ITUTION NOILALILSNI NVINOSHLINS S3IYvVUaIT LIBRARIES

\ si

syugi7 LIBRARIES SMITHSONIAN INSTITUTION

SMITHSONIAN INSTITUTI

« “s/

Mae

f Up

NVINOSHLINS

SMITHSONIAN

NVINOSHLINS

SMITHSONIAN

NVINOSHLIWS |

».

NOILNLILSNI NVINOSHLINS S3Iluvug

NOILNLILSNI

LIBRARIES

i a

TUTION NOILNLILSNI NVINOSHLINS S3IYVYEIT LIBRARIES SMITHSONIAN

S3INVYGIT LIBRARIES SMITHSONIAN

INSTITUTION NOILNLILSNI

S - i

= wo a Lp

- 2 : EE

i = > Y Ye

- s ci re “ a ‘

z m m i

_LIBRARIES INSTITUTION NOILANLILSNI NVINOSHLINS S31uYVds

‘. ® : 3

NVINOSHLIWS

NVINOSHLINS

NVINOSHLIWS

SMITHSONIAN

Ww)

a os

TUTION NOILNLILSNI_ NVINOSHLINS S3IYVYaIT LIBRARIES SMITHSONIAN _

= w z rf ; Z

n aes W w

= o sat w ~

= 5 =—4 <x |

S eS WS & = Se 5 4)

pases (OP Saye CS oe SG Ntnosn

= [ | = =

WUdiT LIBRARIES oon SCN a EN Nea “NVINOSHLINS Saluvua

S S S

5 red = =

a 4 b> « - = :

Z y = os <OESY

TUTION NOILALILSNI S3IYVYdIT_ LIBRARIES, SMITHSONIAN INSTITUTI

NVINOSHLINS S3I¥YVYSIT LIBRARIES SMITHSONIAN

NVINOSHLINS S31a¥vugit

SMITHSONIAN

NVINOSHLIWS

SMITHSONIAN

yugia LIBRARIES SMITHSONIAN INSTITUTION

& t

». .

NOILNLILSNI NVINOSHLINS S31Y¥Vvug

SMITHSONIAN

Z a oh Ss sae 7)

” aa 8 4 ag Wg ANS a AS SN

s, ae. ~% oc 4+ ays & G

a ty < 4 ¥ WN se

c& Z Y “Os ( ez »

4 “ “fy a a “, &

ro) a G = re) a

Fa me a z J

bit a

TUTION NOILNLILSNI NVINOSHLINS S3tuYvugit LIBRARIES SMITHSONIAN

INSTITUTION NOILNLILSNI

saiuvugi

INSTITUTION

SJiyvyugiy

INSTITUTION

iyvudil LIBRARIES

‘9

SMITHSONIAN INSTITUTION NOILOALILSNI NVINOSHLINS S31uVvuyd

NOSHLINS S3iuvyd)

THSONIAN

NOSHLIWS

THSONIAN

zy,

Lip

NOSHLIWS

INLILSN!I NVINOSHLINS S3!1YVYNGIT LIBRARIES SMITHSONIAN INSTITUTION NOILNLIL

<< = en = < “S N =

= 3 z a z Sa

= z FE Zz = Zz

, = i ans. . 2 = a

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S31YNUVYUS th.

= _ 2 4 AWN = =

< . — <x + NX \ <x a

‘ Cc Cc |

On DS. > Ye ro) = re) dig o

, oil = _ z ~! =

ALILSNI _NVINOSHLIWS 34 iyVvuyg ine LIBRARI ES_ INSTITUTION i NOILALIL

—- ro) = Oo a fe)

rr 2 = rs = = ss

> we > = i> =

4b 2 = a ran i a =

RARI ES SMITHSONIAN INSTITUTION NOILNLILSNI_ NVINOSHLINS, $3 i4y4vyugit _t! B RAR

= < = Wy < = evi ta

= z 4 iy, = = <

FA 9 ; yj . z So SY

\ 9 x °O GS i, = oO x WS

WS z = 2 & 4 P as = 2) = ‘

7S > = S >" = > =

ow 2 ow = ri

AALILSNI_ NVINOSHLINS S3iuVvug ret BRARIES SMITHSONIAN _INSTITUTION NOILNLI

Si n = ” i 7 a

‘hl = ce az ead . = ce

+ _< af < sS —- <x

S a = eae = ce

ro) \ = ro) cad ro) S

za ad a = =a |

SMITHSONIAN INSTITUTION NOIJLNLILSNI NVINOSHLINS S31uvWHNgIT LIBRAR

ro oc Fe: ae z ie

= = 2 a fe) o

5. a = 3) = 2

FE > aa = a - a

a a ra a ra on

= mi ” m ” m

= wo < w” . £ ”

NYINOSHLINS S3IN¥VYUSIT LIBRARIES SMITHSONIAN NOILALI.

= ~” = 28s ~” z<. Ww

< = < = < =

= = = ca = 4

: g 3 2 : a a

& a = 2 ‘= S

Z = > = et s >

” z ” a ae Tp) ae

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3INVYHSII

” = * F a 2

= a a w = 7)

. 3 a = o =

< 3 Py = = = e

aj — 3S = Ba eg. S

_! z a 2 me ly z

ALILSNI_NVINOSHLIWS 34 iyVvug are LIBRARI ES_ SMITHSONIAN INSTITUTION _ NOILALI

—_ e re) — oO i = oS

es = = = 2S =

> i > = > i

a = a = 2 E

RAR! ES SMITHSONIAN INSTITUTION NOILNLILSNI_ NVINOSHLINS, S3 iyvVyugit at! B RAR

= aay = < < as |

=a z = = = z Aw

he a te : 2

a = re) TS oO = aN YA

z = =. = = E Ae

(ee eed aya yeas

) a ae r Lae ; ee ae { 4 ‘ ‘ 2 ey ' 2

7. 68

ANNALS OF THE ANNALE VAN DIE

SOUTH AFRICAN MUSEUM SUID-AFRIKAANSE MUSEUM

VOLUME 76 BAND 76

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

VOLUME 76 BAND

HE TRUSTEES OF THE DIE TRUSTEES VAN DIE

SOUTH AFRICAN MUSEUM SUID-AFRIKAANSE MUSEUM

CAPE TOWN KAAPSTAD

1978

% SET, PRINTED AND BOUND IN THE REPUBLIC OF SOUTH AFRICA BY

THE RUSTICA PRESS (PTY.) LTD., WYNBERG, CAPE

764

LIST OF CONTENTS

CLUVER, M. A.

The skeleton of the mammal-like reptile Cistecephalus with evidence for a fossorial

mode of life. (Published September 1978.)

CrowE, T. M.

The evolution of guinea-fowl (Galliformes, Phasianidae, Numidinae), taxonomy,

phylogeny, speciation and biogeography. (Published September 1978.)

Day, J.

Southern African Cumacea. Part 3. Families Lampropidae and Ceratocumatidae.

(Published August 1978.) - ne ;

GRIFFITH, J.

A fragmentary specimen of Saurichthys sp. from the Upper Beaufort Series of

South Africa. (Published September 1978.) :

HENDEY, Q. B.

Preliminary report on the Miocene vertebrates from Arrisdrift, South West Africa.

(Published July 1978.) : 2

HENDEY, Q. B.

Late Tertiary Hyaenidae from Langebaanweg, South Africa, and their relevance

to the phylogeny of the family. (Published September 1978.) ..

HENDEY, Q. B.

Late Tertiary Mustelidae (Mammalia, Carnivora) from Langebaanweg, South

Africa. (Published September 1978.) : ae oF i

MCLACHLAN, A. & Moore, C. G.

Three new species of Harpacticoida (Crustacea, Copepoda) from sandy beaches

in Algoa Bay, South Africa, with keys to the genera Arenosetella, ae

Leptastacus and Psammastacus. (Published August 1978.) 3 :

Moore, C. G. see MCLACHLAN, A.

RAu, R. E.

The development of Xenopus - Rose & Hewitt pone ges ae

October 1978.)

WOOLDRIDGE, T.

Two new species of Gastrosaccus (Crustacea, Mysidacea) from sandy beaches in

Transkei. (Published September 1978.) : we

Page

2A3

299

265

329

191

247

309

VOLUME 76 PART 1 JULY 1978 | ISSN 0303-2515

po 7.68

OF THE SOUTH AF RICAN

MUSEUM

INSTRUCTIONS TO AUTHORS

1. MATERIAL should be original and not published elsewhere, in whole or in part.

2. LAYOUT should be as follows:

(a) Centred masthead to consist of

Title: informative but concise, without abbreviations and not including the names of new genera or species

Author’s(s’) name(s)

Address(es) of author(s) (institution where work was carried out)

Number of illustrations (figures, enumerated maps and tables, in this order)

(b) Abstract of not more than 200 words, intelligible to the reader without reference to the text

(d) Introduction

(e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents

(f) Summary, if paper is lengthy

(g) Acknowledgements

(h) References

@i) Abbreviations, where these are numerous

3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced

with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of

legends for illustrations should be typed separately, their positions indicated in the text. All

pages should be numbered consecutively.

Major headings of the paper are centred capitals; first subheadings are shouldered small

capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered

italics. Further subdivisions should be avoided, as also enumeration (never roman numerals)

of headings and abbreviations.

Footnotes should be avoided unless they are short and essential.

Only generic and specific names should be underlined to indicate italics; all other marking

up should be left to editor and publisher.

4. ILLUSTRATIONS should be reducible to a size not exceeding 12 « 18 cm (19 cm including

legend); the reduction or enlargement required should be indicated; originals larger than

35 x 47 cm should not be submitted; photographs should be rectangular in shape and final

size. A metric scale should appear with all illustrations, otherwise magnification or reduction

should be given in the legend; if the latter, then the final reduction or enlargement should be

taken into consideration.

All illustrations, whether line drawings or photographs, should be termed figures (plates

are not printed; half-tones will appear in their proper place in the text) and numbered in a

single series. Items of composite figures should be designated by capital letters; lettering of

figures is not set in type and should be in lower-case letters.

The number of the figure should be lightly marked in pencil on the back of each illustration.

5. REFERENCES cited in text and synonymies should all be included in the list at the end of

the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable):

(a) Author’s name and te of publication given in text, e.g.:

‘Smith (1969) describes .

‘Smith (1969: 36, fig. 16) aes

“As described (Smith 1969a, 1969b; fies is a

‘As described (Haughton & Broom 1927).

‘As described (Haughton et al. 1927)...’

Note: no comma separating name and year

Dagination indicated by colon, not p.

names of joint authors connected by ampersand

et al. in text for more than two joint authors, but names of all authors given in list of references.

(b) Full references at the end of the paper, arranged alphabetically by names, chronologically

within each name, with suffixes a, 5, etc. to the year for more than one paper by the same

author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a).

For books give title in italics, edition, volume number, place of publication, publisher.

For journal article give title of article, title of journal in italics (abbreviated according to the World list o,

scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part

number (only if independently paged) in parentheses, pagination (first and last pages of article).

Examples (note capitalization and punctuation)

BuLLouUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FISCHER, P. —H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140.

Fiscuer, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs

Zool. exp. Zen. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon.

Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean.

Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische

und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270.

Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

(continued inside back cover)

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 76 Band

July 1978 Julie

Part 1 Deel

PRELIMINARY REPORT ON

THE MIOCENE VERTEBRATES FROM ARRISDRIFT,

SOUTH WEST AFRICA

Q. B. HENDEY

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material

becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid

van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000

OUT OF PRINT/UIT DRUK

1, 203,528), 30-2), 4-5, 8, ts 53 eo

6 tpi), 704), 8, 902,79), 103):

1012, 567; t=p.8); 15425): 240), 27-310 3)e 2s

Kopieregnavrae na die Suid-Afrikaanse Museum

ISBN 0 908407 46 7

Printed in South Africa by In Suid-Afrika gedruk deur

The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk.,

Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap

PRELIMINARY REPORT ON THE MIOCENE VERTEBRATES FROM

ARRISDRIFT, SOUTH WEST AFRICA

Q. B. HENDEY

South African Museum, Cape Town

(With 13 figures and 11 tables)

[MS. accepted 14 March 1978]

ABSTRACT

At least 28 vertebrate species, of which 22 are mammals, are recorded from the early

middle Miocene (c. 16 m.y. old) fossil occurrence at Arrisdrift on the Orange River in South

West Africa. The material postdates Miocene vertebrates previously recorded from the

Namib desert. The mammals include at least 3 new species (a hyracoid, a palaeomerycid and

an ochotonid), while there are at least 8 genera represented which have not hitherto been

known in southern Africa. Austrolagomys simpsoni Hopwood, 1929, is referred to Kenyala-

gomys Whitworth, 1954, and Prohyrax is placed in the Pliohyracinae, a group which apparently

had its origins in southern Africa.

CONTENTS

PAGE

MpPROG UC OU Gre he ee mt ee eh Ne Ween 1

Other Miocene vertebrate occurrences in the Namib desert . 3

Mae tossilevertebrates from Arisdrift .. . . = . «°. 8

PRPCTOlMINelOCCUEEENCE® oan) ie ee ee wee ee Oe

NCOCMVIrOMIMent “4 (aa ilsia = to od te 1 Rs ee ee | 4

PREMMOMICUOCIMEMES i. te. esa ey aes fe 6) ew 8

References 2 Reta Me icc oi eal Re ge i WO AU 38

INTRODUCTION

Until recently the only substantial information on southern African

Miocene terrestrial vertebrates came from several small fossil assemblages

collected in the southern Namib desert (Stromer 1926; Hopwood 1929;

Hamilton & Van Couvering 1977). The described material is limited in both

quality and quantity. The discovery of a new Miocene vertebrate locality in

terrace deposits at Arrisdrift on the Orange River further south in the same

region (Fig. 1) has proved an important addition to the local Miocene fossil

record (South African Journal of Science 1976; Corvinus & Hendey 1978).

The number of fossils already collected at Arrisdrift exceeds the combined

total from the other Namib desert occurrences, although the quality of the

specimens is not necessarily superior. The presence on the subcontinent of

several taxa has been revealed for the first time. These include the deinothere,

Prodeinotherium hobleyi (Harris 1977).

The fossils were discovered in a prospect pit (Pit 2 of Drill-line AD 8)

in deposits being investigated by the Consolidated Diamond Mines of South

Ann. S. Afr. Mus. 76 (1), 1978: 1-41, 13 figs, 11 tables.

D ANNALS OF THE SOUTH AFRICAN MUSEUM

West Africa (Pty) Ltd. Further fossiliferous deposit has since been exposed by

extending the original pit, but the limits of the occurrence have not been

established.

The material already prepared includes remains of at least 28 vertebrate

species, of which 22 are mammals (Table 1). Most have yet to be positively

identified and studies have so far been confined largely to cranial material,

which is much less common than postcranial bones. The condition of specimens

varies considerably, some being well preserved and reasonably complete, but

most having suffered post-mortem damage. The fossils occur in a poorly sorted

TABLE |

The vertebrates from Pit 2/AD 8 at Arrisdrift, South West Africa.

OSTEICHTHYES

gen. et sp(p). indet.

AMPHIBIA gen. et sp. indet.

REPTILIA

Squamata gen. et sp. indet.

Crocodilia ? Crocodylus niloticus

Chelonia . gen. et sp(p). indet.

AVES gen. et spp. indet.

MAMMALIA

Insectivora

Macroscelididae Myohyrax cf. oswaldi

Carnivora

Amphicyonidae .. Amphicyon cf. steinheimensis

Amphicyonidae or Ursidae .

?Ursidae . ?Hemicyoninae gen. & sp. indet.

?Felidae . ? Metailurus sp.

Mustelidae ? Ischyrictis sp.

indet. gen. & sp. indet.

Hyracoidea

Procaviidae Prohyrax n. sp.

Proboscidea

Gomphotherlidae . gen. et sp. indet.

Deinotheriidae Prodeinotherium hobleyi

Perissodactyla

Rhinocerotidae Dicerorhinus sp.

Artiodactyla

Suidae gen. et sp. indet.

Lopholistriodon moruoroti

Tragulidae Dorcatherium cf. pigotti

Palaeomerycidae Climacoceras sp. nov.

Bovidae gen. et sp. indet.

Pecora gen. et sp. indet.

Lagomorpha

Ochotonidae . Kenyalagomys sp. nov.

Rodentia

?Bathyergidae ? Bathyergoides sp.

Thryonomyidae Paraphiomys pigotti

indet. gen. et spp. indet.

gen. et sp. indet.

ee eS a

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 3

fluvial gravel and their imperfections are due mainly to their having been

transported by water in a high-energy environment. Only one instance is

recorded of skeletal elements occurring in articulation and, in addition to

disarticulation and fragmentation, many specimens are abraded and distorted.

Incrustations of gypsum have etched and even destroyed parts of some speci-

mens. Since the deposit incorporating the fossils is consolidated, power tools

were required during their excavation and this has caused further damage to

specimens. In spite of its shortcomings, the Arrisdrift fossil assemblage is

perhaps the most important one of Miocene age yet discovered in southern

Africa.

The purpose of the present report is to place on record some details of the

nature and number of specimens belonging to the various taxa already recog-

nized. With the exception of the deinothere teeth, none of the material has been

thoroughly studied, although such studies will be undertaken by various

authorities in the future. The geological investigation of the deposits in the

vicinity of Arrisdrift is being undertaken by employees of the mining company

prospecting the area.

The specimens discussed in this report are housed in the Department of

Cenozoic Palaeontology at the South African Museum, Cape Town. The full

catalogue numbers begin SAM-PQ., which identify the institution and depart-

ment concerned, but this lettering is omitted in the text and only the site prefix

(AD) and serial numbers of individual specimens are given. The full site

reference is Arrisdrift, Pit 2/AD 8.

OTHER MIOCENE VERTEBRATE OCCURRENCES IN THE

NAMIB DESERT

The first Miocene vertebrates from the southern Namib desert were

discovered during the First World War and were described in a series of papers

by Stromer (1922, 1923, 1924, 1926). This material came from three localities,

namely, Elisabethfeld, 38 km south of Liideritz; a borehole near Plant 4 of

the Kolonial Bergbau Gesellschaft, 20 km south of Liideritz (= Elisabeth Bay

Pan, see Greenman 1966); and Langental near Bogenfels, 80 km south of

Lideritz (Fig. 1).

Subsequently Hopwood (1929) described another small assemblage of

specimens from the same region, but his material lacks precise locality data

and was recorded as being from ‘south of Liideritz Bay’.

Little additional material was collected in the region in the decades which

followed. The South African Museum has an undescribed ruminant mandible

fragment (SAM-PQ-G 8356) from Bogenfels, which may be a synonym of

Stromer’s Langental locality. Some fragmentary material was collected by

Greenman (1966) from Fiskus and Grillental in the Elisabethfeld area. This

material is also in the South African Museum, but includes little that is

diagnostic.

ANNALS OF THE SOUTH AFRICAN MUSEUM

XPLANT 4

XxELISABETHFELD

ARRISDRIE

Alexander

)

Bay

Fig. 1. Location of Arrisdrift and other South West African Miocene fossil occurrences.

Hamilton & Van Couvering (1977) recently revisited the area and collected

more material from the various localities. They have reviewed and supple-

mented the original faunal lists and compared and contrasted a revised list

with others from early Miocene occurrences elsewhere in Africa.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 5

The Namib fossils are generally regarded as early Miocene (‘Burdigalian’)

in age, their source is usually recorded as ‘Namib desert’ or ‘South West Africa’,

and they are treated as if they were a single assemblage. The described material

is here listed as four separate assemblages (Table 2). Since each is limited in

size and each includes unidentified or incompletely identified taxa, the tases

for comparing them to each other, and to assemblages elsewhere, are limited.

The comments which follow are confined largely to the implications of indi-

vidual taxa in respect of the age of the assemblages. Other references to this

material are included in the discussions on the Arrisdrift fossils.

There is one species from Elisabethfeld, Metapterodon kaiseri, which

Savage (1965) believed to be represented in east Africa by specimens from

Karungu and Rusinga, which are between 18 and 20 m.y. old. Savage also

recognized a second species of Metapterodon from Rusinga, namely, M. zadoki.

Van Valen (1967: 252) found ‘that the two east African species distinguished

by Savage are much more similar to each other than are the east and Southwest

African forms of “‘“M. kaiseri”’. He concluded that the east African species

are more advanced than M. kaiseri and synonymized Metapterodon with

Pterodon. Although this material is problematical, Van Valen’s opinion suggests

that the Elisabethfeld species may predate its east African counterparts. Its

age might therefore be greater than 20 m.y., that is, “‘Aquitanian’ rather than

‘Burdigalian’ in terms of the European mammal age nomenclature (Van

Couvering 1972).

Another of the Elisabethfeld species, Myohyrax doederleini, was regarded

as a synonym of the east African M. oswaldi (Whitworth 1954; Patterson 1965),

a species which has a recorded age range of 18 to 22 m.y. (see Whitworth 1954;

Walker 1969; Van Couvering 1972). An ‘Aquitanian’ to ‘Burdigalian’ age is

therefore indicated. Since a Myohyrax resembling M. oswaldi is now also

recorded from Arrisdrift, this species may have survived beyond the ‘Burdi-

galian’ (see below), and appears to be of little use for relative dating purposes.

The Elisabethfeld Propalaeoryx austroafricanus is a primitive ruminant

which is likely to be broadly contemporaneous with the Rusinga P. nyanzae

(Whitworth 1958; Hamilton 1973), but the available material of these species

is too scanty to determine possible differences of temporal significance.

The only other identified species from Elisabethfeld are Parapedetes

namaquensis and Austrolagomys inexpectatus. They are not known elsewhere,

although there are related genera, Megapedetes and Kenyalagomys, recorded

from the early Miocene of east Africa, again from the 18 to 22 m.y. period.

According to MaclInnes (1957) Megapedetes is less specialised than

Parapedetes, but he did not regard this as indicating an age difference, ascribing

it instead to different evolutionary trends on two contemporary lineages. This

interpretation raises the question of whether assemblages such as those from

Elisabethfeld can be dated in a relative sense by comparing them with east

African assemblages. If MacInnes’s interpretation of the pedetids is correct,

then it follows that seemingly primitive taxa such as Pterodon kaiseri may

ANNALS OF THE SOUTH AFRICAN MUSEUM

‘JOpUL eIUSPOY

apuuayooU SAUIOJAdKAIDAD

1yjosid skwoldvsvg

1MdWUOdAJS SKWOLYADAD

sniqnp stuouowog

1Z]lAapan] SAWOJUBDUDIGE

SIMUAY saplodwolg

ijosid stwolydpsavg

SHLADIJAIJOIU SAPlOBsAIAY JV

sisuanbpuupu Sajapadvavpg

lMosduis sdwosvjpduay

snjpjoadxaul SAwospjosjsnp

snuvpdidfoosjsnp xdsoanjpdoig

snuvdidfoosjsno xAsoapjpdodg

‘JOPUL BIODOG

‘ds winlsayJDIAOG

snubvaidfo snpodyovig

"ds snsaoyroavnjododg ¢

SHUDILA{D SNAVOYIOUAX

"JOpUr e[AJORpOnIY

IUIaZUlay WnidsayloddyaDjg

“ds wnisayjsoydwuoy

SMIADIJAI] XDAKYOA

‘JOPUI BIOAIUICD/BJUOPOdID

1AdS1IDY UOPOdAl

12ZJ22q SaplodsayjodaJosg

1ZJaaq SaplosayjodAjosg

IpjoMso xb1‘AYOd PW

IpjoMso xDsAKyod yy

eSNOILIGGV

CNV SNOILV)D

THILNACI GHSIAdY

"j0doI SIY} (LL6]) SULIOANOD UeA WZ UOIIWEPT ,

‘JOPUI BIUIPOY

DIIDAY DIJAUIYIY

MIWOAJS JaJIapodp

quosduis stwospjosjsnp

‘JOPUI OBpI[NSeIy,

‘JOPUI oBpI[NseI

MAOgso xvAKYOAIW

lay 4apaop XDAKYOA AT

eZLIAdd iI

HO HLNOS

‘JOpuUL EUDPOY

sniqnp skuouowog

IZj1Aapan] SAWoOjUuDWIDIGE

SNUDIIA{D SAWOANIISOAN

SHIADIJAIJOIU SAPlOBsaA YO

"ds snyjousojnsosjs “JO

‘ds snsaoyroavnjvdolg ¢

Snuvd1dfo sndyojubupiq

‘JOPUL 9ePHOAD.OUINY

SMIADIJAI] XDAKYOAT

1ZJaaq Saplosayjodaosg

tIVLNHONVT

‘JOPUI BIUDPOY

IsMadpUuD SAULOIY *}9

SUNY Saplodwolg

SHUDIIA{D SAWOANIISOAN

SN1ADIJAIJOIU SAPlOBAIAY JOT

‘ds snyjousojnsoss “jo ¢

“JOpUT B[AJOVPONIYV

IpjoMso xbDAc yon |W

ta TOHHAOd F LNV Id

‘Hosop QIUeN] oY} JO vare sjoyussog/z}Jepn’y oy) Woy speuUUeU sUDD0IT], *Z AIAV I,

(6761) poomdoy ;

(9761) JOWIONS ;

sisuanbouwu sajapadvivg

snpjaadxau1 sdwuosvjosjsnp

SNUDILA{DOAJSND XXAOaDIDAOAG

‘Jopur elAjOeponiy

‘JOPUL VIOAIUILD/eIUOPOdID

14as1py Uoposajdvjayy

llajAapaop xbidyod py

td THHH LA VSITA

a ee

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA i

simply be conservative southern counterparts of ‘advanced’ east African species.

This problem arises again with the ochotonids. There are clear differences

between Austrolagomys and Kenyalagomys (MacInnes 1953), which may be

interpreted as indicating that the former is the more primitive (Cooke 1972).

In this instance the fact that the ochotonid represented at Arrisdrift (which

evidently does postdate both the Elisabethfeld and the east African ‘Rusinga-

type’ faunas) 1s a Kenyalagomys, may be an indication that Austrolagomys

really is a primitive and early form.

To sum up, there is some evidence to suggest that the Elisabethfeld fauna

represents a southern African equivalent of that of the European Aquitanian.

This conclusion is, however, tentative and should be re-examined if more

material becomes available and/or when the zoogeographic relationships of

east and southern African early Miocene faunas are better understood.

The Langental fauna includes five species which are known from the early

Miocene of east Africa. They are Bathyergoides neotertiarius, Paraphiomys

pigotti, Diamantomys luederitzi (Lavocat 1973), Xenochoerus africanus

(Wilkinson 1976) and Brachypotherium heinzelini (Heissig 1971). In addition,

Prohyrax tertiarius is a primitive species and apparently consistent with an

early Miocene date (see p. 33). This hyrax is regarded as one of the more

certain indications that there was some endemism in southern African faunas

during the earlier part of the Miocene and that the complication in comparing

east and southern African taxa mentioned above does have some substance.

The other identified taxa from Langental, Protypotheroides beetzi and

Pomonomys dubius, are apparently known only from the Namib desert and,

since they may be southern endemics, they may not be useful for relative dating

purposes.

The available evidence indicates that the Langental fauna dates from the

early Miocene and that it may be a ‘Burdigalian’ rather than ‘Aquitanian’

equivalent.

The Plant 4 borehole and Langental faunas have three species in common,

namely, Myohyrax oswaldi, Bathyergoides neotertiarius and Paraphiomys

pigotti. This suggests that the former may also be of early Miocene age. On the

other hand, all three taxa are, or may be represented at Arrisdrift as well (see

below), so a slightly younger age (early middle Miocene) is also possible.

The material described by Hopwood (1929) includes three species known

from the east African early Miocene. They are Paraphiomys pigotti, P. stromeri

and Myohyrax oswaldi. In addition, this assemblage includes an ochotonid

which is apparently closely related to the east African Kenyalagomys minor

(see p. 31). Hopwood’s material also includes the large myohyracine, Protypo-

theroides beetzi, which is represented at Langental. Once again an early Miocene

age is indicated. In view of the earlier comments on ochotonids, the presence of

Kenyalagomys in Hopwood’s assemblage may mean that this assemblage, or

part of it, postdates that from Elisabethfeld.

In spite of the uncertainties relating to the four assemblages, there is no

§ ANNALS OF THE SOUTH AFRICAN MUSEUM

justification for the recent practice of treating the faunas as a unit. Even if it

could be established that they are exact contemporaries, it is as well to accord

them individual status. Since the Arrisdrift fauna is younger than some or all

of those from the Lideritz—Bogenfels area, there is certainly no justification

for adding it to the Namib Miocene mixture, and for this reason alone it will

now be inconvenient and inappropriate to refer to ‘the Miocene fauna’ from

this region.

THE FOSSIL VERTEBRATES FROM ARRISDRIFT

CLASS OSTEICHTHYES

One or more species of fish are represented by a few isolated vertebrae

(e.g. AD 668, AD 672) and fin spines (e.g. AD 759, AD 779). The latter

apparently belong to catfish (Clartidae).

CLASS AMPHIBIA

A single postcranial bone (AD 811) belongs to a frog or toad.

CLASS REPTILIA

ORDER SQUAMATA

There are two snake vertebrae (AD 707, AD 1110) in the assemblage.

ORDER CROCODILIA

? Crocodylus niloticus Laurenti, 1768

The most commonly represented lower vertebrate is a crocodile, probably

Crocodylus niloticus, of which many isolated teeth (e.g. AD 71, AD 310) and

scutes (e.g. AD 335, AD 341) are preserved. Postcranial bones and skull frag-

ments are less common. The best specimens include two incomplete dentaries

(AD 344, AD 999). Crocodiles are not uncommon as fossils in east Africa and

elsewhere but have not previously been recorded from Tertiary occurrences in

southern Africa and are rare in Quarternary deposits.

ORDER CHELONIA

A few isolated scutes (e.g. AD 73, AD 832) belong to one or more species

of tortoise.

CLASS AVES

Birds are represented by a few isolated and incomplete postcranial bones

(e.g. AD 725, AD 841) belonging to more than one species.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 9

CLASS MAMMALIA

ORDER INSECTIVORA

Family Macroscelididae

Myohyrax cf. oswaldi Andrews, 1914

A Myohyrax is the most commonly occurring small mammal in the

assemblage and is represented by many isolated teeth and mandible fragments

(e.g. AD 125, AD 1104) (Fig. 2). Most of the mammalian postcranial bones

have yet to be classified and this material probably includes specimens belonging

to Myohyrax.

Stromer (1926) recorded three myohyracines from the Namib, namely,

M. oswaldi, M. doederleini and Protypotheroides beetzi, while Hopwood (1929)

subsequently named an additional species, M. osborni, from the same region.

In the most recent review of this material, Patterson (1965) recognized only

M. oswaldi (including M. doederleini) and P. beetzi (including M. osborni). In

doing so he followed Whitworth (1954), except that Whitworth did not regard

Protypotheroides as a valid genus.

The Arrisdrift Myohyrax is apparently indistinguishable from M. oswaldi,

but the identification is qualified since the teeth are a little smaller than those

Fig. 2. Occlusal and buccal views of Myohyrax cf. oswaldi mandible (AD 971) from Arrisdrift.

10 ANNALS OF THE SOUTH AFRICAN MUSEUM

of typical east African M. oswaldi. This was one of the features which Stromer

(1926) claimed characterized M. doederleini. Whitworth (1954) found that

M. doederleini fell within the size variation observed in east African M. oswaldi,

but since Arrisdrift provides a second sample of specimens from the Namib

in which the teeth are comparatively small, there may be a taxonomically

significant mean difference between the Myohyrax from the two regions. In

addition, since examination of the Arrisdrift Myohyrax during the present

study was cursory, the material may differ from typical M. oswaldi in characters

other than size.

The Arrisdrift Myohyrax is evidently younger than any previously recorded

myohyracine (see p. 32), and this is another factor to be taken into account

when the material is studied in detail.

ORDER CARNIVORA

Although poorly represented, the only identifiable carnivores from

Arrisdrift are Carnivora rather than Creodonta. It is, however, possible that

certain non-diagnostic specimens such as isolated canines and postcranial bones

do belong to creodonts, a group whose presence is to be expected in view of

the apparent age of the Arrisdrift assemblage (see p. 35).

