m Xlftos *x ^ XjyA»*»x m ^ XjyAsq^x rn

OSHlIlAIS^Sa I avda n~LI BRAR I ES^SMITHSONIAN INSTlTUTI0N^N0liniliSNI~NVIN0SHlll/\l$</)j
co z: „ co z * O'* z co

CO

o

2 2 ^ >'*

HSONIAN^INSTITUTION N0IXnilISNI_NVIKI0SHIIIAJSt/>S3 I bVb a ll^LI B RAR l ES"SMITHSONIAN~
~ CO 2 \ 00 ^ 5 00

Q OSJ^X _ CJ ,V«S* ««. Q N^yosv^' ~ C

2! 1 2? i *^y ^ j 2

3SH1I1MS S3lbVban LIBRARIES SMITHSONIAN INSTITUTION NOiXIUUSNI nvinoshxiws

I /^*&\ £ 1 / I « # 1

m x^y- pxx r; no ^ x^om^x rn

HSON IAN INSTITUTION * NOIinillSNl~*NVINOSHlilAIS "$3 I B Vd 8 I “S L I B RAR I E S ^SMITHSONIAN _
2 o*. co 2 ^ ^ Z .**V g 2

o^aTx 4§&On ^ — , v ~ , m\\ 5 -

2
O
GO

z t NgMWgi' 2

- Vo^r 7 ~ 5

osHiiws^sa 1 ava a nz lib rar i es^smithsonian^ institution NoiiniusNi nvinoshiiins4'

CO “ CO ~ co “ (/

W X^^TX 5 Cl

H

CO

O NiW;/ O X^/ x^xixyx O

HSOnIaN^ INSTITUTION 2 NOlinillSNl^NVINOSHlIWS^SB lavas ll^LI B RAR I ES^ SMITHSONIAN*”
i“ z: f~ 2: f“ z r

2 /^2!2&x a,

m X^vos^jX ^ X^VAStj^ rn s* \>^>- ^ X^vas^X" frj ^ ft

OSHIIWS^SB 1 H va a n~LI B RAR I ES $MITHSOMAN~INSTITUTION%OIiniLLSNI “NVIN0SH1IWS

to 2 n, t£ ^ ^ 2: v co 2 t/

< ^ 2 X»\ < V g <

*** * ^A. 3 g £ W%£kf. o / \ a

CO CO CO | :| (/

<wi

ftosv^x > £» ’ .^r > XJVA^X *g " ^ ^ x^osv^x J

HSONIAN^INSTITUTION WNOIXnXIXSNI KIVINOSHXIWS^SB I bVb B I I^Ll BRAR I ES^SMITHSONIAN^

to 5 \ <0 r; cn

- !2 s - 5

0£

, <

_ . _ m — K'^tAkW^/ “ * /

§ — j — z

OSHXIIMS S3lbVban LIBRARIES SMITHSONIAN INSTITUTION NOIXOlliSNI NVINOSHXIWS
z r- z <~ z r~

o ^SvSwTn „ 2 . m /iSSSSx 2 vfttSvJTx „ c

HSONIAN INSTITUTION NOIXOXIXSNI NVINOSHilliMS S3lbVban LIBRARIES SMITHSONIAN

y

<n

m

to ~ co f:

Ql“l LIBRARIES SMITHSONIAN INSTITUTION NOUfUllSNt NVIN0SH1IWS S3 I BVB 8 11 LI B

Z •_ CO ^ ^ 2 * CO 2 'n ^

< V 2 < a 2 <

2

•£1

> '<122^ S ' > 5 z ^

lON^NOliniliSKII NVINOSHilWS^SB I bVaan2LIBRARI ES^SMITHSONIAN INSTITUTION NOl

CO 5 \ 10 -jj W — </>

X UJ w Jk ^ « X*[53x “

05 n''£^5?©'/ o v2f^’ o xi»osi^ — o

3 2 J 2 _j 2 -J

an LIBRARIES SMITHSONIAN INSTITUTION NOlinillSNI NVINOSHilWS S3 I BVd 8 11 LIE

r- z <2 Z !T %r ■ 2

» ‘ >j6^'s /#% 1 ft s 1 mm s

m N^yoc^ ^ 5 rm

CO ~ if) _ £0 \ ~ to

riON NOIlfliliSNI NVINOSHilWS S3 I H VH 8 11 LIBRARIES SMITHSONIAN INSTITUTION NO

CO 2 <£ 2 ~ ^ Z CO

2 .< g ,..< E 5 As s

2 Z^lf^KA HI /'y ' Z $s^V\> “H /^S£w^\ 2 sSRfcs-s ~

X -J

s £

CO ISp 351 CO in "• co -4*1 to

2 tl

- s > '^P^' 2 > XOia^x 2 >

a n2 LI B RAR I ES^SMITHSONIAN INSTITUTION NOliOlllSNI NVINOSHXIWS^S 3 I U V ti 8 11 LU
i “! Z 5 ^ ^ i

O X^N pC^X ~ O XL* _ XOiusgX Q

NON 2 NOIlfliliSNI ^NVINOSHilWS ^"S 3 IBVHBIT^LIBRARI ES^ SMITHSONIAN^* INSTITUTION NOl
2 r- » 2 r~ 2 “

m * w r; x^vas^x m x^qsb^x ^ m

co ‘ ± co £ co

811 LIBRARIES SMITHSONIAN INSTITUTION NOIXOUISNI NVINOSHXIINS SBlBVBail LIE

2 ^ 2 « CO 2 ^ .w 5

< ^ 2 /»\ < A 2 -A < 2 <

rk o

X to

2
O
O)

t: - « _

2 *' > * 2 X^osv^x > 2

co (/> J>v’ 2 to 2 CO

riON NOliniliSNI NVINOSHilWS S3I avuan LIBRARIES SMITHSONIAN INSTITUTION NOl

V) 5 \ 2 U1 jg w

/- H

o

2 «J

lan LIBRARIES SMITHSONIAN INSTITUTION NOlifUIXSNI NVINOSHilWS S3IBVBan LIE
r- 2 r* ^ 2 r* 2 ___ r-

^ w I rn I m

TION00 NOlifliliSNI- NVINOSHilWS SBIHVaBIT- LIBRARI ES^SMITHSONIAN- INSTITUTION NO
to 2 __ to 2 to z

ANNALS

OF

CARNEGIE MUSEUM

VOLUME 56

1987

PUBLISHED BY THE AUTHORITY OF THE
BOARD OF TRUSTEES OF THE CARNEGIE INSTITUTE
PITTSBURGH, PENNSYLVANIA
1987

DAVID R. WATTERS, Acting Director (1/1/87-31/8/87)
JAMES E. KING, Director (1/9/87-31/12/87)

Editorial Staff
Leonard Krishtalka, Editor
John L. Carter, Editor
Mary Ann Schmidt, Editorial Assistant

Publications Committee
C. J. McCoy, Chairman
Anthony D. Bamosky
James B. Richardson III
F. H. Utech

CONTENTS

Contents iii

New Taxa v

Author Index vi

ARTICLE

h The Middle Eocene Arshanto Fauna (Mammalia) of Inner Mongolia. Tao Qi ...... . 1

2. Ecology of small mammals in a gallery forest of central Brazil. Leslie Zuhn Nitikman

and Michael A. Mares 75

3. Reproduction in a Spanish population of Acanthodactylus cry thrums (Reptilia: Lacertilia:

Lacertidae). Stephen D. Busack and Lorrie L. Klosterman ...................... 97

4. Karyotypic analysis of five rodents and a marsupial from Belize, Central Ameri-

ca, David W. Burton, John H. Bickham, Hugh H. Genoways, and Timothy J ,

McCarthy 103

5. Biosystematic studies in Stenanthium (Liliaceae-Veratreae). I. Floral morphology, floral

vascular anatomy, geography and taxonomy of S. occidentale A. Gray. Frederick H.

Utech 113

6. Results of The Carnegie Museum of Natural History Expeditions to Belize. IT Distri-

butional notes on the birds of Belize. D. Scott Wood and Robert C. Leberman. ... 137

7. A review of the crane flies in the Subgenus Tipula ( Papuatipula ) (Diptera: Tipulidae),

with descriptions of five new species. Chen W. Young .......................... 161

8. Taxonomic and geographic variation of Liophis typhlus and related “green” species of

South America (Serpentes: Colubridae). James R. Dixon ....................... 173

9. A new species of Ciinidium Kirby (Coleoptera: Carabidae or Rhysodidae) from Mexico,

and descriptions of the females of two neotropical members of the genus. Ross T.

Bell and Joyce R. Bell 193

10. Biosystematic studies in Stenanthium (Liliaceae-Veratreae). II. Floral morphology, floral

vascular anatomy, geography and taxonomy of the Mexican S. frigidum (Schlecht. &
Cham.) Kunth. Frederick H. Utech 197

11. Systematics of African bats of the Genus Eptesicus (Mammalia: Vespertilionidae). 2.

Karyotypes of African species and their generic relationships. Karen McBee, Duane
A. Schlitter and R. Laurie Robbins . 213

12. Revision of the Wind River faunas, early Eocene of central Wyoming. Part 8. First fossil

lizard egg (?Gekkonidae) and list of associated lizards. Karl F. Hirsch, Leonard Krish-
talka and Richard K. Stucky 223

13. Taxonomic notes on some African warblers (Aves: Sylviinae). Kenneth C. Parkes ... 231

14. Spergenaspis : a new Carboniferous trilobite genus from North America. David K.

Brezinski 245

15. Decompression syndrome in fossil marine turtles. Bruce M. Rothschild ............ 253

1 6. Towards a postglacial history of the northern Great Plains: A review of the paleoecologic

problems. Cathy W. Barnosky, Eric C. Grimm and H. E. Wright, Jr. ............ . 259

17. Fossil crab (Decapoda: Brachyura) fauna from the Late Cretaceous (Campanian - Maas-

tricht :i ant Pierre Shale in Bowman County, North Dakota. Annette B. Tucker, Rod-
ney M. Feldmaen, F. D. Holland, Jr. and Kenneth F. Brinster .................... 275

iii

18. Excavations at the Harney Site slave cemetery Montserrat, West Indies. David R.

Watters 289

19. Description of skeletal remains from a Black slave cemetery from Montserrat, West

Indies. Robert W. Mann, Lee Meadows, William M. Bass and David R. Watters. . 319

iv

NEW TAXA

DESCRIBED IN VOLUME 56

NEW GENERA, SPECIES, AND SUBSPECIES

fSinosinopa, new genus, Mammalia, Insectivora 15

fSinosinopa sinensis, new species, Mammalia, Insectivora . 15

fAsiomys, new genus, Mammalia, Rodentia 17

tAsiomys dawsoni, new species, Mammalia, Rodentia 18

fMetacoryphodon, new genus. Mammalia, Pantodonta 23

tMetacoryphodon luminis, new species, Mammalia, Pantodonta 23

tMetacoryphodon? minor, new species, Mammalia, Pantodonta 26

fPantolambdodon? minor, new species, Mammalia, Pantodonta 28

fGobiatherium? major, new species, Mammalia, Dinocerata 31

fGobiatherium? monolobotum, new species, Mammalia, Dinocerata 32

tHomogalax reliquius, new species, Mammalia, Perissodactyla 33

tHeptodon minimus, new species, Mammalia, Perissodactyla 34

fHelaletes medius, new species, Mammalia, Perissodactyla 37

fHyrachyus neimongoliensis, new species, Mammalia, Perissodactyla 39

fHyrachyus crista, new species, Mammalia, Perissodactyla 42

tTeleolophus primarius, new species, Mammalia, Perissodactyla 54

fTeleolophus? rectus, new species, Mammalia, Perissodactyla . 54

f Microtitan? elongatus, new species, Mammalia, Perissodactyla 59

fForstercooperia huhebulakensis, new species, Mammalia, Perissodactyla 61

Tipula (Papuatipula) insperata, new species, Insecta, Diptera 164

Tipula (Papuatipula) koiari, new species, Insecta, Diptera 164

Tipula (Papuatipula) nigritus, new species, Insecta, Diptera 167

Tipula (Papuatipula) oneili, new species, Insecta, Diptera 168

Tipula (Papuatipula) wibleae, new species, Insecta, Diptera 170

Clinidium (Mexiclinidium) reyesi, new species, Insecta, Coleoptera 193

Cisticola chinlana keithi, new subspecies, Aves 235

Cisticola chiniana mbeya, new subspecies, Aves 238

Sylvietta leucophrys arileuca, new subspecies, Aves 242

fSpergenaspis, new genus, Arthropoda, Trilobita 247

fSpergenaspis salemi, new species, Arthropoda, Trilobita 248

fSpergenaspis easleyi, new species, Arthropoda, Trilobita 250

f Fossil taxa.

v

AUTHOR INDEX

Bamosky, C W. ...................... 259

Bass, W. M 319

Bell, J. R 193

Bell, R. T 193

Bickham, J. H 103

Brezinski, D. K. ..... 245

Brinster, K. F 275

Burton, D. W 103

Busack, S. D 97

Dixon, J. R. 173

Feldmann, R. M. ...... 275

Genoways, H. H 103

Grimm, E. C 259

Hirsch, K. F. . . . 223

Holland, F. D„ Jr 275

Krishtalka, L 223

Leberman, R. C 137

Mann, R. W. . . . 319

Mares, M. A. 75

McBee, K 213

McCarthy, T. J. 103

Meadows, Lee 319

Nitikman, L. Z. 75

Parkes, K. C 231

Qi, T 1

Robbins, R. L 213

Rothschild, B. M 253

Schlitter, D. A. 213

Stucky, R. K. 223

Tucker, A. B. . . . 275

Utech, F. H 113, 197

Watters, D. R. 289,319

Wood, D. S 137

Wright, H. E., Jr 259

Young, C. W 161

vi

issn 0097-4463

Jfa

ANNALS

of CARNEGIE MUSEUM

THE CARNEGIE MUSEUM OF NATURAL HISTORY

4400 FORBES AVENUE ® PITTSBURGH, PENNSYLVANIA 15213

VOLUME 56

6 MARCH 1987

ARTICLES 1-2

CONTENTS

Art. 1. The Middle Eocene Arshanto Fauna (Mammalia) of Inner

Mongolia Tao Qi 1

Art. 2. Ecology of small mammals in a gallery forest of Central Brazil

Leslie Zuhn Nitikman and Michael A. Mares 75

Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS:

L. Krishtalka
J. L. Carter

Office of Scientific Publications
Phone (412) 622-3287

Communications about subscriptions, purchases of individual numbers, rates,
and changes of address should be sent to:

Publications Secretary

The Carnegie Museum of Natural History

4400 Forbes Avenue

Pittsburgh, PA 15213

ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural
History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213, by the authority of the Board of
Trustees of Carnegie Institute.

THE CARNEGIE

MUSEUM OF
NATURAL HISTORY

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 1, Pp. 1-73 6 March 1987

THE MIDDLE EOCENE ARSHANTO FAUNA (MAMMALIA)
OF INNER MONGOLIA

Tao Qi1

Visiting Museum Specialist
Section of Vertebrate Fossils

Abstract

Review of the Eocene deposits in Inner Mongolia has shown that the classic “Arshanto Formation”
forms the lower beds of the Irdin Manha Formation. Mammalian faunas from the Arshanto beds as
well as from the overlying Irdin Manha beds are middle Eocene in age, roughly equivalent to the early
and late Bridgerian, respectively. Two other units, formerly included in the “Arshanto Formation”
are the Lower Eocene Bayan Ulan beds and the Upper Paleocene Nomogen beds, both here referred
to the Nomogen Formation. Comparisons between the Inner Mongolian Arshanto fauna and the
approximately time equivalent faunas of eastern and southern Asia suggest the existence of two
Asiatic paleobiogeographic districts during the middle Eocene.

Introduction

From 1975 to 198 1 a team of specialists on the early Tertiary from the Institute
of Vertebrate Paleontology and Paleoanthropology, Beijing, explored the Paleo-
gene sediments in the Shara Murun region, Inner Mongolia, where the Central
Asiatic Expedition of the American Museum of Natural History had made a series
of investigations between 1922 and 1930. The team found approximately one
hundred species of fossil mammals. Of these, a few were found by geologists
mapping Inner Mongolia whereas others were recovered by the staffs of the Inner
Mongolian Museum and the Institute of Vertebrate Paleontology and Paleoan-
thropology. Most of the discussions in this paper are based on these newly dis-
covered fossils, although some previously collected specimens from the Arshanto
fauna also are included.

The earliest report on an Asiatic middle Eocene fossil mammal ( Schlosseria
magister) was made by Matthew and Granger in 1926, but discoveries of fossil
mammals of this age since that time have been rare. From the late 1950s to the
present, new discoveries have been made in south Asia (Pakistan and India),
central Asia (Kazakhstan), and east Asia (China) which have enhanced, consid-
erably, our knowledge of middle Eocene mammalian faunas in Asia. The Arshanto
fauna of China has a prominent position among these faunas, and the Arshantan
Age has been designated as the Asiatic middle Eocene (Romer, 1966).

Study of the Arshanto has led us to the following conclusions:

1. The classic “Arshanto Formation” can be divided into three beds: the Upper
Paleocene Nomogen beds, the Lower Eocene Bayan Ulan beds, and the Middle
Eocene Arshanto beds. In 1975, the Nomogen mammalian fauna was discovered
in Haliut and Gonghutong (Zhou and Qi, 1978) in strata at the lowest level of
the classic “Arshanto Formation,” but the boundary between the “Arshanto beds”
and the underlying “Nomogen beds” was not clear. In 1976, another mammalian

1 Address: Institute of Vertebrate Paleontology and Paleoanthropology, Academia Sinica, P.O. Box
643, Beijing, People’s Republic of China.

Submitted 3 November 1986.

1

2

Annals of Carnegie Museum

vol. 56

fauna (under study by Zhai), which is quite similar to but shows some differences
from the Nomogen fauna, was found in Bayan Ulan above the level of the Nom-
ogen beds. Fossil mammals also were found in the Arshanto beds in several other
areas such as Arshanto Obo, Irdin Manha, Bayan Ulan, Huhe Bulak, and Ulan
Bulak in 1976 and 1977. These are the main materials described in this paper,
and these faunas comprise the Arshanto fauna.

2. The composition of the Arshanto fauna bears some resemblance to that of
the Irdin Manha fauna. The ages of both are close, i.e., middle Eocene. Therefore,
three middle Eocene mammalian faunas are recognized in the eastern part of
China: the earliest, Xintai (Sintai) fauna (Zhou and Qi, 1982), the Arshanto fauna,
and the latest, Irdin Manha fauna. The Arshanto and Irdin Manha faunas are
quite different from the Inner Mongolian Shara Murun fauna in composition. The
age of the latter is unquestionably late Eocene, on the basis of such mammals as
the hyaenodontid Pterodon, the perissodactyl Deperetella, and the artiodactyl
Archaeomeryx.

3. The age of the Arshanto fauna seems to compare most closely with that of
the early middle Eocene Bridger fauna (Bridger A and B) of North America,
especially on the basis of the helaletids and the brontotheriids. It must be noted,
however, that some elements of the Arshanto fauna are more typically associated
with older faunas (Harpagolestes, Homogalax ), whereas at least Forstercooperia
is more typically younger.

4. Based on the known fossils, the east and central Asiatic middle Eocene faunas
belong to the same paleogeographic district, and both are quite different from
those of the south Asiatic district. The southern fauna has some marine mammals,
such as Cetacea, and some terrestrial animals, such as Proboscidea and Artio-
dactyla, which never or very rarely appear in the Arshanto fauna and in central
Asia. Some relatives (Anthracotheriidae) of middle Eocene mammals found in
south Asia were discovered in late Eocene beds in south China and Burma, but
are rare in northern Asia. Therefore, southern and southeast Asia appear to belong
to another paleogeographic district.

Abbreviations used in this paper are: IVPP, Institute of Vertebrate Paleontology
and Paleoanthropology; AMNH, American Museum of Natural History; L, length;
W, width; ant., anterior; post., posterior.

Older Chinese spelling (Wade-giles) is used in the bibliographic references through
1972, following the Bibliographies of Vertebrate Paleontology.

Geology

Field Work (Fig. 1)

Berkey and Morris (1924, p. 119) first mentioned the Arshanto beds (Forma-
tion) as follows: “At Irdin Manha, twenty miles southeast of Iren Dabasu, the
Houldjin is not found, and the section exposed consists of 40 to 100 feet of grey
sands, with a rich titanothere fauna, which may be late Middle Eocene or even
Upper Eocene. Beneath the titanothere beds there are red clays, provisionally
called Arshanto, and probably to be correlated with the barren beds above the
Iren Dabasu. The Arshanto may prove to be only the lower Irdin Manha, or it
may be separated from the Irdin Manha by a disconformity. The base of these
beds has not been seen.”

In their study of Schlosseria magister and Teilhardia pretiosa, Matthew and
Granger (1926, p. 1) said, “At the base of the Irdin Manha and Shara Murun
formations are red clays generally barren, to which Berkey and Morris have given

1987 Qi— Eocene Mongolian Arshanto Mammals 3

Fig. 1.— Map of Inner Mongolian Paleogene localities (“Naumuhun” = Nomogen in text).

the provisional name of Arshanto formation.The only fossils found are two re-
markable little perissodactyls, apparently related to the lophiodontoid genera of
the Irdin Manha and Ardyn Obo, but much more primitive. As far as the evidence
goes the Arshanto may be Middle Eocene, but there is no correlation apparent
with any Middle Eocene mammals of Europe or North America.”

Later, Berkey and Morris (1927, p. 207) described the Arshanto beds in detail:
“The Arshanto beds are prevailingly red clays and fine silts, which, on weathering,
crumble into small hard chips. Much of the deposit is structureless, almost like
loess. In some places a faint color-banding is seen in the red beds, but no shaly
structure or marked bedding. Color-banding is held by Matthew (1915, pages
395-398) to be not inconsistent with wind deposits, and our observations tend
to support his opinion. In still other places, chocolate brown beds and, more
rarely, thin grey layers appear among the red, and here at least there is definite
bedding. About a mile east of Irdin Manha escarpment lies another broad, un-

4

Annals of Carnegie Museum

vol. 56

drained hollow (Fig. 99, page 198), part of which is called Arshanto, from which
the formation is named. Here some lense-shaped beds of grey and red sandstone
were found in the red clays. A limited collection of small lophiodonts was made,
which has not yet been fully studied; but the preliminary examination by Matthew
and Granger shows that they are quite different from the lophiodonts of the Irdin
Manha beds. The base of the red beds was not seen, the only bottom known for
the formation is at Iren Dabasu, where it rests upon the Cretaceous beds.”

Radinsky (1964, p. 3) considered the age of the Arshanto beds early late Eocene
because, “ Schlosseria is similar enough to Lophialetes to indicate only a very
slightly older age, possible early Late Eocene, for the Arshanto beds.” Later,
Radinsky (1965, p. 201) pointed out again, “ Schlosseria magister is probably not
much older than Lophialetes expeditus and on morphological grounds could equal-
ly well be late Middle Eocene or early Late Eocene in age.” Romer (1966) took
the “Arshanto Age” to designate the Asiatic middle Eocene age.

In 1975, after representative late Paleocene mammals were found by an Inner
Mongolian geological mapping team, Zhou and Qi (IVPP) studied the localities
and suggested on the basis of field evidence that these late Paleocene mammal-
bearing beds were in the Arshanto Formation (Zhou et al., 1976). In August, 1 976,
paleontologists of the IVPP found fossil mammals, later named the “Bayan Ulan
fauna,” at Bayan Ulan, which is located on the north slope of the Nomogen Mesa.
Above these Bayan Ulan beds are the Arshanto beds, which in turn are overlain
by the Irdin Manha beds. The Arshanto beds yielded some middle Eocene mam-
mals including Gobiatherium mirificum and Hyrachyus cristata new species.

In June 1 977, the IVPP team discovered the incisors of Gobiatherium mirificum,
the premolars of insectivores, and some teeth of lophialetids (field nos. 77023,
77026) at Arshanto Obo. Approximately thirty more taxa of fossil mammals were
found in the Arshanto beds at Huhe Bulak and Bayan Ulan, and ten or more
kinds of early Eocene fossil mammals were recovered at Bayan Ulan. Two small
collections were made in these areas by the IVPP in 1978 and 1980.

Irdin Manha Formation and Arshanto Beds

Granger and Berkey (1922) proposed using “Irdin Manha Formation” for a set
of grey-white sandstone beds (bearing numerous titanotheres) 25 miles south of
Iren Dabasu salt lake. Thickness of the beds ranges between 40 and 100 feet. The
age of the beds was considered to be Eocene-Oligocene. Later, Berkey and Morris
(1924) pointed out that the age of the Irdin Manha Formation may be late middle
Eocene or late Eocene. At this time they mentioned a red clay below the “titano-
there beds” and, as mentioned above, gave the provisional name, Arshanto, to
the red bed.

Since 1926, when Matthew and Granger described Schlosseria magister, the
“Arshanto fauna” was little studied until field work, mostly between 1975 and
1978, again produced many fossil mammals. These discoveries, as well as others
in south Asia, central Asia, and other parts of China (Shandong and Henan
provinces and Xinjiang Uighur Autonomous Region) have contributed a great
deal to understanding the middle Eocene Asiatic mammalian faunas, and the
Arshantan Age as an important time in geological and mammalian history.

The “Arshanto Formation” of Berkey and Morris (1927, p. 206) can be divided
into two parts: the upper part, named “Arshanto beds,” is here included as the
lower part of the Irdin Manha Formation; the lower part of Berkey and Morris’

1987

Qi— Eocene Mongolian Arshanto Mammals

5

“Arshanto Formation” is the Nomogen Formation and contains two levels: the
upper Lower Eocene Bayan Ulan beds and the lower Upper Paleocene Nomogen
beds (or celestite-nodule bearing beds).

The thickness of the Irdin Manha beds is variable: at Huhe Bulak, only 5.4 m;
at Bayan Ulan, 41 m; at Arshanto Obo, 10 m; and at Irdin Manha and Ulan
Shireh, 12-30 m and 43 m respectively. It changes rapidly over a short distance
and indicates topography at the time of deposition. In spite of the existence of
an erosional surface between the Irdin Manha beds and the Arshanto beds in
some places, the general characters of the two beds are quite similar. Therefore,
the two beds are regarded as composing the Irdin Manha Formation. The Arshanto
beds are found mainly in Sunid Yuqi (county). Their northeast extent is located
near Arshanto Obo and their southwest boundary occurs at Nomogen Mesa.

Below the Irdin Manha beds in Aliusu, Urtyn Obo area, there is a set of dark
red clay beds. These were previously considered Arshanto beds, but are perhaps
best referred to the Nomogen Formation. They bear a pantodont, Pastoralodon
lacustris, which is otherwise known only in the upper part of the Nomogen For-
mation in the Bayan Ulan fauna.

The following are abbreviated geological sections (numbers in parentheses are
IVPP field numbers).

I. Arshanto Obo section (upper to lower Fig. 2)

Irdin Manha Formation

Middle Eocene Irdin Manha beds

4) upper part: grey-green, grey-yellow sandstone to the top having a
light red medium sandstone and tobacco-yellow medium quartz
sandstone with coarse quartz gravels (diameter generally less than
1 cm, the maximum 3 cm); grains rounded, but not well sorted
and not well bound.

lower part: grey-green and grey-yellow medium and fine sandstone,
well bound together by siliceous cement; fine silt-sandstone can
often be seen in the transverse direction; locally having numerous
muddy clumps, Forstercooperia confluens (77025). 10.13 m

— disconformity (base of channel) — — -

Middle Eocene Arshanto beds

3) dark red clay with few muddy clumps and fine silica grains, not
well bound. 2.05 m

2) upper part: being covered by Quaternary plants. 13.94 m

middle part: grey-green sandstone, mainly silica grains, bound by
siliceous cement.

lower part: variegated sandstone (mainly grey color; secondary
dark red) with numerous muddy clumps at base bearing Gobiathe-
rium mirificum, Schlosseria magister, etc. (77021). 16.88 m

1) dark red sandy clay with few muddy clumps. 1.68 m

— —bottom not visible—

II. Huhe Bulak section (Fig. 3)

Irdin Manha Formation

Middle Eocene Irdin Manha beds

1 ) grey-white on surface, but grey-yellow inside, sandstone and sandy
gravels with many cylinder-like calcareous masses; having grey
sandy conglomerate and sandstone; gravels often black quartz, the
diameter 1 cm or so, with Andrewsarchus (= Paratriisodorv,
McKenna, personal communication) gigas, etc. (77037). 5.4 m

unconformity (=base of channel}

220

6

Annals of Carnegie Museum

vol. 56

1987

Qi- Eocene Mongolian Arshanto Mammals

7

Fig. 3.— Huhe Bulak section (in 4 segments, top to bottom; complete horizontal section connects at
a, b, c). See legend in Figure 2.

Middle Eocene Arshanto beds
1 0) top: dark grey-green-pink sandy clay,
upper part: yellow-green fine sandstone.

bottom: grey-green sandy “conglomerates” (“gravels” mainly being
muddy masses and quartz, etc., diameter 0.2 cm), with Gobia -

8

Annals of Carnegie Museum

vol. 56

therium mirificum, Mesonyx obtusidens, Schlosseria magister, etc.
(77036-2). 6.0

9) upper part: dark red clay with fossil mammal fragments and cal-
careous-nodules, diameter 1.5 cm.

lower part: brown clay with few manganese nodules. 4.04

8) red-grey muddy siltstone (white on surface), with few mammalian
incisors and fragments (77036-H8). 4.85

7) brown-loess-red clay, with manganese nodules, the diameter 0.5
cm; bearing few limb bones of Dinocerata (77036-H7). 5.1

6) light loess-red siltstone. 0.3

5) brown muddy siltstone (white on surface) with few calcareous
nodules, bearing Sinosinopa sinensis (77036-H5). 2.0

4) brown-grey sandy conglomerates (grey on surface), gravels mainly
a lot of muddy masses and colorless, black, or yellow-green quartz
grains, with many fossil tapirs, fragments of turtles and crocodiles.
Fining upward. 3.04

3) dark red or grey-green variegated clay, with few quartz grains and
calcareous nodules, bearing astragali of Dinocerata (77036-H3).
Grey-green clay decreased and manganese nodules apparent up
section. 6.5

disconformity

Nomogen Formation

Lower Eocene Bayan Ulan beds

2) brick-red sandy clay with few nodules and many bone fragments,
bearing fossil tapirs, Dinocerata (Prodinoceras sp. =Mongolo-
theriunv, McKenna, personal communication) (77035). 2.76

1) loess-red and grey-green clay with many calcareous nodules be-
tween 5-10 cm diameter. 3.0

base not visible

III. Bayan Ulan sections (after material of the Inner Mongolian mapping team;

Fig. 4)

Ulan Gochu Formation

Oligocene

1 7) loose coarse sandstone with gravels with fragments of cf. Indri-

cotherium sp. (1P25H19). 6.6

16) variegated clay. 2.6

15) grey-white medium feldspar-quartz bearing sandstone with Tel -
eolophus magnus. 0.5

14) brown-red clay. 1.8

13) grey-white silt-sandstone. 3.3

12) grey- white medium feldspar-quartz bearing sandstone. 4.8

11) light brown-red clay. 1.6

10) grey-white coarse sandstone with gravels. 0.3

— disconformity —

Irdin Manha Formation

Middle Eocene Irdin Manha beds

9) brown-red clay with Gobiohyus sp., IHarpagolestes sp. and Mes-
onychidae indet. (lP^H,,). 27.5

8) loose manganese sandy conglomerates and loose medium-coarse
sandstone. Upper part: grey-white, light yellow medium feldspar-
quartz bearing sandstone with Teleolophus sp., Microtitan mon-
goliensis, and Lophialetes sp. 6.9

7) variegated silty clay, with Microtitan mongoliensis, Lophialetes
sp., Eudinoceras sp., and Rhodopagus sp. (1P25H7). 6.7

— disconformity-

Middle Eocene Arshanto beds

6) brown-red clay with grey-green clay, with Rhodopagus sp., and
Lophialetes sp. 1 0

m

m

m

m

m

m

m

m

m

m

m

m

m

m

m

m

B B B B B B

Mesa

1987

Qi— Eocene Mongolian Arshanto Mammals

9

10

Annals of Carnegie Museum

vol. 56

5) grey-green silty clay. 2.9 m

4) brown-red clay with grey-green siltstone and grey-green clay, bear-
ing IMongolonyx prominentis, Hymchyus cristata, Breviodon mi -
nutus, Pantolambdodon sp., and Lophialetes sp. 25 m

3) gravel-bearing brown-red clay with green clay and siltstone, the
bottom is brown sandy conglomerates, bearing Gobiatherium
monolobatum (1P25H3). 4 m

disconformity

Nomogen Formation

Lower Eocene Bayan Ulan beds

2) variegated clays with Prionessus lucifer, Mimotona borealis, Pseu-
dictops lophiodon, Pachyaena sp., Plagiocristodon serratus (milk
teeth of Dissacus, McKenna, personal communication), Pasto-
ralodon lacustris, Prodinoceras (=Mongolotherium\ McKenna,
personal communication) efremovi, Prodinoceras ( =Mongolothe -
rium) sp., Pyrodon sp., Palaeostylops iturus, Palaeostylops macro-
don, Mongolotherium sp., and ILambdotherium sp. 6.5 m

1) brown-red sandy-silty sandstone, grey-green silty clay appears in
the top. Bottom: grey-white silty clay with gravels and red muddy
sandstone with gravels, bearing Pastoralodon lacustris, Prionessus
lucifer, Sarcodon pygmyaeus, Palaeostylops iturus, and Panto-
lambdodontidae (new genus and species), etc.

bottom not visible

Fossil Mammals from above and below the Arshanto Beds

/. The Overlying Irdin Manha Beds

The Irdin Manha fauna has been collected from at least ten sites, of
which four are major: the Irdin Manha escarpment, Ulan Shireh, Al-
iusu, and Camp Margetts.

1 . Irdin Manha escarpment. This site is located 40 km south of Iren
Dabasu salt lake on the southwest edge of Suji Mesa. This is the type
area of the Irdin Manha Formation and has been called by various
authors “Irdin Manha Formation,” “Irdin Manha beds,” “Irdin Manha
area,” and “near Irdin Manha.” The thickness of the type Irdin Manha
beds is 41 m and the fossil quarry (a titanothere site excavated by
Russian paleontologists) is not far above the Arshanto beds. Appendix
1 lists the fauna from this site.

2. Ulan Shireh. The fossil quarry is located 10 km northeast of
Tukhum Lamasary (Siziwangqi). No fossil mammals were discovered
in other beds at this site.

3. Aliusu. This fossil quarry was found in 1975 by the staff of the
Inner Mongolian mapping team and the author. It is located in the
Urtyn Obo area. These beds appear to be correlatives of the type Irdin
Manha beds, and preserve the following fossil mammals:

Rodentia

Advenimus sp.

Creodonta

Propterodon sp.

Pantodonta

Eudinoceras mongoliensis

Pantolambdodon inermis

Acreodi

IHarpagolestes orientalis

Perissodactyla

1987

Qi— Eocene Mongolian Arshanto Mammals

11

Protitan grangeri
Microtitan mongoliensis
Microtitan sp.

Lophialetes expeditus
Breviodon minutus
Breviodon sp.

Teleolophus medius
Teleolophus sp.

Forstercooperia confluens
Artiodactyla
Gobiohyus orientalis
Dichobunid sp.

4. Camp Margetts. Camp Marge tts was located somewhere in the Ulan Bulak-
Huhe Bulak escarpment, some kilometers to the west of the Irdin Manha escarp-
ment. The AMNH Central Asiatic Expedition collected a few fossil mammals
from two beds called “Houldjin beds” and the underlying “Irdin Manha beds”
in this area. The discovery in 1978 of Lophialetes expeditus in the “Houldjin
gravels” here proved that they represent a lateral facies of the type Irdin Manha
beds, and should be considered part of the Irdin Manha Formation. The underlying
“Irdin Manha beds” at Camp Margetts are probably equivalent to the Arshanto
beds.

On the other hand, some AMNH fossils are labeled as coming from “7 miles
(about 10 km) southwest of Camp Margetts,” or “10 miles (17 km) west of Camp
Margetts,” or “5 miles (8 km) west of Camp Margetts.” Among these sites, the
“west” area may be the Bayan Ulan area. These sites cannot be correlated directly
with the strata in the Camp Margetts area. However, we may guess the provenance
of some fossil mammals by referring to those preserved in the type Irdin Manha
and type Arshanto beds. Cf. Hyrachyus and Helaletes sp., for example, probably
came from the Arshanto level, and Forstercooperia confluens, Mongolonyx doli-
chognathus, and Litolophus ( Grangeria ) gobiensis from Irdin Manha levels.

The fossil mammals from known localities in the Camp Margetts area are listed
below (excluding the new mammals reported in this paper):

A. Irdin Manha level (“Houldjin beds”)

Pantodonta

Hypercoryphodon thomsoni Osborn, 1932
Condylarthra

Andrewsarchus {=Paratriisodon) gigas (Qi, 1980)

Perissodactyla

Lophialetes expeditus Radinsky, 1965
Brontotheriidae indet. Qi, 1980

B. Arshanto beds
Rodentia

?Paramyid sp. Dawson, 1964
Tamquammys wilsoni Dawson, Li and Qi, 1984
Advenimus burkei Dawson, 1964
Dinocerata

Gobiatherium mirificum Osborn, 1932
Perissodactyla

Schlosseria sp. cf. Schlosseria magister Radinsky, 1965
Helaletes fissus Matthew and Granger, 1925
Helaletes fissusl Radinsky, 1965

Metatelmatherium cristatum Granger and Gregory, 1943
Protitan minor Granger and Gregory, 1 943

12

Annals of Carnegie Museum

vol. 56

II. The Underlying Bayan Ulan Beds

The following mammalian fauna of early Eocene age is found in these beds
(Zhai, under study):

Multituberculata

Prionessus lucifer
Anagalida
Mimotona borealis
Pseudictops lophiodon
Notoungulata
Palaeostylops iturus
P. macrodon
Acreodi
Pachyaena sp.

Dissacus ( =Plagiocristodon ) serratus

Pantodonta
Pastoralodon lacustris
Dinocerata

Prodinoceras (= Mongo lot her ium) efremovi
Prodinoceras (=Pyrodon) sp.

Perissodactyla
ILambdotherium sp.

Asiatic Middle Eocene Mammals

Asiatic middle Eocene mammals have been found in east Asia (China: Inner
Mongolia, Shandong, and Henan; Mongolian People’s Republic), central Asia
(China: Xinjiang, and U.S.S.R.: Kazakh region), and south Asia (Pakistan and
India). Areas yielding important discoveries are:

1. Central part of Shandong Province. After the original report (Zdansky, 1930),
Young and Bein (1935) and Chow (1957, 1963) added more middle Eocene
mammals. Zhou and Qi (1982) summarized the Shandong Eocene mammals, and
Qi and Meng (1 983) reported two other species of perissodactyls collected in Wufu
County. Appendix 2 lists the middle Eocene mammals from central Shandong.

2. Henan Province. Zhou et al. (1975, p. 179) concluded of the Lushi fauna,
“It is typically a southern extension of the Irdin Manha fauna (type locality), but
contains some interesting forms which are not known in the Irdin Manha” (Ap-
pendix 1).

Gao (1976) divided the Paleogene in the Wucheng and Xichuan Basins, Henan,
into four formations (lower to upper): 1) Yuhuangding Formation; 2) Dacangfang
Formation; 3) Hetaoyuan Formation; and 4) Shangsi Formation. Only a coryph-
odontid and a dinoceratan were found in the Yuhuangding Formation, the age
of which is considered to be, “not later than Middle Eocene.” Fossil mammals
found in the Dacangfang Formation are:

Rodentia

Sciuravus sp.

Carnivora

Miacis sp. aff. M. invictus

Creodonta

ISinopa sp.

ITritemnodon sp.

Acreodi

lAndrewsarchus sp.

Perissodactyla

Sianodon sp.

1987

Qi— Eocene Mongolian Arshanto Mammals

13

Teleolophus sp. cf. T. medius
Deperetella new species
Breviodon sp. cf. B. minutus
Lophialetes sp.

Colodon sp.

IProtitan sp.

Tillodontia

Chungchienia sichuanica

Gao (1976) concluded that [translated from Chinese], “The age of the majority
of the fossil mammals collected from Hetaoyuan Formation should be the same
as the Irdin Manha fauna found in Inner Mongolia.”

Accordingly, the Yuhuangding Formation may be tentatively correlated with
the Guanzhuang Formation, the Dacangfang Formation with the Arshanto beds,
and the Hetaoyuan Formation with the Irdin Manha beds.

3. U.S.S.R. Since the early 1960s, middle Eocene mammals have been reported
from the Zaysan (Zaisan) Basin, Kazakh region (Appendix 3).

4. Pakistan and India. Following Pilgrim’s description (1940) of middle Eocene
mammals from North-west India, Dehm and Oettingen-Spielberg (1958) de-
scribed the middle Eocene Ganda Kas fauna from Pakistan. Later, several studies
reported on middle Eocene mammals from India and Pakistan (West, 1980, and
Appendix 4).

Additional collections of possibly middle Eocene mammals have been made
in southern provinces of China, including Yunnan, Guangzi, Hubei, and Jiangxi.

Middle Eocene Faunal Zones in East Asia

There are three middle Eocene mammalian faunal zones in east Asia. From
oldest to youngest, there are: 1) Guanzhuang faunal zone (Shandong Province);
2) Arshanto faunal zone (Inner Mongolia); and 3) Irdin Manha faunal zone.

Hyrachyus modestus collected from the Guanzhuang Formation also occurs in
Bridger B beds in North America. The tillodont, Kuanchuanius shantunensis
(Chow, 1963), from Shandong resembles Trogosus and Tillodon (both lower Brid-
ger; Stucky and Krishtalka, 1983). Accordingly, the age of the Guanzhuang and
lower Bridger faunas may be comparable.

In the Arshanto fauna, Hyrachyus sp. cf. H. eximius is close to Hyrachyus
eximius from the Bridger C and D. Other similar tapiroids, such as helaletids,
occur in both faunas. This indicates that the Arshanto fauna is younger than the
Guanzhuang fauna and closer in age to the upper Bridger fauna. Homogalax, a
typical North American Greybullian tapir, survives as a relict in the middle Eocene
Arshanto fauna of Asia.

Although the Guangzhuang and Arshanto faunas have certain endemic taxa,
their composition suggests that faunal exchange took place during the early and
late Bridgerian between east Asia and North America.

The Arshanto and the Irdin Manha faunas appear to be closely related. A main
characteristic of the Arshanto fauna is the dominance of a variety of perissodactyls
and the absence of artiodactyls. Of the 43 species in the Arshanto fauna, about
23 (53%) are perissodactyls. Perissodactyls still dominate the Irdin Manha fauna,
but several new families appear: Leporidae, Oxyaenidae, Hyaenodontidae, Miac-
idae, and Amynodontidae(?) .

The following species occur in both faunas: Breviodon minutus, IForstercooperia
grandis and Pantolambdodon fords. Genera in common are: Hapalodectes, Mes-
onyx, Helaletes, Teleolophus, Protitan, and Telmatherium. Importantly, phylo-

14

Annals of Carnegie Museum

vol. 56

genetic relationships are evident between several taxa in the two faunas such as:
Schlosseria magister and Lophialetes expeditus ; Helaletes medius and H. mon-
goliensis ; Telmatherium cristatum and T. parvum ; Protitan minor and P. robustus;
and Metacoryphodon, Eudinoceras, and Hypercoryphodon.

The similarities between the Arshanto and Irdin Manha faunas are in contrast
to the differences between the latter and the Shara Murun fauna. There are only
three shared genera (Shamolagus, Rhodopagusl and Triplopusl) in the Irdin Man-
ha and Shara Murun faunas, and no shared species. There are some phylogenetic
relationships between several species such as Teleolophus medius and Deperetella
cristata, and Forstercooperia confluens and Juxia sharamurunensis, as well as
within the Rodentia and Lagomorpha. But in the Shara Murun fauna, brontotheres
and artiodactyls (mainly Archaeomeryx optatus) are dominant, whereas they are
absent from the Irdin Manha fauna. The appearance of the Family Anthracothe-
riidae in the Shara Murun fauna marks the beginning of the late Eocene in China.

East Asia and central Asia were in the same paleobiogeographic district (North
Paleobiogeographic District) during the middle Eocene. The central Asian middle
Eocene mammalian faunas occur mainly in the Kazakh and Kirgizia regions,
U.S.S.R., and in Xinjiang, China, but correlating these faunas is difficult. The
fauna from Xinjiang appears to be equivalent to the Irdin Manha fauna. The
central Asian faunas are generally similar in composition to those of east Asia,
and may represent extensions of the eastern Arshanto and Irdin Manha faunas,
although not of the Guanzhuang fauna. Middle Eocene genera common to central
and eastern Asia are: Lophialetes, Helaletes, Teleolophus, Forstercooperia, Tri-
plopus, Pataecops, Eudinoceras, and Tamquammys.

South Asia was another paleobiogeographic district (South Paleobiogeographic
District) in the middle Eocene. At present, the Qinghai-Tibetan plateau stands
between east-central Asia and south Asia. During the middle Eocene, at least
during a large part of that time, Tethys lay, in part, between the two. Recent
discoveries do not support the assumption that south Asia and east-central Asia
were in the same paleobiogeographic district. Unquestionably there were some
links between the two areas. Some genera such as Schlosseria, Teleolophus, For-
stercooperia, and Tamquammys were shared between south Asia and east Asia,
but there are considerable differences at the specific level between the two regions.
The appearance of Gobiohyus orientalis in the two areas indicates that the age of
the south Asian faunas is, in part, equivalent to that of the Irdin Manha fauna
and that during the time of the Irdin Manha fauna the eastern part of the Tethys
Sea had already closed, allowing exchange of some mammals.

Finally, Sahni et al. (1981, p. 693) pointed out that the middle Eocene south
Asian fauna “includes perissodactyls and rodents, most of them being endemic
at the generic level, showing strong and indubitable affinities to the Eocene faunas
of Northern Asia, Mongolia and Kazakhstan. The diversity among these groups
and the endemicity of the taxa strongly suggest an Early Eocene immigration from
the North.” A recent study of the origin of the proboscideans (Domning et al.,
1986) suggests, however, another route for faunal migration between south-east
Asia and south Asia, namely, a South China-Burma-Indo-Pakistan route.

Systematic Paleontology of the Arshanto Fauna

The discoveries of the Arshanto fauna are centered in the Shara Murun area,
i.e., 1 10°30' to 1 12°30'E; 43°00' to 43°40'N. Most of the fossil mammal sites are
located in Sunid Yuqi (county) of Ulan Chabu Mong, Inner Mongolia.

1987

Qi— Eocene Mongolian Arshanto Mammals

15

Grandorder Insectivora Illiger, 1811
Order Soricomorpha Gregory, 1910
Superfamily Palaeoryctoidea Winge, 1917
Family Micropternodontidae Stirton and Rensberger, 1964
Sinosinopa , new genus

Type species. —Sinosinopa sinensis, new genus, new species.

Included species. — Type and only species.

Diagnosis. — Retains M| unlike Prosarcodon and Sar codon; P4 more elongate
and M1-2 with larger hypocone shelf.

Sinosinopa sinensis , new species
(Fig. 5; Table 1)

Holotype. — V5 6 7 7 , incomplete skull and mandible, with left P!-M3, right C1-
P4 (M1-2 broken), and left M2_3 (field no. 77036-H5).

Diagnosis. — Dental formula ?/?, 1/?, 4/7, 3/3; large in size; diastema very short;
nasal bone elongate; post-metacrista fairly long; hypocone shelf (talon basin) very
wide; M3 with two paraconules; trigonids and talonids well developed; trigonid
of M3 very long; metaconulid fairly prominent.

Description. — Skull: ascending ramus of premaxilla present; maxilla relatively high and long; nasal
bone very elongate, its posterior edge above M1; frontal bone relatively broad; orbit large; infraorbital
foramen fairly small; zygoma not deep (posterior part broken); palatine part of maxilla very long; and
foramen incisivum fairly large.

Upper teeth. C1: relatively robust; transverse section oval; postcanine diastema fairly short. P1:
smallest in size; compressed anteroposteriorly; one main cusp, anterior edge of cusp relatively steep.
P2: larger in size; anterior edge of main cusp less steep than that of P1, posterior edge also gentle;
cuspule on posterior edge of tooth, but not clear. P3: crown view triangular; paracone and protocone
clear; small cuspule (parastyle) on anterior edge of tooth; another larger cuspule (metastyle) on posterior
edge. P4 (paracone and metacone broken): becoming molariform; protocone prominent; pre-protocrista
and post-protocrista clear, forming a narrow trigon basin; hypocone relatively prominent; hypocone
shelf very narrow; parastyle small, relatively prominent; metastyle prominent, but not extending very
far posteriorly. M1: protocone prominent; pre-protocrista and post-protocrista not prominent; pre-
paraconule crista relatively clear; post-paraconule crista not clear; trigon basin narrow; parastyle and
metastyle (crowns broken) forming a “V”-shaped ectoflexus (sharp at bottom of “V,” and anterior
wing of “V” longer than posterior wing); a small cingulum in front of paraconule; trigon basin wide;
a prominent hypocone and another small cusp on postero-lingual edge of trigon basin. M2: post-
metacone crista (broken) probably fairly elongate; protocone prominent; trigon basin narrow; para-
conule larger than metaconule; pre-paraconule crista less clear than that of M1; stylar shelf narrow;
parastyle cone-like; metastyle relatively elongate; ectoflexus opening largely; small segment of cingulum
anterior to paraconule; trigon basin wide; prominent hypocone and another cuspule at postero-lingual
edge of trigon basin. M3: protocone prominent; pre-protocrista and post-protocrista clear; with two
paraconules, one on lingual side smaller than other; trigon basin wide; paracone close to metacone;
paracone high; no metastyle, cingulum, and hypocone.

Mandible: horizontal ramus slender; anterior edge of ascending ramus relatively steep; anterior edge
of masseteric fossa at rear of M3.

Lower cheek teeth. P4 (main cusp broken): anterior and posterior cuspules clear; section of crest in
posterior part of tooth. M,: (only two roots left). M2: trigonid and talonid clear, length almost same;
protoconid-paraconid-metaconid forming distinct trigonid; camassial notch between protoconid and
paraconid stronger than that between protoconid and metaconid; positions of hypoconid, entoconid,
and hypoconulid somewhat posterior; cristid obliqua between hypoconid and protoconid relatively
strong; small cingulum on antero-labial side of paraconid. M3: talonid narrow and elongate, hypo-
conulid very prominent; distance between hypoconid and entoconid, and hypoconulid longer than on
M2; cristid obliqua stronger.

Discussion.— When Matthew and Granger (1925 a) described the genus Sar-
codon, only M1 was known. Its systematic position was uncertain until Szalay and
McKenna (1971, pp. 286-293) recognized the lower teeth of Opisthopsalis vetus

,4 1cm

b

Fig. 5 .—Sinosinopa sinensis new genus, new species. Holotype, V5677. (a) skull, lateral and palatal
views, (b) mandible, crown and lateral views.

1987

Qi— Eocene Mongolian Arshanto Mammals

17

Table 1 . —Measurements (mm) of Sinosinopa sinensis (V5677).

l w

C1

P1

P2

P3

P4

M1

M2

M3

pi-4

M'-3

Pl-M3

O-M3

P3

P4

M2

M2 trigonid
M, talonid
M3

M3 trigonid
M3 talonid

Mandibular depth below P<

1.8

1.7

1.8

1.0

2.2

1.1

2.5

2.0

3.2

2.8

2.7

4.1

3.0

4.4

1.5

3.9

10.3

7.3

16.9

20.9

3.0

1.4

2.3

3.3

1.9

1.7

1.9

1.6

1.7

3.7

1.9

1.8

1.9

1.9

1.4

3.4

Matthew, Granger and Simpson, 1929, as referable to Sarcodon, and considered
similarities of upper dentitions of Sarcodon and Micropter nodus a result of con-
vergence.

When McKenna et al. (1984) described Prosarcodon lonanensis, they made
detailed comparisons between Prosarcodon, Sarcodon, and Sinosinopa. Sinosi-
nopa differs from Prosarcodon and Sarcodon in that “it retains all three molars.
In Sinosinopa both M1 and M2 possess a hypocone that juts strongly postero-
lingual; in Prosarcodon M2, the last molar, possesses a narrow cingulum-like
hypocone. P4 of Sinosinopa is more elongate than in either Sarcodon and Pro-
sar codon.” They concluded, “That Prosarcodon, Sarcodon, and Sinosinopa are
lipotyphlan insectivores is suggested by the presence of a piriform fenestra in
Prosarcodon

There is a well developed hypocone shelf on M1 of the type specimen of Sarcodon
(AMNH 20427). Sarcodon or Prosarcodon could not be ancestral to Sinosinopa
because Sinosinopa has M3. Apparently two different lineages of these early in-
sectivores existed in Asia: 1) Prosarcodon and Sarcodon, which lost M3 and de-
veloped the hypocone shelf; 2) Sinosinopa, which kept M3 but “is more specialized
in having jutting hypocones on several upper molars” (McKenna et al., 1984, p.
13).

Order Rodentia Bowdich, 1821
Family Paramyidae Miller and Gidley, 1918
Subfamily Reithroparamyinae Wood, 1962
Asiomys, new genus

Type species . — Asiomys dawsoni, new genus, new species; only known species.

Diagnosis.— Differs from Reithroparamys in having a prominent protoloph on

, j.5 cm

Fig. 6 .—Asiomys dawsoni new genus, new species, (a) holotype, V5684, M1; (b) V5685, Mt; (c)
V5686, M2.

