertainty of francolin taxa (Tables 1.1, 1.2), a formal revision of all francolin taxa is urgently needed. What is taxonomy and how can we possibly study it? A review of species and subspecies concepts Delineating species and/or subspecies boundaries correctly is crucial to the discovery of life’s diversity because it determines whether or not we can recognize when different specimens are members of the same cohesive lineage (Dayrat 2005). Many different types of taxonomic characters (morphological, physiological, chemical, behavioural, ecological, 162 molecular, among others) have been used to delineate taxa. In taxonomy, the fundamental components of biodiversity, species and subspecies, are discovered, described, named, and classified. The naming of taxa, in this case francolins, is to some degree of less concern in the present study. However, the classification of these taxa at the species-level and below, presently lacks consensus and therefore there is a need to establish a sound classification system that can best reflect the evolutionary history of francolins. A number of francolin species are polytypic, being complexes of taxa that exhibit considerable geographical variation in their biological traits and this variation poses serious challenges to classification (Pough 1990). The concept of ‘polytypic’ species is said to have emerged as a way of simplifying classification by lumping several diverse but difficult to delineate taxa into a single species (Clancey 1957, Maclean 1993). The subspecies concept is one that has been debated for over 100 years with various authors presenting different perspectives on the meaning of ‘subspecies’. Many authors interpret subspecies as geographically partitioned variation, which may or may not exhibit intergradation (Winker 2010). Subspecies are often construed as biological entities that provide evidence of adaptation and the early stages of speciation; they are also considered to be important in improving our understanding by alerting us to interesting geographic and behavioural patterns (Cicero 2010). Cicero (2010), Remsen (2010) and Winker (2010) see subspecies as a meaningful rank, others consider subspecies to be a tool of convenience (Mayr 1982a, FitzPatrick 2010). Zink (2004) calls for the classification and rank names to reflect diagnosable units and hence considers subspecies to be of limited utility. Phillimore et al. (2010) contends that the concept has not been applied appropriately and objectively in that there has to be a sound understanding of the processes that govern the origination and extinction of 163 subspecies. Pruett and Winker (2010) highlighted the importance of using multiple lines of evidence in assessing subspecific status even though they acknowledged the challenge that can arise in cases of discord caused by varying evolutionary rates of character change. Another complicated issue that continues to challenge biologists is that of developing a species definition of universal application. As Darwin (1859) articulated there is vast vagueness in what naturalists mean by ‘species’, as a consequence a suite of species concepts have been postulated (reviewed in Mayden 1997, de Queiroz 2007) to date and a consensus is hardly met. de Queiroz (2007) postulated a “Unified Species Concept” which was meant to be a reconciliation of the various species concepts available to date. He defined species as separately evolving metasubspecies lineages or segments (Knowles and Carstens 2007). The properties that were previously treated as necessary properties for species, for example intrinsic reproductive isolation in the case of the Biological Species Concept; occupation of a distinct niche or adaptive zone in the case of the Ecological Species Concept; fixed character state differences in the case of the diagnosable version of the Phylogenetic Species Concept are no longer considered the defining properties of the species category, rather there is a continuum that extends between these alternate definitions of species that may relate to their relative evolutionary age. Taxonomic determinations made in this chapter The polytypic nature of several species of francolins is highlighted above and tabulated in Table 1.2. At the end of this chapter, it is expected that some of these taxa will be elevated to species rank, provided they represent distinct phylogenetic and 164 biogeographical eco-evolutionary entities diagnosable in terms of organismal and molecular characters. Other taxa will likely be subsumed into the same terminal taxon if they show relatively little phylogenetic distinctiveness and geographical and/or ecological partitioning. The primary grouping criterion followed in this study is multifaceted consilience in character variation (sensu Pruett and Winker 2010), because ultimately one has to draw the taxonomic line somewhere with the goal of finding meaningful evolutionary entities. Hybridization is known to occur even long after speciation (Mallet 2005, Mallet 2008) and hence is not over-emphasised as a criterion for taxonomic delineation. Thus, in this study, the decision to rank a taxon as a species, is considered if it is morphologically and/or behaviourally diagnosable (as defined above), ≥ 1-2% divergent in unweighted molecular characters, and confined to a specific eco-region. For a particular taxon to be considered a subspecies, it should exhibit relatively little phylogenetic distinctiveness and geographical and/or ecological partitioning, be < 1% divergent in unweighted molecular characters, and be confined to a specific eco-region. Objectives of the study:  To review the taxonomic status of the ‘true’ francolins, with a particular focus on African francolins.  To re-assess the monophyletic status of the various species groups proposed by Hall (1963).  To investigate the phylogenetic relationship between Hall’s (1963) enigmatic unplaced Asian species F. pondicerianus and F. gularis , and African species (F. lathami) in light of the findings of Cohen et al. (2012) and Forcina et al. (2012). 165  To produce a revised classification of African francolins that takes into account the evolutionary relationships among taxa. Materials and methods Data collection Morpho-behavioural characters of francolins In total, 24 organismal characters reflecting assessment of plumage/integument and colour/pattern (Fig. 5.7), as well as measurements of certain qualitative and quantitative structures (Table 5.1), and several vocal characters extracted from Crowe at al. (1992) and chapter 4, were scored (Table 5.2). Morphometric characters representing billlength, tarsus- and spur-length were obtained using a Vernier Calliper. A stopped wingrule and a normal ruler were used to measure wing- and tail-length, respectively. Winglength was measured with the wing chord flattened and straightened for enhanced accuracy. Molecular characters For within-group molecular analyses of ‘true’ francolins, 41 terminal taxa (including two outgroup species) were sampled (Tables 5.3) for the entire mitochondrial Cytochrome-b gene (CYTB - 1143 base pairs- bp). GenBank accession numbers of taxa sequenced are detailed in Appendix 5.1. Primers used in sequencing molecular markers are listed in Tables 5.4 and 5.5; 67% of DNA samples of francolins sequenced were derived from toe-pads sub-sampled from museum skins. For the toe-pad samples, instead of sequencing the CYTB gene in a single reaction using the available universal primers, multiple sequence fragments (six for each 166 sample) were generated (using primers specifically designed for francolins, Table 5.5) due to the degraded state of the DNA. This meant that 1143 bp for each toe-pad sample would only be recovered by assembling six different PCR-amplified and sequenced fragments. Maps and mapping of distribution records of investigated taxa The maps showing the distribution ranges of francolins were produced (Fig. 5.1 – 5.6). This was a challenging task given that the actual point locality data for most taxa were not available. Ranges of all the traditionally recognized species presented in Snow’s (1978) atlas were used as the basis for generating range maps (Chapter 5 and 6) because these ranges were produced from the point locality data associated with skins housed in museums. The distribution data in Harrison et al. (1997) were then overlaid on Snow’s ranges and this was very helpful in filling in the gaps in the ranges at least of the southern African species. The 2nd step used Hall’s (1963) distributions to cover the ranges of both species and subspecies that she recognized. Step 3 involved the superimposition of Mackworth-Praed and Grant’s (1952, 1962, 1970) ranges on Snow’s ranges to cover both species and subspecies that these authors recognized. Mackworth-Praed and Grant (1952, 1962, 1970) and Hall (1963) are the two main taxonomic revisions that covered the Africa and Africa and Asia species respectively, contrary to revisions in Roberts (1924), Clancey (1967) and Hockey et al. (2005) that covered smaller geographic areas in which francolins occur. In Step 4 Clancey (1967) and Hockey et al. (2005) were used to account for the distributions ranges of other populations as they recognized them. These two revisions covered the southern African region only. Distribution ranges in steps 2-4 were 167 generally superimposed on Snow’s ranges and gaps, which may either translate to the real gaps in distribution or are indicative of under-sampling in the intervening area. It is only with the exception of Peliperdix schlegelii where the range is not broken according to Snow since the species is confirmed to have a continuous distribution from eastern Cameroon through to Central African Republic, to the western Bahr-el-Ghazal in Sudan (Mackworth-Praed and Grant 1970, Hall 1963). With regard to the names of studied localities of taxa, old rural names were used to a certain extent when translating them posed difficulty. In order to prepare distribution ranges of taxa to be used in the spatial analysis of vicariance, distribution records were mapped on a continuous basis using a 2x2 degree grid on the world map which was accessed on [http://earthobservatory.nasa.gov/Features/BlueMarble/]. Data analyses Phylogenetic analyses The generated nucleotide sequences were edited and assembled in the Staden Package (Staden et al. 2003) and aligned in MAFFT (Katoh et al. 2009). All CYTB (coding genes) sequences generated in this study were checked for stop codons before they were analyzed by translating them into amino acids and this was done online at EMBL [http://www.ebi.ac.uk/emboss/transeq/]. Two phylogenetic inference methods with different optimality criteria were employed to generate phylogenetic hypotheses: maximum likelihood (CYTB) and parsimony (CYTB, organismal, combined CYTB/organismal characters). As suggested by the results of the much larger data set in Chapter 2, all data matrices were rooted on Gallus gallus and Bambusicola thoracica. For all analyses, characters were treated as non-additive. Parsimony-based phylogenetic 168 analyses were conducted using PAUP ver. 4.0b10 (Swofford 2002). The following settings were effected: full heuristic search with all characters unordered and with equal weight, starting tree(s) obtained via stepwise addition; tree-bisection-reconnection branch-swapping, 1000 random additions of taxa (Maddison 1991), one tree being held at each step during stepwise addition , branches were collapsed (creating polytomies) if the maximum branch-length was zero. When multiple, equally parsimonious cladograms were recovered, a strict consensus cladogram was constructed. The extent to which each non-terminal node is supported by different character partitions was determined by using the bootsrap (BS) (Felsenstein 1985) resampling strategy with 1000 pseudoreplicates, with 5 random additions of taxa per bootstrap pseudoreplicate. Since different codon positions evolve under different models of evolution, it has been argued that a partitioned, mixed-model approach should be adopted (Ronquist and Huelsenbeck 2003, Nylander et al. 2004). Mixed-model analyses allowed different parameters (base frequencies, rate matrix or transition/transversion ratio, shape parameter, proportion of invariable sites) to vary among the three codon positions. Mixed-model maximum likelihood analyses were performed using the Randomised Axelerated Maximum Likelihood algorithm for High Performance Computing (RAxML) v7.0.4 (Stamatakis 2006, Stamatakis et al. 2008) as implemented on the CIPRES portal. Mixed-model RAxML analyses make use of a GTR++ model partitioned by gene or codon postion. Support at nodes was assessed with 100 nonparametric bootstrap (BS) pseudoreplicates. The use of different methodological approaches (optimality criteria) facilitated the identification of method-based incongruence. 169 Genetic distances Uncorrected pairwise distances (Table 5.6) were calculated in PAUP ver. 4.0b10 (Swofford 2002) for the CYTB data matrix and were converted to percentage sequence divergence. Results and Discussion Separate versus combined phylogenetic analyses The parsimony analysis for the CYTB data set (with 1143 bp characters, yielded 315 parsimony informative characters and 110 variable characters that were parsimony uninformative. Four trees of 1356 steps were recovered. The organismal data set comprised 24 characters, 23 of which were parsimony informative, and one that was parsimony uninformative. In total, 390 trees of 104 steps were recovered. The combined CYTB/organismal data set comprised 1167 characters of which 334 were parsimony informative and 115 that were variable but parsimony uninformative. In total, 24 trees of 1491 steps were recovered. The systematics of Hall’s species groups and unplaced species Striated Group and the unplaced Asian species Traditionally, the Striated Group encompasses two recognized species, a polytypic Dendroperdix sephaena and a monotypic species Dendroperdix streptophora, and 12 subspecies assigned to D. sephaena: sephaena, spilogaster, somaliensis, schoanus, jubaensis, grantii, rovuma, zambesiae, chobiensis, thompsoni, zuluensis and mababiensis. 170 In comparing the phylogenetic trees (Fig. 5.8, 5.9, 5.10, 5.11), none of these analyses ever recovered the monophyly of the Striated Group as hypothesized by Hall (1963). In all the trees, D. streptophora joins the Red-winged Group whereas D. sephaena (labelled Ortygornis in the trees) is well separated from the Red-Winged Group and remains at the basal parts of the trees. These results confirm earlier suggestions of the distinction of these two taxa (Crowe and Crowe 1985 and Crowe et al. 1992). Dendroperdix sephaena (together with rovuma, grantii, see below) form a sister relationship in most analyses (Fig. 5.8, 5.9, 5.11, respectively) with the two unplaced Asian species F. pondicerianus and F. gularis, which in turn emerge as sister species with high support. The organismal tree (Fig. 5.10) does not support the phylogenetic association of Dendroperdix taxa with F. pondicerianus and F. gularis, but instead places D. sephaena with other members of the genus Francolinus. Francolinus gularis differs from F. pondicerianus with 5% sequence divergence (Table 5.6), from sephaena with 8%, and there is 9% divergence between pondicerianus and sephaena. Francolinus gularis and F. pondicerianus diverged from each other 3.1 mya and both species diverged from D. sephaena at 5.5 mya – Fig. 2.7). Francolinus gularis is quite different in colouration and patterning (with barred back) from the other Asian francolins. The white streaks on the belly strongly contrast with the rufous throat and wings. Francolinus pondicerianus is finely patterned with the nominate subspecies pondicerianus being the darkest and the least grey and has the greatest amount of chestnut markings. The consistent grouping of F. pondicerianus and F. gularis with the Dendroperdix taxa (Fig. 5.8, 5.9, 5.11, 2.2, 2.3, 2.5, Bloomer and Crowe 1998) refutes the finding in Forcina et al. 2012 that the five Asian species (F, francolinus, F. pictus, F. pintadeanus, F. pondicerianus, F. gularis) are monophyletic. 171 Therefore, it is recommended that F. pondicerianus and F. gularis be considered congeneric and be placed in the genus Ortygornis Reichenbach, 1853 over Dendroperdix Roberts, 1922 based on priority, and because Ortygornis was once used for pondicerianus (Roberts, 1940). Therefore, there will be Ortygornis pondicerianus (Gmelin, 1789) and O. gularis (Temminck, 1815) (Appendix 5.2). The Dendroperdix taxa form a monophyletic assemblage with sephaena being sister to rovuma and grantii (Fig. 5.8, 5.9, 5.11) with moderate to high support (see rationale below for splitting these taxa from D. sephaena). Rovuma and grantii, are in turn, sister to each with high support in most analyses (Fig. 5.8, 5.9, 5.11, respectively) with the exception of the organismal only dataset (Fig. 5.10), where relationships among these taxa are unresolved. With respect to uncorrected pairwise genetic divergence, rovuma differs from sephaena by 4%, and sephaena differs from grantii by 5%, and rovuma differs from grantii by 4%. Dendroperdix sephaena can be categorized into two distinct types, the coastal types that have fine drop-shaped light chestnut streaks on the belly and the inland type with no streaks on the belly. Grantii has a reduced band of chestnut triangular marks on the breast, narrow greyish somewhat U-shaped streaks on breast and belly with the distal part of the belly being unpatterned. Rovuma has a band of chestnut triangular marks being reduced without covering much of the breast, no barring on the belly, has drop-shaped chestnut streaks on the belly and not much distally. In the north, the Ethiopian subspecies spilogaster is barred and streaked on the breast with barring being extensive as in sephaena. The Somalian subspecies somaliensis is like spilogaster, but the streaking is not extensive. There is confirmed evidence of hybridization taking place in the north between the southern subspecies rovuma and the northern spilogaster (Hall 172 1963, Snow 1978). The distribution range of rovuma and spilogaster is intercepted by that of grantii such that the two streaked subspecies are isolated geographically. It was not possible to fully account for the taxonomic status of spilogaster in the absence of genetic information. Given the difference in morphology between spilogaster and rovuma and the geographic isolation between the two subspecies, spilogaster (in the north) and rovuma (in the south) both separated by grantii, the recommendation is that rovuma (once recognized as a species in Roberts 1924) be considered a separate species within which two subspecies rovuma and spilogaster are recognized. Furthermore, since grantii, jubaensis and schoanus are morphologically inseparable and genetic evidence showed that little genetic distance exists between grantii and jubaensis and schoanus, they are synonymized with grantii which is recognized as a full species. The taxa, sephaena, zambesiae, chobiensis, zuluensis, thompsoni and mababiensis all have buff bellies (with very narrow greyish barring) and a broad band of chestnut triangular markings on the breast and an unpatterned belly. Based on similar morphological attributes and the availability of genetic evidence for some subspecies, zambesiae, chobiensis, zuluensis, thompsoni and mababiensis should be synonymized with sephaena. Based on the phylogenetic association of the Dendroperdix taxa with the Asian species Ortygornis gularis and O. pondicerianus, it is recommended that the genus Ortygornis Reichenbach, 1853 replace Dendroperdix Roberts, 1922 based on priority. Therefore, the taxonomic status of the African members of this genus would be: Ortygornis sephaena (Smith, 1836), Ortygornis rovuma with subspecies Ortygornis rovuma rovuma Gray, 1867 and Ortygornis rovuma spilogaster Salvadori, 1888 and Ortygornis grantii Hartlaubi, 1866 (Appendix 5.2). 173 The unplaced Francolinus lathami This is one of Hall’s (1963) enigmatic species, that she could not place in any of her groups even though she speculated that it is closest to the Red-tailed francolins (Peliperdix spp.) and that morphological similarities between it and the spotted Asian Black Francolin Francolinus francolinus are superficial or plesiomorphic. Two putative subspecies, lathami and schubotzi are recognized. The underparts of nominate lathami are largely black with white spots in the male as opposed to a brown belly colour with white spots in female. Both sexes have a black throat and patterned face (greyish in male and brownish in female). The upperparts are mottled rufous and brown with pronounced white streaks (margined mostly with black) restricted to the lower neck. There is geographical variation in plumage between the nominate subspecies and schubotzi. The males of schubotzi have the black and white pattern extending down the bulk of the belly continuing to the undertail coverts (Chapin 1932, Hall 1963), where the ground colour still predominates. The females cheeks are more rufous than grey. The species exhibits striking plumage dichromatism and is markedly distinct from any of the African francolins. The phylogenetic inferences place F. lathami near the base of the tree in nearly all the analyses (Fig. 5.8, Fig 5.9, Fig 5.11). The two putative subspecies lathami and schubotzi are supported as sister taxa in all the analyses with high support. Wolters (1975-82), Crowe and Crowe (1985) and Crowe et al. (1992) supported Hall’s (1963) suspicion that F. lathami has affinities with members of the Red-tailed Group, with Crowe and Crowe (1985) and Crowe et al. (1992) placing this species in the subgenus Peliperdix. In our analyses, only the organismal data matrix places F. lathami close to members of the Red-tailed Group (Fig. 4.10). Dating of the molecular characters 174 suggests Francolinus lathami and the Peliperdix taxa split a long time ago, c. 7.6 mya (Fig. 2.7). The two recognized subspecies, lathami and schubotzi differ by 1% in sequence divergence (Table 5.6). Francolinus lathami links the relatively basal Asio-Afrotropical francolins and the quail-like Red-tailed Peliperdix and Red-winged Scleroptila taxa. On the basis of its distinct phylogenetic placement, morphology, unique forest habitat, geography and large genetic divergence, lathami should be recognized in a genus of its own and it should continue to comprise two subspecies lathami and schubotzi. Therefore, the genus Francolinus Stephens, 1819 will be replaced with a newly erected genus Afrocolinus gen. nov. Afrocolinus is formed from two words, afro- meaning African and -colinus for ‘quail’. The subspecies should be Afrocolinus lathami lathami Hartlaub, 1854 and Afrocolinus lathami schubotzi Reichenow, 1912 (Appendix 5.2). Red-tailed group This group is at represent represented by three species, Peliperdix coqui, P. albogularis and P. schlegelii. Within P. coqui there are 14 recognized subspecies: coqui, spinetorum, buckleyi, maharao, ruahdae, hubbardi, angolensis, lynesi, vernayi, campbelli, thikae, kasaicus, hoeschianus and stuhlmanni, and five subspecies are recognized within P. albogularis: albogularis, buckleyi, dewittei, meinertzhageni and gambagae. Peliperdix schlegelii is at present considered a monotypic species, but was once considered a subspecies of Peliperdix ‘Francolinus’ coqui (Peters 1934). The Red-tailed Group is recovered as monophyletic in all analyses (Fig. 5.8, 5.9, 5.10, 5.11) with moderate to high support, with the exception of the organismal analyses where support is lacking. This finding corroborates those of Crowe and Crowe (1985) 175 and Crowe et al. (1992). What stands out within the Red-tailed Group is the split between the P. schlegelii/P. albogularis complex and the rest of the Peliperdix taxa in all the analyses. Peliperdix coqui consists of a number of geographical variants and exhibits sexual dichromatism in plumage (Hall 1963). The various subspecies all have bellies mostly narrowly barred (except ruahdae with broader, blacker and more widelyspaced barring), though the degree of barring is variable and the females have a pink wash on their upper belly. Both sexes generally have grey wings, with the exception of kasaicus and vernayi that have redder, rufous and pinkish wash in the wings, respectively. The colour of the belly varies from being buffish white in coqui, vernayi, ruahdae, stuhlmanni, to tawny in kasaicus. The degree of barring in stuhlmanni is intermediate between that in the nominate coqui and hubbardi in that the barring is not pronounced especially on the belly. Within P. coqui, the various subspecies form a cline in which both sexes of the west African spinetorum have unbarred bellies, the east African hubbardi has barring on the flanks leaving the centre of the belly unbarred, the east African maharao is like the southern African P. coqui subspecies in having wholly barred bellies. Among the recognized subspecies of P. coqui, there is a geographical split between the east/west African subspecies (maharao, hubbardi, spinetorum) and central/southern African subspecies (coqui, vernayi, stuhlmanni, kasaicus) (Fig. 5.8, 5.9, 5.11). However, nodal support for the east/west African clade is only recovered in the ML CYTB tree (Fig. 5.8 - 69% BS). What is also supported in the CYTB parsimony tree (Fig. 5.10, 4.11) is the sister relationship between maharao and hubbardi (Fig. 5.9 57%, Fig. 5.11, 5% sequence divergence). In the CYTB ML tree maharao is sister to spinetorum with poor BS support (Fig. 5.9), whereas hubbardi and spinetorum are sister 176 taxa in the organismal tree (55% BS, 6.4% sequence divergence; Fig. 5.10). Coqui pairs with vernayi in the CYTB ML (93% BS, 3% sequence divergence) and CYTB parsimony (70% BS) tree, whereas coqui pairs with ruahdae in both the organismal and CYTB/organismal trees (all 64%). Stuhlmanni and kasaicus are sister taxa in the CYTB parsimony tree but with no BS support; they differ by 5% sequence divergence. Their systematic relationship is unresolved in the ML CYTB and organismal tree. Overall, the nearest non-coqui CYTB phylogenetically close and with the least genetic distance to coqui is hubbardi, differing at 5% sequence divergence. It is recommended that lynesi, campbelli and angolensis be included in coqui based on morphological similarity, phylogenetic affinities and that there exists limited genetic divergence (<1%), and that vernayi (Roberts, 1932) be considered a subspecies within P. coqui based on their morphological differences (the red wings as opposed to grey wings in nominate coqui, as well as the much redder crown as opposed to the less red crown of coqui). their phylogenetic affinity and small genetic divergence from coqui (3%), It should therefore be recognized as a subspecies Peliperdix coqui vernayi (Roberts, 1932). Hoeschianus Stresemann, 1937 should be synonymized with vernayi based on similar overall morphological appearance and geographic proximity. Specimens of ruahdae were not examined directly and as such the morphological description in Mackworth-Praed and Grant, and Hall, was used to classify the Ugandan ruahdae and hoeschianus. The genetic distance between ruahdae (tissue from Rwanda) and the South African nominate coqui is 4%. This is not surprising since the two subspecies are morphologically different (ruahdae having broad, much black and spaced barring on the belly) and they are geographically isolated. This taxon should 177 therefore be considered a subspecies within P. coqui to be named Peliperdix coqui ruahdae van Someren, 1926. Stuhlmanni could be a cryptic taxon that morphologically looks like an intermediate between coqui (strongly barred) and thikae and hubbardi (barring restricted to flanks). The rest of the belly is not strongly barred as in coqui, but the barring covers the rest of belly and not just the flanks as in thikae and hubbardi. Stuhlmanni is supported as sister to kasaicus in the parsimony CYTB and parsimony CYTB/organismal tree even though with no BS support and it is very divergent genetically (~5-8%) compared to any other P. coqui subspecies. The two subspecies should be considered species and hence Peliperdix stuhlmanni Reichenow, 1889. Perhaps kasaicus, like stuhlmanni, is a cryptic taxon based on the genetic distance, geographic isolation, and phylogenetic placement despite the fact that it has similar morphology to coqui. Kasaicus should be considered a species based on the geographical isolation with vernayi and be named Peliperdix kasaicus White, 1945. The western spinetorum and east African maharao and hubbardi form a clade. Spinetorum is unarguably evolutionarily distinct, geographically isolated from the rest of P. coqui subspecies, genetically different from other subspecies by 6-9%, and both sexes have plain unbarred bellies, but have rufous wings as in maharao. It should therefore be considered a species and be named Peliperdix spinetorum Bates, 1928. The isolated Kenyan subspecies maharao has its closest phylogenetic association with hubbardi. On the other hand, there is considerable genetic divergence (5%). Unlike hubbardi, the rest of the belly of maharao is barred. On the basis of the remarkable genetic distance, differences on the belly and the isolated distribution, 178 maharao should be considered a species and be named Peliperdix maharao Sclater 1927. Thikae and hubbardi have a genetic difference of 3%, occur side by side geographically and have the similar belly morphology despite the difference in wing colour (rufous in thikae and grey in hubbardi). Therefore, hubbardi Ogilvie-Grant, 1895 should be recognized as a species and thikae Grant and Mackworth-Praed, 1934 must be synonymized with it and be named Peliperdix hubbardi Ogilvie-Grant, 1895 (Appendix 5.2). Peliperdix schlegelii is the sister species to the P. albogularis complex and it shares the least sequence difference (1%) with buckleyi (Table 5.6). There is no nodal support for the P. schlegelii/P. albogularis complex clade in the organismal tree. Within the P. albogularis complex, albogularis and buckleyi are sister taxa (Fig. 5.8, 5.9, 5.11 – 99%, 100%, 100% BS respectively) whereas albogularis, buckleyi and dewittei are similar in the organismal tree (Fig. 5.10). The taxon dewittei/ ‘meinertzhageni’ forms a sister relationship with albogularis (with 5% sequence divergence) and buckleyi with high support in the molecular analyses (Fig. 5.8, Fig. 5.9, 5.11). The nearest relative of buckleyi is albogularis with 1% sequence divergence. The Peliperdix albogularis complex has a disjunct distribution across west and west-central Africa with the nominate subspecies being restricted to Guinean savanna in West Africa between Senegal and Cameroon (Cotterill 2006). Peliperdix albogularis resembles P. c. coqui, but the quail-type patterning is less well-defined in the females and the shaft streaks and barring are much narrower. The wings are more rufous as in the east Kenyan subspecies of coqui. The females (but not the males) also have the black facial pattern and necklace, but both are poorly-defined. Both sexes have a buff 179 throat contrasting with the ochre cheeks. The rest of the underparts of male albogularis are quite distinct from any other forms of coqui, being chestnut on the upper belly with ochre shaft streaks and richer ochre on the lower belly, lacking any dark barring. The females are barred (to varying degrees) with a faint pinkish wash or rufous on the upper belly similar to the Ethiopian subspecies of P. c. maharao and the east Kenyan P. c. thikae. The subspecies in which the females are less patterned with narrow barring restricted to the upper belly and flank is in Gambia (albogularis – Hall 1963). There is a subspecies that Peters (1934) recognized as gambagae (Mackworth-Praed, 1920) from the type locality Gambaga, Gold Coast Colony) which Hall and Serle (1957) synonymized with buckleyi. The taxon dewittei (including meinertzhageni) occurs south of the Congo forest in moist montane grasslands and dambo floodplains (Cotterill 2006) and it is considered common in some parts, even though it was said to not have been collected on the Marungu plateau since 1931 (Dowsett and Prigogine 1974). Birds of both sexes are larger than albogularis (wing length ranging from 142 to 147 mm – Mackworth-Praed and Grant, 1970) and more richly coloured than both albogularis and buckleyi with the female dewittei being more heavily barred although this is not as extensive as in buckleyi. The eastern Angolan and possibly the northwestern Zambian taxon meinertzhageni comprise the darkest birds with heavily barred females. This subspecies is found in montane grassland and it is restricted to the Upper Zambezi floodplains (Cotterill 2006). The taxon dewittei is indisputably morphologically different from the other subspecies recognized within P. albogularis, and is divergent genetically. The morphology of dewittei (southeastern Congo) resembles that of the Angolan and northwestern Zambian ‘meinertzhageni’ (both sexes are large birds with a wing length 180 of 140-147 mm, rich dark-coloured birds, heavily patterned). Unfortunately, dewittei could not be sequenced. Geographically, the two subspecies occur in different parts (but occupy similar habitat). Hall (1963) speculated that the northwest Zambian meinertzhageni subspecies may be more closely related to the Congo dewittei than the Angolan meinertzhageni subspecies, this speculation may have been based on the geographical proximity. Based on morphology, geographical and molecular evidence for the Angolan meinertzhageni (in the absence of molecular evidence from northwest meinertzhageni subspecies), meinertzhageni White 1944 should be moved to species level. It is not possible at this stage to comment on Hall’s speculation about the relationship of the north-west Zambia meinertzhageni and the Congo dewittei subspecies. Despite the absence of molecular evidence for dewittei Chapin, 1937, its name has to be given the priority and meinertzhageni will be synonymized with it. The name must be Peliperdix dewittei Chapin, 1937. The West African buckleyi and gambagae are inseparable morphologically and they are almost identical genetically with 0.4% difference. The two subspecies should be merged and form a subspecies with the name buckleyi Ogilvie-Grant, 1892 taking priority over gambagae (Mackworth-Praed, 1920). The recognized name must be Peliperdix albogularis buckleyi Grant, 1892. So, Peliperdix albogularis Hartlaub, 1854 will be represented by two subspecies, Peliperdix albogularis albogularis Hartlaubi, 1854 and Peliperdix albogularis buckleyi Ogilvie-Grant, 1892 (Appendix 5.2). Another controversial taxon within the Red-tailed Group is Peliperdix schlegelii. This taxon was considered a subspecies of P. coqui in Chapin (1932) and Peters (1934). Contrary to this, Hall (1963) and Mackworth-Praed and Grant (1970) recognized this 181 taxon as a species. This is a rare bird; its distribution stretches from eastern Cameroon through the Central African Republic to the western Bahr-el-Ghazal in Sudan. The back plumage colouration and patterning of P. schlegelii is closer to that of P. albogularis than P. coqui, though the quail-type patterning is much reduced and the extent of sexual dimorphism is also pronounced. The differences are seen in the male P. schlegelii, which has broad buff shaft streaks and some streaks with few transverse blackish brown bars whereas the female is almost unpatterned. The back in both sexes is vinous chestnut with the belly being buff to off-white. The male P. schlegelii is similar to southern P. coqui subspecies with ochre sides of the head and throat and narrow black and white barring on the upper and lower belly. The female resembles the male on the crown and throat but has the upper belly feathers edged with black, giving a more mottled appearance of simply triangular marks. The flanks are sparsely barred. The findings demonstrated that P. schlegelii is an evolutionarily distinct taxon both morphologically and genetically. It is more closely related to P. albogularis than it is to P. coqui. Therefore, P. schlegelii must be given full specific status and be named Peliperdix schlegelii Heuglin, 1863. These findings suggest that Chapin (1932) and Peters (1934) were conservative in considering this taxon a subspecies of P. coqu (Appendix 4.2). Red-winged Group The Red-winged Group is represented by six species, Scleroptila shelleyi, S. afra, S. levaillantoides, S. levaillantii, S. finschi and S. psilolaema. Scleroptila afra and S. finschi are monotypic species whereas the balance of the species are polytypic. Scleroptila shelleyi comprises seven subspecies (shelleyi, uluensis, whytei, macarthuri, 182 trothae, sequestris, canidorsalis); S. psilolaema four subspecies (psilolaema, ellenbecki, elgonensis, theresae); S. levaillantii five subspecies (levaillantii, kikuyuensis, crawshayi, benguellensis, clayi); S. levaillantoides 12 subspecies (levaillantoides, kalaharica, pallidior, langi, wattii, jugularis, cunenensis, stresemanni, gutturalis, lorti, archeri, ludwigi). The Red-winged Group is monophyletic in all the analyses (Fig. 5.8, 5.9, 5.10, 5.11) with varying levels of support. One further common aspect, is the incorporation of S. streptophora