Family Amphicyonidae

Amphicyon cf. steinheimensis Fraas, 1885

An incomplete left mandible (AD 133) belongs to an amphicyonid (Fig. 3,

Table 3). There are also several postcranial bones which may belong to this

anne

ees

0 Mn

Fig. 3. Occlusal and buccal views of Amphicyon cf. steinheimensis mandible (AD 133) from

Arrisdrift.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA

€°6 S16

OG CVS

Os Cail

6 V1 O'Sr

OIG LOGY,

0°67 .

9°79

COE

709

eeldGV Oocsl dV

TW MOjaq s[qipueBw Jo WIPeaig

TAL Mojaq sIqipuew jo yIdeq

fq MOjaq s[qGIPpUB JO yIPRaIg

fq MOjaq ajqipuewu jo ydeq

‘YUPSLty WOI (p19 GY) SHadyosy ¢ pue

(119 GY) ‘jJopul svuluoATIWOPH “(EE, GY) vodoiydup “(OZS| GY) WOAIUILD osIV POyMUoprUN oy} JO So[qIpueW PUL YJ99} JAMO] 9Y} JO SUOISUDLIG

¢ ATV,

SNJOSA|V

= ate a GEG 2 CO Opi ey 06 ns GO Oe = 1S BEG = SS I" °9 “plo av

aa = os os Say a eS) Se ey | = 468 AQ, 20 oo = 119 dV

Coe S10) SO CEM IGE Oo = 06 2 pert de (CE SG WO GO) ccl GV

— 0°91 ‘9 WG Oe 2 S6l O9€ GW ANG = We? ae ey a ee = Sa OC GYy,

q I q I q I q I q I q | | q |

a “W aN "d “d 'd ©)

iD ANNALS OF THE SOUTH AFRICAN MUSEUM

species, but they were excluded from consideration.

The amphicyonids, which are sometimes regarded as a subfamily within

the Canidae (e.g. Kuss 1965), or as a separate family (e.g. Hunt 1972), were a

successful, diverse and widespread group in the Old World and North America

during the Oligocene and Miocene. They are not well known in Africa and

prior to the Arrisdrift discovery had not been recorded in southern Africa. The

taxonomy of the group is complex and, in spite of recent revisions, the identifi-

cation of specimens such as AD 133 is difficult.

AD 133 lacks the ascending ramus, the incisors and P, to P;. Of the remain-

ing teeth only the Mg is largely intact, although the salient features of the C

and P, to M, are preserved. This specimen indicates that the species was an

unspecialized, slender-jawed Amphicyon of moderate size. The P, to P3, of which

only the roots or alveoli remain, are reduced in size and more or less evenly

spaced between the C and P,. The P, was single-rooted, P, had two roots which

has coalesced at the alveolar margin, while P, was also double-rooted. The P,,

which has lost the principal cusp, has an anterior accessory cusp, a larger

posterior accessory cusp and a tiny cusp on the posterior cingulum. The M,

is a high-crowned tooth with a prominent protoconid and stout metaconid.

The talonid is sectorial, it lacks the entoconid and makes up about one-third

of the length of the tooth. The M, has a double-cusped trigonid and single-

cusped talonid. The Mj has little relief on the occlusal surface and, like M,

and Mg, is relatively narrow. _

AD 133 resembles specimens belonging to the middle Miocene A. stein-

heimensis from Europe (see Kuss 1965), and is tentatively identified with this

species. It may, however, belong to a previously unrecorded African species of

Amphicyon. The only amphicyonids recorded from east Africa are two early

Miocene species, Hecubides euryodon and H. macrodon (Savage 1965). The

Arrisdrift species differs from H. eurydon in several respects, including its

larger size. It cannot be compared with H. macrodon, which is known only

from an isolated M?.

Superfamily Canoidea (senso Savage 1977)

Amphicyonidae or Hemicyoninae gen. et sp. indet.

A largely intact, but somewhat abraded right mandible with well worn

P, to M, (AD 1520) belongs to a very large canoid (Fig. 4). This specimen

compares in size with the mandible of the largest terrestrial carnivore previously

recorded from southern Africa, namely, the early Pliocene Agriotherium

africanum from Langebaanweg, Cape Province (Hendey 1972, 1977).

In some respects AD 1520 resembles the European middle to late Miocene

amphicyonid, Amphicyon major. For example, the mandibles of the two species

are of similar overall size, while the preserved teeth of AD 1520 are morpho-

logically similar to the corresponding teeth of A. major. Like all amphicyonids,

AD 1520 lacks a premasseteric fossa. Its teeth are closest in size to those of

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA is

Fig. 4. Occlusal and buccal views of large canoid mandible (AD 1520) from Arrisdrift.

later varieties of A. major (see Kuss 1965; Table 3, this report). They do, how-

ever, differ in being relatively broad, a feature which applies particularly in

the case of P,. Although lost, the P, and P; of AD 1520 were evidently also

relatively large and, together with P,, formed a closed series.

The large size of the premolars distinguishes the Arrisdrift species from

previously recorded A. major and, indeed, from all other Miocene amphicyonids.

Apparently only in certain Oligocene species are the premolars relatively

large and in the form of a closed series (see Springhorn 1977). If AD 1520 is

indeed related to A. major, it must be more primitive than recorded representa-

tives of this taxon even though it is ‘advanced’ in terms of overall size. The

origins of A. major are obscure (Kuss 1965) and it may well have arrived in

Europe as an immigrant from Africa. The Arrisdrift species may represent the

stock from which A. major was derived.

There is, however, a second alternative which must be considered. The

Hemicyoninae, a group of Miocene ursids which share many characters with

amphicyonids, also include a very large species whose origins are obscure.

This is Dinocyon thenardi of the later middle Miocene of Europe (Hiirzeler

1944). The hemicyonines also have reduced premolars, although the reduction

is not necessarily as marked as in contemporary amphicyonids. They pre-

sumably evolved from forms in which the premolars were relatively large and

14 ANNALS OF THE SOUTH AFRICAN MUSEUM

in this respect the Arrisdrift species may be seen as an appropriate ancestor for

D. thenardi. The lower molars of the former are a little smaller than those of

the Grive St Alban D. thenardi, and in this respect as well the Arrisdrift species

is the less specialized (1.e. more primitive).

AD 1520 does, however, differ from D. thenardi, and other hemicyonines,

in lacking a premasseteric fossa. This would not necessarily exclude it from an

ancestral role, but it does suggest an amphicyonid, rather than hemicyonine

connection.

A third alternative is that the Arrisdrift species represents the stock from

which both A. major and D. thenardi were derived. It has been suggested that

Amphicyon and Dinocyon are closely related (e.g. Matthew 1924), although

more recent interpretations of canoid inter-relationships indicate that similari-

ties between these taxa are due to parallel evolution. Nevertheless, the fact

that neither A. major nor D. thenardi have known immediate ancestors does

raise the possibility that they may have had one in common.

Finally, AD 1520 may belong to a species related to the early Miocene

Afrocyon burolleti from Gebel Zelten in Libya (Arambourg 1961). The holotype

of this species is a mandible fragment with P, to M3, which is similar in overall

size to AD 1520, but which differs in having smaller P, to M, and a double-

rooted M.. If the two forms are indeed related, then A. burolleti is clearly the

more primitive and the Arrisdrift species could still be ancestral to the European

A. major and/or D. thenardi.

In view of the uncertainties about the relationships of AD 1520, it would

be fruitless at this stage to consider the taxonomic implications of the various

alternatives mentioned above. Although unidentified, AD 1520 is still significant

in revealing the presence of a type of carnivore not hitherto known from the

Miocene of southern Africa and in suggesting that the phylogeny of similar

taxa elsewhere may require reinterpretation.

? Family Ursidae

? Hemicyoninae gen. et sp. indet.

A mandible fragment (AD 611) belongs to a carnivore intermediate in

size between the Amphicyon cf. steinheimensis and the large canoid discussed

above (Table 3). Only the anterior part of the mandibular corpus is preserved,

and of the teeth only the roots or alveoli of P, to P, and a small part of the

crown of M, are preserved. It is readily distinguished from the Amphicyon

cf. steinheimensis by its larger size and in having a deep mandibular corpus. In

the latter respect it resembles the large canoid, but the overall size difference

is more than would be expected in intra-specific variation.

Although AD 611 has yet to be positively identified, it matches in size and

other observable respects corresponding parts of the Hemicyon californicus

holotype from the Miocene of North America (Frick 1926: 34, fig. 12B). It is

larger than specimens of European Hemicyon and Harpaleocyon described by

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 1S)

Hiirzeler (1944), but since hemicyonines and other ursids exhibit appreciable

sexual dimorphism (see Colbert 1939), size differences are not necessarily a

reliable criterion for distinguishing species. Even if it could be established that

AD 611 represents a hemicyonine, it is unlikely that the species concerned could

be identified.

Hemicyonines have not previously been recorded from Africa, but

elsewhere they occur in association with some of the characteristically Miocene

taxa which have been found at Arrisdrift. The size of AD 611 suggests a middle

Miocene rather than earlier age.

? Family Felidae

? Metailurus sp.

An isolated upper canine (AD 616) apparently belongs to a large felid of

the group which includes the extinct genera Metailurus and Dinofelis. The

specimen is 16 mm long, 12,6 mm wide and has a crown height of 44 mm.

The species concerned was larger than the only Metailurus hitherto recorded

in Africa, namely, M. africanus from the east African early Miocene (Savage

1965). AD 616 compares in size with the C of M. major from the late Miocene

of China (Zdansky 1924), but differs in being slightly shorter and broader. It

also resembles the C of the early Pliocene Dinofelis aff. diastemata from Lange-

baanweg (Hendey 1974), particularly the specimen SAM-PQ-L 20685. Once

again the only difference is that AD 616 is a little shorter and broader. This

difference suggests that the Arrisdrift species was less advanced than the other

two species.

Since carnivore canines are not necessarily diagnostic, AD 616 is only

tentatively identified and is referred to Metailurus, a Miocene genus, rather

than Dinofelis, a Plio/Pleistocene genus, in view of the age of the Arrisdrift

assemblage.

Family Mustelidae

? Ischyrictis sp.

An incomplete left mandible (AD 614) belongs to a small carnivore with

a high-crowned canine, relatively narrow and high-crowned premolars and

carnassial, and a reduced M, (Fig. 5, Table 3). The M, has a small metaconid,

while the talonid is short and lacks the entoconid. The small M, and sectorial

M, talonid suggests that the relationships of AD 614 may lie with the primitive

gulonine /schyrictis and it is identified accordingly. This genus has not pre-

viously been recorded in Africa, but is known from the Miocene of Europe

and Asia Minor (see Crusafont-Pairo 1972; Schmidt-Kittler 1976).

The high-crowned teeth and relatively narrow premolars of AD 614

would be primitive characters in an Ischyrictis, and the Arrisdrift species may

prove to be an appropriate ancestor for the early Vindobonian J. zibethoides

of Europe.

16 ANNALS OF THE SOUTH AFRICAN MUSEUM

UTVTTUTIVILUTYVTTUOTOQVACETVOUUTOTTEUOCTOT LTO ETOUUTUT TT

Fig. 5. Occlusal and lingual views of ? Ischyrictis mandible (AD 614) from Arrisdrift.

Two other mandible fragments (AD 128, AD 756) apparently belong to

the same species as AD 614.

Other Carnivora

At least one species in addition to those already mentioned is included in

the Arrisdrift carnivore assemblage. An isolated canine (AD 127) belongs to a

species smaller than the ? Jschyrictis. The ‘Carnivora gen. et sp. indet.’ in the

accompanying faunal list (Table 1) refers to this specimen.

Also unidentified are two mandible fragments (AD 139, AD 773), three

canines (AD 55, AD 122, AD 214), an I? (AD 619) and several postcranial

bones, most of which are incomplete. These specimens may include some

belonging to species in addition to those listed.

ORDER HYRACOIDEA

Family Procaviidae

Prohyrax sp. nov.

By far the most commonly represented vertebrate in the Arrisdrift assem-

blage is a hyrax belonging to a group sometimes given subfamily rank, the

Pliohyracinae (Whitworth 1954). This was the most widespread of the hyrax

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA i)

groups and is known from localities in Eurasia as well as Africa. Most of the

recorded species were extremely large in comparison to living hyracoids and

some specimens have been mistakenly identified as rhinoceroses and chali-

cotheres. Later representatives were apparently aquatic or amphibious animals

(see Osborn 1899; Vekua 1972).

As here understood, the Pliohyracinae include the following genera:

Prohyrax—trelatively small; early to middle Miocene of South West Africa

(Stromer 1923, 1926; this report)

Parapliohyrax—large; middle to late Miocene of east and north Africa (Lavocat

1961; Bishop & Pickford 1975)

Pliohyrax—very large; late Miocene and Pliocene of Europe and China (Forsyth-

Major 1899; Osborn 1899; Viret 1949; Viret & Thenius 1952; Tung &

Huang 1974)

Kvabebihyrax—very large; late Miocene of the Soviet Union (Gabunia & Vekua

1966)

Postschizotherium—very large; late Miocene and Pliocene of China (Von

Koenigswald 1966; Tung & Huang 1974)

The relationships of Prohyrax tertiarius from Langental have hitherto

been obscure since it has been known only from fragmentary material, the holo-

type being a maxillary fragment with P® to M? and part of M? (Stromer 1926:

Pl. 41, fig. 33). Only limited comparisons with the abundant Arrisdrift material

are therefore possible, but there can be little doubt that the two forms are

closely related. They are probably not conspecific since the Arrisdrift material

belongs to a larger species, but they are here taken to be congeneric. Since the

Arrisdrift species is undoubtedly a pliohyracine, Prohyrax is accordingly

included in this subfamily.

The Arrisdrift species shares some characters with other pliohyracines but

it is not conspecific with any of them, the most obvious difference being its

smaller size. It is apparently closest to the east and north African Parapliohyrax

and differs appreciably from the three Eurasian genera, which are the youngest

and most highly specialized members of the group.

The Arrisdrift hyrax is here interpreted as a new species of the genus

Prohyrax, probably directly descended from the Langental P. tertiarius and a

likely ancestor of the later pliohyracines from further north in Africa and from

Eurasia.

Although the species will be dealt with in detail elsewhere, some obser-

vations on it are included here since it is such an important element in the

Arrisdrift assemblage.

More than forty individual animals of all ontogenetic ages are represented,

mainly by mandible and maxilla fragments, although isolated teeth and post-

cranial bones are not uncommon. The best specimen is a nearly complete skull

(AD 363) which lacks only the mandible, right I?, left M+, and parts of the right

M? and right zygomatic arch (Fig. 6, Table 4). The skull is slightly distorted

ANNALS OF THE SOUTH AFRICAN MUSEUM

i OT MN a i ny

Fig. 6. Dorsal, lateral and ventral views of Prohyrax skull (AD 363) from Arrisdrift.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 19

TABLE 4

Dimensions of the Prohyrax skull (AD363) from Arrisdrift.

Overall length : ; : : : : : 5 4 ; ; 5 : : E 160,8

Condylobasal length . : : 2 ‘ : : ; i ; : : . 5220

Palate length along midline . : ‘ : ‘ 3 d } : : 97,7

Anterior margin of orbit to anterior marein of 1h se : é : 4 j ; : 68,0

I’ to M? length : : ; ; : ; : : : : 4 : : , : 96,9

P! to P? length : : ‘ . : : : : ; : : : 5 é 32,9

M'to M? length . : : : : ; : : : : ‘ ; ; : Sees AS

M? length ‘ ; : 5 t ; : 3 . : : : ‘ , : ‘ 24,7

Me breadth . : : é : ; : i : : : é ' : s Wee

Interorbital width . : : : : : ; : : : : d : 3657

Postorbital width . 3 ‘ : : ; : : : p ; : : , Sisk

Zygomatic width . : ; : ; : : : : , 3 : ; : : 91,0

Mastoid width : : : t ; ‘ : : ‘ ; : : 2 : 73,8

Palate width at M? : : i : : 5 : : 5 2653

Horizontal diameter of arbi : ; : : : : : : : : , é 24.3

Vertical diameter of orbit . : : ‘ : : : ; A ‘ : i ; 21,0

Ventral margin of orbit to M® alveolar margin : , ' : ; iS 7/

in places, particularly the posterior part of the braincase. Distinctive features,

some of which are characteristic of other pliohyracines, include closed orbits,

naso-maxillary fossae which lead ventrally and posteriorly into antorbital

foramina situated immediately above the infraorbital foramina, a dental

formula of 3.1.4.3 with the I? to M® series closed, a premolariform C and an

elongated M? with an additional (third) lobe situated posteriorly. The preceding

comments on the upper teeth also apply to the lowers. The mandibles lack the

fossae and fenestrae found in some other Tertiary hyracoids.

Apart from its smaller size, the Arrisdrift species is most readily dis-

tinguished from later pliohyracines by its less elevated orbits and narrower skull.

ORDER PROBOSCIDEA

Family Gomphotheriidae

Gen. et sp. indet.

Four largely intact molars and a premolar (e.g. AD 252, AD 257) belong

to at least three individuals of an unidentified gomphothere (Fig. 7). They

resemble, and may be conspecific with, specimens from Maboko in Kenya

which date back about 16 m.y.

This material was originally identified by MacInnes (1942) as Trilophodon

angustidens kisumuensis. Arambourg (1945) believed that two taxa were repre-

sented and named a new genus and species, Protanancus macinnesi, to accommo-

date some specimens. Subsequently Tobien (1973) suggested that the Maboko

material belongs to a Platybelodon (P. kisumuensis), while Maglio (1974)

referred it to Gomphotherium cf. angustidens. Recently Tassy (1977) identified

it with Choerolophodon (C. kisumuensis).

This diversity of opinion is an indication of the difficulties which exist in

identifying fragmentary proboscidean remains, and it was decided to withhold

20 ANNALS OF THE SOUTH AFRICAN MUSEUM

Hi l nn Iti a i Uy Man ee ieee a i

Fig. 7. Occlusal view of Gomphotheriidae third molar (AD 257) from Arrisdrift.

even a tentative identification of the Arrisdrift species. It is, however, unlikely

to be a Platybelodon since none of the numerous tusk fragments from Arrisdrift

are of the Platybelodon type.

Even though the material is unclassified, it is important since the molars

are more advanced than those of early Miocene gomphotheres elsewhere and

they are one of the elements in the assemblage which suggest a late ‘Burdigalian’

or post-‘Burdigalian’ age for the fauna.

Family Deinotheriidae

Prodeinotherium hobleyi (Andrews, 1911)

The Arrisdrift deinothere is represented by three cheekteeth which have

been described by Harris (1977).

Other Proboscidean Material

In addition to the cheekteeth already mentioned, there are many tusk

fragments and a few postcranial bones which evidently belong to either the

gomphothere or the deinothere. They have yet to be studied.

ORDER PERISSODACTYLA

Family Rhinocerotidae

Dicerorhinus sp.

At least three individuals of a rhinoceros are represented by several isolated

cheekteeth (e.g. AD 635, AD 827) and postcranial bones (e.g. AD 251, AD 601),

the latter being mainly elements of the pes.

The metatarsals which are known are relatively long compared with those

of the living Diceros bicornis and Ceratotherium simum, which suggests that

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA

IAAL

Fig. 8. Occlusal and buccal views of Dicerorhinus M? (AD 339)

from Arrisdrift.

iD) ANNALS OF THE SOUTH AFRICAN MUSEUM

the species concerned was either an Aceratherium or a Dicerorhinus (see Hooijer

1966) (Table 5). The cheekteeth of African Miocene representatives of these

genera may be difficult or impossible to distinguish (Hooijer 1966, 1968a),

but two M?’s from Arrisdrift (AD 339, AD 1103) resemble those of Dicerorhinus

rather than Aceratherium in having metacone bulges and unconstricted proto-

cones (Fig. 8). These characteristics, together with the elongated metatarsals,

distinguish the Arrisdrift species from other recorded African Miocene

rhinoceroses, namely, Paradiceros, Brachypotherium and Chilotheridium

(Hooter 1966, 19685, 1971).

Several Miocene species of Dicerorhinus have been recorded in Eurasia

and Africa (Hooyer 1966), including D. leakeyi from the east African early

Miocene. The teeth of the Arrisdrift species are larger than those of D. leakeyi

and in this respect resemble the European middle Miocene D. schleiermacheri

from Eppelsheim (Table 6). The Arrisdrift metatarsals are longer than those of

D. leakeyi, which are themselves ‘remarkable for their length’ (Hooijer 1966:

178), and although the metatarsals of D. schleiermacheri are not known,

Hooijer believed that they probably ‘exceeded those of D. leakeyi in length’.

Once again a similarity between the Arrisdrift and Eppelsheim species is

indicated.

The Arrisdrift species may also be more advanced than D. Jeakeyi in having

a less prominent metacone bulge, but there is doubt as to how much significance

should be attached to variations in this feature (Hooijer 1966: 128).

The available rhinoceros material from Arrisdrift is probably inadequate

for identifying the species concerned, but it does suggest one which was not

conspecific with the early Miocene D. leakeyi and which was perhaps closer

to the middle Miocene D. schleiermacheri in an evolutionary sense.

TABLE 5

Dimensions of Dicerorhinus metatarsals from Arrisdrift.

AD251 AD249 AD253

Mt II Mt III Mt IV

Median length : : : : ; E ; ‘ : a 190 170

Proximal width . p : : ; : : 33 61 43

Proximal antero-posterior diameter ; ; : : 5 48 = —

Middle width . ‘ : : : : : ; 30 50 —

Middle antero-posterior dhanieier . : f : ‘ DS 26 —

Ratio middle width/length . : ‘ 0,17 0,26 —

All measurements approximte owing to condition of specimens.

TABLE 6

Dimensions of Dicerorhinus M*’s from Arrisdrift.

AD339 AD1103

Antero-posterior diameter . 5 ' , ‘ 5 j ? : 55,0 54,3

Transverse diameter . ! ‘ , 3 : ; : 4 5 61,0 54,1

Length of outer surface , ’ : P ; ; : , 65,9 66,3

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 23

ORDER ARTIODACTYLA

Family Suidae

Gen. et sp. indet.

A mandible fragment with two cheekteeth (AD 631) belongs to a small

suid whose identity has yet to be determined.

Family Suidae

Lopholistriodon moruoroti Wilkinson, 1976

Several specimens belong to a small listriodont pig. They include a maxilla

fragment with P* to M? (AD 136) and two isolated M,;’s (AD 135, AD 636)

(Fig..9, Table 7).

TABLE 7

Dimensions of Lopholistriodon moruoroti teeth from Arrisdrift.

Pe M! M? M? M;

] b ] b ] b ] b ] b

AD 136... ee Op dine, S83 9) Ca ae 103" 10:4 A Ost — —

1D 1135) alo — — — as —— — — 14,0 7,9

ADCs — ~~ = —— _ — — 24 orl

In an unpublished thesis, Wilkinson (1972) described and named a new

species of pig, Xenochoerus ? moruoroti, from Moruorot Hill in Kenya. He

subsequently referred it to the genus Lopholistriodon Pickford & Wilkinson,

1975 (Wilkinson 1976). The teeth of the Arrisdrift listriodont are virtually

indistinguishable from those of the Moruorot L. moruoroti and it is identified

accordingly.

According to Pickford & Wilkinson (1975) this species is present at

Moruorot (17 m.y.), Maboko (16 m.y.) and Muruyur (13 m.y.), so its presence

at Arrisdrift is taken as a further indication that this fauna is late ‘Burdigalian’

or post-‘Burdigalian’ in age.

Family Tragulidae

Dorcatherium cf. pigotti Whitworth, 1958

Two mandible fragments (AD 104, AD 262) and a few postcranial bones

belong to a small tragulid. The teeth are morphologically indistinguishable

from those of Dorcatherium and, of the African species of this genus, they are

closest in size to those of D. pigotti of the east African early Miocene (Whit-

worth 1958) (Table 8).

TABLE 8

Dimensions of Dorcatherium teeth from Arrisdrift.

P, P; M, M,

AD104 : : iy OL 24 1.95 23,0 — — --- —

AD262 : = — — — oS Mee) 9.1 6,2

24 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 9. Occlusal and buccal views of Lopholistriodon moruoroti maxilla (AD 136) from

Arrisdrift.

The Arrisdrift specimens are only tentatively identified with this species

since the grounds for distinguishing poorly represented fossil tragulids are

limited. There is little variation in tooth morphology and there has been a

tendency to name distinct species in Europe, Asia and Africa on the basis of

size differences. It is by no means certain that similarly sized species on the

different continents represent different species.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 25

Family Palaeomerycidae (senso Hamilton 1973)

Climacoceras sp. nov.

The Arrisdrift ruminant assemblage includes several fragments of antler-

like frontal appendages (ossicones) (e.g. AD 130, AD 132) (Fig. 10). The

beams of the ‘antlers’ are straight, transversely compressed, and have small

_ knobs situated at irregular intervals both anteriorly and posteriorly. They

evidently also carried some small tines, several detached specimens having

; been discovered (e.g. AD 129, AD 785). The tines are circular in cross-section

| and slightly curved. AD 648 indicates that there was bifurcation and greater

flattening of the ‘antlers’ distally. AD 483 apparently represents the proximal

part of an ‘antler’ and since it lacks a burr, the ‘antlers’ were evidently not

| deciduous. They were, therefore, not true antlers of the kind which characterize

} the Cervidae.

vat

WNIHI

UH lti

iii

Fig. 10. Lateral view of Climacoceras ‘antler’ fragments from Arrisdrift: AD 648 (+ AD 763)—

distal end showing bifurcation (left); AD 130—beam fragment showing knobs (centre);

AD 129—tine (right).

The complete ‘antlers’ must have resembled those of Climacoceras africanus

from Maboko (MaclInnes 1936). Climacoceras is also present at Fort Ternan

(Gentry 1970). The Arrisdrift specimens are distinguished from east African

specimens by their larger size, the difference being of the order of 20 per cent

according to A. W. Gentry (pers. comm.). The beam circumference of specimens

from Maboko varies from 54 to 92 mm (MaclInnes 1936), while the correspond-

26 ANNALS OF THE SOUTH AFRICAN MUSEUM

ing figures for Arrisdrift specimens are 85 to 110 mm. The longest of the Maboko

tines recorded by MacInnes measures 48 mm, whereas two of the Arrisdrift

specimens (AD 129, AD 1177) are about 60 mm long and both are incomplete.

The larger size of the Arrisdrift specimens is taken to indicate that they

belong to a hitherto unrecorded species of Climacoceras, although the morpho-

logical similarity to specimens from Maboko suggests that the Arrisdrift

species was in a comparable evolutionary state to C. africanus.

Many mandible and maxilla fragments, isolated teeth and postcranial

bones probably belong to this species. It is, however, not certain that all the

material provisionally assigned to the Climacoceras belongs only to this species.

A comparison between two of the more complete mandibles (AD 261, AD 612;

Fig. 11, Table 9) revealed differences which may be taxonomically significant.

For example, AD 612 has slightly larger teeth, an appreciably deeper mandibular

corpus, larger basal pillars on the molars and a less expanded P, metaconid.

In addition, the lingual surface of the M, third lobe is directly connected to the

second and is flanked by a small, more or less circular enamel island which

evidently corresponds to the central cavities of the first and second lobes. By

contrast, the M, third lobe of AD 261 is transversely compressed dorsally and

is connected lingually to the second lobe by a deeply indented loop in the

enamel. It also lacks a ‘central cavity’.

TABLE 9

Dimensions of lower teeth and mandibles AD 261 and AD 612, tentatively assigned to

Climacoceras from Arrisdrift.

P, P, M, -M, Ms;

l b l b l b b ] b

AD 261 — — 14,0 7,6 (S33) Og (PNG 12 Sia 95

A DiGi 2 ae ael Ss) ciao Wo SINE te) c.17,0 — 21.53%: CO aseee oS wt 0, 7,

AD: 261 ~~ AD Gl2

Depth of mandible below P, ; f 25,0 29,0

Breadth of mandible below P, . : al 14,5

Depth of mandible below M,_ . ; 28,5 Sie

Breadth of mandible below M; . : 13,0 15,8

Sorting the remaining mandible fragments on the basis of these criteria

was not entirely satisfactory owing to the poor condition of some specimens and

because in some instances ‘characteristics’ of one type occurred in conjunction

with ‘characteristics’ of the second. Nevertheless, AD 259, AD 263, AD 269

and AD 270 are apparently of the AD 612 type and represent at least 4 indi-

viduals, while AD 271, AD 272, AD 346, AD 356 and AD 621 appear to be of

the AD 261 type and represent 6 individuals.

The two sets of specimens are otherwise similar and amongst the shared

characteristics are a giraffoid-like orientation of the diastema region relative

to the cheektooth row, low-crowned cheekteeth, absence of P, and simple,

crescentic central cavities on the lower molars, the posterior one opening

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA DL]

lingually in early wear. Although several symphyseal teeth are known, they

do not include bilobed giraffoid-like canines.

The less numerous maxillae and upper teeth have not been closely examined.

Gentry (pers. comm.) has found that although the large ruminant teeth and

INNNUnI a

yee

inna

Fig. 11. Occlusal and buccal views of ? Climacoceras mandibles, AD 261 (above) and AD 612

(below), from Arrisdrift.

28 ANNALS OF THE SOUTH AFRICAN MUSEUM

dentitions from Arrisdrift share several characters with the early Miocene

Propalaeoryx from Elisabethfeld and the Fort Ternan Climacoceras, they are

in some respects intermediate between the two. For example, the Elisabethfeld

Propalaeoryx is more primitive in retaining P,. On the other hand, at least some

of the Arrisdrift M,’s (the AD 612 type) have ‘central cavities’ on the third

lobes, whereas the Fort Ternan Climacoceras M,’s lack this feature. In addition,

the Arrisdrift teeth are probably less high-crowned than those of the Fort

Ternan Climacoceras, while the metastylids are developed to a degree inter-

mediate between the Elisabethfeld and Fort Ternan species.

Since the Maboko Climacoceras predates the Fort Ternan species and is

younger than the Elisabethfeld Propalaeoryx, a temporal link with the Arrisdrift

species is suggested. Unfortunately the only described teeth assigned to the

Maboko Climacoceras are three lower molars (MacInnes 1936), so the basis

for comparisons with Arrisdrift specimens is limited. The Maboko teeth are,

however, similar to the AD 261 type in size and some morphological characters,

including reduced or absent basal pillars and dorsally compressed third lobe

of Msg.

There are several complete specimens amongst the postcranial bones

tentatively assigned to the Climacoceras. They include a tibia (AD 1100),

several radii (e.g. AD 494, AD 562) and metapodials (e.g. AD 198, AD 199),

as well as elements of the manus and pes. The long bones are slender and

elongated compared with those of living bovids and cervids of similar overall

size (e.g. Damaliscus dorcas, Cervus unicolor). The housing of the extensor

tendon of the distal metatarsals shows the bovid and giraffid rather than the

cervid condition (see Whitworth 1958: 23).

Family Bovidae

Gen. et sp. indet.

Several mandible fragments (e.g. AD 103, AD 106) and postcranial bones

belong to a small ruminant (Fig. 12, Table 10). Morphologically the teeth are

perhaps closest to those of the somewhat larger Walangania africanus

(= Palaeomeryx africanus + Walangania gracilis (see Hamilton 1973)) of the

east African early Miocene. W. africanus was once thought to be a bovid, but

TABLE 10

Dimensions of the lower cheekteeth and mandibles of the Bovidae indet. from Arrisdrift.

E Ee M, M, M;

l b ] b ] b ] b l b

AD OS =: & eS 4,2 8,3 4,8 — — 904 “Gil 13.0) “6,5

AD 106 — — 8,1 4,5 9,1 6,3 9,7 6,6 = =

AD 105 AD 106

Depth of mandible below P, ! E eles 14,3

Breadth of mandible below P, . : ; 6,3 6,1

Depth of mandible below M, . ‘ ; ISS 18,7

Breadth of mandible below M3. : : 7,6 71,9

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 29

Fig. 12. Occlusal and buccal views of Bovidae mandible (AD 105) from Arrisdrift.