M1, a closed trigonid basin on M,; differs from Franimys in having closely ap-
proximated paracone and metacone on the upper molars; unlike Rapamys Mj
posterior cingulid not close to entoconid.

Asiomys dawsoni , new species
(Fig. 6)

Holotype. — V 5 6 8 4 , a right M1 (field no. 77026).

Referred specimens. — V5 6 8 5 , right M, (field no. 77027); V5686, left M2, and
incisor (field no. 77028).

Localities.— Irdin Manha area (type), Huhe Bulak, and Daatein Obo.

Diagnosis. — Medium-sized reithroparamyine; upper and lower cheek teeth nearly
square in crown view; double metaconules on upper molar; protocone-metacon-
ules-metacone forming a crest; ectolophid developed on lower molar.

Description. — M1 (L, 3.8; W, 4.6): protocone prominent, paracone and metacone conical, close
together; paraconule near protocone, paracone-paraconule-protocone forming paraloph; double meta-
conules, that near protocone larger in size, the other not prominent; metaloph connected with protocone
directly, not with hypocone; mesostyle relatively apparent; hypocone prominent; hypocone shelf rel-
atively broad; posterior cingulum low and wide but not enclosing metacone; anterior cingulum narrow
and short, ended at the paraconule lingually; straight valley on lingual wall between protocone and
hypocone but not reaching base. M, (L, 4.0; W, 3.9): double crests from protoconid forming an anterior-
posterior short, closed trigonid basin; ectolophid developed but not closed to labial side and connecting
protoconid and hypoconid; hypoconid relatively prominent, connecting with posterior cingulum; two
small points on posterior cingulum; entoconid isolated. M2 (W, 4.0): similar toM„ but larger in size
and protolophid II shorter, so trigonid basin opens posteriorly.

Discussion.— The discovery of middle Eocene Asiomys dawsoni is the first re-
cord of the Reithroparamyinae in Asia. This subfamily includes three other genera,
Reithroparamys, Franimys, and Rapamys , all from North America (Wood, 1962).

In size, Asiomys is close to both paramyines and reithroparamyines. The upper
molars of the paramyines have only one metaconule, while the reithroparamyines
have double metaconules. The earlier members of the Reithroparamyinae, such
as Reithroparamys, have double metaconules on the upper molars, one larger and
another smaller. In addition to this similarity, the referred lower teeth from Inner
Mongolia resemble reithroparamyines in features such as medium size and ento-
conid completely separate from posterior cingulum.

1987

Qi— Eocene Mongolian Arshanto Mammals

19

Table 2.— Measurements (mm) o/Tamquammys wilsoni.

V5678

L

w

V5679

V5680

L

W

trigonid

w

talonid

L

W

trigonid

w

talonid

dP3

0.7

0.9

dP4

1.2

1.7

P4

1.3

1.7

M1

1.4

1.8

M2

1.5

Diastema

4.0

P4

1.5

1.1

0.9

M,

1.5

1.3

1.5

1.3

1.2

1.4

m2

1.7

1.5

1.6

1.7

1.4

1.5

m3

1.9

1.6

1.5

1.8

1.5

1.4

p4~m3

6.7

m,-m3

5.1

4.5

M1 of the Inner Mongolian species differs from that of Reithroparamys in several
points. The protoloph of the Inner Mongolian specimen is more prominent,
whereas it is indistinct, if present at all, in Reithroparamys. In Reithroparamys
delicatissimus, for instance, the end of the protoloph is anterior to the protocone,
not connected with it. The protolophid II of Reithroparamys is short, not con-
necting with the metaconid to form a closed trigonid basin. The Inner Mongolian
species has a well developed ectolophid, whereas some reithroparamyines, such
as R. debequensis and R. pattersoni, have only a mesoconid or nothing.

The main distinctive characters of Franimys occur in its skull and limbs, but
the genus is characterized also by having paracone and metacone of its upper
molars well separated. This character is quite different from the condition in the
Inner Mongolian species. Besides this, Franimys has a very short protolophid II.

M1 and M2 of Rapamys have two or three metaconules. Another of its main
characters never appeared on other reithroparamyines —having a prominent but
short posterior cingulid that is very close to the entoconid.

Family Cocomyidae Dawson, Li and Qi, 1984
Tamquammys Shevyreva, 1971
Tamquammys wilsoni Dawson, Li and Qi, 1984
(Figs. 7, 8; Table 2)

Holotype.—Y561S , anterior portion of skull with broken incisors, right P4, left
dP3~M2.

Referred specimens.— V 5679 , right jaw with I, P4-M3; V5680, right jaw with
M^; V5681, M^ V5682, M2 (broken); V5683, left lower molar. All specimens
came from one nodule.

Locality.— Huhe Bulak.

Discussion.— Discoveries in recent years have led to clarification of the clas-
sification of Ctenodactyloidea (Dawson et al., 1984). Only one family, Cteno-
dactylidae, was recognized when Simpson (1945) established the superfamily
Ctenodactyloidea. Now, two other ctenodactyloid families, Cocomyidae and Yu-
omyidae, have been established.

Distribution of Tamquammys is middle Eocene in Kazakhstan and Inner Mon-
golia. Revised diagnosis of Tamquammys is as follows (Dawson et al., 1984, pp.

20

Annals of Carnegie Museum

vol. 56

^ .5 cm

Fig. 7 . — Tamquammys wilsoni. Holotype, V5678, skull, left lateral and palatal views.

142-143): “Ctenodactyloid rodent with skull hystricomorphus, lower jaw sci-
urognathous. Cheek teeth increase in size from front to back, have well developed
conules in upper teeth, well developed lophs in lowers. P4 with single buccal cusp.
P4 with relatively short and narrow talonid.” Tamquammys wilsoni has the fol-
lowing diagnosis: “Smaller than T. tantillus ; P4 has talonid less well developed
posterobuccally, hypoconid-hypoconulid ridge more oblique, less transverse in
orientation.”

1987

Qi— Eocene Mongolian Arshanto Mammals

21

4.5 cm

22

Annals of Carnegie Museum

vol. 56

Order Acreodi Matthew, 1 909
Mesonychidae Cope, 1875
Mongolonyx Szalay and Gould, 1966
Mongolonyx dolichognathus Szalay and Gould, 1966
(Fig. 9)

Holotype.— AMNH 26661, lower jaw with several well-preserved teeth, coro-
noid processes broken off; collected during the 1930 Central Asiatic Expedition;
AMNH field no. 895.

Referred specimens.— AMNH 26662 (AMNH field number 907), a complete
left maxilla with P-M2. Newly referred specimens V5690, M2 (field number
77039); V5691, M? (field number 77031).

Localities. — 7 miles west of Camp Margetts (type), Huhe Bulak, Bayan Ulan.

Discussion.— When Szalay and Gould (1966) described this species, they men-
tioned the uniqueness of M2: “The last molar, M2, is the best preserved and most
diagnostic tooth in the maxilla. The paracone is at least twice the size of the
protocone, while the small metacone is connate with the posterior wall of the
paracone. M2 is only about two-thirds as wide as the preceding M1.”

Newly discovered specimens from Inner Mongolia are two teeth: M2 (V5690)
and M2 (Y5691). Characters of M2 are: 1) paracone twice the size of protocone;
2) metacone very small; 3) labial tooth wall convex; and 4) cingula very weak.
M2 has a prominent paraconid (although broken), a very strong protoconid, a
relatively weak hypoconid. The enamel ridge at the base of the tooth is almost a
straight line. M2 (V5690) is quite similar to that of AMNH 26662 in both size
and morphology, and M2 (Y5691) is close also to M2 of the type specimen.
Discovery of these two teeth establishes the geological provenance of Mongolonyx
dolichognathus as the Arshanto beds.

Measurements. — L/W : M2 (V5690) 26.3/27.2 mm, (AMNH 26661) 26.8/27.9 mm; M2 (V5691)
31.2/16.5 mm; (AMNH 26662) 32(?)/16.5.

Mesonyx Cope, 1872

Mesonyx sp. cf. M. obtusidens (Cope, 1872)

(Figs. 10, 11; Table 3)

Referred specimens.— V 5692: left P4, with three roots (field no. 77036-2);
V5696.1, M2 (field no. 77027); V5693: M2, with three roots (field no. 77036-2);
V5694: lower canine (field no. 77027); V5695, 1-2: P3 or P4 (field no. 77027);
and V5696: M3 (field no. 77026).

Localities.— Huhe Bulak and Irdin Manha area.

Description. — P4: paracone larger than protocone; parastyle prominent; metastyle very weak. M2:
protocone slightly larger than paracone, two convex crests on labial wall of protocone, protocone wall
flat between the two crests; parastyle smaller in size; metacone larger; parastyle and metacone oblique;
metastyle low but clear; extremely worn cingulum on buccal wall, clearer near metastyle; anterior edge
of tooth concave, whereas posterior edge is relatively straight.

P3 (or P4): paraconid small cusp; protoconid large, relatively far back; position of hypoconid much
lower than that of protoconid; no cingulum. M3: crown narrow and low; paraconid relatively prominent;
protoconid larger; length of hypoconid two-thirds length of whole tooth.

Discussion. — Cope (1872) first reported Mesonyx obtusidens , but the most com-
plete skull (AMNH 12643) with upper cheek teeth was described by Matthew
(1907). According to the descriptions of Matthew, M1 of this species has a meta-
style. Morphologically, the Inner Mongolian specimen is quite similar to AMNH
12643 at least in crown view. However, the Inner Mongolian specimen is larger

1987

Qi— Eocene Mongolian Arshanto Mammals

23

Table 3 . — Measurements (mm) of Mesonyx sp. cf. M. obtusidens.

L

w

L

w

P4 (V5692)

15.6

14.8

M2 (V5693; V5696.1)

19.7

18.3

20.0

17.5

P3 (V5695-1)

19.2

8.5

P4 (V5695-2)

19.7

9.8

M3 (V5696)

16.9

7.3

and the parastyle and metacone are oblique, whereas those of AMNH 12643 are
almost erect. The M3 found at locality 77026 is unusual and is only tentatively
referred to this species.

Y5 6 9 6. 1, an M2, is similar to V5693 (M2), except for the absence of a metastyle
and being longer (L, 20.0 mm; W, 17.5 mm).

Hapalodectes Matthew, 1909
Hapalodectesl serus (Matthew and Granger, 1925)

(Fig. 12)

Referred specimen. — V7 3 1 6 , a broken lower jaw with posterior part of M2 and
single root of M3 (field no. 77027).

Locality.— Huhe Bulak.

Discussion. — Matthew and Granger (19256) described two mesonychids, Hap-
alodectes serus (only one tooth) and H. auctus, the former being smaller than the
latter. M2 of V7316, which is broken anteriorly, resembles that of Hapalodectes
serus: it is hypsodont and the crest of the posterior part is sharp. There is a shallow
masseteric fossa on the lower jaw.

Measurements. — AMNH 20172 (lower cheek tooth): 5.6 mm; V7316: 5.3 mm (ca.).

Order Pantodonta Cope, 1873
Family Coryphodontidae Marsh, 1876

Metacoryphodon , new genus

Type species.— Metacoryphodon luminis new genus, new species.

Diagnosis. — Larger than Coryphodon\ unlike Eudinoceras, P2-3 protocristae
present, canine more robust and longer.

Metacoryphodon luminis , new species
(Fig. 13; Table 4)

Holotype. — V5697, a broken skull with canines (not complete), right P!-M3 and
left P2=M3 (field no. 1P5H8).

Locality. — Ulan Bulak.

Diagnosis. — Canine extremely robust; P2 and P3 with apparent pre-protocristae;
post-protocristae tend to disappear; M3 has trace of paracone, ectolophs very long
on the molars.

Description. —Nasal bone (broken) wide transversely, postcanine diastema not long; frontal bone
wide and flat; zygomatic arch very strong. C1: very robust; transverse section round; the growing
directions of the two canines not upright, but inclined outward. P1: no protocone; paraloph longer
than metaloph; cingulum developed. P2: protocone strong with apparent pre-protocrista and weak
post-protocrista; pre-protocrista extends downward and becomes a part of anterior cingulum, but not
connected to inner cingulum; paraloph and metaloph very long, the two joining to form a “V”; parastyle
more swollen than metastyle. P3: pre-protocrista and post-protocrista are developed but weak; parastyle

24

Annals of Carnegie Museum

yol. 56

Fig. 9 .—Mongolonyx dolichognathus. (a) ¥5690, left M2; (b) ¥5691, M2
Fig. 10, 1 l.—Mesonyx sp. cf. M. obtusidens. 10. ¥5693, right M2. 11. ¥5696.1, right M2 .
Fig. 1 2. —Hapalodectesl serus. ¥7316, lower jaw. (a) external, (b) internal, (c) crown view.

1987

Qi— Eocene Mongolian Arshanto Mammals

25

Fig. XZ.— Metacoryphodon luminis new genus, new species. Holotype, V5697, right P'-M3, crown
view.

close to metastyle, so the valley between paraloph and metaloph is narrow. P4: pre-protocrista very
weak; some crenulations on the anterior cingulum; post-protocrista weak, so protocone almost isolated.
M1: protocone high; protoloph almost parallel with ectoloph; parastyle relatively prominent; hypocone
small, but distinct; metastyle swollen; cingulum present on anterior, inner, and posterior sides of
crown, but not buccally. M2: quite similar to M1; ectoloph long; hypocone relatively large. M3: wider
but shorter than M2; protoloph very long, with trace of paraconule; parastyle low; paracone isolated;
ectoloph short, its central part convex anteriorly; no metastyle.

Discussion. — Some characters of this species are intermediate between Coryph-
odon and Eudinoceras. The type specimen was found in 1977 by the staff of the
Inner Mongolian Geological Mapping Team. In size it is close to Eudinoceras
mongoliensis but larger than Coryphodon.

Eudinoceras mongoliensis (P3 or P4) was first described from Irdin Manha beds
by Osborn (1924). In 1931, Osborn and Granger described another species, E.
kholobolchiensis, from Kholobolchi Nor Basin, Mongolian People’s Republic and,
in 1932, described ? Eudinoceras mongoliensis from a lower jaw with cheek teeth.
Many years later, Tong and Tang (1977) reported a new species, Eudinoceras
eras sum. During the summer field season of 1975, the author and the staff of
Geological Mapping Team of Inner Mongolia discovered a broken skull and lower
jaw of Eudinoceras mongoliensis with upper and lower incisors, canines, and cheek
teeth on one side.

Although there are some differences between E. mongoliensis and E. kholo-
bolchiensis, protocones of P2 and P3 in both are isolated, that is, there are no pre-
protocristae and post-protocristae. Metacoryphodon luminis has apparent pre-

Table A.— Measurements (mm) of Metacoryphodon luminis (V5697).

L

w

o

37.0

41.0

pi

19.0

17.0

P2

25.5

34.8

P3

27.0

37.9

P4

21 A

43.7

M1

38.0

43.6

M2

46.5

52.4

M3

44.4

57.1

pi— 4

96.0

M1-3

122.7

P'-M3

216.6

26

Annals of Carnegie Museum

vol. 56

Fig. 14 .—Metacoryphodon! minor new species, (a) Holotype, V5698, P2; (b) V5699, P2.
Fig. 15 .—Metacoryphodon sp. V5700, right P,.

Fig. 1 6.— Pantolambdodon fortis. V5687, right M,. (a) crown, (b) internal, (c) external views.
Fig. 17 .—Pantolambdodon*} minor. V5701, broken right M1.

protocristae and weak post-pro tocristae on P2 and P3. The canine of E. mongo -
liensis is inclined laterally and its crown is sword-shaped with saw-like edges,
whereas the canine of M. luminis is more robust and longer and is directed
anteriorly, as are the lower canines of 1 Eudinoceras mongoliensis.

Coryphodon is somewhat smaller than Metacoryphodon. P1-4 of Coryphodon
have apparent pre-protocristae and post-protocristae.

Metacoryphodon is morphologically an evolutionary intermediate between early
Eocene Coryphodon and the late middle Eocene Eudinoceras and Hypercorypho-
don. In the evolution from Coryphodon to Metacoryphodon to Eudinoceras the
pre-protocristae and post-protocristae gradually disappeared.

Metacoryphodon*} minor, new species
(Fig. 14)

Holotype. — V5 6 9 8 , a right P2 (field no. 77027).

Referred specimen.— V 5699, a right P2 (field no. 77039).

Locality.— Huhe Bulak.

Diagnosis.— Smaller than M. luminis; metastyle of P2 more prominent than
parastyle; opening of V-shaped valley formed by paraloph and metaloph is blocked
by cingulum; pre-protocristae well developed, no post-protocristae; P2 lacking
outer and inner cingula.

1987

Qi— Eocene Mongolian Arshanto Mammals

27

Table 5. —Measurements (mm) o/ Pantolambdodon ieemis and P. fortis.

Pantolambdodon inernis

Pantolambdodon fortis

AMNH 21558

AMNH 22100

AMNH 21748

AMNH 26127

¥5687

L

W

L

w

L

w

L

w

L

w

M,

20.0

9.5

20.0

11.0

29.0

13.0

27.8

14.5

M2

21.5

11.0

22.7

11.5

m3

26.0

11.5

23.0

9.0

M. trigonid

11.5

9.0

11.0

10.0

16.0

13.5

14.0

14.0

M, talonid

8.5

9.0

9.0

10.0

16.0

13.0

14.0

12.5

M2 trigonid

11.5

9.5

13.0

10.5

M2 talonid

10.0

9.0

9.5

9.5

M3 trigonid

15.0

11.0

12.5

9.5

M3 talonid

11.0

8.0

11.0

7.0

Discussion. — Both teeth are small P2 has well developed pre-protocristae, but
no trace of post-protocristae. This character is quite different in M. iuminis . The
protocone of M? minor is not as strong as that of M. Iuminis , its parastyle is
relatively weaker, and there are no traces of a parastyle or metastyle. In addition,
the opening of the V-shaped valley in M. Iuminis is not blocked by the cingulum.

Because no lower jaw of M. Iuminis is known, comparison with P2 (V5699) is
not possible. P2 of Metacoryphodoril minor (L, 18.3 mm; W, 14.6 mm) is smaller
than that of Eudinoceras mongoliensis , and following the size of P2 (L, 22.3 mm;
W, 32.2 mm), is certainly smaller than M. Iuminis.

Metacoryphodon sp.

(Fig. 15)

Referred specimen.— V 57 00, a right Fl (field no. 77036H2).

Locality.— Huhe Bulak.

Discussion., -—This tooth was found at a lower level The size of ?! is close to
that of Eudinoceras mongoliensis but it shows several differences from that taxon:
1) paralophid of P, of E. mongoliensis is shorter and has a small cusp on the
anterior end, whereas the paralophid of V5 700 is longer and has two small cusps
(one on the anterior end and another one at the middle of paralophid); 2) the
paralophid and metalophid of E. mongoliensis are almost on a straight line,
whereas the same lophids of V5700 form a V with a large opening; 3) V5700 has
a cingulum behind the metastylid and a small cusp (worn) on the posterior end
of the metastylid, whereas these are absent in Eudinoceras mongoliensis; 4) V5700
has no cingulum, but Pi of E. mongoliensis is enclosed by cingulum at the base
of the crown.

Measurements. —Pj (V5700), L: 22.1 mm. Wi 13.1 mm.

Family Pantolambdodontidae Granger and Gregory, 1934
Pantolambdodon Granger and Gregory, 1934
Pantolambdodon fortis Granger and Gregory, 1934
(Fig. 16; Table 5)

Holotype.— AMNH 26127, fragmentary right ramus of lower jaw with and
alveoli of all anterior teeth.

Referred specimen.— V 5687 , a right (field no. 77027).

Localities. — 8 miles north of Tukhum Lamasery; Huhe Bulak.

Discussion. — The first fossils of this taxon were found in 1925 and 1928, and

28

Annals of Carnegie Museum

vol. 56

the systematic position of this family was reached by roundabout means. When
Granger and Gregory (1934) studied these materials, they first decided that they
“are dealing with a placental mammal of some sort from the fact that the dental
formula: I3, C1? P4, M3, is the classic primitive eutherian formula.” They made
many comparisons between the Mongolian types and many kinds of mammals,
such as some South American mammals, chalicotheres, titanotheres, artiodactyls,
condylarths, and dinoceratans.

At last, Granger and Gregory (1934, pp. 5-6) discovered that these Mongolian
types are related to pantodonts: “When we come to Pantolamhda and Titanoides,
however, we find some apparently reliable indications of remote relationship to
the Mongolian types, especially in the form of the premolars and molars. In spite
of the fact that Titanoides is a graviportal form almost as big as Coryphodon, it
shares the following features with the Mongolian fossils: (1) Dental formula of
primitive placental type; (2) P1? P2 compressed; (3) talonid fossa of premolars
formed between the posterior ridge connected with the main cone and a transverse
metaconid ridge; (4) molar talonids with V-shaped crests; (5) talonid of M3 nar-
rower than trigonid; (6) M3 with reduced or no hypoconulids; (7) Ml5 M2 with no
trace of hypoconulid; (8) molars not crowded but slightly spaced; (9) coronoid
process inclined backward. Titanoides is distinguished from the Mongolian forms
by its relatively gigantic size, powerful, more erect incisors and canines; relatively
shorter, more massive jaw, etc.

’’From Pantolamhda the present form differs in its much more elongate slender
jaw, somewhat procumbent front teeth, more hypsodont cheek teeth, compressed
premolars; the molars have much larger anterior V’s and smaller posterior V’s;
the ascending ramus sloped backward and is distinctly delicate.”

Ye (1983) referred to this family several teeth collected in Ulan Shireh, which
was the type locality of this genus. A broken maxilla with upper cheek teeth was
found in Aliusu. These cheek teeth, now under study, belong to a species of this
genus.

Pantolambdodon has two species, P. inermis, and P. fords. The main differences
between the two species are in size, the former smaller, the latter larger. The newly
discovered specimen (V5687) is similar to P. fords in both size and morphology.
The tooth has a shallow valley anterior to the crista obliqua, which has never
been seen on other specimens.

Pantolambdodon ? minor, new species
(Fig. 17)

Holotype.— V5701, a broken M1 (field no. 77026; L, protocone to parastyle,
21.0 mm).

Referred specimen. — V 67 06, a left {Pantolambdodon inermis of Ye, 1983;

L/W of trigonid, 15/6.5 mm).

Localities.— Irdin Manha area, Ulan Shireh area.

Diagnosis. — Small size, pre-protocrista short; post-protocrista long; distance
between paracone and metacone short, having a small metastylid on M{.

Description.— M1: protocone strong; pre-protocrista short, extending down anteriorly; post-proto-
crista long, extending down posteriorly; paracone and metacone relatively robust; distance between
paracone and parastyle relatively long; mesostyle preserved only as a small piece of crest. M,: angle
of V-shaped crest is 55°, metastylid apparent, smaller in size (Ye, 1983).

Discussion. — The two teeth were found at different localities: V5701 from Ar-
shanto beds in the Irdin Manha area; V6706 from Ulan Shireh (beds). Although

1987

Qi- -Eocene Mongolian Arshanto Mammals

29

V5701 (M1) is broken on the buccal side, the diagnostic part is well preserved,
allowing comparison between this M1 and an undescribed specimen from Aliusu
(see P. forth above). The distance between protocone and parastyle is two-thirds
that in the Aliusu specimen, while the distance between protocone and the base
of mesostyle is only half that in the Aliusu specimen. However, the two specimens
share some features: protocone very well developed; pre-protocrista and post-
protocrista similar; paracone close to metacone; and roots underneath paracone-
parastyle equally robust.

Order Dinocerata Marsh, 1873
Family Uintatheriidae Flower, 1876
Gobiatherium Osborn and Granger, 1932
Gobiatherium rnirificurn Osborn and Granger, 1932
(Figs. 18, 19, 20a-c, 21, 22; Table 6)

Holotype.— AMNH 26624, a complete and nearly perfect skull

Faratype. — AMNH 26630, a nearly complete pair of lower jaws, lacking the
incisors.

Referred specimens. —¥5702: broken right maxilla, with P3-4, M2~3 and left M1
(field no. 77036-2); V5703: broken P4 (77034); V5704: right lower jaw with P2-
M3 (77036-2); Y5705: broken lower jaw with right P2-M3 (M2 and M3 heavily
broken) and left P3=4 (field number 1P5H8); V5706: left M3 (77036-2); V5707: right
M3 (77036-2); V5708: posterior part of left M3 (77036-2); V5709: incisor (77021);
Y5710: lower jaw of a juvenile with dP^M^ (77036-2).

Localities.— 25 miles southwest from Iren Dabasu; Huhe Bulak; Arshanto Obo.

Description. P:‘ in crown view nearly quadrangular, paraloph and metaloph forming “V”; paraloph

somewhat convex anteriorly; paracone very prominent, its apex pointing posterolingually; a lingual
cuspule, the paraconule, near the paracone; metaloph straight; metaconule relatively prominent; an-
tenor and posterior cingula very wide, connecting at base of protocone and inclined somewhat upward;
external cingulum very weak. P4: paraloph more convex anteriorly; paracone prominent, but smaller
than on P3; rib-like crest on posterolingual side of paracone; metaloph straight; relatively prominent
metaconule forming rib-like crest; anterior and posterior cingula very well developed; no external
cingulum; P3-4 retain some trace of cement. M1: paraloph straight; paraconule and metaconule present;
metaloph relatively straight; anterior and posterior cingula narrower than on P3^ or M2-3; weak external
cingulum. M2: paraloph relatively straight; paracone somewhat swollen; metaloph straight; metaconule
present; cusp, possibly hypocone, on central part of posterior wall of protocone; anterior cingulum
gradually widened from labial to lingual side; posterior cingulum also wide; anterior and posterior
cingula separated at base of protocone; no external cingulum. M3: largest of upper cheek teeth; paracone
swollen; paraconule present; distinct, saddle-like pit between paraconule and protocone; metacone
prominent; metaconule present (although broken); anterior cingulum wide; posterior cingulum prob-
ably wide (also broken). I, (more than half preserved) characteristically notched. P2: smallest of lower
cheek teeth; protoconid high; paraconid low; protoconid-metaconid-hypoconid forming a crest; hy-
poconulid forming posterior cingulum.; with trace of cement. P3: paraconid very small; metaconid
high; protoloph straight; mesoconid large and low; cingulum lying anterior to paracone on lingual side;
hypoconulid forming posterior cingulum; trace of cement behind hypoconid. P4: larger in size but
otherwise similar to P3.. M,: paraconid prominent; vertical crest of hypoconulid faint; distance between
posterior cingulum and transverse crest (both formed by hypoconulid) is long, forming a wide posterior
shelf. M2: (V5706) metaconid high; mesoconid prominent; paraconid distinct; hypoconulid forming
posterior cingulum; entoconid very prominent. M3: (V5705) largest tooth; protoconid and paraconid
distinct, joined by low crest; metaconid high; mesoconid prominent; hypoconid parallels hypoconulid;
entoconid somewhat swollen and forming rib-like crest.

Juvenile (V5710): body of lower jaw rather slender; ascending ramus wide anteroposteriorly; mas-
seteric fossa shallow; short dP,-dP2 diastema. dP2: rather long anteroposteriorly; paraconid and me-
soconid distinct; protoconid very high; metaconid relatively prominent; mesoconid clear; hypoconid
forming a transverse crest and connected to mesoconid; hypoconulid forming posterior cingulum. dP3:
crown view quadrangular; paraconid relatively prominent; anterior cingulum extremely weak; hy-

30

Annals of Carnegie Museum

vol. 56

Fig. 1 8. — Gobiatherium mirificum. Holotype, AMMH 26624, dorsal, lateral, and palatal views of skull
(from Osbom and Granger, 1932).

1987

Qi— Eocene Mongolian Arshanto Mammals

31

Fig. 19.— Gobiatherium mirificum. Paratype, AMNH 26630, lateral and crown views. The anterior
alveolar border is restored from the left side (from Osborn and Granger, 1932).

poconulid forming posterior cingulum. M,: similar to type specimen, but distance between hypoconid
and hypoconulid longer, vertical crest of hypoconulid longer and posterior cingulum (formed by
hypoconulid) wider.

Discussion. — The newly discovered specimens are similar to the type specimen,
but because Osborn and Granger (1932) gave only a simple description of the
type, a more complete description is given here.

The type specimen was found in the “Camp Margetts” area; that is, the present
Huhe Bulak-Ulan Bulak area. V5709 was found in the Arshanto Obo area.

Gobiatherium ? major , new species
(Fig. 23)

Holotype. — V5 7 1 2 , two broken lower jaws, associated, with left dP3 and Ml
and right dP4, Ml5 and M3 (posterior part) (field no. 1P5H8).

Locality. — Ulan Bulak.

Diagnosis. — Large size; protoconid very prominent; “shelf’ formed by M3

hypoconulid and entoconid very low, flat, not steep, and relatively wide.

Description.— dP3: larger than dP4; hypoconid crest relatively long; hypoconulid large and wide. dP4:
small; protoconid not prominent; almost no hypoconulid. M,: large; protoconid very prominent;
protolophid straight and wide; hypoconulid transversely wide. M3 (only posterior part): hypoconulid
anteroposteriorly long, transversely wide and forming a wide “shelf’; entoconid distinct; hypoconulid
crest and entoconid crest very weak.

Discussion. — This species differs from G. mirificum in that: 1) dP4 of G. miri-
ficum is larger than dP3, whereas dP4 of this species is smaller than dP3; 2) pro-
toconid of G. mirificum is very weak or absent, whereas the protoconid of this
species is large and prominent; 3) on M3 of G. mirificum , the hypoconulid and
entoconid are very steep and form the posterior cingulum, whereas hypoconulid
and entoconid in this species form a low, flat, wide shelf.

Measurements.— L/W: dP3, 23.0/17.4? mm; dP4, 19.57/18.0 mm; M„ 29.0/24.0 mm.

32

Annals of Carnegie Museum

vol. 56

Fig. 20a-c. — Gobiatherium mirificum. (a) V5702, right P3-4, M2-3, and left M1; (b) V5707, (c) V5706,
both M3. Fig. 20d. Cf. Uintatherium sp. V571 1, right M3.

Gobiatherium] monolobotum, new species
(Fig. 24)

Holotype. — V5 7 1 3 , 1-6: Ij (I2?), two I3 (partly broken); left P2 and P3, M3 (pos-
terior part) (field no. 1 P25H8).

Referred sped mens. — V 5 7 1 3 , 7-8: two incisors (field no. 77036-2).

Localities.— Ulan Bulak; Huhe Bulak.

Diagnosis.— Crown of incisor not notched.

Description.— I,: small; external wall smooth; straight crest from top to base on internal wall; top
of crown pointed; no notch along crown edge. I3: robust in size; in external view, crown caret-shaped;
a relatively robust and straight crest runs from top to base on internal wall (another crest may parallel
this crest, but this is not clear as tooth is broken). P2: metaconid and metastylid distinguishable;
relatively large. P3: protolophid formed by protoconid, metaconid, and metastylid; direction of pro-
toconid oblique to that of lower jaw; hypoconid prominent, extends forward forming oblique crest;
oblique crest not joined to protolophid; posterior cingulum developed. M3: protoconid robust; talonid
long; hypoconulid very strong, but apex not pointed.

Discussion.— P2 and P3 of this species are similar to those of Gobiatherium
mirificum, but the incisors are quite different: the crown edge in G. mirificum is
notched and the internal wall lacks a straight crest from the top to the base. M3
of G. mirificum also is larger than that of this species.

1987

Qi - Eocene Mongolian Arshanto Mammals

33

Fig. 21 . — Gobiatherium mirificum. V5704, lower jaw with P2-M2, crown and external views.

Measurements.— L/W: Il5 — / 1 0.5? mm; I3, 27.0/15.0 mm; P2, 20.57/14.0? mm; P3, 22.0/16.5 mm;
M3, -/23.0 mm.

cf. Uintatherium sp.

(Fig. 2Gd)

Referred specimen.— V 57 11 , right M3 (field no. 1P5H8).

Locality.— Ulan Bulak.

Discussion.— In crown view M3 is apparently narrower than M3 of Gobiather-
ium, its paraconid is more prominent and its entoconid somewhat swollen.

Measurements. — L: 35.7 mm; W: 25 mm.

Order Perissodactyla Owen, 1848
Suborder Tapiroidea Gill, 1872
Family Isectolophidae Peterson, 1919
Homogalax Flay, 1899
Homogalax reliquius, new species
(Fig. 25)

Holotype. — V 5 7 4 8 , right M3 (field no. 77036-2; L, 14.9 mm; W, 7.2 mm).
Referred specimen.— V 57 47 , left M3 (field no. 77034; L, 15.3 mm; W, 6.8 mm).
Locality.— Huhe Bulak.

Diagnosis.— M3 with single robust hypoconulid; metalophid well developed;
trigonid relatively long.

Description.— V 57 48 (M3), crown wide anteriorly, narrow posteriorly, relatively high; posterior part
of robust hypoconulid slightly turned toward lingual side; metalophid relatively well developed; tri-

34

Annals of Carnegie Museum

vol. 56

a 3 cm

Fig. 22 .—Gobiatherium mirificum. V5710, juvenile, lower jaw and cheek teeth, crown and external

views.

gonid relatively long; weak development of cingulum at base of crown on labial side and behind
hypoconulid.

The hypoconulid on V5747 (M3) consists of two cuspules, of which the labial one is larger than the
lingual; crown relatively high; labial cingulum at base of crown weak.

Discussion. — These M3s are unique among all Asiatic fossil tapirs because they
have robust hypoconulids. The North American fossil tapirs Homogalax and
Isectolophus bear robust hypoconulids. The former is early Eocene (Wasatchian),
and the latter is middle and late Eocene (Bridgerian and Uintan). Homogalax is
smaller than Isectolophus. On M3, the hypoconulid of H. protapirinus is smaller
than that of I. latidens, but the hypoconulid of I. annectens is still larger. The
Inner Mongolian specimen is similar to Homogalax. The hypoconulid of V5747
(M3) consists of two cuspules, one labial, one lingual, whereas North American
Homogalax may have an M3 with two cuspules, one anterior, and one posterior.

The metalophid of the Inner Mongolian specimen is separate from the proto-
lophid, while that of Homogalax and Isectolophus is still connected with the
protolophid. This indicates that the Inner Mongolian specimen was more bilo-
phodont. Also, the Inner Mongolian specimen resembles North American I. la-
tidens (such as AMNH 10640) in having a longer trigonid, but the former is higher
crowned.

Family Helaletidae Osborn, 1892
Heptodon Cope, 1882c
Heptodon minimus , new species
(Fig. 26; Table 7)

Holotype. — VS 7 3 2, associated lower jaws, with left Pi-M3 and right P2-M2 (field
no. 77036-2).

Referred specimens. — V5733: left lower jaw, with P2-M3 (77036-2); V5731, left
M3 (77027).

Locality. — Huhe Bulak.

Table 6.— Measurements (mm) of Gobiatherium mirificum.

1987

Qi— Eocene Mongolian Arshanto Mammals

35

VO

VO

<N

© ©
© to

©

rvi

rv|

© © ©

p-' <n d

— < rv| cm

tN © in >n

Os O V'] ©

— < <N

©

d

<N

<N

d

to

in

r-~"

to

© © © © ©
in d in d ni
fS (N n n n

o © © © © ©
d — ' to © in —i
(N <N <N CO CO Os

q © © © © ©

<N o 6 i=i to Os d

*-< *-< fNfSMm

© © © © © ©

© to to oo <-h’ On

nj <n MMnn

O CO o so

fn<noooo^
n n n <n n

o O © fN r- o

m ^ -st os oo

NMMfNfnoO

bosses

to rj vo ©

d os o’ d

— < — < tN rsi to

in m oo (N

-st- i n — • as
(NMfOfO

T3

6!

43 o

M

plh <nPh

36

Annals of Carnegie Museum

vol. 56

Fig. 23 .—Gobiatheriuml major new species. Holotype, V5712, right dP3, M,, crown view.

Fig. 24. — Gobiatheriuml monolobotum new species. Holotype, V5713 (in part), incisor, internal and
external views.

Diagnosis. — Main cusp on P2 strong; P4 entoconid present; hypoconulid of M3
has two crests extending forward and connecting with hypolophid.

Description.— P, (V5732): protoconid high; posterior crest somewhat swollen; roots fused. V5733,
P2: paralophid short; position of protoconid forward; metaconid not swollen; ectolophid relatively
long, connecting with protolophid; entocristid distinct, also connecting with protolophid; no hypo-
lophid; external cingulum extremely weak; no internal cingulum. P3: protoconid-paralophid-metaconid
forming a semi-circle; ectolophid well developed and more lingual; entocristid connecting with meta-
conid; hypoconid small, not connecting with entoconid; external cingulum present; no internal cin-
gulum. P4: metaconid not convex anteriorly, small notch between metaconid and protoconid; ecto-
lophid developed; entocristid distinct; hypoconid small, not connecting with entocristid; external
cingulum extremely weak; no internal cingulum. M,: paralophid distinct; metalophid short; hypo-
conulid triangular; external cingulum very weak; no internal cingulum. M2: triangular hypoconulid
more prominent, connecting upward to top of hypolophid. M3: paralophid hooks ventrally; metaconid
convex anteriorly; ectolophid relatively short; two crests extend from hypoconulid, a labial one an-
teroventrally, and a lingual one anterodorsally; a weak cristid crosses the hypoconulid valley antero-
posteriorly.

Discussion. — The Inner Mongolian specimen (V5732) is quite different from
North American earlier tapirs such as Homogalax and Isectolophus, being smaller,
and bearing smaller hypoconulids.

Several North American tapirs such as Helaletes, Dilophodon, and Colodon
have no P1? nor does the type specimen of Heptodon calciculus (AMNH 4858).

1987

Qi Eocene Mongolian Arshanto Mammals

37

i cm

Fig. 25.— Homogalax reliquius new species, (a) holotype, V5748, right M3; (b) V5747, left M3.

However, Heptodon ventorum (AMNH 294) maintained Pj. Thus species of Hep-
todon may show an intermediate condition between Homogalax and Isectolophus
(with and Helaletes, Dilophodon, and Colodon (no Pj).

In addition to dental formula, there are two other similarities between the Inner
Mongolian specimens and Heptodon calciculus : 1) the metalophid is well devel-
oped (more so in the Inner Mongolian specimens); and 2) M3 is small. However,
M2 and M3 of H. calciculus have weak external cingulids, absent in the Inner
Mongolian species. The ratio of the length of the premolar series to molar series
in Heptodon varies from 0.62 in H. calciculus (no Pl5 type specimen, AMNH
4858) to 0.78 in Heptodon ventorum (AMNH 294, with P^, which is the same
as in the Inner Mongolian species. Differences between H. ventorum and H.
minimus are: 1) the Inner Mongolian species is smaller; and 2) P4 has an entoconid,
not present in H. ventorum.

Helaletes Marsh, 1872
Helaletes medius , new species
(Fig. 27; Table 8)

Holotype — V 5 7 2 9 : broken lower jaws, associated, with left M3, right P4, MY
and M2 (broken), and M3 (field no. 1P5H8).

Referred specimens.— V 57 30, broken left lower jaw with M2 and M3 (1P5H8).

Table 1 .—Measurements (mm) o/ Heptodon minimus and H. calciculus.

Heptodon minimus

V5732 (holotype) Heptodon calciculus

Left Right V5731 V5733 AMNH 4858

L

w

L

w

L W

L

w

L

w

p,

5.0

2.4

p2

5.7

3.7

5.7

3.7

6.2

4.0

5. 7-6. 8

3. 7-5. 3

p3

6.8

4.9

7.1

4.8

7.5

5.0

7. 0-8. 6

4.5-6.0

p4

7.5

5.8

7.5

5.7

7.9

5.9

7. 2-9. 2

5. 2-7.0

M,

8.6

6.4

8.6

6.4

9.7

6.3

8.9-10.7

6. 2-7. 8

m2

10.6

7.5

10.5

7.4

9.9

6.8

10.0-12.4

7. 2-8. 2

m3

12.4

6.9

11.5 5.7

12.6

6.6

13.3-15.7

7. 3-8. 8

Pm 24.4

P ,_3 31.8 32.1

P,-M3 56.7

38

Annals of Carnegie Museum

vol. 56

Fig. 26.— Heptodon minimus new species, (a) holotype, V5732 (in part), lower jaw with P,-M3, crown
and lateral views; (b) V5731, M3.

Locality. — Ulan Bulak.

Diagnosis. — Metalophids of M, and M3 relatively well developed, M3 short
anteroposteriorly, its hypoconulid very small.

Description. — P4: paralophid relatively lingual and oriented anteroposteriorly; paraconid not swollen;
metalophid relatively well developed; top of hypolophid located almost at level of talonid basin; no
entoconid; no internal cingulum. M,: hypoconid relatively well developed. M2: paralophid very short;
protolophid and hypolophid present; two segments of cingulum occurring on the external wall; no
internal cingulum; hypoconulid flexed dorsally. M3: paralophid relatively long; paraconid somewhat
prominent; protolophid parallels hypolophid; metalophid extending forward and connecting with
protolophid; hypoconulid small, but very distinct; buccal and lingual hypoconulid crests extend to
base of hypolophid.

Discussion. — The position of the paralophid of P4 of this species, as in H.
mongoliensis, is closer to the lingual margin of the crown than to the labial margin
as in Schlosseria magister and Lophialetes expeditus. H. medius has a very small
entoconid, so its talonid is nearly flat. On P4 in Schlosseria magister and Lophi-
aletes expeditus the hypolophid is relatively well developed, but separated by a
notch from the entoconid. An incomplete hypolophid occurs also in Helaletes
mongoliensis, but the entoconid is stronger. In addition, H. medius is smaller.
These features indicate that P4 of this species is more primitive in dental structure.

The hypoconulid of M3 of this species is very small, and the style of connection
between the two crests extending from hypoconulid and hypolophid is similar to
that in Helaletes mongoliensis. The hypoconulid of H. mongoliensis is larger, and
the metalophids on the lower molars are not developed.

This species is similar to North American Helaletes nanus in size, but the
trigonids of P4-M3 of the Inner Mongolian species are longer.

1987

Qi— Eocene Mongolian Arshanto Mammals

39

Table 8. —Measurements (mm) o/Helaletes medius, H. mongoliensis, and H. nanus.

Helaletes medius

Helaletes

mongoliensis

AMNH 20155

Helaletes nanus
(Radinsky, 1963)

V592

V5730

L

w

L

w

L

w

L

w

P3

6.5?

9.75

7.30

6. 8=8. 7

4. 5-5. 9

P4

8.9

6.3

10.00

8.77

7. 5=8. 8

5. 1-7.0

M,

9.1?

6.8

12.28

9.05

8.4-10.8

5. 3-7. 3

m2

11.8

8.0

11.4

8.5

14.55

10.35

9.9-11.8

6.3 7. 9

m3

14.5

8.2

15.0

8.1

17.50

10.30

11.5-11.1

6. 5-8.3

M,_ 3

36.4

35.4?

Helaletes fissus (Matthew and Granger, 1925)

(Fig. 28)

Holotype. — AMNH 20161, maxillary fragment with P2-4.

Locality.— Near Camp Margetts.

Diagnosis. — “Slightly smaller than Helaletes mongoliensis, with P2-4 relatively
shorter and wider, and hypocones better separated from protocones than in that
species. P2-4 metalophs not so prominent as protolophs, and directed toward
protocones” (Radinsky, 1965, p. 230).

Discussion. — Radinsky (1965) referred Desmatotherium fissum to Helaletes fis-
sus, a new combination, and, although the geological position of this species was
not clear at that time, Radinsky suggested that H. fissus and H. mongoliensis may
occur in different horizons. His suggestion has proven to be correct.

It is not possible to make comparisons between H. fissus and H. medius because
the former has only P2-4 and latter only lower cheek teeth. However, as Radinsky
(1965, p. 231) pointed out, ‘The upper premolars of Helaletes fissus are more
advanced than those of any other species of Helaletes . . . .” The relatively prim-
itive lower premolars of H. fissus suggest that it and H. medius are not the same
species.

Helaletes fissusl (Radinsky, 1965)

Discussion. —The features of AMNH 8 1 802 (Q-M3) studied by Radinsky (1 965,
p. 231) are advanced for Helaletes although “It is slightly smaller than Helaletes
mongoliensis and agrees in size with the type of H. fissus .” The P3_4 of Helaletes
fissus ? have small entoconids, but nothing on P4 is suggestive of Helaletes medius.

Hyrachyus Leidy, 1871
Hyrachyus neimongoliensis, new species

(Fig. 29)

Holotype. — V 5721: an incomplete skull, with left P2, P4, M1 (partly broken),
and M2-3 (field no. 77036-2).

Locality. — Huhe Bulak.

Diagnosis. — Frontal bone and nasal bone flat and straight; hypsodont; parastyles
on molars very strong; paraloph almost parallel to metaloph; premolars non-
molariform.

Description. — Skull: height low; posterior part of premaxilla narrow, body ascending abruptly, not
gently; nasal bone flat and straight; posterior edge of nasal notch just anterior to P2; position of orbit
low; frontal bone wide and flat; maxilla very long; zygomatic process of temporal bone and zygomatic
process of molar bone join in mid part of zygomatic arch; zygomatic arch very long; suture between

40

Annals of Carnegie Museum

vol. 56

0 1 C M

1 1

Fig. 27 .—Helaletes medius new species. Holotype, V5729, lower jaws with left M3 and right P4 and M3.
Fig. 28.— Helaletes fissus. AMNH 20161, P24 (from Matthew and Granger, 1925c).

parietal and interparietal present; palatine process of maxilla narrow; median palatine suture clear;
horizontal part of palatine narrow; basilar tubercles evident, separating basilar part of occipital and
body of sphenoid; alar canal distinct; paramastoid process relatively large; mastoid process and occipital
condyle relatively small; foramen magnum relatively large.

Upper cheek teeth: P3: paraloph connecting with ectoloph and protocone; attachment of paraloph
and ectoloph lower than apex of protocone; hypocone distinct; metaloph connecting with ectoloph,
and attachment of the two lophs higher than that of protoloph and ectoloph; anterior cingulum stronger
than posterior cingulum; no external cingulum; paracone high, forming a rib-like and relatively robust
crest on external wall; ectoloph straight. P4: larger than P3; protocone robust, forming paraloph, which
connects to ectoloph; paracone more prominent than metacone; two rib-like crests formed by these
cusps on external wall (that is, anterior rib and posterior rib); anterior rib stronger than posterior rib;
metaloph short, not connecting with paraloph; ectoloph straight; anterior cingulum distinct; no external
cingulum. M1: (anterior part broken) middle part of paraloph swollen, may be result of wear of massive
antecrochet; metacone long; angle between metacone and metaloph about 90°; external wall of meta-
cone flat and straight; anterior cingulum more distinct than posterior one. M2: largest of cheek teeth;
paraloph robust; metaloph short; posterior part of ectoloph straight; paracone high; parastyle relatively
robust; anterior cingulum clearer than posterior cingulum; no external cingulum; the angle between
metacone and metaloph about 60°. M3: distance between paraloph and metaloph large, these two lophs
almost parallel; paracone prominent; parastyle relatively strong; metacone (crest) short; angle between
metacone and metaloph about 90°; anterior cingulum wider than posterior cingulum.

Discussion.— Wood (1934, p. 189) pointed out the diagnostic features of Hy-
r achy us: “Hornless; protoloph of upper molars much more prominent than the
metaloph; no tendency for the metalophs of the upper premolars to touch the
crista; attachments of upper premolar metalophs to ectolophs usually higher than

1987

Qi— Eocene Mongolian Arshanto Mammals

41

Fig. 29.— Hyrachyus neimongoliensis new species. Hole* type, ¥5721, skull, lateral, palatal, and dorsal
views.

corresponding attachments of the protolophs; upper cheek teeth subrhinocerine
rather than rhinocerine, except after extreme wear.”

Radinsky (1967a, p. 21) also pointed out that Hyrachyus is: “A conservative
group preserving into the Middle Eocene primitive dental and skeletal features

42

Annals of Carnegie Museum

vol. 56

Table 9.— Measurements (mm) o/Hyrachyus crista.

V5722

V5723

L

w

L

w

P4

13.5

17.5

14.5

18.3

M1

17.5

19.0

19.2

20.0

M2

21.4

21.4

23.5

22.6

M3

18.4

19.0

M>-3

55.0

seen in the earliest Heptodon. Incisors spatulate. Postcanine diastema present. P{
retained; premolars non-molariform to submolariform. No tendency toward bi-
lophodonty. M3 without a hypoconulid. No nasal incision enlargement.”

Unquestionably, the Inner Mongolian species is very close to the North Amer-
ican Hyrachyus eximius (=H. princeps, AMNH 12364). Both species have some
similar characters in both skull and cheek teeth: 1) P4: attachments of metaloph
to ectoloph higher than corresponding attachment of protoloph; 2) M3: parastyles
strong, and paraloph is close to and almost parallels metaloph. The species have
some differences: 1) the P4 of the Inner Mongolian species has no crochet, while
the North American P4 does; 2) widths of the molars of the North American
specimen are greater than lengths, while the Inner Mongolian specimen has the
opposite condition.

M3 of Hyrachyus metalophus (Zhou and Qi, 1982) is somewhat smaller than
this species, but has a very prominent external rib on the ectoloph. This is a
primitive character.

The skull of this species bears some resemblance to that of Forstercooperia
huhebulakensis : skull dimension; skull low and flat; zygomata slender; and the
position of the bulla.

Measurements.— Skull: length between premaxilla and condyles, 365mm; width, across zygomatic
arch, 158 mm; depth of skull above M3, 81 mm; width of skull (postorbital constriction), 52 mm;
width of condyles, 51 mm. Upper cheek teeth (L/W): P2, 12.3/12.3 mm; P4, 16.1/21.3 mm; M1, 26
(ca.)/2 1 mm (ca.); M2, 30.5/27.3 mm; M3, 28.4/26.2 mm; (L): M1-3, 77.0 mm (ca.).

Hyrachyus crista , new species
(Fig. 30; Table 9)

Holotype. — V5 722, fragmentary maxilla with P4-M3 (field no. 77031).

Referred specimen.— V 57 23, fragmentary maxilla with left and right P4-M2
(right P3 broken) (77031).

Locality.— Bayan Ulan.

Diagnosis. — Attachment of P4 metaloph to ectoloph higher than corresponding
attachment of protoloph; internal cingulum distinct at the base of protocone;
crochet on P4-M2 very weak, but distinct on M3.

Description.— P3 (V5722, same below): only three alveoli remaining. P4: attachment of metaloph to
ectoloph higher than corresponding attachment of protoloph; paracone and metacone forming two
clearly rib-like crests (anterior rib and posterior rib); parastyle prominent; internal cingulum distinct
at the base of protocone; crochet weak. On P4 of V5723, no crochet, internal cingulum forming indistinct
semi-circle. M1: parastyle prominent; paracone high, forming anterior rib; metacone relatively long,
posterior rib very weak; ectoloph connected to metaloph, but attachment covered by cement; angle
between metacone (crest) and metaloph smaller than 90°; metaloph higher than paraloph; external
cingulum weak and short at the base of metacone; crochet very weak. On M1 of V5723, posterior rib
very weak; internal side of parastyle notched; no crochet. M2: similar to M1, but parastyle stronger,
crochet clearer, metacone longer. On M2 parastyle on V5723 notched. M3: parastyle notched, not as

1987

Qi— Eocene Mongolian Arshanto Mammals

43

Fig. 30. —Hyrachyus crista new species, (a) Holotype, ¥5722, P4-M3; (b) ¥5723, left and right P4-M2.
Fig. 31.— Hyrachyus sp. cf. H. eximius. ¥5728, left M3.

robust as on M2; crochet distinct; metacone straight; paraloph not close to metaloph but parallel to
it; paracone high, forming prominent anterior rib.

Discussion . — Radinsky (1 967a) referred one genus, Hyrachyus , to the Subfamily
Hyrachyinae, and three species: H. modestus from North America, Europe, and
China; H. eximius from North America and possibly from China; and H. minimus
from North America.

44

Annals of Carnegie Museum

vol. 56

Radinsky’s (1967a, p. 22) diagnoses of these species are as follows:

1) H. modestus : “A small to medium-sized Hyrachyus species: mean length of
M1-3 from about 45 to 50 mm. Premolars non-molariform. Upper molars usually
lack lingual and labial cingula.”

2) H. eximius : “Size large: mean length of M1-3, 64 mm. Premolars non-mo-
lari form. Upper molars usually lack lingual and labial cingula.”

3) H. minimus : “A medium-sized Hyrachyus species: mean length of M1-3,
about 50 mm. Premolars submolariform. Upper molars usually with labial and
lingual cingula. M3 metacone situated more labially than in other Hyrachyus
species. In these features H. minimus is intermediate between other Hyrachyus
species and Chasmotherium cartieri. ”

P4 of the Inner Mongolian species is similar to Hyrachyus modestus, especially
between V5722 and AMNH 12667. Both P4s have very prominent anterior and
posterior ribs, and very prominent anterior and posterior cingula. Some different
characters are apparent also on various P4s of H. modestus : 1) V5722 and AMNH
80183 have crochets, while V5723 and AMNH 12667 have no crochets; 2) V5723
has a semi-circle-like internal cingulum, lacking on AMNH 12667. M1-3 of the
Inner Mongolian species are different from those of Hyrachyus modestus in the
following points: 1 ) parastyle relatively small; 2) higher crown; 3) metacone elon-
gated.