as a basal taxon to the group in all the analyses (Fig. 5.8, 5.9, 5.10, 5.11). This finding is consistent with that of Crowe and Crowe (1985) and Crowe et al. (1992). This group presents a very complicated and variable phylogenetic structure with few nodes are consistently supported in most of the analyses: (1) kikuyuensis and crawshayi and (2) psilolaema and theresae. The major taxonomic change relevant to this group has been the shift of the monotypic Scleroptila streptophora from the Striated to the Red-winged Group (see above). It has a markedly disjunct distribution occurring in savanna-grassland in Cameroon in the west, and northern Uganda and northwestern Kenya and Tanganyika in the east. It differs from most of the other Red-winged francolins in having an unpatterned crown, throat colour buff rimmed with rufous, the absence of a gorget, barred breast, blotched and mottled belly, brown (not red) primaries, unpatterned side of the head, and tiny spur bump. It differs from uluensis and shelleyi by 8% sequence divergence. The monotypic specific status of S. finschi has never been disputed. This species is found disjunctly in western Angola and parts of western Congo (Mackworth-Praed and Grant 1962, Hall 1963). It differs from the other Red-winged francolins in lacking 183 black and white patterning on the face and neck and by having the upper belly and the lower end of the lower belly grey. The centre of the belly is buff with chestnut blotching and it is essentially unbarred. The sides of the face and border of the throat are ochre as in S. levaillantii but this is less extensive on the hind neck. This species is long-billed like S. levaillantii. It shares the least genetic distance with shelleyi (5% sequence divergence – Table 4.6). Scleroptila shelleyi has a buff to white throat, and a distinct black necklace, a moderately developed gorget, blotched and mottled upper belly and a broadly barred lower belly. It occurs in moist, rocky hillsides below 2000 m throughout southeastern Africa from Mozambique in the north to South Africa, and shares the least genetic distance with gutturalis (4% sequence divergence). Gutturalis was considered by Hall as one of the northeastern subspecies (lorti and friedmanni, stantoni, eritreae, archeri) attributed to S. levaillantoides. It occurs in arid savanna steppe in eastern Ethiopia, Eritrea and northeastern Kenya, and differs from other northern Red-winged francolins in having the lower belly largely unpatterned with bars and throat base freckled, with dark above a well-defined necklace and gorget. It shares the least genetic distance with uluensis (3%). Uluensis occurs in central Kenya, southern Uganda and northeastern Malawi. It is similar in form and habitat to S. shelleyi, but the necklace and gorget are somewhat less distinct and the bill markedly shorter. Some specimens have the throat moderately freckled with black indicating possible hybridization with gutturalis. The Malawian taxon whytei is very different from shelleyi and uluensis in having a rufous buff throat and an indistinct necklace. It has a moderately-developed gorget blotched with reddish chestnut distinguishing it from the other Scleroptila spp. The belly is buffy 184 white with broad black bars on a background of rufous buff. Its nearest CYTB relatives are finschi, gutturalis and psilolaema at 5% sequence divergence. Scleroptila afra is endemic to southern Africa generally occurring in montane grasslands above 1800 m. It differs from uluensis, shelleyi and whytei in having a buff to white throat flecked with black. The reddish colour which is apparent in the wings of other scleroptilids is extremely reduced to absent. The degree of black barring on the buffy belly is very narrow whereas the barring on the underparts of the others is broad. The gorget in S. afra is well-defined and eventually grades to the upper belly which forms a band mottled with grey, chestnut and tawny. The former eastern Transvaal province of South Africa subspecies proximus (which does not differ genetically from those to the south) is according to Clancey (1967) “similar to afra, but darker and more saturated”. “On the underside, rather deeper tawny colour to the ground of the lower throat and upper breast, the grey tipping to the feathers more extensive and darker; rest of the underside more yellowish tinged, and chestnut spotting darker and heavier”. The nearest relative is ellenbecki differing at 2% sequence divergence. Ellenbecki is very similar to afra in having a buff throat freckled with black, a poorly developed necklace, but differs in having broad black and white barring on the lower belly. It inhabits montane grassland in southern Ethiopia above 2500 m. Scleroptila psilolaema and theresae are typical Red-winged francolins and are similar in appearance in having the most red in their primaries, but psilolaema is significantly smaller and shorter billed. Both inhabit montane grassland above 2500 m, the former from southern Ethiopia and the latter from Mt. Kenya and the Aberdares. They differ from ellenbecki in having little or no freckling on the throat and have a much more poorly developed necklace and gorget which is replaced by rich chestnut 185 with black indistinct spots. Scleroptila psilolaema differs from theresae with 2.6% and from pallidior at 5% sequence divergence. Scleroptila levaillantii is endemic to rank highland (1600-2000 m) grasslands of southeastern Africa. It is the largest of the southern African francolins and differs from the other Red-winged francolins in having an ochre collar on the sides of the face and edges of the throat and also inside the black and white facial pattern. Its face is like that of S. levaillantoides, but this species has a long bill, it is darker and richer in colour. The black and white patterning of the well-defined necklace (forming a well-developed gorget) extends in a complete collar round the hind neck below the ochre collar and the black and white stripes from above the eye (which in S. levaillantoides and S. shelleyi run down the side of the face) in the South African S. levaillantii subspecies behind the head to join at the back. It also differs from its congeners in having more red on the wings, a longer bill and shorter spurs. It differs from ellenbecki at 6.6% sequence divergence. On the other hand, S. levaillantoides is a species endemic to the arid grasslands of northern Namibia, Botswana and central South Africa, occurring at lower altitudes than shelleyi, levaillantii and afra. It is also smaller than these taxa. It has a moderately developed gorget with the upper belly blotched with buff and reddish brown. The generally paler subspecies pallidior is still tentatively placed with levaillantoides despite its genetic affinity with ellenbecki (9% sequence divergence) and that it is otherwise identical to the nominate subspecies in its overall biology. It shares the least genetic distance with afra at 3% sequence divergence. The taxonomic treatment renders S. levaillantoides paraphyletic in molecular tree. Jugularis is according to Hall (1963) endemic to southeastern Angola and northwestern Namibia and classified as a 186 subspecies of S. levaillantoides. It differs from this taxon in being paler overall and having a well-developed necklace which grades to the breast to form a broad gorget. It differs from crawshayi at 4% sequence divergence. Crawshayi is endemic to northeastern Malawi and western Tanzania and is richer in colour than levaillantii, with still more red in the wing, and with more black marking on the belly, whereas kikuyuensis (which occurs in Angola and southern Congo – Fig. 5.6) is without the stripe on the hind neck, but otherwise similar to crawshayi. Both have similar habitats to levaillantii. The nearest relative of crawshayi is afra which differ at 3% sequence divergence. The nearest relative of kikuyuensis is crawshayi with 2% sequence divergence. The taxon theresae was considered a subspecies of S. shelleyi by Mackworth -Praed and Grant and a synonym of psilolaema by Hall (1953). Since it is sister to psilolaema here and is 3% divergent from it (as opposed to 6% to shelleyi), Hall appears to have been correct. On the other hand, ellenbecki was recognized by MackworthPraed and Grant as a subspecies of S. afra, but synonymized with psilolaema by Hall. The genetic difference between ellenbecki and S. afra is 2%, and 5% between ellenbecki and psilolaema. Moreover, S. afra and ellenbecki indeed have morphological affinities, i.e. they have similar underparts with freckled throat and barred bellies. So, in this instance Hall appears to have erred in her decision. One other controversial aspect is the taxonomic status of psilolaema. Hall considered this a species, whereas MackworthPraed and Grant considered this taxon a subspecies within S. afra. Genetically, psilolaema differs from S. afra by 6% and the two have no apparent phylogenetic affinity. 187 Within the Scleroptila levaillantoides, the genetic distance observed between the three northern subspecies that Hall (1963) attributed to levaillantoides (lorti, archeri, gutturalis) is in the range of ~3.1-3.4% to S. levaillantoides and ~3-4% to S. shelleyi. Among themselves they differ by <2%. These three subspecies are phylogenetically related with archeri and lorti being sisters (results not shown on the cladogram) and they are in turn sister to gutturalis. Since uluensis is genetically closest to guttaralis, both seem to have been incorrectly placed by Hall and warrant species status. The southern jugularis shares sister relationship with whytei in all the analyses (with 77% BS support only in the combined CYTB/organismal tree) except in the organismal tree and it differs from the other southern subspecies (levaillantoides, pallidior), the northern (gutturalis) subspecies, as well as shelleyi by ~7% sequence divergence. It is closest to the nominate levaillantoides, only in the organismal tree but differs by 7% of genetic distance. Despite having a well-developed gorget, jugularis has some similarities with pallidior in that it has a white to buffy white belly with dark chestnut blotches. Pallidior differs by ~3-4% from levaillantoides, gutturalis and S. shelleyi. It has the closest phylogenetic affinities with S. afra and ellenbecki (though not supported) and the two taxa differ genetically by 2% and 3% respectively. Since pallidior has little morphological resemblance to S. afra and ellenbecki it is still tentatively placed with levaillantoides, subject to further phylogeographic study. Scleroptila levaillantii differs genetically from kikuyuensis and crawshayi by 9 and 8% respectively, and a 3% difference exists between kikuyuensis and crawshayi. Scleroptila shelleyi remains a valid species including the Natal subspecies sequestris Clancey, 1960 and the Inhambane, southern Mozambique subspecies canidorsalis (Lawson, 1963). The Malawi subspecies whytei should be recognized as a separate 188 species based on the remarkable difference observed in morphology, genetic distance and distinct phylogenetic placement and geographic isolation, and hence be Scleroptila whytei Neumann, 1908. The phylogenetic position of uluensis does not favour Mackworth-Praed and Grant’s classification system, that is, uluensis is not phylogenetically related to S. afra even though the two are in the same clade in the parsimony CYTB/organismal tree. They have a genetic distance of 4%. Despite that the morphological divergence and the geographic isolation of the distribution ranges of the uluensis and S. afra are difficult to reconcile. Therefore, uluensis has to be recognized as an evolutionarily distinct taxon and is upgraded to species Scleroptila uluensis (Grant, 1892) with macarthuri (van Someren, 1938) being included as a subspecies (Appendix 5.2). Mackworth-Praed and Grant’s classification of psilolaema as a subspecies of S. afra is probably not correct. The two are phylogenetically and geographically apart, have a genetic difference of 6% and differ morphologically except for the presence of the grey flecks on the throat of S. afra and indistinct black spots on the throat of psilolaema. Thus, Hall (1963) was probably correct in recognizing S. psilolaema as a valid species, Scleroptila psilolaema Gray 1867. The taxonomic status of theresae is also uncertain. Mackworth-Praed and Grant considered it a subspecies of S. shelleyi whereas Hall included it in S. psilolaema. Theresae is sister to psilolaema and the two differ genetically by 3% (and 6% when compared to S. shelleyi). Also, theresae is morphologically similar to psilolaema and not to S. shelleyi. It has rich chestnut in wings and differs from S. shelleyi in having some barring on the tips of the primaries and the underparts are rich buff mottled with chestnut and with some less defined brown to black bars. The gorget is poorly-defined 189 and replaced by rich chestnut with black spots and has lighter chestnut throughout the belly. Therefore, theresae should be considered a subspecies, Scleroptila psilolaema theresae (Meinertzhagen, 1936) with elgonensis (Grant, 1891) being synonymized with it. The phylogenetic position of ellenbecki and the small genetic divergence between ellenbecki and S. afra (2%) certainly confirm Mackworth-Praed and Grant’s classification system. Hall included ellenbecki in S. psilolaema (which differ genetically by 5%). Careful examination of the specimens revealed that S. afra and ellenbecki both have similar underparts with freckled throats and barred bellies. The major difference is that the width of barring on breast and belly is finer in S. afra and broader in ellenbecki. Furthermore, ellenbecki has rich darker chestnut in wings (opposed to grey in S. afra) and some barring at the tips of the primaries. It has a poorly-defined gorget (as opposed to a well-developed gorget in S. afra) replaced by rich chestnut and has dark chestnut throughout the belly different from the narrow black barring in S. afra. In consideration of the similarity in morphology, the small genetic distance and the phylogenetic affinities between ellenbecki and S. afra, and geographical isolation, it is recommended that ellenbecki be considered a species closely related to S. afra and hence Scleroptila ellenbecki Erlanger, 1905. Gutturalis should be recognized as a separate taxon, i.e. as a species Scleroptila gutturalis (Rüppell, 1835), and gutturalis takes priority over archeri Sclater, 1927 and lorti Sharpe, 1897. Friedmanni Grant and Mackworth-Praed, 1934 and stantoni (Cave, 1940) should be synonymized with archeri and eritreae Zedlitz, 1910 with gutturalis. Scleroptila levaillantoides levaillantoides Smith, 1836 remains the valid nominate subspecies with ludwigi Neumann, 1920 and gariepensis Smith, 1843 being included in it. This subspecies is strongly marked on the upper and lower belly and 190 there exists large genetic difference when compared to pallidior (4.2%). Pallidior Neumann, 1908, wattii Macdonald, 1953, langi Roberts, 1932 and kalaharica Roberts, 1932 should be considered a single taxon, i.e. subspecies within S. levaillantoides based on its morphological attributes (strongly marked on upper belly with a relatively unmarked lower belly) irrespective of its phylogenetic position and the least genetic difference when compared to S. afra. Scleroptila levaillantoides pallidior Neumann 1908 takes priority over the other names. From the south, jugularis Büttikorfer, 1889 generally looks like the nominate levaillantoides, but it is relatively unmarked on the lower belly, posses a well-developed gorget, there is remarkable genetic difference between the two and they are wellsupported sister taxa (99%). So, jugularis should be recognized as a species, Scleroptila jugularis Büttikorfer 1889. This name receives priority over cunenensis (Roberts, 1932) and stresemanni Hoesch & Niethammer, 1940. There are marked morphological difference between levaillantii and the other subspecies, kikuyuensis, crawshayi, benguellensis, mulemae and clayi. Generally, differences concern whether a taxon has either rich reddish chestnut colour in the belly feathers, a broken or unbroken ochre collar, rufous band across nape (in clayi) and a streak below the eye with less black spotting and some irregular black patches on the belly in benguellensis. Among the subspecies for which there are DNA sequence data, the genetic divergence is remarkable between levaillantii and the two sister subspecies kikuyuensis (9%) and crawshayi (8%) even though they appear similar morphologically. The recommendation is that kikuyuensis and crawshayi be considered as separate taxa taking the status of a species i.e. with crawshayi - Ogilvie-Grant 1896 receiving priority 191 over kikuyuensis Ogilvie-Grant, 1897 and becoming the nominate subspecies and hence the name Scleroptila crawshayi crawshayi Ogilvie-Grant, 1896. Morphologically crawshayi and kikuyuensis are similar except that kikuyuensis is without the stripe on the hind neck. Thus, benguellensis and mulemae are similar to kikuyuensis. There is no genetic evidence for the recognition of clayi, benguellensis and mulemae the decision is based on morphological characters and the locality of the taxa. Since benguellensis Neumann, 1908, clayi White, 1944 and mulemae Ogilvie-Grant, 1903 are said to be similar to kikuyuensis these subspecies should be synonymized with crawshayi due to the geographic proximity and not with kikuyuensis. Kikuyuensis is isolated from crawshayi and has some morphological differences and therefore, it should be considered a subspecies within S. crawshayi O. Grant, 1896, and named Scleroptila crawshayi kikuyuensis Ogilvie-Grant, 1897. Scleroptila finschi and S. streptophora are the two distinct species with no geographic variation and their species status is not disputed (Appendix 5.2). Conclusions This study has attempted to comprehensively revise the systematics of francolins. There is congruence in the topologies resulting from the various analyses, but some discordance in certain areas of the trees, especially within the species groups, remains. What is apparent is that the combined analyses yielded better phylogenetic results. One of the major finding is that the Striated Group is not monophyletic as hypothesised by Hall (1963) whereas the Red-tailed and Red-winged Groups were largely diagnosable. The consistent phylogenetic placement of the two previously unplaced Asian species Ortygornis pondicerianus and O. gularis as the closest relatives of the African 192 Ortygornis taxa strongly rejects the monophyletic status of the five Asian francolin species as hypothesized by Forcina et al. (2012). Francolinus lathami has distinct evolutionary trajectory, hence it is recommended that it is assigned it own genus Afrocolinus gen. nov. 193 Tables and Figures Table 5.1. Morpho-behavioural characters and character states scored for phylogenetic analysis of francolins. 1. Crown patterning: unpatterned = 0; streaked = 1; mottled = 2 2. Chestnut collar: none = 0; chestnut = 1 3. Upper back patterning: spotted = 1; streaked = 2; barred or barred and streaked = 3 4. Back plumage patterning: streaked = 1; barred = 2; quail-like = 3 5. Throat base colour: white-buff = 1; rufous-chestnut = 2; black = 3; buff-rimmed with rufous = 4 6. Throat flecking/freckling: none = 0; present = 1 7. Gorget: absent = 0; present, but poorly developed = 1; moderately developed = 2; well developed = 3 8. Breast patterning: unpatterned = 0; barred = 1; streaked = 2; spotted = 3; blotched/mottled = 4 9. Belly patterning: unpatterned = 0; spotted = 1; barred = 2; streaked = 3; blotched/mottled = 4; streaked and barred = 5 10. Under tail patterning: unpatterned = 0; barred or barred and blotched = 1 11. Wing patterning: unpatterned = 0; blotched = 1 12. Wing base colour: greyish brown = 1; greyish brown with some red = 2; moderate to well-developed red = 3 13. Side of the head patterning: unpatterned = 0; streaked = 1; striped (eye and/or jaw – progresses to gorget) = 2 14. Leg colour: yellow = 1; red, reddish range = 2 15. Spur number: none/poorly developed = 0; one = 1 16. Spur length: < 1mm = 0; 1-10mm = 1; 11-20mm = 2 17. Wing length (mm): < 140 = 1; < 160 = 2; < 180 = 3 18. Bill length (mm) / Wing length: < .15 = 1; < .17 = 2; < .18 = 3; > .18 = 4 19. Tail length (mm) / Wing length: < .50 = 1; <.55 = 2; < .60 = 3; > .60 = 4 20. Sexual plumage dimorphism: 0 = none; 1 = slight; 2 = marked 21. Raucous advertisement call: absent = 0; type 1 = 1; type 2 = 2 22. Simple musical advertisement call: 0 = absent; slow = 1; fast = 2; very fast =3 23. Short complex musical advertisement call: absent = 0; slow ki-bee-tilli = 1; fast ki-bee-tilli = 2 24. Long complex musical call: absent = 0; type 1 = 1; type 2 = 2 194 Table 5.2. A matrix of organismal character and character state scores used in phylogenetic analysis of francolins. Character no. Taxon Gallus gallus Bambusicola thoracica Francolinus francolinus Francolinus pictus Francolinus pintadeanus Ortygornis gularis Ortygornis pondicerianus Ortygornis sephaena Ortygornis grantii Ortygornis r. rovuma Ortygornis r. spilogaster Afrocolinus l. lathami Afrocolinus l. schubotzi Peliperdix coqui Peliperdix kasaicus Peliperdix maharao Peliperdix ruandae Peliperdix stuhlmanni Peliperdix c. vernayi Peliperdix hubbardi Peliperdix spinetorum Peliperdix albogularis Peliperdix a. buckleyi Peliperdix dewittei Peliperdix schlegelii Scleroptila streptophora Scleroptila finschi Scleroptila levaillantii Scleroptila c. crawshayi Scleroptila c. kikuyuensis Scleroptila levaillantoides Scleroptila l. pallidior Scleroptila jugularis Scleroptila p. psilolaema Scleroptila p. theresae Scleroptila afra Scleroptila ellenbecki Scleroptila shelleyi Scleroptila uluensis Scleroptila gutturalis Scleroptila whytei 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 0 2 1 1 1 0 0 1 1 1 1 0 0 2 2 2 2 2 2 2 2 2 2 2 2 0 2 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 2 3 1 1 1 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 1 3 1 1 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 2 3 2 1 2 1 1 1 1 1 3 3 1 1 1 1 1 1 1 1 1 1 1 1 4 1 4 4 4 1 1 1 1 1 1 1 1 1 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 2 0 0 0 0 0 0 3 3 3 1 1 3 2 2 3 2 2 2 2 3 0 2 1 3 1 2 1 2 2 2 2 3 3 0 1 1 0 1 1 1 1 1 1 1 1 1 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 0 0 2 1 2 3 2 2 2 3 3 1 1 2 2 2 2 2 2 0 0 2 2 2 2 4 2 2 2 2 0 0 0 2 2 2 2 2 2 5 2 0 0 1 0 1 0 1 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 2 1 1 1 1 1 1 1 1 2 1 1 1 2 1 1 2 2 2 1 1 2 3 3 3 3 3 2 3 3 1 3 2 2 3 3 0 0 0 0 1 0 1 1 1 1 1 0 0 2 2 2 2 2 2 2 0 2 2 2 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 0 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 195 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 0 1 1 1 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 2 2 1 1 3 1 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 1 1 2 2 4 2 2 2 2 2 2 1 1 2 2 2 2 2 2 2 2 2 2 2 2 3 3 4 4 4 3 3 3 3 3 3 3 3 3 3 3 4 3 4 3 3 4 4 4 4 4 4 2 2 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 2 1 2 0 0 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 1 1 2 2 2 2 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 3 3 3 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 1 1 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 0 0 0 0 0 1 0 2 2 0 2 Table 5.3. Francolin taxa for which DNA sequences were generated. Acronyms. AMNH American Museum of Natural History, TM - Transvaal Museum, BM - British Museum, Natural History Museum at Tring, SAM - Iziko Museums of Cape Town (Natural History), PFIAO - Percy FitzPatrick Institute of African Ornithology, TMC - Timothy M. Crowe, University of Cape Town, South Africa, GB - GenBank, ‘-’ - Unknown, Br. muscle - Breast muscle, Pect. muscle - Pectoral muscle. The scientific names below are as recorded on specimen labels. Taxa name Sample no. Origin Date coll. Sample type AMNH DOT8023 AMNH 776813 GB EU165707 India India China - Br. muscle Toe-pad - TMC 9 TMC 10 TM 6410 BM 1902 1 20 300 TMC 11 Marico River, South Africa South Africa Zimbiti, Beira, Mozambique Hulul, Ethiopia Cameroon 2004 1910 1902 2005 Heart Br. muscle Toe-pad Toe-pad Br. muscle BM 80 1 1 1066 BM 1965 M 2073 AMNH 68972 PFIAO 47 TMC 13 TM 27245 PFIAO 59 TMC 12 TM 13217 SAM 55386 SAM 57819 AMNH 541174 AMNH 541185 AMNH 192573 AMNH 542359 TM 78622 AMNH 406156 TM 23166 AMNH 308887 Kenya Mount Kenya, Kenya Kenya Ayton Farm, South Africa Kenya Mzimba, Malawi Eastern Cape, South Africa Petrus Steyn, South Africa Quickborn, Namibia Botswana Namibia Ethiopia Somalia Ethiopia Angola Sterkspruit, South Africa Kenya Nyika Plateau, Malawi Angola 1965 1949 2002 2002 1923 2005 1940 1941 Toe-pad Toe-pad Liver Br. muscle Toe-pad Liver Liver Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Br. muscle Toe-pad Toe-pad Toe-pad PFIAO 45 TM 25325 BM 1933 7 14 105 TM 17263 BM 1953 54 52 TM 52676 BM 1956 5 14 85 AMNH 261916 TM 23158 BM 1963 6 2 BM 1953 54 49 BM 19339 5 2 BM 1949 30 19 BM 1929 3 13 1 BM 1999 9 20 2 BM 1934 3 16 28 Settlers, South Africa Luluabourg, Dem. Rep. Congo Richmond, Natal, South Africa Botswana Luluaborg, Dem. Rep. Congo Rwinkwaku, Rwanda Mega, Ethiopia Kidong Valley, Kenya Mzimba, Malawi Arusha, Tanzania Luluaborg, Dem. Rep. Congo Azare banchi, Nigeria Mboro, Bahr-El-Ghazel, Chad Farafeni, Gambia Gambaga, Ghana Ejura Ashanti Ghana 1939 1933 1953 1956 1934 1963 1953 1933 1949 1929 1999 1934 Pect. Muscle Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Spotted Group F. francolinus F. pictus F. pintadeanus Striated Group Francolinus sephaena F. sephaena F. s. rovuma F. s. grantii F. streptophora Red-winged Group F. psilolaema psilolaema F. p. theresae F. p. ellenbecki F. shelleyi F. s. uluensis F. s. whytei F. afra F. levaillantoides levaillantoides F. l. pallidior F. l. kalaharica F. l. jugularis F. l. gutturalis F. l. lorti F. l. archeri F. l. cunenensis F. levaillantii levaillantii F. l. kikuyuensis F. l. crawshayi F. finschi Red-tailed Group F. coqui coqui F. c. angolensis F. c. campbelli F. c. vernayi F. c. lynesi F. c. ruahdae F. c. maharao F. c. hubbardi F. c. stuhlmanni F. c. thikae F. c. kasaicus F. c. spinetorum F. schlegelii F. albogularis albogularis F. a. gambagae F. a. buckleyi 196 Taxa name Sample no. Origin Date coll. Sample type F. a. meinertzhageni BM 1957 35 13 Luacano, Angola 1957 Toe-pad GB AM236893 BM 1949 30 17 AMNH DOT8050 GB U90649 Cameroon Benengal Zandi, Sudan India India 1949 - Blood Toe-pad Br. muscle - Ungrouped taxa F. lathami lathami F. l. schubotzi F. pondicerianus F. gularis 197 Table 5.4. DNA markers sequenced and primers used for PCR-amplification and sequencing of preserved francolin tissues. Primer name Primer sequence (5’to 3’) Reference Cytochrome b L14578 MH15364 ML15347 H15915 cta gga atc atc cta gcc cta ga act cta cta ggg ttt ggc c atc aca aac cta ttc tc aac gca gtc atc tcc ggt tta caa gac J.G. Groth (pers. comm.) P. Beresford (pers. comm.) P. Beresford (pers. comm.) Edwards & Wilson (1990) Table 5.5. DNA marker sequenced and primers used for PCR-amplification and sequencing of museum toe-pads for francolins. Primer name Primer sequence (5’ to 3’) Reference Cytochrome-b Francolin-specific primers L14851 (General) cct act tag gat cat tcg ccc t Franc-H1 cag cag aca cyt cyc tyg cct tc MH15145 aag aat gag gcg cca ttt gc Kornegay et al. (1993) R. Bowie P. Beresford Franc-L1 Franc-H2 tgcctcacaacccaaatcctcac R. Bowie agg agr agr att act cct gtg ttt cag g R. Bowie Franc-L2 Franc-H3 gcc tca ttc tty ttc aty tgy atc ttc c ggr tgg aat ggg att ttg tca gag R. Bowie R. Bowie Franc-L3 Franc-H4 tcatcyractcygacaaaatccc gar rgg gat tag rag gag gat R. Bowie R. Bowie Franc-L4 Franc-H5 tat tcg cct ayg cya tcc twc gct c gta gga rag kga tgc tat ttg gcc R. Bowie R. Bowie Franc-L5 HB20 (General) ctc atc ctc ctc cta atc cc ttg gtt cac aag acc aat gtt R. Bowie J. Feinstein (pers. comm.) 198 Table 5.6. The inter- and intraspecific uncorrected ‘P’ distances calculated from CYTB. Taxo n name 1 Franco linus franco linus 2 Franco linus pintadeanus 3 Ortygo rnis gularis 4 A fro co linus latham i 5 A fro co linus l. s chubo tzi 6 Franco linus pictus 7 Ortygo rnis po ndicerianus 8 Ortygo rnis s ephaena grantii 9 Ortygo rnis s ephaena (Kenya) 10 Ortygo rnis s . ro vum a 11 Ortygo rnis s ephaena (SA) 12 Ortygo rnis s . zam bes iae 13 P eliperdix albo gularis 14 P eliperdix a. buckleyi 15 P eliperdix a. gam bagae 16 P eliperdix a. m einertzhageni 17 P eliperdix co qui 18 P eliperdix c. ruahdae (Rwanda) 19 P eliperdix co qui (Zambia) 2 0 P eliperdix c. ango lens is 2 1 P eliperdix c. cam pbelii 2 2 P eliperdix hubbardi 2 3 P eliperdix c. kas aicus 2 4 P eliperdix c. lynes i 2 5 P eliperdix c. m aharao 2 6 P eliperdix c. s pineto rum 2 7 P eliperdix c. s tuhlm anni 2 8 P eliperdix c. thikae 2 9 P eliperdix c. vernayi 3 0 P eliperdix s chlegelii 3 1 S clero ptila afra 3 2 S clero ptila fins chi 3 3 S clero ptila levaillantii 3 4 S clero ptila l. craws hayi 3 5 S clero ptila l. kikuyuens is 3 6 S clero ptila levaillanto ides 3 7 S clero ptila l. kalaharica (Bo ts wana) 3 8 S clero ptila l. jugularis 3 9 S clero ptila l. archeri 4 0 S clero ptila l. gutturalis 4 1 S clero ptila l. lo rti 4 2 S clero ptila l. pallidio r 4 3 S clero ptila ps ilo laem a 4 4 S clero ptila p. theres ae 4 5 S clero ptila p. ellenbecki 4 6 S clero ptila s helleyi uluens is 4 7 S clero ptila s . whytei 4 8 S clero ptila s helleyi (SA) 4 9 S clero ptila s trepto pho ra 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 25 26 27 28 29 30 31 32 33 34 35 36 37 38 # # 41 42 # 44 45 46 47 48 49 11% 9% 12% 9% 12% 9% 10% 12% 9% 1% 7% 10% 9% 9% 9% 9% 13% 5% 10% 10% 8% 10% 14% 9% 10% 11% 9% 9% 10% 14% 9% 10% 11% 9% 9% 1% 10% 14% 9% 11% 12% 10% 10% 4% 4% 9% 13% 8% 9% 10% 11% 9% 5% 5% 6% 10% 12% 8% 9% 9% 11% 10% 5% 6% 4% 0% 10% 14% 12% 11% 11% 14% 12% 12% 12% 12% 11% 13% 11% 14% 12% 11% 11% 14% 12% 12% 12% 12% 11% 12% 1% 11% 13% 12% 11% 11% 14% 12% 12% 12% 12% 11% 12% 2% 0% 11% 14% 11% 10% 11% 10% 11% 11% 12% 12% 11% 10% 5% 5% 6% 10% 13% 10% 10% 11% 10% 11% 11% 11% 11% 10% 8% 8% 8% 9% 8% 11% 17% 15% 13% 14% 9% 13% 15% 14% 15% 12% 9% 8% 7% 7% 10% 4% 11% 14% 11% 11% 11% 11% 11% 11% 12% 12% 11% 7% 8% 8% 9% 8% 0% 4% 11% 14% 11% 11% 11% 11% 12% 12% 12% 12% 11% 8% 8% 8% 9% 8% 0% 4% 0% 10% 14% 10% 10% 10% 10% 11% 11% 11% 11% 10% 8% 8% 7% 8% 8% 2% 4% 2% 2% 10% 13% 11% 10% 10% 9% 11% 11% 11% 12% 11% 10% 8% 8% 9% 8% 5% 5% 5% 5% 6% 10% 11% 10% 10% 10% 10% 11% 9% 10% 10% 11% 10% 9% 9% 7% 8% 7% 2% 8% 8% 6% 9% 11% 14% 11% 11% 11% 11% 11% 12% 12% 12% 11% 7% 9% 8% 9% 8% 1% 4% 0% 0% 2% 5% 8% 10% 13% 11% 10% 11% 9% 11% 12% 12% 13% 12% 11% 10% 10% 10% 9% 6% 11% 6% 6% 8% 5% 9% 6% 11% 14% 11% 11% 11% 10% 11% 11% 12% 12% 10% 11% 9% 9% 9% 10% 8% 7% 8% 8% 8% 6% 10% 8% 7% 10% 11% 10% 10% 10% 10% 10% 10% 10% 10% 10% 9% 8% 8% 7% 8% 5% 9% 6% 5% 7% 6% 5% 6% 8% 8% 10% 12% 10% 10% 11% 9% 11% 10% 11% 12% 10% 10% 9% 8% 9% 8% 6% 9% 6% 6% 6% 3% 7% 6% 5% 7% 7% 11% 17% 13% 12% 13% 9% 12% 14% 14% 15% 11% 9% 7% 6% 7% 9% 3% 1% 3% 3% 3% 4% 0% 3% 10% 6% 7% 8% 10% 13% 10% 10% 10% 8% 11% 10% 11% 12% 10% 9% 8% 8% 7% 8% 9% 8% 9% 9% 8% 8% 7% 9% 9% 8% 8% 9% 8% 9% 12% 10% 11% 11% 9% 10% 10% 11% 10% 10% 9% 9% 10% 8% 10% 9% 13% 10% 10% 9% 10% 5% 10% 10% 10% 5% 10% 12% 9% 10% 13% 10% 11% 11% 9% 10% 11% 11% 11% 11% 10% 11% 11% 10% 10% 10% 13% 10% 11% 10% 10% 8% 10% 11% 10% 8% 10% 12% 11% 6% 9% 13% 10% 10% 10% 9% 11% 10% 10% 10% 10% 10% 10% 10% 10% 10% 9% 13% 9% 9% 9% 9% 9% 9% 10% 11% 8% 10% 12% 11% 7% 8% 9% 12% 10% 11% 11% 7% 11% 11% 12% 11% 11% 11% 10% 11% 9% 10% 10% 12% 11% 11% 11% 10% 5% 11% 11% 12% 6% 10% 11% 9% 2% 6% 8% 10% 14% 11% 11% 11% 8% 11% 11% 12% 11% 11% 10% 11% 11% 11% 11% 9% 12% 10% 10% 10% 9% 8% 9% 12% 11% 8% 9% 11% 10% 5% 7% 9% 2% 10% 12% 11% 11% 10% 8% 11% 11% 11% 11% 11% 8% 10% 10% 9% 10% 9% 11% 10% 10% 9% 9% 7% 10% 10% 10% 6% 10% 10% 9% 3% 5% 8% 5% 5% 11% 15% 14% 15% 17% 9% 12% 15% 15% 16% 14% 9% 13% 12% 10% 12% 12% 9% 12% 12% 12% 11% 5% 12% 14% 11% 12% 12% 9% 14% 7% 10% 13% 7% 8% 4% 10% 13% 11% 11% 11% 9% 10% 9% 8% 11% 6% 9% 10% 11% 12% 12% 12% 11% 12% 12% 11% 11% 12% 12% 12% 12% 11% 10% 11% 12% 8% 10% 10% 8% 4% 7% 5% 9% 11% 9% 10% 9% 4% 9% 10% 11% 11% 10% 9% 10% 9% 8% 9% 9% 11% 9% 10% 9% 9% 5% 10% 10% 9% 6% 9% 10% 8% 4% 6% 7% 5% 5% 3% 6% 7% 10% 11% 9% 10% 10% 8% 9% 11% 11% 11% 11% 8% 11% 10% 9% 10% 10% 11% 10% 11% 10% 10% 6% 10% 10% 10% 6% 10% 11% 9% 4% 6% 7% 5% 6% 3% 6% 8% 1% 10% 11% 9% 10% 10% 8% 10% 11% 11% 11% 11% 9% 10% 10% 9% 9% 10% 12% 10% 10% 9% 10% 6% 10% 10% 10% 6% 9% 10% 9% 4% 6% 7% 4% 6% 3% 7% 8% 1% 2% 9% 12% 10% 11% 11% 8% 11% 11% 11% 11% 10% 9% 9% 9% 9% 10% 9% 12% 9% 9% 9% 10% 5% 9% 11% 10% 6% 10% 11% 10% 2% 7% 7% 3% 4% 4% 7% 8% 4% 4% 4% 10% 12% 10% 11% 10% 10% 11% 11% 11% 11% 11% 10% 11% 11% 11% 11% 11% 14% 12% 12% 12% 11% 9% 12% 11% 10% 9% 11% 13% 11% 6% 6% 8% 6% 7% 6% 9% 9% 5% 5% 6% 5% 11% 12% 10% 10% 10% 10% 10% 11% 11% 11% 11% 10% 11% 11% 11% 11% 11% 13% 11% 11% 11% 10% 9% 11% 10% 10% 8% 11% 12% 11% 6% 6% 8% 7% 8% 6% 9% 9% 6% 6% 6% 7% 3% 10% 13% 10% 12% 12% 9% 11% 10% 11% 10% 10% 9% 10% 10% 10% 11% 10% 12% 10% 10% 10% 10% 8% 10% 10% 11% 7% 9% 12% 11% 2% 5% 7% 3% 5% 3% 5% 7% 4% 4% 3% 3% 5% 6% 10% 12% 9% 10% 11% 9% 10% 10% 11% 10% 10% 9% 10% 11% 9% 10% 10% 14% 10% 10% 10% 10% 6% 10% 11% 11% 6% 10% 13% 9% 4% 6% 8% 6% 7% 4% 9% 8% 3% 3% 3% 5% 6% 6% 4% 9% 12% 9% 10% 9% 5% 10% 10% 10% 11% 10% 10% 10% 10% 10% 10% 11% 14% 11% 11% 10% 10% 8% 11% 10% 10% 8% 10% 13% 9% 6% 5% 8% 7% 6% 6% 9% 6% 5% 6% 6% 7% 5% 6% 6% 6% 10% 12% 10% 10% 10% 9% 10% 10% 11% 11% 11% 10% 10% 10% 9% 9% 9% 11% 10% 10% 9% 9% 7% 10% 10% 10% 7% 10% 10% 9% 3% 5% 7% 5% 6% 2% 7% 7% 4% 4% 3% 4% 6% 6% 4% 4% 6% 10% 13% 9% 10% 10% 9% 10% 11% 11% 11% 10% 9% 10% 11% 11% 10% 9% 12% 10% 10% 10% 11% 9% 10% 11% 10% 9% 11% 11% 10% 8% 8% 8% 9% 8% 8% 12% 10% 8% 8% 8% 8% 9% 8% 8% 8% 9% 8% 199 3 4 3 Striated Group 1. Ortygornis sephaena 2. O. rovuma rovuma 3. O. grantii 4. O. rovuma spilogaster 2 1 1 2 1 Figure 5.1. Distribution ranges of Hall’s Striated Group taxa. 200 1 1 1 2 2 3 Red-winged Group 1. 2. 3. 4. 5. 6. Scleroptila gutturalis S. streptophora S. finschi S. jugularis S. levaillantoides pallidior S. levaillantoides levaillantoides 4 3 5 6 Figure 5.2. Distribution ranges of Hall’s Red-winged Group taxa. 201 3 2 Red-winged Group (cont.) 1. Scleroptila levaillantii 3 2. S. crawshayi crawshayi 3. S. c. kikuyuensis 1 Figure 5.3. Distribution ranges of Hall’s Red-winged Group taxa. 202 6 7 5 4 Red-winged Group (concl.) 1. 2. 3. 4. 5. 6. 7. Scleroptila afra S. shelleyi S. whytei S. uluensis S. psilolaema theresae S. p. psilolaema S. ellenbecki 2 3 2 2 1 Figure 5.4. Distribution ranges of Hall’s Red-winged Group taxa. 203 8 8 8 9 11 10 7 5 6 4 4 Red-tailed Group 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Peliperdix coqui coqui P. c. vernayi P. stuhlmanni P. kasaicus P. c. ruahdae P. hubbardi P. maharao P. spinetorum P. albogularis albogularis P. a. buckleyi P. schlegelii P. dewittei 12 1 3 1 2 1 Figure 5.5. Distribution ranges of Hall’s Red-tailed Group taxa. 204 1 1 2 2 1 2 1. Afrocolinus lathami lathami 2. A. l. schubotzi Figure 5.6. Distribution of Afrocolinus lathami. 205 Figure 5.7. Illustration of francolin body partitions scored to generate the plumage character matrix. C - Crown, HN - Hindneck, M - Mantle, B - Back, TF - Tail feathers, T - Throat, G - Gorget, BR - Breast, BE - Belly, UNTC - Undertail coverts, SOH - Side of the head, W - Wing (primaries). 206 Francolinus pintadeanus 86 Francolinus pictus Francolinus francolinus S P G Bambusicola thoracica Afrocolinus lathami schubotzi 100 Afrocolinus lathami lathami Ortygornis sephaena 82 Ortygornis rovuma rovuma 99 African Ortygornis grantii Ortygornis pondicerianus Ortygornis gularis 100 Asian Peliperdix hubbardi Peliperdix maharao 69 Peliperdix spinetorum 88 93 Peliperdix coqui coqui Peliperdix coqui vernayi Peliperdix stuhlmanni Peliperdix schlegelii 78 94 Peliperdix albogularis buckleyi 99 R T G Peliperdix albogularis albogularis Peliperdix dewittei 99 Peliperdix kasaicus Scleroptila levaillantii 75 Scleroptila crawshayi kikuyuensis Scleroptila crawshayi crawshayi 58 Scleroptila levaillantoides levaillantoides Scleroptila shelleyi Scleroptila finschi 52 Scleroptila jugularis 98 Scleroptila whytei Scleroptila psilolaema theresae Scleroptila psilolaema psilolaema R W G Scleroptila gutturalis Scleroptila afra Scleroptila levaillantoides pallidior Scleroptila uluensis Scleroptila ellenbecki Scleroptila streptophora Gallus gallus Figure 5.8. Maximum likelihood tree obtained from mitochondrial Cytochrome-b characters. Numbers above branches represent boostrap support values (only ≥ 70% are shown). SPG stands for Spotted Group, RTG – Red-tailed Group, RWG – Red-winged Group. 207 Scleroptila streptophora Scleroptila whytei Scleroptila jugularis Scleroptila finschi Scleroptila psilolaema theresae 100 Scleroptila psilolaema psilolaema Scleroptila uluensis Scleroptila ellenbecki Scleroptila crawshayi kikuyuensis 98 Scleroptila crawshayi crawshayi Scleroptila levaillantoides pallidior Scleroptila afra Scleroptila shelleyi 72 Scleroptila levaillantoides levaillantoides Scleroptila gutturalis Scleroptila levaillantii Peliperdix spinetorum Peliperdix hubbardi Peliperdix maharao 70 Peliperdix coqui vernayi Peliperdix coqui coqui R T G Peliperdix stuhlmanni Peliperdix kasaicus Peliperdix schlegelii Peliperdix dewittei Peliperdix albogularis buckleyi 100 Peliperdix albogularis albogularis Francolinus pintadeanus S Francolinus pictus P G Francolinus francolinus 97 Ortygornis sephaena 99 99 100 Ortygornis rovuma rovuma Ortygornis grantii Ortygornis pondicerianus Ortygornis gularis African Asian Afrocolinus lathami lathami Afrocolinus lathami schubotzi Bambusicola thoracica Gallus gallus Figure 5.9. A parsimony tree (1 of 4 most parsimonious trees) obtained from mitochondrial Cytochrome-b characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). SPG stands for Spotted Group, RTG – Red-tailed Group, RWG – Redwinged Group. 