Gentry (pers. comm.) now believes it could be congeneric with the primitive

ruminant Dremotherium from the European late Oligocene to early Miocene.

Gentry (pers. comm.) reports on the Arrisdrift material as follows: “The

smaller metastylids of the lower molars are more advanced towards bovids

than those of Walangania. Similarly, the weaker anterior ribs and the better

developed closure of the central cavities. The Maboko bovid fragment with

M, (see Whitworth 1958: 25, fig. 10a—c) appears to agree with the Arrisdrift

species except in not being smaller than Walangania: its metastylid is weak,

the anterior rib is not localized, and the central cavities are more enclosed within

the tooth.’

He concludes that the Arrisdrift species ‘is a bovid, although finding a

horn core is needed to be conclusive’.

Other Ruminants

The possibility of a second taxon being represented amongst the material

assigned to the Climacoceras has already been mentioned.

In addition, there are several ruminant postcranial bones which are too

small to belong to the Climacoceras, but far too large to belong to the bovid.

They include a distal humerus (AD 39), proximal metacarpals (AD 764, AD 964)

and a first phalanx (AD 895). The ‘Pecora gen. & sp. indet.’ on the accom-

paying faunal list (Table 1) refers to this material

ORDER LAGOMORPHA

Family Ochotonidae

Kenyalagomys sp. nov.

A mandible fragment with M, and M, (AD 813) (Fig. 13) and an isolated

P? (AD 1185) belong to a small ochotonid which resembles species previously

30 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 13. Occlusal and buccal views of Kenyalagomys

mandible (AD 813) from Arrisdrift.

recorded from early Miocene deposits in the Namib desert and east Africa.

The first African Miocene ochotonid to be described was Austrolagomys

inexpectatus from Elisabethfeld (Stromer 1924, 1926). Hopwood (1929) recorded

a second species from the same region, namely A. simpsoni. Subsequently

MacInnes (1963) identified a second genus, Kenyalagomys, on the basis of

material from east Africa and named two species, K. rusingae and K. minor.

MacInnes made no reference to Hopwood’s species, but amongst the

characters used to distinguish Kenyalagomys from Austrolagomys (MacInnes

1953: 20-21) are two of the three characters which distinguish A. simpsoni

from A. inexpectatus (Hopwood 1929: 2). They are a deep external fold on P3

and a marked median angulation (rib) on the posterior walls of the anterior

lobes of P, to Mg. It follows that if the generic distinction is justified, and it is

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 3]

here assumed that it is, then A. simpsoni must be referred instead to

Kenyalagomys.

Judged on the basis of size, neither of the east African species is a synonym

of K. simpsoni. K. rusingae is larger and, although K. minor is similar in overall

size, it differs in having a smaller P,; and larger molars.

Another species of Kenyalagomys, K. mellalensis, was recently recorded

from the middle Miocene (c. 14 m.y.) of Beni Mellal in Morocco (Janvier &

De Muizen 1976).

The molars of the Arrisdrift mandible (AD 813) are almost identical in

size to the corresponding teeth of the K. simpsoni holotype, but the posterior

walls of the anterior lobes lack the prominent ribs which characterize Kenyala-

gomys. There is, however, a faint indication of such ribs and the teeth of AD 813

are closer to the condition in Kenyalagomys than that in Austrolagomys, where

the posterior walls are smoothly curved.

The Arrisdrift P? (AD 1185), which measures 1,6 by 3,2 mm, is similar

in size to that of K. minor, it is smaller than that of K. rusingae and longer, but

narrower than that of K. mellalensis. It also resembles K. minor in having the

postero-external corner pointed rather than rounded as in both A. inexpectatus

and K. rusingae. K. mellalensis is in an intermediate position in this respect.

AD 1185 differs from K. minor in having the posterior border more or less

straight rather than convex and in this respect resembles K. rusingae and

K. mellalensis. Visible on the occlusal surface is a deeply indented enamel fold

which resembles corresponding features in the east and north African species

of Kenyalagomys, but which is absent in Austrolagomys.

The Arrisdrift ochotonid is here regarded as a previously unrecorded

species of Kenyalagomys, whose closest relatives are K. minor and K. simpsoni.

If there is a phyletic relationship between the latter two species and the one

from Arrisdrift, then the reduced median ribs of the anterior lobes of the

lower molars of AD 813 may be interpreted as being either in an incipient or

in a vestigial state. The latter alternative is more likely in view of the probable

younger age of the Arrisdrift species relative to the Rusinga K. minor and its

apparent contemporary from the Namib, K. simpsoni. Since K. mellalensis is

more likely to be related to K. rusingae than the smaller African ochotonids,

its evolutionary state relative to that of the Arrisdrift species is not determinable.

To sum up, the Arrisdrift ochotonid is apparently a new species of Kenyala-

gomys which is more advanced than the smaller species already recorded from

the early Miocene of Africa.

ORDER RODENTIA

? Family Bathyergidae

? Bathyergoides sp.

Two incomplete lower incisors (AD 141, AD 1024) belong to a large

rodent, possibly a bathyergid. They are tentatively attributed to Bathyergoides,

By ANNALS OF THE SOUTH AFRICAN MUSEUM

a genus recorded from the early Miocene of the Namib (Stromer 1926) and east

Africa (Lavocat 1973).

Family Thryonomyidae

Paraphiomys pigotti Andrews, 1914

Two incomplete mandibles (AD 629, AD 1049) belong to a rodent which

appears indistinguishable from that described by Stromer (1922, 1926) as

Neosciuromys africanus. This taxon has since been recognized as a junior

synonym of Paraphiomys pigotti by Lavocat (1973). P. pigotti is one of the more

commonly occurring rodents in deposits of early Miocene age in both east

Africa and the Namib desert. The Arrisdrift P. pigotti evidently postdates all

previous records of this species (see below).

Other Rodents

The Arrisdrift assemblage includes many isolated rodent incisors and,

judged on the basis of size, they represent at least two species in addition to

those mentioned above.

AGE OF THE OCCURRENCE

The fauna from Pit 2/AD 8 at Arrisdrift undoubtedly dates from the

Miocene and it has been suggested elsewhere (South African Journal of Science

1976; Corvinus & Hendey 1978) that it falls within the age limits of 12 to 18

m.y. The present study has tended to confirm the older Hmit, but it has also

suggested that the 12 m.y. limit is too young.

At present the fauna can be dated only in a relative sense by comparing

individual taxa with more securely dated ones elsewhere. The non-mammalian

vertebrates cannot yet be used in this way and the comments which follow are

confined to the mammals.

On the basis of previous records only two of the twenty-two mammalian

species suggest an early Miocene date for the fauna. They are Myohyrax cf.

oswaldi and Paraphiomys pigotti, which in east Africa are recorded from deposits

ranging in age from 18 to 22 m.y. Since little has been published on the midde

Miocene small mammals of east Africa, it cannot yet be assumed that these

taxa did become extinct there 18 m.y. ago. In addition, since there is some

evidence of differences in the patterns of mammalian evolution in east and

southern Africa during the earlier part of the Miocene (see below), it is possible

that taxa such as M. oswaldi and P. pigotti survived longer in southern Africa.

Two other species, the large canoid and ? Ischyrictis sp., may also be

indicative of an early Miocene date, but in both instances doubts about identifi-

cation render them unreliable for relative dating purposes.

Five of the twenty-two species suggest a middle Miocene date, that is,

late “Burdigalian’ at the earliest, but more probably ‘Vindobonian’ or even

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 33

‘Maremmian’ (senso Berggren & Van Couvering 1974). They are Amphicyon

cf. steinheimensis, Gomphotheiidae indet., Dicerorhinus sp., Lopholistriodon

moruoroti and Climacoceras sp. nov. In addition, there are another three

species which are, or probably are, more advanced than previously recorded

African “Burdigalian’ species and would be consistent with a middle Miocene

date. They are Prohyrax sp. nov., Bovidae indet. and Kenyalagomys sp. nov.

These eight species are regarded as the most significant for dating purposes.

Four of the twenty-two species would be consistent with any age from

‘Burdigalian’ to ‘Vallesian’. They are ?Hemicyoninae indet., ? Metailurus sp.,

Prodeinotherium hobleyi and Dorcatherium cf. pigotti.

The remaining six species provide no evidence of age. They are the

unidentified carnivore, suid, pecoran and rodents.

Negative evidence also gives some indication of the probable age of the

fauna. For example, the absence of equids suggests it is pre-‘Vallesian’ (see

Hooijer 1975), while the absence of positively identifiable bovids and palaeo-

tragines suggests that it predates the 14 m.y. old Fort Ternan fauna (see Gentry

1970; Churcher 1970).

There is, in fact, no secure evidence that the Arrisdrift fauna is as young as

that from Fort Ternan, while there is good evidence that it postdates the

‘Rusinga-like’ faunas of east Africa. Consequently, the likely age limits may

be reduced to between 14 and 18 m.y., with the median estimate being about

16 m.y. This is the age of the Maboko fauna and, although it has yet to be fully

described, it includes at least six species which are conspecific with, or closely

related to species from Arrisdrift (Table 11). They are Choerolophodon

kisumuensis, Prodeinotherium hobleyi, Lopholistriedon moruoroti, Dorcatherium

pigotti, Climacoceras africanus and the unidentified bovid. This suggests that

the two faunas are, indeed, of the same order of age. On the other hand, the

deinothere and tragulid are of little use as precise age indicators, while the two

faunas do differ in certain respects. Primates and creodonts are absent or

TABLE 11

The mammals from Maboko, Kenya, and their counterparts from

Arrisdrift, South West Africa.

MABOKO* ARRISDRIFT

Primates _

Paracynohyaeonodon leakeyi Various Carnivora

Megalohyrax championi Prohyrax sp. nov.

Choerolophodon kisumuensis Gomphotheriidae indet.

Prodeinotherium hobleyi Prodeinotherium hobleyi

Aceratherium acutirostratum Dicerorhinus sp.

Lopholistriodon moruoroti Lopholistriodon moruoroti

Brachyodus aequitorialis —

Dorcatherium spp., including D. pigotti D. cf. pigotti

Climacoceras africanus Climacoceras sp. nov.

Bovidae indet. Bovidae indet.

* Bishop (1967); Hooijer (1968a); Pickford & Wilkinson (1975);

Tassy (1977); Van Valen (1967); Whitworth (1958).

34 ANNALS OF THE SOUTH AFRICAN MUSEUM

apparently absent, at Arrisdrift and the hyracoids from the two occurrences

are at least generically distinct. These differences will be discussed again later,

but they indicate that either the two faunas were not exactly contemporaneous,

or that there were regional differences between contemporary east and southern

African faunas at that time.

The final word on the age of the Arrisdrift fauna has yet to come, but

available evidence suggests that it is early middle Miocene, with an inferred

date of about 16 m.y. before present.

PALAEOENVIRONMENT

The Arrisdrift fauna dates from a zoogeographically important period.

Andrews & Van Couvering (1975: 85-87) have discussed the ‘abrupt changes’

which occurred in the faunas of east Africa between 14 and 18 m.y. ago and

ascribe them to the development of a land bridge between Africa and Eurasia

and the consequent immigration of new taxa. According to Berggren &

Van Couvering (1974) the land bridge in question resulted from the closure

of the eastern Tethys between 18 and 20 m.y. ago. Faunal changes must also

have been experienced in southern Africa during this period, although it cannot

be assumed that they were coincident with, or that they were an exact parallel

of those in east Africa. There is, in fact, some evidence that the situation in

southern Africa did differ from that in east Africa.

Andrews & Van Couvering (1975) pointed out that during the early

Miocene the dominant hyracoids in east Africa were geniohyids and that they

were replaced in the middle Miocene (post-Fort Ternan) by procaviids. In the

Namib region the only recorded hyrax is the procaviid Prohyrax, which was

contemporary with geniohyids in east Africa (up to and including Maboko).

It follows that the procaviids are likely to have had southern Africa as their

centre of origin and that they moved into east Africa at a time when this region

was also receiving Eurasian immigrants. Even if there are changes in the classifi-

cation of the hyracoid taxa concerned, the substance of the preceding theory

remains the same since the east and southern African forms clearly belong to

different lineages, and it is the southern African one (a pliohyracine) which

had descendants in east Africa (and elsewhere) in post-early Miocene times.

Also relevant here is MacInnes’s (1957) opinion that the southern African

Parapedetes and east African Megapedetes were contemporary representatives

of different lineages since it, too, suggests that there was some independent

evolution of related taxa in the two regions. Other of the Namib rodents, as

well as the ochotonid Austrolagomys, may also have been southern endemics.

In comparing faunas, it may be unwise to emphasize the absence of certain

taxa since this may be due to sampling deficiencies or even incorrect identifi-

cation of specimens. While the apparent absence of, for example, primates at

Arrisdrift might be due to such factors, if their absence is real then it must be

palaeoenvironmentally significant. The absence or great rarity of primates at

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 35

another southern African late Tertiary locality, namely, Langebaanweg, has

already been mentioned elsewhere (Hendey 1976: 234), while this group is

also not recorded from the other Miocene localities in the Namib.

By contrast, Andrews & Van Couvering (1975: 86) noted that during the

early Miocene of east Africa ‘there was a notable proliferation of hominoid

primates, seven species in all; and there were at least five species of prosimians’.

Primates became much less common in east Africa during the middie Miocene,

although during the 14 to 18 m.y. transition period monkeys were ‘common

at Maboko’, whereas ‘there is little evidence for their presence in the Early

Miocene environments’ (Andrews & Van Couvering 1975: 93). The primates are

thus one group which reflect the ‘abrupt changes’ referred to earlier, but in this

instance the change is not manifested in the southern African record.

The apparent absence of primates in southern Africa during the earlier

part of the Miocene is made even more remarkable by the fact that this group

features in the ‘Burdigalian’ faunal interchange between east Africa and southern

Eurasia. Thus, while a northward movement of primates from east Africa is

documented, there is no record of a corresponding movement to the south.

To sum up, there is some evidence which suggests that early in the Miocene

faunal interchange between east and southern Africa was inhibited and that at

least some related taxa evolved independently of one another in the two regions.

This indicates the existence of an environmental barrier between the two

regions and it is most likely to have been comprised of the extensive river system

of central Africa, together with the Rift Valley lakes (see Kortlandt 1972).

Even in their present form the Congo and Zambezi river systems, especially

in the region of the Congo/Zambezi divide, make up a broad and almost

continuous area of channels and marshes between east and south-west Africa.

Tectonic disturbances in central Africa during the mid-Tertiary may well have

complicated the headwater drainage patterns of these rivers and so have created

an even more effective barrier to limit the crossing of at least certain mammals.

Another possible barrier may have been that of an intervening arid and

semi-arid region. A more extended form of the present Kalahari desert would

have effectively separated east and south-west Africa from one another. This

alternative is perhaps less likely since the present aridity of the south-western

parts of Africa was apparently initiated only in the very late Tertiary

(A. J. Tankard & J. Rogers, unpublished manuscript). Prior to this the climate

and vegetation of Africa may have been of a more uniform nature.

The preceding observations suggest that in assessing the character and

composition of southern African Miocene faunas, allowance must be made

for deviations from the better documented east African pattern because

geography is a complicating factor, the full implications of which have yet to

be established.

In this connection the apparent absence of creodonts at Arrisdrift may

also be significant. Creodonts predominate in the carnivore faunas of the 18

to 22 m.y. period in east Africa and the only identified carnivore from con-

36 ANNALS OF THE SOUTH AFRICAN MUSEUM

temporary occurrences in the Namib is also a member of this group. Similarly

the only recorded carnivore from Maboko is a creodont, while at Fort Ternan

creodonts are common if not predominant. In other words, creodonts were

an important element in the ‘Aquitanian’ to ‘Vindobonian’ faunas of east

Africa. Consequently, it is to be expected that they would not only be present

at Arrisdrift but would be more commonly represented than fissiped carnivores.

Since carnivores are less restricted by environmental factors than herbi-

vores, the dispersal of immigrant taxa may well have been rapid in spite of

barriers which impeded the movement of, for example, primates. Should it

be established that creodonts are, indeed, rare or absent at Arrisdrift, the

situation could be explained in only one of two ways.

Firstly, southern Africa was an important centre of fissiped evolution

and they superseded creodonts in this region before the same happened in east

Africa. This possibility can be dismissed in view of what is known of fissiped

origins and evolution (Savage 1977) and since it would require complete

isolation of southern Africa from east Africa during the early Miocene.

The second possibility is that Arrisdrift is younger than was indicated

previously and dates from a period when the creodonts had been largely or

completely replaced by fissipeds. The carnivores would then be the only obvious

‘advanced’ element in the Arrisdrift fauna, while those taxa which suggest

ac. 16 m.y. date would be ‘primitive’ forms which survived longer in the Namib

region than in east Africa. This interpretation would require, for example,

that early bovids and giraffids such as those found at Fort Ternan had been

prevented from spreading southwards by the hypothetical zoogeographic

barrier, whereas immigrant fissipeds had already surmounted it and become

established in southern Africa by the time that the Arrisdrift fossils were being

deposited.

Perhaps the only firm conclusion to be drawn from the preceding discussion

is that the present state of knowledge of southern African Miocene faunas

leaves much to be desired.

Some information on the nature of the environment in the immediate

vicinity of Arrisdrift at the time that the fossiliferous deposits were laid down is

suggested by both the fossils and the deposits themselves.

There is no doubt that the fossils accumulated in a river channel, the verte-

brate remains simply being an additional element in the coarse sediment

fraction of a fluvial gravel. The fact that a Prohyrax skull and other delicate

fossils were recovered from the deposits indicates that they at least could not

have been transported far in what was evidently a turbulent channel. Even

those specimens which are abraded are not seriously damaged. Thus, most of

the fossils must represent the remains of animals which lived in the immediate

vicinity or a little further upstream.

The only invertebrate in the fossil assemblage, a serpulid polychaete (cf.

Mercierella sp.), is a typically estuarine form (B. Kensley, pers. comm.) and its

presence suggests that at the time of deposition the coastline, which is at present

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA BY)

about 30 km away, was much closer. This, together with the fact that the

fossiliferous deposits are about 50 m above sea-level, suggests that deposition

took place during a period of relatively high sea-level. There is evidence for a

world-wide marine transgression during the Miocene, between 10 and 20 m.y.

ago (Flemming & Roberts 1973) and presumably at least part of the Arrisdrift

terrace sequence can be correlated with this event. Arrisdrift is situated in a

hilly area on the last meander of the Orange River before it reaches the flat

and low-lying coastal section of its valley, which is likely to have been inundated

during the transgression. The mouth of the river was, therefore, probably

only a few kilometres west of Arrisdrift.

The area around Arrisdrift is now very arid, with Alexander Bay at the

mouth of the Orange River having a mean annual rainfall of less than 50 mm

(Dept. of Transport 1965). The Orange River is, however, a large perennial

river and is flanked by a narrow belt of bushes and trees, although the vege-

tation becomes ephemeral a short distance away. There is little or no soil cover in

the area and aeolian sands and bedrock exposures are ubiquitous. Although

the climate is ameliorated by the proximity of the cold Atlantic Ocean, it is

nevertheless an inhospitable area capable of supporting only sparse popu-

lations of a relatively small number of mammalian species (see Shortridge 1934).

The environment at the time that the fossils were deposited must have been

very different. Although poorly preserved plant remains occur in the Pit

2/AD 8 deposits, no direct information on the nature of the vegetation is yet

available. The large mammals, particularly the two proboscideans, suggest a

densely vegetated and probably wooded environment. Harris (1975) has

suggested that both Prodeinotherium and gomphotheres preferred such a habitat.

The low-crowned teeth of the rhinoceros suggest that it, too, was a browser,

while the long limbs and low-crowned teeth of the Climacoceras indicate that

it was not adapted to grazing. Both are likely to have been woodland species.

The smaller herbivores are probably indicative of a dense undergrowth

at least in the immediate vicinity of the river. Living tragulids are forest-dwelling

browsers and their Miocene ancestors, including the Arrisdrift Dorcatherium,

probably had a similar habitat preference. The same is likely to apply to the

bovid and two suids since they resemble the Dorcatherium in both size and

hypsodonty, while the Prohyrax probably occupied a similar habitat, but with

somewhat different vegetable-food preferences.

Since later pliohyracines were aquatic or amphibious animals, early forms

such as the Arrisdrift Prohyrax may already have developed a preference for

life in water-side situations. The fact that it is the most commonly occurring

vertebrate in the assemblage supports the theory that it was a riparian species

since remains of such animals are more likely to be incorporated in fluvial

deposits than those of other terrestrial species. In this connection it is probably

significant that the other commonly occurring vertebrate at Arrisdrift is a

crocodile, which undoubtedly is a riparian species.

The picture which emerges is of a forested riverine setting, probably with

38 ANNALS OF THE SOUTH AFRICAN MUSEUM

dense undergrowth adjacent to the river and with the sea no more than a few

kilometres away. An essentially similar environment was suggested for those

areas further north in the Namib where the early Miocene vertebrates were

discovered (Stromer 1926; Hopwood 1929). In the case of these occurrences

the contrast to the modern environment is even more striking since there are

no rivers in the area today. There is, however, ample evidence of their presence

during the Miocene.

The Arrisdrift fossils, together with the older ones from the Liideritz—

Bogenfels area, provide some of the evidence which supports the theory that

the present Namib desert is relatively young (A. J. Tankard & J. Rogers,

unpublished manuscript). Whether or not a desert existed in the area in pre-

Miocene times is still a matter of dispute, but the earlier part of the Miocene

was evidently a period of relatively high rainfall and more luxuriant vegetation.

Andrews & Van Couvering (1975) believed that during this period a belt

of lowland forest stretched across equatorial Africa and was flanked on either

side by woodlands. The evidence from Arrisdrift and the other occurrences

suggests that these woodlands extended at least as far south as the Orange River.

ACKNOWLEDGEMENTS

I am indebted to the management and staff of Consolidated Diamond

Mines of South West Africa (Pty) Ltd at Oranjemund who were involved in

the collecting of the Arrisdrift fossils and their subsequent donation to the

South African Museum. Particular thanks are due to Dr G. Corvinus who

undertook the excavation of the material, Mr J. O. Richards (General Manager)

and Dr C. G. Stocken (Chief Geologist). A generous grant from the company

has partly covered the costs of preparing the fossils.

I am also indebted to those persons who have assisted the preliminary

study of the Arrisdrift fossils. They are: Drs A. W. Gentry (British Museum

(Natural History)), J. M. Harris (Kenya National Museum), B. Kensley (South

African Museum), R. G. Klein (University of Chicago), M. Pickford (University

of London) and R. J. G. Savage (University of Bristol). Dr Gentry was especially

helpful in sharing his extensive knowledge of ruminants; he is, however, not

responsible for any errors in interpretation of the Arrisdrift material.

The preparation of the fossils has been undertaken at the South African

Museum by Mrs K. Volman and Miss T. Salinger, while Mr N. Eden provided

the photographs which illustrate this report.

REFERENCES

ANDREWS, P. & VAN COUVERING, J. A. H. 1975. Palaeoenvironments in the East African

Miocene. Contrib. Primatol. 5: 62-103.

ARAMBOURG, C. 1945. Anancus osiris un mastodonte nouveau du Pliocéne inférieur d’Egypte.

Bull. Soc. géol. Fr. (5) 15: 479-495.

ARAMBOURG, C. 1961. Note préliminaire sur quelques Vertébrés nouveaux du Burdigalien de

Libye. C. R. Soc. géol. Fr. 1961: 107-109.

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 39

BERGGREN, W. A. & VAN COUVERING, J. A. 1974. The Late Neogene. Amsterdam: Elsevier.

BisHop, W. W. 1967. The later Tertiary in East Africa—volcanics, sediments and faunal

inventory. Jn: BisHop, W. W. & CLARK, J. D. eds. Background to evolution in Africa:

31-56. Chicago: University Press.

BisHop, W. W. & PicKFoRD, M. 1975. Geology, fauna and palaeoenvironments of the

Ngorora Formation, Kenya Rift Valley. Nature, Lond. 254: 185-192.

CHURCHER, C. S. 1970. Two new Upper Miocene giraffids from Fort Ternan, Kenya, East

Africa: Palaeotragus primaevus n. sp. and Samotherium africanum n. sp. Fossil Vertebr.

Afr. 2: 1-105.

COLBERT, E. H. 1939. Carnivora of the Tung-Gur Formation of Mongolia. Bull. Am. Mus.

nat. Hist. 76: 47-81.

Cooke, H. B. S. 1972. The fossil mammal fauna of Africa. In: Keast, A. et al, eds. Evolution,

mammals and southern continents: 89-139. Albany: State University of New York.

CorVINuS, G. & HENDEY, Q. B. 1978. A new Miocene vertebrate locality at Arrisdrift in

Namibia (South West Africa). N. Jb. Geol. Palaont. Mh. 1978: 193-205.

CRUSAFONT-PAIRO, M. 1972. Les Ischyrictis de la transition Vindobonien-Vallesien.

Palaeovertebrata 5: 253-260.

DEPT. OF TRANSPORT. 1965. Climate of South Africa. 9. Average monthly and annual rainfall

and number of rain-days up to the end of 1960. Pretoria: Government Printer.

FLEMMING, N. C. & RosertTs, D. G. 1973. Tectono-eustatic changes in sealevel and seafloor

spreading. Nature, Lond. 243: 19-22.

ForsyTH-Major, C. I. 1899. The hyracoid Pliohyrax graecus (Gaudry) from Samos. Geol.

Mag. (4) 6: 547-553.

Frick, C. 1926. The Hemicyoninae and an American Tertiary bear. Bull. Am. Mus. nat.

Hist. 56: 1-119.

GABUNIA, L. C. & VEKUA, A. C. 1966. Peculiar representative of hyrax in the Upper Pliocene

in Eastern Georgia. Soobshch. Akad. Nauk. gruz. SSR 42: 643-647. (In Georgian.)

GENTRY, A. W. 1970. The Bovidae (Mammalia) of the Fort Ternan fossil fauna. Fossil Vertebr.

Afr. 2: 243-323.

GREENMAN, L. 1966. The geology of Area 2615C, Liideritz, South West Africa. Unpublished

thesis, University of Cape Town.

HAMILTON, W. R. 1973. The Lower Miocene ruminants of Gebel Zelten, Libya. Bull. Br.

Mus. nat. Hist., Geol. 21: 73-150.

HAMILTON, W. R. & VAN COUVERING, J. A. 1977. Lower Miocene mammals from South

West Africa. Namib Bull. (Suppl. 2, Transv. Mus. Bull.): 9-11.

Harris, J. M. 1975. Evolution of feeding mechanisms in the family Deinotheriidae (Mamma-

lia: Proboscidea). Zool. J. Linn. Soc. 56: 331-362.

Harris, J. M. 1977. Deinotheres from southern Africa. S. Afr. J. Sci. 73: 281-282.

Hetssic, K. 1971. Brachypotherium aus dem Miozan von Siidwestafrikas. Mitt. Bayer. Staats-

samml. Paldont. hist. Geol. 11: 125-128.

HENDEY, Q. B. 1972. A Pliocene ursid from South Africa. Ann. S. Afr. Mus. 59: 115-132.

HENDEY, Q. B. 1974. The late Cenozoic Carnivora of the south-western Cape Province. Ann.

S. Afr. Mus. 63: 1-369.

HENDEY, Q. B. 1976. The Pliocene fossil occurrences in ‘E’ Quarry, Langebaanweg, South

Africa. Ann. S. Afr. Mus. 69: 215-247.

HENDEY, Q. B. 1977. Fossil bear from South Africa. S. Afr. J. Sci. 73: 112-116.

Hoover, D. A. 1966. Miocene rhinoceroses of East Africa. Bull. Br. Mus. nat. Hist., Geol.

13: 119-190.

Hoover, D. A. 1968a. A note on the mandible of Aceratherium acutirostratum (Deraniyagala)

from Moruaret Hill, Turkana District, Kenya. Zool. Meded. 42: 231-235.

Hoower, D. A. 19685. A rhinoceros from the late Miocene of Fort Ternan, Kenya. Zool.

Meded. 43: 77-92.

Hoower, D. A. 1971. A new rhinoceros from the Late Miocene of Loperot, Turkana District,

Kenya. Bull. Mus. comp. Zool. Harv. 142: 339-392.

Hoower, D. A. 1975. The hipparions of the Baringo Basin sequence. Nature, Lond. 254:

211-212.

Hopwoop, A. T. 1929. New and little known mammals from the Miocene of Africa. Am.

Mus. Novit. 344: 1-9.

40 ANNALS OF THE SOUTH AFRICAN MUSEUM

Hunt, R. M. 1972. Miocene amphicyonids (Mammalia, Carnivora) from the Agate Springs

Quarries, Sioux County, Nebraska. Am. Mus. Novit. 2506: 1-39.

HUrRZELER, J. 1944. Zur Revision der europaischen Hemicyoniden. Verh. Naturf. Ges. Basel

55: 131-157.

JANVIER, P. & DE MuiIzon, C. 1976. Les Lagomorphes du Miocéne de Béni Mellal, Maroc.

Geol. Medit. 3: 87-90.

KORTLANDT, A. 1972. New perspectives in ape and human evolution. Amsterdam: Stichting

voor Psychobiologie.

Kuss, S. E. 1965. Revision der europdischen Amphicyoninae (Canidae, Carnivora, Mamm.)

ausschlieflich der voroberstampischen Formen. Sitz. heidelb. Akad. Wiss. 1965 (1): 1-168.

LavocaT, R. 1961. Le gisement de vertébrés Miocénes de Beni Mellal (Maroc). Etude

systématique de la faune de mammiféres et conclusions générales. Notes Mem. Serv.

Carte géol. Maroc. 155: 29-94.

LavocaT, R. 1973. Les Rongeurs du Miocéne d’ ASIST orientale. 1. Miocéne inférieur.

Mem. Trav. Inst. Montpellier 1: 1-284.

MaAcINNES, D. G. 1936. A new genus of fossil deer from the Miocene of Africa. J. Linn. Soc.

25: 521-530.

MacInngs, D. G. 1942. Miocene and post-Miocene Proboscidea from East Africa. Trans.

zool. Soc. Lond. 25: 33-106.

MAcINNES, D. G. 1953. Miocene and Pleistocene Lagomorpha of East Africa. Fossil Mammals.

Afr. 6: 1-30.

MacInnss, D. G. 1957. A new Miocene rodent from East Africa. Fossil Mammals Afr. 12:

1-35.

Maaco, V. J. 1974. A new proboscidean from the late Miocene of Kenya. Palaeontology

17: 699-705.

MATTHEW, W. D. 1924. Third contribution to the Snake Creek fauna. Bull. Am. Mus. nat.

Hist. 50: 58-210.

OsBorn, H. F. 1899. Pliohyrax kruppi from the lower Pliocene of Samos. Congr. Zool. 4:

172-173.

PATTERSON, B. 1965. The fossil elephant shrews (Family Macroscelididae). Bull. Mus. comp.

Zool. Harv. 133: 295-335.

PICKFORD, M. & WILKINSON, A. F. 1975. Stratigraphic and phylozen-ue implications of new

Listriodontinae from Kenya. Neth. J. Zool. 25: 132-141.

SAVAGE, R. J. G. 1965. The Miocene Carnivora of East Africa. Bull. Br. Mus. nat. Hist.,

Geol 10: 241-316.

SAVAGE, R. J. G. 1977. Evolution in carnivorous mammals. Palaeontology 20: 237-271.

SCHMIDT-KITTLER, N. 1976. Raubtiere aus dem Jungtertiar Kleinasiens. Palaeontographica

(A) 155: 1-131.

SHORTRIDGE, G. C. 1934. The mammals of South West Africa. London: Heinemann.

SOUTH AFRICAN JOURNAL OF SCIENCE. 1976. Miocene vertebrates from South West Africa.

Se Alfred VSG Iz 250).

SPRINGHORN, R. 1977. Revision der Alttertiaren europadischen Amphicyonidae (Carnivora,

Mammalia). Palaeontographica (A) 158: 26-113.

STROMER, E. 1922. Erste mitteilung tiber Tertiare Wirbeltier-Reste aus Deutsch-Siidwestafrika.