Apparently the genus Hyrachyus not only appeared in Asia, but its later mem-
bers were broadly distributed in east Asia. Hyrachyus modestus appears closely
related to H. crista.

Hyrachyus sp. cf. H. eximius
(Fig. 31)

Referred specimen.— V 57 28: left M3 (field no. 77039).

Locality.— Huhe Bulak.

Discussion. —This tooth is similar to M3 of Hyrachyus eximius (AMNH 12362)
found in Bridger C beds, Wyoming. The main similarities are: 1) paraloph parallel
to metaloph; 2) paraloph not close to metaloph; 3) degree of hypsodonty; 4) length/
width ratio. The metacone of the Inner Mongolian specimen is weaker.

Compared to M3 of Colodon inceptus (found in Inner Mongolia), the M3 referred
here is wider, and the attachment of the paraloph to the ectoloph is closer to that
of the metaloph to the ectoloph. Both M3s have short metacones.

Measurements. -{V 5128): M3 (L/W), 14.4/22.0 mm.

Hyrachyus sp.

(Fig. 32)

Referred specimens.— V 57 24, left M2-3 (field no. 77028H2); V5725, right P4
(77036-2); V5726, right P4 (77036-2); V5775, left M3 (77039); V5776, right M3
(77039).

Localities. — Daatein Obo, Huhe Bulak.

Discussion. — These specimens, collected from several sites, are too incomplete
to be assigned with certainty. V5724 was found in Daatein Obo Cliff along with
other fossil mammals, Microtitanl elongatus, Forstercooperia grandis and Asi-
omys medius. Characters of the two molars are as follows: M2: parastyle, a slender,
prominent pillar; paracone high, forming a clear anterior rib; metacone long,

1987

Qi— Eocene Mongolian Arshanto Mammals

45

without a clear posterior rib; cingulum around the base of the crown (except
anterior edge); no crochet. M3: parastyle as on M2 and separate from protocone;
attachment of paraloph to ectoloph close to that of metaloph; metacone relatively
long; anterior cingulum clear. M2-3 are similar to those of H. crista but they are
smaller, have weaker parastyles that are slender, pillar-like and separate from the
protocone, and lack a crochet on M3. Among these characters, the slender pillar-
like parastyles are rarely seen in tapiroids.

On P4 (V5725 from the Irdin Manha beds) the paracone is very close to the
metacone; the crochet is very well developed; the paraloph is not connected to
the ectoloph and is relatively straight (the paraloph of H. crista is somewhat
curved and hook-shaped); its metaloph is swollen as in H. crista.

P4 (V5726 from the Arshanto beds), in crown view, is nearly square, the crochet
is strong, the metacone relatively long, and the paraloph is longer than and con-
nected to the metaloph. This tooth differs from P4 of Schlosseria magister in its
larger size; nearly square occlusal outline, strong crochet, and long metacone
(crest). The length/width ratio of V5726 (P4) differs from that of North American
Hyrachyus modestus. Also, the P4 metaloph of H. modestus is isolated, that is,
not connected to the paraloph. The attachment of paraloph to ectoloph in V5726
is higher than the corresponding attachment of metaloph, in contrast to the usual
condition.

V5775 (M3) and V5776 (M3) were found in Huhe Bulak area, and are Hyrachyus -
like in that the paraloph of V5775 parallels the metaloph.

Measurements.- L/W (mm): M2 (V5724), 17.0/18.3; M3 (V5725), 14.5/17.4; P4 (V5726), 13.2/
18.7; M3 (V5775), 20.0/22.1; M3 (V5776), 23.0.

Table 10.— Measurements (mm) o/Schlosseria magister.

46

Annals of Carnegie Museum

vol. 56

o no

■'t

p p

od on

o

r~-

>

i^ChO'?

o — ' cl —I

On m On On O
(d Os — < O ci
—i m

p On p
r- ‘ od r -

qq<N

© d <d

p p
rd

£

Os pOOp
i d no od od rd

p p

cl —

r~-

m

>

p p p O

rd oo on d »d

£

i/l o
© ci

c*

p

cl

£

p p
rd rd

O

p pop p

id d d d d

p O p oo
rd oood

rl

■'T

r~-

>

NfON SO p
id id so od

ii

p p H 00 p p

od © — < Tf rd ci
— < — • —i fN so

£

p so p p
O — ' cl — I

£

p p p pppo
cl rd *d ddh rd

p so p «o cl
od d d d d

p go

so t'~

ocoO^NOin
rd © *d <d cl ci

— l — l — < NO

NOOOON
'E rd On

Ml no
NO O I—1 rn

O CI ci Cl

2

s

<

in c- m

oop-H pqpoo
id no od od o do

~ IN

E o

m in in

p p NO ON p

no no rd rd rd

SO ONNOp p

od od o — « >d

a.a.s.%s.ssss

s

- rseC ^ *_r ^ I L

PTPi

M,_3 36.5 35.8 34.9 35.2

Diastema 14.8 14.0 (ca.)

Depth of mandible

below M, 17.8 19.0 (ca.) 17.2

1987

Qi Eocene Mongolian Arshanto Mammals

47

1 C M

j

Fig. 33.— Schlosseria magister. Holotype, AMNH 20241, P'-M1 (from Matthew and Granger, 1926).

Family Lophialetidae Radinsky, 1965

Schlosseria Matthew and Granger, 1926
Schlosseria magister Matthew and Granger, 1926
(Figs. 33, 34; Table 10)

Holotype. — AMNH 20241, upper and lower jaws, fore and hind feet.

Referred specimens. — V5 7 34, associated maxillae with left and right P4-M3,
lower jaw with P2-M3 (with Pt alveoli); V5735, maxilla with P3-M2; Y5736,
maxilla with M2“3; V5737, lower jaw with Mt_3; V5738, maxilla with dP4-M2
(field no. 77039); V5739, right M3; V5740, maxilla with P2-M3; V5741, lower
jaw with P3“M3 (77036-2); V5742, 1-10, cheek teeth; V5743, lower jaw with left
I3, P2-M3 and right C, P2~M;; Y5744, dP3_4; V5745, lower jaw with P, , P3, and
P4 (77034).

Localities. — Irdin Manha area and Huhe Bulak.

Discussion. — Of the known fossil mammals in the Arshanto beds, Schlosseria
magister is most abundant. It occurs throughout the Arshanto beds in the Huhe
Bulak and Ulan Bulak areas. After studying Morris’ field notes for 1 923 and 1 925,
Radinsky (1964, p. 3) determined the precise location of the type locality: “the
Schlosseria quarry ... for the distance from it to Arshanto well is recorded as 3
miles . . . .”

48

Annals of Carnegie Museum

vol. 56

Radinsky (1965, p. 197) found that, “In the Irdin Manha area the beds con-
taining Schlosseria underlie those with Lophialetes, but in the Camp Margetts
area both genera are recorded as coming from the same horizon.” Our discoveries
have confirmed this point at least in the Huhe Bulak (that is, Camp Margetts)
area. In Radinsky’s study of the early Tertiary Tapiroidea of Asia, he listed the
differences between Schlosseria and Lophialetes, as follows: species of Schlosseria
are “Medium sized lophialetids with premolar series long relative to molar series.
Premolar and molar metacones short and slightly convex; lophs and lophids
relatively low and obtuse. P1 with narrow posterolingual shelf. P2-4 with protoloph-
metaloph loop. M2-3 relatively short and wide. P2_3 relatively long and narrow,
with long trigonids. Nasal incision not so enlarged as that in Lophialetes. Manus
tridactyl” (Radinsky, 1965, pp. 199-200).

Although the new specimens are very similar to Schlosseria magister and differ
from Lophialetes expeditus, the two genera share some characters and occur in
the same beds. Therefore, the possibility should not be excluded that these are
sexual differences. This needs more detailed study.

Ranga Rao (1972) described a new tapiroid, Kalakotia, and made detailed
comparisons between Kalakotia and Schlosseria (or Lophialetes). Although there
exist several apparent differences between these genera, such as Kalakotia being
smaller in size and more primitive in tooth pattern, Rao referred this genus to
the Lophialetidae because, “it is clear that Kalakotia bears more similarities in
dental structure to the Mongolian forms: Lophialetes and Schlosseria, than to any
other known tapiroid.”

Schlosseria sp. cf. S. magister
(Fig. 35)

Referred specimen. — V 57 46, right M3 (field no. 77034).

Locality. — Huhe Bulak.

Discussion.— On this M3, two small crests extend from the hypoconulid and
attach to hypolophid in a manner similar to that in S. magister. At the base of
hypoconulid, labially, there is a cingulum that is not present in other species of
Schlosseria, but the tooth is similar otherwise to Radinsky’s (1965) Schlosseria
cf. Schlosseria magister.

Measurements.— (V 5746) L/W: 13.6/6.7 mm.

Lophialetes

Lophialetes expeditus Matthew and Granger, 1925
(Figs. 36, 37; Table 11)

Holotype. — AMNH 19163, maxilla with P*-M3.

Referred specimens.— V 5749, left M1; V5750, right M3; V5751, maxilla with
dP2-3 (field no. 77027); V5752, left M2; V5753, maxilla with dP'-M1 (field no.
77036-2); V5754, maxilla with left dP1 (broken)-M1 and right dP3-4 (broken);
V5755, jaw with left dP2-M1 and dP3-Mj; V5756, lower jaw with dP3 and M3
(field no. 77039).

Localities.— Camp Margetts; Huhe Bulak.

Description.— dP1 (V5754, largely broken): only the ridged internal edge can be seen. dP2 (V5754):
paraloph oblique, not parallel to metaloph; hypocone strong; metaloph straight; paracone high; crochet
relatively clear; cingulum extremely weak. dP3: paraloph straight and long; the position of metaloph
higher than paraloph; paracone and metacone very prominent; paracone rib and metacone rib clear;
parastyle prominent; crochet relatively distinct; metastyle slightly prominent. dP4: paraloph long; a
small crest between parastyle and paracone; paracone rib stronger than metacone rib; on left dP4 a

1987

Qi— Eocene Mongolian Arshanto Mammals

49

Fig. 34 .—Schlosseria magister. V5734, palate with right and left P4-M3 and lower jaw with C, P2-M3.
Fig. 35.— Schlosseria sp. cf. S. magister. V5746, right M3.

secondary cuspule occurs at the base of hypocone. dP! (V5755), single-rooted. dP2: paraconid small
but distinct; protolophid straight; protoconid highest cusp; metaconid close to paraconid; metalophid
distinct. dP3: paraconid very prominent; paraconid not close to metaconid; protolophid relatively
long; entoconid prominent, connects with hypoconid forming hypolophid; a small triangle-like process
at position of hypoconulid. dP4: in crown view wider anteriorly than posteriorly; metaconid strong
and convex anteriorly; distance from protoconid to anterior edge of tooth shorter than on dP3; triangle-
like process at hypoconulid position more prominent.

Discussion. — The differences between Schlosseria and Lophialetes were men-
tioned above. These new specimens are unquestionably Lophialetes. The milk
teeth reported by Radinsky (1965) are also quite similar to the new specimens.

Breviodon Radinsky, 1965
Breviodon minutus (Matthew and Granger, 1925)

(Figs. 38, 39, 40; Table 12)

Breviodon acares Radinsky, 1965, p. 203.

Holotype. — AMNH 20139, upper molar.

Referred specimens. — AMNH 26113 (holotype of Breviodon acares ), left man-
dible with I3 (reversed), Cl5 P3 alveolus, and P4-M3; V5757, maxilla with dP3-4
and M1; V5758, dP3~4 (above field no. 77027); V5759, low jaw with M,_2 (broken)
(1P25H8); V5760, low jaw with dP4 (broken) and (1P25H8).

Localities. — Irdin Manha area (type locality of B. minutus ), Ulan Bulak, Huhe
Bulak, North Mesa, Ulan Shireh (type locality of B. acares).

Description. — dP3: in crown view nearly square; paraloph almost parallel to metaloph; paracone rib
and metacone rib distinct; posterior cingulum very weak; no cingulum at the base of hypocone. dP4:
similar to dP3, but paraloph more prominent; metacone flat and straight. M1: larger in size; lophs

50

Annals of Carnegie Museum

vol. 56

E

o

CM

Fig. 36 .—Lophialetes expeditus. (a) ¥5754, maxilla with dP’-M1; (b) ¥5755, jaw with left dP2-]

1987

Qi— Eocene Mongolian Arshanto Mammals

51

o

q

cd

q

*n

q

in

q in — *

OS o —

m >n 't

os os o

q q q

<n n in

fN o so m vo

^ t-* On Os Os O

rq — .

OS 00
oo os

Os oo
o 6 os

%

t'' m m Tt
^thO\6

Tf

lO

r-

m

>

C^*

qooq

so t"*’ os

q q q
os o rd

m — h

p^7 cC clT &JT «^r

,d'Ox!’OT3222'd',0’d’d,o2

52

Annals of Carnegie Museum

vol. 56

0 1CM

> »

Fig. 37 .—Lophialetes expeditus. Holotype, AMNH 19163.

slightly worn. dP4: (only posterior part preserved) the angle between ectolophid and hypolophid 90°;
the position of hypoconulid somewhat prominent. M,: anterior width almost the same as posterior
width; metaconid distinct and convex anteriorly; a triangle-like process at the hypoconulid position.
M2: (only posterior part preserved) ectolophid and hypolophid developed; slight prominence at hy-
poconulid position.

Fig. 38 .-Breviodon minutus. (a) V5757, dP^-M1; (b) V5758, dP3-4; (c) V5760, dP4-M,.

1987

Qi— Eocene Mongolian Arshanto Mammals

53

0 1 CM

1 i

Figs. 39, 40 .—Breviodon minutus. 39. Holotype, AMNH 20139 (from Matthew and Granger, 1925c,
Fig. 8). 40. AMNH 26113, holotype of B. acares, left lower jaw with I3 (reversed), C1? P3 alveolus and
P4-M3.

Discussion. —Breviodon was erected by Radinsky (1965, p. 203) and diagnosed
as follows: “Very small lophialetids, with premolar series extremely short relative
to molar series; P{_2 lost. Symphysis broad, postcanine diastema long. Lower molar
cusp patterns like those in Lophialetes He named a new species, Breviodon
acares , and referred Lophialetes minutus with question to this genus as Breviodon?

Table 12.— Measurements (mm) o/Breviodon minutus.

V5757

V5758

V5759

V5760

L

w

L W

L

w

L W

dP3

5.8

6.3

5.4 5.5

dP4

6.5

7.3

M1

7.8

8.6

dP4

3.9

M,

7.9?

5.0?

m2

8.8

6.3

54

Annals of Carnegie Museum

vol. 56

minutus. According to the material (associated upper and lower cheek teeth) found
in Mongolia, Reshetov (1975) considered B. acares to be a synonym of Breviodonl
minutus, a conclusion followed here.

V5757 (dP3-4 and M1) fit with V5759 (ML_2), so the two specimens are apparently
associated. M1 of V5757 fits with M1 of Breviodon minutus (no. 3107-32) found
in Mongolia, whereas the M{ of V5759 is similar to that of “Breviodon acares ”
(AMNH 26 1 1 3). This further confirms the synonymy of B . minutus and B. acares.
On the anterior edge of dP3 of Y5757 there seems to be a wear facet, suggesting
the presence of dP2.

Huang and Qi (1982) described a new species of Breviodon, B. lumeiyiensis,
the teeth of which are similar to those of B. minutus, but smaller in size, the cheek
teeth being twenty percent narrower than those of B. minutus.

Family Deperetellidae Radinsky, 1965
Teleolophus Matthew and Granger, 1925
Teleolophus primarius , new species
(Fig. 41; Table 13)

Holotype. — V5 7 6 1 , mandible with P2_4 and M,_2 (field no. 1P5H7).

Referred specimens. — V 57 62, mandible with M2_3 (77027); V5763, 1-21: as-
sociated teeth, several carpi, and phalanges (including left P3, M1? and M2 (77036-
2); V5764, 1-2: two M, (77039); V5764, 1-2: two M, (1P5H8).

Localities. — Ulan Bulak, Huhe Bulak.

Diagnosis. — The metalophid is labial on P2, but not labial on P3 and P4; pre-
molars lacking entoconids; three pairs of incisors.

Description.— P2: metalophid near labial edge of crown; protoconid very close to metaconid; para-
conid somewhat swollen; a rib-like crest on the internal wall of hypoconid; external cingulum relatively
strong; internal cingulum incomplete. P3: protoconid and metaconid clearly separate; no hypolophid;
a rib-like crest on the internal wall of hypoconid; trigonid longer than talonid. P4: paralophid short;
trigonid relatively short; talonid longer than trigonid. M,: protolophid parallel to hypolophid; no
metalophid; anterior cingulum weak; posterior cingulum very short. M2: paralophid very wide antero-
posteriorly; anterior cingulum below the paralophid very weak; no metalophid; protolophid almost
parallel to hypolophid. M3 (V5762): similar to M2 but posterior root curved posteriorly. Radiale (carpal
bone): five facets, of which the facet that articulates with radius is the largest; two facets articulate
with intermediate carpal, the distal of them narrow and long, and the proximal, wide and short; the
other two facets are larger and articulate with the second and third carpal. Intermediate carpal bone:
a total of seven facets— one flat facet articulates with the radius, two facets articulate with radiale (the
proximal one is wide, the distal, narrow), two facets articulate with ulnar carpal(?), and the other two
facets articulate with the third and fourth carpal bones. Fourth carpal bone: roughly triangular in
proximal view; in distal view, wider anteriorly than posteriorly; seven facets for articulations— two
with intermediate and ulnar carpal bones, three with third carpal bone and another two with third
and fourth metacarpal. First phalanges: three different phalanges, under the second, third, and fourth
metacarpals, respectively. Second phalanges: the one under second metacarpus is widest. Third pha-
langes: proximal surface with two facets; dorsal surface rough.

Discussion. — These specimens were found in adjacent sites, so the known dis-
tribution of this species is restricted. The main difference between Teleolophus
and Deperetella is that the premolar series in the former is shorter than the molar
series. Although and M3 were not found on one specimen of Teleolophus, the
length of P2_4 is equivalent to M^. Therefore, the length of the P14 must be shorter
than Mj_3.

The differences between T. primarius and T. medius (found in 1975, now under
study) are as follows: 1) the three incisors are aligned anteroposteriorly in T.
primarius but form a triangle in T. medius ; the third incisor of T. medius is greatly
reduced, and the canine is more robust; 2) the symphysis is narrower in T. pri -

1987

Qi— Eocene Mongolian Arshanto Mammals

55

Fig. 41. — Teleolophus primarius new species, (a), (b) holotype, V5761, mandible with P2-M2, external
and internal views; (c), (d), (e) V5762, mandible, with M2_3, internal, crown, and external views.

56

Annals of Carnegie Museum

vol. 56

p

*C

ed

6

•c

a

a

jo

’o

JD

H

cn

cs vo vq p no
r»’ OO Os 3 cs

as p o cs vq
O vn wd d

6*'o

«S r-~

!>

&^r

Sf2

'0>
m *
a:

3-|

5 a

1 6

6 5?

- „ „:s*s

—■ — ' cs

p O Tf
©o p-’ on

I

&ji

S3 I

l ' Hb ~

1 1

ao

|!

S-9

o — >
on d

oq p

d 3

rn p

&|

g &

t3 o

§3

<u 3

l i
*0 T3
d c

D D

Under fourth carpus 21.2 12.7 14.3

1987

Qi— Eocene Mongolian Arshanto Mammals

57

marius; 3) on P2 of T. primarius the metalophid is more labial, the metalophid
is shorter, the entoconid is absent, and the protoconid and metaconid are not
completely separated; 4) P3 is shorter and wider in T. primarius and there is an
apparent notch separating the protolophid and metalophid, whereas this notch is
absent in T. medius ; a P3 entoconid is lacking in T. primarius but present in T.
medius ; 5) P4 is longer and wider in T. primarius, the notch between protolophid
and metalophid is stronger, the paralophid is shorter and the metalophid longer;
there is no entoconid, and a rib-like crest occurs on lingual wall of hypoconid; 6)
the M,_2 paralophid is stronger in T. primarius ; 7) the posterior root on M3 is
curved posteriorly in T. primarius but is straight in T. medius.

These differences suggest the following phylogenetic trends in the genus Teleo-
lophus : 1) reduction of incisors; 2) entoconid becoming prominent; 3) positions
of metalophids of P2_3 becoming more labial; and 4) P2 protoconid gradually
separating from metaconid.

Teleolophusl rectus , new species
(Fig. 42; Table 14)

Holotype. — V5 7 6 6 : two broken pieces of mandible with left P4 (anterior part
broken) and M1? and right P4 (posterior part broken), P2_3, and M2 only roots left
(field no. 1P5H8).

Locality. — Ulan Bulak.

Diagnosis.— P4: paralophid closer to lingual edge of crown, metalophid aligned
almost with the medial part of the talonid; entoconid indistinct.

Description.— P4: paralophid short, originates from protolophid rather than protoconid, so trigonid
very short; metalophid almost medial on talonid; entoconid indistinct; cingula very weak and no
internal cingulum. M,: paralophid relatively short; metalophid developed, not connected with pro-
tolophid.

Discussion. — The main characters of Teleolophus rectus are the postition of the
paralophid of P4, which is closer to the lingual edge of the crown, the metalophid
located almost medially on the talonid, and presence of a weak entoconid. In
tooth morphology, T. rectus appears intermediate between T. primarius and T.
medius.

Suborder Brontotherioidea Hay, 1 902
Family Brontotheriidae Marsh, 1873

Telmatherium Marsh, 1872
Telmatherium cristatum (Granger and Gregory, 1943)

Metatelmatherium cristatum Granger and Gregory, 1943.

Discussion. — Granger and Gregory (1943, pp. 355-356) discovered that “the
type skull of Telmatherium ultimum Osborn from the Uinta (upper Eocene) of
North America so closely resembles the type of our Metatelmatherium cristatum
in all aspects that one can barely discover specific differences between them . . . .”
Nevertheless, they erected a new genus, Metatelmatherium, from the Arshanto
fauna, believing that it represented “a distinct generic stage characteristic of the
upper Eocene of North America and Mongolia, a stage which is much larger and
has more advanced premolars than the true Telmatherium . . . .” The age of the
Arshanto beds and Irdin Manha beds is now recognized as middle Eocene as is
the North American Uintan. “ Metatelmatherium ” does not warrant recognition.

58

Annals of Carnegie Museum

vol. 56

Fig. 42. — Teleolophusl rectus new species. Holotype, V5766 (in part), right mandible with left P4-Ml5
external, internal, and crown views.

Desmatotitan Granger and Gregory, 1943
Desmatotitan sp.

(Fig. 43)

Referred specimen.— V 57 69: left M3 (field no. 77036-2).

Locality. — Huhe Bulak.

Discussion. — This tooth is wider and 10 mm longer than that of Microtitan
mongoliensis, and is closer in size to that of D. tukhumensis (Ulan Shireh). V5769
also resembles that of D. tukhumensis in having relatively distinct cingula.

Measurements.— V 5769, M3, L/W: 58/28 mm.

Protitan minor Granger and Gregory, 1943

Holotype. — AMNH 26416. A skull, lacking the distal end of the nasals.
Locality.— Camp Margetts.

1987

Qi” Eocene Mongolian Arshanto Mammals

59

Table 14 Measurements (mm) o/Teleolophus? rectus (V5766).

Left

Right

L

w

L

w

p2

9.7?

P.3

10.3?

9.9

P4

13.2

10.3

13.6

10.0

Mj

16.1

10.8

15.0?

Depth of mandible

below P4 and M,

34.2

35.3

Discussion. — Granger and Gregory (1943, pp. 358-359) erected this genus, which
includes six species. Of them, only Protitan minor was collected from “probably
top of Irdin Manha beds,” “Camp Margetts,” that is, from the top of what are
here recognized as the Arshanto beds. Other species are probably based on ma-
terials from the Irdin Manha beds. Specific characters of P. minor are as follows
(Granger and Gregory, 1943): “P1 small, compressed; P2 oblique, asymmetric,
narrow; P3 narrow; P4 not nearly so wide as in Epimanteoceras formosus ; M1 of
moderate width, M2 narrow, M3 fairly narrow; P4 with slight tetartocone ridge
but no cusp. Zygomata not wide. Basi sphenoid pits conspicuous. Lower jaw not
definitely known but the distance from the hypocone to the anterior rim of P2
(2 1 7), which should be equal to the distance from the hypoconulid to the tip of
P2, is nearly matched in a lower jaw, AMNH 26410, from the same locality and
horizon. This jaw agrees well in other measurements and dental characters with
the type skull and is therefore referred to this species. It differs from the type of
D. olseni chiefly in its smaller dimensions but tends to confirm the reference of
that form to Protitan grangeri. AMNH 20108, a maxilla, is close to P. minor in
general, but the M3 is much larger. The latter is in an early stage in lengthening,
a process which was carried to an extreme in Gnathotitan.”

Microtitan Granger and Gregory, 1943
Microtitan! elongates. , new species
(Fig. 44)

Holotype. — V 57 67 , right P3-M3 (M1 and M2 largely broken; M3 preserves only
mesostyle and metacone) (field no. 77028H2).

Locality.— Daatein Obo.

Diagnosis.— Larger than Microtitan mongoiiensis ; P3 and P4 wider than long
(although width of P3 is close to that of P4), mesostyles on P3 and P4 well developed
and forming triangular processes on the external walls of the teeth, cingula de-
veloped.

Description.— P3: protocone prominent; paracone and metacone present; parastyle fairly well de-
veloped; mesostyle high, forming a triangular process on the external wall, which connects with
prominent external cingulum; metastyle also fairly distinct; the base of the tooth rounded by a cingulum
which separates from external cingulum at the parastyle and the metastyle. P4: similar to P3, but wider
and parastyle more prominent. M1: (protocone and hypocone heavily worn) metacone well developed.
M2: apparently larger than M1, with no wear on hypocone. M3: (only ectoloph preserved) the distance
between paracone and mesostyle is smaller than that between mesostyle and metacone.

Discussion. — Microtita.nl elongates is slightly larger than M. mongoiiensis. P3
and P4 of Micmtitanl elongates are wider than long and rectangular in crown
view whereas P3 ’4 are square in M. mongoiiensis. The mesostyles on P3 ~4 of M. ?

60

Annals of Carnegie Museum

vol. 56

1 cm

45

Fig. 43 . — Desmatotitan sp. V5769, right M3.

Fig. 44.— Microtitan? elongatus new species. Holotype, V5767 (in part), P3-M2.
Fig. 45.= Microtitan sp. V5768, right M2.

1987

Qi— Eocene Mongolian Arshanto Mammals

61

elongatus and especially the triangular processes on the external wall are rarely
seen in titanotheres and are absent in M. mongoliensis. On P3 and P4 of Microtitan?
elongatus the basal cingula are more nearly continuous than in M. mongoliensis.

Microtitan ? elongatus differs from Protitan minor (AMNH 26417) as follows:
P34 are apparently smaller; widths of P3 and P4 are subequal (P4 is larger than P3
in P. minor ); P3-4 protocones are more prominent; and the paracone is almost
medial on the ectoloph whereas it is more anterior in Protitan minor.

Measurements.— L/W: P3, 19.0/23.3 mm; P4, 17.7/24.6 mm.

Microtitan sp.

(Fig. 45)

Referred specimen. — V5 7 6 8 , right M2 (field no. 77027).

Locality. — Huhe Bulak.

Discussion. — This tooth was found low in the Arshanto section along with
Microtitan elongatus.

Measurements.— V 57 68, M2, L/W: 37/29 mm.

Suborder Rhinocerotoidea Gill, 1872
Family Hyracodontidae Cope, 1879
Forstercooperia Wood, 1939
Forstercooperia huhebulakensis , new species
(Fig. 46; Table 15)

Holotype. — V5 7 70, a broken skull with P4-M3 (field no. 77036-2).

Referred specimen. — V5771, a broken mandible with P3— M3 (M3 heavily bro-
ken) (77036-2); Y5772, M2 and M3 (77036-2).

Locality. —Huhe Bulak.

Diagnosis. — Skull low; zygomatic arch long; upper cheek teeth similar to those
of Forstercooperia confluens, but M3 with metacone; metalophids long and connect
with protolophids. Differs from F. totadentata in its apparently smaller size and
from F. confluens in having 1) zygomatic arch flatter and straighter; 2) in lateral
view flatter frontal and nasal bones; 3) M2 and M3 without anti-crochets. Differs
from F. sharamurunensis in having less molarized premolars.

Description. — Skull: frontal flat and low; zygomatic arch relatively flat and straight; anterior edge
of orbit anterior to M3; sagittal crest well developed, parietal fairly high; interparietal very distinct;
occipital crest thin; occipital part deeply convex forward; occipital condyle large; paramastoid process
may be very robust; length of bulla, 27 mm. P4: paraloph and metaloph form a semi-circle with wear;
anterior and posterior cingula present; no internal nor external cingula. M1: (heavily worn) paraloph
fairly robust; metaloph more robust; paracone prominent forming a clear paracone rib; metacone
relatively prominent, but metacone rib not present; angle between metaloph and metacone crest about
90°; anterior cingulum weak; no internal nor posterior cingula, external cingulum short, near ectoloph.
M2: larger in size; paraloph robust; dental wall with posteriorly projecting process at the position of
the anti-crochet; paracone prominent; metacone rib fairly strong, the angle between metaloph and
metacone (crest) about 90°; paraloph and metaloph connect at the base; anterior and posterior cingula
weak. M3: no anti-crochet; parastyle not prominent; paracone high; ectoloph connects with metaloph
forming straight crest, but metacone very distinct; paraloph not connecting with metaloph at the base;
anterior cingulum relatively strong.

Mandible: vertical edge fairly flat and straight; the angle between horizontal ramus and ascending
ramus larger than 90°; foramen mentale clear. P3: protoconid close to metaconid; protolophid long;
metalophid fairly short; no cingulum. M,: smaller than M2; protolophid parallel to hypolophid; meta-
lophid not connecting with protolophid; no cingulum. M2: similar to M1, but with extremely weak
cingulum on external wall. M3: (heavily broken) metalophid connects with protolophid.

62

Annals of Carnegie Museum

vol. 56

Fig. 46 .—Forstercooperia huhebulakensis new species. Holotype, V5770, skull, lateral, dorsal, and
palatal views.

Discussion. — The family Hyracodontidae has 11 genera, of which most are
Eocene; Ardynia, Urtinotherium, and Indricotherium are Oligocene.

The shape of the upper cheek teeth agrees with Forstercooperia, especially P4
and M3. The length of M1-3 of F. huhebulakensis is close to that of F. confluens

1987

Qi— Eocene Mongolian Arshanto Mammals

63

Table 1 5 .—Measurements (mm) of Forstercooperia huhebulakensis.

l w

V5770

P4 20.0 23.5

M! 24.5 27.5

M2 29.2 30.3

M3 26.0 30.0

M1-3 79.7

Length, anterior edge of orbital fossa to condyle 213.0

Width, across zygomatic arch 190.0

Height, from lower border of condyle to occipital crest 93.0

V5771

P3 13.7 10.2

P4 16.0 12.9

M, 22.2 17.0

M2 26.3 17.7?

M3 17.0?

V5772

M2 26.8 19.1

M3 29.8 18.0

Depth, mandible below M, 50.0

(79.7 mm and 87. 1-90.5 mm, respectively). F. huhebulakensis differs from species
of Hyracodon in having the paracone higher than the ectoloph whereas in the
latter the paracone is weak and not separated from the ectoloph. F. huhebulakensis
differs from species of Triplopus in that the M3 has a distinct metacone, and from
species of Allacerops in being smaller.

F. huhebulakensis occurs stratigraphically lower than other species of this genus.
Some phylogenetic trends within this genus (from F. huhebulakensis to F. con-
Jluens and F. totadentata to F. sharamurunensis) are: 1) increase in size; 2) nasal
notch more posterior; 3) molarization of premolars. As Radinsky (1967Z?) sug-
gested, Forstercooperia may be the ancestor of Indricotherium.

Forstercooperia ? grandis (Peterson) 1919
(Fig. 47a)

Referred specimen.— V 577 3, right P4 (ectoloph missing).

Discussion. — The cingulum continuous on the internal side of the tooth; the
paraloph is long, the protocone robust, metaloph short, and crista evident. This
tooth is larger than P4 of F. huhebulakensis, and resembles that of North American
Forstercooperia ? grandis (UCMP 69722). The metaloph of the North American
form is not continuous nor is it in the Inner Mongolian specimen after wear. In
dimensions this tooth is close to P4 of F. totadentata, but the latter lacks a con-
tinuous cingulum and crista.

Measurements. —V 577 3, P4, L/W: 19.4/22.1 mm (ca.).

Forstercooperia sp.

(Fig. 47b, c)

Referred specimens. -V 5777 , left M3 (field no. 77028H2); V5774, right M2
(77028H2).

64

Annals of Carnegie Museum

vol. 56

•

1 cm

1 cm

Fig. 41a-c.—Forstercooperia‘ ? grandis. V5773, right P4. b, c.—Forstercooperia sp. (b) V5774, right M2;
(c) V5777, M3.

Locality. — Daatein Obo.

Discussion. — In regard to the dental pattern, M3 (V5777) is quite similar to M3
of Urtinotherium incisivum, but the dimensions are too different between the two
M3s. Besides this, V5777 has a small cingulum while Urtinotherium incisivum
lacks one. The discovery of this tooth provides some clue to the phylogeny of the
Hyracodontidae. The length of V5774 (M2) is close to that of Forstercooperia
confluens but its width is less. The width of the protolophid is larger than that of
metalophid, a phenomenon rarely seen in the family Hyracodontidae.

Measurements.- L/W: M3 (V5777), 40.0/46.5 mm.

Acknowledgments

I thank the following American colleagues who so graciously helped me to complete the study of
the Arshanto Fauna: Drs. M. R. Dawson, L. Krishtalka, and R. K. Stucky, Carnegie Museum of
Natural History; Drs. M. C. McKenna, R. H. Tedford, and M. J. Novacek, American Museum of
Natural History. Special thanks go to Drs. Dawson, Krishtalka, and McKenna, for giving so much of
their time and energy to discussing and correcting the English manuscript. Dr. Dawson also encouraged
me to complete this study. I am very grateful to Miss Elizabeth A. Hill, Carnegie Museum of Natural
History, for typing and proofreading the manuscript several times, and to her and Andy Redline for
remounting the plates.

I would also like to thank the following Chinese colleagues for their help and support: Zhai Ren-
jie, Wang Ban-yue, Huang Xue-shi, Ye Jie, Wang Zhe-fu, Hou Jin-feng, Yang Ming-wen, Hu Hui-
ching, Shou Hua-quan, Chen De-wang, Peng Chuen, Xie Dang, and Xie Shu-hua, IVPP; and especially
Dr. Ming-zhen Zhou, IVPP, my professor, for his patient guidance and help both in field work and
laboratory research.

For discussions in the field I thank the following American colleagues: Drs. M. C. McKenna, W.
A. Clemens, Jr. (University of California) (1981), J. A. Wilson (University of Texas), M. R. Dawson,
and J. A. Schiebout (Louisiana State University) (1983).

Literature Cited

Berkey, C. P., and F. K. Morris. 1924. Basin structure in Mongolia. Bulletin of the American
Museum of Natural History, 5 1(5): 103-127.

. 1927. Geology of Mongolia. American Museum of Natural History, 475 pp.

Chow, Min-chen. 1957. Notes on some mammalian fossils from the Late Cenozoic of Sinkiang.
Vertebrata PalAsiatica, 1(1): 3 3-41.

1987

Qi— Eocene Mongolian Arshanto Mammals

65

— . 1963. Tillodont materials from Eocene of Shantung and Honan. Vertebrata PalAsiatica, 7:

97-104.

Cope, E. D. 1872. Description of some new Vertebrata from the Bridger group of the Eocene.

Proceedings of the American Philosophical Society, 12:460-465.

Dawson, M. R., Chuan-kuei Li, and Tao Qi. 1984. Eocene ctenodactyloid rodents (Mammalia)
of eastern and central Asia. Special Publication Carnegie Museum of Natural History, 9:138-
150.

Dehm, R., and T. zu. Oettingen-Spielberg. 1958. Palaeontologische und geologische Untersuch-
ungen im Tertiar von Pakistan 2. Die mitteleocanen Saugetiere von Ganda Kas bei Basal in
Nordwest Pakistan. Abhandlungen Bayerische Akademie der Wissenschaften, Mathematisch-
Naturwissenschaftliche Klasse, new series 91:1-54.

Domning, D. P., C. E. Ray, and M. C. McKenna. 1986. Two new Oligocene desmostylians and a
discussion of Tethytherian systematics. Smithsonian Contributions to Paleobiology, 59:1-56.
Gao, Yu. 1976. Eocene vertebrate localities and horizons of Wucheng and Xichuan basins, Henan.
Vertebrata PalAsiatica, 14(1):26— 34.

Granger, W., and C. P. Berkey. 1922. Discovery of Cretaceous and older Tertiary strata in
Mongolia. American Museum Novitates, 42:1-7.

Granger, W., and W. K. Gregory. 1934. An apparently new family of amblypod mammals from
Mongolia. American Museum Novitates, 720:1-8.

■ . 1 943. A revision of the Mongolian titanotheres. Bulletin of the American Museum of Natural

History, 80(10):349-389.

Huang, Xue-shi, and Tao Qi. 1982. Notes on fossils of the tapir group from the Late Eocene of
Lunan County, Yunnan Province. Gu Jizhui Dongwe yu Gu Renlei, 20(4):3 15-326.

Matthew, W. D. 1907. The relationships of the “Sparassodonta.” Geological Magazine, (5)4:53 1—
535.

Matthew, W. D., and W. Granger. 1925a. Fauna and correlation of the Gashato Formation of
Mongolia. American Museum Novitates, 189:1-12.

— . 1925 b. New mammals from the Irdin Manha Eocene of Mongolia. American Museum

Novitates, 198:1-10.

— . 1925c. The smaller perissodactyls of the Irdin Manha Formation, Eocene of Mongolia.

American Museum Novitates, 199:1-9.

. 1926. Two new perissodactyls from the Arshanto Eocene of Mongolia. American Museum

Novitates, 208:1-5.

Matthew, W. D., W. Granger, and G. G. Simpson. 1929. Additions to the fauna of the Gashato
Formation of Mongolia. American Museum Novitates, 376:1-12.

McKenna, M. C, Xiangxu Xue, and Ming-zhen Zhou. 1984. Prosarcodon lonanensis, a new
Paleocene microptemodontid palaeoryctoid insectivore from Asia. American Museum Novitates,

2780:1-17.

Osborn, H.F. 1924. Eudinoceras, Upper Eocene amblypod of Mongolia. American Museum Novi-
tates, 145:1-5.

Osborn, H. F., and W. Granger. 1931. Coryphodonts of Mongolia, Eudinoceras mongoliensis
Osborn, E. kholobolchiensis sp. nov. American Museum Novitates, 459:1-13.

— . 1932. Coryphodonts and uintatheres from the Mongolian Expedition of 1930. American

Museum Novitates, 552:1-16.

Pilgrim, G. E. 1 940. Middle Eocene mammals from north-west India. Proceedings of the Zoological
Society of London, 1 10A: 127-1 52.

Qi, Tao, and Zhenya Meng. 1983. The discovery of Hyrachyus modestus and Rhodopagus sp. nov.

(Mammalia, Perissodactyla) in Laiwu County, Shandong. Vertebrata PalAsiatica, 2 1(2): 103-1*08.
Radinsky, L. B. 1 964. Notes on Eocene and Oligocene fossil localities in Inner Mongolia. American
Museum Novitates, 2180:1-11.

— — — . 1965. Early Tertiary Tapiroidea of Asia. Bulletin of the American Museum of Natural
History, 129(2): 18 1-263.

— . 1 967 a. Hyrachyus, Chasmotherium, and the early evolution of helaletid tapiroids. American

Museum Novitates, 2313:1-23.

— . 1967&. A review of the rhinocerotoid family Hyracodontidae (Perissodactyla). Bulletin of

the American Museum of Natural History, 1 36(1): 1—45.

Ranga Rao, A. 1972. New mammalian genera and species from the Kalakot Zone of Himalayan
Foot Hills near Kalakot, Jamma and Kashmir State, India. India Oil and Natural Gas Com-
mission, Directorate of Geology, Special Paper no. 1 , 1 9 p.

Reshetov, V. Y. 1975. [Review of Early Tertiary Tapiroidea of Mongolia and USSR. The Joint

66

Annals of Carnegie Museum

vol. 56

Soviet-Mongolian Paleontological Expedition, Transactions, Vol. 2: Fossil fauna and flora of
Mongolia], pp. 19-53. (In Russian.)

Romer, A. S. 1966. Vertebrate paleontology. University of Chicago Press (3rd edition), 468 pp.

Sahni, A., S. B. Bhatia, J.-L. Hartenberger, J.-J. Jaeger, K. Kumar, J. Suder, and M. Vianey-
Liaud. 1981. Vertebrates from the Subathu Formation and comments on the biogeography of
Indian subcontinent during the early Paleogene. Bulletin de la Societe Geologique de France,
23(6):689-695.

Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bulletin of
the American Museum of Natural History, 85:1-350.

Stucky, R. K., and L. Krishtalka. 1983. Revision of the Wind River faunas, early Eocene of
central Wyoming. Part 4. The Tillodontia. Annals of Carnegie Museum, 52(1 7):375— 39 1 .

Szalay, F. S., and S. J. Gould. 1966. Asiatic Mesonychidae (Mammalian, Condylarthra). Bulletin
of the American Museum of Natural History, 132(2): 1-1 73.

Szalay, F. S., and M C. McKenna. 1971. Beginning of the Age of Mammals in Asia: the Late
Paleocene Gashato Fauna, Mongolia. Bulletin of the American Museum of Natural History,
1 44(4):27 1—3 1 7.

Tong, Yong-sheng, and Ying-jun Tang. 1977. A new species of Eudinoceras. Vertebrata Pal-
Asiatica, 15(2): 139-142.

West, R. M. 1980. Middle Eocene large mammal assemblage with Tethyan affinities, Ganda Kas
region, Pakistan. Journal of Paleontology, 54:508-533.

Wood, A. E. 1 962. The early Tertiary rodents of the family Paramyidae. Transactions of the American
Philosophical Society, 52:1-261.

Wood, H. E. 1934. Revision of the Hyrachyidae. Bulletin of the American Museum of Natural
History, 67:181-295.

Ye, Jie. 1983. Mammalian fauna from the Late Eocene of Ulan Shireh area, Inner Mongolia.
Vertebrata PalAsiatica, 2 1 (2): 1 09—1 1 8.

Young, Chung-chien, and Mei-nian Bien. 1935. Cenozoic geology of the Wenho-Ssushui District
of Central Shantung. Bulletin of the Geological Society of China, 14:221-246.

Zdansky, O. 1930. Die Altertiaren Saugetiere Chinas nebst Stratigraphischen Bemerkungen. Pa-
leontologia Sinica, 6(2): 1-87.

Zhou, Ming-zhen, Chuan-kuei Li, and Yu-ping Zhang. 1973. Late Eocene mammalian faunas
of Honan and Shansi with notes on some vertebrate fossils collected therefrom. Vertebrata
PalAsiatica, 1 1(2): 165-181.

Zhou, Ming-zhen, and Tao Qi. 1982. New materials of fossil mammals from the Middle Eocene
of Xintai County, Shandong Province. Gu Jizhui Dongwe yu Gu Renlei, 20(4):302-3 14. (Chinese,
English summary.)

Zhou, Ming-zhen, Tao Qi, and Yung Li. 1976. Paleocene stratigraphy and faunal characters of
mammalian fossils on Nomogen Commune, Si-zi-wang-qi, Nei Mongol. Vertebrata PalAsiatica,
14(4):228-233.

1987

Qi— Eocene Mongolian Arshanto Mammals

67

Appendix 1 .—List of Irdin Manha fauna.

Huhe Bulak

Irdin

Ulan

(Camp

Lushi

Mon-

Manha

Shireh

Margetts)

(Henan)

golia

Order Primates

Anaptomorphidae

Lushius qinlinensis

X

Order Insectivora

IPantolestes sp.

X

Order Rodentia

Paramyidae
paramyid spp.

Cocomyidae

Tsinlingomys youngi

Yuomyidae

X

X

Advenimus hohlini

X

cf. Advenimus sp.

X

Yuomys weijingensis

X

Order Lagomorpha

Leporidae

Shamolagus grangeri

Lushilagus lohoensis

X

X

Order Creodonta

Oxyaenidae

Sarkastodon mongoliensis
Hyaenodontidae

X

X

Paracynohyaenodon morrisi

X

Propterodon irdinensis

Hyaenodon sp.

X

X

Order Carnivora

Miacidae

Miacis invictus

M. lushiensis

miacid indet.

X

X

X

Canidae

Cynodictis sp.

Felidae

X

cf. Eusmilus sp.

X

Order Acreodi

Mesonychidae

Andrewsarchus mongoliensis

A. henanensis

X

X

A. gigas

Hapalodectes serus
‘ IHapalodectes serus

X

X

X

X

IHarpagolestes orientalis
Honanodon macrodontus

X

X

Lohoodon lushiensis

Pachyaena sp.

X

X

Mesonyx sp.

X

X

mesonychid indet.

X

X

68

Annals of Carnegie Museum

vol. 56

Appendix 1 . — Continued.

Huhe Bulak

Irdin

Ulan

(Camp

Lushi

Mon-

Manha

Shireh

Margetts)

(Henan)

golia

Order Tillodontia

Stylinodontidae

IStylinodon sp.

X

Order Pantodonta

Coryphodontidae

Eudinoceras mongoliensis

X

X

X

E. kolobolchiensis

Eudinoceras sp.

Hypercoryphodon thomsoni

X

X

X

Pantolambdodontidae

Pantolambdodon inermis

X

P. fortis

X

Order Perissodactyla

Brontotheriidae

Telmatherium parvum

X

Protitan grangeri

X

X

P. robustus

X

P. obliquidens

P. bellus

X

X

Rhinotitan grangeri

Microtitan mongoliensis

X

X

X

1 Microtitan sp.

Desmatotitan tukhumensis

X

X

Epimanteoceras formosus

X

Dolichorhinoides augustidens

X

Acrotitan ulanshirehensis
brontotheriid indet.

X

X

Eomoropidae

Lunania youngi

X

Eomoropus sp.

Isectolophidae
isectolophid indet.

Helaletidae

X

X

Helaletes mongoliensis

IColodon sp.

X

X

Lophialetidae

Lophialetes expeditus

X

X

X

Lophialetes sp.

X

X

Breviodon minutus

X

X

X

X

1 Breviodon minutus
cf. Breviodon minutus

X

X

IBreviodon sp.

Parabreviodon dubius (=cf. Brevio-

X

don acares, Radinsky, 1965, prob-
ably Irdin Manha beds.)

X

Simplaletes sujiensis

X

X

S. ulanshirehensis

X

Zhongjianoletes chowi

X

Zhongjianoletes sp.

Pataecops parvus

X

X

1987

Qi— Eocene Mongolian Arshanto Mammals

69

Appendix 1.— Continued .

Huhe Bulak

Irdin

Ulan

(Camp

Lushi

Mon-

Manha

Shireh

Margetts)

(Henan)

golia

Deperetellidae

Teleolophus medius

Deperetella sp.

X

X

X

X

Hyracodontidae

Rhodopagus pygmaeus

Rhodopagus sp.

X

X

X

ITriplopus proficiens

Triplopus sp. ( =Caenolophus or

X

X

Amyndontidae)

X

Forstercooperia totadentata
Forstercooperia cf. grandis
Forstercooperia spp.

Teilhardia pretiosa (“Ulan Shireh”

X

X

X

beds)

X

Prohyracodon sp.

Amynodontidae

X

Lushiamynodon menchiapuensis
ILushiamynodon sharamurunensis

X

X

Sianodon honanensis

X

Order Artiodactyla

Dichobunidae

Dichobune sp.

Choeropotamidae

X

Gobiohyus orientalis

X

X?

X

G. pressidens

X

G. robustus

X

X

Cebochoeridae

achaenodont indet.

Anthracotheriidae

Anthracotherium spp.

Leptomerycidae

X

X

cf. Archaeomeryx indet.

X

Order indet.

Didymoconidae

Mongoloryctes auctus

Kennatherium shirehensis

X

X

70

Annals of Carnegie Museum

vol. 56

Appendix 2.— Middle Eocene mammals in

the central part of Shandong Province.

Xintai Mengyin Laiwu

Order Condylarthra

Hyopsodontidae

IHaplomylus sp.

X

Order Tillodontia

Esthonychidae

Kuanchuanius shangtungensis

X

Order Pantodonta

Coryphodontidae

Coryphodon flerowi

X

Metacoryphodon xitaiensis

X

Order Dinocerata

Uintatheriidae

cf. Uintatheriuml sp.

X

Order Rodentita indet.

X

Order Creodonta

Hyaenodontidae

Thinocyori ? sichowensis

X

Order Perissodactyla

Equidae

Propalaeotherium sinense

X

Propalaeotherium sp.

X

(Hyracotheriinae: two species, no names,

Zdansky, 1930)

X

Brontotheriidae

gen. and sp. indet.

X

Chalicotheriidae

Grangeria canina

X

Helaletidae

Hyrachyus modestus

X

H. modestus' ?

X

H. metalophus

X

Helaletes sp.

X

Lophialetidae

Lophialetes sp.

X

Schlosseria sp.

X

Breviodonl minutus

X

Deperetellidae

Teleo tophus shandongensis

X

Teleolophus sp.

X

Hyracodontidae

Rhodopagus laiwuensis

X

Rhodopagus sp.

X

Order indet.

Heptaconodon dubium

X

Appendix 3.— Middle Eocene mammals in Zaisan Basin , Kazakhstan, U.S.S.R.

Tschai-

Kalmak-

Ulken-

Aksyir

bulak

Obaila

pay

Ulasty

Order Rodentia

Cocomyidae

Tamquammys tantillus

Yuomyidae

X

X

X

Saykanomys chalchas

X

X

Petrokozlovia notas

X

Order Lagomorpha

Leporidae

Shamolagus sp.

Order Condylarthra indet.

Order Creodonta

X

X

Hyaenodontidae

cf. Limnocyoninae

X

deltatheriid indet.

X

Order Pantodonta

Coryphodontidae

Eudinoceras obailensis

X

X

pantodont indet.

Order Perissodactyla

X

Brontotheriidae

Palaeosyops sp.

Isectolophidae

X

Isectolophus bogdulensis

cf. Isectolophus sp.

X

X

Helaletidae

Veragromovia desmat other oides

X

Helaletes cf. mongoliensis

Helaletes sp.
helaletid indet.

X

X

X

Lophialetidae

Lophialetes cf. expeditus

X

Lophialetes minor

Lophialetes n. sp.

X

X

Breviodon sp.

lophialetid indet.

X

X

Deperetellidae

Teleolophus cf. medius
deperetellid indet.

X

X

Hyracodontidae

Rhodopagus cf. pygmaeus

Triplopus cf. implicatus

X

X

X

Triplopus sp.
cf. Triplopus sp.

X

X

X

X

Prohyracodon sp.

hyracodontid indet.

X

X

Order Artiodactyla

Dichobunidae

dichobunid indet.

X

Order uncertain

Paraphenocodus solivagus

X

Aksyiria oligestus

X

72

Annals of Carnegie Museum

vol. 56

Appendix 4. —Middle Eocene mammals from Pakistan and India.

Ganda Chor-

Kas Kalakot lakki Kutch

Order Insectivora indet.

Order Chiroptera indet.

Order Primates
Adapidae indet.

Order Condylarthra(?)
Hyopsodontidae(?)

Dulcidon gandaensis

Order Acreodi
Mesonychidae
cf. Honanodon
indet.

Order Tillodontia

Basalina basalensis

Order Cetacea
Protocetidae
Pakicetus inachus
Gandakasia potens
Protocetus sloani
P. horudiensis
P. attoki
indet.

Indocetus ramani
Ichthyolestes pinfoldi

Agorophiidae
Andrewsiphius kutchensis
A. minor

Order Creodonta
Hyaenodontidae
Paratritemnodon indicus

Order Carnivora
Miacidae indet.

Order Sirenia
Protosirenidae
Protosiren fraasi

Order Proboscidea
Anthracobunidae
Lammidhania wardi
Jozaria palustris (Kohat District)
Anthracobune pinfoldi
A. pilgrimi

Order Perissodactyla

Hyracodontidae
Prothyracodon(I) kalakotensis
Forstercooperia jigniensis
Isectolophidae
Sastrilophus dehmi

X

X

X

X

X

X

X X

X X

X

X

X

X

X

X X

X

X

XXX

X

X

X

XX X

X

X

X

X

Kohat

X

X

1987

Qi Eocene Mongolian Arshanto Mammals

73

Appendix 4.— Continued.

Ganda

Chor-

Kas

Kalakot

lakki

Kutch

Kohat

Helaletidae

Hyrachyus asiaticus

X

H. ( Chasmotherium ) mckennai
Deperetellidae

Teleolophustf) daviesi

Lophialetidae

X

X

Schlosseria radinskyi

Kalakotia simplicidentata

X

X

Aulaxolophus quadrangularis
Brontotheriidae

X

Eotitanops dayi

X

Pakotitanops latidentatus

X

Order Artiodactyla

Helohyidae

Khirtharia dayi

X

X

X

Gohiohyus orientalis

Dichobunidae

X

Kunmunella rajauriensis

Chorlakkia hassani

X

X

Choeropotamidae

IndohyusQ ) kalakotensis

X

Indohyus indirae

Raoellidae

X

X

Raoella dograi

Anthracokeryx sp.