208 R W G Gallus gallus Bambusicola thoracica Francolinus francolinus Francolinus pintadeanus Francolinus pictus S P G Ortygornis sephaena 96 Ortygornis grantii Ortygornis rovuma rovuma Ortygornis rovuma spilogaster Ortygornis pondicerianus African Asian 95 Afrocolinus lathami lathami Afrocolinus lathami schubotzi Peliperdix coqui coqui Peliperdix ruahdae Peliperdix kasaicus Peliperdix coqui vernayi Peliperdix maharao Peliperdix stuhlmanni Peliperdix hubbardi Peliperdix spinetorum R T G Peliperdix albogularis albogularis Peliperdix albogularis buckleyi Peliperdix dewittei Peliperdix schlegelii Scleroptila streptophora Scleroptila finschi Scleroptila levaillantii Scleroptila crawsha. crawshayi Scleroptila craw. kikuyuensis Scleroptila afra 94 74 70 Scleroptila levaillantoi. levaillantoides Scleroptila levaillantoides pallidior Scleroptila jugularis Scleroptila psilolaema psilolaema Scleroptila ellenbecki Scleroptila psilolaema theresae Scleroptila shelleyi Scleroptila uluensis Scleroptila gutturalis Scleroptila whytei Ortygornis gularis Figure 5.10. A parsimony tree (1 of 397 most parsimonious trees) obtained from organismal characters. Numbers above branches represent bootstrap support values (only ≥ 70% are presented). SPG stands for Spotted Group, RTG – Red-tailed Group, RWG – Red-winged Group. 209 R W G Gallus gallus Bambusicola thoracica Francolinus francolinus Francolinus pictus Francolinus pintadeanus Peliperdix coqui coqui Peliperdix coqui ruahdae S P G Peliperdix coqui vernayi Peliperdix maharao Peliperdix hubbardi Peliperdix spinetorum Peliperdix kasaicus Peliperdix stuhlmanni Peliperdix albogularis albogularis 100 Peliperdix albogularis buckleyi Peliperdix dewittei Peliperdix schlegelii 93 71 71 83 77 87 100 87 Scleroptila streptophora Scleroptila levaillantii Scleroptila finschi Scleroptila jugularis Scleroptila whytei Scleroptila psilolaema psilolaema Scleroptila psilolaema theresae Scleroptila crawshayi crawshayi 100 Scleroptila crawshayi kikuyuensis Scleroptila afra Scleroptila uluensis Scleroptila levaillantoides pallidior Scleroptila ellenbecki 85 R T G R W G Scleroptila levaillantoides levaillantoides Scleroptila shelleyi Scleroptila gutturalis Ortygornis gularis Ortygornis pondicerianus Ortygornis sephaena Ortygornis grantii 100 73 97 83 Ortygornis rovuma rovuma Ortygornis rovuma spilogaster Afrocolinus lathami lathami Asian African 71 100 Afrocolinus lathami schubotzi Figure 5.11. A parsimony tree (1 of 24 most parsimonious trees) obtained from combined mitochondrial Cytochrome-b and organismal characters. Numbers above branches represent bootstrap support values (only ≥ 70% are presented). SPG stands for Spotted Group, RTG – Red-tailed Group, RWG – Red-winged Group. 210 Appendix 5.1. List of francolin taxa included in the analysis and the GenBank accession number for CYTB. CYTB GenBank no. Taxon Francolinus francolinus AF013762 Francolinus pictus Francolinus pintadeanus FR694142 NC011817 Ortygornis pondicerianus Ortygornis gularis FR691632 U90649 Afrocolinus lathami lathami Afrocolinus lathami schubotzi AM236893 FR694139 Ortygornis grantii Ortygornis sephaena Kenya FR694144 FR694141 Ortygornis rovuma rovuma Ortygornis sephaena South Africa FR694135 FR694140 Ortygornis sephaena zambesiae Peliperdix albogularis albogularis FR694143 FR694145 Peliperdix a. buckleyi Peliperdix a. gambagae FR694147 FR694146 Peliperdix a. meinertzhageni FR694148 Peliperdix coqui coqui Peliperdix coqui angolensis AM236895 FR694153 Peliperdix coqui campbelii Peliperdix coqui hubbardi FR694156 FR694151 Peliperdix coqui kasaicus Peliperdix coqui lynesi FR694150 FR694155 Peliperdix coqui maharao Peliperdix coqui spinetorum FR691635 FR694154 Peliperdix coqui stuhlmanni Peliperdix coqui thikae FR694152 FR691634 Peliperdix coqui vernayi Peliperdix schlegelii FR694157 FR694149 Scleroptila afra Scleroptila finschi AM236897 FR691607 Scleroptila levaillantii Scleroptila crawshayi crawshayi U90642 FR691605 Scleroptila c. kikuyuensis Scleroptila levaillantoides FR691606 FR691612 Scleroptila levaillantoides archeri Scleroptila levaillantoides gutturalis FR691610 FR691613 Scleroptila jugularis Scleroptila levaillantoides lorti FR691608 FR691611 Scleroptila levaillantoides pallidior Scleroptila psilolaema FR691609 FR691614 Scleroptila psilolaema ellenbecki Scleroptila psilolaema theresae FR691616 FR691615 211 Scleroptila uluensis Kenya FR691620 Scleroptila shelleyi South Africa Scleroptila uluensis AM236898 FR691622 Scleroptila whytei Scleroptila streptophora FR691621 FR691617 212 Appendix 5.2. A revised classification of African francolins that is based on evidence presented in this chapter. An asterisk (*) indicates that the taxonomic status of that particular taxon may change when DNA characters become available and are included in the analyses. Species and subspecies authority appear in Table 1.2. Family: Phasianidae Sub-family: Phasianinae Genus: Francolinus Stephens, 1819 Francolinus francolinus Francolinus pintadeanus Francolinus pictus Genus: Ortygornis Reichenbach, 1853 Ortygornis pondicerianus Ortygornis gularis Ortygornis sephaena Ortygornis rovuma rovuma Ortygornis rovuma spilogaster* Ortygornis grantii Genus: Afrocolinus gen. nov. Afrocolinus lathami Afrocolinus lathami schubotzi Genus: Peliperdix Bonaparte, 1856 Peliperdix coqui coqui Peliperdix coqui vernayi Peliperdix stuhlmanni Peliperdix kasaicus Peliperdix coqui ruahdae Peliperdix hubbardi Peliperdix maharao Peliperdix spinetorum Peliperdix albogularis albogularis Peliperdix albogularis buckleyi Peliperdix schlegelii Peliperdix dewittei Genus: Scleroptila Blyth, 1849 Scleroptila levaillantoides levaillantoides Scleroptila levaillantoides pallidior Scleroptila jugularis Scleroptila gutturalis Scleroptila finschi Scleroptila streptophora Scleroptila levaillantii Scleroptila crawshayi crawshayi Scleroptila crawshayi kikuyuensis 213 Scleroptila afra Scleroptila shelleyi Scleroptila whytei Scleroptila uluensis Scleroptila psilolaema psilolaema Scleroptila psilolaema theresae Scleroptila ellenbecki 214 CHAPTER 6 Taxonomy and phylogeny of spurfowls Abstract Delimitation of taxa continues to be a challenging exercise in the absence of a universal definition of species and subspecies that can be applied consistently across different groups of study. The spurfowls, members of which share a common evolutionary path (genus Pternistis Wagler, 1832, sensu Crowe et al. 2006), which are the group of focus in this chapter, are no exception. Over and above the challenges resulting from suspected high levels of hybridization, discordance between organismal and molecular characters, difficulty in determining the genetic threshold when delimiting taxa, and unavailability of fresh samples of many taxa for DNA analyses, this study sought to establish a classification system of spurfowls which takes into account the evolutionary relationships among taxa, with the goal of bringing stability to the taxa recognized as valid species, subspecies and genera, based on congruent multiple lines of evidence. The phylogenetic inference methods, Maximum likelihood and parsimony were performed on separate and combined DNA and organismal characters. The outcome resulted in some subspecies being elevated to the species level with others being placed into more inclusive entities; no generic recommendations are made. Most of the phylogenetic hypotheses presented by Hall (1963) were rejected. Spurfowls are represented by extremely morphologically and vocally divergent taxa. Thus, it is difficult to identify homologuous characters that can be used to investigate their evolutionary relationships. However, a multi-faceted character approach seems to be a 215 fitting strategy for working out intra-generic relationships within spurfowls. Phylogeographic studies of various species complexes are recommended. 216 Introduction What are spurfowls? The hypothesised monophyly of the genus Francolinus Stephens 1819 (sensu Hall 1963), was rejected resulting in the split of birds traditionally known as ‘francolins’ into two lineages, ‘true’ francolins and spurfowls, that are evolutionary distant (Crowe et al. 2006). Spurfowls are represented by members currently assigned to a single genus Pternistis Wagler 1832 (Table 1.1, 1.2). Spurfowls, like francolins are placed in the sub-family Phasianinae, and together with other Old World partridge- and quail-like gamebirds (e.g. Perdix and Coturnix spp.) are placed in the tribe Perdicini (Chapin 1932, Peters 1934, Wolters 1975-82, Crowe et al. 1986, Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994, Madge and McGowan 2002). This assemblage is comprised of 23 traditionally recognized species with their core distribution mainly in sub-Saharan Africa and no species occur in Asia (Hall 1963, Johnsgard 1988). The West African Pternistis bicalcaratus is the only species that extends its distribution into Morocco, in North Africa. Spurfowl species and their distribution This chapter focuses on spurfowls (Pternistis spp.), which Hall (1963) categorized into four putative monophyletic species groups (see Table 1.1), all represented by African species. These are the Bare-throated Group, Montane, Vermiculated, and the Scaly Group. Generally, spurfowls like francolins are (partridge-like) resident birds that can run or fly over short distances when facing a threat. Most species are sexually monochromatic in plumage, but males of most species have a single spur on their tarsus while females in some instances also have relatively smaller spurs (Johnsgard 1988). Members of the 217 different species groups have complex morphology, ecology, behaviour and distribution patterns (Snow 1978) and occur in different habitats primarily of a tropical or sub-tropical nature (Hall 1963, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). Species group diversity, distribution and morphology Bare-throated Group The Bare-throated Group consists of four traditionally recognized species, the Rednecked Spurfowl Pternistis afer (Müller, 1776), Swainson’s Spurfowl P. swainsonii (Smith, 1836), Yellow-necked Spurfowl P. leucoscepus (Gray, 1867) and Grey-breasted Spurfowl P. rufopictus Reichenow, 1887 (Table 1.1). A number of subspecies are traditionally recognized, 20 in P. afer (afer (Müller, 1776), harterti Reichenow, 1909, nyanzae Conover, 1929, böhmi Reichenow, 1885, intercedens Reichenow, 1909, itigi (Bowen, 1930), cranchii (Leach & Koenig, 1818), punctulatus (Gray, 1834), benguellensis Bocage, 1893, leucoparaeus (Fischer & Reichenow, 1884), humboldtii (Peters, 1854), swynnertoni Sclater, 1921, castaneiventer Gunning & Roberts, 1911, melanogaster Neumann, 1898, loangwae Grant & Praed, 1934, lehmanni Roberts, 1931, notatus Roberts 1924, krebsi Neumann, 1920, cunenensis Roberts, 1932, cooperi Roberts, 1947); five in P. swainsonii (swainsonii (Smith, 1836)), lundazi White, 1947, gilli Roberts, 1932, damarensis Roberts, 1932, chobiensis Roberts, 1932); seven in P. leucoscepus (leucoscepus (Gray, 1867)), infuscatus Cabanis, 1868, holtemülleri Erlanger, 1904, keniensis Mearns, 1911, kilimensis Mearns, 1911, tokora Stoneham, 1930, muhamed-ben-abdullah Erlanger, 1904) and a monotypic P. rufopictus (Table 1.2). 218 These taxa have widespread distribution in eastern Africa from Eritrea to the Cape Province and extend westward south of the Congo forest to Gabon, Angola and northern Namibia (Hall 1963) (Fig. 6.1, 6.2). They inhabit grasslands with cover of trees and scrubs and they are found near water courses. Members of the Bare-throated Group are large with a body mass ranging from 480 – 960 g (Johnsgard 1988). These birds are distinguished from the other spurfowls by having a patch of bare skin on the throat and round the eye, with the underparts being diagnostic in all four species. The presence of fairly well-developed spurs is a remarkable feature (probably linked to the origin of the group name ‘spurfowls’). The males have a long and strong lower spur with a short blunt upper spur being common in P. leucoscepus and P. rufopictus, and less common among individuals of P. afer and P. swainsonii. The “afer” type subspecies have no vermiculation and they are strongly patterned in black, white and grey on the belly and face. The belly of P. afer is broadly streaked in white and black, with feathers having black centres and white edges. In short, what characterizes what Hall (1963) refers to as the cranchii-type subspecies is that the belly feathers have buff central streak vermiculated with blackish grey and margined with broad chestnut (degree of chestnut colour varies from one subspecies to the other) whereas the afer-type subspecies have broad greyish black central streak with buff margins (particularly in the nominate subspecies afer) or with thin greyish black central streak separating the long buff parallel streaks margined with black or sometimes maroon. The feathers on the belly of P. swainsonii have drop-shaped dark brown shaft streaks which increase in size over the lower belly and flanks. Pternistis leucoscepus is the most distinctive member of the group with a yellow bare throat skin and it presents slight variation compared to P. afer. The back is diagnosed by white 219 shaft streaks and the underparts are streaked with white and chestnut in a way that each feather has chestnut with narrow white edges and a triangular white patch at the tip, tapering up the shaft. The primaries are said to have a conspicuous patch which according to Hall (1963) is visible during flight. Unlike the other Bare-throated species, P. rufopictus has orange pink throat skin, grey with black shaft streaks breast with the belly belly having the narrow central black streak separated from rufous chestnut margins by broad buff to white streaks. Montane Group This group consists of seven traditionally species: Erckel’s Spurfowl P. erckelii (Rüppell, 1835), Djibouti Spurfowl P. ochropectus (Dorst & Jouanin, 1952), Chestnutnaped Spurfowl P. castaneicollis Salvadori, 1888, Jackson’s Spurfowl P. jacksoni O. Grant, 1891, Handsome Spurfowl P. nobilis Reichenow, 1908, Mount Cameroon Spurfowl P. camerunensis Alexander, 1909, Swierstra’s Spurfowl P. swierstrai (Roberts, 1929) (Table 1.1). The putative subspecies are: two in P. erckelii (erckelii (Rüppell, 1835)), pentoni Praed, 1920, two in P. nobilis (nobilis Reichenow, 1908, chapini Grant & Praed, 1934); six within P. castaneicollis (castaneicollis Salvadori, 1888, bottegi Salvadori, 1898, gofanus Neumann, 1904, ogoensis Praed, 1920, kaffanus Grant & Praed, 1934, atrifrons (Conover, 1930)); three in P. jacksoni (jacksoni O. Grant, 1891, pollenorum Meinertzhagen, 1937, gurae Bowen, 1931) (Table 1.2). In terms of its distribution, the Montane Group is formed of scattered communities in the mountains of north eastern Africa from Eritrea to Mount Kenya, on the eastern Congo border, in the highlands of Angola, and on mount Cameroon (Hall 1963) (Fig. 6.3). Its members are birds which require some trees for cover and roosting 220 and the majority of species are in or near montane evergreen forest. Pternistis erckelii is the only member which sometimes wanders into grasslands of the high plateau of northern Ethiopia. The Montane species are said to be the least homogeneous (Snow 1978) of the spurfowl groups morphologically, so much so that, it is impossible to designate any group character other than that the males have the crown, lower back, primaries and tail plain brown or red-brown. In case where the sexes are dichromatic, the lower back and tail are vermiculated. Hall (1963) acknowledged that variation in some characters follows geographical trends (Meiri and Dayan 2003, James 1970), with the birds of the extreme north eastern parts being the largest and most heavily spurred with dark bills, yellowish legs, no bare skin round the eyes, and with the sexes alike. Birds of the two isolated western subspecies and the one central subspecies are the smallest with a mass that according to Johnsgard (1988) ranges from ~500-600 g in P. camerunensis and P. swierstrai and 600-890 g in P. nobilis. They are the least heavily spurred with red bills and legs, and with the sexes quite unlike (Snow 1978) except for P. nobilis that exhibits no sexual dimorphism. The Cameroon species, P. camerunensis has an extensive area of red bare skin round the eye. Generally, the diagnostic characters of the Montane spurfowls are found mainly on the back and to a larger extent on the underparts particularly on the belly in the midventral feathers. The patterning of the belly feathers of P. castaneicollis, which has an extensive double-patterning on the back with wing coverts and breast being clearly defined in black and white, with some ochre and chestnut) is generally similar to that in P. erckelii and P. swierstrai with the feathers being made up of a broad buff central streak which is constricted in the middle and expanded distally into a tear drop distally, 221 margined with rufous. The central buff streak in P. ochropectus is margined with greyish black U-shaped streak. The feathers on the belly of P. camerunensis, P. nobilis and P. jacksoni are uniformly coloured with buff margins in P. jacksoni and P. camerunensis, and grey margins in P. nobilis. Scaly Group The Scaly Group comprises three traditionally recognized allopatric species, Ahanta Spurfowl P. ahantensis Temminck, 1854, Scaly Spurfowl P. squamatus Cassin, 1857 and Grey-striped Spurfowl P. griseostriatus O. Grant, 1890 (Table 1.1), with nine putative subspecies recognized within P. squamatus (squamatus Cassin, 1857, maranensis Mearns, 1910, schuetti Cabanis, 1880, usambarae Conover, 1928, uzungwensis Bangs & Loveridge, 1931, doni Benson, 1939, zappeyi Mearns, 1911, tetraoninus Blundell & Lovat, 1899, chyuluensis Someren, 1939); two subspecies in P. ahantensis (ahantensis Temminck, 1854, hopkinsoni Bannerman, 1934) being recognized while Pternistis griseostriatus is a monotypic species (Table 1.2). Members of this group are characterized by having scaly underparts and they inhabit the lowland forest in Upper and Lower Guinea, and cultivations and clearings (Hall 1963) (Fig. 6.4). Compared to other spurfowls, these species have very little defining plumage pattern and strong colouration except for P. ahantensis which shows some distinct patterning with feathers having paler edges and darker centres. The underparts have some parts that are creamy-buff or brown with the colour and pattern varying in different subspecies, but all have narrow darker edges to breast feathers giving a scaly appearance. There is no marked plumage dimorphism except that the females tend to be more vermiculated than the males. 222 Pternistis squamatus is not sexually dimorphic but females tend to be less strikingly vermiculated. This species is characterized by having indistinctly vermiculated upperparts with faint U-patterning on the lower neck and the feathers have blackish centres tinged with red-brown. The belly is plain brown with scaly pattern and ill-defined dark shaft streaks margined buff. In P. ahantensis, the rest of belly feathers are richly coloured and streaked with dark brown chestnut edged buff while the upperparts are vermiculated with some white U-patterning (indistinct on the back but very distinct on the lower neck). On the other hand, P. griseostriatus is the most distinct member of the group in that the lower neck feathers and wing coverts are chestnut and broadly vermiculated. It has plain belly, the breast and flank feathers are chestnut and edged with greyish or creamy buff. Vermiculated Group This is the most widespread of all African spurfowl species groups (Fig. 6.5, 6.6, 6.7) consisting of nine traditionally recognized species (Hall 1963). The species are the Double-spurred Spurfowl P. bicalcaratus (Linnaeus, 1766), Heuglin’s Spurfowl P. icterorhynchus Heuglin, 1863, Clapperton’s spurfowl P. clappertoni (Children & Vigors, 1826), Hildebrandt’s Spurfowl P. hildebrandti Cabanis, 1878, Natal Spurfowl P. natalensis Smith, 1833, Hartlaub’s Spurfowl P. hartlaubi Bocage, 1869, Harwood’s Spurfowl P. harwoodi Blundell & Lovat, 1899, Red-billed Spurfowl P. adspersus Waterhouse, 1838 and Cape Spurfowl P. capensis (Gmelin, 1789) (Table 1.1). The putative subspecies are four within P. hartlaubi (hartlaubi Bocage, 1869, crypticus Stresemann, 1939, bradfieldi (Roberts, 1928), ovambensis (Roberts, 1928)); three in P. adspersus (adspersus Waterhouse, 1838, kalahari de Schauensee, 1931, mesicus 223 Clancey, 1996); two in P. natalensis (natalensis Smith 1833, neavei Praed, 1920); six subspecies in P. hildebrandti (hildebrandti Cabanis, 1878, altumi Fischer & Reichenow, 1884, helleri Mearns, 1915, fischeri Reichenow, 1887, johnstoni Shelley, 1894, grotei Reichenow, 1919); five in P. bicalcaratus (bicalcaratus (Linnaeus, 1766), ogilviegrantii Bannerman, 1922, ayesha Hartert, 1917, adamauae Neumann, 1915, thornei Ogilvie-Grant, 1902); four in P. icterorhynchus (icterorhynchus Heuglin, 1863, dybowski Oustalet, 1892, ugandensis Neumann, 1907, emini Neumann, 1907); eight subspecies in P. clappertoni (clappertoni (Children & Vigors, 1826), sharpii OgilvieGrant, 1892, heuglini Neumann, 1907, gedgii Ogilvie-Grant, 1891, nigrosquamatus Neumann, 1902, konigseggi Madarasz, 1914, testis Neumann, 1928, cavei Macdonald, 1940) (Table 1.2). Pternistis capensis and P. harwoodi are monotypic species. Members of the group are known to exhibit variation in their ecology, with the northern vermiculated species (P. bicalcaratus, P. icterorhynchus, P. clappertoni and P. harwoodi) forming a homogenous assemblage occupying grasslands and savanna habitats (Hall 1963). Within the Vermiculated member species the underparts are diagnostic. Pternistis bicalcaratus has buff underparts distinctly and heavily streaked with blackish and chestnut with small arrow-shaped buff marks on most belly feathers. In P. clappertoni, the upperparts are greyish brown above with barring in flight feathers. It is buff below with black to brownish marks while the hind and lower neck have Upatterning as in P. icterorhynchus whose underparts are buff with dark brown markings. The male P. harwoodi that was examined is grey speckled and finely barred with blackish and buff above. The hind and lower neck, sides of face, and throat are speckled with black and white. It has irregular double-V patterning on the belly which tends to be scattered just on the lower extreme of the belly on the buff background. Pternistis 224 hartlaubi is one of the spurfowls that exhibit striking plumage dimorphism with the difference in the known populations being mainly in the degree of colouration. The male P. hartlaubi has buff throat with black streaks which continues to the sides of neck through to the hind neck and to the breast. The belly of male is like the underparts of female which are pale tawny. The underparts of the female are tawny chestnut while males are darker and have broader black streaks on the belly. Pternistis hildebrandti shows marked plumage dimorphism with the males of the nominate form hildebrandti being greyish brown above with vermiculation, the hind and lower neck streaked black with white margins, the underparts are marked and mottled with black and white almost like in P. natalensis. The females have similar backs to the males but they have different underparts which are rusty. In P. natalensis, the hind neck mottled black and white, the back is highly vermiculated and it is greyish brown with variable black, whitish and buffish markings. The rest of the belly is buff with the upper belly to midbelly being heavily patterned in black and buff. The patterning is concentrated on the upper part of the belly with the extreme lower belly having no or few marks. Pternistis adspersus is a different looking species with the nominate form adspersus consisting of minute vermiculation on the upperparts and with distinct black and white barring on the underparts and variably on the lower neck. It is characterized by having distinctive uniform brown and white double V- or U-patterning on the back and on the belly while the pattern on the throat is reduced to form some irregular black fleckings. The belly patterning also has distinct white shaft streaks. What do we know about their taxonomy and phylogeny? 225 Profound taxonomic disagreements and confusion as reflected in Table 1.2 also affect the spurfowls and this disagreement is largely at the subspecies level and to a limited extent at the species level. Hall’ (1963) monograph has made a significant contribution (Johnsgard 1988, Sibley and Monroe 1990, del Hoyo et al. 1994) towards the understanding and enhancement of our knowledge about the evolutionary relationships and distribution ranges of these birds. The monophyletic status of genus Francolinus as hypothesized in Hall (1963), comprises 37 traditionally recognized species which were divided among the eight putatively monophyletic groups comprising ecologically similar but largely allo- or parapatric species. Even though Hall (1963) did not partition genus Francolinus, she suggested that if generic division were necessary, she would have two major groups, one consisting of members that she referred to as belonging to among others, the spurfowls, encompassing members of the Bare-throated, Montane, Scaly, Vermiculated Group. Another group would be represented by the francolins (see Chapter 5). Contrary to Roberts (1924), Peters (1934) and Mackworth-Praed and Grant (1952, 1962, 1970) and Hall (1963), Wolters (1975-82) is the one who assigned the genus Pternistis (Table 1.2) to all taxa which are known as ‘spurfowls’ today. Peters (1934) recognized the genera Francolinus and Pternistis, and like Roberts he recognized Pternistis for the Bare-throated species only (Table 1.2). Mackworth-Praed and Grant (1952, 1962, 1970) also recognized two genera, Francolinus and Pternistis, with Pternistis being assigned to members of the Bare-throated Group only (Table 1.2). Phylogenetic difficulties have been articulated by Crowe and Crowe (1985), Crowe et al. (1986), Crowe et al. (1992), Bloomer and Crowe (1998), Crowe et al. (2006) and most recently by Forcina et al. (2012), with regard to the monophyletic status of the 226 genus Francolinus, and in particular, the status of the different putative monophyletic groups recognized by Hall (1963). It is important to note that different studies used different types of characters (Table 1.4), focussed on different geographic areas of francolin distribution (Table 1.3), and made use of different methods of phylogenetic analysis. Furthermore, only the Montane Group was once recovered as monophyletic (Crowe and Crowe 1985, Crowe et al. 1992) and the Bare-throated, Vermiculated and the Scaly Groups were either found to be paraphyletic or polyphyletic (Crowe and Crowe 1985, Crowe et al. 1992, Bloomer and Crowe 1998). Based on these results, there is a dear need to test the monophyly of the species groups of spurfowls as delineated by Hall (1963) hypotheses. What is taxonomy and how can we possibly study it? Species and subspecies concept review The species and subspecies concept review is as outlined as it pertains the spurfowls in Chapter 5. Taxonomic determinations made in this chapter Most of the spurfowl species are polytypic, encompassing more than one putative subspecies (see Table 1.2). The same criteria as outline in Chapter 5 are used here. Objectives of the study:  To review the taxonomic status of spurfowls.  To re-assess the monophyletic status of the species groups as delineated by Hall (1963). 227  To investigate the phylogenetic relationship between taxa within Hall’s (1963) putative monophyletic groups.  To produce a revised classification of spurfowls that takes into account the evolutionary relationships among delineated taxa. Materials and methods Data collection Morpho-behavioural characters of spurfowls Spurfowls were divided into discrete sections and scored for variation in plumage patterning and colouration (Fig. 5.7). In total, 24 organismal characters reflecting assessment of plumage/integument, colour/pattern, as well as measurements of certain qualitative and quantitative structures (Table 6.1), and several vocal characters extracted from Crowe at al. (1992) and Chapter 4 were scored (Table 6.2). Morphometric characters representing bill-length, tarsus- and spur-length were obtained using a Vernier Calliper. A stopped wing-rule and a normal ruler were used to measure wingand tail-length, respectively. Wing-length was measured with the wing chord flattened and straightened for maximum accuracy. Molecular characters For within-group molecular analyses of spurfowls, 34 terminal taxa (including the two outgroup taxa) were sampled (Tables 6.3) with respect to mitochondrial Cytochrome-b (CYTB - 1143 base pairs- bp) characters. Primers used in sequencing molecular markers are listed in Table 6.4 and 6.5. The 1143 bp long CYTB gene was sequenced for all taxa and Genbank accession numbers are detailed in Appendix 6.1. 228 72% of DNA samples were toe-pads sub-sampled from museum skins. Due to the degraded state of the DNA, the CYTB gene for the toe-pads was sequenced in multiple fragments (six for each sample) using spurfowl-specific primers – Table 6.5). Maps and mapping of distribution records of investigated taxa The maps showing the distribution ranges of spurfowls (Fig. 6.1 – 6.7) were produced following the same procedure as outlined in Chapter 5. Data analyses Phylogenetic analyses Two phylogenetic inference methods, parsimony and maximum likelihood with different optimality criteria were employed to generate phylogenetic hypotheses as outlined for francolins in Chapter 5. As suggested by the results in Chapter 2 (Fig. 2.1, 2.2, 2.3, 2.5, 2.6), all the data matrices were rooted on Alectoris chukar and Coturnix coturnix. Genetic distances Uncorrected pairwise distances (Table 6.6) were calculated in PAUP ver. 4.0b10 (Swofford 2002) and were transformed into percentages. Results and Discussion Separate versus combined phylogenetic analyses 229 The parsimony analysis for the CYTB data set (with 1143 bp characters, yielded 252 parsimony informative characters, 150 variable characters which are parsimony uninformative and two trees with 995 steps), for the organismal data set (with 33 characters, yielded 33 parsimony informative characters and 397 trees with 192 steps), for combined CYTB and organismal data set (with 1176 characters, yielded 283 parsimony informative characters, 148 variable characters which are parsimony uninformative and two trees with 1191 steps). The systematics of Hall’s species groups of spurfowls Bare-throated Group The Bare-throated Group consists of four traditionally recognized species, polytypic Pternistis afer, P. swainsonii, P. leucoscepus, and a monotypic species P. rufopictus. Pternistis afer is comprised of the following recognized subspecies: (afer, harterti, nyanzae, böhmi, intercedens, itigi, cranchii, punctulatus, benguellensis, leucoparaeus, humboldtii, swynnertoni, castaneiventer, melanogaster, loangwae, lehmanni, notatus, krebsi, cunenensis, cooperi); P. swainsonii (swainsonii, lundazi, gilli, damarensis, chobiensis); and P. leucoscepus (leucoscepus, infuscatus, holtemülleri, keniensis, kilimensis, tokora, muhamed-ben-abdullah) (Table 1.2). All four of the analyses recovered monophyly of the Bare-throated Group, although with varying levels of support (Fig. 6.8, 6.10, 6.11). Pternistis nobilis joins the Bare-throated Group in the organismal tree (Fig. 6.9). These results are consistent with those of Bloomer and Crowe (1998), but contradict those of Crowe and Crowe (1985), Crowe et al. (1992), who suggested the Bare-throated group to be paraphyletic. 