Sber. bayer. Akad. Wiss. 1921: 331-340.

STROMER, E. 1923. Bemerkungen iiber die esten Landwirbeltier-Reste aus dem Tertiar

Deutsch-Stidwestafrikas. Palaeont. Z. 5: 226-228.

Stromer, E. 1924. Ergebnisse der Bearbeitung mitteltertiar Wirbeltier-Reste aus Deutsch-

Siidwest-Afrika. Sher. bayer. Akad. Wiss. 1923: 253-270.

STROMER, E. 1926. Reste land- und stisswasser-bewohnender Wirbeltiere aus den Diamanten-

feldern Deutsch-Siidwestafrikas. In: KAIser, E., ed. Die Diamantenwiiste Siidwestafrikas

2: 107-153. Berlin: D. Reimer.

Tassy, P. 1977. Présence du genre Choerolophodon Schlesinger (Proboscidea, Mammalia)

dans le Miocéne est-africain. C. r. hebd. Séanc. Acad. Sci., Paris 284: 2487-2490.

TosiEN, H. 1973. On the evolution of mastodonts (Proboscidea, Mammalia). 1. The bunodont

trilophodont group. Notizbl. hess. Landesamt. Bodenforsch. Wiesbaden 101: 202-276.

Tuna, Y. S. & Huana, W. P. 1974. A new Pliohyrax from Shansi. Vert. palasiat. 12: 212-216.

(In Chinese.)

MIOCENE VERTEBRATES FROM ARRISDRIFT, SOUTH WEST AFRICA 41

VAN COUVERING, J. A. 1972. Radiometric calibration of the European Neogene. In:

BisHop, W. W. & MILLER, J. A., eds. Calibration of hominoid evolution: 247-271. Edin-

burgh: Scottish Academic Press.

VAN VALEN, L. 1967. New Paleocene insectivores and insectivore classification. Bull. Am.

Mus. nat. Hist. 135: 217-284.

VekuaA, A. C. 1972. Kvabebi fauna of Akchagyu vertebrates. Moscow: Izdatelstvo ‘Nauka’.

VirET, J. 1949. Sur le Pliohyrax rossignoli du Pontien de Soblay (Ain). C. r. hebd. Séanc. Acad.

Sci. Paris 228: 1742-1744.

VirET, J. & THENIUS, E. 1952. Sur la présence d’une nouvelle espece d’Hyracoide dans le

Pliocéne de Montpellier. C. r. hebd. Séanc. Acad. Sci. Paris 235: 1678-1680.

VON KOENIGSWALD, G. H. R. 1966. Fossil Hyracoidea from China. Proc. K. ned. Akad. Wet.

69B: 345-356.

WALKER, A. 1969. Lower Miocene fossils from Mount Elgon, Uganda. Nature, Lond. 223:

591-593.

WHITWoRTH, T. 1954. The Miocene hyracoids of East Africa. Fossil Mammals Afr. 7: 1-58.

WHITWORTH, T. 1958. Miocene ruminants of East Africa. Fossil Mammals Afr. 15: 1-50.

WILKINSON, A. F. 1972. The Lower Miocene Suidae of Africa. Unpublished thesis, University

of Bristol.

WILKINSON, A. F. 1976. The Lower Miocene Suidae of Africa. Fossil Vertebr. Afr. 4: 173-282.

ZDANSKY, O. 1924. Jungtertidre Carnivoren Chinas. Palaeont. sin. (C) 2: 1-149.

6. SYSTEMATIC papers must conform to the /nternational code of zoological nomenclature

(particularly Articles 22 and 51).

Names of new taxa, combinations, synonyms, etc., when used for the first time, must be

followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb.

nov., Syn. nov., etc.

An author’s name when cited must follow the name of the taxon without intervening

punctuation and not be abbreviated; if the year is added, a comma must separate author’s

name and year. The author’s name (and date, if cited) must be placed in parentheses if a

species or subspecies is transferred from its original genus. The name of a subsequent user of

a scientific name must be separated from the scientific name by a colon.

Synonymy arrangement should be according to chronology of names, i.e. all published

scientific names by which the species previously has been designated are listed in chronological

order, with all references to that name following in chronological order, e.g.:

Family Nuculanidae

Nuculana (Lembulus) bicuspidata (Gould, 1845)

Figs 14-15A

Nucula (Leda) bicuspidata Gould, 1845: 37.

Leda plicifera A. Adams, 1856: 50.

Laeda bicuspidata Hanley, 1859: 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (fig. 8a—b).

Nucula largillierti Philippi, 1861: 87. ;

Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9.

Note punctuation in the above example:

comma separates author’s name and year

semicolon separates more than one reference by the same author

full stop separates references by different authors

figures of plates are enclosed in parentheses to distinguish them from text-figures

dash, not comma, separates consecutive numbers

Synonymy arrangement according to chronology of bibliographic references, whereby

the year is placed in front of each entry, and the synonym repeated in full for each entry, is

not acceptable.

In describing new species, one specimen must be designated as the holotype; other speci-

mens mentioned in the original description are to be designated paratypes; additional material

not regarded as paratypes should be listed separately. The complete data (registration number,

depository, description of specimen, locality, collector, date) of the holotype and paratypes

must be recorded, e.g.:

Holeivpe

SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach

Port Elizabeth (33°51’S 25°39’E), collected by A. Smith, 15 January 1973.

Note standard form of writing South African Museum registration numbers and date.

7. SPECIAL HOUSE RULES

Capital initial letters

(a) The Figures, Maps and Tables of the paper when referred to in the text

e.g. *... the Figure depicting C. namacolus...’; *. .. in C. namacolus (Fig. 10)...’

(b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded

by initials or full names

e.g. Du Toit but A.L.du Toit; Von Huene but F. von Huene

e.g. Therocephalia, but therocephalian

Punctuation should be loose, omitting all not strictly necessary

Reference to the author should be expressed in the third person

Roman numerals should be converted to arabic, except when forming part of the title of a

book or article, such as

“Revision of the Crustacea. Part VIII. The Amphipoda.’

Specific name must not stand alone, but be preceded by the generic name or its abbreviation

to initial capital letter, provided the same generic name is used consecutively.

Name of new genus or species is not to be included in the title: it should be included in the

abstract, counter to Recommendation 23 of the Code, to meet the requirements of

Biological Abstracts.

Q. B. HENDEY

PRELIMINARY REPORT ON

THE MIOCENE VERTEBRATES FROM ARRISDRIFT,

| SOUTH WEST AFRICA

i ia \ ‘

VOLUME 76 PART 2 SEPTEMBER 1978 ISSN 0303-2515

‘ee 2

i =a

OF THE SOUTH AFRICAN

~~ MUSEUM

CAPE TOWN

INSTRUCTIONS TO AUTHORS

1. MATERIAL should be original and not published elsewhere, in whole or in part.

2. LAYOUT should be as follows:

(a) Centred masthead to consist of

Title: informative but concise, without abbreviations and not including the names of new genera or species

Author’s(s’) name(s)

Address(es) of author(s) (institution where work was carried out)

Number of illustrations (figures, enumerated maps and tables, in this order)

(b) Abstract of not more than 200 words, intelligible to the reader without reference to the text

(d) Introduction

(e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents

(f) Summary, if paper is lengthy

(g) Acknowledgements

(th) References

(i) Abbreviations, where these are numerous

3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced

with 2,5 cm margins all round. First lines of paragraphs should be indented. Tables and a list of

legends for illustrations should be typed separately, their positions indicated in the text. All

pages should be numbered consecutively.

Major headings of the paper are centred capitals; first subheadings are shouldered small

capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered

italics. Further subdivisions should be avoided, as also enumeration (never roman numerals)

of headings and abbreviations.

Footnotes should be avoided unless they are short and essential.

Only generic and specific names should be underlined to indicate italics; all other marking

up should be left to editor and publisher.

4. ILLUSTRATIONS should be reducible to a size not exceeding 12 x 18 cm (19 cm including

legend); the reduction or enlargement required should be indicated; originals larger than

35 x 47 cm should not be submitted; photographs should be rectangular in shape and final

size. A metric scale should appear with all illustrations, otherwise magnification or reduction

should be given in the legend; if the latter, then the final reduction or enlargement should be

taken into consideration.

All illustrations, whether line drawings or photographs, should be termed figures (plates

are not printed; half-tones will appear in their proper place in the text) and numbered in a

single series. Items of composite figures should be designated by capital letters; lettering of

figures is not set in type and should be in lower-case letters.

The number of the figure should be lightly marked in pencil on the back of each illustration.

5. REFERENCES cited in text and synonymies should all be included in the list at the end of

the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable):

(a) Author’s name and year of publication given in text, e.g.:

‘Smith (1969) describes .. .’

‘Smith (1969: 36, fig. 16) describes...’

‘As described (Smith 1969a, 19696; Jones 1971)’

‘As described (Haughton & Broom 1927)...’

‘As described (Haughton et al. 1927)...’

Note: no comma separating name and year

Dagination indicated by colon, not p.

names of joint authors connected by ampersand

et al. in text for more than two joint authors, but names of all authors given in list of references.

(b) Full references at the end of the paper, arranged alphabetically by names, chronologically

within each name, with suffixes a, b, etc. to the year for more than one paper by the same

author in that year, e.g. Smith (1969a, 19695) and not Smith (1969, 1969a).

For books give title in italics, edition, volume number, place of publication, publisher.

For journal article give title of article, title of journal in italics (abbreviated according to the World list o,

scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part

number (only if independently paged) in parentheses, pagination (first and last pages of article).

Examples (note capitalization and punctuation)

BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FISCHER, P. —H. 1948. Données sur la résistance et de le vitalité des mollusques. J. Conch., Paris 88: 100-140.

FIscHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. Archs

Zool. exp. Zen. 74: 627-634.

Konn, A. J. 1960a. Ecological notes on Canis (Mollusca: Gastropoda) in the Trincomalee region of Ceylon.

Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean.

Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische

und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270.

Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

(continued inside back cover)

wrrveawwrtltenwrvr yy

Fig. 1. Guineafowl taxa recognized in this study.

A. A. meleagrides. B. A. niger. C. Guttera plumifera plumifera. D. G. p. schubotzi. E. G. pucherani

pucherani. F. G. p. verreauxi. G. G. p. sclateri. H. G. p. barbata. I. G. p. edouardi. J. Acryllium

vulturinum. K. Numida meleagris meleagris. L. N. m. sabyi. M. N. m. galeata. N. N. m. somaliensis.

O. N. m. marungensis. P. N. m. reichenowi. Q. N. m. mitrata. R. N. m. coronata.

S. N. m. damarensis.

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 76 Band

September 1978 September

Part 2 ~+=Weel

THE EVOLUTION OF GUINEA-FOWL

(GALLIFORMES, PHASIANIDAE, NUMIDINAE)

TAXONOMY, PHYLOGENY, SPECIATION AND

BIOGEOGRAPHY

T. M. CROWE

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material

becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid

van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000

OUT OF PRINT/UIT DRUK

1, 21-3, 5-8), 3(1-2, 4-5, 8, t.-p.i.), 51-3, 5, 7-9),

6(1, t.-p.i.), 71-4), 8, 91-2, 7), 101-3),

11(1-2, 5, 7, t.—p.i.), 15(4—5), 24(2), 27, 31(1-3), 32(5), 33

Kopieregnavrae na die Suid-Afrikaanse Museum

ISBN 0 908407 52 1

Printed in South Africa by In Suid-Afrika gedruk deur

The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk.,

Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap

THE EVOLUTION OF GUINEA-FOWL

(GALLIFORMES, PHASIANIDAE, NUMIDINAE)

TAXONOMY, PHYLOGENY, SPECIATION AND BIOGEOGRAPHY

T. M. CROWE

FitzPatrick Institute, University of Cape Town

(With 53 figures, 11 tables and 2 appendices)

[MS. accepted 15 March 1978

ABSTRACT

Patterns of qualitative and quantitative character variation in 1 833 museum specimens

encompassing all taxa attributed to the Numidinae are analysed to produce an hypothetical

taxonomy and phylogeny for the subfamily. Quantitatively and ecologically defined genus,

species and subspecies concepts are applied in erecting the taxonomy. A cladistic approach,

based in postulated primitive-derived character sequences, is used in developing the phylogeny.

4 genera, 6 species and 16 subspecies are recognized. Cladistic events are linked to likely

causal geological and palaeoecological events to determine a possible evolutionary chronology.

Genera are thought to have arisen as a consequence of Miocene and Pliocene radiations;

species and subspecies as a result of Pleistocene divergence. An hypothetical map of African

avifaunal zones, based on evolutionary patterns found in guinea-fowl, is offered. Comparisons

of this map with other African avifaunal maps and with distribution maps of selected francolin

taxa suggest that biogeographic patterns found in guinea-fowl reflect broad patterns found

in many African birds. An hypothesis as to the causes of relatively high species richness in

francolins is offered.

CONTENTS

PAGE

Introduction ; : : : , 5 5 . 44

Methods ; ; : j : >

Material and characters ; : ; : oee'§45

Taxonomic philosophy : 3 : : a)

Taxonomic methodology . : : : . 30

Phylogenetic philosophy . : : : Fee)

Phylogenetic Semen . : : san ee

Speciation . . : : ; : i ae

Biogeography : : : : ee

Results, discussion and conclusions . ; : - —36

Taxonomy . : : : 2 ; : S436

Genera . : : : ‘ els

Species and subspecies : : : / “356

Taxonomic summary . : ; : 95

Phylogeny . Se etl:

Primitive and derived character ‘states Be ult

Phyletic analysis . : ; : . 120

Speciation . : Be ull

Origin and derivation of the Numidinae gort2

Evolution of genera. sf l2

Evolution of species and subspecies . a 12D

Biogeography : ; : é : i b 1 128

Results . ‘ ; : : : : 123

Discussion . ; : : : 5 a 18}

Conclusions . ; 5 5 : : ~ §25

Ann. S. Afr. Mus. 76 (2), 1978: 43-136, 53 figs, 11 tables, 2 appendices.

44 ANNALS OF THE SOUTH AFRICAN MUSEUM

PAGE

Synthesis ; : i : : : : ; » 1126

Acknowledgements. é : ; : L130

References . ; ; : . ; ; : . 130

Appendix 1 . ; : : : : : . = 138

Appendix 2 . : : ; : : ; «| h34

INTRODUCTION

There are many reasons why an understanding of the evolution of guinea-

fowl (Numidinae; Sibley & Ahlquist 1972) should provide an insight into

patterns of avian evolution and biogeography in Africa.

|. The Numidinae are endemic to Africa, and presumably evolved and/or

radiated there from a francolin-like ancestor (Ghigi 1936; Cracraft 1973;

Olson 1974).

2. At least one guinea-fowl species has become adapted to life in each

major terrestrial African biome outside of desert, Mediterranean vegetation

and montane forest (Crowe & Snow 1978).

3. Guinea-fowl are sedentary birds (Chapin 1932a; Priest 1933; Archer

& Godman 1937; Elgood et al. 1973), and thus should be more susceptible

to local selection pressures than would be more mobile species (Ehrlich &

Raven 1969).

4. The distributions of some guinea-fowl taxa do not correlate well with

present-day vegetation and topography (Chapin 1932a). Such anomalous

distributions of plants and animals can be used to infer environmental con-

ditions and the distributions of African biomes during the Tertiary and

Quaternary (Chapin 1932a; Moreau 1963, 1966; Roberts 1975; Hamilton

1974; Axelrod & Raven 1978).

5. The genetic basis of morphological variation in several guinea-fowl

species is relatively well understood (Ghigi 1936).

6. Perhaps most importantly, guinea-fowl species are relatively well

represented in museum collections and, accordingly, lend themselves to

quantitative analysis, which is desirable in formulating precise evolutionary

and biogeographic hypotheses (Mayr ef al. 1953; F. Vuilleumier 1975).

Thus, guinea-fowl provide a simple, characteristically African system

from which evolutionary and biogeographic hypotheses may be derived. The

aims of the present study are to:

1. re-examine and revise, if necessary, the rather confused taxonomy within

the subfamily (Table 1), using a repeatable, relatively objective quantitative

methodology;

2. produce a parsimonious phylogeny based on the analysis of shared derived

character states;

3. develop models of speciation, which are consistent with the phylogeny

developed herein and the likely past geological and climatic history of

Africa;

THE EVOLUTION OF GUINEA-FOWL 45

4. suggest tentative avifaunal subregions, provinces and districts for Africa,

based on analysis of the distributions of recognized guinea-fowl taxa.

The possible familial status of the subfamily (Wetmore 1960), the adaptive-

ness of inter- and intra-specific morphological variation in guinea-fowl, and

the predictive value of evolutionary and biogeographic models based on

patterns found in guinea-fowl, will be discussed briefly, if at all, herein. These

topics will be dealt with in greater detail in future papers.

METHODS

MATERIAL AND CHARACTERS

The author examined | 833 museum specimens, including all taxa attributed

to the subfamily in Table 1. All characters investigated (Appendices 1 and 2,

Figs 2-4) appear to have a genetic basis (Ghigi 1936), and have been discussed

in previous studies (e.g. Beddard 1898; Bannerman 1930; Chapin 1932a;

Ghigi 1936; Archer & Godman 1937; Jackson 1938; Boetticher 1954;

Mackworth-Praed & Grant 1952, 1962, 1970). Therefore, they need not be

described again in detail. Characters were chosen to reflect as much of the

phenotype as possible, while eliminating the necessity of close comparisons of

specimens from different collections. Due to logistical constraints (e.g. small

sample sizes of each sex, time allotted for examination of specimens, and

variation in the detail of collectors’ notes and care in specimen preparation)

some of the quantitative characters listed in Appendix | were assessed relatively

subjectively, and data for both sexes were lumped. This introduces a certain

amount of imprecision into the analysis. But, subjective assessments were made

against a set of reference specimens, photographs or drawings encompassing

the range of observed variation. Also, the previous studies mentioned above

have stated that sexual dimorphism is absent, or relatively minor in comparison

with geographical variation in guinea-fowl. Moreover, it was felt that the

advantages accruing from considering a large number of characters for many

specimens outweighed any sampling bias due to imprecise measurement and

sexual dimorphism.

TAXONOMIC PHILOSOPHY

The taxonomic philosophy adhered to in the present study follows that

outlined by Mayr et al. (1953). A species is a group of actually or potentially

interbreeding individuals which has diverged sufficiently in allopatry to have

become reproductively isolated from other such groups. A genus is a mono-

phyletic taxon which is decidedly qualitatively distinct, and is adapted to a

particular mode of life, i.e. a generic ‘niche’.

The subspecies is much more difficult to define. Table 1 shows that much

of the ‘taxonomic variation’ in guinea-fowl systematics is at the subspecies

level. Some systematists (e.g. Wilson & Brown 1953; Moreau 1957; Selander

1971; Gould & Johnston 1972) state that the category is useless, and even

— Xx x x x x x S1Aspajau S11302]0U Dpiunn

Xx x x 1ubsayond

x isuowulds

x x x e2DUI2]YJOY

x x SUJOIIPIAH

x x x x x x DjDqGAvq

x x x 1q4udAS

x x x x x quidpya

x x x x x x x 14210] IS

x x x x x x x 3 IXNDAdJIA

x x x x x x 1YJUus-Yjas

x x x x x x luapajnoyos

x x x DI1aYONS

x x x x x x x ipAonopa ,lp4pnopa

= x x x x ,luvsoyand

5 x x x x x x x izjoqnyas

a x x x x 2K x x vaafmunjd paafiunjd p41ajjny

= x x x x x x x winulAnyjNna wnyjcsop

Zz x x x x x x x Saplisvajau

< x x sasiu Sajspjasp

= x x x Xx x 4aslu SnpIsvyd

hy,

: (OL6I ‘7961

S _ (Apnys (S961) 7661) WURIH (pS61) (9€61) (E61) (Hy UI

fe) sty) SHUM W poeld TOYO IsIYD S19}9q pure v7¢61) exeL

sf OMOID -YIOM YC] uldey)D

fe) t[MOJj-vouIns JO AUIOUOXe} 94} JO AIOISIY VW

: ajay

. I L

“1updsayond *1)

SI poziusooer oureu oyIdeds oy} AyJOLId Jo s[dioutid oy} SUIMOT[OJ ‘Apnys yuasord oy} UL DYIDedsuOd aq 0} POIOPISUOD BIB VX} OAM} 9SOY} BUTS +

z (S961) SHUM 99S «

rf (S961) OMUM 29S “IsDjj0d °a * = Ixnvaddad ‘a “Dz

< (PEG) S19}0d SMOT[OJ ON}V[OUSUUOU “poyioeds osIM194}0 Sso[UY) +

SS eee

eB x“ 1mouayaiad

fy x _NODUD]q

= x x x x x x x sisuaADUUDp

° x x x x x x vsoyjidod

5 x SISUAJDDASUDAJ

| x x sisuaododuy

2 x x x x x x DIDUOAOD

a x muuUosf

a x x QDMY1A

x x sisuayayn

x x x x DUILXDUL

x x x x x x 1JAaDMa]JDI

x x x x x x x sisuasunsDU

x x x x Xx x IMOUIAYI1AA

x x x x x x x DIDAJIU

x x x x x x 1UaSSDAJS

x x x x x 1aYIADUL

x x x x x x x pypajos

x x x x x 14qvSs

sudau1

x 128.40SUD

x SISUdOWO

x quupuinau

x sisuanso]

x pIpau4sajul

TABLE 1

A history of the taxonomy of guinea-fowl.*

Ber as a cv re nm

Chapin Mackworth- : Crowe

Taxa (19324 and Peters Ghigi Boetticher Praed & White (this

in litt.) (1934) (1936) (1954) Grant (1952; (1965) study)

1962; 1970)

Phasidus niger x x x x x : a

elastes niger

ip elena x x x x x x 4

Acryllium vulturinum x x x x x x x

Guttera plumifera plumifera x x x x x x x

schubotzi x x x x x x x

pucherani* x x x x

edouardi* edouardi x x x x x x x

suahelica x x x

schoutedeni x x x x x x

seth-smithi x x x x x x

verreauxi* x x x x x x x

sclateri x x x x x x x

' chapini x x x x x

granti x x x

barbata x x x x x x

lividicollis x x

kathleenae® x x x

symonst x

pucherani x x< x

Numida meleagris meleagris x se axe x x x x

sa CU eS cc ae a Koa ot ae 9 a sears am emmeere— eeen E—

> v im) es ME Pad x ? ~~ a ~ ~~

x x x x x x x

intermedia x x x x x x

toruensis x x x x x x

neumanni x

omoensis x

ansorgei x

inermis

sabyi x x x x et

galeata x x x x y : .

marchei x x x % S

strasseni x x x x ~ <

mitrata x x x se &, 5 =

reichenowi x x x x y) .

marungensis x x x x Me ~ :

callewaerti x x x x Mc .

maxima x x x x

uhehensis x x

rikwae x S

frommi x

coronata x x x % e %

limpopoensis x x

transvaalensis x

papillosa x x x % e ¥

damarensis x x x % S v 2

blancout* x

reichenowi .

1 Unless otherwise specified, nomenclature follows Peters (1934).

2G. e. verreauxi = G. e. pallasi, see White (1965).

3 See White (1965).

2 Se es two taxa are considered to be conspecific in the present study, following the principle of priority the specific name recognized is

. pucnerant.

WngsaAW NVOIdsVv HINOS AHL 40 STVNNY

TMOS-VANIND AO NOILN TOA FHL

48 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 2. Quantitative characters 1, 4-17. A. Lateral view of head, neck and collar of Numida

meleagris. B. Dorsal view of the same.

Fig. 3. Quantitative characters 19-22. A. Secondary remex from Numida meleagris. B. Wing

covert from N. meleagris. C. Mid-dorsal feather of the same.

THE EVOLUTION OF GUINEA-FOWL 49

Oa 39

Jae a ™

ere(° eh

Fig. 4. Quantitative characters 22-26, 29-39. A. Lateral view of head, neck and collar of

Guttera plumifera. B. Mid-dorsal feather of G. pucherani edouardi. C. Spots on mid-dorsal

feathers of Guttera spp.

detrimental to the understanding of geographic variation and intra-specific

evolution. These critics maintain that:

1. characters used to delineate subspecies usually have patterns of geographic

variation which are discordant with each other and with the distribution

attributed to the taxon;

2. indistinguishable phenotypes occur in geographically isolated areas, pre-

sumably due to parallel evolution under similar selective regimes;

3. no objective degree of difference can be offered to distinguish subspecies

from slightly differentiated local populations.

In an attempt to satisfy these legitimate criticisms, Ford (1974) offers

the taxo-evolutionary subspecies concept, the concept adhered to in the present

study. This concept limits the awarding of subspecies status to geographic

ageregates of populations which appear to have undergone genetic and pheno-

typic divergence in allopatry. Past allopatric divergence may be inferred for

50 ANNALS OF THE SOUTH AFRICAN MUSEUM

presently parapatric taxa if their distributions can be derived from concordant

character variation, and if they are separated by a zone of secondary inter-

gradation (Ford 1974).

TAXONOMIC METHODOLOGY

Since both the recognition of taxa and their assignment to taxonomic

categories (with the possible exception of the species) are ultimately subjective

processes (Mayr et al. 1953; Selander 1971; Ford 1974), it is of utmost

importance to outline the methodology underlying taxonomic decisions.

With the foregoing taxonomic philosophy in mind, the procedure used

to determine genera involved four steps:

1. a sorting of specimens into groups which consisted of members of both

sexes, and which possessed unique combinations of qualitative characters;

2. a cluster analysis of these groups according to the number. of shared

qualitative character states;

3. presentation of the clustering similarity matrix in the form of a phenogram;

4. interpretation of the matrix and the phenogram, according to the taxonomic

philosophy outlined above, to determine genera.

Once genera were identified, each was analysed separately to determine

species and subspecies. Two multivariate statistical computer programmes

were used in tandem in these analyses. The first programme, BMDP2M (Dixon

1975), was used to cluster specimens into operational taxonomic units (OTUs,

Sneath & Sokal 1973). In BMDP2M, the clustering algorithm first amalgamates

the two specimens which are most similar. The amalgamated specimens are

then treated as one case in future comparisons. This clustering algorithm con-

tinues until all specimens are grouped into one large cluster. The similarity

measure used is the Euclidean distance, and the clustering method is the

weighted average pair group method (Sneath & Sokal 1973). The printed output

of BMDP2M includes a phenogram which illustrates the results of the cluster

analysis in a hierarchical manner. OTUs were identified from examination of

these phenograms. In a phenogram, OTU status was awarded to any specimen

cluster which contained members of both sexes from several geographically

contiguous localities, and which was morphologically more distinct (i.e. linked

up with other such groups at a lower similarity) than was a group of specimens

from a single, relatively well-sampled locality within the range of an undisputed

taxon listed in Table 1.

OTUs were compared using stepwise multiple discriminant analysis pro-

gramme BMDP7M (Dixon 1975). This programme calculates a series of

linear classification functions in a stepwise manner, such that within-OTU

variance is minimized and between-OTU variance maximized. At each step in

the analysis, the character not yet entered into the functions that best separates

OTUs is included. This procedure continues until all significant (P < 0,05)

characters are included in the functions. The printed output of BMDP7M

THE EVOLUTION OF GUINEA-FOWL 51

includes a probabilistic statement as to the OTU membership of each specimen.

In the final step, BMDP7M computes canonical discriminant functions between

OTUs, and plots the first two so as to give an optimal two-dimensional picture

of the separation of the OTUs. This plot consists of a multivariate centroid

for each OTU, surrounded by a cloud of individual points corresponding to

specimens in that OTU. In this study, an OTU, or group of OTUs, was awarded

species rank if none of the component specimens were intermediate between

two OTUs (i.e. were assigned a probability greater than 0,05 of belonging to

another OTU).

Once species were identified, character variation within each species

consisting of several OTUs was analysed to determine if any component OTUs

merited subspecies status. The taxonomic procedure used in these analyses

involved four steps:

1. division of the species range into uniform areas;

2. computation of mean values and coefficients of variation (COV, Sokal &

Rohlf 1969) for each character for each area;

3. contour mapping of mean values for each character, and of the total COV

for all characters for each area;

4. interpretation of the contour maps to determine OTUs whose distributions

could be derived from concordant character variation, and whose boundaries

were circumscribed by a zone of secondary intergradation.

Subspecies rank was awarded to any OTUs whose distributions could be

derived from concordant variation in at least three characters, and which

enclosed an area of relatively low total COV bounded by an area(s) of high

total COV. Areas with relatively low variability (i.e. low total COV) were

taken to be probable ‘core regions’ for their associated subspecies. Areas with

relatively high variability (i.e. high total COV) were taken to be regions of

secondary intergradation.

The smallest uniform area that produced sample sizes that were statistically

adequate was that enclosed by a block four degrees on a side. Contour mapping

was done with the assistance of a computer, using contouring programme

GPCP (CALCOMP 1971). This programme fits an approximate contour

surface to the data using least squares polynomial analysis.

PHYLOGENETIC PHILOSOPHY

A cladistic approach (Marx & Rabb 1970; Cracraft 1972) was used to

infer phylogenetic relationships between guinea-fowl genera and _ species.

Cladistic analysis involves discernment and use of derived character states.

Derived states are those which are unique or relatively restricted to the taxa

under study, or which could be adaptive in a social or ecological context (Marx

& Rabb 1970). Shared primitive character states, i.e. those commonly inherited

from distant ancestors, cannot be used to unite more recently evolved taxa

(Cracraft 1972). Other than its consistent, relatively objective methodology,

the major advantage of a cladistic analysis is that a proposed phylogeny can be

52 ANNALS OF THE SOUTH AFRICAN MUSEUM

refuted on clearly specified grounds. A proposed phylogeny must be modified

or abandoned if: derived characters used to produce it are shown to be primi-

tive; another interpretation of the same or different combination of derived

character states yields a more parsimonious phylogeny, i.e. requiring fewer

convergences; a fossil form is found which possesses a character suite incom-

patible with the proposed phylogeny; or if it is grossly at variance with known

biogeographic events (Cracraft 1972; F. Vuilleumier 1975).

PHYLOGENETIC METHODOLOGY

Since the presumed ancestor of the guinea-fowl is a francolin-like phasianid

(Ghigi 1936), any character state common among extant francolins or

guineafowl-phasianid hybrids was taken to be primitive. Ghigi (1936),

Mackworth-Praed & Grant (1952, 1962, 1970) and Hall (1963) were used as

sources of information on francolins. Ghigi (1936), Bourke (1967) and R. Chapin

(in litt.) were the sources of information on the phenotypes of hybrids. Any

character state unique to, or relatively common among, guinea-fowl species

was taken to be derived.

Once hypothetical primitive-derived sequences of character states were

determined, a parsimonious phylogeny was produced following methods

outlined by Cracraft (1972).

SPECIATION

There is now widespread agreement among evolutionary biologists and

biogeographers that speciation, extinction and drastic distributional shifts in

plants and animals have been common and world-wide during the Tertiary

and Quaternary (Moreau 1966; B. Vuilleumier 1971; Axelrod & Raven 1978).

As mentioned in the introduction, past geological and climatic events have

almost certainly helped to shape patterns of speciation and biogeography in

Africa. Several authors (Chapin 1932a; Cooke 1962; Howell & Bourliére 1963;

Moreau 1966; Carcasson 1964; Butzer 1967; Hamilton 1974; Axelrod & Raven

1978) have given maps of Africa depicting the hypothetical distribution of

vegetation during relatively wetter and/or drier conditions in the past. Living-

stone (1975), Hamilton (1974) and Axelrod & Raven (1978) have pro-

vided some temporal estimates for climatic and geological events which

may have caused these conditions. As a preface to this study, the various

hypothetical vegetation maps were compared, and compromise ‘wet’ and ‘dry’

maps (Figs 5-6) were drawn which incorporated salient features of each. An

hypothetical pattern and chronology for speciation in guinea-fowl were

developed by relating the phylogeny derived herein to present-day (Fig. 7)

and hypothetical past ‘wet’ and ‘dry’ vegetation maps in the light of temporal

estimates given by authors mentioned above.