X

X

Haqueina haquei

X

Order Rodentia

Cocomyidae

Tamquammys sp.

Chapattimyidae

X

Saykanomys ijlsti

X

S. chalchae

X

X

S. vandermueleni

X

S. sondaari

X

S. lavocati

X

Chapattimys ihrahimshahi

X

C. wilsoni

Petrokoslovia sp.

Metkamys blacki

X

X

X

'

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 2, Pp. 75-95 6 March 1987

ECOLOGY OF SMALL MAMMALS IN A GALLERY
FOREST OF CENTRAL BRAZIL

Leslie Zuhn Nitikman1

Michael A, Mares2
Research Associate, Section of Mammals

Abstract

A 7-month mark-recapture study of small mammals was conducted on a l~ha plot in a gallery forest
associated with the Cerrado grasslands of central Brazil. The natural history of small mammals in
these forests is poorly known. Ten species were caught: two marsupials (. Monodelphis americana and
Marmoset agilis ) and eight rodents ( Oryzomys bicolor, O. capita, O. concolor, O. nigripes , Rhipidomys
mastacalis, Akodon cursor, Bolomys lasiurus, and Oxymycterus roberti ). The present study provides
information on a number of life-history characteristics (body size, sex ratios, age class ratio, biomass
estimates, survivorship estimates, movement patterns, and habitat utilization patterns). M agilis was
the most commonly captured arboreal species (192 times), and B. lasiurus was the most frequently
caught terrestrial species (141 times). M. agilis had the longest mean residency (56.1 days), as well as
the highest mean number of captures per marked individual (12.0 times). O. bicolor traveled the
greatest mean distance between successive captures (45.2 m). O. concolor males moved 45.3 m between
consecutive captures, much further than female conspecifics (25.9 m) Movement patterns were ana-
lyzed for the six most common species (M. americana, M. agilis, O. capita , O. concolor, O. nigripes,
A. cursor, and B. lasiurus ); all exhibited nonrandom movement through the study area. Habitat vari-
ables were analyzed by principal component and cluster analyses in order to evaluate habitat preferences
with respect to six distinct microhabitats: three types of semideciduous gallery forest (dense forest,
vine tangle forest, and forest mosaic), two types of forest ecotones (forest edge and bamboo edge), and
one disturbed area (fem thicket). A. cursor frequented the dense forest, whereas O. capita was caught
primarily in dense forest and forest mosaic. B, lasiurus was captured almost exclusively in the two
ecotonal microhabitats. O. concolor had an apparent preference for the fem thicket. O. nigripes was
trapped more frequently in fem thicket and dense forest than was expected. M. agilis tended to avoid
the fem thicket, but ranged randomly throughout all other microhabitats.

Introduction

The Cerrado is an endemic tropical savanna complex located in the highlands
of central Brazil The Cerrado exceeds 1.8 million km2 in area (Air Saber, 1971),
occupying approximately 25% of Brazil’s land area (Joly, 1970). It ranges from
the northern part of the state of Goias southward to western Minas Gerais state
and from the western edge of the state of Bahia westward to eastern Mato Grosso
and Ronddnia states (Eiten, 1972). It is bordered by the Amazon rainforest to
the north, semiarid Caatinga to the northwest, Atlantic Rainforest to the east,
semiarid Chaco to the south, and Pantanal swamplands to the west.

The Cerrado is composed of four major habitats: cerrado sensu strict© (s.s.),
cerradao, gallery forest, and campo (Eiten, 1972). The cerrado (s.s.) is the dominant
habitat in the Cerrado vegetation complex. It is a sclerophyllous scrub woodland

1 Address: Stovall Museum and Oklahoma Biological Survey and Department of Zoology, University
of Oklahoma, Norman, Oklahoma 73019.

2 Address: Stovall Museum and Department of Zoology, University of Oklahoma, Norman, Oklahoma
73019.

Submitted 10 November 1985.

75

76

Annals of Carnegie Museum

vol. 56

and is readily identified by its open canopy and twisted tree trunks and limbs.
The cerradao, literally “big cerrado,” is a taller, more dense scrub, practically a
forest. The canopy is mostly closed and the trees are usually greater than 10 m
in height. The gallery forest is a narrow band of vegetation, varying from a few
meters to about 200 m in width, that occurs along permanent rivers and streams
throughout the Cerrado; it is composed primarily of evergreen tree species. The
gallery forest transition into savanna is abrupt; often the ecotone is no more than
5 m in width. Campo is a grassland that varies widely in structure from savannas
with scattered low trees or shrubs (campo cerrado), to those with few but con-
spicuous shrubs (campo sujo), to prairies with essentially no woody plants (campo
limpo). Floristically, the Cerrado region is the richest and most diverse savanna
system in the world (Sarmiento, 1983). Heringer (1971) counted more than 300
plant species in 1 ha of cerrado (s.s.) in Distrito Federal. Heringer et al. (1977)
listed 1063 vascular plant species for the entire Cerrado region and this list is not
complete.

The gallery forest is one of the least-known habitats of the Cerrado. Floristically,
the Cerrado in general is relatively well known, but only a few botanists have
surveyed the gallery forest near Brasilia (Barbosa et al., 1984; Ratter, 1980; I.
Schiavini da Silva, personal communication). These preliminary studies suggest
that gallery forests are very complex; neighboring forests can be markedly different
in composition and form (Eiten, 1984). The gallery forest biota is diverse and
shares a close affinity in terms of mammal species composition with the neigh-
boring forests of the Amazon and the Atlantic Coast (Redford and Fonseca, in
press). Although few, if any, mammals are endemic to gallery forests, many species
range into central Brazil only via these mesic forest corridors (Alho, 1 982; Redford
and Fonseca, in press). To these forest mammals, the gallery forest is an extension
of the Amazon or Atlantic rain forests, permitting them to range deeply into the
Cerrado (Fonseca and Redford, 1984).

In spite of the great biotic diversity of the Cerrado, ecological studies of mam-
mals are rare. Several recent investigations have provided valuable information
on the ecology of various mammal species (e.g., Alho, 1979, 1981; Alho and
Souza, 1982; Almeida et al., 1981; Borchert and Hansen, 1983; Dietz, 1981;
Lacher et al., 1984; Mello, 1980; Paula, 1983; Pereira, 1982; Pereira and Alho,
1982; Redford, 1984; Souza, 1979; Souza and Alho, 1980). But even so, these
studies represent relatively short-term efforts. Basic information on more species
is required for advancements in theoretical work, such as analyses of ecosystems
or patterns of adaptive radiation and speciation.

The objective of this study was to examine the ecology and natural history of
small mammals in a Cerrado gallery forest. A small mammal community was
studied using mark-recapture procedures over a 7 -month period in 1984. In
addition, data were obtained on climate and habitat structure.

Study Site

The study site was located in a gallery forest on the Fazenda Agua Limpa, a
research and teaching field station jointly administered since the early 1960s by
the Departmento de Agronomia and the Laboratorio de Ecologia of the Univer-
sidade de Brasilia. The field station (15°58'S, 47°57'W) encompasses 4062 ha and
is located approximately 20 km SSW of downtown Brasilia, Distrito Federal. The
elevation is approximately 1000 m above sea level. Climatically, the region has
a pronounced tropical wet and dry season, with the annual rainfall averaging 1 526

1987

Nitikman and Mares —Brazilian Gallery Forest Mammals

77

mm (Eiten, 1984). The dry season lasts from three to four months during the
Southern Hemispheric winter, during which the mean monthly rainfall is only
10.5 mm (Eiten, 1984). On average, 85% of the rain falls in a 6-month period
from October to March. The mean daily temperature fluctuates little during the
year. Based on a 20-year average for Brasilia (1961-80), the mean seasonal tem-
perature is 18.3°C in the two coldest months (June, July) and 2L5°C in the two
warmest months (September, October) (Eiten, 1984).

Cerrado (s.s.) is the dominant vegetation type of Fazenda Agua Limpa. Campo
limpo also covers considerable area, including the higher ground towards the
eastern boundary of the reserve. Cerradao is relatively rare within the reserve.
Gallery forest occurs along the two permanent watercourses that run through the
reserve (Capetinga Creek and Agua da Oncai. Both creeks eventually flow into
Gama Creek, which in turn empties into Lake Paranoa in Brasilia, The vegetation
of Fazenda Agua Limpa is discussed in more detail by Ratter (1980).

The 1 -ha trapping grid was set in the Capetinga gallery forest, on the west side
of the creek, approximately 3.5 km upstream from its confluence with Gama
Creek. The grid was situated on a terrace that ranged from 2,5 to 10,0 m above
the creekbed. The eastern edge of the grid ran along Capetinga Creek; whereas,
the western edge was bounded by campo limpo habitat (Fig. 1). The Capetinga
gallery forest showed evidence of human disturbance— there was some selective
cutting of trees both during the study and in recent years. An overgrown network
of trails still remained.

Materials and Methods

Field Techniques

Trapping.— A 1-ha grid was established with 100 stations set at approximately 10m intervals. The
grid was irregularly shaped in order to fit the contours of the gallery forest (Fig. 1). Two collapsible
Sherman traps ( 1 6 by 5 by 5 cm) were set at each station; one on the ground, next to low vegetation
or alongside fallen trees, within a 1 -m radius of the stake; the other in trees or vines up to 3 m off the
ground and within a 2.5-m radius of the stake.

Trapping of small mammals was conducted from January through July 1984 (Table 1). There were
seven trapping periods, roughly corresponding to each of the months. Trapping occurred 10 days/
period during the first 5 months. In June, the trapping session was terminated after day 7; in July,
traps were set for 1 1 days. During January, only 57 stations were used and only arboreal traps were
set. A total of 1 2, 1 70 trap-nights was completed on the grid— 5800 trap-nights using ground traps and
6370 trap-nights using arboreal traps.

Live traps were baited with peanut butter, either plain or mixed with rolled oats. Traps were rebaited
every two or three days as needed. When the minimum nighttime temperature was low enough to
affect trap survivorship, cotton was added as insulating material. Animals trapped overnight were
removed each morning. On several occasions, traps were checked in late afternoon to determine
diurnal activity. Captured animals were marked by toe-clipping and examined to determine approx-
imate age, and reproductive condition. They were then released immediately at the point of capture.
Animals were weighed on the initial capture of each trapping period.

Voucher specimens (skin and skeleton) were prepared. They have been deposited in the collection
at the Laboratorio de Ecologia Animal, Universidade de Brasilia and the Stovall Museum of Science
and History, University of Oklahoma.

Data Analysis

Characterization of the habitat. — Microhabitats were defined using cluster and principal components
analyses on 26 habitat variables. Thirteen variables were direct measurements from each of the 100
trapping stations on the grid (Table 2). Most variables are self-explanatory. Density of trees was
calculated as follows. At each station, a pair of trees (with DBH > 9 cm) was marked within a 5-m
radius of the stake at each station. The closest tree to the stake was selected as the “A” tree of the
pair. The “B” tree was chosen as the one closest to the “A” tree. In all, 174 trees were marked. In

78

Annals of Carnegie Museum

vol. 56

order to include variables that described the area that surrounds each station, 1 3 more variables were
calculated. These were the mean values for each of the original variables of the stations immediately
surrounding each station. The number of adjacent stations included in these calculations varied from
1 to 4, depending on the position of the station on the grid. Raw data were standardized (i.e., to make
the character mean 0 and the variance 1), and a dendrogram was constructed using the unweighted
pair-group method using arithmetic averages (UPGMA; Sneath and Sokal, 1973) on an average
distance matrix. Many of these variables were correlated; thus, a principal components analysis was
performed on standardized data to produce a new set of orthogonal components to summarize the
character variance. Trap sites were projected onto the resulting components using the standardized
data.

Computations were performed on the IBM System 3081 and an Apple II at the University of
Oklahoma. Most of the analyses were done using the following computer packages: BIOM (Rohlf,
1982); NT-SYS (Rohlf et al., 1979); and SAS (Barr et al., 1979).

Species accounts. — Only species with a sufficient number of captures were included in analyses
requiring statistical testing. Animals captured initially as subadults and recaptured later as adults were
included in both age class categories for analyses that compared the two classes. Age classes were
estimated using pelage, genital characteristics, and mass.

Population size was estimated by direct enumeration, based on the minimum number of individuals

1987

Nitikman and Mares —Brazilian Gallery Forest Mammals

79

Table l.— Dates of trapping periods and numbers of traps used at Capetinga.

Period

Dates

Duration
(in days)

No. traps

Arboreal Ground

Total

trap-nights

1

1 1-20 Jan

10

56

_

560

2

23 Feb-3 Mar

10

100

100

2000

3

20-29 Mar

10

100

100

2000

4

19-28 Apr

10

100

100

2000

5

13-22 May

10

100

100

2000

6

15-21 Tun

7

100

100

1400

7

14 24 Jul

11

100

100

2200

known to be alive during a particular trapping period. Biomass, defined as the combined mass (in grams)
of conspecifics per hectare, was used as a measure of ecological dominance. Monthly biomass estimates
for each species were determined by multiplying the minimum number of individuals known alive
each month by the mean monthly mass of all animals captured.

Sex ratio and age class distributions were based on monthly capture data. Sex ratios for subadults
and adults were calculated separately but were not reported unless statistically different. Deviations
from the expected 1 : 1 ratio were tested for significance using %2 analysis.

Estimations of survivorship and longevity were made by examination of the mean and maximum
number of days between the first and final captures. All individuals captured more than once were
included in these calculations.

Body mass of species was examined when sufficient data were available. To examine possible seasonal
differences or patterns in the distribution of body mass, data were compiled for each month.

Movement patterns of each species were examined by calculating the mean and maximum distance
traveled between successive capture sites. All individuals captured more than once were included in
these calculations. Reproductive data were based on external examination of captured animals and
were not sufficient to permit analyses of annual cycles.

Patterns of habitat utilization were examined in terms of a species’ use of three-dimensional space
and choice of microhabitat. Use of three-dimensional space was examined in arboreal and semi-
arboreal species only. The following variables were measured: trap location, general characterization
of arboreal support (e.g., horizontal branch, tree fork, etc.); trap slope, relative slope of arboreal trap
support over 1 m preceding the trap entrance; trap height , height above ground of the trap entrance;
support diameter, diameter of arboreal support (to the nearest 10 mm) at the entrance of the trap. A
X2 analysis was used to test whether these variables affected trap selection. Microhabitat preferences

Table 2.— Description and sampling method for 13 habitat variables at Capetinga.

Variable

Sampling method

Canopy height
Density of trees

Number of trees
Number of tree species
Density of ferns

Density of bamboo

Density of grass

Density of vines
Vine size 1
Vine size 2
Vine size 3
Vine size 4
Vine size 5

Relative height of canopy on a 0-3 scale

Sum of the distances between station stake and “A” tree, and be-
tween “A” tree and “B” tree
Number of trees within a 5 m radius of the station stake
Number of tree species within a 5 m radius of the station stake
Relative density of ferns within a 5 m radius of the station stake on
a 0-4 scale

Relative density of individual clumps of bamboo (woody grasses)
within a 5 m radius of the station stake on a 0-4 scale
Relative density of grass ground cover within a 5 m radius of the
station stake on a 0-4 scale
Relative density of vines and lianas on a 0-4 scale
Presence of vines with a maximum diameter of 0-3 cm
Presence of vines with a maximum diameter of > 3-6 cm
Presence of vines with a maximum diameter of >6-9 cm
Presence of vines with a maximum diameter of >9-12 cm
Presence of vines with a maximum diameter of > 12 cm

80

Annals of Carnegie Museum

vol. 56

Table 3 .—Relative composition of each principal component of microhabitat analysis at Capetinga.
The first 13 variables are defined in Table 2; the last 13 are the mean values of surrounding traps.

Components

Variables

i

ii

hi

IV

V

Canopy height

0.247

-0.222

0.079

-0.333

0.287

Density of trees

0.094

-0.400

0.121

-0.239

0.531

Number of trees

0.715

-0.024

-0.116

0.032

-0.306

Number of tree species

0.732

-0.003

-0.138

-0.042

-0.268

Density of ferns

-0.360

0.593

-0.044

-0.077

-0.007

Density of bamboo

-0.264

-0.358

0.060

0.603

-0.153

Density of grass

-0.003

-0.106

-0.297

-0.094

0.637

Density of vines

-0.110

-0.443

0.255

-0.331

0.169

Vine size 1

0.241

-0.523

-0.022

-0.143

0.004

Vine size 2

-0.199

-0.352

0.424

-0.350

-0.068

Vine size 3

-0.249

-0.317

0.543

-0.105

0.173

Vine size 4

-0.118

-0.209

0.649

0.313

0.144

Vine size 5

-0.003

-0.106

-0.297

-0.094

0.637

Sur. canopy height

0.397

-0.266

-0.418

-0.089

0.295

Sur. density of trees

0.058

-0.123

-0.557

0.056

-0.034

Sur. density of species

0.761

-0.121

0.245

0.059

0.024

Sur. density of ferns

-0.139

0.548

0.055

-0.104

0.370

Sur. density of bamboo

-0.252

-0.444

-0.082

0.641

-0.119

Sur. density of grass

0.123

0.395

-0.203

0.042

0.098

Sur. density of vines

-0.391

-0.192

-0.141

-0.423

-0.479

Sur. vine size 1

0.205

-0.497

-0.178

-0.388

-0.322

Sur. vine size 2

-0.421

-0.249

-0.185

-0.535

-0.295

Sur. vine size 3

-0.437

-0.313

-0.422

-0.125

0.164

Sur. vine size 4

-0.065

-0.388

-0.512

0.301

0.198

Sur. vine size 5

-0.126

-0.402

-0.421

0.430

0.048

Percent of character
variance explained

13.6

11.6

10.7

9.0

6.9

were examined by using a x2 analysis to compare observed trapping frequencies in each microhabitat
against the expected (based on the percentage of available traps in each microhabitat type).

Results

Microhabitats

The first five principal components of the initial set of 26 habitat variables
explained 5 1.8% of the variation (Table 3). Generalized trends were identified by
examining the habitat variables with significant loadings (greater than ±0.5). Table
3 can be summarized by noting that high positive loadings on Component I reflect
a high density and diversity of trees, while negative loadings mean more open
forest. Component II positive loadings indicate dense fern patches and the absence
of slender vines. Positive loadings on Component IV indicate dense stands of
bamboo and few surrounding vines, while Component V positive loadings mean
that dense clumps of grass were present, along with a high density of trees, both
of these being surrounded by areas of lower vegetation density. On each com-
ponent, negative loadings reflect the opposite traits of positive loadings. Relative
loadings of each component reflect six distinctive microhabitats (fern thicket,
dense forest, vine tangle forest, forest mosaic, bamboo edge, and forest edge; Fig.
1). Components with loadings greater than ±0.20 were also considered as con-
tributing factors in the overall make-up of a microhabitat.

A cluster analysis of 26 habitat variables was performed to further define the

1987

Nitikman and Mares— Brazilian Gallery Forest Mammals

81

relationship that these microhabitats have with each other. Results were projected
onto a dendrogram that clustered similar trap stations together (Fig. 2). The
analysis shows a distinct dichotomy between the Fern Thicket and Bamboo Edge
and the remaining microhabitats. The primary division separates the Fern Thicket
from the others. The secondary division separates three trap stations that are
characterized as having dense tangles of lianas and vines. These stations were
located within the Vine Tangle Forest (station El 2) and the Forest Mosaic (stations
Cl 3 and 18). The tertiary division separated the Bamboo Edge microhabitat from
the others. Some structure is evident within the remaining group, such as the
presence of a subcluster of Forest Edge trapsites within the larger cluster. The
three types of semideciduous gallery forest (Dense Forest, Forest Mosaic, and
Vine Tangle Forest) were not divisible by cluster analysis, suggesting that these
microhabitats are not as distinctive as the ecotonal microhabitats (Bamboo Edge,
Forest Edge, and Fern Thicket).

The Fern Thicket microhabitat was a cleared area at the north end of the grid
(Fig. 1; 12% of the total area), where bracken fern ( Pteridium aquilinum) formed
extremely dense thickets 2 to 4 m in height. Other plants included low grasses,
small vines, and sparsely scattered emergent trees, many of which were dead. This
area received direct sunlight and experienced extreme daily temperature fluctua-
tions, especially in winter when the temperature ranged from 3°C at night to 34°C
during the day. The ferns formed a closed canopy and provided good cover for
both arboreal and terrestrial mammals.

The Bamboo Edge microhabitat (13% of the grid) was situated along much of
the gallery forest margin, where there were dense stands of woody bamboo (Ich-
nanthus bambusiflorus) forming a canopy at 5 m. This species was present through-
out the grid, but was sparse inside the gallery forest. Trees were widely dispersed.
Dominant tree species included Belangera glabra and Callisthene major.

The Forest Edge microhabitat comprised 10% of the grid and was important
in understanding the distributional patterns of the resident mammals. The tran-
sition from forest to savanna was rather abrupt, with the average width being 7
m. Here the tall grass of the campo intermingled with the thinning trees. Trees
were large and leaned toward the adjoining campo, forming a wall that joined the
gallery forest canopy with the substrate of the campo. The dominant tree species
were Belangera glabra, Callisthene major, and Qualea dichotoma. The under-
growth shrubs and saplings were dense. Herbaceous vines and lianas were abun-
dant and acted as an arboreal network for climbing mammals.

Three types of semideciduous gallery forest occurred over the rest of the grid.
At the south end was Dense Forest, representing 21% of the grid. This was the
least disturbed habitat and plant diversity was high. The forest floor was bare
except for scattered small shrubs and small trees. The 15-m high canopy was
relatively closed, with few emergent trees. Common tree species were Amaioua
guianensis and Protium sp.

The Forest Mosaic, a mixture of forest and light gaps, comprised 27% of the
grid. Light gaps were formed either by natural tree falls or selective cutting by
humans. In light gaps, the canopy was completely open and the sun shone directly
on the forest floor. Undergrowth was thick and, in older gaps, became an impen-
etrable thicket. In between light gaps, the canopy was relatively open and ranged
from 7 to 10 m in height. Large trees emerged above the forest canopy. Dominant
emergent species were Callisthene major, Copaifera langsdorfii, and Piptocarpha
macropoda. Tapirira guianensis was a common canopy-level tree.

The third type of forest was the Vine Tangle Forest (16% of the grid), and was

82

Annals of Carnegie Museum

vol. 56

Bamboo Edge

Forest Edge

V i ne Tangle

Fern Thicket

I I -I L_ I » I -L | J j |

1.0 8 0.6 0.4 0.2 0.0

DISTANCE

Fig. 2.— Distance dendrogram of 100 trap stations based on characters and clustered by unweighted
pair-group method using arithmetic averages. The cophenetic correlation is 0.85.

1987

Nitikman and Mares —Brazilian Gallery Forest Mammals

83

characterized by large trees having dense tangles of vines and lianas to a height
of 4 m. Common tree species were Piptocarpha macropoda and Xylopia aro-
matica. This microhabitat seemed to be a more mature form of the forest mosaic.
Throughout the grid, neither palms nor epiphytes were common. Salacia elliptica
was one of the most common subcanopy trees. Alongside the creek, Piper sp. and
a grass ( Panicum sp.) were common undergrowth plants.

Species Accounts

During seven months of trapping, 118 individuals were caught 731 times,
including two species of marsupials and eight species of rodents. Marmosa agilis
and Monodelphis americana were the marsupials, and the rodents were Oryzomys
bicolor, Oryzomys capito, Oryzomys concolor, Oryzomys nigripes, Rhipidomys
mastacalis, Akodon cursor, Bolomys lasiurus, and Oxymycterus roberti.

Marmosa agilis.— This small nocturnal didelphid marsupial ranges widely in
the Cerrado region. It is generally associated with mesic areas, such as gallery
forests and orchards and was the most commonly caught species at the Capetinga
study site. Although primarily arboreal, 18% of the captures were in terrestrial
traps. Individuals always escaped upon release by way of trees and vines. M. agilis
had the longest mean residency of any of the small mammal species on the grid
and had the greatest mean number of captures per individual (Table 4). Average
mass and biomass estimates are given in Tables 5 and 6. M. agilis is insectivorous
and frugivorous (Nowak and Paradiso, 1983); this was supported by observations
of a captive individual that accepted a wide variety of fruits and insects.

Fourteen males and only two females were caught over six months. One of the
females was captured 37 times. Neither female seemed to reach adulthood during
the study period; however, accurate aging of females was difficult since the vaginal
opening is internal. Females lack a true pouch and teats remain hidden when not
lactating. Neither of the two females was ever observed to lactate during the study
period. During the first 4 months of the study, the only males caught were subadult,
identified by their relatively small size and the hairy condition of their scrotums.
From May onward, males began to show signs of sexual maturity; their growth
rate began to stabilize, the testes enlarged significantly, and the scrotum lost its
hairy appearance and turned a blue-gray color. In June and July, all unmarked
males that appeared on the grid were fully mature adults.

Marmosa agilis was a resident species; 50% of the individuals were caught in
three or more consecutive months. Their ranges tended to shift slightly each month
and overlapped with other individuals at the margins. Adults traveled significantly
further than subadults between successive captures (41.1 m vs. 28.9 m; Student’s
r-test, P < 0.01).

Marmosa agilis was caught in all six microhabitats (Table 7). The x2 analysis
showed that its distribution across the grid differed significantly from that pre-
dicted if habitats had been used randomly (P < 0.01). However, when the fern
thicket microhabitat was removed from the analysis, the distribution was random.
Therefore, it seems that M. agilis was actively avoiding the fern thicket and not
distinguishing among the other microhabitats. Considering that M. agilis was
strongly arboreal, these findings are not surprising. The fern thicket was the only
microhabitat analyzed that was located outside the forest. Arboreal access was
limited in this area.

Analysis of three-dimensional space requirements showed that M. agilis entered

84

Annals of Carnegie Museum

vol. 56

* §•

■t§?

B S

£ S

II

S 5

o' o
-« 3
So 69

c

Ju

P o
S-Q
k5

JJS

E >

3 T3

— - rj- —< O — ; r~- On O — < © no

ri a! oo vo r-‘ r- rn no i <n <n *4 \

^ootNrn'oO'no 1 1

OO-'fNnON^^

+1 +1 +1 +1 +1 +1 +1 +1

— iO\OOM^tmam

oocn^r

Tf Tfr ON

rnoNOGw
r- r-’ on on

m Tt m m

+1 +1 +1 +1
<NC50<N<AJ
t"~" rn fO NO
n (<vcf

<N <N
m 3- on
+1 +1 +1
N- <A1 <N
<N t"-’ ©
m <N —<

i t"- oo
I Tf oo

o o

NO rn

+1 +1

cn ©

ON — ■
N- m

O O O <N
m ©

<N

■3-

cn

<A>

rn

©

<A5

<N

<Al

NO

iri

<N

N"

od

NO

<A5

NO

CN

+1

+1

+1

+1

+1

+1

+ 1

+1

+ 1

+1

O

r-

oo

ON

<AJ

oo

cn

<N

q

no

<N

ro

<N

OO

<A)

<N

*A5

r-‘

<N

On OO ^ OO

<3

K

2

O v.

3~§

II

tlLS

O s -S p

| |t!

^ o o o

"vj- '

m m N“

-h WN -

m m o <N

<N — ' <N

o -3

Et3-~
•fell S’

K O ^

• O

'So

3 &

C/5 ,0

<N g
3

^ a

t-l _H

tP 3

U-4 +-»

— . O

3 3

II

NO <N

— 1 N-

&&
X) X)

— • S c

-d c ° o

05 o O o

wh .2 <u <u

3 -S C/J C/5

• *N . ^

o<r « «

C3 > w o
o Ho-
3 a3

■S* e 6

3

NO — <

>AN -O -a <N ITN
I-. ■> e U

<2 i 3 £ <2

8 -S 2 8 8

C '-H c c c

<30033

."2 o +1 -52 .23

x) ^ „ d d

£§SS£

E

1987

Nitikman and Mares— Brazilian Gallery Forest Mammals

85

Table 5.— Average mass (grams) for each species of small mammal at Capetinga.

Species

Average mass

Adult

Subadult

Marmosa agilis

26.4 ± 5.7a

16.2 ± 3.3

15.0-36.0 (26)b

6.5-24.0 (45)

Monodelphis americana

19.5

14.1 ± 2.6

(1)

9.5-18.0(12)

Oryzomys bicolor

28.4 ± 4.0

17.0 ± 5.3

21.0-53.0 (1 1)

10.0-22.0 (4)

O. capito

58.7 ± 7.7

33.5 ± 16.6

42.0-72.0 (17)

12.0-65 (4)

O. concolor

58.8 ± 7.1

42.6 ± 5.8

41.0-72.0 (19)

36.0-50.0 (9)

O. nigripes

22.6 ± 4.4

12.7 ± 3.0

16.0-33.0 (39)

6.0-18.0(19)

Rhipidomys mastacalis

68.0

27.5

(1)

12.0-43.0 (2)

Akodon cursor c

47.2 ± 4.9

21.5 ± 6.4

39.5-57.0(15)

12.5-26.0 (4)

38.0 ± 1.6

36.0-40.0(14)

Bolomys lasiurus

42.2 ± 7.5

30.8 ± 3.4

33.0-63.0 (33)

21.0-35.0(14)

Oxymycterus roberti

71.0 ± 3.8

55.0

66.0-75.0 (4)

52.0-58.0 (2)

a Mean ± one standard deviation.

b Range, with sample size in parentheses.

c First average mass for adult males; second for adult females.

traps set on fern fronds less frequently than expected (P < 0.01). These results
agree with the macrohabitat data. M. agilis had a preference for traps set on
branches 20-40 mm in diameter ( P < 0.01) and entered traps randomly without
regard to support angle or height.

Cotton balls were provided in all traps for insulation, but M. agilis never used
the material to make temporary nests. During the cool winter mornings, this
species often was found torpid in the traps. Torpidity was apparently an effective
response against low temperatures; despite the lack of an insulating nest, M. agilis
had the lowest rate of trap mortality.

Monodelphis americana.— This small terrestrial didelphid is widespread
throughout eastern Brazil (Nowak and Paradiso, 1 983), although little information
is available on its ecology. Both sexes have three dark stripes on the dorsum, and
females lack a true pouch.

At the Capetinga study site, M. americana was relatively uncommon. It was
highly transient, with only two individuals (22%) caught more than twice. Indi-
viduals were found in traps during most of the late afternoon trap checks, sug-
gesting that they are active during the day. Only one adult (a male) was trapped;
all other individuals were subadults. The sex ratio of M. americana caught in
traps on the grid was 8:1.

Animals were caught during all the months that terrestrial traps were set. The
sample size was too small to test for significance of microhabitat preference.
Nevertheless, this species did not seem to show strong preference for any particular

86

Annals of Carnegie Museum

vol. 56

Table 6.— Monthly biomass estimates (g/ha) based on the minimum number of known individuals for
10 small mammal species from Capetinga. Mean of Marmosa agilis computed for months of January
through July; means for all other species computed for February through July. Number of individu-
als/month given in parentheses.

Species

Jan

Feb

Mar

Apr

May

Jun

Jul

Mean

Monodelphis

—

56

42

10

47

16

20

32

americana

(4)

(3)

(1)

(3)

(1)

(1)

Marmosa agilis

67

79

92

123

122

142

258

126

(5)

(5)

(6)

(7)

(6)

(6)

(9)

Oryzomys bicolor

—

137

103

54

28

29

30

63

(5)

(5)

(2)

(1)

(1)

(1)

O. capito

—

358

319

185

55

63

192

195

(10)

(7)

(4)

(1)

(1)

(3)

O. concolor

—

64

94

161

150

276

309

176

0)

(2)

(4)

(3)

(5)

(5)

O. nigripes

—

18

117

162

136

149

241

137

(1)

(6)

(9)

(7)

(10)

(ID

Rhipidomys mastacalis

—

110

—

—

—

68

—

30

(2)

(1)

Akodon cursor

—

127

84

191

117

205

242

161

(3)

(2)

(5)

(3)

(5)

(5)

Bolomys lasiurus

—

—

—

198

290

467

558

252

(4)

(7)

(12)

(16)

Oxymycterus roberti

__

75

72

58

—

125

55

(1)

(1)

(1)

(2)

Total

67

969

923

1142

945

1415

1975

microhabitat. M. americana did not have a torpidity response to low temperatures,
not did it make use of the cotton provided in each trap. Trap mortality was high
on cool nights.

Oryzomys bicolor.— This nocturnal cricetid rodent ranges from Panama to
tropical South America east of the Andes (Handley, 1976; Honacki et ah, 1982;
Patton et al, 1982). Within this range, it inhabits unflooded evergreen forest
(Alho, 1982) and is sometimes found in native dwellings (Husson, 1978). It has
fleshy thick-padded feet characteristic of arboreal rodents. During this study it
was captured in ground traps only three times ( 1 4% of captures).

Eight individuals (6 males and 2 females) were captured 22 times. Three of the
eight (37%) were caught in more than one trapping period and only one was
captured for more than two periods. In other areas, this species has also been
difficult to recapture. For example, in Venezuela, 71% were never recaptured
(O’Connell, 1979). Of all the species present on the grid, O. bicolor traveled the
greatest mean distance between successive captures (Table 4). The sample size
was not large enough to test for microhabitat preferences. No preference was
shown for specific traps based on their height or slope (x2 test, P > 0.05).

Oryzomys capito.— This species is common throughout the Neotropics and can
be found in nearly all habitats, including campo, cerrado, dry and humid forests,
agricultural fields, as well as in houses (Alho, 1981; Handley, 1976; Mares et al.,
1981 a; Mello and Moojen, 1979; Moojen, 1966). During the seven months of
trapping, six females and eight males were caught a total of 68 times. Only three
percent of the captures were made in arboreal traps. Forty-four percent of the

1987

Nitikman and Mares— Brazilian Gallery Forest Mammals

87

Table 1.— Percent of captures made in each of the following microhabitats: BE, Bamboo Edge; DF,
Dense Forest; FE, Forest Edge; FM, Forest Mosaic; FT, Fern Thicket; VT, Vine Tangle Forest, n =
number of captures, x2 test; * = P < 0.05, ** = P < 0.01, nt = not tested.

Microhabitats

Species

n

BE

DF

FE

FM

FT

VT

X2

Monodelphis americana

34

15

23

15

23

6

18

nt

Marmosa agilis

192

20

21

10

27

4

18

**

Oryzomys bicolor

22

0

14

5

45

0

36

nt

O. capito

68

6

38

7

43

0

6

**

O. concolor

68

1

6

3

15

59

16

*

O. nigripes

133

14

30

11

12

26

7

Akodon cursor

52

2

52

17

8

13

8

**

Bolomys lasiurus

141

42

6

46

6

0

0

**

Percent expected

13

21

10

27

12

17

marked animals were caught in only one trapping period. In February and March
(late wet season), it was the most commonly caught species on the grid; in May
and June (mid-dry season), it was one of the rarest.

Oryzomys capita was caught in all areas of the grid, except the fern thicket. The
X2 analysis indicated that this species was trapped in dense forest and forest mosaic
more frequently (81% of all the captures) than would be expected by chance (Ta-
ble 7).

Oryzomys concolor.— This species is found in tropical forests and marsh-forest
complexes from southern Costa Rica to northern Argentina (Alho, 1982; Honacki
et al., 1982; Paula, 1983; Pine, 1973). At the Capetinga study site, it was the
largest arboreal rodent species caught on the grid.

Six males and two females were captured 68 times. Although O. concolor has
the well-padded feet characteristic of arboreal rodents, 29% of the captures were
in ground traps. All captures were made at night. Six individuals (75%) were
caught in two or more trapping periods. O. concolor had the longest mean length
of residency of all the small rodents on the grid (Table 4). There was a significant
sex and age difference in the mean distance traveled between consecutive captures.
Mean distance for adult males was 45.3 m, whereas for adult females, it was only
25.9 m (Student’s t- test, P < 0.01). Adults of both sexes traveled a mean distance
of 37.2 m; subadults traveled 22.9 m (Student’s /-test, P < 0.05).

Oryzomys concolor was quite specific in its habitat preference. Seventy-five
percent of all captures were made either in the fern thicket or in traps adjacent
to this microhabitat. It was caught more frequently in traps set on fern fronds
than was expected by chance (P < 0.01). It showed a similar preference for traps
that were set on supports less than 10 mm in diameter (P < 0.01); fern petioles
never exceeded 1 0 mm in diameter. Home ranges overlapped only slightly in the
cleared patches of forest.

Analysis of the three-dimensional space usage showed that O. concolor entered
traps set at less than 45° more frequently than expected. Trap height was not
important.

Oryzomys nigripes. — This small cricetid rodent commonly inhabits cerrado,
dense brush, gallery forests, coastal scrub, pastures, agricultural fields, and human
dwellings (Alho, 1981, 1982; Mares et al., 1981a 19816; Mello, 1969, 1977;
Mello and Moojen, 1979; Veiga-Borgeaud, 1982). At Capetinga, it was the smallest

88

Annals of Carnegie Museum

vol. 56

rodent on the grid and one of the most commonly trapped species. During the
course of this study, 23 individuals (8 females and 15 males) were captured 134
times. O. nigripes had one of the shortest mean residencies (28.8 days; Table 4).
Over half (52%) of the marked individuals were caught in only one trapping
period; 26% were caught only once.

Oryzomys nigripes exhibited a significant preference for the fern thicket and
the dense forest (P < 0.01). It entered traps in the forest mosaic less often than
expected. Analysis of traps set at specific heights revealed that O. nigripes preferred
traps set within 1 m of the ground. This species was primarily terrestrial. The
slope of the trap was not important.

Rhipidomys mastacalis. — This species is found primarily in mesic forests (Alho,
1981, 1982; Dietz, 1983; Fonseca and Redford, 1984; Handley, 1976; Mello,
1969; Mello and Moojen, 1979). During this study, all captures were made in
arboreal traps.

Only three individuals (two males and one female) were captured a total of
seven times, making R. mastacalis the least frequently captured small mammal.
None was caught in more than one period and all captures were made at night.
This species was captured only in the months of February and June, corresponding
to early and mid-dry season.

Akodon cursor.— This cricetid rodent is a heavy-bodied, vole-like mouse that
is found throughout the Cerrado region in gallery forests and cultivated fields
(Mello, 1969; Moojen, 1952). It is completely terrestrial. Three females and seven
males were caught a total of 53 times. In spite of the long mean residency (Table
4), 50% of the individuals were never caught a second time and thus were not
included in the calculation of residency length. Only one individual was caught
in more than two trapping periods.

Adult males weighed significantly more than females (47.2 vs. 38.0 g; Student’s
t- test, P < 0.05). There was no significant difference in the mass of subadult males
and females. At Capetinga, subadults were captured only from April through June,
suggesting that young are bom in the late wet season (ca. February and March).
Fifty-two percent of the captures of A. cursor were made in dense forest; this was
statistically higher usage than would be predicted if the animal randomly fre-
quented microhabitats (Table 7).

Bolomys lasiurus ( = Zygodontomys lasiurus).— This cricetid rodent is common
in nearly all habitats, both disturbed and natural (Alho, 1981; Dietz, 1983; Mares
et al., 1981a; Mello and Moojen, 1979; Moojen, 1966). At the Capetinga study
site, it was very common; 22 individuals (13 females and 7 males) were caught
141 times. Every month, 50% of the active individuals were previously unmarked,
suggesting that turnover was high. Only two individuals were not recaptured at
least once.

Bolomys lasiurus is almost exclusively terrestrial. One individual was trapped
twice in the same arboreal trap; this trap was set on a fallen tree less than 1 m
above the ground. Even though there are several previous studies concerning this
species (e.g., Alho, 1979, 1981; Alho and Souza, 1982; Almeida et al., 1981;
Pereira, 1982; Souza, 1979; Souza and Alho, 1980), there are no other reports of
B. lasiurus being captured in arboreal traps.

Bolomys lasiurus had the lowest mean distance traveled of any species in the
study. This was due, in part, to B. lasiurus ’ tendency to enter certain traps more
than others; four adjacent traps located within 1 m of the tall dense grass of the

1987

Nitikman and Mares— Brazilian Gallery Forest Mammals

89

cerrado accounted for 39% of the captures. Adult males traveled significantly
further than adult females (Table 4).

On 12 occasions, the traps were checked twice each day; once early in the
morning and again in the late afternoon. Every afternoon check revealed one to
three B. lasiurus ; generally, these were animals not found in traps during that
day’s morning check, suggesting that most individuals were diurnal. In a related
study, Lacher et al. (in press) found this species to be diurnal in cerrado grassland
habitat.

Bolomys lasiurus was absent from the grid during the first three trapping periods
(January-March 1984) and was first trapped in April. Thereafter, the number of
individuals increased dramatically, until by June it was the most commonly
encountered species. Adults began to appear in April and new adults continued
to be marked in the ensuing months; subadults were not captured before May.
The sex ratio of captured B. lasiurus was 4.1:1.

At Capetinga, B. lasiurus was quite specific in its microhabitat preferences;
distribution across the grid varied significantly from the expected (Table 7). Eighty-
eight percent of all captures were made in two microhabitats: the forest edge and
the bamboo. These two areas were located on the edge of the grid that bordered
the savanna. It was never trapped beyond the forest edge, entering the gallery
forest only at certain points and never very far.

Oxymycterus roberti.— This species is a characteristic resident of brejos, the
permanently inundated savanna that borders gallery forest. Several authors state
that typically it is confined to this habitat and is only rarely caught in adjacent
areas (Borchert and Hansen, 1983; Fonseca and Redford, 1984; Mathews, 1977;
Redford, 1984), although T. E. Lacher (personal communication) caught it fre-
quently in campo limpo. At the Capetinga study site, there was no bordering
brejo, yet O. roberti was trapped on the grid.

Five individuals were caught a total of 13 times; all were males. None of the
marked animals seemed to be resident in the study area. Three were caught only
once and none was caught in more than one trapping period. All but one capture
was made in the part of the grid that was dominated by bracken fern; one individual
was caught inside the gallery forest. Individuals were caught in four of the seven
months; the species was uncommon in both the dry and wet seasons.

Discussion

Small Mammal Diversity

Previous workers have noted that the gallery forest is the most complex habitat
in the Cerrado region (Fonseca and Redford, 1984; Mares et al., 1986; Ratter,
1980). In this study, eight species of rodents and two species of marsupials were
caught in gallery forest habitat. In an earlier live-trapping study at the Fazenda
Agua Limpa, Alho (1981) surveyed the rodents of three other habitats. In the
cerrado (s.s.), three species of rodents were trapped; in the campo, there were four,
and in the cerradao, there were also four. In a related study in a different gallery
forest, Mares et al. (1986) found seven rodent species in the gallery forest and six
in the campo. Thus, gallery forest is generally richer in species than the other
kinds of Cerrado habitats. Of the rodent species Alho (1981) caught, only Calomys
callosus was not caught during this study. O. nigripes (= O . eliurus) was captured
both inside and outside the gallery forest in sizeable numbers; it was considered

90

Annals of Carnegie Museum

vol. 56

to be a habitat generalist (Alho, 1982). B. lasiurus was caught in all four habitats,
but is not considered a regular inhabitant of the gallery forest.

Didelphid Life History

Earlier studies have noted that the life expectancy of some Marmosa species is
approximately 1 year (Hunsaker, 1977; Nowak and Paradiso, 1983). Further
evidence suggests that only one litter is produced in a lifetime (Enders, 1935).
The data presented here imply that M. agilis may breed only once annually. During
the entire study, all males gained weight at the same rate and passed through
various stages of development at the same time (Fig. 3). In the early half of the
study, all males were judged to be subadults. They began to show indications of
maturity in May, and by June and July, all males caught on the grid were fully
mature adults. During the latter months of the study, the only animal remaining
on the grid was a lone female we judged to be a subadult (based on body size).
She weighed 14.0 g on her first day of capture (23 February), reached a mass of
19.0 g by late April, and did not gain weight thereafter. She was captured pe-
riodically until the study ended. M. americana exhibited a similar growth pattern.
During most of the study, only subadults were captured. A single adult male was
captured once in July.

Data for M. agilis and M. americana indicated sex ratios that were strongly
skewed (7:1 and 8:1, respectively). Although this difference was statistically sig-
nificant, it may have been an artifact of the differential activity of each sex. Males
might be more active, inquisitive, or aggressive, and females might not be as rare
as the trapping record suggests. In Venezuela, adult male M. cinerea were trapped
six times as frequently as adult females, and lactating females were never captured
(O’Connell, 1979; August, 1981). A lactating female’s movement may be ham-
pered by the physical burden of young, and her foraging radius may be reduced.
At Capetinga, the only females of either species caught were subadults, suggesting
that M. agilis and M. americana were behaving in a similar manner.

Marmosa agilis was a long-term resident at Capetinga. This contrasted with
what is generalized about didelphids, that they are “nomadic, solitary animals
that do not consistently restrict their activities to any particularly areas” (Hun-
saker, 1977:1 19). During this study, most M. agilis individuals remained on the
grid for three or more months. Their ranges shifted slightly every month and a
few individuals seemed to eventually shift enough to be off the grid. M. agilis did
not seem to have mutually exclusive territories, since ranges overlapped with each
other.

Demographic Patterns

Demographic patterns varied from species to species. R. mastacalis seemed to
be present on the grid on a seasonal basis. At Capetinga it was only caught in
February and June, whereas at Parque Nacional de Brasilia, it was caught in May,
June, and July (dry season; Paula, 1983). Dietz (1983) noted that 90% of the
captures of R. mastacalis at Parque Nacional Serra da Canastra, Minas Gerais,
Brazil, were made in the dry season. If it did not migrate up into the canopy or
to other gallery forest sites during the months when it was absent from the gallery
forest study area, it probably moved to a habitat having a pronounced arboreal
component, such as cerradao.

Oxymycteris roberti was caught at Capetinga within the fern thicket microhab-
itat during the wet season, and twice in the forest mosaic during the dry season.

1987

Nitikman and Mares —Brazilian Gallery Forest Mammals

91

MONTHS

Fig. 3.— Mean monthly weight (grams ± SE) of Marmosa agilis males over course of study.

Alho (1981) caught it a few times in the gallery forest; Borchert and Hansen (1983)
and Paula (1983) noted that it was trapped in low numbers in all habitats except
gallery forest.

Paula (1983) reported on seasonal movements of small mammals in gallery
forest and campo habitats at Parque Nacional de Brasilia, approximately 25 km
from the Capetinga study site. He trapped from February to September, roughly
the same time period as the Capetinga study. Six species of small mammals
(defined as < 1 20 g mean mass) were caught at his study site; these same six were
also caught at Capetinga. O. bicolor, A. cursor, and O. roberti were not captured
in the gallery forest at Parque Nacional de Brasilia. M. agilis was extremely
common, even though Paula only sampled with terrestrial traps. In the wet season,
this species was caught primarily in the dry gallery forest; whereas, in the dry
season, it moved to the wet gallery forest. At Capetinga, seasonal movements
were not evident for M. agilis ; it remained common in the gallery forest during
all the months of the study.

Bolomys lasiurus was captured only from April onward (fourth month of trap-
ping). Even so, its biomass increased rapidly until by July it was 558 g/ha (Table
6); by then, it became the dominant small mammal species on the grid. Since the
forest margin, its preferred habitat, represented only 0.25 ha of the grid, its biomass
per ha of preferred habitat during that month was 2232 g/ha, or approximately
50 individuals per ha of forest margin. This is impressive when one considers
that the number of individuals/ha increased from 0 to 50 in just 4 months. Other
studies have noted that populations of B. lasiurus fluctuated. Mello (1980) found
that this species was absent from her grid during the spring of one year and
abundant the following spring. In Exu, state of Pernambuco, populations were
monitored periodically through several years. From 1967 to 1971, the density of
B. lasiurus reached 187 individuals ha-1 month 1 (Karimi et al., 1976). Years
later, in 1977, three months of trapping in the same general area (29,250 trap-
nights) yielded only one individual (Streilein, 1982). Dietz (1983) reported that

92

Annals of Carnegie Museum

vol. 56

populations were very low in the wet season in southern Minas Gerais state. At
Capetinga, it seemed as if the population bloomed during the study period, forcing
animals to migrate to suboptimal habitat, such as the ecotonal zone between the
gallery forest and the grassland campo.

Distributional Patterns

Six microhabitats were distinguished based on phytophysiognomic characters
and, subsequently, on small mammal distribution patterns. Broadly speaking,
these can be placed into three main groups: the fern thicket, the forest-savanna
ecotone, and the gallery forest itself. The fern thicket was the primary habitat for
O. concolor and was extensively used by O. nigripes. O. roberti was captured only
on rare occasions, but nearly always in the fern thicket. Because O. concolor
seemed to prefer disturbed habitat over the more pristine, one would expect their
numbers to increase through small-scale forest destruction.

The forest-savanna ecotone was divisible into two microhabitats: the forest
edge and the bamboo edge. Both were characterized as narrow strips of vegetation
that contained elements common to both the forest and the savanna. None of the
small mammal species differentiated between the two microhabitats, even though
they were quite different phytophysiognomically. Both constituted the dominant
microhabitats for B. lasiurus. Other studies conducted at the Fazenda Agua Limpa
suggest that the cerrado (s.s.) is the primary habitat for B. lasiurus (Alho, 1981;
Souza, 1979), but it is caught in nearly all major habitats throughout its range
(Alho, 1982). The other small mammal species did not show preferences for the
ecotonal zone. O. concolor was caught in this area less than expected, but this
was probably due to the relatively long distance that lay between the forest margin
and the primary habitat for this species, the fern thicket.

The gallery forest proper was subdivided into three kinds of forest: dense forest,
vine tangle forest, and the forest mosaic. Some small mammals appeared to
distinguish the apparently subtle differences that marked these microhabitats. A.
cursor had a clear preference for the dense forest, the least disturbed part of the
grid. All 10 species were trapped in the forest mosaic and yet no one species
showed a specific preference for it. The forest mosaic microhabitat was the least
definable by statistical classification methods.

Interspecific Avoidance Mechanisms

Several of the mammal species are very similar in appearance and are closely
related taxonomically, creating a potential for resource competition. A. cursor and
B. lasiurus are two such species. Physically they are quite similar, but the micro-
habitat preferences showed that the two species overlapped little. A. cursor pre-
ferred the inner forest, whereas B. lasiurus was a savanna species that entered the
Capetinga gallery forest only at the forest-savanna ecotone. In addition, B. lasiurus
seemed to be primarily diurnal, while A. cursor was nocturnal. Therefore, both
spatial and temporal mechanisms minimize competition for resources.

Four species of Oryzomys coexisted on the study area. Several differences served
to minimize resource competition between them. An obvious division was size:
O. bicolor and O. nigripes are smaller-bodied (28.4 and 22.6 g, respectively),
whereas O. capito and O. concolor are more than twice as large (58.7 and 58.6 g,
respectively). Small and large rodents could select food items of different sizes
and, thus, minimal niche overlap would occur. Within weight classes, spatial
separation reduces the degree of interaction between species. O. bicolor was pri-

1987

Nxtikman and Mares— Brazilian Gallery Forest Mammals

93

marily arboreal (86% of captures in trees); O. nigripes was primarily terrestrial
(70% of captures on ground), thus allowing spatial separation. O. capita and O.
concolor segregate in a similar manner: O. capito was almost exclusively terrestrial
(97% of captures on the ground), whereas O. concolor was arboreal (71% of
captures in the trees). In addition, they showed different habitat preferences. O.
concolor strongly preferred the fern thicket, whereas O. capito was never caught
there. O. capito exhibited a preference for the dense forest and forest mosaic,
while O. concolor ventured into these areas only occasionally. Thus, although
both species were present in high numbers and lived in close proximity, interaction
was probably a rare occurrence.

Our research is not definitive, but it allows us to begin to understand the ecology
of the small mammal fauna of a little-studied habitat, the gallery forest of the
Brazilian Cerrado. Although the gallery forest appears rather homogeneous at first
glance, there are distinct microhabitats within the forest and the small mammals
have responded to these by preferring one over another. This microhabitat pref-
erence and the fact that several species are ground dwellers, others are arboreal
and some occur both in the trees and on the ground, help account for the fairly
rich species composition of mammals in this narrow habitat, that extends only
from 100 m to 1 km in width. Our preliminary findings suggest that habitat
selection by small mammals is quite important in the tropical gallery forest at
permitting coexistence, perhaps by reducing competition. Additional research
should allow us to formulate a more detailed understanding of coexistence and
species richness in this extensive and important tropical habitat.

Acknowledgments

We thank D. D. Gettinger, R. H. F. Macedo, J. M. McPherson, G. D. Schnell, and B. M. Vestal
for their helpful comments on the manuscript; D. J. Hough for assistance with the analysis; C. J. R.
Alho, R. B. Cavalcanti, J. K. Braun, and D. P. S. Dias for logistical help in Brazil and Oklahoma; A.
A. A. Barbosa and 1. 1. Schiavini da Silva for the plant identifications; D. D. Gettinger, R. Gribel, M.
R. Lemes, and V. J. Soares for help with the data collection. C. O. Handley, Jr. and A. L. Gardner
kindly assisted in species identification. This study was part of a larger cooperative research program
sponsored by the University of Oklahoma and the Universidade de Brasilia. Financial support was
provided by grants to M. A. Mares from the National Science Foundation (INT-82- 12576 and DEB-
82- 13675), the OU Associates, the Oklahoma Biological Survey, and the University of Oklahoma
Graduate Student Association.

Literature Cited

Ab’Saber, A. N. 1971. A organizagao natural das paisagens inter- e subtropicais brasileiras. Pp. 1-
14, in III Simposio sobre o Cerrado (M. G. Ferri, ed.), Universidade de Sao Paulo, Sao Paulo,
Brazil, 239 pp.