230 Within the Bare-throated Group, there is a split resulting in P. afer, P. rufopictus and P. cranchii grouping together in most analyses with moderate support (Fig. 6.8, 6.10, 6.11). The putative subspecies afer and humboldtii are sister to each other and they are, in turn sister to P. rufopictus, with P. cranchii being basal in this clade. Pternistis rufopictus is embedded within the afer and cranchii clade in the combined CYTB/organismal, organismal and ML trees, where it is sister to P. afer. The southern P. swainsonii is sister to the eastern P. leucoscepus of which the two subspecies, P. l. leucoscepus and P. l. infuscatus share a sister relationship. Pternistis afer complex This is the most widespread and variable of all the species within the Bare-throated Group and it has an extremely complex distribution (Fig. 6.1). Pternistis afer is mainly diagnosed by having red bare throat skin, bill and legs. Hall (1963) maintained that this species forms a complex which could be split into two types with intervening blocks of hybridization, the “cranchii” and the “afer” type subspecies. The “cranchii” subspecies includes all the subspecies of the southern Congo, northern Angola, Northern Zambia, western Tanzania, Uganda and Lake Victoria shores. These are the subspecies characterized by having vermiculated underparts, with sparse chestnut streaks on the belly. Cranchii has a black and grey facial pattern and it shares the least genetic distance with afer differing at 1% of sequence divergence (Table 6.1). The “afer” subspecies forming the complex have no vermiculation and are strongly patterned in black, white and some grey on the belly and face. In the nominate afer subspecies the face is white and the belly is streaked in white and black, with feathers having black centres and white edges. The black and white subspecies have a 231 broken distribution with the nominate subspecies afer being restricted to south western Angola along the escarpment and the Cunene basin. The underparts of the Benguella, southern Angola subspecies benguellensis has a mix of afer- and cranchii-type features (black, white and chestnut streaks) due to hybridization between the afer- and cranchiitype subspecies. Among the South African subspecies, the face is black and the underparts are streaked with black, white and maroon. The bird from north-central Africa (humboldtii) (Fig. 6.1) is diagnosed by having white jaw feathers and black belly patch. The feathers on the breast are mainly grey with black shaft streaks that contrast with those of the lower belly which form a black patch and those of the flanks that are streaked black and white. The least genetic distance exists between humboldtii and afer differing at 2% of sequence divergence. In short, what characterizes Hall’s cranchiitype subspecies is that the belly feathers have a buff central streak vermiculated with blackish grey and margined with broad chestnut (degree of chestnut colour varies from one subspecies to the other). The belly feather of the afer-type subspecies have a broad greyish black central streak with buff margins (particularly in the nominate subspecies afer) or with a thin greyish black central streak separating the long buff parallel streaks margined with black or sometimes maroon. In Mozambique, Tanzania and south Kenya the afer-type subspecies have very broad black margins and a much narrower central buff to white streak. A range of additional subspecies of intermediate phenotype have been described where these three subspecies are para/sympatric, but they lack the morphological cohesion necessary for recognition. Generally, the nucleotide sequence divergences do not reveal clear cut differences among the P. afer subspecies. Those from the Cunene area of western Angola are 1% sequence divergent from cranchii, suggesting close genetic ties between 232 afer and cranchii, presumably due to hybridization. Indeed, the <1% genetic differences between the subspecies, nyanzae, böhmi, itigi, intercedens and harterti and cranchii indicate that these may be recently diverged lineages. The furthest distance that exists between the cranchii-type is that with humboldtii at 3%, subsequently followed by 3% to the South African afer. The results point to a split of P. afer into what Hall (1963) referred to as the “afer-type” and the “cranchii-type”. The afer-type lineage which include the nominate subspecies afer, castaneiventer/nudicollis, lehmanni, notatus and krebsi and also a group consisting of black and white subspecies that have a black belly patch as in humboldtii (including swynnertoni, melanogaster, leucoparaeus and loangwae). The nominate afer is disjunctly distributed from the South African afer-type allies. The black and white subspecies have a broken distribution with the nominate afer being restricted to the south-western part of Angola along the escarpment and the Cunene basin. The core cranchii-type includes cranchii, nyanzae, harterti, intercedens, itigi and böhmi, cunenensis, benguellensis and punctulatus. These subspecies are diagnosed by having vermiculated underparts and sparse chestnut streaks on the belly (chestnut being replaced with maroon in harterti, and chestnut and a tinge of black in punctulatus). The three subspecies, böhmi, itigi and intercedens have in addition a mixture of white and black tinge on the belly. The subspecies, cunenensis and benguellensis are included in the cranchii-type because they seem to be intermediate between the cranchii- and the afer-type subspecies and this is revealed in their morphological attributes, genetic affinities and in how they relate with the other forms. It is recommended that Pternisti afer (Müller, 1776) be broken into two subspecies, that is, Pternistis afer afer (Müller, 1776) which includes castaneiventer, nudicollis, lehmanni, notatus and krebsi and 233 another subspecies Pternistis afer humboldtii (Peters, 1854) which include swynnertoni, melanogaster, leucoparaeus and loangwae (Appendix 6.2). There is a geographical break between P. a. humboldtii (the black belly patched subspecies) and the south distributed P. a. afer which extends to the north of the Limpopo River basin. Pternistis rufopictus Pternistis rufopictus is a monotypic species that occurs from the south eastern shores of Lake Victoria to the Wembere in Tanzania along the Acacia belt (Fig. 6.2). Unlike the other Bare-throated species, P. rufopictus has orange pink throat skin, grey brown back plumage with dark vermiculation and shaft streaks, and has brown legs. The wing coverts and the feathers on the back are edged with rufous chestnut; the breast is grey with black shaft streaks with the lower belly being richly patterned. The belly feathers have narrow central black streaks separated from rufous chestnut margins by broad buff to white streaks. It is similar to the cranchii-type subspecies in western Tanzania except that it has a partial buff to white jaw feathers (as in humboldtii), and has no vermiculation. This taxon’s distribution range overlaps with that of P. leucoscepus in the southern parts of its distribution. The least genetic distance exists between P. rufopictus and cranchii differing at 2% of sequence divergence. It is an intriguing species that is genetically very different from P. swainsonii and P. leucoscepus (differing from both at 4%). Its phylogenetic placement (embedded within the cranchii/afer clade) points to the possibility that it could be a hybrid between the various taxa belonging to the Bare-throated Group. It is the largest species among the Bare-throated members with a mass of 779-964 g (Johnsgard 1988), has orange pink throat skin which could possibly be an indication of a mix between the red throat skin in 234 P. afer and a yellow throat skin in P. leucoscepus. Its morphologically (and the least genetic distance) places it close to the cranchii-type subspecies but it is phylogenetically close to the afer-type subspecies. Our results provide new evidence that reveals the possibility of P. rufopictus being of hybrid origin. This taxon shows morphological, phylogenetic, genetic affinities and a sympatric distribution with the cranchii-type subspecies. Vocally, it sounds very similar to P. leucoscepus (details in Chapter 4) except that its call is much faster. Despite the potential of a hybrid origin, morphologically, Pternistis rufopictus Reichenow, 1887 still maintains its distinctiveness and has diverged sufficiently from P. leucoscepus, P. swainsonii and from some of the cranchii-type subspecies to be considered a distinct species. Pternistis swainsonii complex Pternistis swainsonii is a polytypic south western African (Fig. 6.2) species diagnosed by having red bare throat skin and black legs. Within P. swainsonii the underparts are less diagnostic and the subspecies (swainsonii, lundazi, gilli, damarensis, chobiensis) are characterized by having a narrow central greyish black streak separated from greyish chestnut margins by broad buff grey vermiculated streaks. The general colour swainsonii is grey brown and lacks the black and white streaking found on the underparts of the nominate subspecies afer. The feathers on the belly have drop-shaped dark brown shaft streaks which increase in size over the lower belly and flanks. The differences among P. swainsonii subspecies are quantitative rather than qualitative such that there are no remarkable plumage differences. Genetically, the smallest divergence is between P. swainsonii and P. rufopictus at 4% sequence divergence, although 235 phylogenetic analyses suggest P. swainsonii is either the basal member of the Barethroated Group or sister to P. leucoscepus. Therefore, Pternistis swainsonii (Smith, 1836) maintains its species status while all the subspecies, lundazi White 1947, gilli Roberts, 1932, damarensis Roberts, 1932 and chobiensis Roberts, 1932 are synonymized with it based on priority. Pternistis leucoscepus complex This is the most distinctive member of the Bare-throated group comprising the north eastern African (Fig. 6.2) subspecies, leucoscepus, infuscatus, holtemülleri, keniensis, kilimensis, tokora, muhamed-ben-abdullah with yellow bare throat skin. These subspecies have similar morphological patterns but differ in the amount of white and chestnut in the feathers of the belly. The back plumage among the different subspecies of P leucoscepus is diagnosed by white shaft streaks and the underparts are streaked with white and chestnut in a way that each feather has chestnut with narrow white edges and a triangular white patch at the tip, tapering up the shaft. The primaries have a conspicuous white patch which is visible during flight (Hall 1963). The northern subspecies infuscatus (differing at 1% sequence divergence from the nominate subspecies) differs from leucoscepus in having more chestnut than white on the underparts contrary to the dominant white over chestnut in leucoscepus. Genetically, P. leucoscepus is close to P. swainsonii, differing by 5% sequence divergence. The morphology and genetic distance point to two subspecies. It is recommended that two subspecies should be recognized: Pternistis leucoscepus leucoscepus (Gray, 1867) and Pternistis l. infuscatus Cabanis, 1868. The rest of the 236 subspecies (holtemülleri, keniensis, kilimensis, tokora, muhamed-ben-abdullah) should be synonymized with P. l. infuscatus Cabanis, 1868 (Appendix 6.2). Vermiculated Group The Vermiculated Group comprises nine traditionally recognized species which are Pternistis hartlaubi, P. bicalcaratus, P. icterorhynchus, P. clappertoni, P. hildebrandti, P. natalensis, P. harwoodi, P. adspersus and P. capensis. The putative subspecies are as follows: for P. hartlaubi (hartlaubi, crypticus, bradfieldi, ovambensis); P. adspersus (adspersus, kalahari, mesicus); P. natalensis (natalensis, neavei); P. hildebrandti (hildebrandti, altumi, helleri, fischeri, johnstoni, grotei); P. bicalcaratus (bicalcaratus), ogilvie-grantii, ayesha, adamauae, thornei); P. icterorhynchus (icterorhynchus, dybowski, ugandensis, emini) and P. clappertoni (clappertoni, sharpii, heuglini, gedgii, nigrosquamatus, konigseggi, testis, cavei) (Table 1.2). In comparing all the phylogenetic trees (Fig. 6.8, 6.9, 6.10, 6.11), the position of P. hartlaubi is consistently basal to all the other spurfowl taxa in the parsimony trees except in the organismal parsimony tree where it forms a sister relationship with the Montane Group member P. camerunensis (Fig. 6.9). In the ML (Fig. 6.11) tree P. hartlaubi is embedded within the Montane group with an unresolved branch. These results are in agreement with those in Bloomer and Crowe (1998). Also Crowe et al. (1992) hinted at the uncertain genetic relationships between P. hartlaubi and the other spurfowls and found that it was morphometrically the most distinct spurfowl and speculated that this taxon shared morphological similarities to the Old World quails. Bloomer and Crowe (1998) concluded that it represents a very early offshoot from the partridge-francolin lineage. 237 The other eight species split themselves geographically into the northern (P. bicalcaratus, P. icterorhynchus, P. clappertoni, P. harwoodi) and southern (P. capensis, P. natalensis, P, hildebrandti, P. adspersus) Vermiculated species in all the trees (Fig. 5.8, 5.9, 5.10, 5.11). The northern Vermiculated clade emerges to be sister to the Barethroated Group in the combined CYTB/organismal (Fig. 6.8) and organismal (Fig. 6.9) tree except in the CYTB (Fig. 6.10) tree where it is only members of the P. clappertoni complex and P. harwoodi that group with the Bare-throated Group while members of the P. bicalcaratus complex and P. icterorhynchus form a paraphyletic clade at the basal parts of the tree. The northern Vermiculated taxa are paraphyletic in the ML tree. Surprisingly P. harwoodi is embedded within P. clappertoni complex forming a sister relationship with P. c. sharpii with 90% and 99 BS in the combined CYTB/organismal (Fig. 6.8), organismal (Fig. 6.9) and CYTB (Fig. 6.10) trees. Pternistis harwoodi and P. clappertoni complex are outliers in the spurfowl clade in the ML tree, The sister relationship between the northern Vemiculated and the Bare-throated clades as consistently recovered in this study, contrasts the finding in Bloomer and Crowe (1998) who found that the southern Vermiculated taxa and not the northern Vermiculated taxa were consistently grouping with each other. This kind of relationship is supported in only one analysis (CYTB – Fig. 6.10) in this study even though the Bare-throated clade is part of a large clade which incorporates two species from the northern Vermiculated clade and also the two Montane sister species (P. jacksoni and P. swierstrai) (CYTB – Fig. 6.10). The southern Vermiculated taxa (P. capensis, P. natalensis, P, hildebrandti, P. adspersus) form a monophyletic clade only in the combined CYTB/organismal (Fig. 5.8), CYTB (Fig. 6.10) and Fig. 6.11 tree (64%, 88%, 95% BS respectively). Pternistis capensis is an outlier in the southern Vermiculated clade forming a paraphyletic 238 relationship with P. adspersus, P. natalensis and P. hildebrandti with P. natalensis being closely related to the two sister subspecies, hildebrandti and fischeri (98% BS CYTB/organismal (Fig. 6.8 tree). Pternistis capensis emerges as a sister taxon to P. adspersus in CYTB (Fig. 6.10) and ML (Fig. 6.11) tree, but this relationship received no support. In the organismal (Fig. 6.9) tree, P. natalensis is sister to P. adspersus but with no nodal support. It is very clear that the Vermiculated Group is polyphyletic just as found in Crowe and Crowe (1985), Crowe et al. (1992) and Bloomer and Crowe (1998). However, the relationships among taxa within the northern Vermiculated clade appear consistent unlike among taxa within the southern Vermiculated clade; the position of P. hartlaubi as a basal taxon seems consistent. Pternistis bicalcaratus complex Pternistis bicalcaratus is distributed from Senegambia to the Central African Republic, except for one taxon, ayesha, that occurs in isolation in Morocco (Fig. 6.5). All the recognized subspecies of P. bicalcaratus are similarly patterned above and below differing in the degree of colouration, patterning, vermiculation and the size of the arrow-shaped buff marks in the centre of the belly feathers. The nominate subspecies bicalcaratus is paler with a more rufous crown, vermiculated with V-patterning and slightly rufous on the lower neck. It has buff underparts distinctly and heavily streaked with blackish and chestnut with small arrow-shaped buff marks on most belly feathers. The heavily patterned ayesha (from Morocco – not mapped) looks very similar to bicalcaratus, but is faintly vermiculated and slightly more rufous on the lower neck, with small arrow-shaped buff marks on belly feathers. The darkest subspecies is adamauae (with 2% sequence divergence from bicalcaratus) with very little rufous on 239 lower neck, and the underparts are more buff with extremely reduced chestnut and have larger arrow-shaped buff marks in centre of belly feathers. The least genetic distance exists between bicalcaratus and P. icterorhynchus differing at 3% of sequence divergence. It is apparent that P. bicalcaratus represents a cline with regard to the overall colouration of back plumage of the various subspecies and the patterning of plumage on the belly even though these subspecies are not strikingly different. The subspecies bicalcaratus is heavily patterned like thornei but the two differ in that bicalcaratus is less dark compared to thornei. Adamauae and Ogilvie-granti are the darkest subspecies but with less patterning. The morphological and genetic evidence reveal that adamauae is distinct from bicalcaratus suggesting that there may be two subspecies within the P. bicalcaratus complex. This should be Pternistis bicalcaratus bicalcaratus (Linnaeus, 1766) which should include ayesha, thornei and ogilvie-granti and Pternistis bicalcaratus adamauae Neumann, 1915. It is possible that ogilvie-granti could well be made a synonym of P. b. adamauae but this decision could not be taken in the absence of genetic evidence. Pternistis clappertoni complex This species has its distribution range covering Cameroon, Central African Republic, Chad, Eritrea, Ethiopia, Mali, Mauritania, Niger, Nigeria, Sudan, and Uganda (Fig. 6.5). Within P. clappertoni, the nominate subspecies clappertoni is greyish brown above with barring on the flight feathers. This subspecies is buff below with black to brownish marks. The hind and lower neck have U-patterning as in P. icterorhynchus. The subspecies sharpii (which differs at 1% sequence divergence clappertoni) has marks on the belly which look streakier than those in clappertoni in a buffy white background 240 with the breast being similarly V-patterned extending on to the back. The primaries are barred as in the nominate subspecies. It is genetically close to P. harwoodi differing at 1% of sequence divergence. The major plumage differences lie in the colour of the background on the belly and the shape of the marks on the belly. There is very little genetic divergence between the nominate clappertoni and sharpii and nigrosquamatus. Therefore, two subspecies are recognized which are Pternistis clappertoni clappertoni (Children & Vigors, 1826) and P. c. sharpii Ogilvie-Grant, 1892. Pternistis c. clappertoni should include heuglini, gedgii and cavei while konigseggi, nigrosquamatus and testis should be synonymized with P. c. sharpii. Pternistis icterorhynchus complex Pternistis icterorhynchus is distributed in the Central African Republic, Democratic Republic of the Congo, Sudan, and Uganda (Fig. 6.5). It consists of four putative subspecies, icterorhynchus, dybowski, ugandensis and emini mainly diagnosed by having distinct U-patterning on the lower neck and more vermiculation on the back than other vermiculated taxa. The underparts are buff with dark brown markings. It shares the least genetic distance with bicalcaratus differing at 3% of sequence divergence. There was no success in sequencing all the subspecies within P. icterorhynchus. Morphologically, there are no striking differences between icterorhynchus and emini except that emini is less patterned on the belly as opposed to icterorhynchus which is extensively patterned whereas ugandensis has some chestnut on the flanks. In the absence of genetic data, no subspecies are recognized. 241 Pternistis harwoodi This is a poorly known species from the Blue Nile gorges (Fig. 6.5), the male bird which was examined is grey speckled and finely barred with blackish and buff above. The hind and lower neck, sides of face, and throat are speckled with black and white. It has irregular double-V patterning on the belly which tends to be scattered just on the lower extreme of the belly on the buff background. Pternistis harwoodi shares the least genetic distance with sharpii differing at 1% of sequence divergence. Phylogenetically, P. harwoodi and sharpii are sister taxa and also share a sympatric distribution. The two are not similar morphologically except that they are both barely patterned distally on the belly. This is a case where possibily with time morphology and DNA will align. While Pternistis harwoodi Blundell & Lovat, 1899 is recognized as a valid species, it could be concluded from the above evidence that P. harwoodi could tentatively be a well-marked subspecies of P. clappertoni or a linking form through sharpii. Pternistis hildebrandti complex Pternistis hildebrandti, a species distributed in central and western Kenya, extending south through Tanzania to south eastern Zaire, northeastern Zambia and Malawi (Fig. 6.6), shows marked plumage dimorphism with the males of the nominate subspecies hildebrandti being greyish brown above with vermiculation, the hind and lower neck streaked black with white margins, the underparts are marked and mottled with black and white almost like in P. natalensis. The females have similar backs to the males but they have different underparts which are rusty. The subspecies johnstoni (differing at 1% of sequence divergence) is like hildebrandti, but differs in that the female has the 242 nape, hind neck and breast uniform with the rest of the plumage. It shares the least genetic distance with P. natalensis differing at 1% of sequence divergence. The genetic and morphological evidence support a split of P. hildebrandti into two subspecies, that is Pternistis hildebrandti hildebrandti Cabanis 1878 and Pternistis hildebrandti fischeri Reichenow, 1887. The subspecies, altumi, 1884 and helleri are synonymized with P. h. hildebrandti, whereas johnstoni and grotei are synonymized with P. h. fischeri. Pternistis natalensis complex Pternistis natalensis is a species complex with the putative subspecies natalensis and neavei from south eastern Africa (Fig. 6.6). The hind neck is mottled black and white, the back is highly vermiculated and it is greyish brown with variable black, whitish and buffish markings. The rest of the belly is buff with the upper belly to mid-belly being heavily patterned in black and buff. The patterning is concentrated on the upper part of the belly with the extreme lower belly having no or few marks. Pternistis natalensis shares the least genetic distance with P. hildebrandti differing at 1% of sequence divergence. However, it is particularly close to altumi, fischeri and johnstoni differing at 1%. The morphological similarities between P. natalensis and male P. hildebrandti could have either been a possibility for P. natalensis and P. hildebrandti interbreeding in the southeast region or that P. natalensis and P. hildebrandti could have been the same taxon that got broken up with some changing in appearance but not genetically. Based on the above-mentioned evidence and the absence of genetic data for neavei and thamnobium, the decision would be to recognize Pternistis natalensis Smith 1833 as a species and synonymize neavei with it. 243 Pternistis adspersus complex Pternistis adspersus consists of three putative subspecies (adspersus, kalahari, mesicus) and it is distributed in Namibia and Botswana (Fig. 6.6). It is a very different looking species with the overall plumage dominated by minute vermiculation on the upperparts and with distinct black and white barring on the underparts and variably on the lower neck. This species shares the least genetic distance with P. capensis differing at 4% of sequence divergence. The subspecies, adspersus and kalahari are inseparable morphologically. The specimens of mesicus were not examined and no description was provided in the literature in which it was recognized. This study acknowledges that Pternistis adspersus Waterhouse 1838 is a valid species, and make no taxonomic decision on the status of other subspecies. Pternistis capensis Pternistis capensis is an undisputed monotypic species which is endemic to the southwestern Cape of South Africa (Fig. 6.6). It is the largest vermiculated species (males weighing ~900 g with females being ~600 g – Johnsgard 1988) and is endemic to the south western part of South Africa (the Cape) (Fig. 6.6). It is characterized by having distinctive uniform brown and white double V- or U-patterning on the back and on the belly while the pattern on the throat is reduced to form some irregular black flecking. The belly patterning also has distinct white shaft streaks. This species shares the least genetic distance with P. adspersus at 4% sequence divergence. Pternistis capensis (Gmelin, 1789) is acknowledged as a valid species due to its morphological and genetic differentiation (4-5%) compared to P. natalensis, P. adspersus and P. hildebrandti and also its phylogenetic position. 244 Pternistis hartlaubi complex Pternistis hartlaubi occupies granite and sandstone outcrops surrounded by semi-desert steppe and is confined to northern Namibia and extreme south western Angola (Fig. 6.7). It is comprised of the subspecies, hartlaubi, bradfieldi, crypticus and ovambensis. Is one of the spurfowl species that exhibits striking plumage dimorphism with the difference in the subspecies being mainly in the degree of colouration. The male hartlaubi has a buff throat with black streaks that continues to the sides of neck through to the hind neck and to the breast. The belly of male birds is like the underparts of the females which is pale tawny. The sides of face and chin and belly of females is pale tawny. It shares the least genetic distance with P. ochropectus differing at 9% of sequence divergence. There is limited genetic and morphological evidence for the separation of hartlaubi and crypticus and therefore, only one taxon is recognized, Pternistis hartlaubi Bocage 1869 and bradfieldi, crypticus and ovambensis are synonymized with it. Nonetheless, P. hartlaubi is clearly an ancient African paleoendemic. The Montane Group The Montane Group encompasses seven species, P. erckelii, P. ochropectus, P. castaneicollis, P. jacksoni, P. nobilis, P. camerunensis and P. swierstrai. The putative subspecies are as follows: within Pternistis erckelii (erckelii, pentoni); P. nobilisis (nobilis, chapini); P. castaneicollis (castaneicollis, bottegi, gofanus, ogoensis, kaffanus, atrifrons; P. jacksoni (jacksoni, pollenorum, gurae) (Table 1.2). The Montane Group is polyphyletic in all the analyses (Fig. 6.8, 6.9, 6.10, 6.11) with all the Montane taxa appearing at the base of the tree. This is in contrast with 245 Crowe and Crowe (1985) and Crowe et al. (1992) wherein the monophyletic status was certain. Despite the Montane Group being polyphyletic, the north eastern species (P. erckelii, P. ochropectus, P. castaneicollis) form a monophyletic clade in all the analyses (Fig. 6.8, 6.9, 6.10, 6.11). Pternistis ochropectus is consistently sister to P. erckelii, and P. nobilis is consistently sister to P. camerunensis. The phylogenetic position of Pternistis swierstrai and P. jacksoni is uncertain in all the trees. The evolutionary history of the Montane spurfowls still needs further scrutiny in line with the challenge that this study rejects Hall’s (1963) speculation that they form a monophyletic assemblage. Further, she also postulated that P. swierstrai and P. camerunensis were the first species to be isolated from the ancestral stocks of the other species. This study consistently reveals that P. camerunensis and P. nobilis were isolated first. Pternistis ochropectus This is a monotypic species endemic to the evergreen forest of Djibouti (Fig. 6.3). A closer analysis of the plumage patterning on the belly feathers of P. ochropectus, P. erckelii and P. castaneicollis renders them almost similar. They have a broad buff central streak constricted in the middle and expanded distally into a tear drop, margined by a greyish black U-shaped streak. This species shares the least genetic distance with P. erckelii differing at 3% of sequence divergence. This study acknowledges Pternistis ochropectus (Dorst & Jouanin, 1952) to be a valid species. Pternistis jacksoni complex This is a large species endemic to Kenya (Fig. 6.3) with a body mass ranging from ~1130-1160 g (Johnsgard 1988). It encompasses three subspecies, jacksoni, pollenorum 246 and gurae. It has a buff throat and greyish lower neck with the proximal part of the lower neck being similarly patterned to the rest of belly. The lower neck feathers are chestnut-coloured edged with buffish to white, but the degree of chestnut and buff and white varies. It shares the least genetic distance with P. griseostriatus with which it differs at 5% of sequence divergence. There is a possibility that there could be two entities (possibly subspecies) within P. jacksoni i.e., jacksoni and pollenorum being one entity and gurae being another entity based only on morphology. However, in the absence of molecular evidence for pollenorum and gurae, our decision is speculation only and hence only Pternistis jacksoni O. Grant, 1891 can be acknowledged as a valid species. Pternistis nobilis complex This species comprises two putative subspecies, nobilis and chapini and is distributed in the mountain forests of the eastern Democratic Republic of the Congo, southwest Uganda and borders between Burundi and Rwanda (Fig. 6.3). The head, uppertail coverts and primaries are grey brown. The wing coverts and the lower neck are deep maroon, with light grey scalloping on the lower neck. The throat colour is buff white with the rest of the belly being chestnut with grey edges replacing the broad buff to white edges found in the nominate jacksoni. It is genetically close to P. camerunensis differing at 7% of sequence divergence. The morphological difference observed between nobilis and chapini are quantitative with regard to colouration and hence only one species is recognized in the absence of molecular evidence even though the two subspecies are found on different 247 mountains. Pternistis nobilis Reichenow, 1908 remains a valid species and chapini should be synonymized with it pending further investigation. Pternistis erckelii complex Pternistis erckelii occurs in the Ethiopian massif (Fig. 6.3) and it is the largest spurfowl (males have a maximum mass of ~1590 g, with females having a maximum mass of ~1136 g) (Johnsgard 1988), and is characterized by its black forehead, eye-stripe and chestnut crown. The lower neck is grey like the upper belly, but has greyish brown margins and a thin central buff streak whereas the upper belly feathers have central greyish black streaks. The lower belly feathers consist of a broad buff central streak constricted in the middle and expanded distally into a tear drop, margined with rufous. It shares the least genetic distance with P. ochropectus at 3% of sequence divergence. The specimens of pentoni were not available for examination. Irrespective of the quantitative differences presented in literature, the recommendation is that pentoni should in the meantime be included in Pternistis erckelii (Rüppell, 1835). Pternistis castaneicollis complex Pternistis castaneicollis is distributed in Ethiopia, Somalia, and possibly Kenya (Fig. 6.3). The subspecies, castaneicollis, bottegi, gofanus, ogoensis, kaffanus and atrifrons are recognized within P. castaneicollis (Table 1.2). Pternistis castaneicollis is generally similar to P. erckelii with the belly feathers being made up of a broad buff central streak which is constricted in the middle and expanded distally into a tear drop, margined with rufous. The eastern Ethiopian subspecies have an extensive double-patterning on the back with wing coverts and the upper belly being clearly defined in black and white, 248 with some ochre and chestnut. The subspecies atrifrons from southern Ethiopia is quite different from the other subspecies with a lesser degree of colouration and patterning on the belly. The less defined belly patterning is in brown and buff, and the throat and the belly cream instead of white as in other subspecies. Despite its distinct morphology, atrifrons was found to have similar voice, habits and environment (Benson 1945) to other subspecies of P. castaneicollis. It shares the least genetic distance with P. camerunensis differing at 8% of sequence divergence. The subspecies, ogoensis, bottegi, gofanus and kaffanus are morphologically similar to the nominate castaneicollis. The southern Ethiopian subspecies, atrifrons is the one that is quite distinct in being less coloured and patterned on the belly as well as having the well-defined patterning and rich colour on the back. Despite the similar voice, habit and habitat that atrifrons shares with other P. castaneicollis subspecies (Hall 1963), it would therefore be crucial to include its DNA in the analysis in order to gain insight into its phylogenetic placement given its distinct morphology. It could be that atrifrons is a valid species sister to P. castaneicollis Salvadori, 1888 which remains a valid species. Pternistis camerunensis This is a monotypic, sexually dimorphic species which is endemic to Mount Cameroon (Fig. 6.3). Its taxonomic status as a monotypic species has never been disputed. The male P. camerunensis has the back (strictly excluding the lower neck) and the wings coloured rich dark brown, crown grey brown and the rest of the belly and the lower neck are plain grey with some black feather centres (Hall 1963). The upper- and underparts of the female are mottled and vermiculated with black, dark brown and buff 249 and some off-white U- to V-patterning on belly and U-patterning on the lower neck. It shares the least genetic distance with P. nobilis differing at 7% of sequence divergence. Pternistis camerunensis Alexander, 1909 is acknowledged as a valid species. Pternistis swierstrai This monotypic species is endemic to the patches of evergreen forest in the highlands and along the escarpment of Angola (Fig. 6.3). It is slightly sexually dimorphic in its plumage and it is very different and isolated from the other members of the group. Male and female birds have white eye stripes not found in any other member of the Montane Group. The male breast is black contrasting with the buff throat while the belly consists of broad buff central streaks with blackish margins. The females were not examined and their underparts are according to Hall (1963), buff with irregular black or brown blotches or bars which are concentrated on the upper belly to form a mottled band and are sparse in the centre of the belly. It shares the least genetic distance with afer differing at 6% of sequence divergence. Scaly Group The Scaly Group comprises three allopatric species, P. ahantensis, P. squamatus, P. griseostriatus, with subspecies, P. squamatus (squamatus, maranensis, schuetti, usambarae, uzungwensis, doni, zappeyi, tetraoninus, chyuluensis) and P. ahantensis (ahantensis, hopkinsoni) being recognized (Table 1.2). The Scaly taxa are paraphyletic, in which case P. squamatus and P. griseostriatus have close phylogenetic association (Fig. 6.8, 6.10, 6.11). Squamatus and schuetti appear as sister taxa in some analyses. 250 (Fig. 6.8, 6.9), and schuetti is sister to P. griseostriatus others (Fig. 6.10, 6.11). Generally, the phylogenetic relationship of the Scaly Group members is unresolved. The general finding is that there are no morphological characters that link up the three species within the group except that they all have some degree of scaliness. There is no qualitative break in characters. There is a cline from the west African P. ahantensis which has clearly defined patterning through to north, east and central African P. squamatus (moderately patterned) and down to southwestern African P. griseostriatus which has less defined patterning. Pternistis griseostriatus is likely to represent the end of the cline. There is a large genetic jump of 4% between P. squamatus and P. ahantensis and 5% between P. squamatus and P. griseostriatus. Pternistis squamatus complex Pternistis squamatus, a species distributed in south central Nigeria, south eastern to western Zaire and east to Uganda, south Sudan and south western Ethiopia, then south to east Zaire, north eastern Tanzania and extreme northern Malawi (Fig. 6.4), consists of the following subspecies, squamatus, maranensis, schuetti, usambarae, uzungwensis, doni, zappeyi, tetraoninus, chyuluensis (Table 1.2). Pternistis squamatus is not sexually dimorphic, although female birds tend to be less strikingly vermiculated. The nominate subspecies squamatus is characterized by having indistinctly vermiculated upperparts with faint U-patterning on the lower neck and the feathers have blackish centres tinged with red-brown. The belly is plain brown with a scaly pattern and ill-defined dark shaft streaks margined buff. Schuetti (with 1% sequence divergence) differs from squamatus in being less vermiculated with the pattern on the lower neck being more clearly defined and with more red-brown at the centre of feathers. The edges of the darker belly feathers 251 have buff edges giving the streaky effect. It shares the least genetic distance with P. griseostriatus differing at 3% of sequence divergence. The two subspecies, schuetti and maranensis have slightly diverged genetically and morphologically from squamatus and warrants only subspecies status for the former. The genetic divergence between schuetti and maranensis is just above 1% and the two subspecies differ in the degree of colouration. Therefore, on the basis of the marked genetic distances and different morphology, two subspecies are recognized within P. squamatus, that is, Pternistis squamatus squamatus Cassin, 1857 and Pternistis squamatus schuetti Cabanis, 1880. Pternistis s. schuetti will include maranensis, chyuluensis, doni, usambarae, uzungwensis, zappeyi and tetraoninus. Pternistis ahantensis complex This species occurs in lowlands of Senegambia to south western Nigeria (Fig. 6.4) and it comprises two putative subspecies, ahantensis and hopkinsoni. The belly feathers are richly coloured and streaked with dark brown chestnut edged buff while the upperparts are vermiculated with some white U-patterning (indistinct on the back but very distinct on the lower neck). It shares the least genetic distance with P. squamatus from which it differs with 4% of sequence divergence. The two subspecies, ahantensis and hopkinsoni show little morphological differentiation. Pternistis ahantensis Temminck, 1854 remains a valid species. Pternistis griseostriatus Pternistis griseostriatus is an indistinctive, monotypic species endemic to Angola (Fig. 6.4). The lower neck feathers and wing coverts are chestnut broadly vermiculated as in 252 P. s. squamatus and P. a. ahantensis, but the belly is plain and the upper belly and flank feathers are chestnut and edged with greyish or creamy buff. It shares the least genetic distance with squamatus differing at 3% of sequence divergence. This study acknowledges Pternistis griseostriatus O. Grant 1890 as a valid species due to its marked genetic distinctiveness (4%) compared to P. squamatus and P. ahantensis. Conclusions The definition of species and subspecies will continue to be contested as more and more data are accumulated, analytical methods continue to be improved and people have freedom to think and present their own views. The good side of this will be that there will be great improvement on how species and subspecies are defined and delimited. The conflict exists between characters from different sources possibly because of hybridization and incomplete lineage sorting and these conditions make taxonomic delineation of spurfowls challenging. Despite the striking diversity within the group, all spurfowls share a single evolutionary path. However, among all of Hall’s putative monophyletic species groups, the Bare-throated Group is the only one to be recovered as monophyletic while others remain largely paraphyletic. The number of species among spurfowls did not increase significantly; P. afer is the only species that was split into two P. afer and P. cranchii. The number of subspecies is reduced (Appendix 6.2). The recommendation is that phylogeographic studies should be conducted for the various species complexes and fresh DNA samples for all the other known subspecies need to be collected and sequenced so that sound taxonomic decisions can be made. 253 Table and Figures Table 6.1. Characters and character states scored and used for phylogenetic analysis of spurfowls. 