BIOGEOGRAPHY

Once the taxonomy and phylogeny of a group are defined, biogeographic

hypotheses based on patterns found in that group may be formulated. Since

THE EVOLUTION OF GUINEA-FOWL 53

14 6

Tes PT

vd Sheet

é tf

XF ae AAA CAYO

a | ae KAKKAA WV. Ay ica i Coe oe

| a es rar

By merce

( aay, een

(KX)

SH sere ag x28 er '

RAR RT EYRE RHR TOGO ITIOOL gto aay

aie SYK sas easels i SERRE gi Rb ai LARK SY

Bat

ORES ee en Se RRR

Ar iged

pe a A ee

aoe

Tropical lowland forest

Forest- savanna mosaic and

deciduous savanna woodland

aA gets |

Re fet t ell |

vas rc) gee Lawl

KX] Bushveld, grassland and steppe ‘6

Temperate grassland

[I] Mediterranean vegetation

[J Desert and sub- desert

[e] Lakes

Swamp

log

i ae

Fig. 5. Hypothetical vegetation map of Africa during a wetter period.

54 ANNALS OF THE SOUTH AFRICAN MUSEUM

boledaa tL Saannee A

LAX

KKK)

iS)

nei,

Chess ms Lt

ee. OR

SRR een

ran abate

ae Rn ERIN ye) hae RXR

OXY aie ROO ry) RRR RRA

YOY ERR ERT ne

ee obs one

ek: claw

e —-

elaees ee Sa

— - a ra 0

Montane and temperate forest neh a

Ae ive

[=] Tropical lowland forest

hte ae i A

(e - 1 eo

aan RR a4

Forest -savanna mosaic and

deciduous savanna woodland

ea Bushveld, grassland and steppe

[| Desert and sub-desert

[e] Lakes

[fa] swamp

Aaa 'a:

at ae

ry,

vil

PNT tee

km Ves canes he

aa CRT ot

Fig. 6. Hypothetical vegetation map of Africa during a drier period.

THE EVOLUTION OF GUINEA-FOWL 35

22 14 6

; ely

Ly Sepuega

Lett TT Lp

- Scie ci

ete TR

id NO,

i /|

VYTY VY

hy Ts roman rae AX

Oh Reh YY ne

Hannan a mi

Bet ROKR

IO XXX

ce PERRIN KS 1 oe

YY)

4 RY

Montane and temperate forest

Bea ee ea

Tropical lowland forest

> 1g

Forest-savanna mosaic and ee |

deciduous savanna woodland [= SSIS

wy Bushveld, grassland and steppe Belew [everclear

VY Yh 16

T rassland }

emperate grass ane

(] Mediterranean vegetation

EASA AAA se

Lal Desert and sub-desert

[e] Lakes >|

km a te |

Be Swamp ———— a — 32

0 600

Fig. 7. Present-day vegetation map of Africa.

56 ANNALS OF THE SOUTH AFRICAN MUSEUM

guinea-fowl are sedentary, stenotopic animals, the distribution of a given taxon

should reflect the geographical limits and temporal stability of its associated

biome. If a taxon has a vicariated distribution, it is likely that the vicars (see

Udvardy 1969) were isolated as a result of past partitioning of its biome.

However, if a taxon shows little geographic variation, its biome probably has

not been fragmented in the past. In this study, the boundaries of recognized

genera, species and subspecies were used to formulate an hypothetical avifaunal

map of Africa. Boundaries of taxonomically homogeneous genera (i.e. with no

subspecies) were used to delimit African avian subregions in this map. In other

words, a subregion is any biome which has presented its associated guinea-fowl

taxa with a sufficiently consistent selective regime to produce a monotypic genus,

or a genus composed of monotypic species. Species boundaries were used to

delimit provinces, and subspecies boundaries to delimit districts, following

a similar reasoning used in defining subregions.

RESULTS, DISCUSSION AND CONCLUSIONS

TAXONOMY

GENERA

There are fourteen groups of guinea-fowl specimens (Table 2) which

possess unique combinations of the qualitative characters listed in Appendix 1.

These groups are termed operational genera (OG). In the light of habitat

preference information summarized by Crowe & Snow (1978), the results of a

cluster analysis of the OG (Fig. 8) suggest that there are four genera in the

Numidinae. These are labelled A—D in Figure 7, and become apparent at the

similarity level of eleven shared characters. In Table 1, genus A corresponds

to the genera Agelastes Bonaparte, 1850, and Phasidus Cassin, 1857; genus

B to Guttera; genus C to Acryllium; and genus D to Numida. Genus A must

be named Age/lastes. The following section consists of a taxonomic summary

including taxonomic conclusions, brief descriptions and mensural statistics,

and a discussion of taxonomic conclusions when deemed necessary. All recog-

nized genera are said to be largely restricted to broad ‘niches’ (Crowe & Snow

1978). The genus Agelastes is found only in dense tropical lowland forest. The

genus Guttera 1s also limited to forest areas, but inhabits riverine forest and the

forest edge as well as tropical lowland forest. The genus Acryllium is confined

to the subdesert steppe of north-eastern Africa. The genus Numida can be found

in virtually all areas of non-forested Africa outside of desert, Mediterranean

and montane vegetation.

SPECIES AND SUBSPECIES

The genus Acryllium appears to be monotypic. A cluster analysis of 43

Acryllium specimens (Fig. 9) according to 7 quantitative characters (nos 1-5,

41-42 in Appendix 2) yields only one OTU. Six individuals from Mt Kunchurro,

Boran, Ethiopia (c. 4°30’N 38°E), link to form a cluster at a level (indicated

THE EVOLUTION OF GUINEA-FOWL

TABLE 2

Operational genera derived from analysis of qualitative character states listed in Appendix 1.

Operational genera

Character

Sy LOSS SIM Gs Il athe a Ql Gs Ws Gs i eli celia @

UO CME SCA CO NO NEN CNN Fe Shi ret mer)

BSTC CUTE SARE ECT CAE CNC CA tat arte

Soe eee NEC EAC CAEN CD CN ONION a ST et ac)

Ose ON eet STON NECN tS a

DANATTANTRANANTANATANANANAAS Tt

CONMANANMAARAAMNANANANAA Tt

mAMATAAINANTRNAMRNMNANNANAA TST

ONMAANANANRBNTAAMNANTDAAA=sST

MAMNMNAANTMANABNMNAAAAAS Tt

FAMNTANTRANDBNMNAANANADWAAT

NANATFTAN RIAN MRUONAANTAAMANN

CU Tey UN sete ce esa en ee em erie

ee et a ee se ws i si me st OO Oe

OANNANHNORODOHAAMNANWON oO

ra es eS st et Sst

* Only in the juvenile bird.

NUMBER OF SHARED CHARACTERS

B. Guttera.

Fig. 8. The results of a cluster analysis of OG in Table 2. A. Agelastes.

C. Acryllium. D. Numida.

58 ANNALS OF THE SOUTH AFRICAN MUSEUM

1 7 6 9 2319 4 1311 2 26 4032 3 14 5 43 10 21 41 27 16 36 18 39 33 35 37 38 286 34 30 29 42 31 22 2517 15 20 8 2412

JONVLISIG NV3GI19N4g

Fig. 9 The results of a cluster analysis of forty-three specimens of Acryllium.

by the dashed line in Fig. 9) lower than does any combination of specimens

from the remainder of the distribution of the genus. Thus, for the characters

investigated, variation in a single population is as great as that found through-

out the range of the genus. Accordingly, in agreement with all previous studies

(Table 1), one monotypic species, Acryllium yulturinum, is recognized. The

distribution of this species is plotted in Figure 10.

The genus Agelastes is composed of two monotypic species. A cluster

analysis of 49 specimens according to 11 quantitative characters (nos 1-5,

23-24, 41-44 in Appendix 2) yields two OTUs, labelled A and B in Figure 11.

The dashed line in the figure indicates the level at which eight individuals from

Kribi, Cameroons (2°50’N 10°5’E), link to form a cluster. A discriminant

functions analysis of the two OTUs reveals no intermediate specimens. There-

fore the OTUs are recognized as species. In Table 1, OTU A corresponds to

TABLE 3

Similarity matrix showing number of shared qualitative character states between OG derived

and listed in Table 2.

OG no. 1 2 3 4 5 6 il 8 9 10° If “2. See

1 — 13 5 2 2 3 3 1 2 5 5) 4 4 5)

2 — 7 3 2 4 4 2 3 5 4 4 4 5

5 — 6 6 6 6 8 9 a / a 6 i

4 — 17 14 13 14 = «13 8 8 9 9 8

5 Sy Sy ele ES ee 8 8 9 9 8

6 == 5.;./ to: > 15 6 6 6 5 6

jj — 15 12 4 4 4 4 4

8 — 15 6 6 6 5 6

9 — 7 i 7! 6 7

10 — 13 5) Assay

11 — 13 14 14

12 — 15 16

13 — 13

THE EVOLUTION OF GUINEA-FOWL 59

one Polen

x y wh |

K x!

wy WY) C x 16

ap, ae, ae are § XXX Xx ye N

Yh werk (eos) ieee yp j x x SAA AA AR AAA AR AS ,

x xX 4 y Ko Ny Es ae fy

g h

J = i

nS x XK XK KY OOOO OOOOS . “XX vk \ x |

YOO OOOO ; xx x «x 4

Be XXX) > 40 4 x x 4 x x xX x dX X X x x ‘

‘\ KX £o,4 A> Od > KXAXKXW NX XX XIX XK KX XW) vw VW YY ¥ ¥ x. VOM |

PY ROOK ( dy eee: i ayy YY Kk x2 > —

Y q -¥ \ ; oan

YY Y , <x

v EN b

\ hp 4 ¢

Montane and temperate forest

Tropical lowland forest

Forest-Savanna mosaic and

deciduous Savanna woodland | .A[- --|- = -

Bushveld, grassland and steppe PNEV Nos ioc IS (Sere Oe beter ae

Temperate grassland

Mediterranean vegetation

Desert and sub-desert

Lakes

Swamp pe ‘ tee: AK

0 600 1200 NU i

Fig. 10. The distributions of Agelastes meleagrides (C1), Agelastes niger (MB), and Acryllium

vulturinum (CO).

Boe Sh eh el eles

60 ANNALS OF THE SOUTH AFRICAN MUSEUM

112 8 4 611 2 9103 5 7 23 49 17 304645 15 14 21 38 18 29 34 42 31 26 28 40 39 22 3635 19 48 47 43 27 33 37 24 25 16 4420 32 41 13

INVLSIG NVAGITONG

Fig. 11. The results of a cluster analysis of forty-nine specimens of Agelastes. A. A. melea-

grides. B. A. niger.

Agelastes meleagrides and OTU B to A. niger. The distributions of these species

are plotted in Figure 10.

The remaining two genera are taxonomically more complex than are the

first two. The results of a cluster analysis of 494 Guttera specimens according

to 24 quantitative characters (nos 1-5, 22-40 in Appendix 2) are summarized

in Figure 12A. The dashed line in Figure 12A indicates the level at which twelve

individuals from Ngayu, Zaire (1°45’N 27°15’E), link to form a cluster. Seven

OTUs are recognized. A discriminant functions analysis (Fig. 13) suggests that

there are two groups of OTUs between which there are no intermediate

individuals. These OTU groups are recognized as species. The first species,

comprising OTUs | and 2, corresponds to G. plumifera in Table 1. The second

species, comprising OTUs 3-7, corresponds to two commonly recognized

species, G. pucherani (Hartlaub), 1860, and G. edouardi (Hartlaub), 1867

(Table 1). Following the law of priority, this species must be named G. pucherani.

The two OTUs comprising Guttera plumifera partition the distribution

of that species into eastern and western portions. The distributions of these

OTUs (Fig. 14) are delineated by patterns of variation in fifteen of the sixteen

quantitative characters which vary in G. plumifera. Contour maps of variation

in these characters, and of the total COV for six areas (Fig. 15) within the distri-

bution of G. plumifera are given in Figures 16 and 17. Two types of character

variation, ‘mountain—valley’ and clinal variation, are apparent in these contour

maps. ‘Mountain-valley’ variation occurs when eastern and western OTUs

have similar values, and are separated by a transition area(s) with higher

(‘mountains’) or lower (‘valleys’) values. Characters which show ‘mountain—

valley’ variation are: bill length (Fig. 16B), wing length (Fig. 16C), tarso-

metatarsus length (Fig. 16D), wattle length (Fig. 16F), crest frontal length

(Fig. 16G), crest basal length (Fig. 17J), dorsal spot number (Fig. 17K), total

spot barbs (Fig. 17M), and total within-spot barbs (Fig. 17N). Characters

which show clinal variation are: occipital fold (Fig. 16A), wattle basal width

(Fig. 16E), crest rear length (Fig. 16H), dorsal spot size (Fig. 17L), ear patch

THE EVOLUTION OF GUINEA-FOWL 61

FONVLSIG NVAGITONA

Fig. 12. The results of a cluster analysis. A. Guttera specimens B. Numida specimens.

t JTAVIYVA TWOINONVO

=6 -4 -2 O +2 +4 +6 +8 +10 +12 +14

CANONICAL VARIABLE |

Fig. 13. A discriminant functions analysis of Guttera OTUs. Circles encompass 90 per cent

of the individuals assigned to each OTU. Only intermediate specimens between non-

overlapping OTUs are plotted. B — intermediate between OTUs 5 and 6; E — intermediate

between OTUs 6 and 7; P — intermediate between OTUs 1 and 2.

62 ANNALS OF THE SOUTH AFRICAN MUSEUM

aia

BB Montane and temperate forest

Tropical lowland forest

Forest-savanna mosaic and

deciduous savanna woodland eS as lise ere ; a | A ee mas

Bushveld, grassland and steppe MEA RS || GSE oe ae

Temperate grassland

Mediterranean vegetation

Desert and sub-desert

X Y ; y\ KAAAA (

a, x x \ A Vi

. } : : Sy

Xx AA A}

¥ ae ae = ty

Lakes (5 0.01, 0,0,0 Na

km i’ Sb. k [|

Fig. 14. The distribution of Guttera plumifera plumifera (W) (= OTU 2), G. p. schubotzi

(VY) (= OTU 1), and a single intergrade (x ).

Fa iO es i 8

THE EVOLUTION OF GUINEA-FOWL 63

Fig. 15. Areas used in contour map analysis of Guttera spp. Those areas marked with an ‘X’

have data for both species. Those marked with a ‘P’ have data for G. plumifera only, and

those without notation for G. pucherani only.

64 ANNALS OF THE SOUTH AFRICAN MUSEUM

14°E

A |

B J

30 28

C 233230 227 223 K

L 11,0

E Mi

F A a

H P

i=)

Fig. 16. Contour maps of character

variation in Guttera plumifera. A. Occipital

fold. B. Bill length. C. Wing length.

D. Tarso-metatarsus length. E. Wattle

width. F. Wattle length. G. Crest frontal

length. H. Crest rear length. OTU Fig. Contour maps of character

boundaries are indicated by thick lines. mee and total COV for Guttera

plumifera. I. Crest central height. J. Crest

basal length. K. Dorsal spot number.

L. Dorsal spot size. M. Total spot barbs.

N. Total within spot barbs. O. Ear patch.

P. Spot barb blueness. Q. Total COV.

OTU boundaries are indicated by thick

lines.

THE EVOLUTION OF GUINEA-FOWL 65

(Fig. 170), and spot barb blueness (Fig. 17P). These well-defined OTUs each

have a region of relatively low total COV (c. 250) within their distributions, and

are separated by a transition area with a relatively high total COV (c.750)

(Fig. 17Q). Thus, both OTUs meet the criteria set for subspecies (see Taxonomic

methodology). In Table 1, the western subspecies (OTU 2) corresponds to

G. p. plumifera, and the eastern subspecies (OTU 1) to G. p. schubotzi. In

statistical comparisons, these subspecies differ significantly (P < 0,05; ¢ test)

in eight characters. Guttera plumifera plumifera has significantly higher values

for bill length, wattle basal width, wattle length, crest rear length, and crest

basal length. Guttera plumifera schubotzi has higher values for occipital fold, ear

patch, and spot barb blueness. Means and standard deviations for these and

other characters are given in the section below.

Approximate geographic distributions of the five OTUs comprising

G. pucherani are shown in Figure 18A. Contour maps of twenty-three characters

which vary in this species (nos 1-5, 22-31, 33-40 in Appendix 2) are given in

Figures 19-30A. In these maps, the distributions of all five OTUs are delineated

from those of their neighbours by statistically significant patterns of variation

in six to seventeen of the characters analysed. The results of all possible

pairwise statistical comparisons (f tests) between neighbouring OTUs are

summarized in Tables 4 and 5. A contour map of variation in total COV for

twenty-five areas (Fig. 15) within the distribution of G. pucherani is given in

Figure 30B. This figure shows that the distributions of all five OTUs enclose

or fall within a region of relatively low total COV (200-400). For all OTU

distributions but one, that of OTU 4, the region of relatively low total COV

is bordered by a region(s) of relatively high total COV (e.g. 600-800). Thus, all

OTUs ascribed to G. pucherani satisfy the criteria specified for subspecies (see

Taxonomic methodology). The lack of a high total COV interface between the

low COV regions of OTUs 3 and 4 is attributed to poor sampling and the small

geographic distribution of OTU 4. Assuming that G. edouardi is a synonym

of G. pucherani in Table 1, and following the law of priority, OTU 3 corresponds

to G. p. verreauxi, OTU 4 to G. p. sclateri, OTU 5 to G. p. pucherani, OTU 6

to G. p. barbata, and OTU 7 to G. p. edouardi. The geographic distributions of

these subspecies, and of intermediate populations, are shown in Figure 31.

The results of a cluster analysis of 704 Numida specimens according to

characters |—22 in Appendix 2 are summarized in Figure 12B. Nine well-defined,

and one borderline OTU are recognized. The dashed line indicates the level at

which eight individuals from the vicinity of Gassam, Senegal (14°50’N 15°20’W),

link to form a cluster. Only 704 specimens could be analysed in this cluster

analysis due to computer storage limitations. However, those specimens analysed

were chosen so as to ensure uniform sampling of the sexes and collection locali-

ties. A discriminant functions analysis including all 1 245 Numida specimens

(Fig. 32) reveals intermediate individuals between all parapatric OTUs. There-

fore only one species is recognized. Following the law of priority, this species

is named Numida meleagris.

66 ANNALS OF THE SOUTH AFRICAN MUSEUM

ast

Fig. 18. The geographic distributions of OTUs comprising A. Guttera pucherani

and B. Numida meleagris.

THE EVOLUTION OF GUINEA-FOWL 67

Fig. 19. Contour maps of character variation in Guttera pucherani. A. Bill length. B. Wing

length. OTU boundaries are indicated by thick lines.

68 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 20. Contour maps of character variation in Guttera pucherani. A. Tarso-metatarsus

length. B. Wattle basal width. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 69

Fig. 21. Contour maps of character variation in Guttera pucherani. A. Wattle length.

B. dorsal spot size. OTU boundaries are indicated by thick lines.

70 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 22. Contour maps of character variation in Guttera pucherani. A. Crest frontal length.

B. Crest rear length. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL Wi

Fig. 23. Contour maps of character variation in Guttera pucherani. A. Crest central height.

B. Crest basal length. OTU boundaries are indicated by thick lines.

12 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 24. Contour maps of character variation in Guttera pucherani. A. Anterior crest curli-

ness. B. Posterior crest curliness. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 73

Fig. 25. Contour maps of character variation in Guttera pucherani. A. dorsal black collar.

B. Ventral black collar. OTU boundaries are indicated by thick lines.

74 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 26. Contour maps of character variation in Guttera pucherani. A. Occipital fold.

B. Throat red. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL TS

Fig. 27. Contour maps of character variation in Guttera pucherani. A. Orbital red. B. Dorsal

spot number. OTU boundaries are indicated by thick lines.

76 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 28. Contour maps of character variation in Guttera pucherani. A. Total spot barbs.

B. Total within spot barbs. OTU boundaries are indicated by thick lines.

— “

THE EVOLUTION OF GUINEA-FOWL Gh

Fig. 29. Contour maps of character variation in Guttera pucherani. A. Spot barb blueness.

B. Chestnut blotch size. OTU boundaries are indicated by thick lines.

78 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 30. Contour maps of character variation in Guttera pucheranj. A. Chestnut blotch

extent. B. Total COV. OTU boundaries are indicated by thick lines. Larger numbers in

B. refer to OTU numbers shown in Figure 18A.

THE EVOLUTION OF GUINEA-FOWL 79

6 6

eo ss eo

D | | ; | | | | \ \ ' eo |

poet at Po | | Ale | + =e a \ aay

ATID (<i ! KOO) | ae | \ \ i

An) 7K eo | ae os | | \ —_\+32

R090 ee oe |

Leo > SD

, He ae ah aN ad be

606. ee x

Cie

RAR NUK OCC EKO

fox

PB Montane and temperate forest

Tropical lowland forest

Forest-savanna mosaic and

deciduous savanna woodland

Bushveld, grassland and steppe Meo. ae of ff.

Temperate grassland

Mediterranean vegetation

Desert and sub-desert

Lakes

Swamp

[ej

Fig. 31. The distributions of Guttera pucherani subspecies. G. p. pucherani (MB). G. p. verreauxi

(A). G. p. sclateri (OQ). G. p. barbata (C0). G. p. edouardi (@). Intergrades (x).

TABLE 4

ANNALS OF THE SOUTH AFRICAN MUSEUM

A comparison of OTU 3 for G. pucherani with OTUs 4, 5 and 7. x — not significantly

different (P < 0,05; ¢ test); + — OTU 3 significantly greater; — — OTU 3 significantly

lower; * — differences delineate OTUs in contour maps.

Character name?

Bill length A :

Wing length . ; ;

Tarso-metatarsus lengt

Wattle basal width

Wattle length

Dorsal spot size

Crest frontal length

Crest rear length .

Crest central height

Crest basal length

Anterior crest curliness

Posterior crest curliness

Dorsal black collar

Ventral black collar

Occipital fold

Throat red

Orbital red

Dorsal spot number

Total spot barbs .

Total within spot barbs

Spot barb blueness

Chestnut blotch size

Chestnut blotch extent

1 See Appendix 2 for character descriptions.

@ AIVIYVA TVIINONVD

CANONICAL VARIABLE |

OTU

Contour map

Figure no.

19A

Fig. 32. A discriminant functions analysis of Numida OTUs. Circles encompass 90 per cent

of the individuals assigned to each OTU. Only intermediate specimens between non-

overlapping OTUs are plotted. M — intermediate between OTUs 4 and 7; R — intermediate

between OTUs 4 and 8; X — intermediate between OTUs 4 and 2.

THE EVOLUTION OF GUINEA-FOWL 81

TABLE 5

A comparison of OTU 6 for G. pucherani with OTUs 5 and 7. x — not significantly different

(P < 0,05; ¢ test); + — OTU 6 significantly greater; — — OTU 6 significantly lower; * —

differences delineate OTUs in contour maps.

Character name* OTU Contour map

5 7 Figure no.

Bill length 4 x x 19A

Wing length . : ; x Le: 19B

Tarso-metatarsus lengt x de; 20A

Wattle basal width x _* 20B

Wattle length x ae 21A

Dorsal spot size . x _* 21B

Crest frontal length -* x DN

Crest rear length . ae —* 22B

Crest central height x 23A

Crest basal length x + * 23B

Anterior crest curliness x _* 24A

Posterior crest curliness x —* 24B

Dorsal black collar JLES _* 25A

Ventral black collar a = 25B

Occipital fold x +* 26A

Throat red aa aL 26B

Orbital red = a 7A

Dorsal spot number x JL 27B

Total spot barbs . x ae 283A

Total within spot barbs x -.* 28B

Spot barb blueness = au 29A

Chestnut blotch size ae = 29B

Chestnut blotch extent ae —* 30A

1 See Appendix 2 for character descriptions.

The approximate geographic distributions of the ten OTUs comprising

N. meleagris are shown in Figure 18B. Contour maps of variation in twenty-two

quantitative characters analysed for this species (nos 1-22 in Appendix 2),

and of the total COV for forty-eight areas (Fig. 33) are given in Figures 34-44.

In these contour maps, the distributions of all ten OTUs but one, that of OTU 3,

are delineated from those of their neighbours by statistically significant patterns

of variation in nine to seventeen of the twenty-two characters analysed (Tables

6-11). Results of all possible pairwise statistical comparisons (¢ tests) between

neighbouring OTUs are summarized in Tables 6-11. The distribution of OTU

3, the borderline OTU in Figure 11B, is delineated from those of its neighbours

(OTUs 2 and 6) by at most two characters (Tables 7, 9). In the contour map of

total COV (Fig. 45), the distribution of all OTUs, except again OTU 3, enclose

a region of relatively low total COV (c. 400-500) bordered by a region(s) of

relatively high total COV (c. 700-1000). Thus, all OTUs ascribed to N. meleagris,

except OTU 3, satisfy the criteria set for subspecies (see Taxonomic methodology).

In Table 1, following the law of priority, OTU 1 corresponds to N. m. sabyi,

OTU 2 to N. m. galeata, OTU 4 to N. m. meleagris, OTU 5 to N. m. somalienses,

Siw 7G to N. m. marungensis, OTU 7 to. N. m. mitrata, OTU 8 to

N. m. reichenowi, OTU 9 to N. m. damarensis, and OTU 10 to N. m. coronata.

ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 33. Areas used in contour map analysis of Numida meleagris.

THE EVOLUTION OF GUINEA-FOWL 83

Fig. 34. Contour maps of character variation in Numida meleagris. A. Bill length. B. Wing

length. OTU boundaries are indicated by thick lines.

84 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 35. Contour maps of character variation in Numida meleagris. A. Tarso-metatarsus

length. B. Wattle basal width. OTU boundaries are indicated by thick lines.

——

THE EVOLUTION OF GUINEA-FOWL 85

Fig. 36. Contour maps of character variation in Numida meleagris. A. Wattle length.

B. Wattle per cent blue. OTU boundaries are indicated by thick lines.

86 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 37. Contour maps of character variation in Numida meleagris. A. Helmet frontal

length. B. Helmet rear length. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 87

Fig. 38. Contour maps of character variation in Numida meleagris. A. Helmet central height.

B. Helmet basal length. OTU boundaries are indicated by thick lines.

88 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 39. Contour maps of character variation in Numida meleagris. A. Helmet thickness.

B. Cere structure length. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 89

Fig. 40. Contour maps of character variation in Numida meleagris. A. Cere structure thick-

ness. B. Collar plumage. OTU boundaries are indicated by thick lines.

90 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 41. Contour maps of character variation in Numida meleagris. A. Nape filoplume length.

B. Nape anteroposterior coverage. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 9]

Fig. 42. Contour maps of character variation in Numida meleagris. A. Nape filoplume lateral

coverage. B. Nape filoplume density. OTU boundaries are indicated by thick lines.

92 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 43. Contour maps of character variation in Numida meleagris. A. Secondary remex

outer web vermiculation. B. Wing covert barring. OTU boundaries are indicated by thick lines.

THE EVOLUTION OF GUINEA-FOWL 93

Fig. 44. Contour maps of character variation in Numida meleagris. A. Dorsal vermiculation.

B. Dorsal spot size. OTU boundaries are indicated by thick lines.

94 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 6

A comparison of OTU 4 for N. meleagris with OTUs 2, 5, 6, 7 and 8. x — not significantly

different (P < 0,05; ¢ test); + — OTU 4 significantly greater; — — OTU 4 significantly lower;

* — differences delineate OTUs in contour maps.

Character name? OTU Contour map

D, 5 6 il 8 Figure no.

Bill length . : < ’ ; : +* S< —* — x 34A

Wing length : : : : ‘ -- < —* —* * 34B

Tarso-metatarsus length . : 2 +* % —* — = 35A

Wattle basal width . : : fi x< < +* ae — 35B

Wattle length. : as ‘ : _ —* —* =+ aL 36A

Wattle per cent blue . ; : ‘ +* +* +* * ae 36B

Helmet frontal length ; : : + * x —* —* —* 37A

Helmet rear length . ; ‘ F +* x< — = = 37B

Helmet central height . ; ‘ +* < — —* —* 38A

Helmet basal length . : ; : < +* —* —* —* 38B

Helmet thickness : : : ‘ +* SE —* —* _ 39A

Cere structure length : : +* —* — a a5 39B

Cere structure thickness . : : +* < =e — =o 40A

Collar plumage . : ‘ ; P +* x x — x 40B

Nape filoplume length . : ; —* —* —* —* —* 41A

Nape filoplume anteroposterior

coverage . ; —* +* —* ‘< x 41B

Nape filoplume aera Coverane , +* +* +* +* +* 42A

Nape filoplume density . : : x S< x = +* 42B

Secondary remex outer web ver-

miculation . : ; : , +* o +* +* +* 43A

Wing covert barring . ‘ , : +* +* +* +* +* 43B

Dorsal vermiculation , : ‘ —* + +* +* + 44A

Dorsal spot size . : +* —* - x —* 44B

1 See Appendix 2 for ee deseanucne

TABLE 7

A comparison of OTU 2 for N. meleagris with OTUs 1, 3 and 6. x — not significantly different

(P < 0,05; ¢ test); + — OTU 2 significantly greater; — — OTU 2 significantly lower; * —

differences delineate OTUs in contour maps.

Character name* OTUs Contour map

1 3 6 Figure no.

Billlength . : : : : ; : : x _ —* 34A

Wing length . 3 : : ; —* — —* 34B

Tarso-metatarsus length : : : —* — —* 35A

Wattle basal width * x +* 35B

Wattle length ; : : : : 5 . * »< _ 36A

Wattle per cent blue . : : : : , +* = —* 36B

Helmet frontal length . ; : : —* — —* 37A

Helmet rear length . : ; ; : : : —* — — 37B

Helmet central height . —* — — 38A

Helmet basal length x — —* 38B

Helmet thickness . : : : 3 : : . * — _ 39A

Cere structure length . x — x 39B

Cere structure thickness x = x 40A

Collar plumage —* —* —* 40B

Nape filoplume length. : —* x —* 41A

Nape filoplume anteroposterior coverage x x =F 41B

Nape filoplume lateral coverage ; : x =e EG 42A

Nape filoplume density * x x 42B

Secondary remex outer web yermiculation x =P ale BA

Wing covert barring . : : é: +* + +* 43B

Dorsal vermiculation . . . .-. eae +* == 2h 44A

Dorsal spot size... : +* — — 44B

1 See Appendix 2 for nietie desiniiods:

THE EVOLUTION OF GUINEA-FOWL 95

TABLE 8

A comparison of OTU 8 for N. meleagris with OTUs 5 and 7. x — not significantly different;

(P < 0,05; ¢ test); + — OTU 8 significantly greater; — — OTU 8 significantly lower; * —

differences delineate OTUs in contour maps.

Character name+ OTUs Contour map

5) 7 Figure no.

Billlength . : : : : : ; ; ‘ : x —* 34A

Wing length . : : : ; ; ‘ : : ; +* +* 34B

Tarso-metatarsus length 5 3 ; , : : ; + * +* 35A

Wattle basal width. ; , : : ‘ . : : — - 35B

Wattle length F ; : : ; ; : ; 3 ; — — 36A

Wattle per cent blue . : ; : ; : ; f > =F —* 36B

Helmet frontal length . j : : : t : 4 e +* =a 37A

Helmet rear length . : : : ; : : 5 : +* * 37B

Helmet central height . : i : : ; 5 , : + +* 38A

Helmet basal length . : ; ; : : : : : +* +% 38B

Helmet thickness . : : : ' x A , : : _ + 39A

Cere structure length . —* _ 39B

Cere structure thickness — + 40A

Collar plumage x — 40B

Nape filoplume length +* = 41A

Nape filoplume anteroposterior coverage + x 41B

Nape filoplume lateral coverage : a _ 42A

Nape filoplume density : : : ; —* x 42B

Secondary remex outer web vermiculation : d : : —* +* 43A

Wing covert barring . : ' : : a + 43B

Dorsal vermiculation . : : : : : : ; : x = 44A

Dorsal spot size . : : : : : ‘ * — 44B

1 See Appendix 2 for seers Tere

TABLE 9

A comparison of OTU 6 for N. meleagris with OTUs 3, 7 and 9. x — not significantly different

(P < 0,05; ¢ test); + — OTU 6 significantly greater; — — OTU 6 significantly lower; * —

differences delineate OTUs in contour maps.