Alho, C. J. R. 1979. Relative exclusiveness of use of space as a measure of spatial distribution of
mammal populations. Revista Brasileira de Biologia, 39:275-299.

— — — . 1981. Small mammals populations of Brazilian Cerrado: the dependence of abundance and
diversity on habitat complexity. Revista Brasileira de Biologia, 41:223-230.

. 1982. Brazilian rodents: their habits and habitats. Pp. 143-166, in Mammalian biology in

South America (M. A. Mares and H. H. Genoways, eds,), Special Publication, Pymatuning Lab-
oratory of Ecology, University of Pittsburgh, Linesville, Pennsylvania, 6:1-539.

Alho, C. J. R., and M. J. Souza. 1982. Home range and use of space in Zygodontomys lasiurus
(Cricetidae, Rodentia) in the Cerrado of central Brazil. Annals of Carnegie Museum, 52: 127-132.

Almeida, C. R., A. M. P. Almeida, and D. P. Brasil. 1981. Observation sur le comportement de
fouissement de Zygodontomys lasiurus pixuna Moojen, 1943. Reproduction au laboratoire (Ron-
geurs, Cricetides). Mammalia, 45:415-421.

August, P. V. 1981. Population and community ecology of small mammals in northern Venezuela.
Unpublished Ph.D dissertation, Boston University, Boston, 218 pp.

94

Annals of Carnegie Museum

vol. 56

Barbosa, A. A. A., M. L. M. Nova da Costa, and A. C. Franco. 1984. Composigao floristica de
uma area de 4.2 ha da floresta de galeria do Corrego Capetinga na Fazenda Agua Limpa-Brasilia-
DF. Unpublished manuscript, Universidade de Brasilia, Brasilia, Brazil, 7 pp.

Barr, A. J., J. H. Goodnight, J. P. Sall, W. H. Blair, and D. M. Chilko. 1979. SAS —statistical
analysis system. SAS Institute, Raleigh, North Carolina.

Borchert, M., and R. L. Hansen. 1983. Effects of flooding and fire on valley side wet campo
rodents in central Brazil. Revista Brasileira de Biologia, 43:229-240.

Dietz, J. M. 1981. Ecology and social organization of the maned wolf ( Chrysocyon brachyurus).
Unpublished Ph.D. dissertation, Michigan State University, East Lansing, 185 pp.

. 1983. Notes on the natural history of some small mammals in central Brazil. Journal of

Mammalogy, 64:521-523.

Eiten, G. 1972. The Cerrado vegetation of Brazil. Botanical Review, 38:201-341.

. 1984. Vegetation of Brasilia. Phytocoenologia, 12:271-292.

Enders, R. K. 1935. Mammalian life histories from Barro Colorado Island, Panama. Bulletin of the
Museum of Comparative Zoology, Harvard University, 78:383-502.

Fonseca, G. A. B. da, and K. H. Redford. 1984. The mammals of IBGE’s ecological reserve and
an analysis of the role of gallery forests in increasing diversity. Revista Brasileira de Biologia, 44:
517-523.

Handley, C. O., Jr. 1976. Mammals of the Smithsonian Venezuelan Project. Brigham Young
University Science Bulletin, Biology Series, 20:1-91.

Heringer, E. P. 1971. Propaga?ao e successao arboreas do Cerrado em fungao do fogo, do capim,
da capina e de aldrim (inseticida). Pp. 167-179, in III Simposio sobre o Cerrado (M. G. Ferri,
ed.), Sao Paulo, Brazil, 239 pp.

Heringer, E. P., G. M. Barroso, J. A. Rizzo, and C. T. Rizzini. 1977. A flora do Cerrado. Pp.
211-232, in IV Simposio sobre o Cerrado (M. G. Ferri, ed.), Universidade de Sao Paulo, Sao
Paulo, Brazil, 405 pp.

Honacki, J. H., K. E. Kinman, and J. W. Koeppl. 1982. Mammal species of the world. The
Association of Systematics Collections, Lawrence, Kansas, 694 pp.

Hunsaker, D., II. 1 977. Ecology of New World marsupials. Pp. 95-1 56, in The biology of marsupials
(D. Hunsaker, II, ed.), Academic Press, New York, 537 pp.

Husson, A. M. 1978. Mammals of Suriname. Zoologische Monographieen van het Rijksmuseum
van Natuurlijke Historie, 2:1-569.

Joly, A. B. 1 970. Conhega a vegetagao brasileira. Editora de Universidade de Sao Paulo, Sao Paulo,
Brazil.

Karimi, Y., C. R. Almeida, and F. Petter. 1976. Note sur les rongeurs du Nord-est du Bresil.
Mammalia, 40:257-266.

Lacher, T. E., Jr., G. A. B. da Fonseca, C. Alves, Jr., and B. MagalhAes-Castro. 1984. Parasitism
of trees by marmosets in a central Brazilian gallery forest. Biotropica, 16:202-209.

Mares, M. A., K. A. Ernest, and D. Gettinger. 1986. Small mammal community structure and
composition in the Cerrado Province of central Brazil. Journal of Tropical Ecology, 2:1-13.

Mares, M. A., R. A. Ojeda, and M. P. Kosco. 1 98 lzz. Observations on the distribution and ecology
of the mammals of Salta Province, Argentina. Annals of Carnegie Museum, 50:151-206.

Mares, M. A., M. R. Willig, K. E. Streilein, and T. E. Lacher, Jr. 1981 b. The mammals of
northeastern Brazil: a preliminary assessment. Annals of Carnegie Museum, 50:80-137.

Mathews, A. G. A. 1 977. Studies on termites from the Mato Grosso state, Brazil. Academia Brasileira
de Ciencias, Rio de Janeiro, 267 pp.

Mello, D. A. 1969. Roedores silvestres de alguns municipios do estado de Pernambuco e suas
regides naturais. Revista Brasileira de Pesquisas Medicas e Biologicas, 2:360-362.

. 1 977. Observances preliminares sobre a ecologia de algumas especies de roedores do Cerrado,

municipio de Formosa, Goias, Brasil. Revista Brasileira de Pesquisas Medicas e Biologicas, 10:
39-44.

. 1 980. Estudo populacional de algumas especies de roedores do Cerrado (norte do municipio

de Formosa, Goias). Revista Brasileira de Biologia, 40:843-860.

Mello, D. A., and L. E. Moojen. 1 979. Nota sobre uma colenao de roedores e marsupiais de algumas
regioes do Cerrado do Brasil Central. Re vista Brasileira de Pesquisas Medicas e Biologicas, 12:
287-291.

Moojen, J. 1952. Os Roedores do Brasil. Institute Nacional do Livro, Biblioteca Cientifica Brasileira.
Ser. A, 2:1-214.

. 1966. Ecological notes on the Brazilian rodents involved in rural plague. Pp. 51-59, in

Seminar on rodents and rodent ectoparasites. 24-28 October 1966. World Health Organization,
Geneva, 222 pp.

1987

Nitikman and Mares— Brazilian Gallery Forest Mammals

95

Nowak, R. M., and J. L. Paradiso. 1983. Walker’s Mammals of the world. Fourth ed. The Johns
Hopkins University Press, Baltimore, Maryland, 1 and 2:1-1362.

O’Connell, M. A. 1979. Ecology of didelphid marsupials from northern Venezuela. Pp. 73-87, in
Vertebrate ecology in the northern Neotropics (J. F. Eisenberg, ed.), Smithsonian Institution Press,
Washington, D.C., 271 pp.

Patton, J. L., B. Berlin, and E. A. Berlin. 1 982. Aboriginal perspectives of a mammal community
in Amazonian Peru: knowledge and utilization patterns among the Aguaruna Jivaro. Pp. 1 1 1-
128, in Mammalian biology in South America (M. A. Mares and H. H. Genoways, eds.), Special
Publications, Pymatuning Laboratory of Ecology, University of Pittsburgh, Linesville, Pennsyl-
vania, 539 pp.

Paula, A. C. de. 1983. Relates espaciais de pequenos mamiferos em uma comunidade de mata
de galeria do Parque Nacional de Brasilia. Unpublished M.S. thesis, Universidade de Brasilia,
Brasilia, 72 pp.

Pereira, L. A. 1982. Uso ecologico do espago de Zygodontomys lasiurus (Rodentia, Cricetidae) em

habitat natural de Cerrado do Brasil Central. Unpublished M.S. thesis, Universidade de Brasilia,
Brasilia, 130 pp.

Pereira, L. A., and C. J. R. Alho. 1982. Uma tecnica de transeptos moveis em gradeado para o
estudo das populates de roedores silvestres sob condi?5es naturais. Resumos IX Congreso
Brasileira de Zoologia, Universidade de Brasilia, Brasilia, 9:257.

Pine, R. H. 1973. Mammals (exclusive of bats) of Belem, Para, Brazil. Acta Amazonica, 3:47-79.

Ratter, J. A. 1980. Notes on the vegetation of Fazenda Agua Limpa (Brasilia, D. F., Brazil). Royal
Botanical Garden, Edinburgh, United Kingdom, 111 pp.

Redford, K. H. 1984. Mammalian predation on termites: tests with the burrowing mouse {Oxy-
mycteris roberti) and its prey. Oecologia, 65:145-152.

Redford, K. H., and G. A. B. da Fonseca. In press. Zoogeography of the cerrado mammalian
fauna and the role of gallery forest. Biotropica.

Rohlf, F. J. 1982. BIOM: a package of statistical programs to accompany the text Biometry. State
Univ. New York, Stony Brook, New York.

Rohlf, F. J., J. Kishpaugh, and D. Kirk. 1979. NT-SYS: numerical taxonomy system of multi-
variate statistical programs. State Univ. New York, Stony Brook, New York.

Sarmiento, G. 1983. The savannas of tropical America. Pp. 245-288, in Ecosystems of the world,
13: tropical savannas (F. Gouliere, ed.), Elsevier Sci. Publ., Amsterdam, 730 pp.

Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy. W. H. Freeman & Co., San
Francisco, 573 pp.

Souza, M. J. 1979. Distribui$ao espacial do roedor silvestre Zygodontomys lasiurus em habitat
natural do Cerrado. Unpubl. M.S. thesis, Universidade de Brasilia, Brasilia, 91 pp.

Souza, M. J., and C. J. R. Alho. 1980. Distribuicao espacial do roedor silvestre Zygodontomys
lasiurus em habitat natural do Cerrado. Brasil Florestal, 44:31-74.

Streilein, K. E. 1982. The ecology of small mammals in the semiarid Brazilian Caatinga. III.
Reproductive biology and population ecology. Annals of Carnegie Museum, 51:251-269.

Vanzolini, P. E., and E. E. Williams. 1970. South American anoles: the geographic differentiation
and evolution of the Anolis chrysolepis species group (Sauria, Iguanidae). Arquivos do Zoologia
Sao Paulo, Brazil, 19:1-298.

Veiga-Borgeaud, T. 1982. Donnees ecologiques sur Oryzomys nigripes (Desmarest, 1819) (Ron-
geurs, Cricetides) dans le foyer natural de peste de Barragao dos Mendes (Etat de Rio de Janeiro,

Bresil). Mammalia, 46:335-359.

1

--

•

•

• ' ‘ $ J

I WM

I

INSTRUCTIONS FOR AUTHORS

ANNALS OF CARNEGIE MUSEUM consist of con-
tributions to the natural sciences and anthropology, in
30 by 46 picas format ( 1 27 by 195 mm or 5 by 75/s inches).
All manuscripts must first be submitted for review to the
Curator in charge of the appropriate scientific Section.
Authors should give particular attention to scientific con-
tent, format, and general style for the ANNALS. Manu-
scripts that do not conform to the style of the ANNALS
will be returned to the author immediately. Every manu-
script will be peer reviewed by at least two outside per-
sons. Authors will be asked to subsidize, if possible, any
or all costs of publication (approximately $ 100/page
printed).

Manuscript style. — Articles should include the follow-
ing items in this order: title page, abstract, text (with
desired headings), acknowledgments, literature cited, ta-
bles, figure captions, and copies of illustrations. All
manuscripts must be typed double-spaced on standard
8>/2 by 1 1 inch white bond paper, with at least one inch
margins all around, and submitted in triplicate— an orig-
inal for the editors, and two review copies. All pages
should be numbered, including tables, literature cited,
and the list of figure captions. Only correspondence qual-
ity or better dot matrix printouts will be accepted; draft
mode is unacceptable.

Title Page. The title, on a separate title page should
be brief, include the animal or plant group involved, and
appear two to three inches below the top margin. Include
the authors name(s) and the affiliations of non-CMNH
authors. In the case of multiple authorship indicate the
address to which proofs should be sent.

Abstracts. Abstracts should be short, double-spaced
and substantive and included at the head of the first page
of text.

Text. Do not right justify text or break (hyphenate) a
word at the end of a line. Footnotes and acknowledgments
as footnotes are unacceptable. All text is double-spaced.

Tables and figure legends. The list of figure legends
and each table should be typed consecutively on indi-
vidual pages separate from the text. Tables must be dou-
ble-spaced throughout with no vertical lines.

Abbreviations. Refer to the CBE Style Manual for cor-
rect abbreviations. Mammalian dentition: use capital let-
ters (I, C, P. M, D, for incisor, canine, premolar, molar,
deciduous, respectively) and superscript/subscript num-
bers (M1 2, P3) to designate upper and lower teeth.

Measurements. Metric units should be used, if possi-
ble.

Literature Cited. All references in text must appear in
the Literature Cited section. The data (author, date and
page) in both citations must agree. Do not abbreviate the
titles of periodicals or serials. The following style, dou-
ble-spaced, should be used in Literature Cited:

1) Two authors in an institutional series:

Matthew, W. D., and W. Granger. 1923. The
fauna of the Houldjin Gravels. American Mu-
seum of Natural History Novitates, no. 97:
1-6.

2) Same authors repeated — use three-em dash:

— . 1923. The fauna of the Ardyn Obo For-

mation. American Museum of Natural
History Novitates, no. 98:1-5.

3) Same authors plus a third author— repeat all authors:

Matthew, W. D., W. Granger, andG. G. Simpson.
1 928. Paleocene multituberculates from Mon-
golia. American Museum of Natural History
Novitates, no. 331:1-4.

4) Chapter in an edited volume:

Rausch, R. L. 1963. A review of the distribution
of Holarctic mammals. Pp. 29-43, in Pacific
basin biography (J. L. Gressitt, ed.), Bishop Mu-
seum Press, Honolulu, Hawaii, xx + 450 pp.

5) Unpublished dissertation:

Smith, J. P. 1976. Review of Eocene mammals.
Unpublished Ph.D. dissert., University of Cal-
ifornia, Berkeley, 302 pp.

6) Book:

White, M. J. D. 1961. The chromosomes. Meth-
uen and Co., Ltd., London, 120 pp.

7) Journal articles with usual volume and issue number:

Anderson, W. I. 1969. Lower Mississippian con-
odonts from northern Iowa. Journal of Paleon-
tology, 43(4):9 16-928.

Illustrations. All illustrations will be called figures, and
are to be numbered in Arabic numerals. Three sets of
illustrations are required, one (original artwork) for re-
production, two for reviewers. Xerox copies of photo-
graphs for reviewers are usually not acceptable but are
adequate for line drawing review copies. All illustrations
must be reducible to a maximum of 127 by 195 mm (30
by 46 picas) without loss of clarity. Line copy should be
designed for reduction to % or V2 of actual size. Type-
written illustration copy will not be accepted. Photo-
graphic figures should be submitted at actual reproduc-
tion size.

Two or more small figures should be combined for
reproduction as a unit or plate, but, if feasible, do not
combine halftones and line drawings. Rectangular half-
tone figures should be abutted, without intervening spaces.
The printer will insert narrow white spaces during the
reproduction process. Halftone figures with solid black
backgrounds will not be accepted unless the author agrees
to pay for the printing costs. All plates must have min-
imally one inch borders all around. Each plate should be
given a protective cover and identified on the back side.

Lettering and/or a magnification scale (linear metric
scale) for rectangular halftone figures should be placed
directly on the photo, not in a blank space between pho-
tos. The scale or lettering for closely cropped photos can
be placed in blank areas close to the figure.

Proof— The author should answer all queried proof
marks and check the entire proof copy. Return corrected
page proof with the edited manuscript promptly to the
editors.

If an author chooses to make extensive alterations to
a paper in proof stage the author or appropriate section
will bear the cost. Original manuscripts will not be re-
turned unless requested. Illustrations will be returned to
the author.

issn 0097-4463

of CARNEGIE MUSEUM

THE CARNEGIE MUSEUM OF NATURAL HISTORY

4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213

VOLUME 56

15 MAY 1987

CONTENTS

ARTICLES 3-7

Art. 3. Reproduction in a Spanish population of Acanthodactylus

erythrurus (Reptilia: Lacertilia: Lacertidae)

Stephen D. Busack and Lorrie L. Klosterman 97

Art. 4. Karyotypic analysis of five rodents and a marsupial from Belize,

Central America

. . . David W. Burton, John H. Bickham, Hugh H. Genoways and

Timothy J. McCarthy 103

Art. 5. Biosystematic studies in Stenanthium (Liliaceae: Veratreae). I.

Floral morphology, floral vascular anatomy, geography and
taxonomy of S. occidentale A. Gray Frederick H. Utech 1 13

Art. 6. Results of The Carnegie Museum of Natural History Expeditions

to Belize. III. Distributional notes on the birds of Belize

D. Scott Wood and Robert C. Leberman 137

Art. 7. A review of the crane flies in the Subgenus Tipula ( Papuatipula )

(Diptera: Tipulidae), with descriptions of five new species

Chen W. Young 161

Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS:

L. Krishtalka
J. L. Carter

Office of Scientific Publications
Phone (412) 622-3287

Communications about subscriptions, purchases of individual numbers, rates,
and changes of address should be sent to:

Publications Secretary

The Carnegie Museum of Natural History

4400 Forbes Avenue

Pittsburgh, PA 15213

ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural
History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213, by the authority of the Board of
Trustees of Carnegie Institute.

mp

THE CARNEGIE

MUSEUM OF
NATURAL HISTORY

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 3, Pp. 97-102

15 May 1987

REPRODUCTION IN A SPANISH POPULATION OF
ACANTHODACTYLUS ER YTHR UR US
(REPTILIA: LACERTILIA: LACERTIDAE)

Stephen D. Busack1 2

Research Associate, Section of Amphibians and Reptiles

Lorrie L. Klosterman1 2 3

Abstract

Reproduction in male Acanthodactylus erythrurus from La Algaida, Cadiz Province, Spain, is char-
acterized by spring spermiogenesis, copulation between June and August, and post-mating spermato-
genesis that begins shortly after the testes are evacuated and continues through the fall. Vitellogenic
growth of ovarian follicles begins in the spring, prior to mating. Vitellogenesis proceeds for a period
of about one to two months, oviductal eggs are first seen in early June, and hatchlings first appear
during August. Clutch size ranges between one and five (x = 3) and larger females tend to produce
larger clutches. No direct evidence for multiple clutches was found.

Introduction

Several aspects of the ecology of Acanthodactylus erythrurus at La Algaida,
Spain, have been described (Busack, 1976; Pough and Busack, 1978; Busack and
Jaksic, 1982), but no reproductive data have been published. We undertook this
study of reproduction in the population to fill this gap in our knowledge of its
life-history, and to compare reproductive parameters with those of the Moroccan
population of this species (Bons, 1962, 1963, 1964, 1967, 1969, 1972).

Materials and Methods

Specimens were collected from La Algaida (25.5 km [AIR] NW Jerez de la Frontera, Cadiz Province,
Spain) between May 1970 and November 1971 (see Busack and Jaksic, 1982, for details). Each
individual was preserved immediately in 10% formalin and later transferred to 70% ethyl alcohol for
permanent storage. Busack and Jaksic (1982) determined that all males >61 mm snout-vent length
(SVL), and all females >57 mm SVL, in this population were adult; only males and females of these
SVLs or larger were used to assess seasonal reproductive changes.

Testes from individuals collected from March through October were prepared following procedures
described by Berg (1953), sectioned, and examined microscopically for the cellular stages of sper-
matogenesis. Length of the left testis was measured in situ with an ocular micrometer, and testis
volume in 1 40 adult males was determined by fluid displacement. Seminiferous tubule diameter was
measured in 15 adult males collected from March-October, and the diameters of up to 30 tubule
cross-sections were used to calculate average tubule diameter. We did not examine efferent ducts or
epididymides for the presence of spermatozoa. In this species spermatozoa appear to pass rapidly
through these structures (Bons, 1969), thus testis condition serves as a reliable indicator of the timing
of mating.

Ovarian follicles were measured in situ with an ocular micrometer in 95 adult females. Larger,
slightly elongated, follicles were measured along the axis of elongation, and follicles in which the

1 Order of authorship is alphabetical.

2 Address: 3711 Garvin Avenue, Richmond, CA 94805.

3 Address: Biology Department, Boston University, 2 Cummington Street, Boston, MA 02215.
Submitted 1 August 1985.

97

98

Annals of Carnegie Museum

vol. 56

diameter was >0.5 mm were recorded. The lengths of oviductal eggs were measured in situ with dial
calipers accurate to 0. 1 mm.

Estimates of clutch size are usually obtained by counting growing or preovulatory follicles present
in ovaries during the breeding season (Fitch, 1970). Degeneration of large follicles has been observed
in A. erythrurus (Bons, 1964) but atretic follicles were not conspicuous in females containing large
yolky eggs examined during this study. By assuming that all vitellogenic follicles with a diameter > 3
mm would be ovulated in a single clutch, we obtained an estimate of clutch size. Numbers of oviductal
eggs and corpora lutea provided additional clutch size data.

Spearman rank correlation procedures were used to assess the relationship between lizard SVL and
the number of developing follicles, corpora lutea, and oviductal eggs present. Resulting values of rs
were transformed according to Fisher’s z transformation procedures and the probabilities presented
are those for committing a Type I error in a two-tailed test.

Specimens examined.— The Carnegie Museum of Natural History (CM) 53245, 53282-53288, 53330,
53333, 53350, 53373, 53383, 53397, 53425, 53438, 53471, 53885-53886, 53888-53890, 53916,
54218-54220, 54223, 54225, 54256-54257, 54277, 54279, 54561, 54565, 54592, 54594, 54676,
54789, 54794, 54873, 55305-55328, 55337-55341, 55474, 55659, and 55699-55701.

Results

Male Reproductive Cycle

Changes in both testis size and spermatogenic activity indicate a reproductive
period from April to July. Testis length and volume reach a maximum between
April and June and dramatically decrease during July. A gradual increase in size
begins again in August and September and continues through October. By October,
mean testis size had not yet reached that seen in the earliest spring collections
(March), so we infer that additional testis growth took place during the winter
when adults were inactive (Busack, 1976). Fig. 1 illustrates changes in testicular
length and volume over the collecting period.

Histological examination confirmed that annual fluctuation in testis size and
changes in diameter of seminiferous tubules are correlated with spermatogenic
activity (Fig. 2). Maximal development of germinal epithelium and maximum
tubule diameter occur from May through June. Seminiferous tubules contain
several layers of primary spermatocytes in various stages of prophase, as well as
numerous spermatids and spermatozoa (Fig. 2B), and spermatozoa are found only
in tubule luminae during these months. Marked reductions in the height of the
germinal epithelium and tubule diameter coincide with decrease in testis size in
July, and seminiferous tubules often contain cellular debris (possibly the result
of recent evacuation of spermatozoa) within the lumen. In most specimens the
epithelium consists predominantly of a basal layer of spermatogonia and Sertoli
cells (Fig. 2C). In a few specimens in which the tubule lumina contained cellular
debris during June, primary spermatocytes were also abundant.

Proliferation of spermatogonia and development of primary spermatocytes oc-
curs again during August. Although some August specimens contain only sper-
matogonia and Sertoli cells, others undergo active spermatogenesis (Fig. 2D).
During the remaining months of the year, numbers of spermatogonia and sper-
matocytes present in seminiferous tubules increase. Large primary spermatocytes
typically are distributed in basal clumps during September. Although some sper-
matocytes reach metaphase and telophase at this time, most spermatocytes are
in earlier (leptotene and pachytene) stages of prophase during October. Although
spermatogenesis resumes in some individuals as early as August, there are neither
spermatozoa nor spermatids evident throughout the remainder of the year. The
earliest appearance of spermatids is in April, just preceding the breeding season
(Fig. 2A).

1987

Busack and Klosterm an — Reproduction in Acanthodactylus erythrurus

99

8

0

Fig. 1.— Seasonal changes in testis length and volume in adult male lizards. There were no statistically
significant differences (Student’s two-sample t-test) in testis length or volume between years, so monthly
samples from both years are combined. Horizontal lines indicate monthly means, ± two standard
errors are indicated by the vertical bars, and sample sizes are indicated above the bars (* = months
in which some individuals contained testicular spermatozoa, ** = months in which individuals showed
evidence for the recent evacuation of spermatozoa).

Female Reproductive Cycle

Changes in maximum follicle size (Table 1) indicate a distinct seasonality in
reproductive condition. Follicles measuring at least 2 mm in diameter are found
during all months of the year in both ovaries of all adult females, and in most
months 2 mm is the maximum size of any follicle (Table 1). Follicles larger than
this are vitellogenic, and are found in 20 of 22 females collected between late
April and early July. One to three vitellogenic follicles are distinct (2.4-4. 5 mm)
in each ovary by mid-May. By late May and June follicles markedly larger than
all others are presumably destined for ovulation. The maximum diameter attained
by a vitellogenic ovarian follicle is 9.5 mm, undoubtedly near size at ovulation.
Vitellogenic follicles are absent in ovaries containing corpora lutea.

Because vitellogenic ovarian follicles appear in some individuals as early as 1 8
April, we infer that the period of vitellogenesis is about two months in duration.

100

Annals of Carnegie Museum

vol. 56

Fig. 2. —Seasonal changes in the seminiferous tubules of adult male A. erythrurus. Lower diagram
illustrates the most differentiated stage of spermatogenesis in at least 50% of the individuals sampled.
Numbered stages indicate the presence of: 1, spermatogonia and Sertoli cells only; 2, spermatogonia
and primary spermatocytes; 3, developing spermatids in addition to the preceding cellular stages; 4,
mature or nearly mature spermatozoa. Photomicrographs illustrate the appearance of the germinal
epithelium in representative individuals collected on the dates indicated by the arrows. Scale bar
below panel A = 0.05 mm; L identifies tubule lumen.

Oviductal eggs were first seen on 19 June. Shelled oviductal eggs measuring be-
tween 11.5 and 14.8 mm were found on 19 July, and the first hatchling lizards
(SVL <31 mm) were found between 13 and 30 August during both years.

We found preovulatory follicles in 22 individuals, oviductal eggs in 6, and
corpora lutea in 7 individuals. Estimates of mean clutch size derived from these
data are 2.7 (range 1-5), 3.0 (2-4), and 2.7 (1-4), respectively. Females in Spain
lay from one to a maximum of five eggs per clutch and the average clutch is about
3 eggs.

The relationship between SVL and number of vitellogenic follicles demonstrates
probable differences in reproductive potential for lizards of different sizes. When
all follicles with a minimum diameter of >3 mm are considered presumptive
eggs, a significant relationship between SVL and clutch size is identified (z = 1.12,
N = 22, P c 0.05). The number of corpora lutea, which probably provides a
more accurate estimate of clutch size, also shows a statistically significant cor-
relation with SVL (z = 2.03, N = 7,P< 0.05). No significant correlation between
SVL and the number of oviductal eggs, however, was identified (z = 0.64, N =
6, P > 0.2).

Discussion

Reproduction in A. erythrurus at La Algaida occurs in late spring and summer.
Although we did not observe mating in the field, the reproductive condition of
both males and females indicates breeding capability between late May and early
July. Seasonal changes in testis size (length and volume) and spermatogenic ac-
tivity indicate a single copulatory period for males annually. Males begin repro-
ductive activity (spermatozoa present in seminiferous tubules) during April, and

1987

Busack and Klosterm an— Reproduction in Acanthodactylus erythrvrvs

101

Table l. —Monthly summary of follicle development and the distribution of oviductal eggs and corpora
luiea among adult female A. erythraras at La Algaida, Spain .

Number of adult females in which maximum

Number of oviductal eggs

Number of corpora lutea per individ-

developing follicle diameter (mm) was:

per individual

ual

Month

I

2

3

4

>5

2

3

4

1

2

3

4

Mar

Apr

-

4

14

1

-

-

-

-

-

-

-

-

-

May

—

2

__

4

5

—

—

_

—

—

—

Jun

_

—

4

4

3

4

1

—

2

1

I

Jul

3

5

—

1

—

1

—

—

1

__

—

1

Aug

14

6

_

_

—

—

—

—

—

—

1

_

Sep

3

6

—

—

—

— ■

—

—

—

—

—

Oct

3

11

—

—

—

—

—

—

—

—

—

—

Nov

—

2

—

—

—

—

—

—

—

—

—

__

reproductivdy capable individuals are present each month from April through
mid- June. The subsequent sharp decrease in testis size and the absence of sper-
matozoa suggest that the majority of males copulate during June and early July.
The gradual increase in overall testis size (Fig. 1), as well as the presence and
proliferation of early germinal stages (spermatogonia and primary spermatocytes;
Fig. 2), from. August to October show that spermatogenesis resumes shortly after
copulation. Ovarian follicles undergoing vitellogenic growth (>2 mm diameter)
first appear in mid- April. Between April and July all females (but two) of adult
size contained vitellogenic follicles (generally two per ovary) in both ovaries,
indicating that virtually all adult females reproduce each season.

Moroccan and Spanish male A. ery thrums show similar cycles of reproductive
maturation, evacuation of testes, and spermatogenesis. Based on changes in the
mean number of spermatids and spermatozoa in the seminiferous tubules of
individuals collected during a single breeding season, Bons (1969) suggested that
individual males in Morocco experience two waves of sperm production per
season. Histological examination does not confirm a similar phenomenon in males
from Spain. Testes from a few individuals in our study did contain, in addition
to cellular debris indicative of recent evacuation of spermatozoa, numerous pri-
mary spermatocytes in the seminiferous tubules. Whether these spermatocytes
continue to develop into a second “wave” of spermatozoa in the same season, or
simply degenerate, is unknown. The most notable difference between these two
populations concerns the period during which reproductivdy mature males are
present. This period is apparently shorter, but is followed by earlier resumption
of spermatogenesis, in Spain.

Vitellogenesis and ovulation likewise occur over the same timespan in the two
populations. Bons (1 972) suggests two phases of vitellogenic growth during a single
breeding season for larger female A. erythrurus in Morocco. Older (i.e. larger)
females produce two dutches each year and younger individuals produce only
one. Evidence for the production of two dutches during a single season can be
inferred by: (1) simultaneous presence in an ovary of vitellogenic follicles of two
distinct size classes, (2) asymmetrical vitellogenic growth of follicles between
ovaries (as in allochronic ovulation [Jones, 1978]), or (3) the presence of young
corpora lutea in addition to vitellogenic follicles. None of these conditions was
observed during our study. While we cannot rule out the possibility that second

102

Annals of Carnegie Museum

vol. 56

clutches are deposited by Spanish females, our data do not favor this hypothesis
over one which infers reproduction over an extended period.

Mature females in Morocco lay between two and six eggs in early clutches, but
second clutches contain only one to three eggs (Bons, 1962). In Spain, mature
females may produce as many as five eggs, the average clutch size is three eggs,
and there is no direct evidence for the production of multiple clutches. Thus the
Moroccan population produces more and (sometimes) larger clutches per female
than the Spanish population.

Acknowledgments

This contribution is dedicated to the memory of the late L E. Hidalgo Gibaja of Sanlucar de
Barrameda. We thank W. R. and L. T. Maxwell of Wellesley, Massachusetts for field assistance; L. J
Cullen of the Registry of Tumors in Lower Animals, and M. E. Melville of the Department of Vertebrate
Zoology, Smithsonian Institution, for preparation of histological material; G. M. Christman for as-
sistance in preparing figure 1; D. B. and M. H. Wake for reviews of earlier versions of the manuscript;
G. R. Zug for facilities; and J. A. Visnaw for assisting with data collection and recording. We are also
indebted to C. J. McCoy and E. I. Censky of The Carnegie Museum of Natural History for numerous
loans of pertinent specimens and review of the manuscript.

Literature Cited

Berg, J. W. 1953. Differential staining of spermatozoa in sections of testis. American Journal of
Clinical Pathology, 23:513-515.

Bons, N. 1962. Le cycle de ponte du reptile Lacertide: Acanthodactylus erythrurus lineo-maculatus.

Compte rendu hebdomadaire des Seances de V Academic de Sciences, Paris, 255:165-167.

- — — . 1963. Le cycle d’activite sexuelle du male du lezard Acanthodactylus erythrurus lineoma-
culatus. Compte rendu hebdomadaire des Seances de l’Academie de Sciences, Paris, 256:1021-
1023.

— — 1964. Degenerescence des follicules ovariens chez un lezard du Maroc: Acanthodactylus
erythrurus lineomaculatus. Bulletin de la Societe de Sciences naturelles et physiques du Maroc,
44:75-82.

— — . 1967. Evolution des cellules interstitielles du testicule chez un Lacertide: Acanthodactylus
erythrurus lineomaculatus. Bulletin de la Societe de Sciences naturelles et physiques du Maroc,
47:207-213.

. 1 969. Le cycle sexuel de male chez Acanthodactylus erythrurus lineomaculatus Dum. et Bibr.

(Sauria, Lacertidae). Bulletin de la Societe de Sciences naturelles et physiques du Maroc, 49:161-
204.

. 1972. Variations histophysiologiques du tractus genital femelle du lezard Acanthodactylus

erythrurus lineomaculatus Dum. & Bibr. au cours du cycle annuel. Bulletin de la Societe de
Sciences naturelles et physiques du Maroc, 52:59-120.

Busack, S. D. 1976. Activity cycles and body temperatures of Acanthodactylus erythrurus. Copeia,
1976:826-830.

Busack, S. D., and F. M. Jaksic. 1982. Autecological observations of Acanthodactylus erythrurus
(Sauria: Lacertidae) in Southern Spain. Amphibia-Reptilia, 3:237-255.

Fitch, H. S. 1970. Reproductive cycles in lizards and snakes. Miscellaneous Publications, Museum
of Natural History, University of Kansas, 52:1-247.

Jones, R. E. 1978. Ovarian cycles in non-mammalian vertebrates. Pp. 731-762, in The vertebrate
ovary (R. E. Jones, ed.), Plenum Press, New York, 853 pp.

Pough, F. H., and S. D. Busack. 1978. Metabolism and activity of the Spanish fringe-toed lizard
(Lacertidae: Acanthodactylus erythrurus). Journal of Thermal Biology, 3:203-205.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 4, Pp. 103-112

15 May 1987

KARYOTYPIC ANALYSIS OF FIVE RODENTS AND A
MARSUPIAL FROM BELIZE, CENTRAL AMERICA

David W. Burton4

John W. Bickham1

Hugh H. Genoways3 4
Research Associate, Section of Mammals

Timothy J. McCarthy2
Abstract

Karyotypes and chromosomal banding patterns of six species of small mammals from Belize are
presented and discussed in light of relevant data in the literature. Individuals of Heteromys desma-
restianus had FN values of 72, which differs from previous reports. G-bands were obtained and
compared to fluorescent bands (DAPI) and with the location of NOR sites. The karyotypes of Pero-
myscus mexicanus, Sigmodon hispidus, Oryzomys couesi, and Ototylomys phyllotis are identical to
the karyotypes for these species described in the literature. Active NORs were located on four pairs
of chromosomes in O. phyllotis and fluorescent bands obtained with the stain Hoechst 33258. The
karyotype of Marmosa robinsoni appears similar to that described by Hsu and Benirschke (1971).
Autosomal C-bands were located at the centromeres. The X had interstitial C-bands and the Y was
entirely heterochromatic. Silver stained NORs were present on five chromosomes corresponding to
chromosomal pairs 4, 5, and 6.

Introduction

The small mammal fauna of Belize has not been extensively surveyed (Hersh-
kovitz, 1951; Laurie, 1954). Kirkpatrick and Cartwright (1975) compiled a list
of the mammals known to occur in Belize, but the list must be considered in-
complete due to the limited amount of study the fauna has received. During the
spring of 1983 the Section of Mammals, The Carnegie Museum of Natural His-
tory, conducted a collecting trip to obtain specimens, primarily of bats, rodents,
and marsupials. This paper presents the findings of karyotypic analyses of six
species of mammals collected throughout Belize.

Materials and Methods

All karyotypes were obtained from bone marrow preparations made in the field by the methods
described by Patton (1967), Lee (1969), and Baker et al. (1982). Individuals to be karyotyped were
usually injected with yeast to increase the mitotic index according to the methods of Lee and Elder
(1979). Several slides were prepared in the field, with the remaining cell suspension frozen in liquid
nitrogen to permit slide preparation in the lab. C-banding was performed on air dried slides by a
modification of the technique of Stefos and Arrighi (1971). G-banding was accomplished using the

1 Address: Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station,
TX 77843-2258.

2 Department of Mammalogy, The American Museum of Natural History, Central Park West at 79th
Street, New York, NY 10024.

3 Address: University of Nebraska State Museum, Lincoln, NE 68588.

4 Address: Department of Zoology, Brigham Young University, Provo, Utah 84602.

Submitted 30 May 1986.

103

104

Annals of Carnegie Museum

vol. 56

technique of Seabright (1971). Nucleolus organizer regions (NORs) were located by the silver staining
method of Gold and Ellison (1983). Fluorescent banding was done with the stains 4,6-diamidino-2-
phenylindole (DAPI) and Hoechst 33258 as described by Disteche and Bontemps (1 974), and Schweizer
(1976), respectively. Fluorescent banding permits the identification of homologous chromosomes and
identifies regions with specific DNA base composition. Hoechst 33258 and DAPI are specific for AT
(adenine and thymine) rich regions and produce a banding pattern analogous to G-bands.

Specimens Examined

All specimens were collected in Belize, Central America, and deposited in The Carnegie Museum
of Natural History (CM).

Heteromys desmarestianus. —CM 91973, female, Stann Creek District, 7.0 km WNW Quam Bank,
Cockscomb Basin; CM 91997, female, Toledo District, 1.8 km NNW Forestry Camp (Salamanca),
Columbia Forest Reserve; CM 91988, male, Stann Creek District, 3.4 km WNW Quam Bank, Cocks-
comb Basin.

Ototylomys phyllotis. — CM 92020, male, CM 92014, female, Belize District, Altun Ha ruins; CM
92034, female, Toledo District, 0.7 km NNE Forestry Camp (Salamanca), Columbia Forest Reserve.

Sigmodon hispidus.~C M92082, female, Stann Creek District, Quam Bank, Airstrip, Cockscomb
Basin.

Oryzomys couesi.—C M92005, female, Stann Creek District, Cabbage Haul Ridge, 2.8 km SSE Quam
Bank, Cockscomb Basin.

Peromyscus mexicanus. — CM 92039, female, and CM 92037, male, Toledo District, Jimmy Cut
2.1 km NW Forestry Camp (Salamanca), Columbia Forest Reserve.

Marmosa robinsoni.—CM9\62>^, male, Toledo District, 2.1 km N Forestry Camp (Salamanca),
Columbia Forest Reserve.

Results and Discussion
Heteromys desmarestianus (2n = 60, FN = 72)

The karyotype presented in Fig. la shows chromosomes 1, 2, 4, and 5 are large
and subtelocentric and chromosome 3 is large and submetacentric. There are 22
acrocentric pairs that grade in size from large to small, and two pairs of small
metacentric autosomes. The X chromosome is medium-sized and metacentric.
Fig. lb and lc display the G-bands and fluorescent bands obtained with DAPI,
respectively. It appears that the centromeric regions are not A-T rich and that
dark G-bands correspond to bright regions with DAPI as expected. Fig. Id shows
active NOR sites on at least six pairs of autosomes. Genoways (1973) and Eng-
strom et al. (in press) report 2n = 60, FN = 82 and 2n = 60, FN = 66, respectively,
for H . desmarestianus . Thus, variation in fundamental number is extensive within
this species.

Ototylomys phyllotis (2n = 48, FN = 78)

The big-eared climbing rat is found throughout the Yucatan Peninsula and
south through Belize, Guatamala, Honduras, El Salvador, Nicaragua, and the
northern half of Costa Rica (Hall, 1981). Lawlor (1969) recognized three subspe-
cies, which are separated altitudinally. Ototylomys is most closely related mor-
phologically to Tylomys and is allied to Neotoma by similarities of the accessory
reproductive glands (Lawlor, 1969). The nondifferentially stained karyotype, pre-
sented in Fig. 2a, contains one pair of large submetacentric, three pairs of large
metacentric, one pair of large acrocentric, one pair of medium-sized submeta-
centric, one pair of medium-sized acrocentric, five pairs of medium-sized meta-
centric, six pairs of small metacentric and five pairs of small acrocentric auto-
somes. The X and Y are large and metacentric. The karyotype of O. phyllotis
reported here is identical to the nondifferentially stained karyotype originally
reported by Hsu and Benirschke (1973). Fig. 2b shows a spread stained with the

1987

Burton et al. — Karyotypic Analysis of Belize Mammals

105

MM 90 ARflO MflUo no to

la ** •• Oa ** id

a

x x

» i i

5

lb

iS ii !S

**

*

■ i

*!*

ii 1 R I* m n

it ?i «i ??** |i

«* 1

X X

Fig. 1.— a) Nondifferentially stained karyotype of Heteromys desmarestianus. b) G-banded karyotype,
c) DAPI stained karyotype, d) Silver-stained NOR sites.

fluorescent dye Hoechst 33258. The banding pattern indicates that much of the
X chromosome is AT rich but not the centromere. Active NORs (Fig. 2c) are
present on one pair of small acrocentric and three pairs of medium-sized meta-
centric autosomes.

106

Annals of Carnegie Museum

vol. 56

*

u f A t X x x s x »s
2a

Ik IS II 11 XX t A

US IN »«

X X

LEE mo EBbed

i'i.

it

t#

2c

^ *

c

r # ri m

I'f f!« •» ff ** ft* '#* ffS-

x x

• m r

2d

Ifllkm si iii of tt *•

AIM SiS s'S HI

i s | ■' •

X X

? *

iipasoi

we ‘>f* 1« ni<l M ftc *« « ' I

2©

X Y

Fig. 2.— a) Nondifferentially stained karyotype of Ototylomys phyllotis. b) Karyotype stained with
Hoechst 33258. c) Silver stained NOR sites. O. phyllotis. d) Nondifferentially stained karyotype of
Sigmodon hispidus . e) Nondifferentially stained karyotype of Oryzomys palustris.

Sigmodon hispidus (2n = 52, FN = 52)

The karyotype presented in Fig. 2d contains 48 acrocentric and two very small
metacentric autosomes. The X chromosome is large and subtelocentric and the
Y small and metacentric. This karyotype is identical to that described by Zim-

1987

Burton et al. — Karyotypic Analysis of Belize Mammals

107

merman (1970) and Kiblisky (1969) who stated this karyotype is found from the
eastern United States to Venezuela.

Elder (1980) studied chromosomal evolution in the genus Sigmodon and found
that two morphologically similar but karyotypically different species, S. arizonae
(2n = 22, 24, FN = 38) and S. mascotensis (2n = 28, FN = 29), share a high
degree of G-band homology. Chromosomal evolution in the genus has been toward
lower diploid numbers by the rearrangement events of tandem and centric fusion.

Oryzomys couesi (2n = 56, FN = 56)

The karyotype presented in Fig. 2e is identical to that reported by Benson and
Gehlbach (1979) and Haiduk et al. (1979) for O. couesi. There are 26 pairs of
acrocentric and one small pair of metacentric autosomes. The X is large and
submetacentric and the Y, as reported by Haiduk et al. (1979), is small and
subtelocentric.

Honacki et al. (1982) list over 50 species of Oryzomys ranging from South
America to the eastern United States making this one of the most complex mam-
malian genera. Diploid numbers within the genus range from 52 to 80 and possess
FNs from 62 to 1 12 (Gardner and Patton, 1976). Haiduk et al. (1979) observed
extensive G-band differences between O. melanotis and O. palustris, which possess
very similar nondifferentially stained karyotypes, and concluded that karyotypic
evolution in Oryzomys is very rapid compared to other vertebrates. The analysis
of G-banding patterns by Baker et al. (1983) found that current models of chro-
mosomal evolution do not adequately explain the variation found within and
between the peromyscine-neotomine-oryzomine lineages.

Peromyscus mexicanus (2n = 48, FN = 58)

Chromosomal banding patterns have been studied for 12 species of the Pero-
myscus mexicanus group (Robbins and Baker, 1981; Rogers et al., 1984; Stangl
and Baker, 1984; Smith et al., 1986). These species possess a high degree of
karyotypic conservatism, which is unusual for the genus. The karyotypes of mem-
bers of the P. mexicanus group are characterized by having 2n = 48 and FN =
58 with biarmed chromosomes 1-3, 9, 22, and 23. Seven members of the P.
mexicanus group have been described as having heterochromatin restricted to
the centrometric regions (Huckaby, 1980; Rogers et al., 1984; Stangl and Baker,
1984).

Smith et al. (1986) state that there is a considerable amount of variation in the
morphology of the X chromosome within the P. mexicanus group. They list seven
species with submetacentric and two species with nearly acrocentric X chromo-
somes, and state that the differences are caused by heterochromatic short arm
variations. The Y chromosome in members of the P. mexicanus group are either
biarmed or acrocentric (Smith et al., 1986). Rogers et al. (1984) and Smith et al.
(1986) conclude that the P. mexicanus karyotype does not represent the primitive
autosomal karyotype for the genus, but is an intermediate form that has stabilized
relative to more highly derived karyotypes, possibly due to selective forces.

The karyotype of P. mexicanus presented in Fig. 3a has the autosomal com-
plement typical of P. mexicanus, but the Y chromosome seems to be considerably
smaller. The G-band pattern (Fig. 3b) of the autosomal complement is generally
similar to that of other members of P. mexicanus. This includes biarmed auto-
somes 1-3, 9, 22, and 23 and acrocentric autosomes 4-8 and 10-21. The G-band

108

Annals of Carnegie Museum

vol. 56

tt;§

i) iff fit Ii a§ ii in

%<s k *

pi

§

X Y

3a

„ *§ «f , £

«-u ** -s ^ t?.

I

*

% .

i:; ii

<#

an

1 III ;

* ■■

^

|

|pf§

Hi* "I- I*

m **

%|'*5 * n

'

Ml

** vp *'*

♦ 1 jt

>*• *•

X

3b

i

ED DDEtf GQDD

fin QQ £Q DO DE3 DO CD DG □□ ^

X Y

HIM

3c

Fig. 3.— a) Nondifferentially stained karyotype of Peromyscus mexicanus. b) G-banded karyotype of
P. mexicanus, homolog to chromosome five missing, c) DAPI stained karyotype of P. mexicanus.

chromosomes (Fig. 3b) are placed in numerical order (decreasing size) and iden-
tified according to the nomenclature of Committee (1977).

The banding pattern obtained with the fluorescent stain Hoechst 33258 is shown
in Fig. 3c. This stain indicates that the centromeric heterochromatin in P. mex-
icanus is not particularly AT rich. Chromosome 1 contains bright regions at the
end of both the long and short arms and in the area directly below the centromeric
region. Chromosomes 2 and 3 also contain bright interstitial bands. Chromosomes
4 through 1 2 often have bright regions at the ends of the long arms and may also
have bright centromeric bands. The small acrocentric chromosomes of P. mex-
icanus appear to be bright primarily in their centromeric regions, and dark distally.

Marmosa rohinsoni (2n = 14 , FN — 24)

There is considerable information on the nonbanded karyotypes of marsupials
(see Hayman’s 1977 review), however, there are only a few papers (Curcuru-

1987

Burton et al.— Karyotypic Analysis of Belize Mammals

109

Giordano et al, 1974; Femandes-Donoso et al., 1979; Merry et al., 1983; Seluja
et al., 1984; Sinha and Kakati, 1976; Sinha et al., 1972; Yonenaga-Yassuda et
al., 1982) that present banded karyotypes. The reason that more species have not
been examined is that it is difficult to obtain quality preparations with marsupial
chromosomes (Seluja et al., 1984). Marsupials have a slow rate of karyotypic
evolution. Previous studies have revealed only three different diploid numbers,
2n = 14, 18, 22 (Hayman, 1977). Sharman (1961) hypothesized that the evolution
of the Australian marsupials may have been associated with reduction in the
number of chromosomes. This hypothesis also has been considered to apply to
the evolution of American marsupials (Sharman, 1973, 1974; Reig et al, 1977).

The karyotypes of the 2n = 14 American didelphids are all very similar with
only minor differences between species representing different subfamilies (Reig et
al., 1977). Hayman (1977) felt that the possibility existed that the standard non-
banded karyotype would not separate some species and that banding techniques
would be required for karyotypic identification.

The murine opossums (. Marmoset ) are the most diverse Neotropical marsupials,
but of the 37 recognized species almost nothing is known about them karyo typ-
ically (Reig, 1970). There are only three “mouse” opossums known to occur in
the country of Belize (Kirkpatrick and Cartwright, 1975), these are Marmosa
alstoni, M. robinsoni, and M. mexicanus. Curcuru-Giordano et al. (1974) obtained
Q-bands on M. mitis, a junior synonym for M. robinsoni (O’Connell, 1983) and
were able to identify every homologous pair of chromosomes. Femandes-Donoso
et al. (1979) reported the location of NORs in M. elegans’, no other member of
the genus Marmosa has been studied for banded karyotypes.

Nondifferentially stained karyotypes have been obtained from eight species of
Marmosa and three species of Caluromys (Reig, 1968; Reig et al., 1977; Biggers
et al., 1965; Hsu and Benirschke, 1971). All were found to have 2n = 14 and
FN = 24. Nearly all have the same proportions (1:3:2) of metacentric, sub m eta-
centric, and subtelocentric chromosomes, with the only two exceptions being M.
elegans with only one subtelocentric and a single acrocentric, and M. fuscata with
four submetacentric and no subtelocentric chromosomes (Reig et al., 1977). Reig
(1970) pointed out that the similarities are noteworthy because Marmosa and
Caluromys are considered to be in different subfamilies of the didelphids. He
stated that the differences between the karyotypes of other genera in the Didel-
phinae and Marmosa are greater than the differences between Caluromys and
Marmosa and that the first four pairs of autosomes of M. fuscata , M. robinsoni
and C. derbianus are indistinguishable.

The morphology of the X chromosome ranges from acrocentric in C. derbianus,
C. lanatus, and M. alstoni, to subtelocentric in C. philander, M. elegans, and M.
robinsoni, to metacentric in M. murina, M. pusilla, and M. cinerea. The Y chro-
mosome is acrocentric in all these species (Reig et al., 1977).

In this paper we report on the Ag-NOR sites and C-bands of M. robinsoni. The
nondifferentially stained karyotype of our specimen is the same as that described
by Hsu and Benirschke (1971). C-bands are presented in Fig. 4a and show het-
erochromatin in the autosomes to be confined to the centromeric regions. The Y
chromosome is entirely heterochromatic, and the X appears to have two interstitial
C-bands. NORs are shown in Fig. 4b and are present on chromosomes 4, 5, and
6.

Centromeric heterochromatin has also been found in the autosomes of Philander
opossum, Didelphis albiventris (Yonenaga-Yassuda et al., 1982; Seluja et al.,
1984), and Monodelphis domestica (Merry et al., 1983). The X chromosome of P.

110

Annals of Carnegie Museum

vol. 56

mJ vJL

Ac

j> \/

L

\*LtX

* j

4a

' 4b

Fig. 4.— a) C-banded spread of Marmosa robinsoni. b) Silver stained NOR sites (arrows).

opossum has a single band in the long arm and a completely heterochromatic Y
chromosome. In Didelphis albiventris the X chromosome has a centromeric band.
Both D. marsupialis and Lutreolina crassicaudata lack centromeric heterochro-
matin on the autosomes but possess C-band positive material on the X chro-
mosomes. The Y chromosomes of both these species are totally heterochromatic
(Seluja et al., 1984).

Nucleolar organizer regions (NORs) have been studied in only one other mem-
ber of the genus Marmosa. In M. elegans NORs were located at secondary con-
strictions on the short arm of a single autosomal pair (Femandes-Donoso et al,
1979). A few studies on the distribution of NORs have been published for several
other species of marsupials. Seluja et al. (1984) found NORs located on chro-
mosome pairs 4, 5, and 6 in Didelphis albiventris and on the short arm of chro-
mosome pair 5 and the long arm of chromosome pair 7 in Lutreolina crassicau-
data. In D. marsupialis the NORs were terminally located in the long arms of
two pairs and in the short arm of one other pair. The NOR bearing chromosomes
could not be specificially identified, but there were specimens with up to eight
active NORs and as few as four (Y onenaga- Y assuda et al., 1982). In P. opossum
the NORs were located terminally in the short arm of pair 5 and the long arm of
pair 7 (Y onenaga- Y assuda et al., 1982). NOR bearing sex chromosomes have
been found in Potorous tridactylis (Goodpasture and Bloom, 1975), Lagorchestes
conspicillatus (Hayman and Sharp, 1981), Trichosurus vulpecula (Murray, 1977),
and in Monodelphis domestica (Merry et al., 1983).