1. Crown margins: unmargined = 0, grey = 1, buff = 2, grey brown = 3 2. Nares: black = 1, chestnut = 2, grey brown = 3, buff or white =4 buff = 4, white = 5 3. Hindneck patterning: unpatterned = 0, mottled = 1, streaked = 2 4. Hindneck base colour: grey brown = 1, grey black = 2, grey chestnut = 3 rufous brown = 4, black = 5 5. Hindneck margins: unmargined = 0, grey =1, buff = 2, grey brown = 3 6. Lower neck patterning: streaked = 1, mottled = 2, barred = 3 7-10. Back plumage: absent = 0, streaked = 1 (7), mottled = 1 (8), vermiculated = 1 (9), barred =1 (10) 11-13. multistate Uppertail coverts: absent = 0, barred = 1 (11), vermiculated = 1 (12), streaked = 1 (13) 14. Throat: feathered = 1, yellow skin = 2, orange = 3, red = 4 15-18. multistate Undertail coverts: absent = 0, barred = 1 (15), streaked = 1 (16), vermiculated = 1 (17), mottled = 1 (18) 19. Bare skin around eye: none = 0, red = 1, yellow = 2 20. Leg colour: yellow = 1, red = 2, orange red = 3, orange = 4, olive green = 5, orange yellow = 6, black = 7 21. Number of spurs: one = 1, two = 2 22. Wing length (males): <160 mm = 1, <180 mm =2, <200 = 3, >200 =4 23. Culmen length/wing length: <.16 = 1, <.20 = 2, >= .20 =3 24. Tail length/ wing length: <.54 = 1, >= .54 = 2 25. Sexual dimorphism (plumage): absent = 0, present = 1 26. Sexual dimorphism (wing length): female >= .9 of male = 0, < .9 = 1 27. Vocalization strophe duration: <= 0.3 sec = 1, > 0.3 < 0.6 = 2, > 0.6 = 3 28. Number of elements: 1 = 1, 2 = 2, >2 = 3 29. Inter-element interval: absent or indistinct = 1, distinct = 2 30. Cackle-trill: absent = 0, present = 1 31-32. multistate Strophe character: tonal = 0 (31), trill = 1 (32) 33. ‘KO-WAAARK’ advertisement call: absent = 0, present = 1 254 Table 6.2. Organismal character matrix used in phylogenetic analysis. ? indicates information which could not be sourced. Character no. Taxon 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 123456789 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 Pternistis hartlaubi Pternistis camerunensis Pternistis nobilis Pternistis erckelii Pternistis swierstrai Pternistis castaneicollis Pternistis c. atrifrons Pternistis ochropectus Pternistis jacksoni Pternistis squamatus Pternistis s. schuetti Pternistis ahantensis Pternistis griseostriatus Pternistis b. bicalcaratus Pternistis b. ayesha Pternistis b. adamauae Pternistis icterorhynchus Pternistis clappertoni Pternistis c. sharpii Pternistis harwoodi Pternistis h. hildebrandti Pternistis h. fischeri Pternistis natalensis Pternistis adspersus Pternistis capensis Pternistis leucoscepus Pternistis l. infuscatus Pternistis rufopictus Pternistis afer afer Pternistis cranchii Pternistis afer humboldtii Pternistis swainsonii 011101110 0 1 1 0 1 1 0 0 0 0 1 2 1 3 2 1 0 3 3 2 0 1 1 0 011202011 0 1 1 0 1 1 0 0 0 1 2 2 2 1 1 1 0 2 2 2 0 1 0 0 011201100 0 0 0 0 1 0 0 0 0 1 2 2 3 1 1 0 0 2 2 2 0 1 1 1 021301001 0 0 0 0 1 0 1 1 0 0 2 2 4 1 2 0 1 3 3 2 1 0 1 0 010501100 0 0 0 0 1 0 0 0 0 0 2 2 3 2 2 1 0 3 3 2 1 0 1 0 010301101 0 0 0 0 1 0 1 0 0 1 2 2 4 1 2 0 1 3 3 2 1 0 1 0 012121101 0 0 0 0 1 0 1 0 0 1 2 2 4 1 2 0 1 3 3 2 1 0 1 0 021301001 0 0 0 0 1 0 1 1 0 0 2 2 4 1 2 0 1 2 2 2 0 1 0 0 010301101 0 0 0 0 1 0 1 0 0 0 2 2 4 1 2 0 0 ? ? ? ? ? ? 0 030102011 0 1 1 0 1 0 0 0 1 0 3 2 2 1 2 0 0 3 3 2 0 1 1 0 031102011 0 1 1 0 1 0 1 0 0 0 3 2 2 1 2 0 0 3 3 2 0 1 1 0 040101101 0 0 1 01 0 1 1 0 0 4 2 2 1 2 0 0 2 2 2 0 1 1 0 030101101 1 1 0 0 1 0 0 0 0 0 3 1 1 1 2 0 0 3 1 1 0 0 1 0 011401101 0 1 1 0 1 0 1 0 0 0 5 2 2 2 1 0 0 1 2 1 0 01 0 011401101 0 1 1 0 1 0 1 0 0 0 5 2 2 2 1 0 0 1 2 1 0 0 1 0 011202101 0 1 1 0 1 0 1 0 0 0 5 2 2 2 1 0 0 1 2 1 0 0 1 0 011201001 1 1 1 0 1 1 0 0 0 2 6 2 2 2 1 0 0 1 2 1 0 0 1 0 011401100 0 1 1 0 1 0 1 0 0 1 2 2 2 2 1 0 0 2 2 1 0 0 1 0 010101100 0 1 1 0 1 0 1 0 0 1 2 2 2 2 1 0 0 2 2 1 0 0 1 0 011401100 0 1 1 0 1 0 1 0 0 1 2 2 2 2 2 ? 1 2 2 1 0 0 1 0 112211001 1 1 1 0 1 1 0 1 0 0 2 2 2 1 2 1 0 1 3 2 0 0 1 0 112211001 1 1 1 0 1 1 0 1 0 0 2 2 2 1 2 1 0 1 3 2 0 0 1 0 030101001 1 1 1 0 1 1 0 0 0 0 2 1 2 1 2 0 02 3 2 0 1 1 0 010203000 1 1 0 0 1 1 0 0 0 2 2 1 2 1 2 0 0 3 3 2 1 1 1 0 232121100 0 1 0 0 1 0 1 0 0 0 2 2 4 1 2 0 0 3 3 2 0 1 0 0 342131101 0 1 1 0 2 0 0 1 0 1 7 2 3 2 1 0 0 3 2 2 0 0 1 1 012101001 0 1 1 0 2 0 0 1 0 1 7 2 3 2 1 0 0 3 2 2 0 0 1 1 030101101 0 0 1 1 3 0 1 0 0 1 7 2 4 2 1 0 0 2 2 2 0 0 1 1 352101100 0 0 0 0 4 0 1 0 0 1 2 2 3 2 1 0 0 2 2 2 0 1 1 1 010101101 0 0 1 0 4 0 1 1 0 1 2 2 3 2 1 0 0 2 2 2 0 1 1 1 342131100 0 0 0 0 4 0 0 0 0 1 2 2 3 2 1 0 0 2 2 2 0 1 1 1 332131101 0 0 1 1 4 1 0 1 0 1 7 1 3 2 1 0 0 3 2 2 0 1 1 1 255 Table 6.3. Taxa for which DNA sequences were generated. AMNH stands for American Museum of Natural History, FHHM = Franch Natural Histiry Museum, TM = Transvaal Museum, BM = British Museum - Natural History Museum at Tring, SAM = Iziko Museums of Cape Town (Natural History), PFIAO = Percy FitzPatrick Institute of African Ornithology, TMC = Timothy M. Crowe, University of Cape Town, South Africa, GB = GenBank, ‘-’ = Unknown, Br. muscle = Breast muscle, Pect. muscle = Pectoral muscle. The specific and susubpecific epithets are as recorded on specimen label. Taxa name Sample No. Origin Date coll. Sample type PFIAO 108, AMNH 267682 AMNH 541485 BM 1903.10.14.91 BM 1932.5.10.214 TM 28584 TM 20341 BM 1953.54.56 AMNH 202502 AMNH 416180 AMNH211906 TMC 40 SAM 2055756a SAM 2003501 AMNH 202503 PFIAO 109 AMNH 419169 AMNH 541581 Tudor East, Watervalboven Mombola Russisi River E. Transvaal S. Tanganyika Cunene River Selindu, Mabsettler Mwinilunga, N. Rhodesia Poona Singida Tukuyu Buhumbiro Marico River Deka Victoria falls Gagayo, Muranza KenyaTana River, Kenya colony - 2004 1903 1932 1957 1935 1953 2004 1969 1904 2004 - Liver Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Liver Toe-pad Toe-pad Toe-pad Heart Toe-pad Toe-pad AMNH 541471 Badaltino, Shoa AMNH DOT11039 Ethiopia FMNH 1971-1072 Djbouti GB AMNH541435 Rafissa, Abyssinia AMNH541426 Lower Sheikh AMNH261929 East slope, Mt. Kenya AMNH1759 West Ruwenzori TMC 42 Mount Cameroon AMNH 419126 Angola TMC 67 Angola, 14.49 S 13.23 E - 30/06/2010 Toe-pad Liver Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Toe-pad Liver Toe-pad Blood AMNH 541409 PFIAO 117 AMNH 541407 AMNH 763912 AMNH 541411 Nr York Pass, Sierra Leone Kilimanjaro district Tshibati, Dem. Rep. Congo Ndalla Tanda - Toe-pad DNA Toe-pad Toe-pad Toe-pad TM 14682 AMNH 541250 AMNH 541280 AMNH 704359 Gold Coast, Hinterland 1901 Forest of Mamora Kavene District, Sierra Leone Cameroon - Bare-throated Group P. afer P. afer benguellensis P. afer harterti P. afer nudicollis P. afer böhmi P. afer cunenensis P. humboldtii swynnertoni P. cranchii cranchii P. cranchii itigi P. cranchii intercedens P. cranchii nyanzae P. swainsonii P. s. lundazi P. s. chobiensis P. rufopictus P. leucoscepus P. l. infuscatus P. l. muhamed-ben-abdullah Montane Group P. erckelii P. erckelii P. ochropectus P. castaneicollis P. c. bottegi P. c. ogoensis P. jacksoni P. nobilis P. camerunensis P. swierstrai P. swierstrai - Scaly Group P. ahantensis P. squamatus P. s. maranensis P. s. schuetti P. griseostriatus Vermiculated Group P. b. bicalcaratus P. b. ayesha P. b. thornei P. b. adamauae 256 Toe-pad Toe-pad Toe-pad Toe-pad Taxa name Sample No. Origin Date coll. Sample type AMNH 541305 TMC 68 AMNH 541324 AMNH 541341 AMNH 156922 GB AMNH 551345 AMNH 261945 AMNH 347277 AMNH 207771 TMC 120 TMC 121 AMNH 703654 PFIAO 229 PFIAO 206A BM 1927.11.5.18 Takoukout, Camergon Cameroon Adarte S. Ethiopia Fanadji Gilgil River N. Tanganyika Territory Mafinga Mt., N. Rhodesia Neng Marico River, South Africa Namibia Erungo Plateau Kakamas, South Africa - 2005 2004 2006 1927 Toe-pad Br. muscle Toe-pad Toe-pad Toe-pad Blood Toe-pad Toe-pad Toe-pad Toe-pad Liver Br. muscle Toe-pad Heart Liver Toe-pad Vermiculated Group P. clappertoni P. clappertoni P. c. sharpii P. c. nigrosquamatus P. icterorhynchus P. hildebrandti P. h. altumi P. h. fischeri P. h. johnstoni P. h. helleri P. natalensis P. hartlaubi P. h. crypticus P. capensis P. adspersus P. harwoodi 257 Table 6.4. DNA markers sequenced and primers used for PCR amplification and sequencing of preserved tissues. Primer name Primer sequence (5’to 3’) Reference cta gga atc atc cta gcc cta ga act cta cta ggg ttt ggc c atc aca aac cta ttc tc aac gca gtc atc tcc ggt tta caa gac J.G. Groth (pers. comm.) P. Beresford (pers. comm.) P. Beresford (pers. comm.) Edwards & Wilson (1990) Spurfowls Cytochrome b L14578 MH15364 ML15347 H15915 Table 6.5. DNA marker sequenced and primers used for DNA amplification and sequencing of museum toe-pads. Primer name Primer sequence (5’ to 3’) Reference Cytochrome b Spurfowl-specific primers L14851 (General) cct act tag gat cat tcg ccc t Pt-H195 ttt cgr cat gtg tgg gta cgg ag Pt-H194 cat gtr tgg gct acg gag g MH15145 aag aat gag gcg cca ttt gc Kornegay et al. (1993) R. Moyle & T. Mandiwana-Neudani R. Bowie P. Beresford Pt-L143 Pt-H361 gcc tca tta ccc aaa tcc tca c gtg gct att agt gtg agg ag R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani Pt-L330 Pt-H645 tat act atg gct cct acc tgt ac ggg tgg aat ggg att ttg tca gag R. Bowie R. Moyle & T. Mandiwana-Neudani Pt-L633 Pt-H901 ggc tca aac aac cca cta ggc agg aag ggg att agg agt agg at R. Moyle & T. Mandiwana-Neudani R. Moyle & T. Mandiwana-Neudani L2-2312 H15696 cat tcc acg aat cag gct c aat agg aag tat cat tcg ggt ttg atg R. Bowie Edwards et al. (1991) Pt-L851alt Pt-H1050 cct att tgc cta cgc cat cct ac gat gct gtt tgg ccg atg R. Bowie R. Bowie Pt-L961 Pt-L961alt HB20 (General) cga acc ata aca ttc cca c ctc atc cta ctc cta atc ccc ttg gtt cac aag acc aat gtt R. Moyle & T. Mandiwana-Neudani R. Bowie J. Feinstein (pers. comm.) 258 Table 6.6. The inter- and intraspecific uncorrected ‘P’ distances calculated from CYTB data partition only. Taxon names are as known prior to this study. Taxo n name 1 P ternis tis hartlaubi 2 P ternis tis h. crypticus 3 P ternis tis cam erunens is 4 P ternis tis no bilis 5 P ternis tis cas taneico llis 6 P ternis tis c. bo ttegi 7 P ternis tis c. o go ens is 8 P ternis tis erckelii 9 P ternis tis o chro pectus 10 P ternis tis s wiers trai 11 P ternis tis jacks o ni 12 P ternis tis ahantens is 13 P ternis tis s quam atus 14 P ternis tis s . m aranens is 15 P ternis tis s . s chuetti 16 P ternis tis gris eo s triatus 17 P ternis tis b. bicalcaratus 18 P ternis tis b. adam auae 19 P ternis tis b. ayes ha 2 0 P ternis tis b. tho rnei 2 1P ternis tis ictero rhynchus 2 2 P ternis tis clapperto ni 2 3 P ternis tis c. nigro s quam atus 2 4 P ternis tis c. s harpii 2 5 P ternis tis harwo o di 2 6 P ternis tis capens is 2 7 P ternis tis ads pers us 2 8 P ternis tis natalens is 2 9 P ternis tis hildebrandti 3 0 P ternis tis h. altum i 3 1 P ternis tis h. fis cheri 3 2 P ternis tis h. helleri 3 3 P ternis tis h. jo hns to ni 3 4 P ternis tis afer (Ango la) 3 5 P ternis tis afer (SA) 3 6 P ternis tis a. nudico llis 3 7 P ternis tis a. benguellens is 3 8 P ternis tis a. bo hm i 3 9 P ternis tis a. cranchii 4 0 P ternis tis a. cunenens is 4 1 P ternis tis a. harterti 4 2 P ternis tis a. intercedens 4 3 P ternis tis a. itigi 4 4 P ternis tis a. nyanzae 4 5 P ternis tis a. s wynnerto ni 4 6 P ternis tis rufo pictus 4 7 P ternis tis leuco s cepus 4 8 P ternis tis l. infus catus 4 9 P ternis tis l. m uh-ben-abdullah 5 0 P ternis tis s wains o nii 5 1 P ternis tis s . cho biens is 5 2 P ternis tis s . lundazi 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1 4 2 4 3 4 4 4 5 4 6 4 7 4 8 4 9 50 51 52 0% 11% 10% 11% 11% 7% 10% 10% 8% 8% 10% 10% 8% 8% 0% 10% 10% 8% 8% 1% 1% 9% 9% 7% 8% 4% 4% 4% 8% 8% 8% 8% 4% 4% 4% 3% 11% 11% 8% 9% 7% 7% 7% 6% 8% 11% 11% 9% 10% 8% 8% 7% 7% 7% 7% 9% 7% 8% 8% 7% 5% 7% 6% 8% 7% 8% 10% 10% 8% 9% 7% 7% 6% 7% 7% 7% 6% 4% 9% 9% 8% 8% 6% 7% 6% 6% 7% 7% 6% 4% 3% 8% 8% 8% 8% 6% 7% 6% 5% 6% 5% 7% 4% 4% 1% 10% 10% 8% 9% 6% 6% 6% 6% 6% 7% 5% 5% 4% 4% 3% 10% 10% 8% 9% 7% 7% 7% 6% 7% 7% 7% 6% 6% 5% 6% 5% 10% 10% 8% 8% 7% 7% 7% 6% 7% 7% 7% 6% 6% 6% 6% 5% 2% 11% 10% 8% 9% 7% 7% 7% 7% 7% 7% 7% 6% 6% 6% 7% 5% 1% 1% 11% 10% 8% 9% 7% 7% 7% 7% 7% 7% 7% 6% 6% 6% 7% 5% 1% 1% 1% 11% 11% 9% 9% 8% 8% 8% 7% 7% 8% 8% 6% 7% 7% 8% 6% 3% 3% 3% 3% 9% 9% 8% 9% 6% 6% 6% 5% 6% 6% 5% 6% 5% 5% 5% 4% 5% 6% 6% 6% 6% 9% 9% 8% 9% 6% 6% 6% 6% 7% 7% 6% 7% 5% 5% 5% 5% 6% 6% 6% 6% 7% 1% 9% 9% 8% 9% 6% 6% 6% 6% 7% 7% 6% 7% 5% 5% 5% 5% 6% 6% 6% 6% 7% 1% 1% 9% 9% 8% 9% 6% 6% 6% 6% 7% 7% 6% 7% 5% 5% 5% 5% 6% 6% 6% 6% 7% 1% 1% 1% 10% 10% 8% 9% 6% 7% 7% 7% 7% 6% 6% 8% 5% 5% 5% 4% 6% 6% 6% 6% 7% 5% 5% 5% 5% 10% 10% 9% 10% 7% 7% 7% 7% 7% 7% 6% 8% 6% 5% 6% 5% 6% 7% 7% 7% 7% 5% 5% 6% 6% 4% 10% 10% 8% 9% 7% 7% 7% 7% 7% 7% 6% 6% 5% 5% 5% 5% 6% 6% 6% 6% 7% 4% 5% 5% 5% 4% 4% 10% 10% 9% 9% 7% 7% 7% 8% 8% 7% 7% 5% 5% 5% 4% 6% 7% 6% 7% 7% 8% 6% 6% 6% 6% 5% 5% 3% 10% 10% 9% 10% 7% 7% 8% 8% 8% 7% 7% 7% 6% 5% 5% 6% 7% 6% 7% 7% 8% 6% 6% 6% 6% 5% 5% 3% 1% 10% 10% 8% 9% 6% 6% 7% 7% 7% 7% 6% 7% 5% 5% 6% 5% 6% 6% 6% 7% 7% 5% 5% 5% 5% 4% 4% 1% 3% 3% 9% 9% 9% 8% 7% 8% 7% 7% 7% 6% 7% 7% 6% 4% 4% 6% 6% 6% 6% 7% 7% 5% 6% 6% 6% 4% 4% 2% 2% 1% 3% 10% 10% 8% 9% 7% 7% 7% 7% 7% 7% 6% 7% 5% 5% 6% 5% 6% 6% 7% 7% 7% 5% 5% 5% 5% 4% 5% 1% 3% 3% 0% 3% 10% 10% 8% 8% 7% 8% 7% 7% 7% 7% 6% 7% 5% 6% 5% 6% 7% 7% 8% 8% 7% 5% 5% 5% 5% 6% 7% 6% 6% 6% 6% 6% 6% 10% 10% 8% 9% 7% 7% 7% 7% 7% 7% 6% 6% 6% 6% 6% 6% 7% 7% 8% 7% 8% 6% 6% 6% 6% 5% 6% 6% 6% 6% 6% 6% 6% 3% 9% 9% 8% 9% 6% 7% 6% 6% 7% 8% 7% 6% 6% 5% 5% 6% 7% 7% 7% 7% 8% 5% 5% 5% 5% 6% 6% 6% 6% 6% 6% 6% 6% 3% 3% 9% 9% 8% 8% 6% 6% 6% 6% 6% 7% 6% 7% 5% 5% 5% 6% 6% 7% 7% 7% 7% 5% 5% 4% 5% 5% 6% 5% 5% 5% 5% 5% 5% 1% 3% 2% 9% 9% 7% 8% 6% 6% 6% 5% 5% 7% 5% 6% 4% 5% 3% 5% 6% 6% 6% 6% 6% 4% 5% 4% 5% 5% 5% 5% 5% 5% 5% 5% 5% 2% 3% 2% 1% 9% 9% 8% 8% 6% 7% 6% 6% 6% 7% 6% 6% 5% 5% 5% 6% 7% 7% 7% 7% 7% 5% 5% 5% 5% 5% 6% 5% 5% 5% 5% 5% 5% 2% 4% 2% 0% 1% 8% 7% 7% 7% 6% 7% 6% 5% 6% 6% 5% 7% 5% 5% 5% 5% 6% 6% 7% 7% 6% 5% 5% 5% 5% 5% 5% 4% 5% 5% 5% 4% 5% 3% 3% 3% 1% 1% 1% 9% 9% 8% 8% 6% 6% 6% 6% 7% 7% 6% 6% 5% 5% 5% 6% 7% 7% 7% 7% 7% 5% 5% 4% 5% 5% 6% 5% 6% 6% 5% 5% 5% 2% 3% 2% 0% 1% 1% 1% 9% 9% 8% 8% 6% 6% 6% 6% 7% 7% 6% 6% 5% 5% 5% 6% 7% 7% 7% 7% 7% 5% 5% 5% 5% 5% 6% 5% 6% 6% 5% 5% 5% 2% 4% 2% 1% 1% 1% 1% 0% 10% 9% 8% 8% 7% 7% 6% 6% 7% 7% 6% 7% 5% 5% 6% 6% 7% 7% 7% 7% 8% 5% 5% 5% 5% 6% 6% 6% 6% 6% 6% 6% 6% 2% 4% 3% 1% 0% 1% 1% 1% 1% 9% 9% 8% 8% 6% 7% 6% 6% 6% 7% 6% 6% 5% 5% 5% 6% 7% 7% 7% 7% 7% 5% 5% 5% 5% 5% 6% 5% 5% 5% 5% 5% 5% 2% 4% 2% 0% 1% 0% 1% 1% 1% 1% 10% 9% 9% 9% 7% 7% 7% 7% 7% 8% 7% 7% 6% 6% 6% 7% 7% 7% 8% 7% 8% 6% 6% 6% 6% 6% 6% 6% 6% 6% 6% 6% 6% 3% 4% 1% 2% 2% 3% 3% 2% 2% 3% 3% 9% 9% 8% 8% 6% 6% 6% 6% 7% 7% 6% 6% 5% 5% 4% 6% 7% 7% 7% 7% 8% 5% 5% 5% 5% 5% 6% 6% 6% 6% 6% 5% 6% 2% 3% 2% 2% 2% 2% 2% 2% 2% 2% 2% 2% 10% 10% 9% 9% 7% 7% 7% 7% 7% 8% 7% 7% 6% 6% 6% 6% 7% 7% 7% 7% 8% 5% 5% 5% 5% 6% 7% 6% 6% 7% 6% 6% 6% 4% 6% 5% 4% 4% 4% 4% 4% 4% 4% 4% 5% 4% 9% 9% 8% 9% 7% 7% 7% 6% 7% 7% 6% 6% 5% 6% 5% 6% 6% 7% 7% 7% 7% 5% 5% 5% 5% 6% 6% 6% 6% 6% 6% 5% 6% 5% 6% 5% 4% 4% 4% 4% 4% 4% 4% 4% 5% 4% 1% 9% 9% 8% 9% 6% 6% 6% 6% 6% 7% 6% 7% 5% 10% 9% 7% 9% 6% 6% 6% 6% 6% 7% 6% 8% 6% 5% 5% 5% 6% 7% 6% 7% 7% 5% 5% 4% 4% 6% 6% 6% 6% 6% 6% 5% 6% 4% 5% 5% 4% 4% 4% 4% 4% 4% 4% 4% 5% 4% 2% 2% 5% 5% 6% 7% 7% 7% 7% 7% 5% 5% 5% 5% 6% 6% 6% 6% 6% 6% 6% 6% 5% 5% 4% 4% 4% 4% 5% 4% 4% 4% 4% 5% 4% 5% 4% 4% 11% 11% 9% 11% 8% 8% 8% 8% 8% 9% 6% 7% 8% 8% 7% 7% 8% 8% 9% 8% 9% 7% 8% 7% 7% 7% 8% 7% 9% 8% 7% 9% 7% 8% 7% 7% 7% 7% 7% 7% 7% 7% 7% 7% 7% 7% 7% 8% 7% 5% 10% 10% 9% 11% 8% 8% 8% 7% 8% 9% 8% 8% 7% 7% 6% 8% 8% 8% 8% 8% 9% 6% 6% 6% 6% 7% 8% 7% 8% 8% 7% 9% 7% 8% 7% 6% 6% 6% 6% 6% 6% 6% 6% 6% 6% 6% 7% 7% 6% 3% 2% 259 1 2 1 1 Bare-throated Group 1. Pternistis cranchii 2. P. afer humboldtii 3. P. a. afer 3 2 1 1 2 3 3 Figure 6.1. Distribution ranges of Hall’s Bare-throated Group taxa. 260 1 2 3 Bare-throated Group (concl.) 1. 2. 3. 4. P. leucoscepus leucoscepus P. l. infuscatus P. rufopictus P. swainsonii 4 4 4 Figure 6.2. Distribution ranges Hall’s Bare-throated Group taxa. 261 2 3 2 4 5 1 6 7 Montane Group 1. 2. 3. 4. 5. 6. 7. 8. Pternistis camerunensis P. erckelii P. ochropectus P. castaneicollis castaneicollis P. c. atrifrons P. jacksoni P. nobilis P. swierstrai 8 Figure 6.3. Distribution ranges of Hall’s Montane Group of spurfowls. 262 2 1 3 1 2 2 2 Scaly Group 1. 2. 3. 4. Pternistis ahantensis P. squamatus squamatus P. s. schuetti P. griseostriatus 2 4 Figure 6.4. Distribution ranges of Hall’s Scaly Group taxa. 263 4 4 3 1 3 5 2 6 3 6 Vermiculated Group 1. 2. 3. 4. 5. 6. P. bicalcaratus bicalcaratus P. b. adamauae P. clappertoni clappertoni P. c. sharpii P. harwoodi P. icterorhynchus Figure 6.5. Distributional ranges of Hall’s Vermiculated Group. 264 1 1 Vermiculated Group (cont.) 1. 2. 3. 4. 5. P. hildebrandti hildebrandti P. h. fischeri P. natalensis P. adspersus P. capensis 2 3 4 2 4 3 4 5 Figure 6.6. Distributional ranges of Hall’s (1963) Vermiculated Group. 265 Vermiculated Group (concl.) 1. Pternistis hartlaubi 1 Figure 6.7. Distributional ranges of Hall’s (1963) Vermiculated Group. 266 Coturnix coturnix Alectoris chukar Pternistis hartlaubi Pternistis camerunensis 93 Pternistis nobilis Pternistis erckelii 95 M T t Pternistis ochropectus Pternistis castaneicollis castaneicollis 95 Pternistis castaneicollis atrifrons Pternistis jacksoni Pternistis squamatus squamatus Pternistis squamatus schuettii Pternistis griseostriatus S C t 90 Pternistis hildebrandti hildebrandti 73 Pternistis hildebrandti fischeri 98 S V t Pternistis natalensis Pternistis adspersus Pternistis capensis Pternistis ahantensis 83 73 100 Pternistis bicalcaratus bicalcaratus Pternistis bicalcaratus ayesha Pternistis bicalcaratus adamauae N V t Pternistis icterorhynchus Pternistis clappertoni clappertoni 100 Pternistis clappertoni sharpii 90 Pternistis harwoodi 100 Pternistis leucoscepus leucoscepus Pternistis leucoscepus infuscatus Pternistis swainsonii Pternistis rufopictus 89 71 93 84 B T G Pternistis afer afer Pternistis afer humboldtii Pternistis cranchii Pternistis swierstrai Figure 6.8. A parsimony tree (1 of 2 most parsimonious trees similar to Strict Concensus tree) obtained from combined mitochondrial Cytochrome-b and organismal characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). MTt stands for Montane taxa SCt – Scaly taxa, SVt – Southern Vermiculated taxa, NVt – Northern Vermiculated taxa and BTG – Bare-throated Group. 267 Coturnix coturnix Alectoris chukar Pternistis hartlaubi Pternistis camerunensis Pternistis squamatus squamatus 70 Pternistis squamatus schuettii Pternistis nobilis Pternistis afer humboldtii Pternistis afer afer Pternistis leucoscepus leucoscepus Pternistis leucoscepus infuscatus B T G Pternistis swainsonii Pternistis cranchii Pternistis rufopictus Pternistis bicalcaratus bicalcaratus Pternistis bicalcaratus ayesha Pternistis bicalcaratus adamauae Pternistis icterorhynchus 98 92 Pternistis hildebrandti hildebrandti Pternistis hildebrandti fischeri Pternistis clappertoni clappertoni N V t Pternistis harwoodi Pternistis clappertoni sharpii Pternistis ahantensis 65 Pternistis erckelii Pternistis ochropectus Pternistis castaneicollis castaneicollis Pternistis castaneicollis atrifrons M T t Pternistis jacksoni Pternistis swierstrai Pternistis capensis Pternistis natalensis Pternistis adspersus S V t Pternistis griseostriatus Figure 6.9. A parsimony tree (1 of 397 most parsimonious trees) obtained from organismal characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). MTt stands for Montane taxa, SCt – Scaly taxa, SVt – Southern Vermiculated taxa, NVt – Northern Vermiculated taxa and BTG – Bare-throated Group. 268 Coturnix coturnix Alectoris chukar Pternistis adspersus Pternistis capensis 88 Pternistis hildebrandti hildebrandti 98 Pternistis hildebrandti fischeri S V t Pternistis natalensis Pternistis afer afer 76 Pternistis afer humboldtii 70 Pternistis rufopictus 87 B T G Pternistis cranchii 100 89 Pternistis leucoscepus leucoscepus Pternistis leucoscepus infuscatus Pternistis swainsonii Pternistis clappertoni clappertoni 100 Pternistis clappertoni sharpii 99 Pternistis harwoodi Pternistis jacksoni Pternistis swierstrai Pternistis squamatus squamatus S C t Pternistis griseostriatus Pternistis squamatus schuettii Pernistis bicalcaratus bicalcaratus 80 N V t Pternistis bicalcaratus ayesha 100 Pternistis bicalcaratus adamauae Pternistis icterorhynchus Pternistis castaneicollis castaneicollis 91 92 94 Pternistis erckelii Pternistis ochropectus 79 89 Pternistis camerunensis M T t Pternistis nobilis Pternistis hartlaubi Figure 6.10. A parsimony tree (1 of 2 most parsimonious trees similar to Strict Consensus tree) obtained from combined mitochondrial Cytochrome-b and organismal characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). MTt stands for Montane taxa, SCt – Scaly taxa, SVt – Southern Vermiculated taxa, NVt – Northern Vermiculated taxa and BTG – Bare-throated Group. 269 Pternistis swainsonii Pternistis cranchii Pternistis afer humboldtii B T 89 Pternistis afer afer G Pternistis rufopictus Pternistis leucoscepus infuscatus 100 Pternistis leucoscepus leucoscepus Pternistis swierstrai Pternistis nobilis 100 Pternistis camerunensis Pternistis hartlaubi Pternistis castaneicollis castaneicollis 88 M T t Pternistis ochropectus 90 Pternistis erckelii Pternistis jacksoni 99 98 Pternistis hildebrandti fischeri Pternistis natalensis Pternistis hildebrandti hildebrandti 95 S V t Pternistis adspersus Pternistis capensis Pternistis squamatus schuettii 79 Pternistis griseostriatus Pternistis squamatus squamatus S C t Pternistis bicalcaratus ayesha Pternistis bicalcaratus adamauae 99 Pternistis icterorhynchus Pernistis bicalcaratus bicalcaratus 91 Pternistis clappertoni clappertoni N V t Pternistis harwoodi Pternistis clappertoni sharpii Alectoris chukar Coturnix coturnix Figure 6.11. Maximum likelihood tree obtained from mitochondrial Cytochrome-b characters. Numbers above branches represent boostrap support values (only ≥ 70% are presented). MTt stands for Montane taxa, SCt – Scaly taxa, SVt – Southern Vermiculated taxa, NVt – Northern Vermiculated taxa and BTG – Bare-throated Group. 270 Appendix 6.1. GenBank accession numbers for taxa sequenced for mitochondrial Cytochrome-b. CYTB GenBank no. Taxon Pternistis hartlaubi hartlaubi FR691618 Pternistis hartlaubi crypticus FR691619 Pternistis adspersus FR691623 Pternistis afer Angola FR694158 Pternistis afer benguellensis FR694159 Pternistis afer böhmi FR694162 Pternistis cranchii cranchii FR694164 Pternistis afer cunenensis FR694160 Pternistis afer harterti FR694161 Pternistis afer intercedens FR694165 Pternistis afer itigi FR694166 Pternistis afer nudicollis FR694163 Pternistis afer nyanzae FR694167 Pternistis afer South Africa AM236908 Pternistis afer swynnertoni FR694168 Pternistis bicalcaratus bicalcaratus FR691624 Pternistis bicalcaratus adamauae FR691626 Pternistis bicalcaratus ayesha FR691625 Pternistis bicalcaratus thornei FR691627 Pternistis camerunensis FR691591 Pternistis capensis AM236909 Pternistis castaneicollis castaneicollis AM236903 Pternistis castaneicollis bottegi FR691629 Pternistis castaneicollis ogoensis FR691628 Pternistis clappertoniclappertoni FR691602 Pternistis clappertoni nigrosquamatus FR691604 Pternistis clappertoni sharpii FR691603 Pternistis erckelii FR691589 Pternistis griseostriatus AM236905 Pternistis harwoodi FR691600 Pternistis hildebrandti hildebrandti FR691595 Pternistis hildebrandti altumi FR691597 Pternistis hildebrandti fischeri FR691598 Pternistis hildebrandti helleri FR691599 Pternistis hildebrandti johnstoni FR691596 Pternistis icterorhynchus FR691601 Pternistis jacksoni FR691594 Pternistis leucoscepus AM236906 Pternistis leucoscepus infuscatus FR691587 Pternistis leucoscepus muhamed-ben-abdullah FR691586 271 CYTB GenBank no. Taxon Pternistis natalensis AM236911 Pternistis nobilis FR691592 Pternistis ochropectus FR691590 Pternistis rufopictus FR691588 Pternistis squamatus squamatus AM236904 Pternistis squamatus maranensis FR691630 Pternistis squamatus schuettii FR691631 Pternistis swainsonii swainsonii AM236907 Pternistis swainsonii chobiensis FR694170 Pternistis swainsonii lundazi FR694169 Pternistis swierstrai FR691593 272 Appendix 6.2. A revised classification of spurfowls which was based on multiple lines of evidence presented in this chapter. The status of taxa with an asterisk (*) might change when DNA sequence data (or with increased number of nucleotide bases) are included in the analyses. Species and subspecies authority appear in Table 1.2. Family: Phasianidae Sub-family: Phasianinae Genus: Pternistis Wagler, 1832 Pternistis hartlaubi Pternistis camerunensis Pternistis erckelii Pternistis ochropectus Pternistis castaneicollis castaneicollis Pternistis castaneicollis atrifrons* Pternistis jacksoni Pternistis nobilis Pternistis swierstrai Pternistis bicalcaratus bicalcaratus Pternistis bicalcaratus adamauae Pternistis clappertoni clappertoni Pternistis clappertoni sharpii Pternistis harwoodi Pternistis icterorhynchus Pternistis hildebrandti hildebrandti Pternistis hildebrandti fischeri Pternistis natalensis Pternistis adspersus Pternistis capensis Pternistis ahantensis ahantensis* Pternistis squamatus squamatus Pternistis squamatus schuetti Pternistis griseostriatus Pternistis cranchii Pternistis afer afer Pternistis afer humboldtii Pternistis leucoscepus leucoscepus Pternistis leucoscepus infuscatus Pternistis rufopictus Pternistis swainsonii 273 CHAPTER 7 Historical biogeography of francolins and spurfowls (Galliformes: Phasianidae) Abstract The biogeography of francolins has not received much attention since the monograph of Hall (1963). Apart from the dispute about their monophyly, uncertainties in the taxonomic designations and their phylogenetic relationships have generated opposing hypotheses ("out of Asia" or "out of Africa") pertinent to the origin of the genus Francolinus. Francolins (sensu lato) are galliform birds that occur in Africa, and are mainly restricted to sub-Saharan and the Indian Sub-continent; inhabiting primarily tropical/sub-tropical areas with some species occurring in forested habitats. They were recently split into two distantly related assemblages, francolins and spurfowls (divergence date going back to c. 33.6 mya) which are split among five francolin genera Francolinus, Ortygornis, Afrocolinus, Peliperdix, Scleroptila as opposed to the single genus Pternistis assigned to all spurfowl taxa. This chapter seeks to understand the history of the current geographic distribution patterns of francolins and spurfowls in light of their phylogeny and to test the two opposing hypotheses on their origin. A range of organismal and DNA characters were analyzed using parsimony and Bayesian methods. A new biogeographical reconstruction method, spatial analysis of vicariance, was used to detect disjunctions and to infer barriers while the ancestral area and habitat were reconstructed in Mesquite. The colonization of Africa by ancestral species may have 274 been through dispersal from Asia, which resulted in the formation of disjunct distributions and the somewhat rapid diversification of francolins and spurfowls within Africa. The Rift Valley system, Lake Chad, Upper Guinea and the Congolian forest, major rivers such as Limpopo, Zambezi, Rovuma, Volta and Rufiji, the Okavango Swamp and Sahara Desert emerged as the major physical breaks that may have created and maintained the distributions observed today and promoted speciation among the African francolins and spurfowls. Sharp diversity gradients in habitat appear to also have played an important ecological role in facilitating diversification among these taxa. 275 Introduction General background Cracraft (1994) argued that a scientific problem that is central to systematists and ecology has been to explain the spatio-temporal patterns of species diversity and attempt to understand why taxa occur in particular areas and not others, as well as how they assembled or spread to those regions. This becomes a difficult exercise to do given that historical biogeography is at best poorly known. Some authors, e.g. de Quieroz (2005), Donoghue and Moore (2003), Lamm and Redelings (2009) argue that the challenge of historical biogeography is to distinguish dispersal and vicariance whereas (Hovenkamp 1997, 2001) argued for the recognition of barriers being an important task in a framework that can examine both inter and intra-continental distributional patterns. In this chapter the focus is on francolins and spurfowls (Table 1.1), here considered two independent radiations of taxa (Crowe et al. 2006, Chapter 2). These assemblages are francolins, divided among the genera Francolinus, Ortygornis, Afrocolinus, Peliperdix and Scleroptila (details in Chapter 5) and spurfowls, which are attributed to a single genus Pternistis (Crowe et al. 2006, Chapter 6). The closest relatives of each of these independent radiations occur in Asia and Europe: the genera Gallus, Bambusicola spp., which are the close relatives of the francolins, and Coturnix, Margaroperdix, Alectoris, Perdicula, Tetraogallus, Excalfactoria and Ammoperdix spp., which are closely related to the spurfowls. Habitat preferences of francolins and spurfowls Francolins and spurfowls have complex distribution patterns (Snow 1978), occurring in different habitats primarily of a tropical/sub-tropical nature, with some 276 species occurring in forested habitats (Hall 1963, Johnsgard 1988, del Hoyo et al. 1994, Madge and McGowan 2002). The three Spotted Middle Eastern-Asian Francolinus species, F. francolinus, F. pictus and F. pintadeanus, thrive in thick bush and patches of scrub jungle (Johnsgard 1988). The Ortygornis taxa, O. sephaena and O. grantii occur in scrubby African woodlands with grass habitats. The Asian francolin O. gularis is a swamp-dwelling species, whereas Asian O. pondicerianus thrives in woodland and bushes. The Red-tailed Peliperdix spp. inhabit open grassland and woodland savannas. Afrocolinus lathami remains the only African francolin to be restricted to deeply forested habitats (del Hoyo et al. 1994, Madge and McGowan 2002). Although the Red-winged Scleroptila spp. occur in varied habitats at different latitudes, they primarily occupy grassland (Johnsgard 1988, Madge and McGowan 2002), which could be open lowland grasslands in which S. levaillantoides, S. gutturalis, S. finschi thrive, or the open hilly (S. shelleyi) and highland grasslands preferred by S. afra, S. psilolaema, S. levaillantii, and S. streptophora. With regard to the spurfowls, the Bare-throated Group includes four relatively large species, P. swainsonii which inhabits thickets and dry savannas, P. afer which thrives in thickets, riparian scrub and adjoining grassland and forest edges, P. leucoscepus occupies thicket and open savanna scrub areas and adjacent fields, and P. rufopictus which occurs in thickets and in Acacia scrub and woodland. Members of the Vermiculated Group (P. hartlaubi, P. capensis, P. natalensis, P. adspersus, P. hildebrandti, P. clappertoni, P. icterorhynchus, P. bicalcaratus, P. harwoodi) generally occupy dense bushy thickets and shrubby grasslands (del Hoyo et al. 1994, Madge and McGowan 2002). In contrast to the habitat preferences of the Bare-throated and the Vermiculated spp., members of the Montane Group (P. camerunensis, P. swierstrai, P. 277 nobilis, P. ochropectus, P. jacksoni, P. castaneicollis) inhabit largely montane forest patches and scrubby slopes (Madge and McGowan 2002) with the exception of P. erckelii being the only member that ventures into adjacent grasslands. The Scaly Group spp., P. squamatus, P. griseostriatus and P. ahantensis inhabit forest edges with dense undergrowth (Madge and McGowan 2002). The origin of the genus Francolinus Stephens, 1819 The relationships between francolins with Palaearctic and Asian genera influenced Hall’s (1963) hypothesis on the origin of the genus Francolinus. Even though Hall (1963) produced commendable work in describing the geographical distribution patterns of francolins, the origin of the genus ‘Francolinus’; however, continued to be highly disputed (Hall 1963, Crowe et al. 1992). Based on the fact that francolins and spurfowls share the closest relationships with Palearctic and Asian genera, Hall (1963) strongly argued for this genus to be of Asian origin and she speculated that its age extended back to the Oligocene 25-35 mya (Hall 1963, Sibley and Alquist 1985). She further maintained that speciation of francolins in Africa was likely promoted by other factors such as reduced competition. In contrast, Crowe and Crowe (1985) and Bloomer and Crowe (1998) argued that the ancestor of francolins was African, inferring that francolins dispersed from Africa to Asia. Despite the difference in opinion about the origin of this lineage, Hall (1963), Crowe and Crowe (1985) and Bloomer and Crowe (1998), all hypothesized that the ancestral francolin (traditionally including spurfowls) was a small, quail-like phasianid based on the observation that the plumage and other integumentary features of immature francolins resemble those of quails (Crowe et al. 1986). 278 A brief overview of historical biogeographical approaches One way to investigate the biogeography of taxa is by tracing ancestral patterns i.e., the history of character evolution and also by reconstructing the evolution of geographic ranges of taxa on phylogenetic trees. In addressing the evolution of the distribution patterns of francolins (sensu lato), Hall (1963) made a direct comparison of the ranges of species with vegetation types and “tentatively tried to correlate some of the climatic eras postulated with those known” as outlined in the monograph in Appendix 1, on page 173-174. Her attempts to explain the evolution of species involved two assumptions. The first one being that the degree of divergence shown by two isolated populations can be correlated with the length of time spent in isolation and secondly, that the species assemblage present originated in some part or parts of their current range. Until recently a commonly utilized approach to investigate historical biogeography was that of dispersal-vicariance analysis (Ronquist 1997), which is implemented in the software DIVA (Ronquist 1996). This approach requires one fullyresolved phylogenetic tree and minimizes the number of dispersal and extinction events required to explain the species’ distributions by way of optimizing the ancestral areas onto internal nodes of a phylogeny. Jønsson et al. (2010) used what they call a newly developed Bayesian approach to dispersal-vicariance analysis, ‘Bayes-diva’ (Nylander et al. 2008), which unlike Ronquist’s DIVA approach, accounts for phylogenetic uncertainty. Ree and Smith (2008) developed a dispersal-extinction-cladogenesis (DEC) model that reconstructs the evolution of geographic ranges. The DEC model has been criticized for "ignoring vicariance” and assuming that the rate of evolution of geographic ranges occurs independently, an assumption that may not be applicable to some biogeographic inferences. 279 Spatial analysis of vicariance by Arias et al. (2011) is a new method which is based on the ideas of Hovenkamp (1997, 2001) as implemented in VIP (Arias et al. 2011), which instead of looking at pre-defined areas, uses direct geographic/georeferenced data to detect disjunctions between sister groups and infers barriers associated with them. It uses an optimality criterion framework which maximizes the number of possible pairs of disjunct sister nodes while minimizing the number of eliminated distributions (Arias et al. 2011). This method requires a phylogeny. In this chapter, the aim was to understand the history of the current geographic distributional patterns of francolins and spurfowls in light of the species groups suggested by Hall (1963), with modifications as outlined in Chapters 5 and 6, in order to detect disjunctions and infer barriers, as well as to resolve the contrasting hypotheses on the origin of the genus Francolinus sensu lato outlined by Hall (1963; our of Asia) and Crowe and Crowe (1985; our of Africa), respectively. Materials and methods Data collection Sampling of taxa and characters Only francolin and spurfowl taxa that were recognized as valid terminal taxa in Chapter 5 (francolins) and Chapter 6 (spurfowls) were analyzed in this chapter. Maps and mapping of distribution records of investigated taxa The maps depicting the distribution ranges of francolins and spurfowls were produced as outlined in Chapter 5 and 6. 280 Ancestral state reconstructions The multistate matrices for francolins and spurfowls were generated for two traits: area and habitat. The francolin area states were scored as follows: Southern Asia (0), West Africa (1), Southern Africa (2), Central Africa (3) and East Africa (4), whereas those for habitat were: open grasslands (0), scrubby grasslands (1), wooded grasslands (2), Montane grasslands (3) and Forest (4). The states for area and habitat of the spurfowls were coded as Southern Africa (0), West Africa (1), North Africa (2), Central Africa (3) and East Africa (4) and those for habitat were: savanna and arid woodlands (0), mesic woodlands (1), forest/forest edge (FFE) (2), lowland fynbos (3) and macchia (a scrubland vegetation of the Mediterranean region which is composed primarily of leathery, broad-leaved evergreen shrubs). Spatial analysis of vicariance In the absence of direct georeferenced distribution data, the distribution ranges of taxa were used to detect disjunctions and infer barriers between sister groups. Data analyses Chapter 2 demonstrated that instead of francolins being considered a single evolutionary lineage, they represent two lineages and as a result francolin and spurfowl analyses were conducted separately Ancestral state reconstructions To reconstruct the ancestral patterns, a maximum likelihood approach (based on a Bayesian topology) which seeks to find the ancestral states that maximize the 281 probability that the observed states would evolve under a stochastic model of evolution (Schluter et al. 1997, Pagel 1999) was used. What differentiates this approach and makes it a better method in reconstructing ancestral states over the parsimony approach is that it incorporates both topology and branch-lengths. Another advantage of a Bayesian reconstruction is that the marginal probabilities are assigned to nodes and thus the level of confidence can be gained for any assignment. Ancestral area and habitat reconstructions were implemented in Mesquite (Maddison and Maddison 2006) with the idea of tracing the evolution of area and habitat among francolins and spurfowls. Bayesian and parsimony phylogenetic trees were appended to the input files and the analyses were performed under the mk1 model (Lewis 2001). Spatial analysis of vicariance Distributional ranges of taxa were analysed and an overlap of up to 25% was accepted as a disjunct distribution. The cost of overlapped (i.e. non-disjunct) distributions was set to the default of 1.00 and the cost for removal of distribution was 2.00, which means that a removal was only accepted if it was found at least in one additional pair of disjunct sister groups. To detect barriers, a simple hill climbing heuristic search was performed starting with a sector of 20 nodes over 1000 iterations during which a “flipping nodes” strategy was used. This procedure is quite effective when the distributions are close together, but it becomes less accurate for distributions placed far away, as it is placed in the line that is equidistant to each point. 