Character name} OTUs Contour map

3 7 9 Figure no.

Bill length at == aes 34A

Wing length . a= a5 a 34B

Tarso-metatarsus length x +* +* 35A

Wattle basal width — +* == 35B

Wattle length é ; : , ' _ x = 36A

Wattle per cent blue . : : : : 5 : +* +%* ++ 36B

Helmet frontal length . : ‘ ; 3 : + +* +* 37A

Helmet rear length ‘ — — —* 37B

Helmet central height . : : . : + x — 38A

Helmet basal length . : ; : ; ; : + +* +* 38B

Helmet thickness . Pee hae Mee: + a5 ae 39A

Cere structure length . : : : : : . + sé —* 39B

Cere structure thickness : ; + * —* 40A

Collar plumage . : : : : == * —* 40B

Nape filoplume length + +* a= 41A

Nape filoplume anteroposterior coverage : : — +* +* 41B

Nape filoplume lateral coverage : j : _ -- +* 42A

Nape filoplume density : : : x +* +* 42B

Secondary remex outer web venmiculation : ‘ — — —* 43A

Wing covert barring . ; : : — x x 43B

Dorsal vermiculation . : : : : — _ —* 44A

Dorsal spot size . : n : ; == == x 44B

1 See Appendix 2 for shanadict HeccEDHOAS:

96 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 10

A comparison of OTU 10 for N. meleagris with OTUs 7 and 9. x — not significantly different

(P < 0,05; ¢ test); + — OTU 10 significantly greater; — — OTU 10 significantly lower;

* — differences delineate OTUs in contour maps.

Character name* OTUs Contour map

7 9 Figure no.

Bill length . 5 : ; : 3 i : : . ; = ar 34A

Wing length . : ‘ : : : : ; ; SS = 34B

Tarso-metatarsus jensth f 3 ; ; : ; 5 ; x + 35A

Wattle basal width . : ; ; : : : : : x x 35B

Wattle length ; ; : ; : : ; A +* x 36A

Wattlespemcentblucy i) i) fy see ey a eR ee —* 36B

Helmet frontal length . f 5 ‘ ; ; ; : +* +-# 37A

Helmet rear length . ‘ : ‘ : : : : ‘ +* -+-* 37B

Helmet central height . : : : 2 : ; . : +* +* 38A

Helmet basal length . : A : : : : Emer +* 38B

Helmet thickness . : : : : 4 : : : : + x 39A

Cere structure length . : ‘ ; E : . : + —* 39B

Cere structure thickness. : ; ‘ ; : : + —* 40A

Collar featheration . : : ; : : : i melee —* 40B

Nape filoplume length : a ee a 41A

Nape filoplume anteroposterior coverage co og Dee eS a= 41B

Nape filoplume lateral coverage : : i : : : = =." 42A

Nape filoplume density ; ; : : ; — +* 42B

Secondary remex outer web vermiculation ‘ : : » ees = 43A

Wing covert barring . : : , 3 : : x x 43B

Dorsal vermiculation . ‘ : F : i i : : +* x 44A

Dorsal spot size . : WME 9 ie el —* 44B

1 See Appendix 2 for Charice ee ee

TABLE 11

A comparison of OTU 7 for N. meleagris with OTU 9. x — not significantly different (P <

0,05; ¢ test); + — OTU 7 significantly greater; — — OTU 7 significantly lower; * — differences

delineate OTUs in contour maps.

Character name? OTU Contour map

9 Figure no.

Bill length : : 3 . ; 5 : : : 4 +

Wing length . : : 2 s , ; : 3 : : + 34B

Tarso-metatarsus jeneth : ; : ' 4 : +

Wattle basal width x

Wattle length =

Wattle per cent blue 36B

Helmet frontal length . ; : : : ; : : 37A

Helmet rear length . é ; : : ? ; : : : —* 37B

Helmet central height . : ‘ ’ ‘ ; ; t ; : —* 38A

Helmet basal length . : : : , : : : : ; +* 38B

Helmet thickness . : A : : : : : . —* 39A

Cere structure length . : 5; : : : : —* 39B

Cere structure thickness. ‘ : : : : : ‘ s —* 40A

Collar featheration . : ; : ; ; : : —* 40B

Nape filoplume length 3 ; : . , : +* 441A

Nape filoplume anteroposterior coverage : ; : : ; +* 41B

Nape filoplume lateral coverage : 5 ; : ; ; : +* 42A

Nape filoplume density : : : : : +* 42B

Secondary remex outer web vermiculation : ; ; : : —* 43A

Wing covert barring . , Fj ‘ : ‘ : ‘ ; : x 43B

Dorsal vermiculation . : : : : ; —* 44A

Dorsal spot size . . } ' ; é ‘ —* 44B

1 See Appendix 2 for cece escape

THE EVOLUTION OF GUINEA-FOWL 97

Fig. 45. A. Contour map of total COV in Numida meleagris with OTU boundaries indicated

by thick lines. B. The approximate distribution of zones of intermediacy between Numida

OTUs. Larger numbers in A refer to OTU numbers shown in Figure 18B.

98 ANNALS OF THE SOUTH AFRICAN MUSEUM

The distributions of these subspecies are plotted in Figure 46, and zones of

intermediacy are shown in Figure 45B.

TAXONOMIC SUMMARY

In the following classification, verbal descriptions are brief, focusing on

more obvious differences and similarities with respect to the qualitative and

quantitative characters analysed. For each subspecies, reference is made to its

OG or OTU number and the appropriate page(s), tables and figures summarizing

its taxonomic relationships and quantitative comparisons with other subspecies.

If a synonymized taxon is not discussed, it must be assumed that it did not

emerge as an OTU in the cluster analyses. In other words, it was not taxo-

nomically distinct above a level attributable to individual variation in a relatively

homogenous population. Subspecies which are mentioned in Table 1, but are

not recognized or synonymized, are taken to be intergrades. Herein, an inter-

grade is a group of taxonomically indeterminable, phenotypically highly variable

populations. A population is taken to be taxonomically indeterminable if

many of its component individuals show affinities to two or more subspecies

(i.e. are classed as intermediates) in discriminant analyses. A population is

taken to be highly variable phenotypically if it has a high total COV. See the

section on Taxonomic methodology for a detailed discussion of OTUs, dis-

criminant analyses and total COV.

Genus Agelastes

Agelastes Bonaparte, 1850

Fig. 8(0GI1-2)

Agelastes meleagrides Bonaparte, 1850

Figs 1A, 8(0G1)

Agelastes meleagrides Bonaparte, 1850: 145.

Description

Small overall size; no crown, occipital, cere, nape or throat adornments;

rudimentary red gape wattles; no feathers on head or neck, skin colour of head

and neck red; collar plumage white; body plumage black with faint vermicu-

lations; tarso-metatarsus covered with imbricated scales in rows, and usually

with a well-developed spur(s); iris brown; outer margins of secondaries black

with faint vermiculations; furcula blade-shaped; abdominal plumage white.

Statistics for quantitative characters (N = 12)

Character X S.D.

BIpleMmeo tie 2a fon i. abo cube eM cele sige ea Cm eases eae 16,8 mm 17

Wine len@th . sleek. . 20. oe 2 ZO 4,8

iarso-metatansus lemeth. 2° 202°. al. 2 ee 80,9 mm 3,6

Waitt Tel OV GOIN aoe esis eh Selene’ “Gone Eek Vey ure alee ence 6,5 mm 2

THE EVOLUTION OF GUINEA-FOWL ‘09

ile us Ly Py aad

x mi A 2 ' ‘ ~ ee a pg

Seas aoe Ne aA HE Mie a nace Oh aA

: = 1s —oo

a and temperate forest

Tropical lowland forest

Forest-Savanna mosaic and

deciduous savanna woodland

| i Bushveld, grassland and steppe

Ey Temperate grassland

Mediterranean vegetation

Desert and sub-desert

Lakes

K Mm

Swamp _— rs)

0 600

pk = =

Fig. 46. The distribution of Numida meleagris. N. m. meleagris (A). N. m. sabyi (s).

galeata (L)). N. m. somaliensis (O). N. m. marungensis (HB). N. m. reichenowi (VY).

N. m. mitrata (t). N. m. damarensis (A). N. m. coronata (@).

100 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character x S.D.

NWiattleslemetla 08 aay an ere ene eal ee ee aati mt 2,7/mm OW

Tarsalsinucrune mulches ee 0,8 0,6

ansall structune lemetine (hy yn) eo eee 4,8 mm 85)

White collar) js % pe, 0 Aa ais ie, eg oem ea KOON OR A 0,0

Facial floplamess (3.4560) ian) Man eee oe ne 0,0 0,0

Distribution

See Figure 10.

Discussion

White collar and facial filoplumes are used as quantitative characters in

cluster and discriminant analyses to demonstrate the qualitative distinctness of

this taxon from A. niger.

Agelastes niger (Cassin, 1857)

Figs 1B, 8(0G2)

Phasidus niger Cassin, 1857: 322.

Description

As A. meleagrides except: crown surmounted by a short crest of feathers;

nape covered by short, sparsely distributed, black downy feathers; face covered

with sparsely distributed filoplumes; collar and abdominal plumage black with

faint vermiculations, abdominal plumage white in juveniles.

Statistics for quantitative characters (N = 39)

Character X S.D.

Bil Memetln wees PIC voi nee ee aig omn 3 Car leew 16,9 mm igs)

Wing length ae ae A en en enue ee ())3) 0) Sclida 13

Marso-metatansus lemorh: oan eae en ee 79,2 mm 3,4

Wattlevbasale wig. "eu 2h spose cela 6,5 mm 12

Wattle lemetin “is ys cam ac ee Val sian To [ite 2,4 mm 0,6

Mansa eseeuceune mune |) eae) tee) ene IL Il 0,7

Tarsal structure length Oe te Nek a ee RR BN i 3,4 mm 2,4

Wonite collar, agus, ooo. Ue uoey eds i Way gti ae wa OO, 0,0

PactalMlOplwMies th) Mee a ee Oe Weta ee, 9,8 0,2

Distribution

See Figure 10.

Discussion

Hall (1961) also advocates the synonomy of Phasidus in Agelastes, basing

her argument largely on analysis of juvenile characters.

THE EVOLUTION OF GUINEA-FOWL 101

Genus Guttera

Guttera Wagler, 1832

Fig. 8(0G4-9)

Guttera plumifera plumifera (Cassin, 1857)

Figs 1C, 12A(OTU2), 13(0TU2)

Numida plumifera Cassin, 1857: 321.

Description

Larger than Agelastes spp.; crown surmounted by a crest of long, straight,

bristly feathers; small occipital fold of blue-black skin; no nape, throat or

cere adornments; well-developed, pointed, blue gape wattles; orbital and

throat skin blue-black; collar plumage spotted; body plumage spotted without

vermiculations; tarso-metatarsus without spurs, scales pentagonal, not in

rows; iris brown; outer margin of secondaries white; furcula blade-shaped;

caecum up to 150 mm; abdominal plumage spotted bluish-white without

vermiculations.

Statistics for quantitative characters (N = 26)

(See p. 65 for statistical comparisons with G. p. schubotzi.)

Character D4 S.D:

Bill length 24,5 mm 17

Wing length 225,6 mm 5.3:

Tarso-metatarsus length 81,7 mm at

Wattle basal width . 10,7 mm £2

Wattle length 11,6 mm 22

Dorsal spot size 10,7 units 1,0

Crest frontal length 24,6 mm 3.1

Crest rear length 46,9 mm 56

Crest central height 31,8 mm G7

Crest basal length . 30,0 mm 2,6

Anterior crest curliness 1,0 0,0

Posterior crest curliness 1,0 0,0

Dorsal black collar OO Y 0,0

Ventral black collar 0:0 7 0,0

Occipital fold ops evA 1D

Ear patch 0,4 mm 0,6

Throat red On, 0,0

Orbital red . C057, 0,0

Dorsal spot number Dies |

Total spot barbs 5,4 0,6

Total within spot barbs 35 0,6

102 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character

Spot barb blueness .

Chestnut blotch size

Chestnut blotch extent

Distribution

See Figure 14.

0,8

0,0 units

L004

Guttera plumifera schubotzi Reichenow, 1912

Figs 1D, 12A(OTU1), 13(°0TU]1)

Guttera plumifera schubotzi Reichenow, 1912: 320.

Description

S.D.

0,6

0,0

As G. p. plumifera, except that the base of the nape and an area anterior

to the ear are covered by patches of orange-yellow skin.

Statistics for quantitative characters (N = 71)

(See p. 65 for statistical comparisons with G. p. plumifera)

Character

Bill length

Wing length

Tarso-metatarsus length

Wattle basal width .

Wattle length

Dorsal spot size

Crest frontal length

Crest rear length

Crest central height

Crest basal length .

Anterior crest curliness

Posterior crest curliness

Dorsal black collar

Ventral black collar

Occipital fold

Ear patch

Throat red

Orbital red .

Dorsal spot number

Total spot barbs

Total within-spot barbs

Spot barb blueness .

Chestnut blotch size

Chestnut blotch extent

22,8 mm

227,2 mm

80,7 mm

9,6 mm

8,9 mm

‘10,0 units

24,3 mm

38,4 mm

32,4 mm

25,9 mm

1,0

COZ

0,0 %

24,0 %

Heysilsreotan

ONO

0:0 FY

2hg2

Dil

eT)

1,6

0,0 units

0,0 %

S.D.

1,4

8,2

3,4

2,4

1,4

5,9

4,8

6,4

2,4

0,0

8,5

2,8

0,0

2,9

0,6

0,5

0,6

0,0

THE EVOLUTION OF GUINEA-FOWL 103

Distribution

See Figure 14.

Discussion

The one intermediate specimen between G. p. schubotzi and G. p. plumifera

is more schubotzi in appearance, but the patches of orange on the nape and ear

are much reduced in extent.

Guttera pucherani pucherani (Hartlaub, 1860)

Figs 1E, 12A(OTUS), 13(O0TUS), Tables 4-5 (OTUS)

Numida pucherani Hartlaub, 1860: 341.

Description

Larger than G. plumifera; crown surmounted by a long crest of relatively

curly, downy feathers; well-developed occipital fold of blue-black skin; no

nape and cere adornments; orbital and throat skin red; throat skin folded;

rudimentary, relatively short blue wattles at gape; collar plumage spotted; body

plumage spotted without vermiculations; tarso-metatarsus without spurs,

scales pentagonal, not in rows; iris red; outer margin of secondaries white;

furcula hollow; caecum length up to 150 mm; abdominal plumage spotted

bluish-white without vermiculations.

Statistics for quantitative characters (N = 64)

(See Tables 4-5 for statistical comparisons with other subspecies.)

Character xX S.D.

SUL IShG. 0 a 25,5 mm 1,8

Wing length ee eee ee ae 2 OleOr mn 14,8

Warse-metatarsus length . . . . 3. 6 el 90,2 mm 6,9

Miiasmcubasdiwidth. . . . . .« . . « a 4 9,2 mm 1S)

We SE So a en 2.7) ton 0,7

OS] SOCT.S 4 11,3 units [3

MSiecsminomialiometh . . . . ». « » ws «2 « 19,4 mm 3.5)

LPSS Sar le 1 rr rr a re 29,5 mm Bit

Mecemecntralheisht . . . 9 2 . os 6 let. % 26,4 mm 6,6

Mewemssaienemy 2 2. fe Ol 34,7 mm a

Pemctiom crest Curliness:. . . . . «=. »«» «© « « 1,0 Oo?

Ememem@merest CUrIIMESS . - . - 6 s « ss 20 0,2

Dorsal black collar Py CRIS poe Eee: RAT er end ae Deny A ES

Mimeraieplack-collar =. sw 7,4 % 10,4

Mesias es kl i 74,6 % 195

Ef O20Gf -.0 2 a a no earns 0,0 mm 0,0

“erceih rad. a. ne 9330774 A

SeGCISRCORP MR et et OR a) SS we 95,4 % 2,4

104 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character X S.D.

Dorsalspotmumber f° = ess Sh, 2 eee 26,0 4,3

Wotalispotibarbstas 23° 74. 2/e) os eee lee 6,3 13

Total. within=spot barbs, 454 3 a eee 4,3 0,7

Spot barb blueness ee Vee, Ones ere Cee a ae 9 0,9

Chestnutbloteh sizes... ee ee ee 0,0 units 0,0

Chestnut blotchrextenG@ ) 45 (29 eh =e 0.07 0,0

Distribution

See Figure 31.

Discussion

From breeding experiments with captive birds, Ghigi (1936) demonstrated

complete interfertility between individuals ascribed to G. edouardi and

G. pucherani. He also found that red throat and orbital skin, characters used

to distinguish G. pucherani, were invariably absent among FI hybrids. This

suggests that these character states may be recessive. Ghigi (1936) further

hypothesized that Guttera edouardi suahelica (Neumann 1908: 14) and

G. e. granti (Elliot 1871: 584), subspecies described from specimens collected in

southern Tanzania, are in fact intergrades between G. p. pucherani and

G. p. barbata. The results of this study support that hypothesis, since pucherani-

barbata intermediates in discriminant analysis (Fig. 13) almost invariably fit

descriptions of these taxa, and the distribution attributed to them falls in the

high total COV region between the suggested parental subspecies (Fig. 30B).

Guttera pucherani verreauxi (Elliot, 1870)

Figs 1F, 12A(OTU3), 13(0TU3), Table 4(0TU3)

Numida verreauxi Elliot, 1870: 300.

Guttera cristata sethsmithi Neumann, 1908: 13.

Guttera pallasi Stone, 1912: 208.

Guttera edouardi schoutedeni Chapin, 1923: 73.

Guttera edouardi chapini Frade, 1926: 139.

Guttera edouardi kathleenae White, 1943: 19.

Description

As G. p. pucherani, except crest longer; less well-developed occipital fold;

orbital skin colour blue; collar plumage black with no vermiculations; spotting

with stronger blue hue; iris brown.

Statistics for quantitative characters (N = 159)

(See Table 4 for statistical comparisons with other subspecies.)

Character xX S.D.

emotive tes. Seat iry gute As, Ce ew a 23,8 mm 6

Wing length OR ay ik ie, OE, 0a SR ee (Pe 11.2

THE EVOLUTION OF GUINEA-FOWL

Character X

Weno-metatarsusiengsth « . £ wok ke 89,2 mm

Piseeesoasdl Wit i a 9,2 mm

ES [SPC eh” ee a ee 2,7 mm

PieuseecmOU size ka 9,5 units

Macveummetarteneth. ke DA i wanes:

SS) Pdi? LE ee 37,4 mm

Meesmecminralheiont =. we 26,5 mm

MeemsIGMOOM ek a a 31,9 mm

Pemceiaisckest CUTIINeSS © . 9. ef lk ll it

Beistomi@r Chest CUTIMESS 2 5 3 6 kk wk ek 2,0

Mepssemisckecollar 3... kl a Dit

DeeMrmOaCKee@at! ge 8 a ea LES

DEC MEEOIC NG te 47,1 %

20 20S pe ee ee ee ne cee 0,0 mm

SOE SC. 1 eee re PT or

EE PSC co ee OO

MRORSESMOLMUMIDEL <6 24,1

DUEL SOC SST a os ane 6,0

formmmaumm-spot barbs . . . . . . . «. 39

POMMIMEHONOINCMESS. 2 . . we lk lk Des)

Micsmiimeiloren Size... kw 0,0 units

@isccmuboloteh extent . . 9 . 3. 2 8 we 8 O02,

Distribution

See Figure 31.

Guttera pucherani sclateri Reichenow, 1898

Figs 1G, 12A(OTU4), 13(0TU4), Table 4(0TU4)

Guttera sclateri Reichenow, 1898): 115.

Description

As G. p. verreauxi, except anterior crest much shorter.

Statistics for quantitative characters (N = 12)

(See Table 4 for statistical comparisons with other subspecies.)

Character x

IRE ey RP 25.2 7am

Der ee ee we ey ee ASO

fmse-metatarsus leneth . . . ke 89,9 mm

iemceaSMWwiGt 2 a we Ree 10,4 mm

Ee S ep” ae ee sar 2,3 mm

MIS IESPOUSZCr 8 OR 11,5 units

Macwainoniailenea: |. 6 kg we 5,8 mm

105

155

S.D.

DED.

10,4

4.5

0,5

Zeal

106 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character X

Grestircar leneth ss 57, ce te eee Oe eee 40,8 mm

Crest centralshetomt.| 4. 2740 See ce Sa ee ee ee 18,5 mm

@rest basalalensth... “5 8. “sea ee 32,1) iam

AntehOr Crest CUMIMESS, 5 Sells, | Stun een ens el

Posterior Chest: cunlmess: 5 ea. aimee ne Den

Dorsal blackecollak =) xs a ee 18,0 %

Ventral blackecolllar 1%) 4) 2) 8G: Seen ee 23,5074

Occipitalfolde Wats Ga 4) ue eae en ae O2SigA

Barpateh. : foe. ee OO a oh lg eee 0,0 mm

WhEOAt Ted: =. BEE aT Oh Ri ne ne a 95,374

Orbitalted. 6) 222 6 ee ey ae ee 0:0, 7,

Dorsal spot numbery. 2.42 6 Ge Ge elles ee Boe ee 23

Lotal:spot barWs "j2k 5 Gaye awe, Sa oe 6,3

Rotalswithmespor barbs ss) aya es | ee eee 4,3

Spot-barbsbluenesss.. . Go. 4G eee See 1,8

Chestnut blotchisize “bk Roe Ses eee 0,0 units

Chesinutiblotchiextent 85) 25 eis, = eae oe 00%

Distribution

See Figure 31.

Guttera pucherani barbata Ghigi, 1905

Figs 1H, 12A(OTU6), 13(0TU6), Table 5(OTU6)

Guttera barbata Ghigi, 1905: 194.

Description

1/6

0,0

As G. p. pucherani, except crest longer; throat and orbital skin blue;

collar plumage black with no vermiculation; body plumage spots occasionally

interspersed with chestnut blotches.

Statistics for quantitative characters (N = 28)

(See Table 5 for statistical comparisons with other subspecies.)

Character X

BillSlemethy.. sae <x Wieu yn Merge i le i, se ees 25,3 mE

Wing lemsth Gf= = . an Oe a ae

iarso-metatarsus tempeh) ).c.w 64. a) 2 ee ee 90,2 mm

Wattle*basalewadthe.s "798-539". nw wh ee Gk 9,4 mm

NVieretlelemethah ey Pie. kw STM og gh we ae 2,8 mm

Dorsalspotisizere.. Aieel en hoo ce GA no Ro 10,9 units

Crest trontaliensthy Vy ee ye ho Ge ee 23,3 mm

Cresimicurlenetlin eG. te Ge “se. LN ne eM ee 32,2 mina

Grest.cemrateneiolt 3.0 a. 2k te 6 se bs ee BES 26,9 mm

Crest basalsienothy 0-5) wn Bee ee ae ee 34,6 mm

Sap),

1,3

11,0

4,5

1,3

0,8

3,9

8,5

2,9

THE EVOLUTION OF GUINEA-FOWL 107

Character X SD:

PUMGMOL GREStCUGINESS 6 2. 2 1 . «6 se A 0,3

Memamcrest CURIINESS 2 9. 5 « sh ke 1,9 0,3

Mensigolackicollar 2. °. 3. «ot elu hl 21,4 % 4,5

Meemmeemlack collar. 3. 2 05 6 » 8 s + 20D/ 255)

EGMECMEROIG 2 ye 76,4 % feles

eer kk 0,0 mm 0,0

Sse each 2) Sa ee er a aaa 6,9 25,8

MOCERCOMI fe ES ee 0,3 39

Manairspotmumber 2. 2 « + + 8 + es 2959 4,9

MeveseO@tbarbS: 2... + se ee kk 539 0,7

Mormiavicnim-spot barbs . . . . . + . «= . 4,0 0,5

SaemiathonbleNesS. . 6°. 3 klk ke eS) 0,7

Mmesinmimoloreh Size . 2. wk tk ls 2,7 units ligil

Micsmuigolotch extent 2 2 2. 7 ek le AS 7, 2,6

Distribution

See Figure 31.

Discussion

Ghigi (1936) mentions G. p. barbata specimens with brown irides. This is

further evidence of gene flow between the brown-eyed western and red-eyed

eastern subspecies of G. pucherani.

Guttera pucherani edouardi (Hartlaub, 1867)

Figs 11, 12A(OTU7), 13(OTU7), Tables 4-5(OTU7)

Numida edouardi Hartlaub, 1867: 36.

Guttera lividicollis Ghigi, 1905: 195.

Guttera edouardi symonsi Roberts, 1917: 3.

Description

As G. p. barbata, except crest curlier; occipital fold is of whitish skin;

no throat fold; black collar plumage extensively covered with chestnut blotching.

Statistics for quantitative characters (N = 34)

Character X S.D.

[eli HSICGEID ~ SS earet a aee 25,6 mm 4,9

rier MMe ae he we ee st 25538 mM 7,4

Wearsoometatarsus length. . . . . . 2-2. 4 86,9 mm 4,8

Remletoasal width. . .« .« «7. 2 - is ww 9,9 mm 1,0

SeeeemeCMOtMbet ie oe sw we eS P 2:5. mm 0,6

MRS ESPOIMGIVCR I 0 ee Me a we 9,9 units 1,0

Ciaesemomtawiensth” . 2. 6s wk Oe 22,6 mm 4,7

TSS) PER era) ee en rn 38,0 mm 4,1

108 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character Xx S.D.

Crest central erates Wan 23.0 it cee we ee ce 25,9 mm 5

Crestibasall lemoehnas hz) 20 a sek ee 32,8 mm 2,4

AMtenlor CreSE.CURNMESS =) a en eee 5 OFF

POStEhIOk CRESL CURMIMCSS omar ye ye en 2 0,3

Dorsal black-collar™ 1 \: Aag)Ancde ke ee ee 214-9 4,8

Ventral black collar 22°...) 09? te Se ey aes 3258 VA 4,7

Occipital fold? a". iva Sh Ree? Se Dae D558) 4 96

Bar pate: -. GN Seok he ar Aare ae 0,0 mm 0,0

Pheo@at red: Byres tee jh iano. aa ee Be 0,0 % 0,0

Orbital Tee). Sey ee ae a a O10 0,0

Dorsab:spot mumber:: 45) Whee a ha pein ty aes ee 2354 4,2

Totalspot barbs.2 2. 2: 7 oe Se ee ee ee ee D52 0,7

otal within=spot,barbs “27 6) 2 a ee ae 35 0,6

Spot barb blMemess se .. Seen La lye Ge eae ae 0,8 0,5

@hestnut blotchisize. 9". 0 ee Ge ee 10,8 units 3,4

Chestnut blotch extent ei.) Ju uae Cott Vik arene eee 342-7 LST

Distribution

See Fisure 31.

Genus Acryllium

Acryllium Gray, 1840

Fig. 8(0G3)

Acryllium vulturinum (Hardwicke, 1834) °-

Fig. 1J

Numida vulturina Hardwicke, 1834: 52.

Description

The largest guinea-fowl species; no crown, nape, throat or cere adorn-

ments; occiput covered by short, dense, downy chestnut-coloured feathers;

rudimentary blue-grey wattles at gape; orbital and throat skin blue-grey;

well-developed collar hackle; body plumage spotted with vermiculations;

tarso-metatarsus usually with bump(s), scales pentagonal, not in rows; iris red;

outer margins of secondaries lavender; furcula blade-shaped; caecum longer

than 200 mm; abdominal plumage blue.

Statistics for quantitative characters (N = 43)

Character X S.D.

BUMeHeCne ee Ae 8 I) eM Ee ine oer ae 28,5 mm 2,0

Wing length Veen Ie, ae ee a eR 2S) 3 aE 10,1

Tarso-metatarsus leneth . oo. .° . . 2 « ©» WOG@aama 6,9

Wattle basakwidth' § 5-0. (Bey Mao! ee gaa & 7,2 mm EA

THE EVOLUTION OF GUINEA-FOWL 109

Character xX S.D:

RULE LEGS aye en 2,2 mm 0,6

Pee SeRUGLUTE MUINOCE fk Ok 1,8 1,4

Tarsal structure length een PRR eT Pn fcc la ee At 2,6 mm 1,9

Distribution

See Figure 10.

Genus Numida

Numida Linne, 1766

Fig. 8(0G10-14)

Numida meleagris meleagris (Linne, 1758)

Figs 1K, 12B(OTU4), 32(O0TU4), Table 6(OTU4)

Phasianus meleagris Linne, 1758: 158.

Numida ptilorhyncha var. major Hartlaub, 1884: 217.

Numida ptilorhyncha omoensis Neumann, 1904: 407.

Numida ptilorhyncha toruensis Neumann, 1904: 410.

Numida ptilorhyncha macroceras Erlanger, 1904: 97.

Numida ptilorhyncha neumanni Erlanger, 1904: 97.

Numida ptilorhyncha var. inermis DuBois, 1915: 18, 27.

Description

Larger than Guttera spp.; crown surmounted by a bony helmet; no occipital

or throat adornments; well-developed, rounded, blue gape wattles; nape

covered by short, downy feathers; orbital and throat skin blue; cere sur-

mounted by a tuft of cartilaginous bristles; collar black, finely barred with

white; body plumage spotted with vermiculations; tarso-metatarsus lacks

spurs, scales pentagonal, not in rows; iris brown; outer margins of secondaries

banded black and white with vermiculations; furcula blade-shaped; caecum

less than 150 mm; abdominal plumage spotted with faint vermiculations.

Statistics for quantitative characters (N = 311)

(See Table 6 for statistical comparisons with neighbouring subspecies.)

Character xX SD:

PRI eS ee em A 25,0 mm 7

See oe AS as 8 2622 om (ey

Masso-mictatatsus length 2. . . . 6 we a 84,3 mm 532

SeeteomoasolewiGth., 9-0 3. 6 we we 14,1 mm 7

Pe mermICI ah beh a 13,5 mm 3,0

Pemrerperecnt ole 2 2 kk lt ee wed Sal oy, Da

mcimicmMROMmAINCHOIA . 5. a 22,6 mm 10,2

Pemiirencanicneti: . 2 6k ee el 11,8 mm 15

memmemcenttalhereht .-.. 2). 8 oh ee 11,7 mm 8,0

110 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character xX S.D.

elmet basalvlemegin . 5995 e incu eee ee 1952 mim 3:0

Helinet thickness’ 42. a ee a 6,4 mm ei

Cerestructunce ieneth’ . % oa, bee eee 6,0 mm 3,6

Cere structure thickness: == <2 3) ea eee 1,1 mm 0,4

Collar plumage. . ee GIRS, eee, Wade ee ge vee Ds 0,6

Nape filoplume iederie sng ee OP ays ta: eee 14,9 mm ead

Nape filoplume anteroposterior coverage . . . . 3853 fA 17,4

Nape filoplume lateraltcoveracey =) 2) | ae 955657 5,7

Nape filoplume density . . . ir, ate See Ds 0,3

Secondary remex outer web Gemniculation eRe 558 0,7

Wine Covert barrin® | 9 7 2 ee ee eee 155 0,7

Dorsalvermiculation e+ 509) Seka 5 eee ee 3,0 0,7

Worsal spot Sizee "9s Pe eae ths) ee 18,8 units 7,

Distribution

See Figure 46.

Discussion

Two taxa, N. m. strasseni (Reichenow, 1911: 82) and N. m. blancoui

(Grote 1936: 158), have been described from the region in which N. m. galeata

and N. m. meleagris meet. Specimens attributed to these taxa are invariably

intermediate between galeata and meleagris in discriminant analysis, and their

collection sites fall in the region of high total COV between the parental forms

(Fig. 45A). They are treated as intergrades.

Numida meleagris sabyi Hartert, 1919

Figs 1L, 12B(OTU1), 32(;0TU1), Table 7/;0TU1)

Numida sabyi Hartert, 1919: 69.