Acknowledgments

We thank Mr. Michael Reed, Dr. Alan Rabinowitz (New York Zoological Society) and members
of the Ornithological and Herpetological sections of The Carnegie Museum of Natural History for aid
in fieldwork, the staff of the Belize Zoo, Mr. Herman Pastor, and Mr. D. Owen -Lewis for lodging,
and Mr. Oscar Rosado, Principal Forestry Officer, Ministry of Natural Resources, Belize, for collecting
permits. Funded in part by NSF Grant PCM-8202794, the R. K. Mellon Institute of North American
Mammal Research, and O’Neil field funds, The Carnegie Museum of Natural History. D. S. Rogers
and D. J. Schmidly offered helpful comments and discussion regarding the manuscript.

1987

Burton et al. — Karyotypic Analysis of Belize Mammals

111

Literature Cited

Baker, R. J., M. W. Haiduk, L. W. Robbins, A. Cadena, and B. G. Koop. 1982. Chromosomal
studies of American bats and their systematic implications. Pp. 303-327, in Mammalian biology
in South America (M. A. Mares and H. H. Genoways, eds.). Special Publication Series, Py-
matuning Laboratory of Ecology, University of Pittsburgh, 6:303-327.

Baker, R. J., B. F. Koop, and M. W. Haiduk. 1983. Resolving systematic relationships with
G-bands: a study of five genera of South American cricetine rodents. Systematic Zoology, 32:
403-416.

Benson, D. L., and F. R. Gehlbach. 1979. Ecological and taxonomic notes on the rice rat ( Oryzomys
couesi ) in Texas. Journal of Mammalogy, 60:225-228.

Biggers, J. D., H. I. Fritz, W. C. D. Hare, and R. A. McFeely. 1965. Chromosomes of American
marsupials. Science, 148:1602-1603.

Committee for the Standardization of Chromosomes of Peromyscus. 1977. Standardized
karyotype of the deer mouse, Peromyscus (Rodentia). Cytogenetics and Cell Genetics, 19:38-43.

Curcuru-Giordano, F. M., R. G. Weed, and E. C. Jenkins. 1974. Banding analysis of the chro-
mosomes of Marmosa mitis (Murine Opossum). Canadian Journal of Genetics and Cytology, 16:
31-38.

Disteche, C, and J. Bontemps. 1974. Chromosome regions containing DNAs of known base pair
composition, specifically evidenced by 2,7 -di-t-butyl-proflavine. Chromosoma, 36:263-281.

Elder, F. F. B. 1980. Tandem fusion, centric fusion, and chromosomal evolution in the cotton rats,
genus Sigmodon. Cytogenetics and Cell Genetics, 26:199-210.

Engstrom, M. D., H. H. Genoways, and P. K. Tucker. In press. Morphological variation, kary-
ology, and systematic relations of Heteromys gaumeri (Rodentia: Heteromyidae). In Studies in
Neotropical mammalogy: essays in honor of Phillip Hershkovitz (B. D. Patterson and R. M.
Timm, eds.), Fieldiana: Zoology, new series.

Fernandes-Donoso, R., S. Barrios, and J. Pincherira. 1979. Position of the nucleolus within the
nuclei of the pachytene spermatocytes of Dromiciops australis and Marmosa elegans (Didel-
phoidea: Marsupialia). Experientia, 35:1021-1023.

Gardner, A. L., and J. L. Patton. 1976. Karyotypic variation in Oryzomyine rodents (Cricetinae)
with comments on chromosomal evolution in the Neotropical cricetine complex. Occasional
Papers of the Museum of Zoology, Louisiana State University, 49:1-48.

Genoways, H. H. 1973. Systematic and evolutionary relationship of the spiny pocket mice, genus
Liomys. Special Publications, The Museum, Texas Tech University, 5:1-368.

Gold, J. R., and J. R. Ellison. 1983. Silver staining for nucleolar organizing regions of vertebrate
chromosomes. Stain Technology, 58:51-55.

Goodpasture, C, and S. E. Bloom. 1 975. Visualization of nucleolar organizer regions in mammalian
chromosomes using silver staining. Chromosoma, 53:37-50.

Haiduk, M. W., J. W. Bickham, and D. J. Schmidly. 1979. Karyotypes of six species of Oryzomys
from Mexico and Central America. Journal of Mammalogy, 60:610-615.

Hall, E. R. 1981. The mammals of North America. John Wiley and Sons, New York, second ed.,
1:1-606 + 90.

Hayman, D. L. 1977. Chromosome number— constancy and variation. Pp. 27-48, in The biology
of marsupials (B. Stone and D. Gilmore, eds.), MacMillan Press Ltd., 486 pp.

Hayman, D., and P. Sharp. 1981. Verification of the structure of the complex sex chromosome
system in Lagorchestes conspicillatus Gould (Marsupialia: Mammalia). Chromosoma, 83:263-
274.

Hershkovitz, P. 1951. Mammals from British Honduras, Mexico, Jamaica and Haiti. Fieldiana
Zoology, 31:547-569.

Honacki, J. H., K. E. Kinman, and J. W. Koeppl. 1982. Mammal species of the world. The
Association of Systematics Collections, Lawrence, Kansas, 694 pp.

Hsu, T. C, and K. Benirschke. 1971. An atlas of mammalian chromosomes. Springer- Verlag, New
York, 5:folio 202.

. 1973. An atlas of mammalian chromosomes. Springer- Verlag, New York, 7:folio 317.

Huckaby, D. G. 1980. Species limits in the Peromyscus mexicanus group (Mammalia: Rodentia:
Muroidea). Contributions in Science, Natural History Museum, Los Angeles County, 326:1-24.

Kiblisky, P. 1 969. The chromosomes of the hispid cotton rat, Sigmodon hispidus, from two localities
in Venezuela. Journal of Mammalogy, 50:810-81 1.

Kirkpatrick, R. D., and A. M. Cartwright. 1975. List of mammals known to occur in Belize.
Biotropica, 7:139-140.

Laurie, E. M. O. 1954. Rodents from British Honduras, Mexico, Trinidad, Haiti and Jamaica
collected by I. T. Sanderson. Annals and Magazine of Natural History, 12th series, 6:382-394.

112

Annals of Carnegie Museum

vol. 56

Lawlor, T. E. 1969. A systematic study of the rodent genus Ototylomys. Journal of Mammalogy,
50:28=42.

Lee, M. R. 1 969. A widely applicable technic for direct processing of bone marrow for chromosomes
of vertebrates. Stain Technology, 44:155=158.

Lee, M. R., and F. F. B. Elder. 1979. Yeast stimulation of bone marrow mitosis for cytogenetic
investigation. Cytogenetics and Cell Genetics, 26:36=40.

Merry, D. E., S. Pathak, and J. L. VandeBerg. 1983. Differential NOR activities in somatic and
germ cells of Monodelphis domestica (Marsupialia, Mammalia). Cytogenetics and Cell Genetics,
35:244-251.

Murray, J. D. 1977. Nonrandom sex-chromosome association and constitutive heterochromatin
in the brush-tailed opossum Trichosurus vulpecula (Marsupialia: Phalangeridae). Cytogenetics and
Cell Genetics, 18:90=96.

O’Connell, M. A. 1983. Marmosa robinsoni. Mammalian Species, 203:1=6.

Patton, J. L. 1967. Chromosome studies of certain pocket mice, genus Pergonathus (Rodentia:
Heteromyidae). Journal of Mammalogy, 48:27=37.

Reig, O. A. 1968. The chromosomes of the didelphid marsupial Marmosa robinsoni Bangs. Exper-
ientia, 26:185=186.

. 1970. The chromosomes of Marmosa fuscata Thomas, from northern Venezuela (Marsu-
pialia, Didelphidae). Experientia, 26:199-201.

Reig, O. A., A. L. Gardner, N. O. Bianchi, and J. L. Patton. 1977. The chromosomes of the
Didelphidae (Marsupialia) and their evolutionary significance. Biological Journal of the Linnaean
Society, 9:191-216.

Robbins, L. W., and R. J. Baker. 1981. As assessment of the nature of chromosomal rearrangements
in 18 species of Peromyscus (Rodentia: Cricetidae). Cytogenetics and Cell Genetics, 31:194=202.

Rogers, D. S., I. F. Greenbaum, S. J. Gunn, and M. D. Engstrom. 1984. Cytosystematic value
of chromosomal inversion data in the genus Peromyscus (Rodentia: Cricetidae). Journal of Mam-
malogy, 65:457-465.

Schweizer, D. 1976. Reverse fluorescent chromosome banding with chromomycin and DAPI.
Chromosoma, 58:307-324.

Seabright, M. 1971. A rapid banding technique for human chromosomes. Lancet, 2:971-972.

Seluja, G. A., M. V. DiTomaso, N. Brum-Zorrilla, and H. Cardoso. 1984. Low karyotypic
variation in two didelphids (Marsupialia): karyogram and chromosome banding analysis. Journal
of Mammalogy, 65:702-707.

Sharman, G. B. 1961. The mitotic chromosomes of marsupials and their bearing on taxonomy and
phylogeny. Australian Journal of Zoology, 9:38-60.

. 1973. The chromosomes of non-eutherian mammals. Pp. 485-530, in Cytotaxonomy and

vertebrate evolution (A. B. Chiarelli and E. Capanna, eds.), Academic Press, London, 783 pp.

. 1974. Marsupial taxonomy and phylogeny. Australian Journal of Mammalogy, 1:137-154.

Sinha, A. K., and S. Kakati. 1976. C- and G-bands of the opossum chromosomes: terminal
sequences of DNA replication. Canadian Journal of Genetics and Cytology, 18:195-205.

Sinha, A. K., S. Kakati, and S. Pathak. 1972. Exclusive localization of C-bands within opossum
sex chromosomes. Experimental Cell Research, 75:265-267.

Smith, S. A., R. D. Bradley, and I. F. Greenbaum. 1 986. Karyotypic conservation in the Peromyscus
mexicanus group. Journal of Mammalogy, 67:584-586.

Stangl, F. B., and R. J. Baker. 1984. Evolutionary relationships in Peromyscus: congruence in
chromosomal, genic, and classical data sets. Journal of Mammalogy, 65:643-654.

Stefos, K., and F. E. Arrighi. 1971. Heterochromatic nature of W chromosomes in birds. Exper-
imental Cell Research, 68:228-231.

Y onenaga- Y assuda, Y„, S. Kasahara, M. J. Souza, and M. L. Abbate. 1982. Constitutive het-
erochromatin, G-bands and nucleolus-organizer regions in four species of Didelphidae (Marsu-
pialia). Genetica, 58:71-77.

Zimmerman, E. G. 1970. Karyology, systematics and chromosomal evolution in the rodent genus
Sigmodon. Publications of the Museum, Michigan State University, Biology Series, 4:385-454.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 5, Pp. 1 13-135

15 May 1987

BIOS Y STEM ATIC STUDIES IN STENANTHIUM
(LILIACEAE: VERATREAE).

I. FLORAL MORPHOLOGY, FLORAL VASCULAR ANATOMY,
GEOGRAPHY AND TAXONOMY OF S. OCCIDENTALE A. GRAY

Frederick H. Utech

Associate Curator, Section of Botany

Abstract

The floral morphology, continuous pedicel to stigma floral vascularization, distribution and lecto-
typification of Stenanthium occidentale, a western North American bulbous lily, are presented. The
perfect, protandrous flowers with dimorphic filaments and bilobed or v-shaped tepal nectaries which
promote outbreeding are arranged in andromonoecious racemes. Raphide idioblasts were not observed.
Total floral vascularization is derived from three lower pedicel bundles. A spiral series of divisions
and fusions in the upper pedicel produces three compound outer tepal and three compound inner
tepal bundles. From the former, an outer tepal median, outer tepal laterals, an outer stamen and a
dorsal bundle result, while from the latter, an inner tepal median, inner tepal laterals and an inner
stamen bundle result. Their formation is above that at which the ventral and ovular supplies are
established due to an inferior gynoecium with a central carpellary hole and epigynous perianth. The
ventral supply is spirally derived from continuing bundles following the formation of the compound
inner tepal bundles. Within each septal arm, a shared septal axial is associated with but remains free
of a compound ventral that divides into two ventrals. The bitegmic, basitropic, campylotropous ovules
develop into flat, oblong, winged seeds.

Introduction

This is the first of several papers which will discuss the biosystematics of the
genus Stenanthium Kunth (1842), a conserved generic name (Farr et al., 1979;
Voss, 1983). The four currently recognized species of Stenanthium, which are S.
occidentale A. Gray, S.frigidum (Schlecht. & Cham.) Kunth, S. gramineum (Ker-
Gawl) Kunth (including S. robustum S. Watson) and S. sachalinense F. Schmidt,
have traditionally and consistently been placed in the Liliaceae, subfamily Me-
lanthioideae, tribe Veratreae (Bentham and Hooker, 1883; Engler, 1889; Gates,
1918; Krause, 1930; Zimmerman, 1958; Kupchan et al., 1961; Melchior, 1964;
Hutchinson, 1973; Schulze, 1978; Dahlgren et al., 1985).

This paper presents the floral morphology, floral vascular anatomy, distribution
and taxonomy of S. occidentale. Several common names have been used for this
species— “western Stenanthium” (Carter and Newcombe, 1921; Abrams, 1940;
Peck, 1941; Brockman, 1947; Hitchcock and Cronquist, 1973; Ferlatte, 1974),
“bronze bells” (Carter and Newcombe, 1921; Moss, 1959; Larrison et al., 1974;
Clark, 1976; Kuijt, 1982), and “mountain bells” (St. John and Hardin, 1929;
Clark, 1976; Taylor and MacBryde, 1977; Scoggan, 1978), in reference to its
western location and greenish brown, narrowly campanulate flowers.

Materials and Methods

Three different populations of Stenanthium occidentale were studied in western North America—
Oregon: Lane County, Willamette National Forest, Walker Creek, Lowder Mountain, 5 July 1978,

Submitted 26 October 1986.

113

114

Annals of Carnegie Museum

vol. 56

Fig. 1 . — Distribution and elevation of Stenanthium occidentale A. Gray based on herbarium specimens.
Arrows indicate regions where the range approaches sea level. Flowering frequency is summarized by
week and month.

Utech 78-275 (CM); Washington: Lewis County, Mt. Rainier National Park, Silver Falls, near Stevens
Canyon, 7 July 1977, Utech 77-201 (CM); Washington: Whatcom County, Mt. Baker National Forest,
Austin Pass, 11 July 1978, Utech 78-300 (CM). Herbarium collections examined are cited in the
taxonomy review.

Flowering inflorescences of varying age collected from the three study sites were fixed in acetic-
ethanol (1:3) for 10 hours with subsequent storage in 70% ethanol. Standardized transverse and
longitudinal paraffin sectioning (14-16 /x) and staining (saffarin-methylene blue) techniques (Johansen,
1940; Sass, 1958) were used on floral samples of varying stages from all three populations. As an
additional check on these sections, whole flowers were cleared and stained in a NaOH-1% fuchsin
mixture (Fuchs, 1963).

Distribution based on herbarium specimens (Fig. 1) and various aspects of floral morphometries
(Fig. 3, 4) are presented. Composite photomicrographs (Fig. 5-9) present the vascular floral anatomy
and morphology of S. occidentale, whereas Fig. 10 is a summary floral vascularization diagram. No
teleological implications are intended in the descriptive ascent and departure of the various floral
bundles and traces which are letter-coded for ease in comparison. This coding parallels that used in
our previous liliaceous studies (Utech, 1978*2, 1978/?, 1978c, 1978c/, 1978c, 1979*2, 1979 b, 1982,
1984, 1986; Utech and Kawano, 1975, 1976, 1980, 1981), but it does not imply homology.

Observations

Morphology and Floral Vascular Anatomy

Stenanthium occidentale is a glabrous perennial with narrowly ovoid, tunicated
bulbs (1 .0) 1 .5-3.5 cm long, and 2-4 mostly basal, linear to narrowly oblanceolate

1987

Utech— Stenanthium occidentale

115

Table 1 . — Degree of inflorescence branching in Stenanthium occidentale based on herbarium collections

and field study plots (2x2 meter).

Material

Branch points within inflorescence

0

1

2

3

Total

N

(%)

N

(%)

N

(%)

N

(%)

N

(%)

Herbarium collections

796

(75.4)

219

(20.7)

39

(3.7)

2

(0.2)

1056

(100)

Utech 78-275

152

(73.1)

40

(19.2)

16

(7.7)

0

208

(100)

Utech 77-201

135

(77.1)

33

(18.9)

7

(4.0)

0

175

(100)

Utech 78-300

210

(71.4)

62

(21.1)

21

(7.2)

1

(0.3)

294

(100)

leaves (10) 1 5-30 cm long, 5-20 (25) mm wide. The basal leaves are lax, gradually
narrowed at both ends, and sheathing at their bases. The weak, erect and usually
simple stems measure (2.5) 3. 0-4. 5 (6.0) dm in length. The few stem leaves are
progressively reduced apically.

The floral portion of an unbranched inflorescence typically measures 1. 0-2.0
dm long. Occasionally on larger plants, there are floral branches at the lower
nodes. When a second, rarely a third, branch occurs, a simple panicle or compound
raceme results (Fig. 3, 11; Table 1). Each pedicel irrespective of the degree of
branching is associated with a scarious lanceolate bract which measures 6-10 mm
at the lower nodes and is progressively shorter upwards (Fig. 2). The total number
of flowers per plant varies from 7 to 30 depending on the degree of branching
(Fig. 3; Table 1). Without branching, the average flower number is 10 in a simple
raceme, and with one branch point, the average increases to 14 flowers in a simple
panicle. The typical raceme is weakly andromonoecious due to a declining array
of protandrous flowers. Beginning at the base, most flowers do set fruit. Floral
phenology lasts for a relatively long time period, based on the number of flowers
present and their flowering sequence. From both field observations and herbarium
collections, the average plant size appears to be directly proportional to the total
number of flowers, seeds, or fruits produced (Fig. 3).

The flowering pedicels are normally 7-17 mm long with pendulous flowers (Fig.
2A-C, 4). In fruit, on the other hand, the pedicels are elongated and ascending
with upright fruit (Fig. 2D, 4). Flowering and fruiting pedicels have the same
number of bundles and vascular arrangement and differ only in that the latter is
longer and has a sclerenchymatous sheath surrounding the central bundles. In
cross-section, the lower pedicels are generally circular with three large, centrally
arranged bundles. These three bundles establish the complete floral vasculature
(Fig. 10) and are located on radii which are designated the outer tepal (OT) radii.
The three radii between the OT radii are designated the inner tepal (IT) radii.

Each of the three lower pedicel bundles along the OT radii undergoes a tri-
parted, radial division with three resulting bundles. However, these tri-parted
divisions occur at slightly different levels, that is, they are not co-planar, but rather
follow a spiral pattern. A gap is created by the outward departure of the central
bundle product of each tri-parted division. The three central bundles which depart
along OT radii are designated the compound outer tepal or OT bundles (dorsal-
compound bundle; Sterling, 1982). They remain free of other vasculature and
eventually establish the outer tepal median (OTM), outer tepal lateral (OTL),
outer stamen (OS) and dorsal (D) vasculature (Fig. 7D, E, 8, 10).

The two remaining lateral products adjoining a gap following a tri-parted di-
vision fuse with similar adjacent laterals along the IT radii (Fig. 10). These new

116

Annals of Carnegie Museum

vol. 56

Fig. 2.— Flowering habit, fruit and seeds of S. occidentale. A-C. Three different views of simple
flowering racemes, bracts (br) and nodding flowers. D. Mature fruit on an erect pedicel. The lower
epigynous zone (ep) with persistent tepals, freed styles and zones of septicidal separation are evident.
E. Flat, narrowly oblong, winged seeds.

fusion bundles are formed at progressively higher levels in a spiral pattern. Each
of these three fusion fundles undergoes a tri-parted division similar to that ob-
served at a lower level among the three original bundles. The central bundle
products of this second set of tri-parted divisions depart along the IT radii and
establish the compound inner tepal or IT bundles (“zwischenbundel”; Sterling,
1982).

Subsequent divisions among the compound OT and IT bundles will be discussed
later. However, it should be stressed that their formation was in a spiral pattern.

Plant Height -cm

1987

IJtech — Stenanthkjm occidentale

117

40

30

o

Stenonthium occidentale A. Gray

f r , , , , + , , , , L_

10 20 30

Number of Flowers/Plant

Fig. 3. —Scatter diagram comparing plant height (cm), number of flowers/plant and degree of inflo-
rescence branching of S. occidentale. Each dot represents an individual plant selected randomly from
range-wide herbarium specimens. Average inflorescence models are indicated for unbranched and
branched plants. (Sample size: N = 120.)

A given pedicel cross-section will usually show one bundle departing, another
being formed and a gap in an area where a third will be formed (Fig. 5B, C, 10).

The two remaining lateral bundles of the second, above mentioned tri-parted
divisions fuse laterally with adjacent laterals. These three fusion products close
the gaps along the OT radii which were formed when the compound OT bundles
were established and departed. These three new fusion bundles establish the total
ventral supply network of the gynoecium (Fig. 10). Dorsal formation is not as-
sociated with these bundles. These new bundles along OT radii undergo radial
divisions at progressively higher levels. The two resulting products of each division
fuse laterally with a similar adjacent lateral. Three new fusion bundles result along
the IT radii. These bundles in turn close the gaps formed during the departure of
the compound IT bundles. The three newly formed fusion bundles which are
along the IT radii form the points of a central, triangular vascular zone (“stele”).

The inferior gynoecium (Fig. 2D, 4) and its associated vasculature can best be
described in two parts: that of the lower epigynous zone and that of the upper
freed zone. Formation of the ventral supply, locule opening, ovule placentation
and gynoecium subdivision all occur within the lower epigynous zone (Fig. 5B-

118

Annals of Carnegie Museum

vol. 56

0 5 10 15 20 5 10 15 20 25

Length of Freed Tepal |mm| Length of Freed Gynoecium |mm|

Fig. 4.— Floral scatter diagrams for S. occidentale. A. Comparison between freed tepal length (L in
mm) to the height (H) of the flowering epigynous zone (mm). B. Comparison between the length (L)
of the freed gynoecium (mm) to the height (H) of the epigynous zone (mm). While the epigynous zone
height nearly doubles from the nodding flowering condition to that of the erect fruiting condition, the
fruit appears only weakly inferior. The degree of epigyny is approximately 1 6% in fruit.

F, 6, 7 A, B) in a spiral, not co-planar, pattern. The dorsal bundles which are
derived from compound OT bundles are not established until the upper peripheral
limits of the epigynous zone.

Following the formation and departure of the compound OT and compound
IT bundles, a triangular vascular zone remains in the central area (Fig. 5D-F,
6A-D). The three large bundles which lie along the IT radii establish the triangle’s
comers and close the gaps formed by compound IT bundle departure (Fig. 6A-
E). These three bundles designated as septal axials (SA) are formed at slightly
different levels. The three locules open perpendicular to OT radii once the three
comer bundles are established (Fig. 5D-F, 6). Also at this level, a central opening
or “hole” appears (Fig. 7A, B) which indicates that the three locules will be inter-
connected along OT radii and three septal arms or wings freed along IT radii.
The lower carpellary hole is internally continuous with the three locules and along
the floral axis into the upper stylar canal.

The three dorsals (D) are the last carpellary bundles to be formed in the upper
epigynous zone. Furthermore, as the outer carpellary walls are freed from the
surrounding perianth, the septal arms are subdivided internally along the IT radii
(Fig. 7 A, B). Septal glands were not observed. The three septal axials (SA) end in
this zone without any further connection to the central ventral supply. The three
carpels are freed from one another (Fig. 7) in what can best be described as a
weakly appressed apocarpous condition. Each freed carpel has a dorsal (D) and
two ventrals (V) which continue into the stylar zone. There is no terminal car-
pellary fusion between the ventrals or between the ventrals and the dorsal. The

1987

UTECH — STENANTHIUM OCC1DENTALE

119

Fig. 5. —Transverse sections through the upper pedicel and lower receptacle of S. Occident ale showing
the formation and departure of both compound outer tepal (OT) bundle and compound inner tepal
(IT) bundle as well as spiral locule opening. A. Mid-pedicel section showing the departure of three
compound OT bundles. B. Enlargement of A detailing the progressive formation of three compound
IT bundles (arrows). C. Upper pedicel section above B showing compound OT bundle departure and
further formation of compound IT bundles (arrows). D. Section above C showing transition zone
vascularization between the upper pedicel and lower receptacle. With first locule (L) opening and
outward departure of compound OT and IT bundles, the remaining central bundles establish the
ventral supply. E. Section above D with two open locules. F. Section above E showing three open
locules. Formation of the ventral supply occurs in the central placental area before the dorsal supply
is established.

stylar canal formed by the septal wing tips are lined with non-papillate cells. Each
free upper style forms a stylodium (Fig. 2D). Raphide idioblasts were not observed
in either carpellary or perianth tissue.

Each of the three carpels usually has eight to twelve bitegmic, basitropic, cam-
pylotropous ovules. The placental (P) bundles supplying these ovules depart quite

120 Annals of Carnegie Museum vol. 56

Fig. 6.— Transverse sections through the lower receptacle to upper epigynous zone of S. occidentale.
A. Overview section showing the peripheral compound OT and IT bundles, the central ventral supply
and ovules in all locules. B. Enlargement of A detailing the ventral supply. A pair of ventrals (V) is
associated with each locule and a septal axial (SA) with each septal arm. C. Section above B showing
ovular attachment. D. Enlargement of section above C detailing the ventral supply, compare with B.
E. Overview section above D showing compound OT and IT bundles in outer epigynous zone and
septal arm formation with simple septal axials (SA). F. Section above E showing further central zone
subdivision and septal arm formation. Each carpel usually has four rows of ovules supplied via two
ventrals (V). The dorsal supply is not established at this level.

horizontally in two ranks from a given ventral (V) (Fig. 6E, F, 7 A, B, 10). Four
rows of ovules per carpel are typical. Several flat, oblong, winged seeds are usually
found in each carpel (Fig. 2D, E). The seed’s distal appendage is directed upwards
and away from its attachment point. Following anthesis, the gynoecium far sur-
passes the withered, persistent perianth and greatly reduces the observable floral
epigyny and the semi-inferior gynoecium. Septicidal dehiscence begins in the free

121

1987

U TECH — S TENANTHIUM OCC1DENTALE

Fig. 7. —Transverse sections through the upper epigynous and freed carpellary zones of S. occidentale
showing various views of the bitegmic, basitropic and campylotropous ovules. A. Section showing
the central carpellary hole as the septal arms with septal axials (SA) are established. B. Section above
A showing the freed gynoecium with a carpellary hole and lacking septal axials. Septal arms subdivision
occurs along the IT radii (arrows). C. Overview section above B showing freed floral parts and the
central carpellary hole. D. Section above C showing the subdivided septal arms with ventrals (V) near
the inner appressed septal wing tips. E. Young floral bud section showing floral epigyny around a
separating gynoecium (arrows), compare with A and E. F. Section above E showing a freed, immature
gynoecium surrounded by fused stamens and tepals. Protandry is evident when compared to C.

stylar or stylodial area of the capsule (Fig. 2E) and continues along the zone where
the septal arm tips meet (Fig. 7D).

In most liliaceous species with a superior ovary, tepals, stamens and their
vasculatures are well established below the level at which locules open and ovules
are supplied. However, this is not the case in S. occidentale and other members
of the Veratreae due to an inferior gynoecium and varying degrees of perianth

122

Annals of Carnegie Museum

vol. 56

Fig. 8. —Series of transverse sections from an inner tepal of S. occidentale showing tepal vascularization
and the v-shaped adaxial nectary. (The outer tepal nectary and vascularization is similar.) A. Transverse
section showing inner tepal and stamen epitepaly and lateral fusion to outer tepals (arrows), compare
to Fig. 7E, F. An inner tepal median (ITM), two inner tepal laterals (ITL) and an inner stamen (IS)
bundle are shown. B. Origin of nectiferous zone between inner tepal and stamen (arrow). C. Inner
tepal freed from the dilated inner stamen, arrows indicated adaxial nectiferous zone. D. Increased
width of nectiferous zone (arrows) with division of two ITL shown. E. Lateral localization of nectiferous
surface into a v-shape is associated with enlarged margins and an abaxial mid-rib and four ITLs. F.
Further lateral localization of nectiferous surface along the adaxial tepal margins (arrows).

fusion. In S. occidentale, formation of tepal and stamen vasculatures occurs above
the level at which the ovules are supplied (Fig. 6, 7 A, B). The insertion and
departure of the tepals and stamens follow a spiral pattern in forming an imbricated
floral tube. Throughout much of the narrowly campanulate flower, the freed outer
tepals overlap the inner (Fig. 7C, 7E, F, 8, 9). The pale greenish yellow to purplish
brown nodding flowers have narrowly lanceolate tepals with spreading, recurved,
acuminate tips (Fig. 2A-C, 4). Basally above the epigynous zone the tepals are
connate for a short distance. Epitepaly occurs among both the outer and inner
stamens. A bilobed or v-shaped nectiferous zone occurs adaxially on the freed
tepals. The stamens, which average 7.0-8. 5 mm long, are included. The anthers
are located at the level at which the tepal tips recurve. The gynoecium and styles

1987

TJ TECH — S TENANTHIUM OCCIDENTALE

123

Fig. 9.— Stamen vascularization and dimorphic filament morphology of a young bud of S. Occident ale.
A. Transverse section above the epitepaly zone showing six basally dilated filaments. Adaxial nectif-
erous zones on both the outer and inner tepals are evident as well as three open styles. B. Transverse
section above A showing three inner filaments (IS) surrounding three freed styles adaxially and three
outer filaments (OS) abaxially. C. Section above B showing the central inner filaments (IS) and divergent
outer filaments (OS). D. Section above C showing adaxial inner anther attachment as the outer filaments
(OS) remain peripherally. The four apparent thecae are confluent terminally. E. Section above D
showing adaxial attachment of outer anthers and the centrally located inner anthers. F. Section above
E showing the upper confluent thecae of the three outer anthers (OS), arrows indicate dehiscence
zones.

are much shorter than the stamens when the anthers dehisce. However, the mature
fruiting gynoecium far surpasses the withered, persistent perianth (Fig. 2D) and
the degree of observable inferiority is greatly reduced (Fig. 1 1).

Six complex, compound bundles established in the lower receptacle levels are
ultimately responsible for the complete tepal and stamen vascularization. These
compound bundles have been designated the compound OT bundles (dorsal-

124

Annals of Carnegie Museum

vol. 56

1987

Utech — Stenanthium occidentale

125

compound bundle; Sterling, 1982) and the compound IT bundles (“zwischen-
bundel”; Sterling, 1982), respectively, for they are positioned along those respec-
tive radii. A description of the resulting vascularization from one compound OT
and IT bundle will illustrate the patterns for both sets of tepals and stamens since
the six bundles involved are all free from each other (Fig. 10).

In the upper epigynous zone, a compound OT bundle appears triangular in
cross-section (Fig. 5B-F, 6 A, C, E, F). Several complex subdivisions occur within
the compound OT bundles which results in the formation of an outer tepal median
(OTM), two outer tepal laterals (OTL), an outer stamen bundle (OS) and a dorsal
(D). The OTM is established first and after its departure, the remaining product
undergoes a rapid subdivision in which two laterals (OTL), outer stamen (OS)
and dorsal (D) bundles are established. The OTM, OS and D bundles all lie along
the same radius with normally arranged xylem (adaxial) and phloem (abaxial)
(Fig. 5C). It should be noted that the outer tepal laterals (OTL) are derived from
the remaining product bundle, not the outer tepal median (OTM). Each outer
tepal receives three bundles or traces, an outer median (OTM) and two laterals
(OTL). The laterals undergo further radial divisions to establish additional laterals.
In a freed, mature outer tepal, a seven, rarely more, bundled condition is typical,
that is three outer tepal laterals (OTL) + outer tepal median (OTM) + three outer
tepal laterals (OTL).

The origin of the inner tepal median (ITM), the two inner tepal laterals (ITL),
and an inner stamen bundle (IS) from a common IT bundle is similar to that of
the outer series vascularization (Fig. 5B-F, 6A, C, E, F, 7A-D). The ITM, two
ITLs and IS bundles all have normally arranged conducting elements. The only
difference in outer and inner vascularization which arose from compound OT
and IT bundles, respectively, is that a dorsal (D) is associated with subdivision
of the compound OT parental bundle, and there is no counterpart with the com-
pound IT bundle.

Each inner tepal is supplied with three bundles or traces, an inner median (ITM)
and two laterals (ITL). The laterals undergo further radial divisions, as in the
outer cycle, to establish a seven bundled condition (Fig. 7C, E, F, 8). A freed,
mature inner tepal has three inner tepal laterals (ITL) + inner tepal median
(ITM) + three inner tepal laterals (ITL). There is no fusion between any of the
laterals or between laterals and medians in either set of tepals. Each bundle or
traces follows a parallel course and ends near the laminal tepal margins.

The outer filaments are ca. 0.75-1.5 mm longer than the inner filaments (Fig.
9). The dimorphic filaments present a greater vertical zone of pollen dispersal
within the imbricated tepals, than if the filaments were the same length. The
extrorse anthers which are basifixed have a valvular dehiscence between the con-
fluent thecae that opens into a reniform, peltate disc (Fig. 9). The endothecium

Fig. 10. “-Roll-out longitudinal summary diagram for the floral vasculature of S. occidentale. A spiral
pattern is indicated in the formation and departure of the various letter coded and text discussed
bundles: OT = compound outer tepal, IT = compound inner tepal, V = ventral, P = placental and
SA = septal axial. The following bundles which are not shown result from the subdivision of the
following two bundles, respectively: from the compound OT bundle results the OS (outer stamen),
OTM (outer tepal median), OTL (outer tepal laterals) and D (dorsal), and from the compound IT
bundle results the IS (inner stamen), ITM (inner tepal median) and ITL (inner tepal laterals).

126

Annals of Carnegie Museum

vol. 56

Pulliser'a Drit. N. Am. Expl. Expedition.

ROCKY MOUNTAINS.

coll. E. Bourgt aa. 18$$.

Dr. I.yajt.
1K51>.

Fig. 1 1.— Lectotype (right element) and paratype (left element) of S. occidentale. (Courtesy of Gray
Herbarium, Harvard University)

has wall thickenings or bands of the girdle type (Dahlgren and Clifford, 1982).
Protandry occurs within each flower of S. occidentale , since both outer and inner
anthers dehisce at the same time, and this is earlier than gynoecial maturity (Fig.
7C, E, F, 9).

1987

Utech — Stenanthium occidentale

127

Geography and Taxonomy

Stenanthium occidentale occurs in British Columbia from 52°N Latitude south-
ward into the Rocky Mountain Trench to northern Idaho and Montana as well
as southward at higher elevations in the Cascade and Coast Ranges of western
Washington, western Oregon and northwestern California (Fig. 1). The species
occurs between 3800-6500 ft in Del Norte, Humboldt, Siskiyou and Trinity
counties, California (Jepson, 1921, 1923; Munz, 1973; Ferlatte, 1974), but near
sea level on Vancouver Island, the Olympic Peninsula and the Columbia River
Gorge. Neither specimens nor reports are known from Alaska (Anderson, 1945;
Hulten, 1968), Wyoming (Dorn, 1977) and Queen Charlotte Island, British Co-
lumbia (Carter and Newcombe, 1921; Taylor, 1966; Calder and Taylor, 1968).
Frequent in coniferous to subalpine zones, this species occurs on wet cliffs, mossy
stream banks, rocky crevices, moist montane thickets, and meadows. In northern
areas, it is commonly an understory associate of Abies, Thuja, Tsuga and Pseu-
dotsuga.

In describing S. occidentale, Asa Gray (1873) listed a syntypic series of five
collections: Rocky Mountains, 1858, Bourgeau; Cascade Mountains, 49°N Lati-
tude, 1859, Lyall; Kootenay, 1860, Lyall; Rocky Mountains, 49°N Latitude, 6000
ft, 1861, Lyall; and foothills of Cascade Mountains, mossy banks of stream, 1871,
Hall. Baker’s review (1 879) cited only Bourgeau, Lyall and Hall collections without
selecting a type. Rydberg (1900), in transferring S. occidentale and S. sachalinense
to Stenanthella on the basis of a superior ovary, cited numerous specimens for
Stenanthium occidentale, but only Bourgeau’s collection was included from Gray’s
syntypic series. Piper (1906), Jepson (1921), and Hitchcock et al. (1969) in their
typification stated “Bourgeau, Rocky Mountains.” Since the Bourgeau specimen
at the Gray Herbarium (Fig. 1 1) is stamped “type” and was mentioned first by
Gray (1873) in his syntypic series, it should be designated the lectotype.

Stenanthium occidentale A. Gray, 1873

Stenanthium occidentale A. Gray, Proc. Amer. Acad., 8:405. 1873.

Lectotype. — Rocky Mountains, Palliser’s British North American Exploratory Expedition, E. Bour-
geau s.n., 1858 (GH!, here designated; photo CM)— Fig. 1 1.

Isolectotypes.—{ B, K, NY!; photo CM).

Paratypes. — Oregon Boundary Commission, Cascade Mountains, 49°N Lat., 1859, Lyall s.n. (GH!,
photo CM; Fig. 1 1); Oregon Boundary Commission, Cascade Mountains to Fort Colville, about 49°N
Lat., Cascade Mountains, July 1860, Lyall s.n. (GH!; photo CM); Oregon Boundary Commission,
from Fort Colville to Rocky Mountains, Rocky Mountains, Lat. 49°N at 6000 ft above sea level,
1861, Lyall s.n. (GH!; photo CM); Oregon, foothills of the Cascade Mountains, mossy banks of stream,
1871, is. Hall 535 (F!, GH!, MO!-2 sheets, NY!-3 sheets; photos CM).

Stenanthella occidentalis (A. Gray) Rydberg, Bull. Torrey Bot. Club, 27:531. 1900.

Stenanthium rhombipetalum Suksdorf, Werdenda, 1:6. 1923.

Lectotype. — Washington, Skamania County, Cape Horn along Columbia, 27 May, 29 June 1920,
W. Suksdorf 10,466 (GH!; photo CM).

Isolectotypes.~{ DS!, MO!, NY!-2 sheets, UC!, WTU!; photos CM).

Representative specimens examined. — CANADA: ALBERTA: Near Banff, Scott s.n. (DAO), Canby
s.n. (GH, US); Banff, Olson s.n. (GH), Rusby s.n. (NY), Phillips s.n. (CM), 4550 ft, Hermann 12657
(US); Baker Creek, Kojima s.n. (DAO); Waterton Lakes, Macoun 13850 (GH), Moss 1003 (DAO);
Crypt Lake, 6400 ft, Senn 2639 (DAO); Red Rock Canyon, 4500 ft, Breitung 16098 (F, NY), 5000
ft, de Vries 2127 (DAO); Red Rock Nature Trail, Sudol 115 (DAO); Moraine Lake, Hallis 5727 (MO),
McCabe 5227 (UC); Bertha Lake, 4200-4600 ft, Malte & Watson 282 (GH), 4400 ft, Hermann 13064
(US); Devil’s Lake, Sanson 22126 (CM); Alderson Lake, Moss 510 (DAO, GH); Summit Lake, 5500
ft, Moss 547 (GH); Headwaters of Saskatchewan & Athabasca Rivers, S of Summit, Brown 1547 (GH);

128

Annals of Carnegie Museum

vol. 56

Rowe Lake, Scatter 10014 (DAO); Sulphur Mt., van Brunt 97 (NY), Sanson s.n. (DAO), Jenkins 1424
(DAO), 5000 ft, McCalla 2217 (NY, US); Crow Nest Pass, Macoun 25101 (NY); Ptarmigan Valley
to Pipestone Valley, 6000 ft, Brown 432 (GH, MO, NY); Lake Louise, Olson s.n. (GH), Jones s.n.
(WTU), Fraser 4 (DAO), Dempster 1237 (JEPS), Big Beehive, Beattie 5249 (DAO); Valley of Ten
Peaks, MacCallum s.n. (CAS); Spray River, Degener 2409 (NY); Tunnel Mt., Malte & Watson 835
(GH); Marble Canyon, Turner s.n. (DAO), Groh s.n. (DAO); Sundance Canyon, McCalla 4555 (DAO);
Sundance Camp, McCalla 4555 (GH); Johnson Canyon, 1500-1700 m, Boivin 4741 (DAO); Johnston
Creek, Hillsdale, Mair 153 (UBC); Cascade River, Lake Minnewanka, Jenkins 1600 (DAO); N of
Coleman, W flank Livingstone Range, Norris 81 (DAO); Livingstone Falls, 5900 ft, Hermann 12766
(US); Sugarloaf Mt., NW of Livingstone Gap, 7700 ft, Dore et al. 20524 (DAO); Morley, Packer s.n.
(DAO); Eisenhower Junction, Vermilion Pass, 5500 ft, Ogilvies.n. (UBC); Between Dolomite & Cirque
Peaks, Calder & MacKay 32752 (DAO); Larch Valley, 7200 ft, Vrugtman s.n. (DAO); Consolation
Valley, 6200 ft, Beamish & Vrugtman 620159 (DAO, UBC); Pigeon Creek, 5500 ft, Whitehorn 504
(DAO); SW of Exshaw, NW of Pigeon Mt., Brunton 1137 (DAO).

BRITISH COLUMBIA: Rocky Mts., 5000 ft, Macoun s.n. (MO); Rocky Mts. & Pacific Coast,
Freeman s.n. (GH); Selkirk & Rocky Mts., 7500 ft, Shaw 338 (CM, GH, MO, NY), 7000 ft, Macoun
338 (US); Skagit Valley, 5000 ft, Macoun 70219 (CM); Chilliwack Valley, Macoun 54078 (GH, NY,
US); Bow River Pass, Macoun s.n. (GH, NY); Emerald Lake, 4300 ft, Macoun 118 (NY), 4200 ft,
Brown 332 (GH, MO), Marquand 53 (NY); Bella Coola, Clayton Falls, Leamon 145 (UBC); Lastsquaw
Creek, McCabe 1476 (UC); Natural Bridge, 4000 ft, Ulke 6 (CAS); Little Yoho Valley, Bostock s.n.
(DAO), Comte 1721 (MO); Yoho Valley, Eastham s.n. (UBC), Fleming s.n. (UBC); Point Lace Falls,
Lane s.n. (CM), McCalla 7144 (UBC); Takakkaw Falls, 4900 ft, Calder & Saville 12020 (DAO); Yoho
Pass, Walcott s.n. (US), Vermilion Pass, McCabe 5180 (UC); Kicking Horse Lake, Macoun s.n. (GH,
US, NY); Glacier, McCallum s.n. (CAS), Doanes.n. (DS); Mt. Cheam, Fletcher 1872 (DAO); Spuzzum
Creek on Hope-Lytton highway, Calder & Saville 8411 (DAO, UBC); Penticton, Green Mt. Road,
Stonor s.n. (UBC); Mt. Brent, W of Penticton, 5500 ft, Calder et al. 2057 (DAO), 5000 ft, Eastham
s.n. (UBC); SW of Penticton, 6800 ft, Calder & Saville 12269 (DAO, NY), 6000 ft, Krause 681951
(UBC); Hope-Princeton area, 5000-6500 ft, Fedor s.n. (UBC); Hope, Eastham s.n. (DAO, UBC);
Tulameen River, Kemp s.n. (NY); Manning Park, W on Hope-Princeton highway, 4300 ft, Calder &
Saville 11562 (DAO), Krajina s.n. (UBC), Orchid Meadow, 3700 ft, Beamish et al. 9084 (GH, UBC),
Ellis Peak turn-off, 6500 ft, McLean s.n. (DAO), Paddy Lake road, 6000 ft, McLean s.n. (DAO); Mt.
Apex, E of Nickel Plate Mine, 5600 ft, Calder & Saville 10741 (CAS); Hedley, Nickel Plate Mine,
Vrugtman 5 (UBC), McNulty Creek road, 4000 ft, Grant s.n. (DAO); Nickel Plate townsite, SE of
Hedley, 5700 ft, Beke & Marchand s.n. (DAO); Goodchap Mt., 5000-7000 ft, Martin s.n. (UBC);
Apex Alpine Ski area, Spellenberg & Soreng 5633 (NY); Ashnola District, Cathedral Lakes, 7000 ft,
Taylor 1311 (UBC, UC); Cathedral Park, Ewart Creek Valley, 1 200 m, Hainault 7598 (DAO), Hainault
7902 (DAO); Noisy Creek, Ashnola Forks, 5000 ft, Barr 9320 (UBC); Ashnola Range, 49°07'N,
120°08'W, 5600 ft, Calder et al. 19559 (DAO, UC, WTU); SW of Okanagan Falls, 3800 ft, Calder &
Saville 9933 (DAO, UC, GH); Nelson, Six Mile Lakes, Eastham s.n. (UBC), Silver King Mine, 5000-
6000 ft, Eastham s.n. (UBC); NNE of Nelson, Duhamel Creek, 3100-3300 ft, Calder & Saville 9389
(DAO); SW of Nelson, Copper Mt., 7000 ft, Calder & Saville 12271 (DAO); Kootenay, Floe Lake,
4000 ft, Hume 704 (UBC); Kootenay, Sinclair Nature Trail, 3475 ft, Seel 39 (DAO); Stanley Peak
Valley, 51°1 l'N, 1 16°05'W, Calder & MacKay 32795 (DAO); Moyie, between Creston & Cranbrook,
3100 ft, Calder & Saville 9235 (DAO, NY, UC, WTU); Mouse Creek, Fort Steele, Fodor 53 (UBC);
Femie, Coal Creek, Eastham s.n. (DAO, UBC); Estelle Mine, Wasa, 5000 ft, Grant s.n. (DAO); N of
Bull River, 49°37'N, 1 15°25'W, 5200 ft, Taylor & Ferguson 2717 (DAO); Elk River Valley, 20 mi N
of Natal, Weber 2257 (GH); Starvation Peak area, 49°02'N, 1 14°12'W, 6300 ft, Taylor et al. 3412
(DAO); Flathead, 4000 ft, Bell & Davidson 42 (DAO); SE of Flathead townsite, Flathead River, 5200
ft, Taylor & Ferguson 2057 (DAO); Flathead Summit, McCabe 4997 (UC); Mt. Stephen, Fogg s.n.
(UBC); NE of Canal Flats, N of junction of Albert & Ralliser Rivers, 3900 ft, Taylor & Ferguson
1488 (DAO); Fairmont Hot Springs, Flat Creek, Eastham s.n. (UBC); W of Windermere, Paradise
Mine, Toby Creek, 5200 ft, Calder & Saville 11346 (DAO); Paradise Mine, Invermere, Hardy s.n.
(DAO); Radium Hot Springs, Eastham s.n. (UBC), Calder & Saville 11177 (DAO); NW of Takakkaw
Falls, 6000 ft, Hitchcock & Martin 7700 (DS, GH, MO, NY, UC, WTU); Othello, Mt. Ogilvie, 3500
ft, Brayshaw s.n. (UBC); Mt. Copley, 3000-3500 ft, Beamish 232 (UBC); Mt. Assiniboine, Sunburst
Lake, 7200 ft, Scamman 6600 (GH); Frosty Mt., 4400 ft, Calder et al. 15673 (GH); Canim Mine,
Beamish & Stone 7574 (UBC); NE of Pollock, 49°22'N, 1 14°34'W, 6500 ft, Taylor & Ferguson 2916
(DAO); Upper Sage Creek Valley, 5200 ft, Beamish et al 747 (UBC); San Joseph Bay, 20 ft, Anderson
& Dejardin 9 (UBC); Moss Lake, Pajar & Pinder-Moss 64 (UBC). VANCOUVER ISLAND, Macoun
s.n. (GH), 5000 ft, Calder & MacKay 32586 (NCU); Port Alice, 2200 ft, McComb & Watson s.n.
(UBC); Kyuquot, 100 m, Taylor & Szczawinski 359 (UBC); Power River, 50°12'N, 127°30'W, 560

1987

Utech— Stenanthium occidentale

129

m, Jones s.n. (DAO); Crest Creek along Gold River road, Calder & MacKay 31786 (DAO); Great
Central Lake, Buckland s.n. (UBC); Forbidden Plateau, Gumming s.n. (UBC); Mt. Becher, 2900 ft,
Calder & MacKay 31906 (DAO); Mt. Arrowsmith, Carter s.n. (GH, US), Baker 938 (CAS), Howell
7619 (CAS), 3500 ft, Jeffery s.n. (UBC), 3400 ft, Schmidt s.n. (DAO); Markale, Taylor & Szczawinski
211 (UBC); Cameron Lake, McBey s.n. (UBC), 2500 ft, Eastham s.n. (UBC); Port Albemi, Calder et
al. 19427 (DAO); Home Lake, Macoun s.n. (MO); Port Albemi-Sarita road, N of Franklin River,
Calder & MacKay 30329 (DAO); Albemi, Carter s.n. (GH); Mt. Joan, Beaufort Range, 2500-3000
ft, Calder & MacKay 32334 (DAO); Deep Lake, 3500 ft, Schmidt s.n. (DAO); Between Kennedy &
Taylor Rivers on Port Albemi-Tofino road, Calder & MacKay 30661 (DAO); Sprout Lake, Meyer
1543 (GH, MO, NY, UC, US), Taylor 1252 (UBC), Calder & MacKay 30682 (CAS); Fourth Nanaimo
Lake, 300 m, Krajina et al. 4741 (UBC); Nanaimo River Valley, 930 ft, Mueller- Dombois 101-1
(UBC), 300 m, Krajima et al. 5126 (UBC); Englishman’s River, 3300 ft, Schmidt s.n. (DAO); Estelle,
300 m, Krajina et al 4801 (NY, UBC); Moat Lake, 4000 ft, Underhill s.n. (UBC, WTU), Calder &
MacKay 32280 (DAO); Elk River Falls, W of Campbell River, Calder & MacKay 30416 (DAO); Elk
River, 1000 ft, Young & Hubbard 579 (DAO); Elk Pass, 50°34'N, 1 1 5°04'W, 6200 ft, Taylor & Ferguson
3611 (DAO); Junction Upper Campbell Lake & Battle Lake, Calder & MacKay 30583 (DAO, NCU);
Gail Falls on Kelsy Bay-Campbell River road, 50°16'N, 125°49'W, Calder & MacKay 32439 (DAO);
Burman Lake, 4400 ft, Calder & MacKay 32491 (DAO, UBC, UC), 5000-6000 ft, Krajima et al
68081 (UBC); Below Mt. Burman, 5000 ft, Calder & MacKay 32586 (DAO); Kennedy River, 450 ft,
Young & Hubbard 285 (DAO); Kennedy Lake, Wade et al. 200 (UBC), 200 ft, Pajar 144 (UBC);
Cowichan Lake, Rosendahl 1805 (GH, MO, NY, UC), Spilsbury s.n. (DAO).

UNITED STATES: CALIFORNIA: Del Norte Co.: Siskiyou Mts., Brandegee s.n. (GH); Waldo
to Black Butte, Eastwood 2150 (CAS); S Fork Smith River, 300 ft, Adams s.n. (HSC); Preston Peak,
Bear Basin, Gasquet, 4701 ft, Clifton & Griswold 5732 (HSC). Humboldt Co.: Trinity Summit Range,
Happy Camp Mt., 4200 ft, Tracy 15558 (DS, GH, JEPS), 4200 ft, Tracy 18172 (UC), 4200 ft, Tracy
18517 (UC), Box Camp, 4800 ft, Tracy 17901 (UC); Near Tish Tang A Tang Creek, S Fork, Hoopa,
4649 ft, Gilbert & Muth 11524 (HSC); Salmon Mts., Mill Creek, Middle Fork, Hoopa, 4400 ft, Overton
12360 (HSC), 5551 ft, Clifton & Griswold 11935 (HSC); Whiteys Peak, 4998 ft, Kearns 434 (HSC).
Siskiyou Co.: Preston Peak, Rattlesnake Meadow, 5400 ft, Montalvo & Ackerman 425 (HSC), 5000
ft, van Deventer s.n. (HSC); White Mt., T 47N, R 10W, S 10, Wheeler 3181 (CAS, GH); Canyon
Creek, 6300 ft, Tucker et al. 3747 (CAS, UC); Paradise Lake, 6300 ft, Alexander & Kellogg 5857 (DS,
UC, WTU); Hammond, Kelsey Creek, 5000 ft, Hoffman 4043 (UC); Marble Gap Trail, 6000 ft, Muth
427 (CAS); Marble Mts., Spirit Lake, 6000 ft, Howell 15070 (CAS); Marble Valley Guard Station,
Stillman 364 (HSC); Klamath National Forest, Siskiyou Mts., Cook & Green Pass, 3500 ft, Smith
9962 (HSC), 4000 ft, Roderick 6487 (UC), Roderick 65106 (JEPS); Black Mt., Muth s.n. (CAS); Black
Marble Mt., Pacific Crest Trail, Sawyer & Mesler 4050 (HSC). Trinity Co.: Trinity Alps Wilderness
Area, Union Creek, 6100 ft, Hall 8626 (UC), 5700 ft, Goforth 280 (HSC); Sapphire Lake, 6100 ft,
Sawyer 2359 (HSC).