282 Phylogenetic analyses Bayesian and parsimony methods were used in analysing combined mitochondrial and nuclear DNA data partitions from Chapter 2, 4 to 6, and the combined organismal (Chapter 4, 5 and 6) and DNA data partitions, respectively. The francolin and spurfowl trees were each rooted with the most basal taxon among members of these assemblages, that is, F. francolinus for the francolins and P. hartlaubi for the spurfowls. Combined organismal and DNA parsimony analyses consisting of 5141 characters of francolins and 5203 characters of spurfowls were performed and these characters were used to build the strict consensus trees (the boldest hypotheses of relationship) implemented in TNT (Tree analysis using New Technology - Goloboff et al. 2008). All characters were equally weighted and treated as non-additive. It must be noted that, in order to reduce the denseness of taxa on the tree, the phylogenetic trees presented only included taxa that are recognized in Chapter 5 and 6. The settings for the analyses followed the option of a full heuristic search with the starting tree(s) being obtained via stepwise addition and random addition of sequences (Maddison 1991) over 1000 replicates. One tree was held at each step during stepwise addition. The tree-bisection-reconnection option was implemented and Farris et al.’s (1996) jackknife measure was used to assess branch support over 1000 pseudoreplicates. Bayesian analyses of the combined mitochondrial and nuclear DNA characters for francolins (5116 bp) and spurfowls (5170 bp) were implemented in MrBayes 3.1.2 (Ronquist and Heulsenbeck 2003). The francolin and spurfowl Bayesian analyses excluded one taxon each, O. r. spilogaster and P. castaneicollis atrifrons, respectively due to unavailability of DNA characters. Over and above parsimony analyses, the use of a Bayesian approach was to investigate the influence of mixed models (parameter-rich) 283 on the phylogeny. Modeltest 3.7 (Posada and Crandall 1998) was used to search for the best-fit model for each molecular marker and each codon data partition for CYTB and ND2. The parameters for the model considered were those under the Akaike Information Criterion (AIC). The best-fit model of nucleotide substitution suggested for all the data partitions was the GTR+I+G model. These analyses involved two independent runs (with random starting trees) and they proceeded without a molecular clock being imposed upon the rate of sequence evolution. The Markov Chain Monte Carlo (MCMC) for the spurfowl and francolin data set was run for 3 x 106 and 1.5 x 106 generations, respectively, sampling every 100th generation using four chains (one cold and three heated). Tracer v 1.5 (Rambaut and Drummond 2007) was used to check the ‘burn-in’ phase and the state of convergence of the two parallel runs: 300 000 and 150 000 trees from the burn-in phase for the spurfowls and francolins, respectively, were discarded and convergence was attained when the average deviation of split frequencies converged towards zero. Finally, the 50% majority rule consensus of sampled trees (excluding the burn-in) was obtained with posterior probabilities (PPs) appearing above nodes. Results Ancestral state reconstructions Even though the parsimony ancestral area and habitat reconstructions of francolins and spurfowls were performed, the results were used for comparison with those for the Bayesian reconstructions irrespective of the differences inherent in the methodology of these two approaches. The results discussed in subsequent sections are based solely on Bayesian reconstructions. 284 Francolins – area and habitat Figure 7.1 and 7.2 respectively show the Bayesian reconstructions through which the ancestral area and habitat among Asian and African francolins were traced. The marginal probabilities are assigned to each node and these are estimated based on the regional and habitat data that were incorporated into the input file and they indicate the confidence that each state can be assigned at a particular node. Figure 7.1 showed mixed results, e.g. some nodes such as 1, 7, 9, 10, 11, 12 have probabilities above 0.90; node 2, 6, 14 were assigned with probabilities between 0.80 and 0.90 and the rest of the nodes showed great uncertainties. The root, node 1, points to southern Asia being the ancestral area that gave rise to all other francolins. Node 3 links the Asian francolins and the rest of Africa, but with very low probabilities (0.52 and 0.36) being assigned to the rest of Africa and Asia, respectively. The immediate link of Asia and Africa is at node 4, which links the Asian O. gularis/O. pondicerianus and the African Ortygornis clade (Fig. 7.1). The probabilities are very low with the proportions 0.48 and 0.40 pointing to southern Africa and southern Asia respectively as being the ancestral areas. These species possibly evolved in scrubby grassland habitats (probabilities of 0.99). Afrocolinus lathami is the only West African (and forest endemic) species supported with a probability of 0.99. This species probably made its way to the forest habitat from a scrubby grassland ancestor. The Red-tailed and the Red-winged clade (node 6) are truly southern African clades supported with 0.88 probability of which the probability assigned to their ancestral habitat (Fig. 7.2) is uncertain. Within the Redtailed Group clade (node 7), members share a most recent common ancestor in southern African (0.99 probability) that likely thrived in wooded grasslands from which these clades diversified in other parts of Africa i.e., two species spread to West Africa (P. 285 albogularis and P. spinetorum), one to Central (P. schlegelii) and two to East Africa (P. maharao and P. hubbardi). When these species diversified in different parts of Africa they retained the ancestral preference of wooded habitats with a slight difference being that P. albogularis occurred in mesic woodlands, whereas the other three species occurred in arid woodlands. Seemingly, the most recent common ancestral area of the Red-winged taxa (node 10) was in southern Africa (0.91 probability), although this clade underwent its greatest diversification in East Africa. Most species inhabit open grasslands, three species, with two taxa (S. psilolaema psilolaema and S. ellenbecki) that went into montane grasslands in East Africa and one species S. afra made its way to montane grasslands in southern Africa. Scleroptila streptophora spread to the scrubby mesic habitat in Cameroon and around Lake Victoria that differs from that of its former group member O. sephaena that mainly occupies arid scrubby grasslands and woodlands. Thus, the Red-winged Group diversified into mesic, arid, and montane grasslands. The Striated members spread from the south to East Africa and the probability assigned to this node is 0.80. O. grantii and O. r. spilogaster occupy arid scrubby habitat. The nominate O. r. rovuma occupies mesic scrub habitat. It seems that, in Africa, francolins diversified out of southern Africa, spreading mainly to the East, then West and P. schlegelii (inhabitant of savanna and arid woodlands) evolved in central Africa where it remained isolated. The Red-tailed diverged to occupy both mesic and arid woodland habitats. 286 Spurfowls – area and habitat The likelihood reconstruction for area (Fig. 7.3), unlike that for habitat (Fig. 7.4) exhibited ambiguity at several nodes with respect to the ancestral distribution of spurfowl. Although with a low probability of assignment the root node suggests southern Africa as the potential ancestral area. Habitat reconstruction (Fig. 7.4) showed some support (0.64) for tracing all spurfowls back to forest (node 1). The Bare-throated Group/P. clappertoni/P. leucoscepus/P. swainsonii (node 7 which combines node 8 and 9) is assigned with a low probability values with east Africa (0.47) being favoured to be an ancestral area over southern Africa (0.40). With regard to habitat, taxa that thrive in mesic woodlands (P. rufopictus, P. cranchii, P. a. afer, P. a. humboldtii and P. l. leucoscepus, P. l. infuscatus, P. swainsonii) are reconstructed as occupying savanna and arid woodlands with high probability (0.95; Fig. 9.12). Node 4 joins the Montane Group (forest inhabitants) to P. clappertoni (savanna and arid woodland), P. a. afer (mesic woodlands), P. c. sharpii, P. harwoodi and the Barethroated Group, and is reconstructed to likely have had a forest ancestor with a probability of 0.70. The most recent common ancestor of P. b. bicalcaratus, P. b. adamauae, P. icterorhynchus, ‘ayesha’ has a savanna and arid woodland ancestor (0.87 probability). In Fig. 7.4, node 10 connects the Scaly Group and the southern Vermiculated clade has the highest probability of 0.72 and this is in support of a forest ancestor, while the most recent common ancestor of the Scaly Group is supported with 0.99 probability to have lived in forest in southern Africa (0.68). The probability of 0.89 supports the southern African origin of the Vermiculated species P. adspersus, P. capensis, P. hildebrandti, and P. natalensis. Pternistis hildebrandti has its distribution in southern 287 and eastern Africa. Most of the nodes on the ancestral area reconstruction showed great uncertainty in the assignment of marginal probabilities. It is, however, clear that colonization of areas happened more than once with the exception of North Africa where P. b. ayesha has an isolated distribution in Morocco in the macchia type of habitat. Areas too were colonized more than once except for the lowland fynbos, a home for P. capensis. Spatial analysis of vicariance The spatial analysis of vicariance of francolins yielded two reconstructions from which a consensus reconstruction analysis was performed yielding the best score associated with a cost of 9.0, and 320 hits of 1000 iterations. Thirty optimum disjunct sister groups were recovered and these disjunctions are mapped at nodes on the parsimony tree (Fig. 7.5). The spurfowl distribution ranges resulted in one reconstruction with 28 optimum disjunctions (Fig. 7.6) accompanied by a best score with a cost of 6.0 for 471 hits of 1000 iterations. The history of evolution of distribution ranges Francolins - The phylogenetic results confirmed that the Asian francolins are the most basal of all francolins and the main geographic barriers exist around the Indian subcontinent (Fig. 7.7a-b). The first break being near the Windhya Range of Mountains (Fig. 7.7a), that are located in Rajasthan, south of Delhi, between F. francolinus and F. pictus (node 1, Fig. 7.5). According to Hall (1963), Rajasthan has no clear natural barrier. Thus, it remains uncertain whether the Windhya Range and possibly the Ganges tributaries should be considered effective physical barriers. The Ganges River in 288 northeast India separates Francolinus pictus from F. pintadeanus (Fig. 7.7b, node 2 in Fig. 7.5) as well as the Asian/African francolin clade represented by all taxa restricted to the sub-Saharan region. The Naga Hills in Burma in addition to the Chin Hills and the Manipur Range as alluded to by Hall (1963) act as the possible barriers between F. pintadeanus and the Asian+African francolin clade (Fig. 7.7c, node 3 in Fig. 7.5). The disjunction (node 4 in Fig. 7.5) between the clade that includes members of Ortygornis and the African clade circumscribed by A. lathami/Peliperdix/Scleroptila is associated with a barrier that runs along the Rift Valley system down to the south and in the west meets the Congolian forest, setting itself as the main barrier in the south separating the African Ortygornis taxa on the east, from the rest of the African francolins in the west (Fig. 7.7d). Another barrier is set by the Limpopo River in the south separating the African Ortygornis clade from specifically S. levaillantii and S. afra in the area where they overlap (Fig. 7.7d). The barrier between the O. gularis/O. pondicerianus clade and the O. sephaena/grantii/rovuma clade (node 5 in Fig. 7.5) is quite ambiguous, because in this case the barrier could be any intervening feature, possibilities include the: Arabian Sea, and the large rivers and mountain ranges in Pakistan and Iran (Persia) (Fig. 7.7e). The disjunction that exists within the clade that includes O. pondicerianus + O. gularis and O. sephaena/grantii/rovuma clade might be an indication of colonization of Africa by the clade ancestor from Asia or a return of the clade ancestor to Asia. In colonizing Africa, the clade ancestor landed itself on the east side of the Rift Valley system thereby inhabiting arid habitats. The O. pondicerianus/O. gularis/O. sephaena clade diverged c. 7.2 mya (Chapter 2). The estimated age for the formation of the Rift Valley is c. 6 mya. It could be speculated that the clade ancestor possibly crossed the Rift system 289 diversifying in central and west Africa occurring along the mesic Congolian and Upper Guinea forests. Within the O. sephaena/grantii/rovuma clade, the Zambezi River acts as the major barrier that separates O. sephaena (in the south) and the O. grantii clade in the north (Fig. 7.7f). The Rufiji River separates the southern African O. r. rovuma from O. grantii (Fig. 7.7h) whereas the northeastern O. r. spilogaster is separated from O. grantii by the Ganane Basin and possibly the Ethiopian Rift (Fig. 7.7g). In addition, the difference in habitat, xeric versus mesic scrub, seems to have an influence within the O. sephaena/grantii/rovuma clade. The xeric habitats differ in the amount of mean annual rainfall received, ~400-800 mm where O. sephaena and O. grantii occurs, ~200-400 mm where O. r. spilogaster occurs (Clark 1967). The mesic scrubs receive the mean annual rainfall of ~800-1600 mm and this is occupied by O. r. rovuma. The mesic versus xeric habitat preference seems to set itself as the possible barrier between the southern O. r. rovuma and its northern counterpart O. r. spilogaster (Fig. 7.7h). There is a disjunction (node 9 in Fig. 7.5) between the A. lathami/Peliperdix and Scleroptila clades (Fig. 7.7i). The Congolian and the Upper Guinea forest emerged to be effective breaks between the two clades. While there is no detected barrier to separate A. lathami and the Peliperdix clade, the west African forest population A. l. lathami is set apart from the central A. l. schubotzi by the Congo Basin (Fig. 7.7j). Both taxa occur in high rainfall areas of above 3200-4000 mm annually (Clark 1967). Due to the presence of a disjunction in just one of the reconstructions, there was no barrier detected between node 11 and node 14 in Figure 7.5 which links the P. schlegelii/P. albogularis and P. coqui clade. Therefore, there are no definitive barriers to infer between P. schlegelii and P. albogularis (Fig. 7.7k, node 11 in Fig. 7.5), 290 between P. a. dewittei and P. albogularis albogularis/P. a. buckleyi clade (Fig. 7.7l), node 12 in Fig. 7.5) and also between P. a. albogularis and P. a. buckleyi (Fig. 7.7m). Both clades occur in habitats that receive similar amounts of annual rainfall (800-1600 mm). The occurrence of P. a. dewittei on montane wooded grasslands south of the Congo forest (Cotterill 2006) could be attributed to the fact that during the time when Africa was wet the forest may have formed a continuous belt from west Africa through to central and down to south central Africa (Clark 1967, Crowe and Crowe 1982). During this time the proto-albogularis may have had a continuous distribution along the forest belt. During a dry period (possibly glaciation), forest contracted and formed pockets leaving P. a. dewittei being isolated south of the Congo. Morphologically, there are subtle differences between P. a. albogularis and P. a. dewittei which relate to the size and extent of barring on the underparts. The predicted mean annual rainfall in areas where P. a. albogularis and P. a. dewittei occur is 100-1400 and 1400-1800 mm, respectively. Within the clade that includes P. stuhlmanni/P. kasaicus and P. c. coqui/P. c. vernayi/P. c. ruahdae/P. hubbardi/P. maharao/P. spinetorum (node 14 in Fig. 7.5), the Rovuma River was detected as an effective barrier that exists on the east at least between P. stuhlmanni/P. hubbardi and P. maharao in the south the Limpopo River for a barrier separating P. stuhlmanni from P. c. coqui/P. c. vernayi (Fig. 7.7n). P. kasaicus is separated from the other taxa in this clade, but there is no definite barrier attributable to this disjunction. Peliperdix kasaicus and P. stuhlmanni both occur in mesic conditions which differ in areas where the mean annual rainfall is 1000-1400 mm. The mesic woodland P. c. ruahdae (800-1600 mm of mean annual rainfall) on the west is 291 separated from P. hubbardi, which occurs in much more xeric habitat to the east around Lake Victoria and Lake Tanganyika (400-800 mm of mean annual rainfall) (Fig. 7.7o). The southern African P. c. coqui and P. c. vernayi are separated by the Okavango swamp (Fig. 7.7p) with P. c. ruahdae and P. c. vernayi being separated by the mesic and arid habitat respectively (Fig. 7.7q). The area between the Ethiopian and the East African Rift might possibly be the barrier separating P. hubbardi from P. maharao in Malawi (400-800 mm of mean annual rainfall), whereas all the remaining taxa occur in arid savanna (including P. spinetorum 200-400 mm of mean annual rainfall) (Fig. 7.7r). Habitat separation definitely has an influential role among these disjunctions. The first divergence within the Scleroptila clade led to the separation of S. streptophora/S. levaillantii from the rest of the Scleroptila taxa (Fig. 7.7s, node 21 in Fig. 7.5). The Scleroptila taxa mainly occur in xeric environments. In the south the Limpopo and Zambezi Rivers act as barriers between S. levaillantii and the other Scleroptila taxa (Fig. 7.7t), whereas northern Lake Tanganyika and the Rufiji River separate the east African S. streptophora population from the other Scleroptila taxa (Fig. 7.7u). Mesic habitat links the two west and east S. streptophora populations that occur on either side of the Congo Basin. It is quite strange that not even habitat links the disjunct west and east African S. streptophora populations with those of southern African S. levaillantii. Scleroptila streptophora has a puzzling distribution being found in grasslands in Cameroon where the annual rainfall reaches ~2600-3200 mm and in east Africa (800-1600 mm) and it was included in the Striated Group together with O. sephaena by Hall (1963). From examination of museum skins of the two populations of S. streptophora, there were no differences observed. The two species S. streptophora 292 and O. sephaena are distinct morphologically and at least ecologically (Hall 1963), but the east African population shares almost the same type of habitat with that of O. sephaena. Scleroptila streptophora occurs in mesic habitat (at an altitude ranging from 1050–1200 m in Cameroon and 600-1800 m in scrub-covered hillsides in east Africa – del Hoyo et al. 1994), whereas S. levaillantii is found in xeric to mesic habitats (Fig. 7.7t) at an altitude of 600-2 250 m (Hall 1963, Snow 1978) and the mean annual rainfall of 600-800 mm. The disjunction between the S. psilolaema and S. uluensis/S. shelleyi is not associated with any detectable barrier. However, the African Great lakes (Lake Victoria, Lake Tanganyika, Lake Malawi) form a barrier that separate S. psilolaema from S. finschi (Fig. 7.7v), with S. finschi being separated from S. whytei/S. jugularis (Fig. 7.7w, node 25 in Fig. 7.5). The African Great lakes appear to be effective barriers between S. uluesis and S. shelleyi (Fig. 7.7x), whereas S. gutturalis and S. uluensis are separated by the Ethiopian Rift, with but taxa occurring in xeric grassland habitats (Fig. 7.7y). Among members of the S. shelleyi/S. levaillantoides and the S. ellenbecki clade, the distribution ranges of S. shelleyi and S. levaillantoides are maintained by two breaks, Lake Malawi in the north and the Limpopo River in the south (Fig. 7.7z). Scleroptila shelleyi and S. levaillantoides overlap in the southern part of their ranges and there was no detectable barrier. The mean annual rainfall is the area according to Clark (1967) is 600-800 mm, i.e. in areas where S. levaillantoides, S. shelleyi, S. levaillantii occur and 400-600 mm in areas where S. afra occurs. The impact of the African Great Lakes is also observed between S. c. crawshayi and S. c. kikuyuensis (Fig. 7.7aa, node 30 in Fig. 7.5). The ranges of the taxa that belong to S. ellenbecki have no barriers detected among them and they are completely disjunct. The Zambezi and 293 Limpopo valleys appear to be effective barriers among some of the species and this confirms the findings in Benson et al. (1962). Spurfowls - the results in Figure 7.8a reveal that the most basal disjunct node (1 in Fig. 7.6) is associated with a barrier that separates the clade comprising Ammoperdix heyi from the rest of the African spurfowls. The Red Sea emerges to be the main barrier with some overlap along the Bab al Mandab Mountains that lie between the Red Sea and the Gulf of Aden. The barrier may have resulted from the opening of the Red Sea. The barrier between A. heyi and P. asiatica (Fig 9.8b, node 2 in Fig. 7.5) could be anything between Saudi Arabia and India with the possible barriers being the Arabian Sea, and the rivers and mountain ranges in Pakistan and Iran. It is possible that dispersal route may have started in Asia, into Saudi Arabia and subsequently into Africa or vice versa. It is somewhat difficult to explain clearly how Africa was colonized by the ancestral-spurfowl, but to some extent it could have been that the ancestral-spurfowl dispersed from Asia along the desert of Saudi Arabian coast and settled in the desert in the south western part of Africa i.e., in Namibia and south western Angola where P. hartlaubi occurs (Fig. 7.8c). The ancestral-hartlaubi diversified to other parts of Africa from the desert habitat. Whereas P. hartlaubi is said to occur in savanna and arid woodlands where the mean annual rainfall is below 100 mm (Clark 1967), it has been observed in boulder-strewn slopes and rocky outcrops (Komen 1987, Sinclair and Ryan 2005). The subsequent Africa spurfowl radiation involved the evolution of the montane lineage that includes species that were confined to the mountain forests. The first break, the African Great Lakes is detected between the clades comprising P. nobilis and P. swierstrai (node 4 in Fig. 7.6) with habitat acting as 294 another barrier (Fig. 7.8d). Pternistis nobilis and P. camerunensis inhabit the mesic montane forest occurring on the extreme side of the Congo Basin, whereas the other taxa that they are being compared with in this context are found in arid savannas and they are both separated by the Congo forest (Fig. 7.8e). Pternistis nobilis occurs in an area where the mean annual rainfall is ~ 800-1600 mm, whereas P. camerunensis is endemic to an area which receives an average of more than 3200 mm of rain annually. The disjunction exists between the Angolan highlands endemic species P. swierstrai and the north eastern P. c. castaneicollis clade, and this separation is mainly promoted by habitat differences, mesic versus xeric environment, even though they all occur in forests (Fig. 7.8f). The separation seems large such that vegetation alone may not explain the present distribution. The main barrier between the P. c. castaneicollis clade and the other spurfowls (especially P. jacksoni as other ranges are removed) is the area that forms part of the Rift Valley system between the Ethiopian and the East African Rift as they are generally found in similar habitat (Fig. 7.8g, node 7 in Fig. 7.6). There is a slight overlap between the clades comprising P. c. castaneicollis and P. ochropectus along the Ethiopian Rift (Fig. 7.8h). The distribution ranges of P. c. castaneicollis and P. c. atrifrons are possibly maintained by part of the Ethiopian Rift (Fig. 7.8i). Pternistis erckelii and P. ochropectus prefer similar habitat, they are montane forest species, with the exception of P. erckelii, which sometimes wanders into montane grasslands. Their distribution ranges are also maintained by part of the Ethiopian Rift (Fig. 7.8j). The disjunction that exists between P. ahantensis and P. icterorhynchus cannot be explained geographically because the sister groups were removed (Fig. 7.6). Figure 7.8k shows the detected barrier between P. ahantensis and P. griseostriatus clade (node 12 in 295 Fig. 7.6) which cuts through east of Lake Victoria, along Lake Tanganyika (where they slightly overlap) separating the two clades into mesic versus somewhat xeric habitat taxa with the Congolian and Upper Guinea forest being effective barrier between the northern and the southern taxa. The Niger River strongly separates P. ahantensis which occurs in the far west in the Upper Guinea forest from P. squamatus (Fig. 7.8l) which is mainly confined to the Congolian forest while P. s. squamatus is separated from P. s. schuetti by the area between the Ethiopian and the East African Rift (Fig. 7.8m). Pternistis griseostriatus is endemic to Angola and occurs in riverine forest (Sinclair and Ryan 2005). The P. griseostriatus clade also include the southern Vermiculated species P. capensis, P. natalensis, P. hildebrandti and P. adspersus that occur along the arid belt which acts as a barrier between them and P. griseostriatus (Fig. 7.8n). In Figure 7.8o, it is apparent that P. capensis and P. natalensis clades (node 16 in Fig. 7.6) are dissected by the Okavango swamps along the area of slight overlap whereas they are all found in the intervening arid habitats. Between P. capensis and P. adspersus the Orange River emerges as an effective barrier (Fig. 7.8p) whereas the Zambezi River and Lake Malawi separate P. hildebradti from P. natalensis (Fig. 7.8q). Lake Malawi also acts as a barrier between P. h. hildebrandti and P. h. fischeri along the area where they overlap slightly (Fig. 7.8r). Although some parts of the mesic and arid savannas are covered, the Upper Guinea and the Congolian forest act as a barrier between the clades that include P. icterorhynchus and P. clappertoni/P. leucoscepus along the area where they overlap (Fig. 7.8s, node 20 in Fig. 7.6) with P. icterorhynchus and P. bicalcaratus being separated from each other by the Niger River and Lake Chad (Fig. 7.8t). The Volta and Niger River likely both play a role in breaking P. b. adamauae from P. b. bicalcaratus 296 clade (Fig. 7.8t) with the isolated Moroccan P. b. ayesha being strongly separated from P. b. bicalcaratus by the Sahara Desert (Fig. 7.8t). The spurfowls occur in almost the rest of the sub-Saharan region with the exception of P. bicalcaratus that expanded its range to Morocco in north Africa where ayesha occurs (Fig. 7.8a). Spurfowl typically occur in various montane/lowland habitats as well as xeric/mesic habitats that vary in mean annual rainfall. The Ethiopian Rift may have been effective in maintaining the break between the P. clappertoni and P. leucoscepus clades (Fig. 7.8u). The eastern branch of the East African Rift and possibly the Rufiji River separate the distribution range into P. leucoscepus clade in the north and P. swainsonii clade in the south (Fig. 7.8v, node 25 in Fig. 7.6), with P. l. leucoscepus and P. l. infuscatus being separated by the Ethiopian Rift (Fig. 7.8w). No barrier was detected between the P. swainsonii and the P. rufopictus clade, whereas P. cranchii is separated from P. rufopictus by the Rift Valley system in the east and the Congolian forest in the west (Fig. 7.8x). Pternistis cranchii primarily occurs in mesic woodlands. Habitat seems to be playing a role in maintaining the break between P. rufopictus and P. a. afer (Fig. 7.8y). According to Hall (1963), they all occur along the acacia belt that is differentiated by different Acacia species. Pternistis rufopictus occupied largely the mesic part with P. a. afer and P. a. humboldtii getting into the dry and mesic parts. The Zambezi River strongly separates the two subspecies P. a. afer and P. a. humboldtii (Fig. 7.8z). 297 Discussion Reconstruction of ancestral area and habitat Francolins and spurfowls On the basis of the results, it was difficult to determine where the ancestor of francolins could have lived based on the evolution of area, whereas in terms of habitat, it could be concluded that the ancestor of francolins likely evolved in scrubby grassland. Even though the spurfowl area reconstruction had the poorest probabilities assigned to nodes, habitat reconstruction suggests that the ancestral distribution was restricted to forest. However, it was difficult to draw definitive conclusions on the basis of the area and habitat maximum likelihood reconstructions due to uncertainties in nodal probabilities assigned to states among francolins and spurfowls. The two biogeographic methods, one focusing on ‘detecting barriers’ and the other one focusing on ‘mapping of ancestral areas and habitat’ both seek to explain the history of distribution ranges of taxa, but differ in their underlying theoretical applications. Even though both applications made use recovered phylogeny, the ancestral reconstructions seem uncertain when there is poor phylogenetic resolution. In the context of francolins and spurfowls, the spatial analysis of vicariance results provided a much better picture on the history of their distribution patterns. Spatial analysis of vicariance Francolins Some barriers that were detected among the Asian species are geologically and geographically associated with the Indian sub-continent. While it may be possible that F. francolinus and F. pictus are ecologically separated, the Wandhya Range seems 298 unlikely to be a significant barrier. The Asian species generally have different habitat preferences. Ortygornis gularis is confined to marshes and reeds in the plains of the Ganges River, O. pondicerianus occurs in the Gulf of Oman in the plains of India, F. pictus inhabits the plains of India, whereas F. pintadeanus occurs in marshes and hills such as the Naga and Chin and Manipur range. Francolinus francolinus prefers mountainous areas for example, the Himalayan Mountains in Nepal, and extending to the west reaching Iran and Afghanistan. The results here suggest that habitat type could have been an effective ecological break between these F. francolinus and F. pictus. The Naga and Chin Hills as well as the higher ranges of Manipur are effective barriers along most of the boundary between F. francolinus and F. pintadeanus even though the two species might be expected to meet along the coastal strip. It could be possible that speciation may have started around the India–Eurasia border even though the actual age of this event has always been uncertain. The latest evidence points to c. 40 mya (Bouilhol 2013) while the split between Gallus/Bambusicola/francolins and Coturnix/Alectoris/spurfowls is at around 33.6 mya which is Oligocene. Plate tectonics may have led to the formation of different topographical features such as the Himalayan Mountains and the Kūhha-ye-zāgros in Iran. At the same time during collision the ancestral species may have been able to disperse into Africa possibly cutting through south eastern Saudi Arabia into Africa. The unfavourable era may have pushed the range of the ancestral species southwards, isolating the ancestral stock of the three species in south western Asia that is where F. francolinus currently occurs, southern India (F. pictus) and south eastern Asia where F. pintadeanus occurs. 299 The Oligocene to Miocene is said to have marked the start of generalized cooling, with glaciers forming in Antarctica for the first time. The increase in ice sheets led to a fall in sea level while the tropics diminished, giving way to cooler woodlands and grasslands. Diversification of habitat has occurred with A. lathami being the only francolin species that is endemic to primary forest. Among the Scleroptila taxa, there is a somewhat weaker, but possible retreat of some of the species to montane areas possibly seeking the favourable habitat. Examples of these are the southern African S. afra, the east African S. p. psilolaema, S. p. theresae and P. ellenbecki. Scleroptila afra and S. levaillantii have sympatric distributions, but they are divided altitudinally (del Hoyo et al. 1994, Little and Crowe 2000, Sinclair and Ryan 2005). The Peliperdix taxa are all found in arid or mesic wooded savannas. Habitat preference seems to play an influential role among members of the P. coqui clade. They all occur along the dry belt differing in the amount of annual rainfall that is received in those areas with the exception of P. c. kasaicas and P. stuhlmanni which occur in mesic habitats with mean annual rainfall of 800-1600 mm. The Rift Valley system which includes the Great African lakes (Lake Victoria, Lake Tanganyika, Lake Malawi), the Upper Guinea and the Congolian forest, major rivers such as the Limpopo, Zambezi and Rovuma may have played a crucial role in maintaining the distribution ranges of francolins. The two streaked subspecies, the northern O. r. spilogaster (xeric scrubs) and the southern O. r. rovuma (mesic scrubs) are separated by habitat types which are characterized by difference in the amount of rainfall received in a year. 300 Spurfowls A number of nodes have disjunctions that can be explained by geography and had detectable barriers associated with them. It is possible that the tectonic effect that resulted from the collision of Africa with Asia may have caused the clade ancestor to disperse along the Saudi Arabian coast into Africa moving to the desert in southwest Africa. Pternistis hartlaubi is the most basal spurfowl which probably is a paleoendemic species that occurs in the desert of Namibia and southwestern Angola where the mean annual rainfall is below 100 mm. The India-Saudi Arabia-Africa route and vice versa is quite possible with difficulty arising with regard to understanding how different parts of Africa including different habitats were colonized as the spurfowls diversified within various bushy habitats. However, it seems possible that, in Africa, diversification may have been initiated in southwestern Africa where P. hartlaubi occurs and subsequently followed by the radiation into forests of montane areas (Clark 1967). During this dry period the range of forest was reduced leading to montane forest pockets being occupied by the few species that preferred these kinds of habitat. Some species such as P. squamatus, P. ahantensis and P. griseostriatus remained in lowland forests, while others remained in other lowland habitats such as mesic and arid bushy and wooded grassland habitats and diversified there. Habitat type is clearly linked to the rainfall amount that each habitat type receives and this clearly separate out parts even within species. The Rift Valley system is acting as an effective barrier that is maintaining the ranges of most of the spurfowl distributions, for example among the north eastern and east African montane spurfowls, among the Bare-throated taxa P. leucoscepus/P. swainsonii/P. rufopictus/P. afer, between P. clappertoni/P. harwoodi and the Bare- 301 throated clade, as well as between the P. squamatus/P. ahantensis and P. griseostriatus/southern Vermiculated species clade. Major rivers such as the Niger, Volta, Zambezi, Orange and Rufiji River have played significant roles as barriers. The Upper Guinea and Congolian forest, Okavango swamp and also the mesic versus arid habitats were found to be effective barriers. Conclusions The history of taxa can never be explained with ease because history is generally unknown. The colonization of Africa by the ancestral species may have been through dispersal which resulted in the formation of disjunct distributions and the somewhat rapid diversification of francolins and spurfowls in Africa. Even though francolins and spurfowls evolved at almost the same time that is, during the Miocene francolins (c. 7.6 mya) are slightly younger than spurfowls (c. 8.7 mya). However, the main divergence between these two lineages occurred at around 33.6 mya in the Oligocene and this is probably the initial speciation of genus Francolinus around India. The major physical breaks that may have maintained the distributions and at the same time promoted speciation among the African francolins were the Rift Valley system which includes the Great African lakes (Lake Victoria, Lake Tanganyika, Lake Malawi), Lake Chad, the Upper Guinea and the Congolian forest, and major rivers such as Limpopo, Zambezi and Rovuma River. The diversity in habitat for example, xeric/mesic savanna versus mesic savanna, montane and lowland grasslands and forest habitat has also played an important ecological role in facilitating diversification among francolins. Among spurfowls, similar barriers as those among francolins were inferred with the addition of 302 the Niger and Rufiji River, Okavango Swamp and Sahara Desert acting as barriers between some taxa. In conclusion, the overall possible biogeographic explanation among francolins is that the ancestral species may have dispersed from Asia to Africa and later possibly returned to Asia. If Asian and African species are monophyletic, it could be fitting that during the collision of Asia and Africa some individuals remained in Asia while others remained in Africa resulting in the two groups differentiating with subsequent speciation occurring on both continents. In this case, the Asian Ortygornis gularis clade is sister to the O. sephaena complex clade and this fits the former explanation. The direction of dispersal is difficult to determine for the spurfowls. One possible explanation on the origin of the spurfowls is that the ancestor may have dispersed from Asia, into the desert of Saudi Arabia along its coast and then into Africa where it eventually arrived in the desert in south western Africa. Generally, there is support for Hall’s (1963) hypothesis on the evolution of the genus Francolinus having dispersed from Asia where its related genera occur, as opposed to Crowe and Crowe (1985) and Bloomer and Crowe’s (1998) hypothesized African origin of the genus Francolinus. Despite that the ancestral area and habitat from the maximum likelihood reconstructions were highly undecided, the spatial analysis of vicariance results at least clearly demonstrated that the ancestral francolin was Asian and they lived in dry and mesic scrubby and wooded grassland habitat. It is important to note that since the two approaches (spatial analysis of variance, ancestral character reconstructions) are dependent on the phylogeny, poor phylogenetic resolution influences the pattern. In other words, phylogenetic signal plays a core role in the outcome of the analyses and in the absence of well-resolved phylogenies, informed 303 ancestral patterns cannot be traced and the colonization route can hardly be determined. The maximum likelihood ancestral area and habitat reconstruction provided little and decisive insight into the history of the distribution of ranges of francolins and spurfowls, and a genomic approach to data collection may be required to fully resolve spurfowl relationships. 