Description

As N. m. meleagris, except gape wattles red; nape featheration long

filoplumes restricted to the mid-dorsal line; no cere adornment.

Statistics for quantitative characters (N = 4)

(See Table 7 for statistical comparisons with neighbouring subspecies.)

Character X S.D.

Bll enechifher «sy ts.-° ok team ek) Gee ce ne eee 24,0 mm 0,8

Wing lenothit “00 40 ave - t Ge 6 oe SO onoame 10,3

iarso-metatarsusiiensthy «5 “aes! Le an ee 88,3 mm all

Wiattle basaliwidtiie= Gx tan Wea “es eA woe, ee 15,5 mm 1,0

NVacleleneuhe ay ae A Pe a ae 17,8 mm a7

WiatelespemceminOlUGy e: “al Ps ee ee ge 0,0 % 0,0

THE EVOLUTION OF GUINEA-FOWL 11]

Character Xx S.D.

memctinontanieneth . . . .« . . « « s . 22,5 mm 3,0

Pemvencamiemetm . 2 i fo bd 8 Sl 15,0 mm G35

Eemicwcenttal hersht 9 16,3 mm 26

feeimeebusaliiiensth 8. - 2. 6k oe ke 20,0 mm 3,4

BMMCINEMIGKMICSS 2 5 kk a 5,3 mm 0,5

Mecermucimuneieneth 2... 6k kee 0,0 mm 0,0

Mrnoesmctune MMICKMESS 2. 6s ee 0,0 mm 0,0

Collar plumage. . OE Ok es Se ar aA e 25 0,6

Nape filoplume lencia . es ee fe he ot) Lee 34,0 mm Tel

Nape filoplume anteroposterior coverage . . . . 100,0 % 0,0

iINapemloplume lateral coverage . . . . . . . SOA 17,9

Nape filoplume density... Si Se Nails Ate 3,0 0,0

Secondary remex outer web mexmiculanon Se ie 3,8 eS

Dyameecovers barming . . . . 2. .« . ws ws 0,0 0,0

Dersaiayenmiculation . .. . . .« 2. 2 wo we 29 0,6

DOES) SOCE SZ en 7 Se. (Os5)

Distribution

See Figure 46.

Discussion

Although this taxon is represented by only four specimens, its validity

was upheld owing to its isolation and correspondence with criteria set for

subspecies.

Numida meleagris galeata Pallas, 1767

Figs 1M, 12B(OTU2), 32(0TU2), Tables 6-~7(OTU2)

Numida galeata Pallas, 1767: 13, 15.

Numida marchei Oustalet, 1882: 1.

Description

As N. m. sabyi, except smaller with collar plumage grey to blue-grey.

Statistics for quantitative characters (N = 137)

(See Tables 6-7 for statistical comparisons with neighbouring subspecies.)

Character xX S.D.

PCCM me ee) we ae) Ges 22,3 mm 1,8

eMPMCMPOM NTS ee ee a we S| 624 im 14,6

emsopmetalansusseneth 2... ee RR 81,8 mm 6,4

ReeilerbaSalewiGd . 6 ee 14,2 mm 1,9

SES ead 1 a ee re 15,5 mm 29

SremelespemcentONIG a Sony, 4,3

112 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character X S.D.

Felmet irontallenethy 4 ee ee ee 15,5 imam 5,8

iclmet reandenethh) cays. a. 15) wee ee re 7,0 mm 259

Helmet central heisht (4 5. e 5,8 mm 2,9

Helmet-basaljlengthe 4) 3) 20 2 ee ee 18,6 mm 353

Helmet thickmesse 0." SV ee eo a ee ae eee 4,9 mm He)

Cére Structumevlem atlas 1) iy pa ee ce ee 0,3 mm 0,9

Gére structure thickmessiaur.. a5 ees, ee ee eee 0,3 mm 1,6

Collar plumage. . A eae ts yy Re cg eae 0,1 0,3

Nape filoplume leva. LaO Stee in + eee 20,4 mm 4,3

Nape filoplume anteroposterior coverage . . . . DONA 16,0

Nape tloplume lateral coverage’ =). . 2 Seeee TOSS 5303

Nape filoplume density . . ee 2,8 0,4

Secondary remex outer web Tecmiculagan <i Cae 4,4 13

Wine covert baring.) 29 5000 7 ee ee ee 0,6 0,2

Dorsalbvermiculation ». 92). tao ee 3,6 0,6

OKSANSPOb SIZE hy ae et okra Sh 8, ee ene 12,2 units 3,8

Distribution

See Figure 46.

Numida meleagris somaliensis Neumann, 1899

Figs 1N, 12B(OTUS), 32(O0TUS), Tables 6, 8(OTUS)

Numida somaliensis Neumann, 1899: 25.

Description

As N. m. meleagris, except gape wattles blue with red tips; cere tufts

much longer and more numerous; and nape featheration long filoplumes

restricted to the middorsal line.

Statistics for quantitative characters (N = 44)

(See Tables 6, 8 for statistical comparisons with neighbouring subspecies.)

Character Xx S.D.

ler SO eee he cian he eile Rime Me 25,1 mm 1,4

Wine Tenet 3) Oe ee ment 11,0

arso=metatarusmenctiny-) 2 eo eh 2 fe ee 84,2 mm 4,2

Wattle basanites. « a GOW say Bees © cn) ie 13,6 mm 2,4

Wrattlewemetlag Bete. es-C iii yee ie Yen We ee 14,6 mm 353

Wattle per cemtiOlicaine <2 trian 26th ie ne 87,5 7% 14,9

elimet trontalliengtiay. 2 oS re a 20,0 mm 7,6

RSet reat eMm tite OG, ee ee ie Pale! an eee i 10,3 mm 6,0

elmer centralthieicht a ee ee 12,9 mm 12,3

THE EVOLUTION OF GUINEA-FOWL

Character

Helmet basal length

Helmet thickness

Cere structure length

Cere structure thickness

Collar plumage .

Nape filoplume ange

Nape filoplume anteroposterior coverage

Nape filoplume lateral coverage .

Nape filoplume density

Secondary remex outer web Tenmiculation :

Wing covert barring

Dorsal vermiculation

Dorsal spot size

Distribution

See Figure 46.

17,4 mm

6,6 mm

13,3 mm

1,0 mm

3,0

15,8 mm

Aled: 7%

53,01 74

2,8

0,8

2,4

23,6 units

Numida meleagris marungensis Schalow, 1884

Figs 10, 12B(OTU6), 32(0TU6), Tables 6-7, 9(;0TU6)

Numida coronata marungensis Schalow, 1884: 105.

Numida marungensis maxima Neumann, 1898: 21.

Description

S.D.

153

156

0,1

0,7

20,7

34,0

0,5

2,0

0,6

0:9

6,5

The largest subspecies; helmet characteristically thicker and longer basally;

long gape wattles, blue with red tips; no cere adornment, and wattles pointed.

Statistics for quantitative characters (N = 97)

(See Tables 6-7, 9 for statistical comparisons with neighbouring subspecies.)

Character

Bill length

Wing length

Tarso-metatarsus length

Wattle basal width .

Wattle length

Wattle per cent blue

Helmet frontal length .

Helmet rear length .

Helmet central height .

Helmet basal length

Helmet thickness

Cere structure length

Cere structure thickness

26,7 mm

283,5 mm

90,2 mm

11,8 mm

16,6 mm

(ERO YA

47,7 mm

13,4 mm

18,7 mm

39,2 mm

12,4 mm

0,3 mm

S.D.

1,4

4,7

155

3,1

8,5

3,8

337

256

0:7

0,9

114 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character X S.D.

Collar plumage. . Uae cay ter mayer gre Ps 2) 0,7

Nape filoplume ‘engi. Ae fl, wore Ai see a 26,3 mm 3;3

Nape filoplume anteroposterior suas Se eges oe T1533) Ya se)

iINape hlloplumelateralcoverase =) eee 2901 / 75S

Nape filoplume density . . ee eee D5) 0,4

Secondary remex outer web sepinioulle on Zncleny aera 0,9 1,0

Wine covert Daring. «2. ¢ 2. 2 . Re eee 0,0 0,0

Dorsal vermiculation <2 S45 32> 45 5. eee LS 0,8

IDOrsal SPOtsizZeug —o— as df “wees ee ee eee 23,5 units 4,5

Distribution

See Figure 46.

Discussion

N. m. frommi (Kothe 1911: 13) and N. m. rikwae (Reichenow 1900: 40)

have been described from the region between N. m. marungensis and N. m.

mitrata, and N. m. callewaerti (Chapin 19326: 1) from the region between

N. m. galeata and N. m. marungensis. Specimens attributed to these three forms

are invariably intermediate between two subspecies in discriminant analyses,

and their collection localities fall within regions of high total COV (Fig. 45A).

They are treated as intergrades.

Numida meleagris reichenowi Ogilvie-Grant, 1894

Figs IP, 12B(OTU8), 32(;0TU8), Tables 6, 8;OTU8)

Numida reichenowi Ogilvie-Grant, 1894: 536.

Description

Similar to, and nearly as large as NV. m. marungensis; except wattles rounded

and red, helmet taller and sabre-shaped, and nape featheration less dense.

Statistics for quantitative characters (N = 121)

(See Tables 6, 8 for statistical comparisons with neighbouring subspecies.)

Character x S.D.

DiMlMlenote A? WP vad captinae cahpsal le Seeeae Mere Semeye 24,7 mm 1,6

Wing length a ee eee ee Mee 252.2) ean 9,1

iharso-metatansus lenethe . a. oly Wen oe ee 90,7 mm 11,4

Wattle basalivviditii.,, “s se “aia, ae a, av aah ve os 11,9 mm 1,4

Wrattlelenetht yo" ac fe «Gallen. eas ahha <P eg, (ee 12,8 mm Vis)

Wattle percent ble. ois )aer ar at acer ae ae We) 7. 29

Fieinettromtallemeth 4. 4) sw sa a a 44,2 mm 10,9

elimetrearlenet se =) oO. = Sue = Suh 29,5 mm 9,8

Ficlimet centralhecicht. 2. = “ae = oS 4 eee 29,5 mm 9,0

THE EVOLUTION OF GUINEA-FOWL 115

Character Xx S.D.

Meumetbasaliensth =. . 2. . Ww . ww ee 24,9 mm 29

BrcmMeetMeKNeSSy 6 ee we 8,3 mm 17

Memecmuctineleneth 2 2. eee 0,7 mm 1,8

Meremsmucture thickness . . © . . 2 os es 0,4 mm 0,8

MOMMIES 218) 0,5

iNapemloplumelensth. . . . ee ee eA 20,5 mm 3,6

Nape filoplume anteroposterior coverage . . . . O94 49,2

iNepealoplume lateral coverage . . . . . «= . 30,0 % DAES

Nape filoplume density . . vee Sg ote 2 0,6

Secondary remex outer web o iniculanion eee ee 23) ee)

meseovernt barrime 2 6 6 kk ke O2 0,5

Moralvermeculation . . . «1. . .« .« « . 2,4 0,7

AGSESOINSIZE 5 ek 23,2 units 35

Distribution

See Figure 46.

Discussion

N. m. intermedia (Neumann 1898: 21), and N. m. ansorgei (Hartert 1899:

331) have been described from the transition area between N. m. meleagris

and N. m. reichenowi. N. m. uhehensis (Reichenow 1898a: 88) has been described

from the transition area between N. m. mitrata and N. m. reichenowi. For

reasons given in other such instances, these forms are treated as intergrades.

Numida meleagris mitrata Pallas, 1767

Figs 1Q, 12B(OTU7), 32(O0TU7), Tables 6, 8-10(OTU7)

Numida mitrata Pallas, 1767: 18.

Description

Smaller than the last two subspecies; phenotype as N. m. marungensis

except that helmet less well developed.

Statistics for quantitative characters (N = 293)

(See Tables 6, 8-10 for statistical comparisons with neighbouring sub-

species. )

Character 4 S.D.

ES Teng." i on re eer 25,7 mm KS

Ser MeMOUME I ele wt ae | 2A 8,2

Narno-metatarus length 2... ee 86,1 mm 4,2

Breulemoasa@lwidti. 2 6 ow ek 10,9 mm 163)

Pr micwlemetni ee ek 16,8 mm 3,0

Boil pemicentDMe Fo 2 wk te 64,6 % O83)

116 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character X S.D.

Helmet irental lemeth 2. 2) Go er. eee 38,3 mm SLs

Helmet neartenetine .. Wie os a ee 18,8 mm Syl

Felmet centralheight sede) ae ee ee 19,4 mm Je

Ricimetebasalmlengtha “awe i a ae Se 23,3) taiia 3,8

FISiMetthickMessiyos 7% AG eo cok ew Oe het ee 8,3 mm 2,0

Cere Structinenleneth 4s, 1 ee ee 0,2 mm 0,8

Gere structirethickmess. = 319 > 2955 Wale 0,2 mm 0,7

Collar plumasey, We 15. 20 Sy - a & Sl 0,6

Nape tiloplumetensth: © 2 <2 ih) «as =) sae ae 23,0 mm 38)

Nape filoplume anteroposterior coverage . . . . 59a oeT

Nape floplumedateralicoverase: 9 5 =) = “eee 23,0074 8,3

Nape filoplume density . . h5 cule DES 0,7

Secondary remex outer web senmieniagon £5) Sa eae iPS 3

Wineecovert Darume 2 oy 20a ue Ra) el eae eee 0,0 0,0

Dorsal vermiculaGion YO. Seis Le 0,7

Dorsal Spotsize: i+ TU FeO eee ee eee 19,4 units 4,5

Distribution

See Figure 46.

Numida meleagris coronata Gurney, 1868

Figs 1R, 12B(OTU10), 32(0TUI10), Table 10(0TU10)

Numida coronata Gurney, 1868: 253.

Numida transvaalensis Neumann, 1899: 26.

Numida papillosa limpopoensis Roberts, 1924: 77.

Description

As N. m. marungensis, except smaller overall size; decidedly thinner and

taller helmet, collar plumage more streaked than barred with white.

Statistics for quantitative characters (N = 136)

(See Table 10 for statistical comparisons with neighbouring subspecies.)

Character xX S.D.

BMPR Ot ey ek td oe ge eh 2531 mama 7

Wing length a RE ig a ae aS oo Wh Ge | a 8)

Parso-metatansus lemethy “5.5 ° =@ i& Get ee 85,6 mm 552

Wattle basalewidthiv. et! Ly Ge ae 2 els PR 11,0 mm SS

Wiattle tenet ey ac acy Uae Be a Le Sk et 18,7 mm 3,2

Wale percenivbluc. .. etl ar aw An aS ol Se 10,0

Melmetirontaldenethy {18 se 7 Tete Oe 54,8 mm 9,6

icimetteamiemetiane ie “wel eo ou oo Awl ew) el” wah oe 25,1 mm 5,8

lncimet central mersat’ .. Fie eo e e 24,7 mm 5,3

Pelmetbasailenoen io Ge en ee a ee 25,8 mm Zo

THE EVOLUTION OF GUINEA-FOWL 117

Character x S.D.

BACMMGEEMIGKMESS Ph we 8,8 mm Oa

MGeESeRMCIURe (EMC oo ae 0,4 mm 0,7

Mcrcesenuctune tMIGKMESS. 2 5. kk 0,3 mm 0,4

Collar plumage. . RT ee eM i Pree na 4,2 Ji

Nape filoplume lene. Bee eS sia! ED ag chen 18,5 mm 4,6

Nape filoplume anteroposterior coverage . . . . 40,0 % 22,8

iINapewmleplume lateralcoverage . . . . . . . be 304 4,9

Nape filoplume density . . Te ios ee 0,8

Secondary remex outer web Penniculanion Wee, a Aue Des 1,4

Puiemcovert DaTTIN® 5 6 2 ok ye oe ek 0,0 0,0

Monstevermiculation 2. 5. 5 eee a 2,6 0,6

(OS, SSCE SZ a ae 20,4 units 5,6

Distribution

See Figure 46.

Discussion

Numida meleagris papillosa (Reichenow 1894: 145) has been described from

the region south of Lake Ngami, i.e. the transition area between NV. m. damarensis

and NV. m. coronata. For reasons given in other instances this form is taken to

be an intergrade.

Numida meleagris damarensis Roberts, 1917

Figs 1S, 12B(OTU9), 32(0TU9), Tables 9-11(OTU9)

Numida papillosa damarensis Roberts, 1917: 2.

Description

As N. m. coronata, except well-developed papilli at cere; collar spotted;

helmet less well developed.

Statistics for quantitative characters (N = 102)

(See Tables 9-11 for statistical comparisons with neighbouring subspecies.)

Character x S.D.

allo es a Wa 24,6 mm 1,6

Wing length Me er, eae erie ee TR ait ae oh Se anita 16

iiaeso-metatarsus lensth 2. 2... C. 83,4 mm 3,9

Piavlesmasalowidtly; <0 ee ee 1 noma 188)

SvciMememeihe fo 6 kk a 18,9 mm DES

Deaulcapemeent blue. 2. we eG 64,4 % 8,5

memichimontalicneth 2. 9.3 ke Bw eed 45,2 mm 1

PicMeincammlCnebhye NY yy he te we 22,9 mm 4,6

Incimcecembralheieht =. 0. os 8. Ge ak 22.3- mm 4,3

118 ANNALS OF THE SOUTH AFRICAN MUSEUM

Character x S.D.

Helmet basalilemeth@ 9% “4.5 eee 19,7 mm Syl

Flelmet UhickineSsisy © 0) 22) pen oe) ae ees ree 9,2 mm 1,3

Cereistructunelemethin.. Ua 4 eee oe 3,7 mm hs)

@eke Simuctunestinickmesss 2) ee ene 229 sia 0,3

Collar plumage. . git Htoy (ogee. cee 4,9 1,0

Nape filoplume lena - ee oe ee 12,3 mm 8,5

Nape filoplume anteroposterior coverage . . . . 24,7 % DD

Nape filoplumetlateralicoverace ene ON ay 8,0

Nape filoplume density. . ee 0,5 0,7

Secondary remex outer web ceaaioniauon es 358 1,4

Wins covert-barting ) 2 xe, ee 0,0 0,0

Dorsaltvermiculation’) 49.) 4). a ee 2,6 0,5

DorsaluSpOt:SiZe® ciens oo econ Se Se ee 22,5 units 8,5

Distribution

See Figure 46.

PHYLOGENY

PRIMITIVE AND DERIVED CHARACTER STATES

The following hypothetical primitive-derived sequences are postulated for

the characters listed in Appendix 1. The number in parentheses following each

character name is its number in Appendix 1.

Crown (1), occipital (2), nape (3), and throat (6) adornments

All francolins and guineafowl-phasianid hybrids have feathered crowns,

occiputs and napes. Very few francolins have crests, and then only rudimentary

ones. No francolin has a throat fold. Therefore, a naked crown, nape or occiput

is taken to be derived relative to the feathered condition. Among guinea-fowl

with feathered crowns, a well-developed crest of downy feathers is taken to be

derived relative to a short crest of downy feathers, and a well-developed crest

of bristly, erect filoplumes derived relative to a well-developed crest of downy

feathers. The latter assumption is based on the idea that a modification of a

downy, drooping crest to an erect bristly crest allows species or individual

recognition in social encounters. Also, among guinea-fowl which have these

regions unfeathered, any likely secondary elaboration of the unfeathered con-

dition (e.g. a helmet, long filoplumes, patches of skin of contrasting

colour, folds of skin) is taken to be derived, since these structures may be

adaptive in a social (individual and species identification) or physio-ecological

(thermoregulation, crypsis) context (Brown 1963; Reynolds 1977). Finally,

folded throat skin is taken to be derived relative to the unfolded condition.

THE EVOLUTION OF GUINEA-FOWL 119

Orbital (4), throat (5), and gape wattle (8) skin colour

In nearly all francolins, areas of naked skin on the head and throat are

red. Therefore, red is taken to be the primitive condition relative to any other

colour for the above characters.

Gape adornment (7)

No francolin has a gape wattle, and guineafowl-phasianid hybrids have at

most rudimentary wattles. Therefore, well-developed gape wattles are taken

to be derived.

Cere adornment (9)

No francolin or guineafowl-phasianid hybrid has an adornment at the

cere. Therefore, any cere adornment is taken to be derived.

Collar (10), body (11) and abdominal (18) plumage

The plumage of francolins and guineafowl-phasianid hybrids is usually

dark, often streaked, barred or vermiculated with lighter colours. Moreover,

none of the above possesses an elaborate collar-hackle. Therefore, among

guinea-fowl, white or blue plumage is taken to be derived relative to dark

plumage, and spotted plumage derived relative to barred, streaked or vermicu-

lated plumage. Among guinea-fowl with spotted plumage, spotted plumage

without peripheral vermiculation is taken to be derived relative to that with

vermiculation. Also, regardless of plumage pattern, an elaborate collar hackle

is taken to be derived relative to an undifferentiated collar.

Tarso-metatarsal scales (12) and adornment (13)

All francolins have their posterior tarso-metatarsus covered by small

imbricated scales in rows. In nearly all francolins, the male at least possesses

spurs. Therefore, among guinea-fowl any modification of the spurred condition

(e.g. naked tarsi or tarsal bumps), and any scalation pattern other than that

described above are taken to be derived conditions.

Tris colour (14)

Nearly all francolins have brown eyes. Therefore, among guinea-fowl, a

red iris is taken to be derived.

Outer margins of secondaries (15)

In most francolins the secondaries are brown to grey, sometimes faintly

streaked or vermiculated with white. In only a few francolins do the outer

margins of these feathers appear to be strikingly different from the general

body plumage. Therefore, among guinea-fowl, secondaries that contrast with

the general body plumage pattern, 1.e. character states 2-4, are taken to be

derived.

Furcula (16)

All francolins have blade-shaped furculas. Therefore, a hollow furcula is

taken to be derived.

120 ANNALS OF THE SOUTH AFRICAN MUSEUM

Caecum length (17)

In francolin and guinea-fowl of the genera Numida and Guttera caecum

length is about 16 per cent that of the large and small intestines combined

(Beddard 1898). There is no information on caecum length for Agelastes spp.

In Acryllium yulturinum, the length of this organ is more than 23 per cent

that of the intestines (Beddard 1898). Therefore, among guinea-fowl a relatively

long caecum is taken to be derived.

PHYLETIC ANALYSIS

The most parsimonious guinea-fowl phylogeny based on primitive-derived

character sequences is given in Figure 47. In this figure, the four shared derived

character states comprising character suite | delineate the subfamily Numidinae

from francolin-like phasianids. These derived character states, common to all

guinea-fowl taxa (Ghigi 1936), and their primitive counterparts (in parentheses)

are:

1. large size (small size);

2. third and fourth sacral vertebrae with robust transverse processes (only

third vertebra with such a process);

3. second metacarpal lacks a backward process (process present);

4. largely naked head with at least rudimentary wattles (feathered head, and

no wattles).

Agelastes sp., with their preponderance of primitive character states,

probably more closely resemble the proto-guinea-fowl than does any other

member of the subfamily. These two species share only one derived character

state, white abdominal plumage. Adult A. niger (Fig.. 1B) shows primitive

character states for all characters investigated except for its naked occiput.

The white abdominal plumage linking it to A. meleagrides (Fig. 1A) is present

only in juvenile birds. It is also possible that the additional derived character

states attributable to Agelastes due to the naked head and neck of A. meleagrides

are the result of convergent evolution, and do not suggest any closer affinity

to other guinea-fowl genera.

Character suite 2 consists of three shared derived character states (spotted

body plumage, non-imbricated, pentagonal scalation of the posterior tarso-

metatarsi, unspurred tarso-metatarsi) which link the three remaining genera.

Character suite 3 consists of three shared derived character states (unfeathered

crown, blue orbital and throat skin) which link Acryllium (Fig. 1J) and Numida

(Fig. 1K-S). Character suite 4 consists of four shared derived character states

(folded occipital skin, white outer secondary margins, body plumage spotted

without peripheral vermiculation, hollow furcula) which link the two Guttera

species (Fig. 1C—I), and distinguish them from Acryllium and Numida. Character

suite 5 consists of four derived character states (helmeted crown, naked occiput,

well-developed wattles, cere with tufts or papilli) which distinguish Numida

Fig. 1K-S) from Acryllium (Fig. 1J). Character suite 6 consists of three

THE EVOLUTION OF GUINEA-FOWL 194

Agelastes Guttera Acryllium Numida

Fig. 47. An hypothetical phylogeny for the Numidinae. Each monophyletic lineage is charac-

terized by a suite of derived character states (Mf), the primitive counterparts of which (LJ)

are found in members of the co-ordinate sister-group.

derived character states (hackled collar, long caecum, red eye) which distinguish

Acryllium from Numida.

SPECIATION

ORIGIN AND DERIVATION OF THE NUMIDINAE

Several hypotheses have been offered concerning the origin and evolution

of the Numidinae. Ghigi (1936) states that the proto-guinea-fowl originated

in Africa, and probably was derived from a francolin ancestor. This hypothesis

is based on the fact that guinea-fowl are found only in Africa (Arabian and

Malagasy populations being probably the result of introductions by man),

and phenotypic similarities between the most primitive guinea-fowl (Agelastes

spp.) and Francolinus spp. Cracraft (1973), in support of his hypothesis of a

Gondwanaland origin for the Galliformes, suggests a North American origin

from New World quails. Olson (1974), suggests a possible Asiatic origin from a

pheasant-like bird, based on analysis of a single Eocene fossil femur from

Mongolia. This bone is intermediate in shape between femurs of extant

pheasants and Agelastes niger. However, it is only about 70 per cent the length

of an Agelastes femur, i.e. well within the size range of many extant francolins

(Mackworth-Praed & Grant 1952, 1962, 1970).

There is no compelling evidence favouring any of these three hypotheses.

When, and if, sufficient information comes to the fore, it will probably support

122 ANNALS OF THE SOUTH AFRICAN MUSEUM

elements of all three. Accordingly, the working hypothesis taken herein is:

the Numidinae are derived from a francolin-like Asiatic phasianid; but evolution

and radiation of extant guinea-fowl has occurred solely in Africa.

EVOLUTION OF GENERA

The first opportunity for colonization of Africa by Asiatic faunal elements

arose with the mid-Miocene (c. 17-18 m.y.B.P.) union of the African and Asian

plates (Axelrod & Raven 1978). At this time, forest was much more extensive

in Africa than at present, possibly even exceeding limits depicted in Figure 5.

However, the colonization corridor connecting these two continents was

covered by relatively arid savanna vegetation in the mid-Miocene, and has

almost certainly not had more lush vegetation since then (Axelrod & Raven

1978). Thus, any Asiatic ancestral guinea-fowl was probably a bird which

lived in savanna habitat, and, upon its arrival in Africa, encountered vast

forest, and much less extensive savanna adaptive zones. Since Agelastes spp.

possess so few derived character states (see above), it is likely that radiation of

proto-Agelastes into the forest took place soon after colonization.

With the joining of the two continents, the mild climate that favoured

widespread forest vegetation throughout the late Cretaceous and early Tertiary

began to deteriorate. Africa became progressively more arid; and savannah and

desert-steppe biomes expanded, at the expense of forest, throughout the latter

Miocene and Pliocene (Axelrod & Raven 1978). Moreover, this period was

characterized by widespread uplifting, rifting and tectonic activity adding

considerable topographic diversity to the continent, and partitioning its

expanding and contracting biomes (Axelrod & Raven 1978). These conditions

would have favoured radiation in and into expanding forest-edge, savanna

and desert-steppe biotopes; and it is possible that proto-Guttera, Numida, and

Acryllium were the result of such Mio-Pliocene radiations.

EVOLUTION OF SPECIES AND SUBSPECIES

After the uniformly arid Pliocene, the relatively rapid wet—dry climatic

fluctuations and continuing rifting and mountain building during the Pleistocene

provided additional opportunities for radiation in Africa (Moreau 1966;

Hamilton 1974; Livingstone 1975; Axelrod & Raven 1978). Expanding

forest and wetlands in relatively moist phases would have divided savanna

biome into more or less isolated tracts, and restricted desert to relatively small

refugia in Somalia, northern and south-western Africa. If wet-phase forest

and wetlands were distributed as in Figure 5, Acryllium vulturinum (Fig. 10)

and sub-Saharan subspecies attributed to Numida meleagris herein (Fig. 46)

would have had the opportunity to diverge in isolation, since portions of their

present-day ranges are encompassed by isolated tracts of savanna and desert—

steppe refugia. Also, expanding forest may have favoured a second radiation

into relatively widespread lowland forest, culminating in Guttera plumifera.

During arid phases, forest-living guinea-fowl, and Moroccan N. meleagris,

THE EVOLUTION OF GUINEA-FOWL 123

would have been restricted to island-like refugia. If vegetation were distributed

as in Figure 6, N. m. sabyi (Fig. 46), Agelastes meleagrides (Fig. 10), A. niger

(Fig. 10), Guttera plumifera plumifera and G. p. schubotzi (Fig. 14) and, to a

lesser extent, forest-edge taxa (subspecies attributed to G. pucherani) had the

opportunity to diverge in allopatry, since portions of their present-day ranges

would have had access to refuges of suitable biotope. Thus, we need look no

further than the Pleistocene for biogeographic events which could have allowed

allopatric evolution of extant guinea-fowl species and subspecies.

BIOGEOGRAPHY

RESULTS

A map of hypothetical African avifaunal zones drawn from evolutionary

patterns found in guinea-fowl is given as Figure 48. Zonal boundary lines in

this map agree remarkably well with those in Chapin’s (1932a) faunal map

(Fig. 49) based on distribution patterns of “many species and races of birds’,

and with species and subspecies boundaries in distribution maps of selected

francolin species and subspecies (Figs 50-51). In the hypothetical avifaunal

map, subregion boundaries are those of the relatively phenotypically homo-

geneous (i.e. no subspecies) genera, Agelastes and Acryllium (Fig. 10). Provincial

boundaries coincide closely with those of species, and district boundaries with

those of subspecies (Figs 14, 31 and 46). Chapin neither lists the taxa whose

ranges form the basis of his map, nor specifies criteria used in distinguishing

subregions, provinces and districts. The hypothetical map differs markedly

from Chapin’s in that it:

1. restricts the forest subregion (commonly labelled @) in Figs 48-49) to an

area somewhat larger than Chapin’s province no. |;

2. divides his province no. 4 into eastern and western, rather than northern

and southern districts;

divides his district 1B into two districts (1A and 1B in Fig. 48).

reapportions territory within the commonly labelled province no. 6;

divides his district no. 6E into two provinces (7 and 8 in Fig. 48);

does not recognize montane provinces.

The only differences between the hypothetical map and the francolin

distribution maps are relatively minor shifts in boundary lines, and a still

finer subdivision of districts by francolins. Also, district boundaries, which

Separate guinea-fowl subspecies, often delimit francolin species.

DISCUSSION

It is impossible to resolve differences between the two avifaunal maps in

zonal boundaries and hierarchical assessments, since Chapin does not specify

the data base and methodology underlying his map. His decision to unite the

continuous block of lowland forest with the surrounding forest—savanna

mosaic to form subregion @) (Fig. 49) is probably due to the abundance, in

124 ANNALS OF THE SOUTH AFRICAN MUSEUM

exes $Ubregions

provinces

lA ——-—- districts

Fig. 48. Hypothetical African avifaunal zones based on evolutionary patterns found in

guinea-fowl.

the latter, of relict patches of lowland forest and gallery forests, which provide

suitable habitats for forest birds. Indeed, when distributional data are lacking

or equivocal, Chapin (1932a) and other zoogeographers (e.g. Davis 1962;

Moreau 1966) seem to have relied on the distribution of vegetation as a predictor

of bird distributions. The division of Chapin’s province no. 4 and district 1B

into east—west districts in the hypothetical map is due to effects (on guinea-fowl

evolution) of probable past forest—wetland and savanna barriers that bisected

these zones during the Pleistocene (see Figs 5-6). The reapportionment of

territory to districts in the commonly labelled province no. 6 may reflect a lack

of clear-cut avian distributional patterns within that province, a possibility

already suggested by Benson & Irwin (1966). The necessity of partitioning

Chapin’s district no. 6E was anticipated by that author (Chapin 1932a: 89),

and has been done by other authors (Moreau 1952; Davis 1962). The lack of

montane avifaunal zones in the hypothetical map is certainly due to the sub-

montane altitudinal limitation of guinea-fowl.