IDAHO: Adams Co.: S side Ruth Lake, ca. 16 mi SSW of Riggins, T 22N, R 2W, S 15, 7250 ft,
Bingham & Miller 446 (ID). Bonner Co.: Hope, Leiberg s.n. (MO, ORE); Green Monarch Mt., Dunkle
s.n. (ID); Pack River, Colburn, 2100 ft, Ripley & Barneby 10940 (CAS). Boundary Co.: Snowy Top,
Christ 5995 (ID, NY); Port Hill, Smith Creek, 4500 ft, Rossbach 273 (DS, WTU); Smith Peak, 6000
ft, Rossbach 274 (DS); N Fork Hunt Creek, Daubenmire 44372 (NY, WTU); Three Ponds area, T
6 IN, R 2W, S 1 1, 3400 ft, Wellner 2659 (ID); Upper Priest River Valley, Upper Priest Falls, 2900
ft, Henderson & Cholewa 6789 (ID). Clearwater Co.: N Fork Clearwater River, junction main fork,
Sharsmith 3574 (GH, UC, WTU); Clearwater National Forest, Aquarius, 1670 ft, Wellner 1755 (ID).
Idaho Co.: Headwaters of Squaw Creek, 5400 ft, St. John & Muller 8389 (UC); Lowell, Three Devils
Creek, Middle Fork Clearwater River, Constance et al. 1084 (MO, UC); Lowell, Three Devils Camp,
Meyer 916 (GH, MO), Davis 3443 (CAS), Baker 14805 (ID, NY); Selway River, SW of Lowell, Christ
12094 (ID), Christ & Ward 12094 (NY); Falls of Selway River, Thompson 1668 (WTU), McMullen
1517 (ID); Meadow Creek, Gail s.n. (ID), Botany 54 Class (ID); Seven Devils Mt., Brundage Trail at
head of Sheep Creek, 6600 ft, Packard 454 (UC), Christ 12698 (NY); Hells Canyon Recreation Area,
1 1 mi WSW of Riggins, T 23N, R 2W, S 12, 7100 ft, Bingham & Miller 407 (ID); W of Riggins,
Baker 11999 (ID, NY). Kootenai Co.: Leiberg s.n. (F, NY); Watson Woods, Leiberg 125 (WTU).
Shoshone Co.: Coeur d’Alene Mts., Stevens Peak, E of Wallace, 1750 ft, Leiberg 1479 (GH, MO,
NY, ORE); Priest Lake, Piper 3691 (GH); Creek Pass, Baker 16939 (WTU); Upper Priest River, 3000
ft, Epling 7071 (MO); St. Joe River Canyon, Avery, 3300 ft, Henderson & Cholewa 6653 (CM).

MONTANA: Flathead Co.: Columbia Falls, Flathead River, Bad Rock Canyon, Rogers 995 (MO,
NY, WTU); Loneman Mt., 4400 ft, Parke 15 (UC); Whitefish, Stickney 1228 (GH). Glacier Co.:
Glacier National Park, Russell s.n. (GH), van Dyke s.n. (CAS), Phelps s.n. (CM); Mineral Park, Jones
s.n. (DS); Iceberg Lake, Young s.n. (TEX); Logan Pass, Schreiber 1333 (UC), Henry s.n. (CM); Kootenai

130

Annals of Carnegie Museum

vol. 56

Lake to Waterton Ranger Station, Harvey & Marshall 4283 (MO, NCU); Canyon Creek, 5800 ft,
Brant 66 (WTU). Lincoln Co.: Cabinet Mts. Wilderness Area, Leigh Lake, 6500 ft, Woodland 694
(DS), Thomas 12280 (DS). Missoula Co.: Ridge between Hillgate & Big Blackfoot Rivers, Watson
428 (GH); Big Blackfoot River, Canby 332 (NY); Woodruffs Falls, Upper Meanas Pass, Canby 332
(GH); Headwaters of E Fork Flathead River, Canby 332 (NY); Columbia Falls, Williams 917 (US),
Williams s.n. (NY); Glacier National Park, Hitchcock 11938 (US), 1650-1950 m , Standley 16705
(US); Sperry Glacier, Umbach 18553 (WIS); Blackfoot Glacier, Jones s.n. (DS, MO); Gunsight Pass,
Dunkle s.n. (ID); Gunsight Lake, Maguire 644 (UC); Lake Josephine, Standley 15236 (US); Lake
McDermott, Standley 15911 (US); Lake McDonald, Umbach 363 (WIS, MO); Little St. Mary’s Lakes,
Umbach 821 (WIS, MO); Grinnell Lake, Hollis 5728 (MO); Cracker Lake Trail, McLaughlin 763
(MO); Granite Park Chalets, Standley 16256 (US); Dawson Pass, Somes 51 (NY); Woodman, 4500
ft, Kirkwood 1477 (GH, UC); Copper Cliff, 5000 ft, Hitchcock 1826 (CM, DS, GH); Midvale, Umbach
247 (DS, MO, NY, US).

OREGON: Willamette Slough, Howell s.n. (GH); Cliffs along Columbia River, Cascades, Howell
s.n. (ORE); Columbia River, Howell 894 (CAS, US); Cascades, Howell 579 (CM); Parry 1016 (US).
Clackamas Co.: Eagle Creek, Henderson s.n. (ORE, OSC); Clackamas River, S of Rhododendron
Creek, 2400 ft, Wright 1710 (OSC). Clatsop Co.: Astoria, Anderson s.n. (DS, GH); Saddle Mt., 2900
ft, Patterson 45 (ORE), Ifft 15 (DS; OSC), Rossbach 392 (DS, WTU), Gilkey s.n. (OSC, WTU), 3250
ft, Detling 6456 (DS, ORE, UC), Dennis 2545 (DAO, DS, US), 2650 ft, Dennis 2592 (OSC); W peak
of Saddle Mt., Howell 28443 (DAO); Onion Peak, 3064 ft, Chambers 3227 (NY, OSC); Sugarloaf Mt.,
SE of Cannon Beach, 2800 ft, Chambers 3789 (OSC, WTU). Curry Co.: Onion Camp, 4500 ft, Leach
2960 (ORE, OSC); Siskiyou National Forest, Kalmiopsis Wilderness Area, 3760-4160 ft, Denton
2341 (ID, NY, WTU). Hood River Co.: Mt. Hood National Forest, Lost Lake, Gooding & Evinger
45 (OSC), Peck 9914 (WILLU), Henderson 800 (MO); Dry Creek Falls, Cascade Falls, Gustafson 151
(OSC); Columbia Gorge, W of Wyeth, 100 ft, Detling 7253 (DS, ORE). Jackson Co.: Siskiyou Mts.,
Whisky Peak, Star Ranger Station, 4000 ft, Robbins 3588 (JEPS). Josephine Co.: Howell s.n. (ORE);
S of Takilma, Peck 8024 (WILLU); E Fork Illinois River, Peck 8024 (GH); W Fork Williams Creek,
Cave Camp, Applegate 5102 (DS), Applegate 8725 (DS), Cave Trail, Applegate 5116 (DS); Grayback,
Layton ditch, Savage s.n. (ORE); Grayback Mt., Williams Creek, Henderson 12826 (ORE); Peavine
Lookout, 3000 ft, Priebe s.n. (OSC); Oregon Caves National Monument, No Name Creek, Baker &
Ruble 450 (ID, NY); W of Selma, Store Gulch, Steward s.n. (OSC). Lane Co.: Belknap Springs, Peck
2616 (WILLU); Dorena, Bohemia, Sharp Creek, Constance s.n. (UC); E Fork McKenzie River, T
16S, R 5E, S 5, 1400 ft, Detling 2849 (ORE); Big Fall Creek, Hette Camp, Henderson 18729 (ORE);
S of Oakridge, Peck 22072 (UC); Lookout Creek Basin, T 15S, R 6E, S 30, 3000 ft, Franklin & Dyrness
167 (OSC); Willamette National Forest, Walker Creek Trail, Lowder Mt., Utech 78-275 (CM). Marion
Co.: N Silver Creek, below Falls, Foster s.n. (ORE); Detroit, Santiam River, Peck 1667 (WILLU);
Silver Creek Falls, Nelson 2240 (GH). Multnomah Co.: Sandy River, Henderson s.n. (ORE, OSC);
Columbia River, Multnomah Falls, Howell s.n. (MO), Peck 1386 (WILLU), Peck 2035 (GH), Piper
6267 (GH, US), Thompson 2703 (DS, WILLU); Bridal Veil, Jackson s.n. (OSC); Latourelle Falls,
Thompson 4350 (DS, MO, WTU); Wakenah, Leach 1325 (ORE); McCord Creek, Warrendale, Jaques
s.n. (OSC), 3000 ft, Thompson 11872 (DS, MO, NY, WTU); Oneonta Gorge Trail, 65 ft, Ireland
3242 (ORE); Wahkeena Falls, Buddenhagen 5 (OSC); Crown Point, 210 ft, Bacigalupi et al 7885
(JEPS). Tillamook Co.: Cascade Mts., Kellogg & Harford 1025 (GH); Columbia River, Howell 374
(CAS, CM, NY), Eggert s.n. (MO), Parry 1016 (US), Dickson 4853 (NY), Dickson 8432 (MO); Castle
Rock, Dunn & Harford 1025 (NY); Klamath Falls, Applegate 738 (GH); Coast Mts., Lloyd s.n. (NY);
Bonneville, Palmer s.n. (US); Tillamook, Wilson River, Thompson 3116 (DS, MO); N of Wilson
River, SE of Blue Lake, 3300 ft, Chambers 4137 (OSC).

WASHINGTON: Cascades, Dunn s.n. (DS); Cascade Mts., Tweedy 340 (NY); Cascade Mts., Yakima
Region, Bandegee 438 (NY, UC), Tweedy s.n. (NY); Olympic Mts., Piper 2226 (GH, US), Piper s.n.
(WTU); Heart Lake, 4500 ft, Dickinson 62 (WTU). Chelan Co.: Bridge Creek, Chelan Lake, 3000
ft, Jones s.n. (DS); Wenatchee National Forest, Chiwaukum Lake, Eggleston 13549 (US); Ice Creek,
6000 ft, Morrill 349 (WTU); Mt. Stuart District, Mt. Ingalls, Eyerdam s.n. (MO, UC); Bryan Butte,
6200 ft, Ward 51 (NY, WTU); Rainy Creek junction with Bridge Creek, Ward 633 (NY, WTU);
Clallam Co.: Olympic Mts., Elmer 2305 (CAS, DS, MO, NY, ORE, WTU); Mt. Olympus, Flett s.n.
(WTU); Mt. Angeles, 5000 ft, Thompson 5604 (DS, MO, WTU), 5000 ft, Thompson 7516 (MO,
WTU), 4000 ft, Thompson 7843 (DS, GH, MO, WTU), 5300 ft, Thompson 10082 (NY, WTU), 6000
ft, Howell 7519 (CAS, DS), Flett 3352 (UC, US, WTU), Webster s.n. (WTU); Hoh River Trail to
Hoh Lake, 3400 ft, Otis 1326 (WTU); Canyon Creek, 3000 ft, Jones 5910 (WTU), Jones 8368 (WTU),
3200 ft, St. John 4774 (GH, MO, UC, WTU); Bogachiel Mt., Eyerdam 6309 (WTU). Jefferson Co.:
Constance Ridge, 4000 ft, Thompson 6569 (WTU); Lake Constance, 5500 ft, Thompson 7906 (WTU);
Marmot Pass, 5000 ft, Thompson 9918 (MO, NY, WTU); Mt. Anderson, 5000 ft, Meyer 687 (MO,

1987

Utech — Stenanthium occidentale

131

WTU). King Co.: Snoqualmie, Smith s.n. (MO, WTU); Snoqualmie Pass, Guy Peak, 3500 ft, Thomp-
son 9692 (NY, WTU); S ridge between Hester Lake & Godmeyer Hot Springs, 3500 ft, Jones s.n .
(DAO). Kittitas Co.: Martin to Lake Kichelas, Henderson s.n. (WTU); Yakima Valley, 2000 ft,
Wiegand et al. 522 (MO); Morse Creek, branch of American River, Peirson 11853 (UC); Wenatchee
Mts., Iron Peak, N Fork Teanaway River, Kruckeberg 2583 (ID, NY, UC, WTU), DeRoux Forest
Camp, Kruckeberg 5314 (WTU). Lewis Co.: Mt. Rainier National Park, King s.n. (CAS); Narada Falls,
Baker 743 (CAS), Pope s.n. (WTU); Silver Falls, Stevens Canyon, 2000 ft, Utech 77-201 (CM); Goat
Rocks Wilderness Area, Jordon Basin, 6150 ft, Franklin 598 (OSC); Baldy Peak, 4000 ft, Lamb 1322
(MO); Randle, W of Fawn Creek, Spellenberg & Sutherland 1222 (HSC). Mason Co.: Mt. Elinor Trail,
2000 ft, Freer 403 (WTU). Okanogan Co.: Okanogan Forest, Ashnola River, Sheep Mt. to Bald Mt.
Trail, 1630 m, Eggleston 13396 (US); Tiffany Range, Salmon Creek to Muckamuck Lookout, 4000
ft, Thompson 6952 (DS, F, GH, MO, OSC, UC, WTU); Big Craggy, 3000 ft, Thompson 10867 (CM,
DS, MO, NY, WTU); Crescent Mine, W of Twisp, Hitchcock & Muhllick 21595 (NY, WTU); N of
Whistle Pass, NW of Conconully, 7200 ft, Douglas 334 (DAO); Salmon Meadows Camp, NW of
Conconully along Meadow Creek, 4500 ft, Woodley 164 (GH); SE slope Mt. Chooaka, Cold Creek
Camp, 6000 ft, Kruckeberg 6543 (WTU). Pierce Co.: Longmire Springs, Wiegand et al. 523 (NY);
White River to Owyhigh Lakes, 4500 ft, Raven 8646 (CAS). Skamania Co.: Cape Horn, Howell s.n.
(GH), Piper 4966 (US), Gorman 2036 (ORE), Suksdorf 10394 (WTU); Mt. Adams, Misery Gulch,
Lloyd s.n. (NY); Spirit Lake, 3700 ft, Meyer 770 (MO); Hamilton Mt., 1400 ft, Detling 7065 (DS,
ORE, UC). Snohomish Co.: Savage et al, s.n. (MO); Whitechuck Mt., 3000 ft, Eyerdam s.n. (MO,
UC); Mt. Pugh, 7000 ft, Thompson s.n. (WTU); Cascade Mts., Perry Creek Trail, 4000 ft, Thompson
14746 (GH, MO, WTU). Wahkiakum Co.: Skamokawa, St. John 8752 (WTU). Whatcom Co.: Mt.
Paddo, Suksdorf s.n. (DS, US); Goat Mt., 5500 ft, Allen 233 (CAS, DS, GH, MO, NY, UC, US, WTU);
Mt. Baker, Gorman 2545 (ORE), 3700 ft, Mason 3847 (GH, UC), Mason 7414 (F, MO, NY, UC,
WTU); E side Deming Glacier, 5500 ft, Gorman 2893 (ORE); Meadow along Ruth Creek, Muenscher
7713 (GH, UC); Austin Pass, 3000 ft, Utech 78-300 (CM, WIS); Mt. Baker Lodge, 4000 ft, Thompson
5370 (DS, WTU), Thompson 5401 (DS, GH, WTU); Cliffs at Mt. Baker Lodge, 4200 ft, Thompson
11011 (MO, WTU), Thompson 11293 (WTU); N Fork Nooksack River, 3500 ft, Benson 2401 (DS,
MO, NY); Mt. Sauk, N of Rockport, 5000 ft, Douglas 1308 (DAO).

Discussion and Conclusions

Based on the current northern hemisphere distribution, the genus Stenanthium
has a pattern characteristic of descendant members of the so-called Arcto-Tertiary
Geoflora. There is a wide-spread species in both eastern North America [S. gra-
mineum (=S. robustum)] and western North America ( S . occidentale ; Fig. 1) as
well as species in the central volcanic belt of Mexico (S. frigidum) and eastern
Asia (S. sachalinense). The closely related S. occidentale and S. sachalinense,
which is endemic to Sachalin Island, form a circum-northem Pacific disjunct pair.
Their relationship will be investigated in this series.

The proposed segregation of S. occidentale and S. sachalinense into Stenanthella
by Rydberg (1900) based largely on a superior ovary has not been accepted, nor
has Stenanthium rhombipetalum Suksdorf (1923), a proposed robust segregate of
S. occidentale with most open flowers male. As shown here, the flowering gy-
noecium is inferior, the flowers protandrous and the perianth epigynous. Zim-
merman (1958) and Kupchan et al. (1961) proposed and used “for convenience
in summarizing information only” Section Eustenanthium (S. gramineum ) and
Section Stenanthella (S. occidentale, S. frigidum and S. sachalinense ).

Epigyny in S. occidentale is most evident in early anthesis because locule open-
ing and ovule placentation occur well below the level at which the adnate perianth
and stamens are freed. Within the epigynous zone, tepal, stamen and dorsal
vasculature are only represented by peripheral compound or composite bundles.
At maturity, the gynoecium is only slightly inferior.

The most common inflorescence in the Veratreae (Melanthium, Zigadenus and
Veratrum) is a large polygamous panicle or compound raceme with the terminal
flowers, especially on the side branches, staminate and those lower on the branches

132

Annals of Carnegie Museum

vol. 56

perfect. Stenanthium gramineum and S. frigidum have this type, while the simple
andromonoecious raceme of S. occidentale is among the most reduced in the tribe.
Outbreeding within this raceme is facilitated by the protandrous flowers with
dimorphic filaments and bilobed or v-shaped tepal nectaries. While S. frigidum
is also reported to have tepal nectaries and S. gramineum not (Zimmerman, 1958;
Kupchan et al., 1961), detailed comparison awaits future study. Amianthium
muscaetoxicum (Utech, 1986) and several species of Schoenocaulon which lack
nectaries are notable exceptions to the varying kinds of tepal nectaries and glands
that are common throughout the Veratreae. The need to document carefully the
relationships between inflorescence type and floral morphology and phenology in
the Veratreae is important in future comparative work.

Three meiotic chromosome counts of n = 8 have been reported for S. occidentale
by Cave (1966, 1970) from Siskiyou County, California and Preece (1956) from
Selway Falls, Idaho. The two other counts for the genus, both 2n = 20 for S.
gramineum (=S. rohustum) (Miller, 1930; Sato, 1942), need critical reexamina-
tion, since a x = 8 base number is common among the Veratreae (Cave, 1970;
Preece, 1956; Fedorov, 1969).

In describing the vascular floral anatomy and carpel morphology within the
Veratreae, Anderson (1940), El-Hamidi (1952), Ambrose (1975, 1980), Sterling
(1982) and Utech (1986) have noted characteristic similarities and differences. A
common vascularization pattern consistently observed involves three compound
vascular bundles composed of fused tepallary and staminal bundles (compound
IT bundles or “zwischenbundel”) and three compound bundles, similar to those
above, fused to a dorsal bundle (compound OT bundle or dorsal-compound
bundle). Major anatomical and morphological differences which are frequently
used to differentiate genera (and species) include inflorescence type, floral ge-
ometry, tepal nectaries, presence or absence of sutural openings and a central
carpellary hole at the lowermost level of ovular insertion, the degree of carpellary
separation below the locular apex (an apocarpous tendency), the number of car-
pellary lateral or septal axial bundles and a hypogynous or perigynous versus an
epigynous condition at the lowermost level of ovular insertion.

The complete pedicel to stigma vasculature of S. occidentale is derived through
a complex series of divisions and fusions from three lower pedicel bundles. In
the upper pedicel three compound OT bundles and three compound IT bundles
are formed via tri-parted divisions which is typical of the Veratreae. Due to an
inferior gynoecium and epigynous perianth, the tepallary, staminal and dorsal
bundle formation is above the level at which the ventral and ovular supplies are
established. Each tepal receives three bundles and the stamens one. There is no
inter-connection between any of the tepal laterals or between laterals and medians.
The ventral supply originates via successive divisions and fusions of continuing
lateral branches following the compound IT bundle formation. Within each of
the three undivided septal arms that subtend the locules in the epigynous zone,
a fusion septal axial bundle with normally arranged xylem and phloem is asso-
ciated with a compound ventral that divides into two ventrals. Each carpel receives
a dorsal, two ventrals and a shared septal axial in the common septal arm. There
is no terminal inter-connection between the ventrals and septal axials which end
in the upper epigynous zone. Each of the two ventrals per carpel supplies two
rows of bitegmic, basitropic, campylotropus ovules which develop into winged
seeds.

Ambrose (1975) and Sterling (1982) reported two simple septal (axial) bundles

1987

U TECH — S TENANTHIUM OCCIDENTALE

133

per septum and two to four accessory placental (ventral) bundles with some carpels
having up to six or more in S. gramineum (=S. robustum) and that the septal
bundles united with the placental bundles below the style. This, however, was
not observed in S. occidentale where a shared fusion septal (SA) bundle and two
placental (V) bundles occur without terminal fusion. That such inter-connection
occurs in the Veratreae is known (Sterling, 1982; Utech, 1986), but which pattern
occurs in the other species of Stenanthium remains to be investigated.

The occurrence in S. occidentale of an inferior gynoecium with an epigynous
perianth at anthesis plus a weakly apocarpous fruiting gynoecium is a most unusual
set of liliaceous characters. Furthermore, the insertion of freed floral parts as well
as their vascularization formation is in a spiral pattern. Above the epigynous zone,
the tepals are connate for a short vertical distance in which staminal epitepaly
also occurs. A central carpellary hole at the gynoecium base is internally contin-
uous with the three locules and the open stylar canal which is lined with non-
papilloid cells. The three carpels are essentially free above the epigynous zone to
the stylodium area. Carpellary separation along the inner septal margins results
in septicidal dehiscence and dispersal of the winged seeds. No septal glands or
nectaries were observed. Raphide idioblasts were not observed in carpellary or
perianth tissue of S. occidentale. However, they were reported to be numerous in
S. gramineum (=S. robustum) (Ambrose, 1975; Sterling, 1982) and occasional in
S. frigidum (Sterling, 1982). The homogenous tannins which characterized the
complete floral epidermis of Amianthium muscaetoxicum (Utech, 1986) were not
observed in S. occidentale.

Acknowledgments

The author thanks the M. Graham Netting Research Fund and the O’Neil Botany Field Fund of
The Carnegie Museum of Natural History for supporting the field work and the lab related preparation
of materials and photographs in the new Biosystematics Laboratory. Mr. William W. Brown deserves
special thanks for his artistic aid in producing the figures. The author gratefully acknowledges the loan
and use of specimens from the following herbaria (CAS, CM, DAO, DS, F, JEPS, HSC, GH, ID, MO,
NCU, NY, ORE, OSC, TEX, UBC, UC, US, WILLU, WIS, and WTU) whose acronyms are cited in
Holmgren et al., 1981.

Literature Cited

Abrams, L. 1940. Illustrated flora of the Pacific states. Vol. 1: Ophioglossaceae to Aristolochiaceae.
Stanford University Press, Stanford, p. 374.

Ambrose, J. D. 1975. Comparative anatomy and morphology of the Melanthioideae (Liliaceae).
Unpublished Ph.D. dissertation, Cornell University, Ithaca, vii + 240 pp.

-. 1980. A re-evaluation of the Melanthioideae (Liliaceae) using numerical analyses. Pp. 65-

81, in Petaloid monocotyledons (C. D. Brickell, D. F. Cutler, and M. Gregory, eds.), Academic
Press, London, ix + 222 pp.

Anderson, C. E. 1940. Some studies on the floral anatomy of the Liliales. Unpublished Ph.D.
dissertation, Cornell University, Ithaca, iv + 142 pp.

Anderson, J. P. 1945. Flora of Alaska and adjacent parts of Canada. Part III: Cyperaceae to
Orchidaceae. Iowa State College Journal of Science, 19(2): 135-205.

Baker, J. G. 1879. Synopsis of the Colchicaceae and aberrant tribes of the Liliaceae. Journal of the
Linnean Society, Botany, 17:405-413, 469-485.

Bentham, G., and J. D. Hooker. 1883. Genera Plantarum. Vol. III. L. Reeve, London, pp. 834-
836.

Brockman, C. F. 1947. Flora of Mount Rainier National Park. United States Government Printing
Office, Washington, D.C., pp. 54-58.

Calder, J. A., and R. L. Taylor. 1968. Flora of the Queen Charlotte Island. Part 1: Systematics
of the vascular plants. Queen’s Printer, Ottawa, xii + 659 pp.

Carter, W. R., and C. F. Newcombe. 1921. A preliminary catalogue of the flora of Vancouver and
Queen Charlotte islands. W. H. Cullin, Victoria, 88 pp.

134

Annals of Carnegie Museum

vol. 56

Cave, M. 1966. Documented chromosome number of plants. Madrono, 18:245-246.

. 1 970. Chromosomes of California Liliaceae. University of California Publications in Botany,

57:2, 20-22, 45, tab. 13.

Clark, L. J. 1976. Wild flowers of the Pacific Northwest from Alaska to northern California. Gray’s
Publishing, Sidney, pp. 28-29.

Dahlgren, R., and H. T. Clifford. 1982. The monocotyledons. A comparative study. Academic
Press, London, xiv + 378 pp.

Dahlgren, R., H. T. Clifford, and P. F. Yeo. 1985. The families of monocotyledons. Structure,
evolution, and taxonomy. Springer- Verlag, New York, xii + 520 pp.

Dorn, R. D. 1977. Manual of the vascular plants of Wyoming. Vol. 2: Haloragaceae to Zygophyl-
laceae. Garland Publishing, New York, pp. 942-960.

El-Hamidi, A. 1952. Vergleichend-morphologische Untersuchungen am Gynoeceum der Unterfam-
ilien Melanthioideae und Asphodelioideae der Liliaceae. Arbeiten aus dem Institut fur allgemeine
Botanik, Universitat Zurich, ser. A, no. 4: 1-50.

Engler, A. 1889. Liliaceae. Pp. 1 0-30, in Die natiirlichen Pflanzenfamilien (A. Engler and K. Prantl,
eds.), Engelmann Verlag, Leipzig, 2:10-91.

Farr, E. R., J. A. Leussink, and F. A. Stafleu. 1979. Index Nominum Genericorum (Plantarum).
Vol. Ill: Pegaeophyton-Zyzygium. Regnum Vegetabile, 102:1671.

Fedorov, A. A. (ed.) 1969. Chromosome numbers of flowering plants. Nauk, Leningrad, 926 pp.
(in Russian).

Ferlatte, W. J. 1974. A flora of the Trinity Alps of northern California. University of California
Press, Berkeley, pp. 169-171.

Fuchs, C. 1 963. Fuchsin staining with NaOH clearing for lignified elements of whole plants or plant
organs. Stain Technology, 38:141-144.

Gates, R. R. 1918. A systematic study of the North American Melanthaceae from a genetic stand-
point. Journal of the Linnean Society, Botany, 44:131-172.

Gray, A. 1873. Botanical contributions by Asa Gray. 2. Determination of a collection of plants
made in Oregon by Elihu Hall during the summer of 1871, with characters of new species and
various notes. Proceedings of the American Academy Arts and Sciences, 8:405.

Hitchcock, C. L., and A. Cronquist. 1973. Flora of the Pacific Northwest. University of Wash-
ington Press, Seattle, pp. 678-697.

Hitchcock, C. L., A. Cronquist, M. Ownbey, and J. W. Thompson. 1969. Vascular plants of the
Pacific Northwest. University of Washington Press, Seattle, Vol. 1:801-802.

Holmgren, P. K., W. Keuken, and E. K. Schofield. 1981. Index Herbariorum. Part I: The herbaria
of the world. Ed. 7, Regnum Vegetabile, 106:1-452.

Hulten, E. 1968. Flora of Alaska and neighboring territories. Stanford University Press, Stanford,
pp. 304-312.

Hutchinson, J. 1973. The families of flowering plants. Vol. II: Monocotyledons. Clarendon Press,
Oxford, Ed. 3, 268 pp.

Jepson, W. L. 1921. A flora of California. University of California, Berkeley, 1(6):263.

. 1923. A manual of the flowering plants of California. University of California, Berkeley, p.

221.

Johansen, D. A. 1 940. Plant microtechnique. McGraw-Hill Book Company, New York, xi + 523 pp.

Krause, K. 1930. Liliaceae. Pp. 227-276, in Die natiirlichen Pflanzenfamilien (A. Engler and K.
Prantl, eds.), Engelmann Verlag, Leipzig, 2(15a):227-390.

Kuijt, J. 1982. A flora of Waterton Lakes National Park. University of Alberta Press, Edmonton,

pp. 400-401.

Kunth, C. S. 1842. Melanthaceae. Enumeratio Plantarum. Hinrichs, Tubingen, 4:189-190.

Kupchan, S. M., J. H. Zimmerman, and A. Alfonso. 1961. The alkaloids and taxonomy of Veratrum
and related genera. Lloydia, 24:1-26.

Larrison, E. J., G. W. Patrick, W. H. Baker, and J. A. Yaich. 1974. Washington wildflowers.
Durham and Downey, Portland, pp. 35-44.

Melchior, H. 1964. Syllabus der Pflanzenfamilien. Vol. II: Angiospermen. Gebruder Bomtraeger,
Berlin, vi + 666 pp.

Miller, E. W. 1930. A preliminary note on the cytology of the Melanthioideae section of the Liliaceae.
Proceeding of the University of Durham Philosophical Society, 8:267-271.

Moss, E. H. 1959. A flora of Alberta. University of Toronto Press, Toronto, pp. 156-161.

Munz, P. A. 1973. A California flora with supplement. University of California Press, Berkeley, p.
1333, Supplement p. 173.

Peck, M. E. 1941. A manual of the higher plants of Oregon. Binfords and Mort, Portland, pp. 1 88—
189.

1987

Utech— Stenanthium occidentale

135

Piper, C. V. 1906. Flora of the State of Washington. Contributions from the United States National

Herbarium, 11:197.

Preece, S. J., Jr. 1956. A cytotaxonomic study of the genus Zigadenus. Unpublished Ph.D. disser-
tation, State College of Washington, Pullman, v + 167 pp.

Rydberg, P. A. 1900. Studies on the Rocky Mountain flora II. The Rocky Mountain species of
Melanthaceae. Bulletin of the Torrey Botanical Club, 27:530-531.

St. John, H., and E. Hardin. 1929. Flora of Mt. Baker. Mazama, 1 1:52-102.

Sass, J. E. 1958. Botanical microtechnique. Iowa State University Press, Ames, xi + 228 pp.

SatO, D. 1942. Karyotype alteration and phylogeny in Liliaceae and allied families. Japanese Journal
of Botany, 1 2:57—16 1 .

Schulze, W. 1978. Beitrage zur Taxonomic der Liliifloren IV. Melanthiaceae. Wissenschaftliche
Zeitschrift der Friedrich Schiller Universitat Jena, Mathematisch-naturwissenschaftliche Reihe,
27(1): 10-21.

Scoggan, H. J. 1 978. The flora of Canada. Part 2: Pteridophyta, Gymnospermae, Monocotyledoneae.

National Museum of Natural Sciences, Publications in Botany, 7(2): 505.

Sterling, C. 1982. Comparative morphology of the carpel in the Liliaceae: Veratreae. Journal of
the Linnean Society, Botany, 84:57-77.

Suksdorf, W. 1923. Washingtonische Pflanzen III. Werdenda, 1(2):6.

Taylor, R. L., and B. MacBryde. 1977. Vascular plants of British Columbia. Technical Bulletin,
4. University of British Columbia Press, Vancouver, pp. 497-508.

Taylor, T. M. C. 1 966. The Lily family (Liliaceae) of British Columbia. British Columbia Provincial

Museum Handbook, 25:68-70, 100.

Utech, F. H. 1978a. Floral vascular anatomy of Medeola virginiana L. (Liliaceae-Parideae = Tril-
liaceae) and tribal note. Annals of Carnegie Museum, 47:169-191.

-. 1978 b. Comparison of the vascular floral anatomy of Xerophyllum asphodeloides (L.) Nutt.

and X. tenax (Pursh) Nutt. (Liliaceae-Melanthioideae). Annals of Carnegie Museum, 47:147-167.
. 1978c. Vascular floral anatomy of Helonias bullata L. (Liliaceae-Helonieae) with a com-
parison to the Asian Heloniopsis orientalis. Annals of Carnegie Museum, 47:169-191.

. 1 978c/. Floral vascular anatomy of Pleea tenuifolia Michx. (Liliaceae-T ofieldieae) and its
reassignment to Tofieldia. Annals of Carnegie Museum, 47:423-454.

- — — . 1978c. Floral vascular anatomy of the monotypic Japanese Metanarthecium luteoviride
Maxim. (Liliaceae-Melanthioideae). Annals of Carnegie Museum, 47:455-477.

— — . 1979a. Floral vascular anatomy of the Himalayan Theropogon pallidus Maxim. (Liliaceae-
Convallarieae). Annals of Carnegie Museum, 48:25-41.

— — 19796, Floral vascular anatomy of Scoliopus bigelovii Torrey (Liliaceae-Parideae = Trilli-
aceae) and tribal note. Annals of Carnegie Museum, 48:43-71.

— — . 1982. Floral vascular anatomy of the east Asian Rohdea japonica (Thunb.) A. W. Roth
(Liliaceae-Convallarieae). Journal of Phytogeography and Taxonomy, 30:27-36.

- — — . 1984. Floral vascular anatomy of Japonolirion osense Nakai (Liliaceae) and its tribal rela-
tionship. Annals of Carnegie Museum, 53:447-461.

- — — . 1986. Floral morphology and vascular anatomy of Amianthium muscaetoxicum (Walter) A.
Gray (Liliaceae- Veratreae) with notes on distribution and taxonomy. Annals of Carnegie Museum,
55:481-504.

Utech, F. H., and S. Kawano. 1975. Biosystematic studies in Erythronium (Liliaceae-Tulipeae) II.

Floral anatomy of E. japonicum Decne. Botanical Magazine (Tokyo), 88:177-185.

— — . 1976. Biosystematic studies in Maianthemum (Liliaceae-Polygonatae) VIII. Vascular floral
anatomy of M. dilatatum, M. bifolium and M. canadense. Botanical Magazine (Tokyo), 89:145-
157.

— — -. 1 980. Vascular floral anatomy of the Japanese Paris tetraphylla A. Gray (Liliaceae-Parideae).

Journal of Phytogeography and Taxonomy, 28:17-23.

. 1981. Vascular floral anatomy of the east Asian Heloniopsis orientalis (Thunb.) C. Tanaka

(Liliaceae-Helonieae). Botanical Magazine (Tokyo), 94:295—3 1 1 .

Voss, E. G. (ed.) 1983. International Code of Botanical Nomenclature adopted by the Thirteenth
International Botanical Congress, Sydney, August 1981. Regnum Vegetabile, 111:324.
Zimmerman, J. H. 1958. A monograph of Veratrum. Unpublished Ph.D. dissertation, University
of Wisconsin, Madison, iii + 32 1 pp.

131!

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 6, Pp. 137-160

15 May 1987

RESULTS OF THE CARNEGIE MUSEUM OF NATURAL
HISTORY EXPEDITIONS TO BELIZE.

III. DISTRIBUTIONAL NOTES ON
THE BIRDS OF BELIZE

D. Scott Wood
Assistant Curator, Section of Birds

Robert C. Leberman

Bird Bander, Section of Birds

Abstract

Much new information on the temporal and geographic distributions of the birds of Belize (Central
America) has been gathered since the publication of Russell’s (1964) monograph. We present here our
data on bird distributions gathered during five expeditions to Belize from 1971 to 1985.

Introduction

Belize (until 1973 the British Crown Colony called British Honduras) has be-
come in recent years an increasingly attractive area of Central America to students
of ornithology. Relatively close to the United States, safe, English-speaking, and
possessing a wealth of tropical life forms, it is, perhaps surprising that so little
has appeared in print concerning the avifauna of the country. This may be ex-
plained in part by the high reputation enjoyed by Russell’s (1964) excellent mono-
graph on the distribution of the birds of the area. However, considerable additional
fieldwork has been conducted in Belize in the past 20 years, yielding much new
information on the distribution and status of the birds there.

Among the more important published records of Belizean birds since Russell’s
(1964) report are the annual Christmas Bird Counts (Young, 1973-1985) which
cover two areas (Belize City and Belmopan) near the center of the country. Im-
portant additions have also been provided by Barlow et al. (1969, 1970, 1972),
Erickson (1977), and Kiff and Kiff (1974). Several publications have reported the
observations of British bird watching groups (Jenkins, [1983]; Hallchurch, 1982;
Sherrard-Smith, 1982). A minimally annotated checklist (Weyer and Young, 1983)
has recently been considerably revised and expanded by Wood et al. (1986). Other
publications have dealt primarily with specific ornithological research projects in
Belize (Kricher et al., 1984; Miller and Tilson, 1985; Orians, 1983; Stacey, 1981).

Representatives of the Section of Birds, The Carnegie Museum of Natural
History, participated in two multidisciplinary expeditions to Belize in March-
April 1984 and in June-July 1985. In addition, Leberman visited Belize on three
other occasions between 1971 and 1983 in conjunction with fieldwork conducted
in association with Manomet Bird Observatory. These five trips resulted in sig-
nificant collections and observations of the country’s avifauna. A primary objec-
tive during the 1984 and 1985 expeditions was to collect anatomical materials
(skeletons and specimens preserved in alcohol). We preserved 858 specimens of

Submitted 30 October 1986.

137

138

Annals of Carnegie Museum

vol. 56

191 species in alcohol and 839 specimens of 188 species as skeletons. For most
of the skeletal preparations a spread feathered wing was also saved. In addition,
we preserved 96 specimens of 73 species as traditional study skins; most of this
latter group were either specimens in molt or species unusual to Belize. Also, our
records include field observations of more than 350 species.

The purpose of this report is to comment on the distributions of the birds of
Belize, especially where our observations supplement or modify Russell’s (1964)
findings. Thus, the accounts are presented below only for those species for which
our information adds to or significantly changes what has already been reported
in the literature.

Gazetteer of Localities

Russell’s (1964) monograph provided an extensive ornithological gazetteer for
Belize, many sites of which were visited during the course of our own fieldwork.
However, we also visited and collected at a number of sites not listed by him.
The locations listed below are mentioned in the individual species comments
(with one exception, Cubetas, which is a previously unidentified Russell location).
Each site is identified by latitude and longitude and referenced to a location on
a major world atlas (National Geographic Society, 1981), to identified sites on
both the 1:750,000 Belize map (Directorate of Overseas Surveys, United King-
dom, 1981) and on the 1:50,000 topographic map series (quadrangles numbered
1-42; Directorate of Overseas Surveys, United Kingdom, 1966-1980), and to one
of Russell’s (1964) numbered gazetteer localities. These maps are referred to in
the locality descriptions as NG Atlas, Belize map, Topo map (number), and Russell
(number), respectively. Three sites are of camps set up by Carnegie Museum
teams away from any named locality; each is identified as a CM camp.

Altun Ha: Belize District; 17°45'N, 88°21'W; marked on Topo map 11 and NG Atlas; 2 km W of
Cowhead Creek (on old Northern Highway) on Belize map; 14 km S of Maskall (Russell 6); same
location as Rockstone Pond (Barlow et al. 1969, 1970). This is an excavated Mayan min main-
tained as an archaeological preserve. The habitat is primarily heavy second growth forest with
several large stands of cohune palms ( Orbignya cohune). The area around the main temples is
kept cleared of heavy vegetation. A small pond (cenote) surrounded by forest is situated near the
excavated sites.

Big Fall: Toledo District; 16°15'N, 88°52'W; indicated on Belize map and Topo maps 38 and 42; 18
km N, 8 km W of Punta Gorda (NG Atlas); village at the crossing of the Southern Highway over
the Rio Grande. Our collecting locality was 3 km SE of the bridge and on the south side of the
river. This site is gallery forest along the river with mixed cultivation, milpa, huamil, and second
growth forest away from the flood plain. This site should not be confused with another important
birding locality, Big Falls, Belize District, a rice plantation 40 km W of Belize City on the Belize
River, referred to by Barlow et al. (1972); Hallchurch (1982); Jenkins [1983], and others.

Chaa Creek: Cayo District; 17°07'N, 89°04'W; a private resort 5 km S of San Ignacio on the western
bank of the Macal River (= Eastern Branch of the Belize River [Belize map]). The area is very
hilly and much of the land has been cleared for pasture. A large stand of cohune palms is located
just west of the resort property. Uncleared land is primarily young second growth forest.

Chan Laguna: Orange Walk District; 18°07'N, 88°29'W; a small fresh water lagoon and cattail marsh
marked on Topo map 5; 3 km S, 1 km E of San Estevan (Belize map); 7 km E, 3 km N of Orange
Walk Town (NG Atlas, Russell 2). The area was originally scrub forest but has largely been cleared
for the growth of sugar cane.

CM Bladen Camp: Toledo District; 1 6°33'N, 88°43'W; point at which the Bladen Branch of the Monkey
River exits the foothills of the Maya Mountains; 2 km NW of Chun Bank (Topo map 35); 12
km N, 1 km E of Medina Bank (Belize map); 26 km W, 22 km N of Monkey River Town (NG
Atlas, Russell 85). This point is in the ecotone between the low savanna and the steep forested
foothills. The elevation of the river at this site is approximately 40 m with the immediately
adjacent hills rising to 140 m. Most of the forest is high second growth with a few very large trees
remaining, especially around the site of an old sawmill.

1987

Wood and Leberman — Birds of Belize

139

CM Chiquibul Camp: Cayo District; 16°36'N, 89°00'W; elevation approximately 600 m; point where
a new logging road south from Millionario and Grano de Oro Camp crosses the Chiquibul River;
close to spot marked Moses Head Camp on Topo map 33; 12 km E, 1 1 km S of Round Hole
Bank (Belize map); 60 km S, 5 km E of San Ignacio (NG Atlas); 16 km S, 1.5 km W of Millionario
(Russell 73). The camp was in high forest with many large trees although much of the more
valuable timber had selectively been removed.

CM Columbia Forest Camp: Toledo District; 16°17'N, 89°01'W; elevation approximately 140 m; 1
km N of site marked Forestry Camp on Topo map 37; 1.5 km E, 0.5 km S of Jimmy Cut (Belize
map); 5 km N of San Antonio (NG Atlas); 7 km W, 2 km N of San Pedro Columbia (Russell
86). This area was originally high forest (well logged) but is now milpa and huamil (except for
the steepest and rockiest hillsides) for a distance of 1-2 km from the main roads. Our collecting
localities were largely at the edges of the huamil and into the forest.

Coco Plum Cay: Stann Creek District; 16°53'N, 88°07'W; marked on Belize map; 1 1 km E, 9 km S
of Stann Creek Town [Dangriga] (NG Atlas); 4 km W of Tobacco Cay (Russell 60).

Columbia Forest Station: Toledo District; 16°17'N, 89°01'W; marked Forestry Camp on Topo map
37; 1 km S of CM Columbia Forest Camp (q.v.). Also called Salamanca, this site is now a military
training camp.

Cubetas: Cayo District; 16°47'N, 89°02'W; marked on Topo map 28; 3 km N of Millionario (Russell
73). Mentioned by Russell in his text but not included in his gazetteer.

The Dump: Toledo District; 16°14'N, 88°57'W; marked on Topo map 42; 1 km WNW of junction of
Southern Highway and Punta Gorda-San Antonio Road (Belize map, NG Atlas); considerable
open marsh and wet meadow area traversed by the Punta Gorda-San Antonio Road (large quan-
tities of fill had to be dumped there to make the road, hence the name). This has become a
frequent stop for birdwatchers.

Garbutts Cay: Stann Creek District; 16°59'N, 88°05'W; marked on Belize map; 14 km E of Stann
Creek Town [Dangriga] (NG Atlas); 13 km N of Tobacco Cay (Russell 60).

Guanacaste Park: Cayo District; 17°16'N, 88°46'W; park on the banks of the Belize River and Roaring
Creek at the junction of the Western and Hummingbird highways; just east of the town of Roaring
Creek. The park is mostly second growth forest.

Jimmy Cut: Toledo District; 16°17'N, 89°02'W; marked on Topo map 37 and Belize map; 6 km N,
1 km W of San Antonio (NG Atlas). Formerly high forest, all areas near roads are now milpa.

Las Lomitas: Toledo District; 16°29'N, 88°35'W; fire lookout station 3 km S, 2 km W of junction of
the Southern Highway and the Swasey Branch of the Monkey River (Belize map); 1 7 km NW of
Monkey River Town (NG Atlas, Russell 85). This is an area of low hills covered in pine forest-
pine savanna with dense brush along the waterways.

Milepost 24-30: Western Highway: Belize District; 17°22'N, 88°32'W (Milepost 30); 24 km E, 12 km
N of Belmopan. An old stretch of the original Western Highway parallels the present highway to
the south between these mileposts; this old road is passable for most of its length and is used as
access for several properties. The area between the old and new highways is mostly pine savanna
and includes a fairly extensive grass and sedge marsh near Milepost 24. Our collecting sites were
primarily between the old and new highways.

Salamanca: Toledo District; Alternate name for Columbia Forest Station (q.v.).

Union Camp: Toledo District; 16°24'N, 89°09'W; elevation 720 m; marked on Topo map 37; just SW
of Little Quartz Ridge (Belize map); 1 6 km N, 1 1 km W of San Antonio (NG Atlas). This camp
was visited by the Royal Air Force group (see below; Jenkins, [1983]).

Species Accounts

Within each of the accounts, localities are either as described above or are keyed
to identified sites in the National Geographic Atlas (National Geographic Society,
1981), or to sites on the 1:750,000 Belize map (Directorate of Overseas Surveys,
United Kingdom, 1981).

Three organized groups are mentioned in several of the following accounts: The
[British] Royal Air Force Ornithological Society mounted an expedition to Belize
from February through April 1981; their records, summarized in an extensive
report (Jenkins, [1983]), are credited to the RAFOS. Also included in this report
is a summary of mist-netting at several locations from April through June 1979.
The [British] Army Bird Watching Society mounted an expedition to Belize during
February and March 1982, and published their records in a journal of very limited

140

Annals of Carnegie Museum

vol. 56

distribution (Hallchurch, 1982); their sightings are credited to the ABWS. Several
groups connected with Manomet Bird Observatory have participated in banding
operations in Belize; their observations are credited to the Manomet group or to
Mrs. Erma J. Fisk who organized the first two trips.

The classification and nomenclature used in this report follow the sixth edition
of the Check-list of North American birds and supplement (American Ornithol-
ogists’ Union, 1983, 1985). Specimens listed are in the collections of The Carnegie
Museum of Natural History, Pittsburgh, Pennsylvania. The state of preparation
of a specimen may be inferred from the catalog number given: A = complete
specimen stored in alcohol; P = study skin; S = skeleton, usually with an associated
spread feathered wing.

TIN AMOUS T INAMIDAE

Thicket Tinamou, Crypturellus cinnamomeus

One seen at close range in pine savanna habitat in Orange Walk District near
the new Northern Highway just north of the Belize District line, 1 July 1985.
Very few records of Thicket Tinamous exist for Belize. Russell (1964) listed the
species as uncommon near Hill Bank (approximately 45 km SW of the present
locality) in dense second growth forest and rare near Gallon Jug, also in forest.
No other published records are available; the status listed in Wood et al. (1986)
reflects Weyer’s belief that the species occurs (very rarely) in the southern hard-
wood forests of the country and in the Mountain Pine Ridge.

GREBES PODICIPEDIDAE

Pied-billed Grebe, Podilymbus podiceps

One seen on Crooked Tree Lagoon, 11 June 1985; one seen on Chan Laguna,
27 June 1985. Although Pied-billed Grebes are common winter residents through-
out much of Belize, few summer records exist. Russell (1964) listed only two,
both from May. The present records further support the belief that this species
breeds in Belize. Monroe (1968) reported breeding of Pied-billed Grebes in Hon-
duras.

CORMORANTS PHALACROCORACIDAE

Olivaceous Cormorant, Phalacrocorax olivaceus

Common throughout Belize except on the cays. Records include: one collected
(CM-A3837) of three seen on the Sibun River just north of the Hummingbird
Highway bridge, 20 March 1984; two seen repeatedly near the CM Bladen Camp,
3-7 April 1984; two seen 5 km south of San Ignacio on the Macal River (Eastern
Branch, Belize River), 16 June 1985; two seen repeatedly near the Guacamallo
Bridge on the Macal River, 23-24 June 1985. Russell (1964) lists relatively few
records for the country. The RAFOS also reported these birds to be common,
and saw large numbers (50) on Cay Chapel.

ANHINGAS — — ANHINGIDAE

Anhinga, Anhinga anhinga

One seen flying over savanna at Milepost 29, Western Highway, 19 February
1983. One seen repeatedly at Chan Laguna, 28-30 June 1985. Russell (1964) lists
few records for Belize. The RAFOS recorded up to 30 at Northern Lagoon and

1987

Wood and Leberm an— Birds of Belize

141

many elsewhere; the ABWS recorded four to six daily at Big Falls and the Airport
Camp.

FRIG ATEBIRDS — — FREG ATIDAE

Magnificent Frigatebird, Fregata magnificens

A very common species along the coast and on the cays, these birds occasionally
stray inland: one was seen at Milepost 29 on the Western Highway, 4 July 1985.

BITTERNS AND HERONS ArDEIDAE

Pinnated Bittern, Botaurus pinnatus

Breeding male collected (CM-S8655; testes 25 x 12 mm) at Milepost 21 on
the Western Highway, 14 March 1984; non-breeding male found dead by Dora
Weyer (CM-S 10296; flat skin and skeleton; testes 9.2 mm), Milepost 14 on the
Northern Highway, 21 April 1985; two seen several times at Milepost 24 on the
Western Highway, July 1985. All records were from open savanna, especially
marshy areas. This species is cryptic and difficult to observe. Wood flushed the
specimen collected in 1984 from a distance of less than 10 m; it was unseen until
it flew. In addition, these birds appear to be restricted to a habitat not usually
frequented by birders. This, perhaps, accounts for the lack of published records
for the country: Russell (1964) did not list the species (although he mentions
Yucatan records). The first record and specimen for Belize was one found by Dora
Weyer in 1970, also in savanna habitat (Barlow et al. 1972). Two were also seen
during the 1982 Belize City Christmas Bird Count (Young 1973-1985). The
bitterns seen by the ABWS (1982) at Big Falls (as many as five in one day) were
probably this species. We feel certain that Pinnated Bitterns are much less rare
than the records indicate; additional fieldwork is needed in wet savanna habitats.

American Bittern, Botaurus lentiginosus

One wing salvaged by Meg Price Craig (CM-P 144864) from a bird shot by a
local hunter near Mussel Creek (approx. 25 km WNW Belize City) in early Feb-
ruary 1972. This is the first specimen record of this bittern for Belize; Russell
(1964) listed two sight records. The bittern reported by the ABWS (1982) from
The Dump may also have been this species.

Least Bittern, Ixobrychus exilis

Female collected (CM-S 10235; largest ovum 6 mm) at Chan Laguna, 29 June
1985. Russell (1964) listed only two records: a transient collected on Half Moon
Cay and a winter sight record for Hill Bank. Both the ABWS and RAFOS recorded
at least three within 25 km of Belize City. Weyer (personal communication) has
observed this species several times in fresh water lagoons along the coastal plain.
We consider it to be an uncommon and local permanent resident with a small
augmentation of winter visitors from the north.

Bare-throated Tiger-Heron, Tigrisoma mexicanum

Common in coastal pine ridges, especially in the wetter areas, but somewhat
secretive. Among our records is one of four birds repeatedly seen in a marsh near
Milepost 24 of the Western Highway, 26 June to 4 July 1985. This species is
listed as uncommon in northern forests by Russell (1964). Between one and 10
have been seen on all but two Belize City Christmas Counts (Young, 1973-1985).

142

Annals of Carnegie Museum

vol. 56

Great Blue Heron, Ardea herodias

One found dead (CM-S8644) in the Belize River near Burrell Boom, 5 March

1984, is the first specimen of the species for Belize. One seen at Crooked Tree
Lagoon, 1 1 June 1985; one seen at Milepost 24 of the Western Highway, 26 June

1985. No previous published records exist for the period June through October,
although Weyer (personal communication) believes that a few breed each year on
the northern cays. Griscom (1926) found the species breeding on the Culebra
Keys, Quintana Roo, in January 1925.

Great Egret, Casmerodius albus

An estimated 1200 seen at Crooked Tree Lagoon, 1 1 June 1985. Russell (1964)
reported no documentation of breeding, but the present record strongly suggests
that these birds nest in the country. Breeding is also to be expected since it has
been reported from surrounding countries (Monroe, 1968; Paynter 1955).

Snowy Egret, Egretta thula

An estimated 100 seen at Crooked Tree Lagoon, 1 1 June 1985. Russell (1964)
listed no records for June.

Little Blue Heron, Egretta caerulea

An estimated 250 seen at Crooked Tree Lagoon, 1 1 June 1985. Russell (1964)
listed no records for the period 1 5 May-19 July. Weyer (personal communication)
has recorded this species breeding on the northern cays.

Tricolored Heron, Egretta tricolor

Three seen at Crooked Tree Lagoon, 11 June 1985. Russell (1964) listed no
records for June or July.

Reddish Egret, Egretta rufescens

Two seen at Crooked Tree Lagoon, 11 June 1985. Russell (1964) listed only
one sight record and Kiff and Kiff (1974) reported the first specimen for the
country. The ABWS (1982) recorded two on Cay Chapel. Weyer (personal com-
munication) has recorded this species breeding on two northern cays.

Cattle Egret, Bubulcus ibis

Abundant throughout the country in suitable habitat. Russell (1964) listed only
a few records and thought that the species had only just become established as
breeding birds at that time.

Green-backed Heron, Butorides striatus

An estimated 25 seen at Crooked Tree Lagoon, 11 June 1985. Russell (1964)
listed no June or July records and knew of no breeding records although he believed
it to nest.

Black-crowned Night-Heron, Nycticorax nycticorax

Three seen along Black Creek (near Crooked Tree Lagoon), 5 March 1984; two
seen at Sand Hill, Belize District, 9 March 1984. Russell (1964) considered this
species to be a rare transient and winter resident and listed only four records.
However, we consider the bird to be fairly common; Young (1 973-1 985) recorded
the species in most years for the Belize City Christmas Count (2-13 individuals
per count) and both the RAFOS and ABWS encountered several birds. Weyer

1987

Wood and Leberman —Birds of Belize

143

(personal communication) believes some individuals may remain throughout the
summer months.