304 Francolinus francolinus F. pictus F. pintadeanus Ortygornis gularis O. pondicerianus 4 14 O. grantii O. r. rovuma O. sephaena Afrocolinus lathami lathami A. l. schubotzi Peliperdix a. albogularis P. a. buckleyi 8 P. dewittei P. schlegelii P. c. coqui 7 P. c. ruahdae 9 P. c. vernayi P. hubbardi 1 P. maharao 2 P. spinetorum 3 P. stuhlmanni P. kasaicus 5 Scleroptila afra 6 S. ellenbecki S. crawshayi crawshayi 13 12 11 S. jugularis S. levaillantoides pallidior Scleroptila l. levaillantoides S. shelleyi 10 Character: Area S. gutturalis S. uluensis Southern Asia S. finschi West Africa Scleroptila whytei Southern Africa S. p. psilolaema Central Africa S. p. theresae S. crawshayi kikuyuensis East Africa S. levaillantii S. streptophora Figure 7.1. The Bayesian reconstruction of ancestral area of francolins inferred from Bayesian topology and branch-lengths. Numbers 1-14 identifies the nodes they are associated with. The marginal probabilities are not mapped on the tree but references to them are in the main text. 305 Francolinus francolinus F. pictus F. pintadeanus Ortygornis gularis O. pondicerianus 4 14 O. grantii O. sephaena O. r. rovuma Afrocolinus lathami lathami A. l. schubotzi Peliperdix a. albogularis P. a. buckleyi 8 P. dewittei P. schlegelii P. c. coqui 7 P. c. ruahdae P. hubbardi P. c. vernayi I 1 P. maharao 2 P. spinetorum 3 P. stuhlmanni P. kasaicus 5 Scleroptila afra 6 S. ellenbecki 13 12 S. crawshayi crawshayi S. l. levaillantoides S. l. pallidior 11 Character: Habitat Open grasslands S. jugularis S. shelleyi 10 S. gutturalis S. uluensis Scrubby grasslands S. finschi Wooded grasslands S. whytei Montane grasslands S. crawshayi kikuyuensis Forest S. p. theresae S. p. psilolaema S. levaillantii S. streptophora Figure 7.2. The Bayesian reconstruction of ancestral habitat of francolins inferred from Bayesian topology and branch-lengths. Numbers 1-14 identifies the nodes they are associated with. The marginal probabilities are not mapped on the tree but references to them are in the main text. 306 Pternistis hartlaubi P. bicalcaratus bicalcaratus 2 P. b. adamauae P. icterorhynchus P. b. ayesha P. camerunensis P. nobilis P. erckelii 5 6 P. ochropectus P. castaneicollis P. swierstrai P. jacksoni P. clappertoni clappertoni 4 P. c. sharpii 8 P. harwoodi 1 P. leucoscepus leucoscepus 7 P. l. infuscatus 3 P. swainsonii P. rufopictus 9 P. cranchii P. afer humboldtii P. a. afer P. squamatus squamatus Character: Area 11 Southern Africa P. s. schuetti P. griseostriatus West Africa North Africa P. ahantensis 10 Central Africa P. hildebrandti hildebrandti East Africa 12 Pternistis h. fischeri P. natalensis P. capensis P. adspersus Figure 7.3. The Bayesian reconstruction of ancestral area of spurfowls inferred from Bayesian topology and branch-lengths. Numbers 1-12 identifies the nodes they are associated with. The marginal probabilities are not mapped on the tree but references to them are in the main text. 307 Pternistis hartlaubi P. bicalcaratus bicalcaratus 2 P. b. adamauae P. icterorhynchus Pternistis b. ayesha P. camerunensis P. nobilis 5 Pternistis erckelii 6 P. ochropectus P. castaneicollis P. swierstrai Pternistis jacksoni P. clappertoni clappertoni 4 P. c. sharpii 8 1 P. harwoodi Pternistis leucoscepus leucoscepu 7 P. l. infuscatus 3 P. swainsonii P. rufopictus 9 P. cranchii P. afer humboldtii P. a. afer P. squamatus squamatus 11 Character: Habitat P. s. schuetti P. griseostriatus Savanna & arid woodlands P. ahantensis 10 Mesic woodlands P. hildebrandti hildebrandti Forest/forest edge 12 Lowland fynbos P. h. fischeri P. natalensis Macchia P. capensis Pternistis adspersus Figure 7.4. The Bayesian reconstruction of ancestral habitat of francolins inferred from Bayesian topology and branch-lengths. Numbers 1-12 identify the nodes they are associated with. The marginal probabilities are not mapped on the tree but references to them are in the main text. 308 Gallus sp. Francolinus francolinus Francolinus pictus Francolinus pintadeanus Ortygornis pondicerianus Ortygornis gularis Ortygornis sephaena Ortygornis grantii Ortygornis rovuma spilogaster Ortygornis rovuma rovuma Afrocolinus lathami schubotzi Afrocolinus lathami lathami Peliperdix schlegelii Peliperdix dewittei Peliperdix albogularis buckleyi Peliperdi albogularis albogularis Peliperdix stuhlmanni Peliperdix kasaicus Peliperdix coqui coqui Peliperdix coqui vernayi Peliperdix coqui ruahdae Peliperdi hurbbadi Peliperdix spinetorum Peliperdix maharao Scleroptila streptophora Scleroptila levaillantii Scleroptila psilolaema theresae Scleroptila psilolaema psilolaema Scleroptila finschi Scleroptila whytei Scleroptila jugularis Scleroptila uluensis Scleroptila gutturalis Scleroptila shelleyi Scleroptila levaillantoides levaillantoides Scleroptila ellenbecki Scleroptila levaillantoides pallidior Scleroptila afra Scleroptila crawshayi kikuyuensis Scleroptila crawshayi crawshayi Figure 7.5. The francolin parsimony tree showing disjunct nodes which are associated with barriers (filled black squares) and disjunct nodes indicative of removed distribution (unfilled squares). The 309 unmarked nodes could not be explained geographically due to large overlap in distributions. Ammoperdix heyi Perdicula asiatica Pternistis hartlaubi Pternistis nobilis Pternistis camerunensis Pternistis swierstrai Pternistis castaneicollis atrifrons Pternistis castaneicollis castaneicollis Pternistis ochropectus Pternistis erckelii Pternistis jacksoni Pternistis ahantensis Pternistis squamatus schuetti Pternistis squamatus squamatus Pternistis griseostriatus Pternistis capensis Pternistis adspersus Pternistis natalensis Pternistis hildebrandti fischeri Pternistis hildebrandti hildebrandti Pternistis icterorhynchus Pternistis bicalcaratus adamauae Pternistis b. bicalcaratus ‘ayesha’ Pternistis bicalcaratus bicalcaratus Pternistis clappertoni clappertoni Pternistis harwoodi Pternistis clappertoni sharpii Pternistis leucoscepus infuscatus Pternistis leucoscepus leucoscepus Pternistis swainsonii Pternistis cranchii Pternistis rufopictus Pternistis afer humboldtii Pternistis afer afer Figure 7.6. The spurfowl parsimony tree showing showing disjunct nodes which are associated with barriers (filled black squares) and disjunct nodes indicative of removed distribution (unfilled squares). The unmarked nodes could not be explained geographically due to large overlap in distributions. 310 (a) (b) (c) Figure 7.7 (a-aa). The range maps of francolins indicating where barriers lie between each the two circumscribed clades that are being compared. The range in blue is a reference taxon/clade which is compared to the range of the taxon/clade in red. The yellow lines are barriers detected, whereas the red and dark blue lines are connecting lines. Green grids indicate area of overlap of distribution ranges. 311 (d) (e) Figure 7.7. (cont.) 312 (f) (g) (h) (i) (j) (k) (l) (m) Figure 7.7. (cont.) 313 (n) (o) (p) (q) (r) (s) (t) (u) Figure 7.7 (cont.) 314 (v) (w) (x) (y) (z) (aa) Figure 7.7 (concl.). 315 (a) (b) Figure 7.8 (a-z). The range maps of spurfowls indicating where barriers lie between two clades that are being compared. The range in blue is a reference taxon/clade which is compared to the range of the taxon/clade in red. The yellow lines are barriers detected whereas the red and dark blue lines are connecting lines. Green grids indicate areas of overlap of distribution ranges. 316 (c) (d) (e) (f) (g) (h) (i) (j) Figure 7.8. cont. 317 (k) (l) (m) (n) (o) (q) (p) (r) Figure 7.8. cont. 318 (s) (t) (u) (v) (w) (y) (x) (z) Figure 7.8. concl. 319 CHAPTER 8 Summary and Synthesis Aims of this thesis: 1. To undertake a general review of the taxonomy of ‘francolins’ [Chapter 1]. 2. To contrast alternative hypotheses concerning the monophyly of the genus ‘Francolinus’ (e.g. Hall 1963 versus Bloomer and Crowe 1998) and to further explore DNA-based, vocalization and behavioural evidence in order to test the phylogenetic affinities of the Stone Partridge Ptilopachus petrosus and Nahan’s Francolin Francolinus nahani as suggested in Crowe et al. (2006) [Chapter 2]. 3. To investigate if there are any syringeal features (anatomical) that can be used to further our understanding of phylogenetic relationships among francolin and spurfowl taxa [Chapter 3]. 4. To test the validity of the francolin-spurfowl taxonomic dichotomy based on vocal characters [Chapter 4]. 5. To test Hall’s (1963) hypotheses on the monophyly of the suggested eight species groups within her circumscription of the genus Francolinus, and thereby provide a modern systematic revision of the terminal taxa and genera of francolins and spurfowls [Chapter 5 & 6]. 6. To describe and explore the patterns of distribution of francolins and spurfowls using Hall’s (1963) monograph as a null hypothesis [Chapter 7]. 320 The key findings of this thesis A general review of the taxonomy of Francolinus (sensu lato) There has been persistent confusion regarding the taxonomic status of francolins sensu lato since the description of the genus Francolinus in the mid-1830s. This is readily demonstrated in the outline of the various taxonomic reviews carried out on the genus (Chapter 1). The confusion centred on the taxonomic and phylogenetic status of the genus, the taxonomic status and the number of valid genera, terminal species and subspecies of francolins, the phylogenetic position of Hall’s (1963) putative monophyletic species groups, as well as the phylogenetic position of Ortygornis sephaena and Hall’s (1963) four unplaced enigmatic species (now Ptilopachus nahani, Afrocolinus lathami, Ortygornis pondicerianus, Ortygornis gularis), and finally the origin of the genus Francolinus which involves two conflicting hypotheses postulated by Hall (1963) and Crowe and Crowe (1985), respectively. One of the major motivations for conducting this study was to establish a classification system that accounts for the evolutionary history of francolin and spurfowl taxa. Most species of francolins and spurfowls vary, often strikingly, in morphology, and they form complexes of species that make delimitation of taxa and the endeavour to define species complicated and difficult to achieve. The conclusions drawn from the taxonomic reviews were that taxon and character sampling procedures, types of characters (organismal versus molecules) and analytical methods (quantitative statistical methods versus model-/non-model-based phylogenetic methods) have significantly improved our understanding of the systematics of francolins and spurfowls, that have important implications for designating taxonomic rank: e.g. whether a taxon should be considered a species, subspecies or a synonym. These results in turn have improved our understanding of the biogeography of francolins and spurfowls. 321 Contrasting ideas: Monophyly of the genus Francolinus (sensu Hall 1963) and the species groups The advent of molecular data has revolutionized the understanding of evolutionary aspects of francolins and spurfowls. One of Hall’s major challenges was the difficulty of determining the phylogenetic position of Ortygornis sephaena and hence she was forced to place 41 traditionally recognized species in a single genus Francolinus. Without the use of discrete molecular data and modern phylogenetic inference methods, as well as the need for excellent taxon sampling, early studies could not decisively reject monophyly of Francolinus and had to simply allude to the possibility of francolins forming two major lineages, and hence often opted to retain all species in a single genus Francolinus. This study decisively demonstrates that the genus Francolinus is not monophyletic, instead it can be split into two distantly related assemblages (Chapter 2): ‘francolins’ (partitioned among the genera Francolinus, Ortygornis, Afrocolinus, Peliperdix, Scleroptila) and ‘spurfowls’ (all placed in the genus Pternistis). The Bayesian reconstruction for the time of divergence between these two lineages, that is, the Coturnix/Alectoris/spurfowls and Gallus/Bambusicola/francolins, was recovered at c. 33.6 mya. The estimated evolutionary date for the time-to-most-recent-common-ancestor for francolins and spurfowls was recovered at c. 7.6 and 8.7 mya, respectively. Monophyly of four (the Red-tailed, Red-winged, Spotted and the Bare-throated Groups) of eight species groups, as delimited by Hall (1963), was recovered. Significantly, this study was able to decisively place the problematic species O. sephaena within the ‘true’ francolins, despite its possession of mixed francolin and spurfowl features. This study also revealed that its closest evolutionary relatives are the two unplaced Asian species O. 322 pondicerianus and O. gularis. Ortygornis sephaena may well be a derived francolin that emerged early in the history of the genus Francolinus. Another major revelation was that Hall’s (1963) difficulty in assigning Nahan’s Francolin Francolinus ‘Ptilopachus’ nahani to any specific species group made sense since this study unequivocally revealed that Nahan’s Francolin is not a francolin but a partridge, and sister to the African Stone Partridge Ptilopachus petrosus (having diverged c. 9.6 mya details in Chapter 2), and the duo are in turn, sister to the New World quails. The remaining enigmatic morphological taxon of Hall (1963), Afrocolinus lathami was in most analyses, found to represent its own evolutionary trajectory and a new genus name Afrocolinus gen. nov. was assigned in Chapter 5. Further investigation into the phylogenetic affinities of the Stone Partridge Ptilopachus petrosus and Nahan’s Francolin Francolinus nahani (sensu Crowe et al. 2006) Francolinus nahani is a taxonomically enigmatic African galliform restricted to the interior of remnant primary forests of the eastern equatorial lowlands of the Democratic Republic of the Congo and Uganda. This thesis extended the earlier result of Crowe et al. (2006) by confirming that the closest extant relative of Francolinus nahani is Ptilopachus petrosus with an estimated divergence time at c. 9.6 mya. The two species are in turn most closely related to the New World quails (Odontophoridae) than any other Old World galliform. On the basis of the close phylogenetic relationship between F. nahani and P. petrosus, and their shared vocal and behavioural characters, this study recommended that F. nahani be moved to the genus Ptilopachus Swainson, based on name priority, subfamily Ptilopachinae subfam. nov. (Bowie et al. 2013), and family Odontophoridae. 323 Francolin-spurfowl dichotomy – syryngeal and vocalization perspective With ever increasing data suggesting a deep divergence between the francolin and spurfowl assemblages, the syringes (Chapter 3) of selected francolins and spurfowls and the calls (Chapter 4) of the various species were analyzed in an effort to further validate the divergence between these two lineages of primarily African galliforms. It was found that there were clear-cut vocal differences between (and among) francolins and spurfowls with francolins generally rendering longer tonal strophes with distinct elements that have detectable harmonics. Spurfowls, in contrast, generally give shorter atonal strophes with elements that generally have no harmonics. The structure of the syringes also supports this distinction, with the shape of the tympanum placing O. sephaena decisively with other francolins despite the lack of the mineralized bronchial half rings that are present in other francolins, and the size of the pessulus corroborates the francolin-spurfowl dichotomy. The amount of connective tissue places species with tonal and whistling calls together (Scleroptila levaillantii and S. afra) to the exclusion of those taxa that have atonal, raucous and grating calls (Pternistis capensis, P. natalensis and O. sephaena). The validity of terminal taxa - francolins and spurfowls In Chapter 5 and 6, the focus was to establish an evolutionary-based classification system to enable the identification of those francolin and spurfowl taxa that represent valid species and subspecies, as well as to determine the validity of various proposed generic names. The major disagreement recorded was at the subspecies level, followed by the generic and to some degree at the species level. As outlined in Chapters 5 and 6, this study proposes a tremendous decrease in the number of subspecies traditionally recognized in the earlier reviews (see Chapter 1) with most of these taxa being synonymized (with those deemed valid subspecies) 324 whenever they were found to not represent distinct evolutionary units and fail to meet the criteria as outlined in Chapter 5. In summary, the number of francolin and spurfowl species increased while additional genera were recognized for francolins. Five francolin genera Francolinus, Ortygornis, Afrocolinus gen. nov., Peliperdix and Scleroptila are recognized and all spurfowl species are placed in a single genus Pternistis. A number of traditionally recognized subspecies were elevated to species in this revision, resulting in 31 and 24 species of francolins and spurfowls, respectively. Although some subspecies are morphologically distinct, combined phylogenetic evidence from organismal and DNA characters, together with sequence divergence estimates based on Cytochrome-b characters, suggests that these taxa have yet to reach an independent evolutionary trajectory that warrants species status, hence they are synonymized with other taxa. What differentiates the taxonomic approach adopted in this study from those in the previous revisions is that the evolutionary relationships among taxa played a central role. The history of distributional patterns of francolins and spurfowls and the origin of the genus Francolinus The focus in Chapter 7 was to interpret the current distributional patterns of francolins and spurfowls using historical information. Further, in light of the two alternate hypotheses for the origin of francolins and spurfowl, this study sought to determine the origin of the ancestral francolin and spurfowl. Between the two approaches that were used, the disjunctiondetecting and barrier-inferring spatial analysis of vicariance or the ancestral reconstruction of area and habitat, it could be concluded that the spatial analysis of vicariance provided a better picture of the history of distribution patterns of francolins and spurfowls. So, the colonization of Africa by the ancestral species may have been through dispersal which resulted in the 325 formation of disjunct distributions and the somewhat rapid diversification of francolins and spurfowls in Africa. The direction of dispersal might have gone either way, i.e. from Asia to Africa, and vise versa. The sister relationship between the Asian O. gularis clade and the O. sephaena complex could fit the explanation that, during the collision of Asia and Africa, some individuals may have remained in Asia while others remained in Africa resulting in the two groups differentiating with subsequent speciation occurring on both continents. On the other hand, one other possible explanation on the origin of the spurfowls is that the ancestor may have dispersed from Asia, into the desert of Saudi Arabia along its coast and then into Africa where it eventually arrived in the desert in south western Africa. Generally, there is strong support for Hall’s (1963) hypothesis for the evolution of the genus Francolinus being of Asian origin. The Rift Valley system, Lake Chad, Upper Guinea and the Congolian forest, major rivers such as Limpopo, Zambezi, Rovuma and the Volta River were found to be maintaining the distribution patterns and possibly promoting speciation among African francolins. In addition, the Niger and Rufiji River, Okavango swamp and the Sahara desert have maintained the distribution ranges of the spurfowl taxa. Habitat heterogeneity e.g. xeric/mesic savanna versus mesic savanna, montane and lowland grasslands and forest habitat also play an important ecological role in facilitating diversification among francolins. Among spurfowls, similar habitat barriers to those detected among francolins were inferred. Future prospects Taxonomy remains a challenging discipline especially in the absence of a universal definition, which can be applied at both the species and subspecies level. The long-standing debate on the definition of species (as well as subspecies) is one that will remain difficult to resolve. The 326 conflict exists between character partitions from different sources possibly due to the high level of hybridization and incomplete lineage sorting and these conditions make taxonomic practice complicated. Thus, there has been profound taxonomic confusion. Future prospects to specifically focus on the phylogeographic studies of selected species complexes of francolins and spurfowls and also to investigate the nature of hybrid zones, are brought forth. The role of various syringeal features in voice production among francolins and spurfowls should be investigated. 327 CHAPTER 9 References Akaike H 1973 Information theory as an extension of the maximum likelihood principle. Pp. 276-281. In: Second International Symposium on Information Theory. (Petrov BN, Csaki F, eds). Akademiai Kiado, Budapest. Alexander, 1909 Francolinus camerunensis. Bulletin of the British Ornithologists´ Club 25, p. 12. Alström P 2001 The use of sounds in bird systematics. Introductory Research Essay No. 2. Uppsala University, Uppsala, Sweden. Alström P, Rasmussen PC, Olsson B, Sundberg P 2008 Species delimitation based on multiple criteria: the Spotted Bush Warbler Bradypterus thoracicus complex (Aves: Megaluridae). Zoological Journal of the Linnean Society 154, pp. 291307. Ames PJ 1971 The morphology of the syrinx in passerine birds. Bulletin of the Peabody Museum of Natural History 37, pp 1-94. Appel FW 1929 Sex dimorphism in the syrinx of the fowl. Journal of Morphology 47(2), pp. 497-517. Arias JS, Szumik CA, Goloboff PA 2011 Spatial analysis of vicariance: a method for using direct geographical information in historical biogeography. Cladistics 27, pp. 1-12. Armstrong MH, Braun EL, Kimball RT 2001 Phylogenetic utility of avian Ovomucoid intron G: a comparison of nuclear and mitochondrial phylogenetics in Galliformes. The Auk 118, pp. 799-804. 328 Aubrecht G, Holzer G 2000 Die Stockenten: Biologie, O¨ kologie, Verhalten. 1. Östechischer, Agrarverlag, Vienna, Austria. Bangs & Loveridge, 1931 Pternistis squamatus uzungwensis. Proceedings of the New England Zoological Club 12, p. 93. Bannerman, 1934 Francolinus ahantensis hopkinsoni. Bulletin of the British Ornithologists´ Club 55, p. 5. Bannerman D, 1934 Francolinus ahantensis hopkinsoni. [A new race of the Ahanta Francolin from Gambia and Portuguese Guinea]. Bulletin of the British Ornithologists´ Club 55, pp. 5-6. Bannerman, 1922 Francolinus bicalcaratus ogilviegranti. Bulletin of the British Ornithologists´ Club 42, p. 132. Bancroft JD, Gamble M 2002 Theory and practice of histological techniques (5th ed). Churchill Livingstone, NY. Bates, 1928 Francolinus coqui spinetorum. Bulletin of the British Ornithologists´ Club 49, p. 33. Benson CW, 1939 Francolinus squamatus doni Benson. A new Francolin from Nyasaland. Bulletin of the British Ornithologists´ Club 59, pp. 42-43. Benson, 1939 Pternistis squamatus doni. Bulletin of the British Ornithologists´ Club 59, p. 42. Benson CW 1945 Notes on the birds of sourthern Abyssinia (part). Ibis 87, pp. 366400. Benson CW, Irwin MPS, White CM 1962 The significance of valleys as avian zoogeographical barriers. Symposium on causes and problems of animal distribution with special reference to southern Africa. Annals of the Cape 329 Provincial Museum 2, pp. 155-189. Bininda-Edmonds ORP, Jones KE, Price SA, Cardillo M, Grenyer R, Purvis A 2004 Garbage in garbage out: data issues in supertree constructions. Pp. 267-280. In: Phylogenetic supertrees: combining information to reveal the Tree of Life (Binida-Edmonds ORP, ed). Kluwer, Dordrecht. Bloomer P, Crowe TM 1998 Francolin phylogenetics: molecular, morphobehavioral, and combined evidence. Molecular Phylogenetics and Evolution 9(2), pp. 236-254. Blundell & Lovat, 1899 Francolinus squamatus tetraoninus. Bulletin of the British Ornithologists' Club 10, p. 22. Blundell & Lovat 1899 Francolinus harwoodi. Bulletin of the British Ornithologists´ Club 10, p. 22. (Blyth, 1843) Francolinus pintadeanus phayrei. Journal of the Asiatic Society of Bengal 12, p. 1011. Blyth, 1852 Scleroptila. Catalogue of the birds in the museum [of the] Asiatic, p. 250. Bocage, 1869 Francolinus hartlaubi hartlaubi. Jornal de Sciencias mathematicas, physicas e naturas, publicado sob os auspicos da Academia real das sciencias da Lisboa 2(8), p. 350. Bocage, 1881 Francolinus finschi. Ornithologie d'Angola pt2, p. 406. Bocage, 1893 Pternistis afer benguellensis. Jornal de Sciencias mathematicas, physicase naturas, publicado sob os auspicos da Academia real das sciencias da Lisboa 3(2), p. 154. Bock WJ, Farrand J 1980 The number of species and genera of Recent birds. American Museum Novitates 2703, pp. 1-29. Bonaparte, 1856 Francolinus francolinus asiae. Comptes Rendus 42, p. 882. 330 Bonaparte, 1856 Francolinus francolinus henrici. Comptes Rendus 42, p. 882. Bonaparte, 1856 Peliperdix. Comptes Rendus 42, p. 882. Bouilhol P, Jagoutz O, Hanchar JM, Dudas FO 2013 Dating the India–Eurasia collision through arc magmatic records. Earth and Planetary Science Letters 366 (2013), pp. 163-175. Bowen, 1930 Pternistis afer itigi. Proceedings of the Academy of Natural Sciences of Philadelphia 82, p. 86. Bowen, 1931Francolinus jacksoni gurae. Proceedings of the Academy of Natural Sciences of Philadelphia 83, p. 302. Bowie RCK, Fjeldså J 2005 Genetic and morphological evidence for two species in the Udzungwa Forest Partridge Xenoperdix udzungwensis. Journal of East African Natural History 94, pp. 191-201. Bowie RCK, Cohen C, Crowe TM 2013 Ptilopachinae: a new subfamily of the Odontophoridae (Aves: Galliformes). Zootaxa 3670(1), pp. 097-098. Brown C, Ward D 1990 The morphology of the syrinx in the Charadriiformes (Aves): possible phylogenetic implications. Bonner Zoologische Beiträge 41, pp. 95-107. Büttikofer, 1889 Francolinus levaillantoides jugularis. Notes from the Leyden Museum 11, p. 76. Cabanis, 1868 Pternistes leucoscepus infuscatus. Journal of Ornithology 16, p. 413. Cabanis, 1878 Pternistis hildebrandti hildebrandti. Journal of Ornithology 26(142), p. 206. Cabanis, 1880 Pternistis squamatus schuetti. Journal of Ornithology 28(152), p. 351. Cannell PF 1988 Techniques for the study of avian syrinxes. Wilson Bulletin 10, pp. 289-293. 331 Chapin JP 1926 A new genus Acentrortyx proposed for Francolinus nahani Dubois. The Auk 43, pp. 235. Chapin JP 1932 The birds of the Belgian Congo. Bulletin of the American Museum of Natural History. Vol. LXV, New York. Chapin JP, 1937 Francolinus albogularis dewittei. A new race of Francolinus albogularis from Marungu. Revue de zoologie et de botanique africaines 29, pp. 395-396. Chappuis, C 2000 African bird sounds. Birds of North, West and Central Africa. Audio CD set. With the collaboration of the British Library National Sound Archive (London). Société d Études Ornitholgiques de France, Paris. Cassin, 1857 Pternistis squamatus squamatus. Proceedings of the Academy of Natural Sciences of Philadelphia 8, p. 321. (Children & Vigors, 1826) Francolinus clappertoni clappertoni. Narrative of Travels and Discoveries in Northern and Central Africa App. 21, p. 198. Cicero C 2010 The significance of subspecies: a case study of sage sparrows (Emberizidae, Amphispiza belli). The American Ornithologists’ Union 67, pp. 103-113. Clancey PA 1957 On the range and status of Certhilauda falcirostris Reichenow, 1916: Port Nolloth, N.W. Cape. Bulletin of the British Ornithilogists’ Club 77, pp. 133-137. Clancey PA, 1960 Francolinus shelleyi sequestris. Miscellaneous taxonomic notes on African birds. XV. 1. A new race of Shelley’s Francolin Francolinus shelleyi Ogilvie-Grant from Natal and Zululand. Durban Museum Novitates 6(2), pp. 1314. 332 Clancey, 1996 Francolinus adspersus mesicus. Bulletin of the British Ornithologists´ Club 116, p. 106. Clancey PA 1967 Gamebirds of Southern Africa. Purnell & Sons, Cape Town. Clancey PA 1986 Endemicity in the southern African avifauna. Durban Museum Novitates 13(20), pp. 245-284. Clark JD 1967 Atlas of African prehistory. The University of Chicago, London. Cohen C, Wakeling JL, Mandiwana-Neudani TG, Sande E, Dranzoa C, Crowe TM, Bowie RCK 2012 Phylogenetic affinities of evolutionarily enigmatic African galliforms: the Stone Partridge Ptilopachus petrosus and Nahan’s Francolin Francolinus nahani, and support for their sister relationship with New World quails. Ibis 154, pp. 768-780. Conover, 1929 Pternistes afer nyanzae. Auk 46, p. 345. Conover, 1928 Pternistis squamatus usambarae. Auk 45, p. 356. (Conover, 1930) Francolinus castaneicollis atrifrons. Proceedings of the Biological Society of Washington 43, p. 3. Cotterill FDP 2006 Taxonomic status and conservation importance of the avifauna of Katanga (south-east Congo Basin) and its environs. Ostrich - Journal of African Ornithology 77(1&2), pp. 1-2. Cox WA, Kimball RT, Braun EL 2007 Phylogenetic position of the New World quails: eight nuclear loci and three mitochondrial regions contradict morphology and the Sibley/Ahlquist tapestry. The Auk 124(1), pp. 71-84. Cracraft J 1994 Species diversity, biogeography, and the evolution of biotas. American Zoologist 34, pp. 33-47. Cracraft J 1981 Toward a phylogenetic classification of the recent birds of the world. 333 The Auk 98, pp. 681-714. Crowe TM, Crowe AA 1985 The genus Francolinus as a model for avian evolution and biogeography in Africa. Pp. 207-231. In: Proceedings of the International Symposium on African vertebrates (Schuchmann KL, ed). Museum Alexander Koenig, Bonn. Crowe TM, Keith S, Brown LH 1986 Phasianidae, guineafowl, Congo peacock, quail, partridges and francolins. Pp. 1-75. In: The Birds of Africa (Urban EK, Fry CH, Keith S, eds). Vol II. Academic Press, London. Crowe TM 1992 Morphometric research on Miocene and Pliocene fossil phasianids from the west coast of southern Africa: some preliminary taxonomic and phylogenetic conclusions. Pp. 261-267. In: Proceedings of the 7th Pan-African Ornithological Congress (Bennun L, ed). PAOC Committee, Nairobi. Crowe TM, Short LL 1992 A new gallinaceous bird from the Oligocene of Nebraska, with comments on the phylogenetic position of the Gallinuloididae. Natural History Museum of Los Angeles County Science Series 36, pp. 179-185. Crowe TM, Harley EH, Jakutowicz M 1992a Phylogenetic relationships of southern African francolins (Francolinus spp.) as suggested by the structure of their mitochondrial DNA and their morphology, behaviour and ecology. Proceedings of the 7th Pan-African Ornithological Congress, pp. 261-267. Crowe TM, Harley EH, Jakutowicz MB, Komen J, Crowe AA 1992b Phylogenetic, taxonomic and biogeographical implications of genetic, morphological, and behavioural variation in francolins (Phasianidae. Francolinus). The Auk 109(1), pp. 24-42. Crowe TM, Bowie RCK, Bloomer P, Mandiwana TG, Hedderson TAJ, Randi E, 334 Pereira SL, Wakeling J 2006 Phylogenetics, biogeography and classification of, and character evolution in, gamebirds (Aves: Galliformes): Effects of character exclusion, data partitioning and missing data. Cladistics 22, pp. 1-38. Darwin C 1859 On the origin of species. Facsmile of first edition (1964). Harvard University Press, Cambridge, Massachusetts. Dayrat B 2005 Towards integrative taxonomy. Biological Journal of the Linnean Society 85, pp. 407-415. Delacour J, Mayr E 1945 The Family Anatidae. Wilson Bulletin 57, pp. 3-55. del Hoyo J, Elliott A, Sargatal J 1994 Handbook of the birds of the world: New World Vultures to Guineafowl. Vol. II. Lynx edicions, Barcelona. de Queiroz A 2005 The resurrection of oceanic dispersal in historical biogeography. Trends in Ecology and Evolution 20, pp. 68–73. de Queiroz K 2007 Species concept and species delimitation. Systematic Biology 56(6), pp. 879-886. De Schauensee, 1931 Francolinus adspersus kalahari. Proceedings of the Academy of Natural Sciences of Philadelphia 83, p. 453. Dickinson EC (ed) 2003 The Howard and Moore complete checklist of the birds of the World. 3rd ed. Christopher Helm, London. Dimcheff DE, Drovetski SV, Krishnan M, Mindell DP 2000 Cospeciation and horizontal transmission of avian sarcoma and leucosis virus gag genes in galliform birds. Journal of Virology 74, pp. 3984-3995. Dimcheff DE, Drovetski SV, Krishnan M, Mindell DP 2002 Phylogeny of Tetraoninae and other galliform birds using mitochondrial 12S and ND2 genes. Mololecular Phylogenetics and Evolution. 24, pp. 203-215. 335 Dinesen L, Lehmberg TJ, Svendsen O, Hansen TA, Fjeldså J 1994 A new genus and species of perdicine bird from Tanzania: a relict form with Indo-Malayan affinities. Ibis 136, pp. 3-11. Donoghue MJ, Moore BR 2003 Toward an integrative historical biogeography. Integrative and Comparative Biology 43, pp. 261-270. (Dorst & Jouanin, 1952) Pternistis ochropectus. L'Oiseau et la Revue Francaise d'Ornithologie 22, p. [71]. Dowsett RJ, Prigogine A 1974 The avifauna of the Marungu Highlands. Exploration Hydrobiologique du Bassin du lac Bangweolo et du Luapula 14, pp. 1-67. Drummond AJ, Rambaut A 2007 BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7, pp. 214. (Dubois, 1905) Francolinus nahani. Annales du Musée du Congo 1(pl.10), p. 17. Dyke GJ 2003 The phylogenetic position of Gallinuloides Eastman (Aves: Galliformes) from the Tertiary of North America. Zootaxa 199, pp. 1-10. Dyke GJ, Gulas BE, Crowe TM 2003 Suprageneric relationships of galliform birds (Aves, Galliformes): a cladistic analysis of morphological characters. Zoological Journal of the Linnean Society 137, pp. 227-244. Edwards SV, Wilson AC 1990 Phylogenetically informative length polymorphism and sequence variability in mitochondrial DNA of Australian songbirds (Pomatostomus). Genetics 126, pp. 695-711. Edwards SV, Arctander P, Wilson AC 1991 Mitochondrial resolution of a deep branch in the genealogical tree for perching birds. Proceedings of the Royal Society 243, pp. 99-107. 