THE EVOLUTION OF GUINEA-FOWL (25

@) == subregions

] —— provinces

1A ---7— districts

Fig. 49. African avifaunal zones based on analysis of bird species and subspecies distributions

(after Chapin 1932a).

CONCLUSIONS

The results of comparisons of the hypothetical avifaunal map with Chapin’s

map and francolin distribution maps suggest three tentative conclusions, which

can serve as hypotheses in future evolutionary and biogeographic studies.

1. Distribution patterns found in guinea-fowl can be used as models for

broad patterns exhibited by African bird species and subspecies other than those

dependent on montane habitats.

2. At least some francolin species and subspecies have evolved as a result

of factors that have been important in the evolution of guinea-fowl.

3. Physical and ecological barriers which have allowed only subspeciation

in guinea-fowl have been sufficient to bring about speciation in francolins.

The first hypothesis is being tested by a cluster analysis (Hagemier &

Stults 1964; Sneath & Sokal 1973) of 119 equally sized areas of Africa according

to 1 099 passerine species and well-marked subspecies in Hall & Moreau (1970)

(Crowe in prep.). The preliminary results of this study, summarized in Figure 52,

are consistent with that hypothesis. The second and third hypotheses can be

tested, if patterns of character variation in francolin taxa are analysed using

126 ANNALS OF THE SOUTH AFRICAN MUSEUM

——Subspecies boundaries

— species boundaries

Y extralimital areas

=| zones of sympatry

Fig. 50. Distributions of selected non-forest francolin species and subspecies (after Hall

1963 and Mackworth-Praed & Grant 1952). A. Francolinus bicalcaratus ayesha. B. F. clapper-

toni. C. F. icterorhynchus. D. F. leucoscepus. E. F. afer cranchii, intercedens and harterti.

F. F. hildebrandti. G. F. hartlaubi. H. F. adspersus. 1. F. natalensis. J. F. capensis.

methodology outlined herein, and the results are compared to those for

guinea-fowl.

SYNTHESIS

Taxonomy, phylogeny, speciation and biogeography are intimately related

aspects of guinea-fowl evolution. Their necessary separation under different

headings in the present study has been somewhat detrimental to the under-

standing of each. Accordingly, this section attempts to synthesize the author’s

conception of evolution in guinea-fowl (summarized in Fig. 53).

Guinea-fowl are characteristically African birds. Although the likely

ancestral guinea-fowl was an Asiatic francolin-like phasianid which could live

in arid savanna habitat, the evolution that has led to extant guinea-fowl taxa

occurred solely in Africa. Moreover, biogeographic patterns derived from

THE EVOLUTION OF GUINEA-FOWL 127

a ——

Ee Re,

<r Gee eae

——— Sy ay

= Francolinus ahantensis

ESTE squamatus squamatus

[J F.s. schuetti

Fig. 51. Distributions of selected forest-living francolin species and subspecies (after Hall 1963).

guinea-fowl species and subspecies boundaries closely parallel broad patterns

found in African birds as a whole. The Asiatic ancestral guinea-fowl probably

traversed the arid-savanna corridor linking Asia and Africa soon after the

mid-Miocene union of the two continents. This savanna-living bird encountered

an African continent dominated by forest, possibly unoccupied by potential

competitors. Such conditions favoured radiation into the forest, and it is

likely that Agelastes, the most primitive (i.e. most francolin-like) guinea-fowl

genus is a result of an early radiation into forest.

Relatively soon (on a geological time scale) after this successful invasion

of forest, the climate of Africa became more arid. Throughout the latter Miocene

and Pliocene savanna and desert biomes expanded at the expense of forest.

Such a situation favoured radiation in non-forest biomes, and into the expanding

forest-edge biotope, and it is possible that proto-Numida, Acryllium and

128 ANNALS OF THE SOUTH AFRICAN MUSEUM

@) ee subregions

1 —— provinces

1A ----— districts

Fig. 52. Hypothetical African avifaunal zones based on a cluster analysis of 119 equally

sized blocks of Africa according to 1 099 species and subspecies in Hall & Moreau (1970).

Guttera were the result of such radiations. The relatively uniformly arid con-

ditions of the Pliocene subjected these four lineages to strongly divergent

selective pressures, and it is probable that the genera recognized herein were

already well defined at the beginning of the Pleistocene.

The arid climate of the Pliocene was replaced by a fluctuating wet—dry

climate in the Pleistocene. These climatic fluctuations had profound effects on

the distribution of African biomes. During moist phases, the forest biome

expanded considerably beyond its present extent, partitioning non-forested

biomes into more or less isolated tracts. Desert biome was confined to relatively

small areas, and savanna biome bridged the western Sahara, allowing dispersal

of N. meleagris into north Africa. Sub-saharan subspecies of N. meleagris are

the result of divergence in these wet-phase isolated tracts. Also, expanded forest

during mesic phases could have allowed a second radiation into lowland forest,

culminating in Guttera plumifera. During arid phases, the forest contracted

into island-like refugia, and N. m. sabyi was isolated, and presumably diverged

from sub-Saharan populations. The species Agelastes meleagrides and A. niger,

129

"[MOJ-eouINS JO ATO\SIY ATBUOTINJOAS [eoNoyJOdAY UY ‘ES “SIF

1S34¥043 JO NOITLOVYLNOD CNV NOISNWdxXa GIdvY NOI LOWYLNOD WOWIXWW lv 1S3404 LOVYLNOD OL SNI9DSG 193404 1SauO4 AVaNdS30IM

THE EVOLUTION OF GUINEA-FOWL

&&4 44

Ya9IN

$301 49W313W

Tauwnoda ‘d *5

Vivauvd 'd ‘9

Taalv19s ‘d 'S—) INWYSHONd

TXNVAYNRA *d *5

INVUSHONd 'd '9

IZLOGNHDS ‘d '9

VuasIWNId 'd ‘9

1S3404 G4ALIVYLNOD 40

SGOIY3d ONIYNG NOTLVID3dS

QNV S3ON39YN3SAIC 91 Yl vdo11V-

1S3a4y04

GALIVYLNOD 40 SAOIYAd

SNIYNG SAIDAdSANS VHSsLINS

dO JINS9N3AIC DIYLvd011V-

1s3¥04 daqNvdxa Jo

SGOIYad ONTYNG WHsdIWNId '9

NI ONILVNIWIND SWOIA 1S3y04

OLNI NOILVIGWY AYVONOD3S-

daN1430-114M SIISVTASV-

TMOAVANINS-OLOYd WONS

JINASYNFAIC JVLLII-

NOTLWIYVA

Q141DadSANS Tay WW-

“dS Wu311NS daNI430-T13M 3NO-

3dO1O1@ 3903-1S3404

ANOHSNOYHL TWSY3adSId-

SV149V-O010Ud

Ol 3S1Y ONIAIS

vas 1inS-O10ud

Ol 3SIY ONIAIS

3dOLOI1d 3903-1S34N04

OLNI NOILVIGVY

aWOId 1S3yO4

OLNI NOILVIGVY

QINVISWHd JILVISV

3W11-NITOONVYS

ONTAI1-VNNVAVS

3dd31S-1uasad ie

vuNvAvs //]

isauod fi

NOILVL393A dO

NOITLNATYLSIG FHL

VOIuAV SAZINO109

1MO4VININS-OL0Ud

Sassa20ud ONV

SLN3A3 AYWNOILN1OAS

STSNST 1WWOS ‘'W

TMONSHOISY *W NOI LOVYLNOD

VIVULIW 'W 1S34¥04_ JO SdO0] Yad AWWITD NOILVIYVA-

VLVNOUWOD 'W SNIYNG IAGVS 'W 'N 4O @3N 1430-113 VOIWNN-OLOYd AWNIT9 NOILVIUWA-

ORR 4" JONFONAAIC JIYLVdOTIV- STYSVITIN VOTHN- OL 3S1Y ONIAIS S3WOId LS4YO4-NON

BOTSHAd YR OSs! 1s3¥o4 3WOIG VNNVAVS 3WOI VNNVAVS NI NOTLVIavu-

SISNZONNYWW 'W (30NVdX3 43O SGOIYad ONIYNG OL NOILVidvav GNV OLNI NOILWIGVY

Viva 1v9 'W S3193dS8NS NVYWHYS-aNs 4O “1NOHSNOYHL TWSuadsia-

IAdvs ‘W SINASYSAIG JIelWdOT1V-

SIYSV313W 'W

G3N1430-113M WNITTANDV-

SWOIG 3dd31S-143S30 WIT TIXYSV-OL0ud OL

SWOIG 3dd3iS 1Y¥3aS3qd-ans 3SIY SNIAIS JWOld

Ol NOILVidvay GNY EE ERC

WONTYNLINA "NI 39N3 QNIMNOAVS 3YNSS3Yd ddd3lS-143s3aq

SuaAT YaHLund 3AI19373S LNVLSNOD OLNI NOTLVIGWY

ATSAILV134 “SNOULS-

I "da A'W CT "da ATW BT (3719S OL LON)

3N3901S 131d 3N39011d SN350IW ane

130 ANNALS OF THE SOUTH AFRICAN MUSEUM

and subspecies in the genus Guttera, are also a result of divergence in these

refugia. The fact that Guttera subspecies are much more well marked than are

those of N. meleagris suggests that isolation in forest refugia has been more

effective than in tracts of savanna partitioned during wet phases.

ACKNOWLEDGEMENTS

I am grateful to Mr P. Clancey of the Durban Museum and Art Gallery,

Dr F. Gill of the Philadelphia Academy of Natural Sciences, Mr M. Irwin of

the National Museum of Rhodesia (Bulawayo), Dr A. Kemp of the Transvaal

Museum, Dr M. Louette of the Musee Royal de lAfrique Centrale,

Dr R. Paynter of the Museum of Comparative Zoology, Dr L. Short of the

American Museum of Natural History, Dr D. Snow of the British Museum

(Natural History), and Major M. Traylor of the Field Museum of Natural

History for allowing me access to the guinea-fowl in their institutions’ collec-

tions. Mrs J. Adams painted the guinea-fowl shown in Figure 1. Mr S. Piper

helped with contour maps. Dr L. Throckmorton provided initial encouragement

to begin this analysis. Prof. W. R. Siegfried provided the necessary stimulus

to finish the analysis. Mr R. Brooke commented critically on several drafts of

the paper. My wife, Anna, helped in all phases of the study, particularly with

illustrations and data processing. The De Beers and University of Cape Town

Computer Centres gave valuable advice on data processing and computer time.

Generous grants from De Beers Consolidated Mines Ltd and the University

of Cape Town were essential to the project. The inclusion of Figure | was

made possible by a grant from the University of Cape Town Editorial Board.

I should like to thank the South African Museum for publishing this paper.

REFERENCES

ARCHER, G. A. & GODMAN, E. 1937. The birds of Somaliland and the Gulf of Aden 2 London:

Gurney & Jackson.

AXELROD, D. I. & RAVEN, P. H. 1978. Late Cretaceous and Tertiary vegetation history of

Africa. In: WERGER, M. J. A., ed. Biogeography and Ecology of Southern Africa. Mono-

graphiae biol. 31. The Hague: Junk.

BANNERMAN, D. A. 1930. Birds of tropical west Africa 1 London: Crown Agents.

BEDDARD, F. E. 1898. The structure and classification of birds. London: Longmans, Green

& Co.

BENSON, C. W. & IRWIN, M. P. S. 1966. The brachystegia avifauna. Ostrich, Suppl. 6: 297-321.

BorTTICHER, H. V. 1954. Die Perlhiihner. Wittenberg—Lutherstadt: Ziemsen.

BONAPARTE, C. 1850. A new species of guinea fowl from West Africa. Proc. zool. Soc. Lond.

145.

BourkE, D. 1967. The Kazabo. Niger. Fld 32: 17-21.

Brown, J. L. 1963. Ecogeographic variation and introgression in avian visual signal: the

crest of the Steller’s Jay, Cyanocitta stelleri. Evolution 17: 23-39.

Butzer, K. W. 1967. Hypothetical rainfall and vegetation zones. In: CLARK, J. DESMOND, ed.

Atlas of African Prehistory. London: University of Chicago Press.

CaLcomp. 1971. GPCP: a general purpose contouring program. Anaheim: CALCOMP Com-

puter Products Inc.

CARCASSON, R. H. 1964. A preliminary survey of the zoogeography of African butterflies.

E. Afr. wildl. J. 2: 122-157.

THE EVOLUTION OF GUINEA-FOWL 131

CASSIN, J. 1857. Birds collected in West Africa by P. B. du Chaillu. Proc. Acad. nat. Sci.

Philad. 8: 322.

CuaPIN, J. P. 1923. The crested guineafowl of the southern Congo basin. Revue zool. afr.

11: 71-77.

CHAPIN, J. P. 1932a. The birds of the Belgian Congo. Part I. Bull. Am. Mus. nat. Hist. 65:

1-756.

CHAPIN, J. P. 19325. Fourteen new birds from Tropical Africa. Am. Mus. Novit. 570: 1-18.

Cooke, H. B. S. 1962. The environment in Southern Africa during the Pleistocene. Ann. Cape

Prov. Mus. 2: 11-15.

CRACRAFT, J. 1972. The relationships of the higher taxa of birds: Problems in phylogenetic

reasoning. Condor. 74: 379-392.

CRACRAFT, J. 1973. Continental drift, paleoclimatology and the evolution of birds. J. Zool.

Lond. 169: 455-545.

CrowE, T. M. & SNow, D. W. 1978. Numididae. In: SNow, D., ed. An atlas of speciation in

African non-passerine birds. London: British Museum (Natural History).

Davis, D. H. S. 1962. Distribution patterns of Southern African Muridae, with notes on some

of their fossil antecedents. Ann. Cape Prov. Mus. 2: 56-76.

Dixon, W. J. ed. 1975. Biomedical computer programs. University of California Publications

In Automatic Computation, No. 5. Berkeley: University of California Press.

DuBois, F. 1915. Remarques sur l’ornithologie de l’etat independent du Congo suivies d’une

loste des especres jusqui’n cet etat. Annls Mus. Congo., Zool., 1: 18, 27.

EHRLICH, P. & RAVEN, P. 1969. Differentiation of populations. Science 165: 1228-1232.

ELGoop, J. H., Fry, C. H. & DowsetTrT, R. J. 1973. African migrants in Nigeria. Jbis 115: 1-45.

ELuiot, D. 1870. A description of a new species of crested guinea fowl. Jbis: 300-301.

ELLioT, D. 1871. Description of a supposed new species of guinea-fowl. Proc. zool. Soc.

Lond.: 584.

ERLANGER, C. 1904. Neue Afrikanische arten. Orn. Mber. 12: 97-98.

Forp, J. 1974. Concepts of subspecies and hybrid zones, and their application in Australian

ornithology. Emu 74: 113-123.

FRADE, F. 1924. Notes d’ornithologie Africaine. Bull. Soc. port. Sci. nat. Lisbonne 9: 136.

Guici, A. 1905. Revisione del genera Guttera Wagler. Mem. R. Accad. Sci. Inst. Bologna

2: 189-198.

Guiai1, A. 1936. Galline di Faraone e Tacchini. Milano: V. Hoepli.

GOULD, S. J. & JOHNSTON, R. F. 1972. Geographic variation. A. Rev. Ecol. Syst. 3: 457-498.

Gray, G. R. 1840. Generic list for birds: 61. London.

Grote, H. 1936. Das perlhuhn von Noroostkamerun. Orn. Mber. 44: 156-158.

GuRNEY, J. 1868. Notes on Mr. Layard’s ‘Birds of South Africa’. bis: 253-271.

HaAGemier, E. M. & STULTS, C. D. 1964. A numerical analysis of the distributional patterns

of North American mammals. Syst. Zool. 13: 125-155.

HALL, B. P. 1961. The relationship of the guineafowls Agelastes meleagrides Bonaparte and

Phasidus niger Cassin. Bull. Br. Orn. Club 81: 132.

HALL, B. P. 1963. The Francolins, A study in speciation. Bull. Br. Mus. nat. Hist. 10: 105-204.

HALL, B. P. & MoreEAu, R. E. 1970. An atlas of speciation in African passerine birds. London:

British Museum (Natural History).

HAMILTON, I. 1974. The significance of patterns of distribution shown by forest plants and

animals in tropical Africa for the reconstruction of upper Pleistocene palaeoenviron-

ments: a review Palaeoecology 9: 63-97.

HARDWICKE, T. 1834. Description of a new species of the genus Numida. Proc. zool. Soc.

Lond.: 52.

HARTERT, E. 1899. In: ANSORGE, W. Under the African Sun. New York: Longmans, Green.

HARTERT, E. 1919. Numida Sabyi, a new species of helmeted guinea-fowl. Bull. Br. Orn.

Club 39: 69.

HARTLAvB, G. 1867. Veber eine neue Numida. J. Orn., Lpz. 15: 36-37.

HARTLAvuB, G. 1860. Drie neue afrikansche Vogel der Pariser Sammlung. J. Orn., Lpz. 8:

340-341.

HARTLAvuB, G. 1884. Ornithologie der Ostlich-aquatorialen Gebiete Africas. AbA. naturw.

Ver Bremen 8: 183-222.

Howe LL, F. C. & BOURLIERE, F. eds. 1963. African ecology and human evolution. Chicago:

Aldine.

132 ANNALS OF THE SOUTH AFRICAN MUSEUM

JACKSON, F. J. 1938. The birds of Kenya Colony and the Uganda Protectorate 1: London:

Gurney & Jackson.

KoTHE, K. 1911. Numida frommi; n. sp. Orn. Mber. 19: 13-14.

LINNE, C. 1758. Systema Naturae. 10th ed. Stockholm.

LINNE, C. 1766. Systema Naturae. 12th ed. Stockholm.

LIVINGSTONE, D. A. 1975. Late Quaternary climatic change in Africa. A. Rev. Ecol. Syst.

6: 249-280.

MACKWORTH-PRAED, C. W. & GRANT, C. H. B. 1952. Birds of eastern and north-eastern

Africa. London: Longmans, Green.

MACKWORTH-PRAED, C. W. & Grant, C. H. B. 1962. Birds of the southern third of Africa.

London: Longmans, Green.

MACKWORTH-PRAED, C. W. & GRANT, C. H. B. 1970. Birds of central and western Africa.

London: Longmans, Green.

Marx, H. & Rass, G. B. 1970. Character analysis: an empirical approach applied to advanced

snakes, 2 zone. Lond. 161: 525-548.

Mayr, E., LINSLEY, E. G. & USINGER, R. L. 1953. Methods and principles of systematic

Zooey. New York: McGraw-Hill Co.

MoreEAu, R. E. 1952. Africa since the Mesozoic: with particular reference to certain biological

problems. Proc. zool. Soc. Lond. 121: 869-913.

MorEAu, R. E. 1957. Variation in Western Zosteropidae (Aves). Bull. Br. Mus. nat. Hist.

7: 309-433.

MoreEAu, R. E. 1963. The distribution of tropical African birds as an madicater of post climatic

changes. In: HowELL, F. C. & BOURLIERE, F. eds. Africa Ecology and Human Evolution,

Viking Fund Publications in Anthropology 36: 28-42. Chicago: Aldine.

MorEAu, R. E. 1966. The birds faunas of Africa and its islands. New York: Academic Press.

NEUMANN, O. 1898. Die Helmperlhiiner. Orn Mber. 6: 17-22.

NEUMANN, O. 1899. Neue und wenig bekannte afrikanische Végel. Orn. Mber. 7: 23-26.

NEUMANN, O. 1904. Vogel von Schoa uns stid Aethiopien. J. Orn., Lpz. 53: 410.

NEUMANN, O. 1908. New subspecies of Guttera cristata. Bull. Br. Orn. Club 23: 13-14.

OGILVIE-GRANT, W. 1894. On a new species of guineafowl. Ibis: 535-538.

OLSON, S. L. 1974. Telecrex restudied: a small Eocene guineafowl. Wilson Bull. 86: 246-250.

OUSTALET, E. 1882. Description d’une nouvelle espece de pintade du Gabon. Annis Sci. nat.

Zool. 13: 1.

PALLAS, P. 1767. Spicilegia Zoologica 1: 18. Berolina: Gottl. Aen Lange.

PETERS, J. L. 1934. Check-list of the birds of the world 2. Cambridge: Harvard University Press.

PRIEST, C. D. 1933. The birds of Southern Rhodesia 1: London: William Clowes & Sons, Ltd.

REICHENOW, A. 1894. Das Helmperhithn von Damaraland. Orn. Mber. 2: 145.

REICHENOW, A. 1898a. Nachrichten. Orn. Mber. 6: 88.

REICHENOW, A. 18985. Guttera sclateri Rchw n. sp. Orn. Mber. 6: 115.

REICHENOW, A. 1900. Neue Forschungen in Deutsch Ostafrika. Orn. Mber. 8: 38-40.

REICHENOW, A. 1911. Neue arten aus Afrika Neuguinea. Orn. Mber. 19: 82-83.

REICHENOW, A. 1912. Neue arten ans dem Velle-Gebied in Mittel-Afrika. J. Orn., Lpz. 60:

320-321.

REYNOLDS, J. F. 1977. Thermo-regulatory problems of birds nesting in East Africa: a review.

Scopus 1: 57-68.

Roserts, A. 1917. Descriptions of a new species and genus of flycatchers from East Africa

and two new subspecies of guinea-fowls from South Africa. Ann. Transv. Mus. 6: 1-3.

Roserts, A. 1924. Classification of South African birds. Ann. Transv. Mus., 10: 77-86.

Roserts, T. R. 1975. Geographical distribution of African freshwater fishes. J. Linn. Soc.,

Zool, 57: 249-320.

SCHALOwW, H. 1884. Numida coronata marungensis subsp. nov. Z. Orn. prakt. Gefliigezl: 105.

SELANDER, R. K. 1971. Systematics and speciation in birds. In: FARNER, D. S. & KING, J. R.,

eds. Avian Biology 1: 57-147.

SipLey, C. G. & AHLouisT, J. E. 1972. A comparative study of the egg white proteins of

non-passerine birds. Bull. Peabody Mus. nat. Hist. 39.

SNEATH, P. H. A. & SOKAL, R. R. 1973. Numerical Taxonomy. San Francisco: W. H. Freeman

& Co.

SOKAL, R. R. & Rout, F. J. 1969. Biometry: the principles and practice of statistics in biological

research. San Francisco: W. H. Freeman & Co.

THE EVOLUTION OF GUINEA-FOWL 133

STONE, W. 1912. Vroeg’s catalogue. Auk. 29: 205-208.

Upvarpy, M. D. F. 1969. Dynamic zoogeography with special reference to land animals.

New York: Van Nostrand Reinhold Co.

VUILLEUMIER, B. S. 1971. Pleistocene changes in the fauna and flora of South America. Science

173: 771-780.

VUILLEUMIER, F. 1975. Zoogeography. In: FARNER, D. S. & KING, J. R. eds. Avian Biology

5: 421-496. New York: Academic Press.

WAGLER, J. 1832. Isis. London: Willoughby Society.

Wetmore, A. 1960. A classification for the birds of the world. Smithson. misc. Collns 139.

Wuite, C. 1943. Three new races from Northern Rhodesia. Bull. Br. Orn. Club 64: 19-22.

Waite, C. M. N. 1965. A revised check-list of African non-passerine birds. Lusaka: The

Government Press.

Witson, E. O., BRown, W. L., JR. 1953. The subspecies concept and its taxonomic appli-

cation. Syst. Zool. 2: 97-111.

APPENDIX |

Qualitative characters and character states analysed in this study.

Character name States

Crown adornment . : . 1 -—short crest of feathers

2 — long crest of feathers

3 — none, only naked skin

4 — bony helmet

Occipital adornment . . . 1 —short, dense, chestnut-coloured downy feathers

2 — long filoplumes confined to a mid-dorsal line

3 — fold of blue to black skin

4 — fold of whitish skin

5 — none, only naked skin

Nape adornment : , . 1 -—short downy feathers

2 — long filoplumes confined to a mid-dorsal line

3 — patch of orange-yellow skin at base, similar

patch anterior to ear

4 — none, only naked skin

Orbital skin colour. b . 1 — pink to red

2 — light blue to black

Throat skin colour. , . 1 — pink to red

2 — blue to black

Throat adornment. : . 1—no fold of skin

2 — fold of skin

Gape adornment : ; : 1 — rudimentary wattles

2 — well-developed and pointed wattles

3 — well-developed and rounded wattles

Gape wattle colour . d . 1-— pink to red

2 — blue to blue-grey

3 — blue with red tips

Cere adornment . : : eo none

2 — cartilaginous tufts or papilli

Collar plumage . : : . 1 — black with faint vermiculations

2 — black with no vermiculations

3 — black, finely barred with white

4 — grey to blue-grey

5 — spotted

6 — well-developed hackle

7 — white

Body plumage . : : . 1 — black with vermiculations

2 — spotted with vermiculations

3 — spotted without vermiculations

4 — as 3, with chestnut blotching between spots

134 ANNALS OF THE SOUTH AFRICAN MUSEUM

No. Character name States

12 Tarso-metatarsal scales . . 1 —imbricated and in a row

2 — pentagonal, not in rows

13 Tarso-metatarsal adornment . 1 W— spurs or bumps

2 — none

14 Iriscolour . : : : . 1 — brown

2 — red

15 Outer margins of secondaries . 1 -— brown to black with faint vermiculation

2 — white

3 — lavender

4 —alternating bands of black and white with

varying degrees of black and white vermicula-

tions

16 + Furcula A ee bs: . . 1 — blade-shaped

2 — hollow, cup-shaped; found in guineafowl with

long crests (Chapin 1932a)

17 Caecum length . 4 5 ‘ 1 — up to 150 mm

2 — greater than 200 mm; in all specimens con-

forming to the description of Acryllium

vulturinum (Beddard 1898)

18 Abdomen plumage . : . 1 — white

2 — blue

3 — white spots with faint vermiculations

4 — bluish-white spots without vermiculations

APPENDIX 2

Quantitative characters analysed in this study. See Figures 1-4 for a pictorial representation

of many of the characters.

No. Name Units Description

1 Billlength . : ; : mm the chord measured from the base of the

cere to the tip of the maxilla

2 Wing length ; ' : mm the chord of the unflattened folded wing

from the farthest anterior tip of the wrist

joint to the tip of the longest primary

3 Tarso-metatarsus length . mm the diagonal chord from the posterior point

of articulation of the tarsometatarsus with

the tibia to the most distal undivided tarsal

scute on the dorsal surface of the middle toe

4 Wattle basal width .. mm the chord from the most anterior to posterior

points of juncture of the wattle with the

cheek

ae Wattle lencth 995 ; mm the chord from the juncture line of the wattle

to the most distal point along the wattle

margin

6 Wattle percent blue... Wh a subjective estimate of the surface area of

the wattle covered by blue pigment. It is

possible to assess the amount of this colour

in preserved material since, although natural

colour is lost soon after death, the demar-

cation between red and blue can be deter-

mined because red areas revert to a yellow

or translucent amber state, and blue areas

to an opaque blue-grey

7 Helmet frontal length : mm the curvilinear distance, as measured with a

flexible tape, along the anterior margin of

the bony helmet from the point of juncture

with the skull to the apex

THE EVOLUTION OF GUINEA-FOWL

Name

Helmet rear length

Helmet central height

Helmet basal length

Helmet thickness

Cere structure length .

Cere structure thickness

Collar plumage

Nape filoplume length

Nape ffiloplume antero-

posterior coverage

Nape filoplume lateral

coverage

Nape filoplume density

Secondary remex outer web

vermiculation

Wing covert barring

Dorsal vermiculation .

Dorsal spot size .

Crest frontal length

Crest rear length

Units

0-4

135

Description

the curvilinear distance along the posterior

margin of the helmet from the point of

juncture with the skull to the apex

the chord perpendicular to the line of

juncture with the skull to the highest point

along the margin of the helmet

the chord from the anterior to posterior

juncture points of the helmet with the skul]

the maximum lateral width of the helmet at

its base

the chord from the base of the cere to the

distal tip of the longest tuft or papilla

the maximum thickness of the thickest tuft

or papilla

0 = grey; 1 = blue/violet; 2 = black, barred

white with a blue wash; 3 = black, barred

white; 4 = black, barred white with faint

longitudinal streaking; 5 = black, barred

white with some spotting; 6 = spotted

the chord from the base to the tip of the

longest straightened nape filoplume

the percentage of the nape, at the mid-dorsal

line, from the occiput to the upper-most

collar feathers, covered by filoplumes

the percentage of the nape covered laterally

by nape filoplumes at the mid-point of the

anteroposterior coverage

a subjective estimate, based on a comparison

with reference specimens encompassing the

range of variation, of density of the nape

filoplumes on a scale of increasing density.

0 = filoplumes not present, 4 = covered by

a mat of filoplumes

0 = absent; 1 = extends +

white bands; 2 = extends

white bands; 3 = extends

white bands; 4 = extends = way along the

white bands; 5 = extends to full length of

white bands; 6 = outer edge of white bands

obliterated by vermiculation

O= not present; 1 — present but faint:

Di hesene

a subjective estimate, as with character no.

18, O = faint, graded subjectively to 4 =

dense

the maximum width, as measured with a

dissecting microscope fitted with an ocular

grid, of a particular spot on a randomly

selected feather from the mid-dorsal region

the chord from the base to the tip of the

longest straightened crest feather within

5 mm of the cere

the chord from the base to the tip of the

longest straightened crest feather within

5 mm of the most posterior extent of the crest

way along the

Cal enleo expo

136

ANNALS OF THE SOUTH AFRICAN MUSEUM

Name

Crest central height

Crest basal length

Anterior crest curliness

Posterior crest curliness

Dorsal black collar

Ventral black collar

Occipital fold

Ear patch

Throat red

Orbital red

Dorsal spot number

Total spot barbs .

Total within spot barbs

Spot barb blueness

Chestnut blotch size

Chestnut blotch extent

Tarsal structure number

Tarsal structure length

White collar

Facial filoplumes

Units

count

0-1

Description

the chord perpendicular to the line of crest

juncture with the skull from the base of the

crest feathers to the highest point along the

margin of the unstraightened crest

the chord from the anterior to posterior

juncture points of the crest with the skull

a subjective estimate as with character no. 18.

1 = straight; 2 = moderately curly; 3 = very

curly

as with character no. 27

the percentage of black plumage extending

from the most anterior dorsal aspect of the

collar to the base of the tail

as with character no. 29 but the ventral

surface

a subjective estimate of the lateral extent of

the occiput covered by a fleshy fold

the maximum width of any patch of non-

blue/black skin anterior to the ear

a subjective estimate of the anteroposterior

extent of red pigmented skin between the

throat and the ventral base of the neck. As

with character no. 6, this character may be

assessed since the red colour reverts to a

yellow state after preservation.

a subjective assessment of the amount of

red pigmented skin around the eye

the maximum number of spots falling within

a circle of 1 cm radius superimposed over

the dorsal feather discussed in character

non22

the number of non-black barbs associated

with the spot measured in character no. 22

the number of barbs encompassed by the

spot measured in character no. 22

a subjective assessment of the amount of

blue in the spot. 0 = white; 1 = faint blue;

2 = medium blue; 3 = darkest blue

the maximum width of any chestnut blotch,

as measured with a dissecting microscope

fitted with an ocular grid, found on the

feather examined for character no. 22

the extent of the spotted area of the feather

examined for character no. 22 covered by

chestnut blotching

the number of spurs or bumps on the tarsus

the chord measured from the juncture line

of the longest tarsal bump or spur with the

tarso-metatarsus to the apex of the structure

the percentage of white plumage in the collar

a subjective estimate of the extent of the

head, other than the occiput, nape and

crown, covered by filoplumes

6. SYSTEMATIC papers must conform to the /nternational code of zoological nomenclature

(particularly Articles 22 and 51).

Names of new taxa, combinations, synonyms, etc., when used for the first time, must be

followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb.

nov., syn. nov., etc.