336 Elzanowski A, Stidham TA 2011 A gallofanserine quadrate from the Late Cretaceous Lance formation of Wyoming. The Auk 128(1), pp.138-145. Enggist-Düblin P, Birkhead TR 1992 Differences in the calls of European and North American Black-billed Magpie and the Yellow-billed Magpie. The International Journal of Animal Sound and its Recording 4, pp. 185-194. Eberle JJ Greenwood DR 2012 Life at the top of the greenhouse Eocene world – a review of the Eocene flora and vertebrate fauna from Canada’s High Arctic. Bulletin of the Geological Society of America 124, pp. 3-23. Eo SH, Bininda-Evans ORP, Carroll JP 2009 A phylogenetic supertree of the fowls (Galloanserae, Aves). Zoologica Scripta 38(5), pp. 465-481. Farais IP, Ortí G, Meyer A 2000 Total evidence: molecules, and the phylogenetics of cichlid fishes. Journal of Experimental Zoology (Molecular and Developmental Evolution) 288, pp. 76-92. Farnsworth A, Lovette IJ 2008 Phylogenetic and ecological effects on interspecific variation in structurally simple avian vocalizations. Biological Journal of the Linnean Society 94, pp. 155-173. Farris JS, Albert VA, Källersjö M, Lipscomb D, Kluge AG 1996 Parsimony jackknifing outperforms neighbor-joining. Cladistics 12, pp. 99-124. Felsenstein J 1985 Confidence limits on phylogenies: an approach using the Bootstrap. Evolution 39(4), pp. 783-791. (Fischer & Reichenow, 1884) Pternistis afer leucoparaeus. Journal of Ornithology 32(166), p. 263. Fischer & Reichenow, 1884 Francolinus hildebrandti altumi. Journal of Ornithology 32(165), p. 179. 337 FitzPatrick JW 2010 Subspecies are for convenience. The American Ornithologists’ Union 67, pp. 54-61. Fjeldså J, Bowie RCK 2008 New perspectives on the origin and diversification of Africa's forest avifauna. African Journal of Ecology 46, pp. 235-247. Forcina G, Panayides P, Guerrini M, Nardi F, Gupta BK, Mori DE, Al-Sheikhly EF, Mansoori J, Khaliq GI, Rank DN, Parasharya BM, Khan AA, Hadjigerou P, Barbanera F 2012 Molecular evolution of the Asian francolins (Francolinus, Galliformes): a modern reappraisal of a classic study in speciation. Molecular Phylogenetics and Evolution 65(2012), pp. 523-534. Frank T, Walter I, Probst A, König HE 2006 Histological aspects of the syrinx of the male Mallard (Anas platyrhynchos). Anatomia, Histologia, Embryologia 35, pp. 396-401. Frank T, Probst A, König HE, Walter I 2007 The syrinx of the male Mallard (Anas platyrhynchos): special anatomical features. Anatomia, Histologia, Embryologia 36, pp. 121-126. Friesen VL, Congdon BC, Walsh HE, Birt TP 1997 Intron variation in Marbled Murrelets detected using analyses of single-stranded conformation polymorphisms. Molecular Ecology 6, pp.1047-1058. Fumihito A, Miyake T, Takada M, Ohno S, Kondo N 1995 The genetic link between the Chinese bamboo partridge (Bambusicola thoracica) and the chicken and junglefowls of the genus Gallus. Proceedings of the National Academy of Sciences of the United States of America 92, pp. 11053–11056. 338 Gaban-Lima R, Höfling E 2006 Comparative anatomy of the syrinx in the tribe Arini (Aves: Psittacidae). Brazilian Journal of Morphological Sciences 23, pp. 501-512. Gatesy J, Mathee C, DeSalle R, Hayashi C 2002 Resolution of a supertree/supermatrix paradox. Systematic Biology 51, pp. 652-654. Gaunt AS, Gaunt SLL, Casey RM 1982 Syringeal mechanics reassessed: evidence from Streptopelia. The Auk 99, pp. 474-494. Gibbon G 1995 Southern African Bird sounds. Southern, Durban, South Africa, African Birding cc. Gill FB 1990 Ornithology. 2nd ed. WH Freeman and Company, New York. Gill F, Donsker D (eds) 2013 IOC World Bird List (version 3.4). Available at http://www.worldbirdnames.org/ [Accessed "20/08/2013"]. (Gmelin, 1789) Francolinus capensis. System Naturae 1 (pt2), p. 759. (Gmelin, 1789) Francolinus pondicerianus pondicerianus. Systema Naturae 1(pt2), p. 760. Goloboff PA, Farris JS, Nixon KC 2008 TNT, a free program for phylogenetic analysis. Cladistics 24, pp. 774-786. Grant CHB & Mackworth-Praed CW, 1934 Francolinus coqui thikae. Descriptions of two new races of Francolin]. Bulletin of the British Ornithologists’ Club 54, pp. 173-174. Grant & Mackworth-Praed, 1934 Pternistis afer loangwæ. Bulletin of the British Ornithologists' Club 55, p. 17. Grant CHB & Mackworth-Praed CW, 1934 Francolinus castaneicollis kaffanus. [Descriptions of two new races of Francolin.]. Bulletin of the British 339 Ornithologists´ Club 54, pp. 173-174. Grant CHB & Mackworth-Praed CW, 1934 Francolinus nobilis chapini. On a new subspecies of Francolin from eastern Africa. Bulletin of the British Ornithologists´ Club 55, pp. 62-63. Grant CHB & Mackworth-Praed CW, 1934 Francolinus sephaena somaliensis. [The eastern African races of Francolinus sephæna (Smith).]. Bulletin of the British Ornithologists´ Club 54, pp. 170-173. (Gray JE, 1831) Francolinus pictus pallidus. Illustrations of Indian Zoology [Hardwicke] 1(pt8, pl.55). (Gray JE, 1834) Pternistis afer punctulatus. Illustrations of Indian Zoology[Gray & Hardwicke] 2(pl.43, f.2). Gray GR, 1867 Francolinus psilolaema psilolaema. List of the specimens of Birds in the British Museum (pt5), p. 50. (Gray GR, 1867) Pternistis leucoscepus leucoscepus. List of the specimens of Birds in the British Museum (pt5), p. 48. Gray GR, 1867 Francolinus sephaena rovuma. List of the specimens of Birds in the British Museum (pt5), p. 52. Griffiths CS 1994a Monophyly of the Falconiformes based on syringeal morphology. The Auk 111(4), pp. 787-805. Griffiths CS 1994b Syringeal morphology and the phylogeny of the Falconidae. The Condor 96, pp. 127-140. Gunning & Roberts, 1911 Pternistis afer castaneiventer. Annals of the Transvaal museum 3(2), p.110. Hackett SJ, Kimball RT, Reddy S, Bowie RCK, Braun EL, Braun MJ, Chojnowski 340 JL, Cox WA, Han K-L, Harshman J, Huddleston CJ, Marks BD, Miglia KJ, Moore WS, Sheldon FH, Steadman DW, Witt CC, Yuri T 2008 A phylogenomic study of birds reveals their evolutionary history. Science 320, pp. 1763-1768. Hall BP 1963 The francolins, a study in speciation. Bulletin of the British Museum (Natural History) 10, pp. 8-204. Hall R 2001 Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model and animations. Journal of Asian Earth Sciences 20, pp. 353-431. Harrington GJ, Eberle J, Le-Page BA, Dawson M, Hutchison JH 2012 Arctic plant diversity in the Early Eocene greenhouse. Proceedings of the royal society B - Biological Sciences 279, pp. 1515-1521. Harrison JA, Allan DG, Underhill LG, Herremans M, Tree AJ, Parker V, Brown, CJ (eds) 1997 The Atlas of Southern African Birds: Non-passerines. Vol. 1. BirdLife South Africa, Johannesburg. Hartert, 1917 Francolinus pondicerianus interpositus. Novitates Zoologicae 24, p. 288. Hartert, 1917 Francolinus bicalcaratus ayesha. Novitates Zoologicae 24, p. 291. Hartlaub, 1854 Francolinus lathami lathami. Journal of Ornithology 2(9), p. 210. Hartlaub, ["1865"]1866 Francolinus sephaena grantii. Proceedings of the Zoological Society of London (pt3, pl.39), p. 665. Heuglin, 1863 Francolinus schlegelii, Journal of Ornithology 11(64), p. 275. Hockey PAR, Dean WRJ, Ryan PG (eds) 2005 Roberts - Birds of Southern Africa. 7th ed. The Trustees of the John Voelcker Bird Book Fund, Cape Town. Hoesch W & Niethammer, G 1940 Francolinus gariepensis stresemanni. Die 341 Vogelwelt Deutsch-Südwest-Afrikas namentlich des Damara- und Namalandes. Journal of Ornithology 88, pp. 1-404. Hogg DA 1982 Ossification of the laryngeal, trachea and syringeal cartilages in the domestic fowl. Journal of Anatomy 134, pp 57-71. Hovenkamp P 1997 Vicariance events, not areas, should be used in biogeographical analysis. Cladistics 13, pp. 67-79. Hovenkamp P 2001 A direct method for the analysis of vicariance patterns. Cladistics 17, pp. 260-265. Huelsenbeck JP, Ronquist F 2001 MrBayes: Bayesian inference of phylogeny. Bioinformatics 17, pp. 754-55. Hume, 1888 Francolinus francolinus melanonotus. Stray Feathers 11, p. 305. Humphrey PS 1955 The relationships of the seaducks (Tribe Mergini). Unpublished Ph.D. thesis, University of Michigan-Ann Arbor. James FC 1970 Geographic size variation in birds and its relationship to climate. Ecology 51(3), pp. 365-390. (Jardine & Selby, 1828) Francolinus pictus pictus. Illustrations of Ornithology 1(pl.50). Johnsgard PA 1961 Tracheal anatomy of the Anatidae and its taxonomics significance. Wildfowl 12, pp. 58-69. Johnsgard PA 1973 The Grouse and Quails of North America. University of Nebraska Press, Lincoln. Johnsgard PA 1986 The Pheasants of the World. Oxford University Press, Oxford. Johnsgard PA 1988 The Quails, Partridges, and Francolins of the World. Oxford University Press, Oxford. 342 Johnsgard PA 1999 The Pheasants of the World. 2nd ed. Smithsonian Institution Press, Washington, D.C. Jolles J, Schoentgen F, Jolles P, Wilson AC 1976 Amino acid sequence and immunological properties of chachalaca egg white lysozymes. Journal of Molecular Evolution 8, pp. 59-78. Jolles J, Ibrahimi IM, Prager EM, Schoentgen F, Jolles P, Wilson AC 1979 Amino acid sequence of pheasant lysozyme. Evolutionary change affecting processing of pre-lysozyme. Biochemistry 18, pp. 2744-2752. Jønsson KA, Bowie RCK, Moyle RG, Les Christidis JAN, Benz BW, Fjeldså J 2010 Historical biogeography of an Indo-Pacific passerine bird family (Pachycephalidae): different colonization patterns in the Indonesian and Melanesian archipelagos. Journal of Biogeography 37, pp. 245-257. Källersjö M, Farris JS, Chase MW, Bremer B, Fay MF, Humphries CJ, Petersen G, Seberg O, Bremer K 1998 Simultaneous parsimony jackknife analysis of 2538 rbcL DNA sequences reveals support for major clades of green plants, land plants, seed plants and flowering plants. Plant Systematics and Evolution 213, pp. 259-287. Katoh K, Asimenos G, Toh H 2009 Multiple alignment of DNA sequences with MAFFT. Methods in Molecular Biology 537, pp. 39-64. Kimball RT, Braun EL 2008 A multigene phylogeny of Galliformes supports a single origin of erectile ability in non-feathered facial traits. Journal of Avian Biology 39, pp.438-445. Kingdon J 1989 Island Africa. Collins, London. Knowles LL, Carstens BC 2007 Delimiting species without monophyletic gene trees. 343 Systematic Biology 56(6), pp. 887-895. Komen J 1987 Preliminary observations of the social pattern, behaviour and vocalization of Hartlaub’s francolin. South African Journal of Wildlife Research Supplement 1, pp. 82-86. Kornegay JR, Kocher TD, Williams LA, Wilson AC 1993 Pathways of lysozyme evolution inferred from the sequences of cytochrome b in birds. Journal of Molecular Evolution 37, pp. 367-379. Krakauer AH, Tyrrell M, Lehmann K, Losin N, Goller F, Patricelli GL 2009 Vocal and anatomical evidence for two-voiced sound production in the greater sagegrouse Centrocercus urophasianus. Journal of Experimental Biology 212, pp 3719-3727. Kriegs JO, Matzke A, Churakov G, Kuritizin A, Mayr G, Brosius J, Schmitz J 2007 Waves of genomic hitchhikers shed light on the evolution of gamebirds (Aves: Galliformes). BMC Evolutionary Biology 7, pp.190-212. Ksepka DT 2009 Broken gears in the avian molecular clock: new phylogenetic analyses support stem galliform status for Gallinuloides wyomingensis and rallid affinities for Amitabha urbsinterdictensis. Cladistics 25, pp. 173-197. Lamm KS, Redelings BD 2009 Reconstructing ancestral ranges in historical biogeography: properties and prospects. Journal of Systematics and Evolution 47, pp. 369-382. Lanyon WE 1969 Vocal characters and avian systematics. Pp. 291-310. In: Bird Vocalizations: their relation to current problems in Biology and Psychology (Hinde RA). Cambridge University Press, London. Lanyon WE 1986 A phylogeny of the thirty-three genera in the Empidonax assemblage 344 of tyrant flycatchers. American Museum Novitates 2846, pp. 1-64. Larsen ON, Goller F 2002 Direct observation of syringeal muscle function in songbirds and a parrot. Journal of Experimental Biology 205, pp. 25-35. (Latham, 1790) Francolinus afra. Index ornithologicus, sive systema ornithologiae 2, p. 648. (Lawson, 1963) Francolinus shelleyi canidorsalis. Durban Museum Novitates 7, p. 77. (Lawson WJ, 1963) Francolinus shelleyi canidorsalis. A contribution to the Ornithology of Sul Do Save, southerm Moçambique. Durban Museum Novitates 7, pp. 73-124. (Leach & Koenig KD, 1818) Pternistis afer cranchii. Narrative of an expedition to explore the river Zaire[Tuckey] App.4, p. 408. Legge, 1880 Francolinus pictus watsoni. Birds Ceylon 3, p. 745. Lei F, Wang A, Wang G, Yin Z 2005 Vocalizations of Red-necked Snow Finch, Pyrgilauda ruficollis on the Tibetan Plateau, China – a syllable taxonomic signal?. Folia Zoologica 54(1-2), pp. 135-146. Lewis B 1983 Bioacoustics: a comparative approach. Academic Press, London. Lewis PO 2001 A likelihood approach to estimating phylogeny from discrete morphological character data. Sytematic Biology 50(6), pp. 913-925. (Linnaeus, 1766) Francolinus francolinus francolinus. Systema Naturae, edition 12, p. 275. (Linnaeus, 1766) Francolinus bicalcaratus bicalcaratus Systema Naturae, edition 12 p. 277. Little R, Crowe T 2000 Gamebirds of southern Africa. Struik publishers, Cape Town. Livezey BC 1986 A phylogenetical analysis of recent Anseriform genera using 345 morphological characters. The Auk 103, pp 737-754. Macdonald JD, 1940 Francolinus clappertoni cavei. A new race of Francolin and a new race of Lark from the Sudan. Bulletin of the British Ornithologists´ Club 60, pp. 57-59. Macdonald JD, 1953 Francolinus levaillantoides wattii. The races in South West Africa of the Orange River Francolin. Bulletin of the British Ornithologists´ Club 73, pp. 34-36. Mackworth-Praed CW, Grant CHB 1952 African Handbook of Birds. Series I: Birds of Eastern and North Eastern Africa. Vol. I. Longmans, London. Mackworth-Praed CW, Grant CHB 1962 African Handbook of Birds. Series II: Birds of the Southern Third of Africa. Vol. II. Longmans, London. Mackworth-Praed CW, Grant CHB 1970 African Handbook of Birds. Series III: Birds of West Central and Western Africa. Vol. I. Longmans, London. Mackworth-Praed, 1920 Francolinus castaneicollis ogoensis. Bulletin of the British Ornithologists´ Club 40, p. 141. Mackworth-Praed, 1920 Francolinus erckelii pentoni. Bulletin of the British Ornithologists´ Club 40, p. 141. Mackworth-Praed, 1920 Francolinus natalensis neavei. Bulletin of the British Ornithologists´ Club 40, p.140. Mackworth-Praed, 1920 Francolinus sephaena zambesiae. Bulletin of the British Ornithologists´ Club 40, p.140. Maclean GL 1985 Roberts’ Birds of Southern Africa. 5th ed. The Trustees of the John Voelcker Bird Book Fund, Cape Town. Maclean GL 1993 Roberts’ Birds of Southern Africa. 6th ed. The Trustees of the John 346 Voelcker Bird Book Fund, Central News Agency, Cape Town. Madarasz 1914 Francolinus clappertoni koenigseggi, Annales Historico-Naturales Musei Nationalis Hungarici 12, p. 560. Maddison DR 1991 The discovery and importance of multiple islands of most parsimonious trees. Systematic Biology 40, pp. 315-328. Maddison WP, Maddison DR 2006 Mesquite: a modular system for evolutionary analysis. Version 1.12. Available from: <http://mesquiteproject.org>. Madge S, McGowan P 2002 Pheasants, Partridges, and Grouse: a guide to the Pheasants, Partridges, Quails, Grouse, Guineafowl, Buttonquails, and Sandgrouse of the world. Princeton Univ. Press, Princeton, N.J. Mallet J 2005 Hybridization as an invasion of the genome. Trends in Ecology & Evolution 20, pp. 229-237. Mallet J 2008 Hybridization, ecological races and the nature of species: empirical evidence for the ease of speciation. Philosophical Transactions of the Royal Society B: Biological Sciences 363, pp. 2971-2986. Marler P 1969 Tonal quality of bird sounds. Pp. 5-18. In: Bird Vocalizations: Their Relation to Current Problems in Biology and Psychology (Hinde RA, ed). Cambridge University Press, London. Mayden RL 1997 A hierarchy of species concepts: the denouement in the saga of the species problem. Pp. 381-424. In: Claridge MF, Dawah HA, Wilson MR. Species: the units of biodiversity. Chapman & Hall, UK. Mayr E 1982 The growth of biological thought. Belknap P. of Harvard U.P.: Cambridge (Mass.) Mccracken KG, Sheldon FH 1997 Avian vocalizations and phylogenetic 347 signal. Proceedings of the National Academy of Sciences of the United States of America 94, pp. 3833-3836. McGowan PJK 1994 Family Phasianidae (Pheasants and Partridges). In: del Hoyo J, Elliott A, Sargatal J. Handbook of the birds of the world: New world Vultures to Guineafowls, Vol. 2. Lynx edicions, Barcelona. McLachlan GR, Liversidge R 1957 Roberts Birds of South Africa. 2nd ed. The Trustees of the South African Bird Book Fund, Cape Town. McLachlan GR, Liversidge R 1970 Roberts Birds of South Africa. 3rd ed. The Trustees of the John Voelcker Bird Book Fund, Central News Agency, Cape Town. McLachlan GR, Liversidge R 1978 Roberts Birds of South Africa. 4th ed. The Trustees of the John Voelcker Bird Book Fund, Cape Town. Mearns, 1910 Pternistis squamatus maranensis. Smithsonian miscellaneous collections 56(14), p. 1. Mearns, 1911 Francolinus squamatus zappeyi. Smithsonian Miscellaneous Collections 66(20), p. 4. Mearns, 1911 Pternistes leucoscepus keniensis. Smithsonian Miscellaneous Collection 66(20), p. 1. Mearns, 1911 Pternistes leucoscepus kilimensis. Smithsonian Miscellaneous Collections 66(20), p. 2. Mearns, 1915 Francolinus hildebrandti helleri. Proceedings of the U.S. National Museum 48, p. 381. Meinertzhagen, 1937 Francolinus shelleyi theresae. [Six new races from Mt. Kenya, Kenya Colony.]. Bulletin of the British Ornithologists´ Club 57, pp. 67-70. Meinertzhagen R, 1937 Francolinus jacksoni pollenorum. [Six new races from Mt. 348 Kenya, Kenya Colony.]. Bulletin of the British Ornithologists´ Club 57, pp. 6770. Meiri S, Dayan T 2003 On the validity of Bergmann’s rule. Journal of Biogeography 30, pp. 331–351. Milstein P le S, Wolff SW 1987 The over-simplification of our “francolins”. South African Journal of Wildlife Research Supplement 1, pp. 58-65. Mlikovsky J 1989 A new guineafowl (Aves: Phasianidae) form the late Eocene of France. Annalen des Naturhistorischen Museums in Wien 90, pp.63-66. Mobley JA, Prum RO 1995 Phylogenetic relationships of the Cinnamon Tyrant, Neopipo cinnamomea, to the tyrant flycatchers (Tyrannidae). The Condor 97, pp. 650-662. Moore WS, Sheldon FH, Steadman DW, Witt CC, Yuri T 2008 A phylogenomic study of birds reveals their evolutionary history. Science 320, pp. 1763-1768. Morton ES 1975 Ecological sources of selection of selection on avian sounds. American Naturalist 109, pp. 17-34. Morony JJ, Bock WJ, Farrand J 1975 Reference List of the Birds of the World. American Museum of Natural History, New York. Moum T, Johansen S, Erikstad KE, Piatt JF 1994 Phylogeny and evolution of the auks (subfamily Alcinae) based on mitochondrial DNA sequences. Proceedings of the National Academy of Sciences of the United States of America 91, pp. 7912-7916. Mourer-Chauvire C, Pickford M, Senut B 2011 The first Palaeogene galliform from Africa. Journal of Ornithology 152, pp. 617-622. (Müller PLS, 1776) Pternistis afer afer. Natursystems Supplements, p. 129. Myers JA 1917 Studies on the syrinx of Gallus domesticus. Journal of Morphology 349 20(1), pp. 165-215. Neumann, 1898 Pternistis afer melanogaster. Journal of Ornithology 46(2, pl.3, f.1.), p. 299. Neumann, 1902 Francolinus clappertoni nigrosquamatus. Ornithologische Monatsberichte 10, p. 8. Neumann, 1904 Francolinus castaneicollis gofanus. Journal of Ornithology 52, p. 353. Neumann, 1907 Francolinus clappertoni heuglini. Ornithologische Monatsberichte 15, p. 199. Neumann, 1907 Francolinus icterorhynchus emini. Ornithologische Monatsberichte 15, p. 98. Neumann, 1907 Francolinus icterorhynchus ugandensis. Ornithologische Monatsberichte 15, p. 199. Neumann, 1908 Scleroptila levaillantoides pallidior. Bulletin of the British Ornithologists´ Club 21, p. 45. Neumann, 1908 Francolinus levaillanti benguellensis. Bulletin of the British Ornithologists' Club 21, p. 44. Neumann, 1908 Scleroptila shelleyi whytei. Bulletin of the British Ornithologists´ Club 21, p. 76. Neumann, 1915 Francolinus bicalcaratus adamauae. Ornithologische Monatsberichte 23, p. 73. Neumann, 1920 Pternistis afer krebsi. Journal of Ornithology 68, p. 78. Neumann, 1920 Francolinus gariepensis ludwigi. Journal of Ornithology 68, p. 79. Neumann, 1928 Francolinus clappertoni testi. Journal of Ornithology 76, p. 784. Newman K 2002 Newman’s Birds of Southern Africa. Struik Publishers, Cape Town. 350 Nylander JAA, Ronquist F, Huelsenbeck JP, Nieves-Aldrey J 2004. Bayesian phylogenetic analysis of combined data. Journal of Systematic Biology 53(1), pp. 47-67. Nylander JAA, Olsson U, Alström P, Sanmartín I 2008 Accounting for Phylogenetic Uncertainty in Biogeography: a Bayesian approach to Dispersal-Vicariance analysis of the Thrushes (Aves: Turdus). Systematic Biology 57(2), pp. 257-268. Ogilvie-Grant, 1890 Francolinus griseostriatus. Ibis (pl.10), p. 349. Ogilvie-Grant, 1890 Francolinus shelleyi shelleyi. Ibis, p. 348. Ogilvie-Grant, 1891 Francolinus clappertoni gedgii. Ibis, p. 124. Ogilvie-Grant, 1891 Pternistis jacksoni jacksoni. Ibis, p. 123. Ogilvie-Grant, 1891 Francolinus psilolaema elgonensis. Ibis, p. 126. Ogilvie-Grant, 1891 Francolinus streptophora. Ibis, p. 126. Ogilvie-Grant, 1892 Francolinus coqui buckleyi. Ibis, p. 41. Ogilvie-Grant, 1892 Francolinus shelleyi uluensis Ibis, p. 44. Ogilvie-Grant, 1892 Francolinus clappertoni sharpii. Ibis p.47. Ogilvie-Grant, 1896 Francolinus levaillantii crawshayi. Ibis (pl. 12), p.482. Ogilvie-Grant, 1897 Francolinus levaillantii kikuyuensis. Bulletin of the British Ornithologists´ Club 6, p. 23. Ogilvie-Grant, 1902 Francolinus bicalcaratus thornei. Bulletin of the British Ornithologists´ Club 13, p. 22. Olson SL 1974 Telecrex restudied: a small Eocence guineafowl. Wilson Bulletin 86, pp. 246-250. Oustalet, 1892 Francolinus icterorhynchus dybowski. Le Naturaliste 6(2), p. 232. Päckert M, Martens J, Kosuch J, Nazarenko A, Veith M 2003 Phylogenetic signal in 351 the song of crests and kinglets (Aves: Regulus). Evolution 57(3), pp. 616-629. Pagel M 1999 The maximum likelihood approach to reconstructing ancestral character states of discrete characters on phylogenies. Systematic Biology 48, pp. 612– 622. Payne RB1971 Duetting and chorus singing in African birds. Ostrich Supplementary 9, pp. 125-146. (Peters, 1854) Pternistis afer humboldtii. Monatschrifte der Deutschen Akademie der Wissenschaften zu Berlin, p. 134. Peters JL 1934 Check-list of Birds of the World. Vol II. Harvard University Press, Cambridge. Phillimore AB 2010 Subspecies origination and extinction in birds. The American Ornithologists’ Union 67, pp. 42-53. Posada D, Crandall KA 1998 Modeltest: testing the model of DNA substitution. Bioinformatics 14, pp. 817-818. Posada D, Buckley TR 2004 Model selection and model averaging in phylogenetics: advantages of Akaike information criterion and Bayesian approaches over likelihood ratio tests. Systematic Biology 53, pp. 793-808. Pough FH 1990 Vertebrate life. MacMillan, New York. Prager EM, Wilson AC 1976 Congruency of phylogenies derived from different proteins. A molecular analysis of the phylogenetic position of cracid birds. Journal of Molecular Evolution 9, pp. 45-57. Price JJ, Lanyon SM 2002 Reconstructing the evolution of complex bird song in the oropendalas. Evolution 56(7), pp. 1514-1529. Primmer CR, Borge T, Lindell J, Saetre GP 2002 Single nucleotide polymorphism 352 characterization in species with limited available sequence information: high nucleotide diversity revealed in the avian genome. Molecular Ecology 11, pp. 603-612. Pruett CL, Winker K 2010 Alaska Song Sparrows (Melospiza Melodia) demonstrate that genetic marker and method of analysis matter in subspecies assessments. The American Ornithologists’ Union 67, 162-171. Prum RO 1992 Syringeal morphology, phylogeny, and evolution of the Neotropical manakins (Aves: Pipridae). American Museum Novitates 3043. Prum RO, Lanyon WE 1989 Monophyly and phylogeny of the Schiffornis group (Tyrannoidea). The Condor 91, pp. 444-461. Rambaut A, Drummond AJ 2007 Tracer v1.4. [http:// tree.bio.ed.ac.uk/software/tracer/]. Ree RH, Smith SA 2008 Maximum likelihood inference of geographic range evolution by dispersal, local extinction and cladogenesis. Systematic Biology 57(1), pp. 4-14. Reichenbach, ["1852"]1853 Ortygornis. Avium systema natural das natürliche System der Vogel mit hundert Tafeln grosstentheils Original-Abbildungen der bis jetzt entdecken fast zwölfhundert typischen Formen, p. 28. Reichenow, 1885 Pternistes afer böhmi. Journal of Ornithology 33, p. 465. Reichenow, 1887 Francolinus hildebrandti fischeri. Journal of Ornithologist 67, p. 51. Reichenow, 1887 Pternistis rufopictus. Journal of Ornithology 35(177), p. 52. Reichenow, 1889 Francolinus stuhlmanni. Journal of Ornithology, p. 270. Reichenow, 1908 Francolinus nobilis nobilis. Ornithologische Monatsberichte 16, p. 81. 353 Reichenow, 1909 Pternistis afer harterti. Ornithologische Monatsberichte 17, p. 41. Reichenow, 1909 Pternistis afer harterti. Ornithologische Monatsberichte 17, p. 41. Reichenow, 1909 Pternistes afer intercedens. Ornithologische Monatsberichte 17, p. 88. Reichenow, 1912 Francolinus lathami schubotzi. Journal of Ornithology 60, p.320. Reichenow, 1919 Francolinus hildebrandti grotei. Journal of Ornithology 67, p. 334. Remsen Jr. JV 2010 Subspecies as a meaningful taxonomic rank in avian classification. The American Ornithologists’ Union 67, pp. 62-78. Roberts A 1924 Synoptic check list of the birds of South Africa. Annals of the Transvaal Museum. Vol. X. South Africa. (Roberts, 1929) Francolinus swierstrai. Annals of the Transvaal Museum 13, p. 72. Roberts, 1922 Dendroperdix. Annals of the Transvaal Museum 8, p. 194. Roberts, 1924 Francolinus sephaena zuluensis. Annals of the Transvaal Museum 10, p. 78. Roberts, 1924 Francolinus sephaena thompsoni. Annals of the Transvaal Museum 10, p. 78. Roberts, 1924, Francolinus afer notatus, A. “Synoptic Check-list of Birds of South Africa.” Annals of the Transvaal Museum 10, pp. 89-125. (Roberts, 1928) Francolinus hartlaubi bradfieldi. Annals of the Transvaal Museum 12, p. 292. (Roberts, 1928) Francolinus hartlaubi ovambensis. Annals of the Transvaal Museum 12, p. 293. Roberts, 1931 Pternistis afer lehmanni. Annals of the Transvaal Museum 14, p. 238. Roberts, 1932 Francolinus sephaena chobiensis. Annals of the Transvaal Museum 16, p. 21. 354 Roberts, 1932 Francolinus jugularis cunenensis. Annals of the Transvaal Museum 15, p. 22. Roberts, 1932 Pternistis swainsoni gilli. Annals of the Transvaal Museum 15, p. 23. Roberts, 1932 Francolinus gariepensis kalaharica. Annals of the Transvaal Museum 16, p. 22. Roberts, 1932 Francolinus gariepensis langi. Annals of the Transvaal Museum 16, p. 22. Roberts, 1932 Francolinus sephaena mababiensis Annals of the Transvaal Museum 16, p. 22. Roberts, 1931 Pternistis swainsoni damarensis. Annals of the Transvaal Museum 14, p. 238. Roberts, 1932 Pternistis swainsoni chobiensis. Annals of the Transvaal Museum 15, p. 23. Roberts, 1932 Pternistis afer cunenensis. Annals of the Transvaal Museum 15, p. 22. Roberts, A 1940 The Birds of South Africa. 1st ed. Riverside Press, Edinburgh. Roberts A, 1947 Pternistis cooperi. A new species of Pternistis from Salisbury, S. Rhodesia. The Ostrich 18, p. 197. Ronquist F 1996 DIVA, version 1.1. Computer program and manual available by anonymous FTP from Uppsala University, Uppsala, Sweden. Available at: ftp.uu.se or ftp.systbot.uu.se. Ronquist F 1997 Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Systematic Biology 45, pp.195-203. Ronquist F, Huelsenbeck JP 2003 MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, pp.1572-1574. 355 Rosen DE 1978 Vicariant patterns and historical explanation in biogeography. Systematic Zoology 27, pp. 159-188. Rothschild, 1902 Francolinus coqui angolensis. Bulletin of the British Ornithologists' Club 12, p. 76. (Rüppell, 1835) Francolinus levaillantoides gutturalis. Neue Wirbelthiere zu der Fauna von Abyssinien gehörig, p. 13. (Rüppell, 1835) Francolinus erckelii erckelii. Neue Wirbelthiere zu der Fauna von Abyssinien gehörig (pl.6), p. 12. Salvadori, 1888 Francolinus castaneicollis castaneicollis. Annali del Museo Civico di Storia Naturale di Genova 26, p. 542. Salvadori, 1888 Francolinus sephaena spilogaster. Annali del Museo Civico di Storia Naturale di Genova 26, p. 541. Salvadori, 1898 Francolinus castaneicollis bottegi. Annali del Museo Civico di Storia Naturale di Genova 38, p. 652. Sande E, Dranzoa C, Wegge P, Carroll JP 2009 Home ranges and survival of Nahan’s Francolin Francolinus nahani in Budongo Forest, Uganda. African Journal of Ecology 47, pp. 457-462. Schluter, Price T, Mooers AØ, Ludwig D 1997 Likelihood of ancestor states in adaptive radiation. Evolution 51, pp. 1699-1711. Sclater WL, 1921 Pternistis afer swynnertoni. Bulletin of the British Ornithologists´ Club 41, p. 134. Sclater WL 1927 Francolinus levaillantoides archeri. Bulletin of the British Ornithologists´ Club 48, p. 51. Sclater WL, 1927 Francolinus coqui maharao. Bulletin of the British Ornithologists´ 356 Club 48, p. 51. Sclater WL, 1932 Francolinus coqui lynesi. Bulletin of the British Ornithologists' Club 62, p. 143. (Scopoli, 1786) Francolinus pintadeanus pintadeanus. Deliciae florae faunae insubricae 2, p. 93. Scotese CR 2001 Atlas of Earth History. Vol. 1. Paleogeography. PALEOMAP project, Arlington, Texas. Seller TJ 1983 Control of sound production in birds. Pp. 93-124. In: Bioacoustics: a Comparative Approach (Lewis B, ed). Academic Press, London. Serle W 1957 A contribution to the ornithology of the eastern region of Nigeria. Ibis 99, pp. 371-418, 628-685. Sharpe, 1897 Francolinus levaillantoides lorti. Bulletin of the British Ornithologists´ Club 6, p. 47. Shelley, 1894 Pternistis hildebrandti johnstoni. Ibis, p. 24. Shen Y-Y, Liang, L, Sun Y-B, Yue B-S, Yang Xiao-J, Murphy RW, Zhang Y-P 2010 A mitogenomic perspective on the ancient, rapid radiation in the Galliformes with an emphasis on the Phasianidae. BMC Evolutionary Biology 10, pp 132 doi:10.1186/1471-2148-10-132. Shields GF, Wilson AC 1987 Calibration of mitochondrial DNA evolution in geese. Journal of Molecular Evolution 24, pp. 212-217. Sibley CG, Ahlquist JE 1985 The relationships of some groups of African birds, based comparisons of the genetic material, DNA. Pp. 115-161. In: Proceedings of the International symposium on African vertebrates (KL Schuchmann, ed). Museum Alexander Koenig, Bonn. 357 Sibley CG, Ahlquist JE 1990 Phylogeny and Classification of Birds. Yale University Press, New Haven. Sibley CG, Monroe Jr. BL 1990 Distribution and Taxonomy of Birds of the World. Yale University Press, New Haven. Sinclair I, Ryan P 2003 Birds of Africa South of the Sahara. Struik publishers, Cape Town. Smith A, 1833 Francolinus natalensis natalensis. South African Quarterly Journal 2(1), p. 48. (Smith A, 1836) Francolinus coqui coqui. Report of the expedition for exploring central Africa, p. 55. Smith A, 1836 Francolinus levaillantoides levaillantoides. Report of the expedition for exploring central Africa, p. 55. (Smith A, 1836) Francolinus sephaena sephaena. Report of the expedition for exploring central Africa, p. 55. (Smith A, 1836) Pternistis swainsonii swainsonii. Report of the expedition for exploring central Africa, p. 54. Sorenson MD, Ast JC, Dimcheff DE, Yuri T, Mindell DP 1999 Primers for a PCRbased approach to mitochondrial genome sequencing in birds and other vertebrates. Molecular Phylogenetics and Evolution 12, pp. 105-114. Snow DW 1978 An Atlas of Speciation in African Non-passerine birds. British Museum, London. Staden R, Judge DP, Bonfield JK 2003 Analysing sequences using the Staden package and EMBOSS. Introduction to Bioinformatics. A Theoretical and Practical Approach. (Stephen A, Krawetz SA, Womble DD, eds). Human 358 Pressinc., Totawa, NJ. Stamatakis A 2006 RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, pp. 26882690. Stamatakis A, Hoover P, Rougemont J 2008 A rapid bootstrap algorithm for the RAxML web-servers. Systematic Biology 75, pp. 758-771. Stephens, 1819 Francolinus. General Zoology [Shaw] 11(pt.2), p. 316. Steven P, Paulsen F, Tillmann B 2000 Orcein-Picroindigocarmine - a new multiple Stain. Archives of Histology and Cytology 63(4), pp. 397-400. Stevenson T, Fanshawe J 2002 Field Guide to the Birds of East Africa. T & AD Poyser, London Stidham TA 2008 The first fossil of the Congo peafowl (Galliformes: Afropavo). South African Journal of Science 104, pp. 511-512. Stoneham, 1930 Pternistis leucoscepus tokora. Bateleur 2, p. 113. Stresemann E, 1937 Francolinus coqui hoeschianus subsp. nova. Ornithologische Monatsberichte 45(2), pp. 66-67. Stresemann E, 1939 Francolinus hartlaubi crypticus. Zwei neue Vogelrassen aus Südwest-Afrika. Ornithologische Monatsberichte 47, pp. 61-62. Swainson, 1837 Chaetopus. On the natural history and classification of birds[D Lardner]. The Cabinet Cyclopaedia 2(92), p. 344. Swofford DL 2002 PAUP*: Analysis Using Parsimony, Version 4. Sinaeur Associates, Sunderland, Massachusetts. (Temminck, 1815) Francolinus gularis. Histoire naturelle generale des pigeons et des gallinaces 3, pp. 401, 731. 359 Temminck, 1854 Francolinus ahantensis ahantensis. Bijdragen Tot De Dierkunde 1(pt6), p. 49. Templeton AR 1989 The meaning of species and speciation: a genetic perspective. Pp 3-27. In: Speciation and its consequences (Otte D, Endler JA, eds). Sinauer Associates, Sunderland, Massachusetts. Tsukahara N, Yang Q, Sugita S 2008 Structure of the syringeal muscles in Jungle Crow (Corvus macrorhynchus). Anatomical Science International 83, pp. 152158. (Valenciennes, 1825) Francolinus levaillantii levaillantii. Dictionnaire des Sciences Naturelles 38, p. 441. Van Someren, 1926 Francolinus coqui ruahdae. Journal of the East Africa and Uganda Natural History Society 25, p. 34. Van Niekerk JH 2010 Vocal behaviour of Crested Francolin Ortygornis sephaena in response to playback calls. Ostrich - Journal of African Ornithology 81(2), pp. 149-154. Van Someren VGL, 1938 Francolinus africanus macarthuri. A new race of Grey-wing Francolin from Kenya Colony. Bulletin of the British Ornithologists´ Club 59, p. 7. Van Someren VGL, 1939 Francolinus squamatus chyuluensis. Reports on the Corydon Museum Expedition to the Chyulu Hills. 2. The birds of the Chyulu Hills. Journal of the East Africa and Uganda Natural History Society 14, pp. 15-129. Verheyen R 1956 Contribution a l’anatomie et a la systematique des Galliformes. Bulletin de l'Institut Rroyal des Sciences Naturelles de Belgique 32, pp. 1-24. 360 Voelker G, Outlaw RK, Bowie RCK 2010 Pliocene forest dynamics as a primary driver of African bird speciation. Global Ecology and Biogeography 19, pp. 111-121. Von Erlanger, 1904 Pternistes leucoscepus höltemulleri. Ornithologische Monatsberichte 12, p. 98. Von Erlanger, 1904 Pternistes leucoscepus muhamed-ben-abdullah. Ornithologische Monatsberichte 12, p. 98. Von Erlanger, 1905 Francolinus psilolaema ellenbecki. Journal of Ornithology 63, p. 151. Von Heuglin, 1873 Francolinus sephaena schoanus. Systematische Uebersicht der Vögel Nord-Ost-Afrika's mit Einschluss der arabischen Kueste des Rothen Meeres und der Nil-Quellen-Länder 2(pt.l, pl.29), p. 891. Wagler, 1832 Pternistis. Isis von Oken 25, col. 1229. Wang N, Kimball RT, Braun EL, Liang B, Zhang Z 2013 Assessing phylogenetic relationships among Galliformes: a multigene phylogeny with expanded taxon sampling in Phasianidae. Public Library of Science One 8(5), pp. e64312. doi:10.1371/journal.pone.0064312. Waterhouse, 1838 Francolinus adspersus adspersus. An Expedition of the discovery into the Interior of Africa[Alexander] 2 App., p. 267. Wetmore A 1960 A classification for the birds of the world. Smithson. Smithsonian Miscellaneous Collections 139(1), pp.1-37. Whistler H, 1941 Francolinus pondicerianus ceylonensis. Recognition of new subspecies of Birds in Ceylon. Ibis 5(14), pp. 319-320. White CMN, 1944 Francolinus albogularis meinertzhageni. A new race of 361 Francolin from Northern Rhodesia. Bulletin of the British Ornithologists' Club 65, pp. 7-8. White CMN, 1944 Francolinus levaillantii clayi. A new race of Scrub Robin and a New race of Red-winged Francolin from Northern Rhodesia. Bulletin of the British Ornithologists´ Club 64, pp. 49-50. White CMN, 1945 Francolinus coqui kasaicus. The ornithology of the KaondeLunda province, Northern Rhodesia. Part III. Systematic List. Ibis 87, pp. 309345. White CMN, 1947 Pternistis swainsonii lundazi. Bulletin of the British Ornithologists´ Club 67, pp.72-73. Winker K 2010 Subspecies represent geographically partitioned variation, a gold mine of evolutionary biology, and a challenge for conservation. The American Ornithologists’ Union 67, pp. 6-23. Wolters HE 1975-82 Die Vogelarten der Erde. Paul Parey, Hamburg & Berlin. Wortley AH, Scotland RW 2006 The effect of combining molecular and morphological data in published phylogenetic analyses. Systematic Biology 55 (4), pp. 677-685. Zarudny, 1906 Francolinus francolinus bogdanovi. Ornithologische Monatsberichte 14, p. 151. Zarudny & Härms, 1913 Francolinus pondicerianus mecranensis. Ornithologische Monatsberichte 21, p. 53. Zarudny & Härms, 1913 Francolinus francolinus arabistanicus. Ornithologische Monatsberichte 21, p. 54. Zedlitz, 1913 Francolinus sephaena jubaensis. Ornithologische Monatsberichte 21, p. 362 59. Zimmer KJ, Robbins MB, Kopuchian C 2008 Taxonomy, vocalizations, syringeal morphology and natural history of Automolus roraimae (Furnariidae). Bulletin of the British Ornithologists' Club 128(3), pp 14-33. Zink RM 2004 The role of subspecies in obscuring avian biological diversity and misleading conservation policy. Proceedings of the Royal Society of London, Series B 271, pp. 561-564. 363