IBISES AND SPOONBILLS THRES KIORNITHIDAE

White Ibis, Eudocimus albus

Eight seen near Sand Hill, Belize District, 17-19 March 1984; 15 seen just west
of Dangriga Airport, 21 March 1984; 200 seen at Crooked Tree Lagoon, 1 1 June
1985; small numbers seen repeatedly southwest of Belize City flying over the
mangroves, 1 1 June-8 July 1985. Russell (1964) stated that there were no recent
records of White Ibis for Belize. However, Erickson (1977) found 1 0 near Dangriga
on 26 December 1976 and Young (1973-1985) listed the species for the first three
Belize City Christmas Counts. Our records, those of the RAFOS and ABWS (up
to 100 individuals per day), and records from more recent Christmas Counts (244
recorded on the 1983 Belize City count), suggest that these birds are increasing
in numbers in Belize.

Roseate Spoonbill, Ajaia ajaja

Fifteen seen at Crooked Tree Lagoon, 1 1 June 1985. Russell (1964) listed only
two records for the country. Young (1973-1985) recorded one bird on the Belize
City Christmas Count in 1983 and again in 1984. The RAFOS recorded three
just outside of Belize City. These birds are probably more common than the
records indicate, but may be largely confined to the northeastern part of the
country.

SWANS, GEESE AND DUCKS -AnATIDAE

Black-bellied Whistling-Duck, Dendrocygna autumnalis

Six seen at Crooked Tree Lagoon, 1 1 June 1985; two seen Milepost 29 on the
Western Highway, 27 June 1985; two seen at Chan Laguna, 30 June 1985. Russell
(1964) considered this species rare in Belize but remarked that it seemed to be
increasing. Subsequent published notes (Barlow et al., 1969, 1972) and conver-
sations with Dora Weyer and others in Belize, as well as our own observations,
suggest that the species is now relatively common in the northern half of the
country.

KITES, EAGLES, HAWKS AND ALLIES— ACCIPITRIDAE

Hook-billed Kite, Chondrohierax uncinatus

One seen each date at Columbia Forest Station, 5 December 1971 (Leberman
and K. S. Anderson), 25 March 1972 (Leberman). Russell (1964) listed only two
records for this rare hawk, neither from the southern forests. The only other
published records are from the Belmopan Christmas Counts (Y oung, 1973-1985):
one seen in 1982 and one in 1984.

Double-toothed Kite, Harpagus bidentatus

One seen along Hummingbird Highway northwest of Caves Branch, 23 February
1983 (Leberman and Manomet group); male collected (CM-S8722; testes 8 x 4.5
mm) 1 km north of Columbia Forest Station, 22 March 1984. The only other
published records for this species in Belize are one in Russell (1964) and two
from the Belmopan Christmas Counts (Young, 1973-1985): one seen in 1978 and
three in 1980.

144

Annals of Carnegie Museum

vol. 56

Solitary Eagle, Harpy haliaetus solitarius

One seen soaring above Guacamallo Bridge, 24 June 1985. Russell (1964) does
not list this species for Belize. Weyer (personal communication) indicates that
the bird we saw is probably one of a pair that has nested near Guacamallo Bridge
for the past several years.

Broad-winged Hawk, Buteo platypterus

One captured, photographed (photo on file at CM) and released along the
Hummingbird Highway south of Belmopan, 24 February 1983 (Leberman and
Manomet group). The only other published records for this very rare transient
(and winter visitant?) are the specimen (taken 22 October 1906) mentioned by
Russell (1964) and one bird seen on the 1973 Belize City Christmas Bird Count
(Young, 1973-1985). Broad-wings should be looked for during the fall since they
migrate through Central America in large numbers (Monroe, 1968).

Swainson’s Hawk, Buteo swainsonii

One seen at Caves Creek, 23 February 1983 (Leberman and Manomet group);
one seen soaring over CM Columbia Forest Camp each day, 23-24 March 1984;
one seen along Hummingbird Highway, 6 April 1984 (Manomet group). This
species is not listed for Belize by Russell (1964), but has been frequently recorded
(1-4 individuals) on both Belizean Christmas Counts (Young, 1973-1985).

Red-tailed Hawk, Buteo jamaicensis

One immature seen flying at close range just northwest of Belize City, 16 March
1984. Russell (1964) listed only Mountain Pine Ridge records. The only other
published record is of one seen on the 1981 Belmopan Christmas Count (Young,

1973-1985).

Harpy Eagle, Harpia harpyja

One immature seen in high forest northwest of Columbia Forest Station near
Jimmy Cut by Leberman and K. S. Anderson, 28 March 1972. The only other
published record is from Gallon Jug (Russell, 1964).

Black-and-white Hawk-Eagle, Spizastur melanoleucus

One seen perched at Beaver Dam Creek (near Milepost 38 of the Western
Highway, Cayo District), 25 March 1984 (Manomet group); one seen at the CM
Bladen Camp, 3 April 1984; one seen at Milepost 29 of the Western Highway,
26 June 1985; one seen at Milepost 24 of the Western Highway, 3 July 1985 (the
latter two records are likely of the same bird). This rare hawk seems to prefer the
forest edge. Russell (1964) lists only three records but the species has also been
seen three times on the Belmopan Christmas Counts (Young, 1973-1985).

Black Hawk-Eagle, Spizaetus tyrannus

One seen flying over Altun Ha, 18 March 1984. Black Hawk-Eagles are the
most common of the three hawk-eagles in Belize despite the single record of
Russell (1964); one or two individuals have been recorded on most of the Bel-
mopan Christmas Counts and the species has also been recorded twice on the
Belize City Christmas Count (Young, 1973-1985). Erickson (1977) saw one near
Belmopan and the RAFOS recorded two in the Columbia Forest region. The
present sighting is farther north than other published records for Belize; we think
this species could be encountered anywhere in the forested regions of the country.

1987

Wood and Leberm an— Birds of Belize

145

Ornate Hawk-Eagle, Spizaetus ornatus

One seen at Milepost 30 of the Western Highway, 1 9 February 1983 (Leberman);
one seen over the Sibun River southeast of Milepost 36 on the Western Highway,
5 July 1985. Although Russell ( 1 964) considered this species as merely uncommon
we think the birds quite rare and even less frequent than Black Hawk-Eagles.
Ornate Hawk-Eagles have been recorded only once (Belmopan, 1984) on the
Christmas Counts (Young, 1973-1985). The present records support Russell’s
observation that they occasionally forage in pine savannas, away from hardwood
forests.

CARACARAS AND FALCONS— — FALCONIDAE

Barred Forest-Falcon, Micrastur ruficollis

One seen CM Bladen Camp, 7 April 1984; breeding adult female (CM-S9081;
ovaries 14 x 7, 16 x 8 mm) collected near Rio Frio Cave, Augustine, Cayo
District, 16 April 1984; immature molting into adult plumage captured, photo-
graphed and released at Guacamallo Bridge, 24 June 1985; one seen near San
Estevan, Orange Walk District, 29 June 1985. This secretive falcon is undoubtedly
more common than the few published records in Russell (1964) indicate. Most
of our encounters with the species occurred when the falcon was attracted by
struggling mist-netted birds.

Aplomado Falcon, Falco femoralis

At least five individuals seen hunting through smoke of a pine savanna fire near
Milepost 29 of the Western Highway, 31 March 1984 (Manomet group); one seen
at Las Lomitas, 9 April 1984; two seen in pine forest 4 km N and 1.5 km W of
Melinda Forest Station, 19 June 1985. We consider this species to be more
common in Belize than reported by Russell (1964). Young (1973-1985) has re-
ported sightings for five of the Belize City Christmas Counts.

Orange-breasted Falcon, Falco deiroleucus

Adult and two immatures seen at close range sitting together in a dead snag
near our CM Chiquibul Camp, 21 June 1985. Haney (1983) recorded this rare
falcon not far north of our site and several other observations of these birds have
been made in the Chiquibul in the past few years (Weyer, personal communi-
cation); the species apparently breeds in this area. The RAFOS encountered one
individual at Lubaantun, the Maya ruin near San Pedro Columbia, 4 April 1981.
These are the only published records of the species from Belize.

RAILS, GALLINULES AND COOTS RALLIDAE

Common Moorhen, Gallinula chloropus

Two seen at Crooked Tree Lagoon, 25 February 1983 (RCL and Manomet
group); 20 seen at Sand Hill, Belize District, 9 March 1984; two seen on the New
River at Orange Walk Town, 30 June 1985. Russell (1964) reports only a few
sight records for this species. However, Barlow et al. (1969) reported three spec-
imens taken from Rockstone Pond in April, Erickson (1977) noted the species
between Belize City and Orange Walk in December, and Young (1973-1985)
recorded them (sometimes in considerable numbers) on the Belize City and Bel-
mopan Christmas Counts. The present summer record suggests these birds occur
throughout the year, as they do in Honduras (Monroe, 1968).

146

Annals of Carnegie Museum

vol. 56

PLO VERS —— CHARADRIIDAE

Black-bellied Plover, Pluvialis squatarola

Two or three seen on Tobacco and Coco Plum cays, 1 1-12 March 1984. Russell
(1964) lists only three records for the country but Young (1973-1985) records the
species on all Belize City Christmas Counts except two and the RAFOS and
ABWS found them at Belize City, Bangriga and Cay Chapel. This species is
probably common on the cays during winter and migration periods.

Wilson’s Plover, Charadrius wilsonia

Two seen and photographed on Garbutts Cay, 10 March 1984. Russell (1964)
lists very few records but Barlow et al. (1969) cite a number of additional records
and specimens. The RAFOS encountered single birds at Bangriga and Punta
Gorda, and the ABWS encountered 40 on Cay Chapel.

STILTS AND AVOCETS— ReCURVIROSTRIDAE

Black-necked Stilt, Himantopus mexicanus

Approximately 80 seen at Crooked Tree Lagoon, 27 March 1984 (Manomet
group); six seen at Crooked Tree Lagoon, 1 1 June 1985. Russell (1964) listed only
three records for Belize, all from spring, and Young (1973-1985) recorded the
species on three Belize City Christmas Counts. The RAFOS and ABWS recorded
considerable numbers (up to 40) in the Belize City area. Weyer (personal com-
munication) reports that the species has been found nesting within the country
in recent years.

SANDPIPERS AND ALLIES— SCOLOP ACID AE

Greater Yellowlegs, Tringa melanoleuca

Two seen at Crooked Tree Lagoon, 11 June 1985. Russell (1964) listed very
few records for this species, all from the period August-March. It was unrecorded
on the Christmas Counts until Becember 1983 when 15 were seen in Belize City
(Young, 1973-1985). The RAFOS and ABWS recorded the species from Big Falls
Ranch.

Lesser Yellowlegs, Tringa flavipes

Four seen at Sand Hill, 9 March 1984. Russell (1964) listed very few records
for this species (all from the period February-March) and it has been recorded
only four times on the Christmas Counts (Young, 1973-1985). However, the
RAFOS and ABWS encountered as many as 60 at a time at Big Falls Ranch.

Willet, Catoptrophorus semipalmatus

Four seen at Bangriga, 20 March 1984. Russell (1964) listed very few records
for this species and it has been recorded only twice on the Christmas Counts
(Young, 1973-1985). Erickson (1977) noted it at Belize City and Bangriga in
Becember. The RAFOS and ABWS recorded small numbers at many coastal
localities and on Cay Chapel.

Upland Sandpiper, Bartramia longicauda

Female collected (CM-P1 62522; ovary 9x5 mm) of two of this species seen
in dry savanna just south of the CM Bladen Camp. This is the first published
record in Belize since 1889 (Russell, 1964) but small numbers undoubtedly stop
each year during migration; additional fieldwork is needed in grassland areas.

1987

Wood and Leberman— Birds of Belize

147

Long-billed Curlew, Numenius americanus

One seen in flight (cinnamon wing linings and enormous bill clearly visible) at
Dangriga, 21 March 1984. This is the first published record for Belize but the
species has been noted occasionally in the past (Weyer, personal communication).
Belize is near the southern limit of the winter range of this species (American
Ornithologists’ Union, 1983).

SKUAS, GULLS, TERNS AND SKIMMERS- LaRIDAE

Caspian Tern, Sterna caspia

Five seen at Crooked Tree Lagoon, 11 June 1985. Russell (1964) listed only
three records and Erickson (1977) added a fourth (at Dangriga) but Young (1973-
1985) recorded the species on every Belize City Christmas Count (as many as 32
individuals) except two.

Sandwich Tern, Sterna sandvicensis

Ten seen just north of Belize City, 9 March 1984. Erickson (1977) summarized
the few records for Belize to that time. The RAFOS and ABWS encountered this
species frequently along the coast and considered it common.

PIGEONS AND DOVES— COLUMBIDAE

Red-billed Pigeon, Columba flavirostris

Two to five seen at Chaa Creek each day, 20-23 April 1984; two collected (CM-
A4458, 4485) at Chaa Creek, 2 1 and 23 April 1984; four collected (male, 2 females,
immature female; CM-S 10051, 10052, 10053, 10054) of at least 40 seen at Chaa
Creek, 16 June 1985. Russell (1964) listed only one record (Corozal). Barlow et
al. (1970) listed several records from Rockstone Pond and Erickson (1977) noted
one near Dangriga. This species has been recorded on most Belize City Christmas
Counts and twice on the Belmopan count (Y oung, 1973-1985). The RAFOS also
recorded several individuals at Guacamallo Bridge. Despite the scarcity of records
from the western part of the country, it was the commonest pigeon in the Chaa
Creek area. This species deserves more attention in Belize since it is considered
rare elsewhere on the Caribbean slope of Central America (American Ornithol-
ogists’ Union, 1983).

White-tipped Dove, Leptotila verreauxi

Common in the northern half of the country; specimens from Milepost 30 of
the Western Highway, Chaa Creek, Chan Laguna. Russell (1964) listed only two
records and Barlow et al. (1969) reported on three specimens from the Rockstone
Pond area. Young (1973-1985), however, has noted the species on every Christ-
mas Count (both Belize City and Belmopan) except one. This dove appears to be
largely absent from the southern half of Belize, the only record being one sighting
by the RAFOS in the Columbia Forest area (Aguacate).

CUCKOOS AND ALLIES- CUCULIDAE

Striped Cuckoo, Taper a naevia

One seen near San Antonio, Toledo District, 4 December 1971 (Leberman, K.
S. Anderson, D. Weyer); one seen near Milepost 29 of the Western Highway, 1 7
and 27 February 1983 (Leberman and Manomet group); one seen near Milepost
33 of the Western Highway, 21 March 1984 (Manomet group). Russell (1964)

148

Annals of Carnegie Museum

vol, 56

noted only two records of this very shy species. Young (1973-1985) listed the
species on five of the Belmopan Christmas Counts.

TYPICAL OWLS— — — SXRIGIDAE

Yermiculated Screech-Owl, Otus guatemalae

Two nestlings brought by a local Indian to Columbia Forest Station, March 28
1972, and subsequently released; one heard at CM Columbia Forest Camp, 26
March 1984; rufous phase male found dead on road (CM-P 162528; testes 5x3
mm) at Milepost 36 of the Western Highway (Cayo District), 4 April 1984. This
latter bird is the first specimen of the species for Belize. The only other published
records are two sightings noted by Russell (1964), one bird reported on the 1981
Belmopan Christmas Count (Young, 1973-1985), individual birds heard by the
RAFOS north of Guacamallo Bridge and in the Columbia Forest, and two birds
mist-netted in the Columbia Forest (Jenkins, [1983]).

Least Pygmy-Owl, Glaucidium minutissimum

Immature found dead (CM-P1 62527; ovary(?) 3x2 mm) in Belmopan, 23
March 1984; one mist-netted and photographed east of the Sibun River (in Belize
District) southeast of Milepost 35 on the Western Highway, 14 April 1984 (Man-
omet group); two to four seen each day at Chaa Creek, 20-23 April 1984: two of
these were seen copulating near their presumed nest hole on 21 April Russell
(1964) listed only one other record for the country.

Ferruginous Pygmy-Owl, Glaucidium brasilianum

Non-breeding male (testes 3x2 mm; CM-S 10251) collected in Orange Walk
District near the new Northern Highway 0.5 km N of the Belize District line, 1
July 1985. Russell (1964) listed few records outside of the Mountain Pine Ridge
but this species is widespread and fairly common in Belize.

Stygian Owl, Asio stygius

One photographed by Ford Young at Milepost 1 5 of the Western Highway (near
Hattieville), 8 November 1971 (fide K. S. Anderson); one seen at Milepost 30 of
the Western Highway, March 1983 (Manomet group); one found dead on road
(CM-A5210) by Dora Weyer at Milepost 30 of the Western Highway, 10 March
1986. Only one published record exists for Belize, a bird collected in the Mountain
Pine Ridge (Russell, 1964).

SWIFTS ApODIDAE

Chimney Swift, Chaetura pelagica

Female (CM-P1 62504; ovary 3.5 x 4 mm) collected from a flock of approxi-
mately 100 swifts 3 km SE of Big Fall, Toledo District, 29 March 1984. This and
similar flocks seen in the same area on 30 March and 1 April contained both
Chimney and Vaux’s (C. vauxi) swifts (as well as two Lesser Swallow-tailed Swifts);
on 30 March at least six Chimney Swifts were present. The size difference between
the two Chaetura species was quite obvious. Russell (1964) doubted the validity
of the two alleged records of Chimney Swift for Belize. However, the Chimney
Swift clearly migrates through the country, at least occasionally. Young (1973—
1985) has recorded a few individuals in the past several years on Christmas Counts.
Monroe (1968) considered this species an uncommon to common fall migrant
and common spring migrant in the coastal lowlands of Honduras.

1987

Wood and Leberman— Birds of Belize

149

Lesser Swallow-tailed Swift, Panyptila cayennensis

Two seen in the large flock of swifts noted above on 30 March. Russell (1964)
considered the species regular and probably nesting at Gallon Jug, but there are
few records from elsewhere in Belize.

HUMMINGBIRDS -T ROCHILIDAE

Band-tailed Barbthroat, Threnetes ruckeri

Two mist-netted Columbia Forest Station, 6, 14 December 1971 (E. J. Fisk);
male (CM-P1 62501, testes 2.5 x 2 mm) collected at CM Columbia Forest Camp,
24 March 1984; male and two of unknown sex (CM-S8787, testes 2.2 mm; CM-
A3956; CM-A3977) collected at CM Columbia Forest Camp, 25 March 1984;
male (CM-P162506, testes 2.5 x 2 mm) collected 3 km SE of Big Fall, Toledo
District, 29 March 1984. Barlow et al. (1972) reported the first record of this
species for the country. The RAFOS and the ABWS mist-netted several in the
Columbia Forest area. Our experience indicates that this species is not particularly
rare in the southernmost parts of Belize.

Scaly-breasted Hummingbird, Phaeochroa cuvierii

One mist-netted at the Columbia Forest Station, 9 December 1971 (E. J. Fisk
et al.); one collected (CM-A3905) at CM Columbia Forest Camp, 23 March 1984;
one collected (CM-A4126) 3 km SE of Big Fall, Toledo District, 31 March 1984;
two collected (CM-S8963 female, ovary 4x4 mm; CM-A497 1) at the CM Bladen
Camp, 4, 5 April 1984. Russell (1964) reported only a few records of this incon-
spicuous hummingbird, all from the northern half of the country. However, the
RAFOS and ABWS each mist-netted one in the Columbia Forest area. The present
records indicate that it is at least as common in the south and perhaps more so
than in the northern forests.

Wedge-tailed Sabrewing, Campylopterus curvipennis

One non-breeding male collected (CM-S8998, testes < 1 mm) of several seen
at the CM Bladen Camp, 6 April 1984; one collected (CM-A4971) of five seen 6
km S, 6 km W of Progress©, Corozal District, 30 June 1985; several others collected
at localities covered by Russell (1964). Russell was the first ornithologist to record
this species in Belize; he considered it “moderately common” but local, whereas
we found it nearly everywhere there was hardwood forest. However, despite their
concentration on hummingbirds, neither the RAFOS nor the ABWS encountered
this species. Hutson and Lyal mist-netted seven in the Columbia Forest area
(Jenkins, [1983]).

Crowned Woodnymph, Thalurania colombica

Three collected (CM-P162502, male, testes <1 mm; CM-S8777, male, testes
2 x 1.5 mm; CM-A3951) at CM Columbia Forest Camp, 24, 25 March 1984.
Russell (1964) reported only one record for Belize, a male taken near our locality.
The RAFOS and ABWS each mist-netted one individual in the Columbia Forest
area and Hutson and Lyal also netted two individuals there (Jenkins, [1983]).

Buff-bellied Hummingbird, Amazilia yucatanensis

Male collected (CM-P 166042, testes 3 x 1.5 mm) 3 km N, 1 km W of the
Melinda Forest Station; male collected (CM-S 10269, testes 1 x 0.5 mm) 1.5 km
east of Milepost 29 of the Western Highway; one seen at close range near Chan

150

Annals* of Carnegie Museum

vol. 56

Laguna, 30 June 1985. Russell (1964) listed only a few records, all from the
lowland pine ridges. Our localities are all lowland but only one was in pines
(Milepost 29 of the Western Highway). The Melinda Forest Station bird was taken
in an island of broadleaf forest along Big Creek, surrounded by pine ridge. The
Chan Laguna sighting was from northern hardwood forest remnants in an area
dominated by sugar cane plantations.

Stripe-tailed Hummingbird, E up herns a eximia

Female collected (CM-P14485 1) by D. Weyer, E. J. Fisk et al. at the Columbia
Forest Station, 6 December 1971. This species was previously recorded only from
the Cockscomb Mountains (Russell, 1 964) and from Union Camp by the RAFOS;
the present record is the first for the lowlands of Belize.

TROGONS f ROGONIDAE

Collared Trogon, Trogon collaris

Male captured, photographed and released near Rio Frio Cave, Augustine, Cayo
District, 15 April 1984; at least three males and one female seen and heard near
the CM Chiquibul Camp, 21 June 1985. Russell (1964) listed only five records
for the country; the RAFOS added one record from the Columbia Forest area
(Aguacate). This species appears to be quite rare north of the Mountain Pine Ridge
and southeast of the Maya Mountains. However, in the Chiquibul it appears to
be somewhat more common.

MOTMOTS -MOMOTIDAE

Tody Motmot, Hylomanes momotula

One captured, photographed and released of two seen repeatedly 1 km S of the
Guacarnallo Bridge, 23, 24 June 1985. Russell (1964) listed few records for Belize
of this uncommon bird.

TOUCANS RAMPHASTIDAE

Emerald Toucanet, Aulacorhynchus prasinus

Female collected (CM-P 166050, ovary 11x7 mm) 12 km S, 1 km W of
Millionario, 22 June 1985. Russell (1964) considered this species very uncommon.
The largest populations may occur in the Chiquibul region.

WOODPECKERS- PlCIDAE

Red- vented Woodpecker, Melanerpes pygmaeus

One adult and one immature male collected (CM-P 166064, testes 5x3 mm;
CM-P166065, testes 1.5 x 1 mm) of at least five seen west of Chan Laguna, 29
June 1985. Russell (1964) listed only two records, but the species has subsequently
been reported by Barlow et al. (1969), Erickson (1977), and on two recent Belize
City Christmas Counts (Young, 1973-1985). All encounters have been in the
coastal lowland forests from Belize City north. We believe this species to be more
common than the published records indicate, although restricted to the north-
eastern portion of the country, an area that has received relatively little ornitho-
logical investigation. The birds we encountered were in remnant patches of hard-
wood forest surrounded by sugar cane plantations.

1987

Wood and Leberman-— Birds of Belize

151

Ladder-backed Woodpecker, Picoides scalar is

Female collected (CM-A3860) at Las Lomitas, 21 March 1984; one seen at
close range in Orange Walk District 0.5 km N of the Belize District line near the
new Northern Highway, 1 July 1 985. The March record lies well within the species’
range as described by Russell (1964). However, the July record considerably
extends to the north the known range of this woodpecker in Belize, narrowing the
gap between the Belize and Yucatan populations.

OVENBIRDS — FURNARIIDAE

Scaly-throated Leaftosser, Sclerurus guatemalensis

Two banded at the Columbia Forest Station, one each day, 13 December 1971
and 24 March 1972 (Manomet group); three collected (male, CM-S8769; female,
CM-S8771; unknown sex, CM-A3949) at the CM Columbia Forest Camp, 24-
25 March 1984. Russell (1964) listed few records for this elusive species and only
one additional sighting has been published: Hutson and Lyal netted one bird in
the Columbia Forest area (Jenkins, [1983]). Our specimens were mist-netted at
the edge of high forest bordered by huamil.

WOODCREEPERS- DENDROCOLAPTIDAE

Strong-billed Woodcreeper, Xiphocolaptes promeropirhynchus

Male collected (CM-S8794: testes 10x6 mm) at CM Columbia Forest Camp,
25 March 1984 in high forest. Russell (1964) listed very few records and none
from the southern parts of the country. This species probably occurs in very low
numbers wherever extensive high forest remains.

Streak-headed Woodcreeper, Lepidocolaptes souleyetii

Male collected (CM-S8953: testes 12x7 mm) of two seen at the CM Bladen
Camp, 4 April 1984. This species was considered quite uncommon by Russell
(1964) and the above records represent the only individuals which we positively
identified during our fieldwork (we collected and saw large numbers of Ivory-
billed Woodcreepers, Xiphorhynchus flavigaster). Our experience echoed that of
the RAFOS who found very few Streak-headed but large numbers of Ivory-billed
Woodcreepers. In contrast, the Streak-headed Woodcreeper is recorded as fre-
quently on both Christmas Counts as the Ivory-billed (Young, 1973-1985). Since
distinguishing these two species in the field is difficult, requiring cautious iden-
tification, we consider the Streak-headed to be much less common than the Ivory-
billed, Christmas Count data notwithstanding.

ANTBIRDS- -FORMICARIIDAE

Great Antshrike, Tar aba major

Three mist-netted along the Sibun River near Milepost 36 of the Western
Highway, 22 February 1983 by the Manomet group; male collected (CM-A3842)
by the Sibun River just east of the Hummingbird Highway bridge, 20 March
1984; two males collected (CM-S8732: testes 2 x 1 mm; CM-A3986) of four
birds seen at CM Columbia Forest Camp, 23-26 March 1984; female collected
(CM-S8909: ovary 5x3 mm) at Big Fall, Toledo District, 1 April 1984. Russell
(1964) listed only two records for the country. Young (1973-1985) recorded the
species on three separate Christmas Counts. The ABWS recorded one in the

152

Annals of Carnegie Museum

vol. 56

Columbia Forest Region. Great Antshrikes appear to be very local and confined
primarily to the southern half of Belize.

TYRANT FLYCATCHERS TYRANNIDAE

TYRANNULETS, ELAENIAS AND ALLIES BlAENIINAE

Ochre-bellied Flycatcher, Mionectes oleagineus

In contrast to Russell’s (1964) comments (“widely distributed . . . but not com-
mon or at least not seen regularly”) we found this species to be the most common
flycatcher in the southern forests of Belize (we encountered as many as ten in-
dividuals in one day in the Columbia Forest as opposed to a maximum of eight
for any other flycatcher).

Sepia-capped Flycatcher, Leptopogon amaurocephalus

One collected (CM-A3863) at CM Columbia Forest Camp, 22 March 1984;
one seen at CM Bladen Camp, 5 April 1984; two collected (CM-A4945; female,
CM-S10125: ovary 4x5 mm) just south of Guacamallo Bridge, 23, 24 June
1985. An uncommon species (Russell, 1964), it probably occurs throughout the
southern forests. The RAFOS recorded this flycatcher three times in the Columbia
Forest area.

FLUVICOLINE FLYCATCHERS FlUVICOLINAE

Ruddy- tailed Flycatcher, Terenotriccus erythrurus

Two banded at Columbia Forest Station, 13 December 1971 (E. J. Fisk et al.);
four collected (CM-A3883; CM-A3926; CM-A4014; female, CM-S8795: ovary
4x3 mm) at CM Columbia Forest Camp, 22-27 March 1984; two seen at CM
Bladen Camp, 3 April 1984 and one seen (same bird?) the next day. Russell (1964)
listed very few records for Belize, mostly from the Columbia Forest area.

Acadian Flycatcher, Empidonax virescens

One banded along the Sibun River southeast of Milepost 35 of the Western
Highway, 20 March 1984 (Manomet group); male collected (CM-P1 62533: testes
2 x 1 mm) 1 km west of Augustine, 15 April 1984, in broadleaf forest. Russell
(1964) listed few records and none from the western part of the country.

“Traill’s” Flycatcher, Empidonax alnorum or E. traillii

The first record of either species for Belize was an individual captured on 8
October 1971 in Barclay’s Bank, Belize City, by James Waite and subsequently
released. This bird had been banded by Leberman at The Carnegie Museum’s
Powdermill Nature Reserve, 4 km S of Rector, Pennsylvania on 12 September
1971. Unfortunately, these two species are only safely separable by voice so we
cannot make positive identification. We know of no subsequent records of either
species from Belize.

White-throated Flycatcher, Empidonax albigularis

Female collected (CM-P 144852) at Columbia Forest Station, 9 December 1971;
one mist-netted at Columbia Forest Station, 23 March 1972, by E. J. Fisk, RCL
and others. This species is not listed for Belize by Russell (1964); the specimen
is the first (and only) for the country. Young (1973-1985) lists one record on the
1979 Belmopan Christmas Count.

1987

Wood and Leberm an— Birds of Belize

153

Vermilion Flycatcher, Pyrocephalus rubinus

A pair seen copulating near Milepost 35 of the Western Highway, 29 June 1985.
This is a rather late date for such activity, judging by Russell’s (1964) data, and
may represent a second nesting.

TYRANNXNE FLYCATCHERS I Y RANNINAE

Rufous Mourner, Rhytipterna holerythra

One seen or collected each day (CM-A3963; CM-A4000) at CM Columbia
Forest Camp, 25-27 March 1984; three collected (CM-A4216; CM-A4240; fe-
male, CM-S8954: ovary 11x6 mm) at CM Bladen Camp, 4-6 April 1984; one
collected of four seen (male, CM-S 10092: testes 10x4 mm) at CM Chiquibul
Camp, 21 June 1985. This species was listed by Russell (1964) from few localities.
Both the RAFOS and ABWS found a few individuals in the Columbia Forest
area. Young (1 973-1 985) recorded the species only twice on the Christmas Counts.

Great Crested Flycatcher, Myiarchus crinitus

Male collected (CM-P 162539: testes 3 x 1.5 mm) 1.5 km north of Augustine,
17 April 1984. Russell (1964) considered this species to be a rare migrant. The
present record is the latest spring date for the species in Belize.

Streaked Flycatcher, Myiodynastes maculatus

One collected of three seen (CM-A4452) at Chaa Creek, 20-21 April 1984; two
collected of ten seen (CM-A4927; male, CM-S 10073: testes 12x7 mm) at CM
Chiquibul Camp, 20 June 1985; male collected of two individuals seen (CM-
S 10232: testes 12 x 7, 10 x 6 mm) near Chan Laguna, 29 June 1985. The only
other published localities for this species are Gallon Jug and Ballerina Camp
(Russell, 1964). This species seems to be much more common at the western edge
of Belize, and in the Chiquibul it outnumbered the normally much more common
Sulphur-bellied Flycatcher {Myiodynastes luteiventris).

TITYRAS AND REGARDS TlTYRINAE

White-winged Becard, Pachyramphus polychopterus

Male collected by Dora Weyer and John Dunning (CM-P 144861) at Columbia
Forest Station, 26 April 1970; female collected by Robert Askins (CM-P 144862)
at Columbia Forest Station, 21 August 1971; one banded at Columbia Forest
Station, 10 December 1971 by E. J. Fisk et al.; male collected (CM-S8852: testes
2.5 x 1.5 mm) at Big Fall, Toledo District, 29 March 1984; one captured, pho-
tographed, and released at CM Bladen Camp, 6 April 1984. These constitute the
only records known to us of this species for the country.

Gray-collared Becard, Pachyramphus major

One mist-netted near Milepost 31 of the Western Highway, 28 February 1983
by the Manomet group. Russell (1964) listed only three records for this species:
two from the far south and one from Middlesex. One individual was recorded on
the 1977 Belmopan Christmas Count (Young, 1973-1985).

Rose-throated Becard, Pachyramphus aglaiae

One mist-netted along the Hummingbird Highway just south of Belmopan, 24
February 1 983 by the Manomet group; two collected (female, CM-P 1 66066: ovary

154

Annals of Carnegie Museum

vol. 56

7x4 mm, brood patch; male, CM-P 166071: testes 11x6 mm) near Chan
Laguna, 29, 30 June 1985; male collected (CM-P 166069: testes 10 x 4.5 mm) 6
km south and 6 km west of Progresso, 30 June 1985. Russell (1964) listed eight
specimens and considered the species uncommon. Barlow et al. (1969) reported
an additional two specimens from Rockstone Pond but the Royal Ontario Museum
has obtained several more specimens since then (Ross James, personal commu-
nication). Rose-throated Becards have been recorded on all but one Belize City
Christmas Count and on three Belmopan Christmas Counts (Young, 1973-1985).
The RAFOS also found this becard in the northeast (at Altun Ha) but, in addition,
recorded individuals at Guacamallo Bridge and in the Columbia Forest area. This
species appears to be considerably more common in the northeastern parts of the
country than elsewhere.

Black-crowned Tityra, Tityra inquisitor

Two collected (CM-A3874; CM-A3999) of four seen at CM Columbia Forest
Camp, 22-27 March 1984; at least two seen at CM Bladen Camp, 4-6 April 1984;
two seen at Chaa Creek, 16 June 1985; two collected (male, CM-S 10221: testes
7.5 mm; female, CM-S 10222: ovary 9x4 mm, largest ovum 2.5 mm) of at least
four seen near Chan Laguna, 28-30 June 1985. Russell (1964) listed only a few
localities where this species had been recorded. Like Russell, we found this species
in close association with the Masked Tityra T. semifasciata.

SWALLOWS HlRUNDINIDAE

Tree Swallow, Tachycineta bicolor

Ten seen in Belize City, 15 March 1984; two or three seen at Big Fall, Toledo
District and at The Dump (just west of the southern terminus of the Southern
Highway), 29, 30 March 1984. These are relatively late dates for this species
according to Russell’s (1964) data.

Bank Swallow, Rip aria rip aria

One seen at close range over the Sibun River at Hummingbird Hershey Plan-
tation, 20 March 1984. This is the farthest inland record; most sightings are from
the cays (Russell, 1964).

WRENS T ROGLODYTIDAE

White-bellied Wren, Uropsila leucogastra

One banded at Guanacaste Park, 9 February 1983 (Manomet group); one seen
at Big Fall, 1 April 1984; encountered several times throughout western Cayo
District; one captured and released at Milepost 29 of the Western Highway, 3
July 1985. Russell (1964) listed only a few localities for this species: Gallon Jug,
Hill Bank, and several from western Cayo District. The RAFOS mist-netted two
individuals, one in the Mountain Pine Ridge and one near Jimmy Cut. Young
(1973-1985) has recorded a few individuals on Christmas Counts (both locations).

Nightingale Wren, Microcerculus philomela

Male collected (CM-P144854) at Columbia Forest Station, 13 December 1971
(E. J. Fisk, Leberman et al.); male collected (CM-S8763: testes 4 x 2.5 mm) CM
Columbia Forest Camp, 24 March 1984. Not listed by Russell (1964). These are
the first specimens for the country. Nightingale Wrens appear to be restricted to
the southernmost parts of Belize.

1987

Wood and Leberm an— Birds of Belize

155

MUSCICAPIDS- MUSCICAPIDAE

GNATCATCHERS -SYLVIINAE

Blue-gray Gnatcatcher, Polioptila caerulea

One seen at CM Columbia Forest Camp, 22 March 1984; one seen at Big Fall,
Toledo District, 1 April 1984; very common in coastal pine savannas throughout
the year: five collected (two immature females, CM-P1 66057, 166060: almost no
skull pneumatization; two adult males, CM-P 166056, 166058: testes 4x3 mm
on each; adult female, CM-P1 66059: ovary 4x2 mm) near Milepost 24 of the
Western Highway, 26 June 1985; one collected (adult male, CM-P1 66076: testes

3.5 x 3 mm) of four seen just north of the Belize District— Orange Walk District
line along the new Northern Highway, 1 July 1985. Russell (1964) noted that this
species had been seen between 28 July and 5 April but he doubted that they bred
in Belize. However, he evidently did not visit the northern coastal pine savannas
during May and June where these birds nest quite commonly (Weyer, personal
communication). Most earlier Belizean specimens in collections have been iden-
tified as migrants of the northern subspecies P. c. caerulea (Russell, 1964).

SOLITAIRES, THRUSHES AND ALLIES- TURDINAE

Brown-backed Solitaire, Myadestes occidentalis

One seen near the CM Chiquibul Camp, 21 June 1985. Eye-ring and contrasting
back and nape color noted. This species was not listed by Russell (1964) but has
been observed on a very few occasions near the western border of the country
(Weyer, personal communication). Although usually considered to be restricted
to high elevations (Land, 1970; Peterson and Chalif, 1973), Monroe (1968) found
this species as low as 600 m (the same elevation as our sighting).

Veery, Catharus fuscescens

Three collected (CM-A4369; male, CM-P 162534: testes 4 x 2.5 mm; male
[destroyed by an opossum in camp]) near the Rio Frio Cave, 16-19 April 1984.
Russell (1964) listed only two records for the country, both in the fall. Young
(1973-1985) recorded one bird on the 1980 Belize City Christmas Count. We
think small numbers of Veeries migrate through Belize both in spring and fall but
are likely to be detected only through mist-netting.

Gray-cheeked Thrush, Catharus minimus

Three males collected (but subsequently destroyed by an opossum in camp)
near the Rio Frio Cave, 18-19 April 1984; male collected (CM-P 162543: testes

2.5 x 1.5 mm) at Chaa Creek, 22 April 1984. Russell (1964) lists only a few
records, all from near Belize City or the cays. Young (1973-1985) recorded this
species on the 1982 Belmopan Christmas Count. We think this species, like the
Veery, is more common than indicated by the available records but is difficult to
observe and thus usually overlooked.

Swainson’s Thrush, Catharus ustulatus

This species was collected or seen frequently from 23 March through 23 April
1984 in all parts of the country we visited. In addition to localities mentioned by
Russell (1964), we encountered these birds throughout the southern forests.

156

Annals of Carnegie Museum

vol, 56

VIREOS VlREONI DAE

Yellow-throated Vireo, Vireo flavifrons

One seen in San Ignacio, 7 March 1984. Female collected (CM-S8874: ovary
3x4 mm) near Big Fall, Toledo District, 30 March 1984. Russell (1964) noted
this species as uncommon, but Young (1973-1985) has recorded this species on
most Belize City Christmas Counts as well as on several Belmopan Christmas
Counts.

Philadelphia Vireo, Vireo philadelphicus

One male collected (CM-P1 62538: testes 1 mm) of two birds seen near Au-
gustine, 16 April 1984. Russell (1964) lists only one record from Belize, and
Young (1973-1985) recorded one on the 1984 Belmopan Christmas Count.

Green Shrike- Vireo, Vireolanius pulchellus

Female collected (CM-S10103: ovary 7x5 mm, largest ovum 2 mm) of at
least five heard or seen at our CM Chiquibul Camp, 21 June 1985. Russell (1964)
considered this species to be characteristic of the tall forests but not generally
common in Belize. Young (1973-1985) has recorded one or two individuals on
four of the last five Belmopan Christmas Counts. We found it only in the Chiquibul
but it appeared to be common there.

Rufous-browed Peppershrike, Cyclarhis gujanensis

One seen at the edge of high hardwood forest just northeast of Columbia Forest
Station, 14 December 1971 (Leberman); two collected (female in breeding con-
dition: CM-P 166077, ovary 10 x 5 mm; bird of unknown sex: CM-A5010) of at
least four birds seen between Milepost 24 and 29 of the Western Highway, 1-4
July 1985. In the Yucatan Peninsula, this species inhabits treetops in thin decid-
uous forest (Paynter, 1955; Parkes, personal communication), but Russell (1964)
noted that in Belize these birds are restricted to dense undergrowth in lowland
pine forests (and perhaps in mangrove swamps). While this species may be most
common in the lowland pine forests, our experience indicates that they are not
restricted to this habitat. Young (1973-1985) has recorded from one to five in-
dividuals on each of the last six Belize City Christmas Counts.

EMBERIZIDS — — EmBERIZIDAE
WOOD WARBLERS— PARULINAE

Blue-winged Warbler, Vermivora pinus

Individuals either seen or collected at each of the locations we visited during
March and April 1984 except Chaa Creek. A banded individual collected at the
CM Bladen Camp is reported elsewhere (Wood and Adams, 1985). We consider
the species to be more common than suggested by Russell (1964; 26 scattered
records). Young (1973-1985) listed one to seven individuals on every Christmas
Count except one.

Golden-winged Warbler, Vermivora chrysoptera

One seen near Rio Frio Cave, 14 April 1984. Russell (1964) listed only five
records for Belize and Young (1973-1985) recorded one individual each on the
1979 and 1982 Belmopan Christmas Counts.

1987

Wood and Leberman— Birds of Belize

157

Tennessee Warbler, Vermivora peregrina

One or two seen each day at CM Columbia Forest Camp 23-27 March 1984;
four collected (CM-A4068, 4092, 4106, 4137) of at least 12 seen at Big Fall,
Toledo District, 29 March-1 April 1984; other individuals seen elsewhere in the
country. Contrary to Russell (1964) we consider this species to be a common
transient in Belize, at least in the forested sections.

Cape May Warbler, Dendroica tigrina

One seen near Punta Gorda, 4 December 197 1 (Leberman and K. S. Anderson);
one seen along Hummingbird Highway near the Sibun River, 24 February 1983
(Manomet group, Leberman); one seen on Coco Plum Cay, 1 1 March 1984. Russell
(1964) lists only two records. Young (1973-1985) recorded two on the 1974 Belize
City Christmas Count. However, this species is probably not as rare as the limited
number of sightings suggests (Weyer, personal communication).

Black-throated Blue Warbler, Dendroica caerulescens

One seen at CM Columbia Forest Camp, 22 March 1984, the first published
record from the southern parts of Belize. Young (1973-1985) recorded individuals
of this species on three Christmas Counts.

Cerulean Warbler, Dendroica cerulea

Male collected (CM-P 162530; testes 5x3 mm) near Rio Frio Cave, 14 April
1984. Russell (1964) listed very few records for this rare transient and none from
Cayo District. Young (1973-1985) recorded the species on the 1978 and 1984
Belize City Christmas Counts.

Swainson’s Warbler, Limnothlypis swainsonii

One banded at Columbia Forest Station, 28 March 1972 (E. J. Fisk et al.); one
banded along Sibun River southeast of Milepost 35 on the Western Highway, 4
February 1983 (Manomet group); six banded at Guanacaste Park, one each on 7
and 8 February 1983, two each on 23 and 24 March 1984 (Manomet group); two
collected (CM-A3824, 3831) at Altun Ha, 18, 19 March 1984. Russell (1964)
listed only four records for this rare warbler. The only other published records
from Belize is of two seen on the first Belize City Christmas Count (Y oung, 1 974—
1985) and of one seen on Cay Chapel by the ABWS.

Ovenbird, Seiurus aurocapillus

We found this species to be one of the most abundant birds in the southern
forests during our March 1984 visit; up to ten individuals were observed each
day. Russell (1 964) listed it as a moderately common transient and winter resident.

Louisiana Waterthrush, Seiurus motacilla

Several seen at Columbia Forest Station in December 1971 and March 1972
(E. J. Fisk et al.); one collected (CM-A4166) of two seen at the CM Bladen Camp,
3, 5 April 1984; female collected (CM-P1 62541, ovary 6 x 2.5 mm) near Rio
Frio Cave, 16 April 1984. Russell (1964) listed very few records but we think this
species is more common than he suggested. Young (1973-1985) recorded the
species on six Christmas Counts and both the RAFOS and ABWS reported in-
dividuals from the Columbia Forest area.

158

Annals of Carnegie Museum

vol. 56

Gray-throated Chat, Granatellus sallaei

Female collected (CM-P 144869, ovary not enlarged) at Columbia Forest Sta-
tion, 24 March 1972 (E. J. Fisk et ah); male captured, photographed, and released
at Chaa Creek, 16 June 1985. The Columbia Forest Station specimen is the first
record from Toledo District. Only two locations (Gallon Jug and Cayo) are listed
for this rare resident by Russell (1964).

B AN AN AQUITS — COEREBINAE

Bananaquit, Coereba flaveola

Three seen at close range in Guanacaste Park, 6 March 1984; one seen at
Dangriga, 21 March 1984; two collected (CM-S8824, female, ovary 5x4 mm;
CM-S8825, male, testes 7x4 mm) at Big Fall, Toledo District, 28 March 1984;
one collected (CM-P162510, male, testes 7 x 4.5 mm) of six seen at Big Fall,
Toledo District, 31 March 1984; two seen at CM Bladen Camp, 4 April 1984.
This species was not encountered by the Louisiana State University personnel,
and Russell (1964) knew of only two old records. The fact that Russell spent
relatively little time in the south of Belize may account for the lack of sightings,
but Bananaquit populations may also be increasing; we consider the species com-
mon in the southern forests. Both the RAFOS and ABWS encountered this species
in the Columbia Forest area. Young (1973-1985) reported up to four individuals
on every Belize City Christmas Count except in 1977 and also listed it on four
Belmopan counts.

TANAGERS T HRAUPINAE

Common Bush-Tanager, Chlorospingus ophthalmicus

One seen and carefully identified by Leberman and K. S. Anderson indepen-
dently at Columbia Forest Station, 9 December 1971, is only the second record
for Belize (Russell, 1964). Although this is thought to be a montane bird, some
individuals apparently venture to lower elevations along the foothills of the Maya
Mountains.

CARDINALS AND ALLIES CARDINALINAE

Blue Grosbeak, Guiraca caerulea

We agree with Russell’s (1964) comments for this species, except we found
these birds common into the early part of April (rather than late March) and
recorded four at Augustine as late as 16 April 1984 and one the following day.

EMBERIZINES— EMBERIZINAE

Olive Sparrow, Arremonops rufivirgatus

Five collected of many seen in open pine forest between Milepost 24 and
Milepost 30 of the Western Highway, 13, 14 June 1985; at least six seen in open
pine forest along the new Northern Highway at the Belize-Orange Walk District
boundary. Russell (1964) apparently spent very little time in the pine forests
between Belize City and Belmopan, otherwise he would have had to encounter
this species which is common there. Russell’s few records are from the north-
eastern portion of Belize.

Grassland Yellow-Finch, Sicalis luteola

One collected (CM-P144865) at Mussel Creek Rice Station, 10 October 1971,
the first record for Belize. Young (1973-1985) recorded a flock of 100 on the 1981

1987

Wood and Leberman — Birds of Belize

159

Belize City Christmas Count and a flock of 16 on the 1984 Belize City Christmas
Count.

Acknowledgments

We are indebted to the staff of the Ministry of Natural Resources of the Government of Belize for
their assistance and cooperation with our fieldwork in Belize. We also appreciate the support and
assistance of the Belize Audubon Society; their conservation efforts have greatly enhanced the prospects
for bird study in Belize. We are likewise grateful for help provided us by International Expeditions,
Inc., Birmingham, Alabama, and in particular to Mr. Ray E. Ashton, Jr. Dora C. Weyer and Kathleen
S. Anderson generously provided information from their field records. Kenneth C. Parkes made many
helpful suggestions. We thank Erma J. Fisk and the Bradley Fisk Fund for making Leberman’s 1971
and 1972 trips to Belize possible, and Manomet Bird Observatory for funding his 1983 visit. Our
1984 and 1985 expeditions were supported by the O’Neil Museum Trust, The Carnegie Museum of
Natural History. Finally, we wish to thank the other members of our expedition teams, especially C.
J. McCoy, Stephen P. Rogers, Ellen J. Censky, Paul and Mara Freed, and Timothy McCarthy; our
accomplishments would have been far fewer without them.

Literature Cited

American Ornithologists’ Union. 1983. Check-list of North American birds. Sixth edition. Amer-
ican Ornithologists’ Union.

- — — . 1985. Thirty-fifth supplement to the American Ornithologists’ Union Check-list of North
American birds. Auk, 102:680-686.

Barlow, J. C., J. A. Dick, D. H. Baldwin, and R. A. Davis. 1969. New records of birds from
British Honduras. Ibis, 1 1 1:399-402.

Barlow, J. C., J. A. Dick, and E. Pendergast. 1970. Additional records of birds from British
Honduras (Belize). Condor, 72:371-372.

Barlow, J. C., J. A. Dick, D. Weyer, and W. F. Young. 1972. New records of birds from British
Honduras (Belize) including a skua. Condor, 74:486-487.

Erickson, R. 1 977. First record of the Knot, Calidris canutus, and other records, from Belize (British
Honduras). Bulletin of the British Ornithologists’ Club, 97:78-81.

Griscom, L. 1926. The ornithological results of the Mason-Spinden Expedition to Yucatan. Part
I. —Introduction; birds of the mainland of eastern Yucatan. American Museum Novitates, no.
235.

Hallchurch, T. T. 1982. Exercise King Vulture: Belize: 23 February-16 March 1982. Adjutant
[Journal of the Army Bird Watching Society], 12:3-36.

Haney, J. C. 1983. First sight record of Orange-breasted Falcon for Belize. Wilson Bulletin, 95:31 4—

315.

Jenkins, P. G., ed. [1983]. Report on the Royal Air Force Ornithological Society Expedition to
Belize 1981. Royal Air Force Ornithological Society. 92 pp.

Kiff, L. F., and J. L. Kiff. 1974. First specimen of Reddish Egret from British Honduras. Auk, 9 1 :

170-171.

Kricher, J., N. Brokaw, and T. Lloyd-Evans. 1984. Belize: a quick profile; vegetation profiles;

banding and census results. Manomet Bird Observatory, August 1984:2-1 1.

Land, H. C. 1970. Birds of Guatemala. Livingston Publ. Co., Wynnewood, Pennsylvania.

Miller, B. W., and R. L. Tilson. 1985. Snail Kite kleptoparasitism of Limpkins. Auk, 102:170-
171.

Monroe, B. L., Jr. 1968. A distributional survey of the birds of Honduras. Ornithological Mono-
graphs, 7.

National Geographic Society. 1981. Atlas of the world. National Geographic Society, Washington,
D.C.

Orians, G. H. 1983. Notes on the behavior of the Melodious Blackbird (Dives dives). Condor, 85:
453-460.

Paynter, R. A., Jr. 1955. The omithogeography of the Yucatan Peninsula. Bulletin of the Peabody
Museum of Natural History, Yale University, 9.

Peterson, R. T., and E. L. Chalif. 1973. A field guide to Mexican birds. Houghton Mifflin Co.,
Boston.

Russell, S. M. 1964. A distributional study of the birds of British Honduras. Ornithological Mono-
graphs, 1.

Sherrard-Smith, D. W. 1982. Reports from Belize. Big Falls Ranch. Bulletin of the Army Bird
Watching Society, 1982(2):F2-F3.

160

Annals of Carnegie Museum

vol. 56

Stacey, P. B. 1981. Foraging behavior of the Acom Woodpecker in Belize, Central America. Condor,
83:336-339.

Weyer, D., and W. F. Young. 1983. A checklist of the birds of Belize. International Expeditions,
Inc., Birmingham, Alabama.

Wood, D. S., and R. J. Adams, Jr. 1985. First Central American recovery of Blue-winged Warbler.
Journal of Field Ornithology, 56:424-425.

Wood, D. S., R. C. Leberman, and D. Weyer. 1986. Checklist of the birds of Belize. Carnegie
Museum of Natural History Special Publication, 12.

Young, W. F. 1973-1985. Christmas Bird Count summaries for Belize City and Belmopan areas,
1972-1984. American Birds, 27-39.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 7, Pp. 161-171

15 May 1987

A REVIEW OF THE CRANE FLIES IN THE SUBGENUS
TIPULA {PAPUATIPULA) (DIPTERA: TIPULIDAE),

WITH DESCRIPTIONS OF FIVE NEW SPECIES

Chen W. Young

Collection Manager, Section of Invertebrate Zoology

Abstract

The subgenus Tipula (. Papuatipula ) is reviewed. Twenty-three species are recognized with five species
from Papua New Guinea being newly described: Tipula {Papuatipula) insperata, Tipula (. Papuatipula )
koiari, Tipula {Papuatipula) nigritus, Tipula {Papuatipula) oneili, and Tipula {Papuatipula) wibieae.
A key is provided to identify all species for which types have been examined. Female brachyptery
and gynandromorphy in Tipula {Papuatipula) koiari, n. sp., are discussed.

Introduction

The subgenus Papuatipula (Alexander, 1935) of the large and extensive genus
Tipula (Linnaeus, 1758) contains a small group of species from New Guinea. This
review was stimulated by the discovery of five new taxa collected during a recent
field trip to Papua New Guinea. It is an effort to clarify the systematic limits of
this subgenus, and to investigate its relationship to other subgenera within Tipula.

Species of Papuatipula are poorly represented in collections, and specimens
other than type materials are lacking for comparative study. Many of the older
types are in poor condition, having been improperly mounted or subsequently
mishandled. Species of subgenus Papuatipula are the most frequently encountered
and the most diverse species of Tipula in New Guinea. Many more species will
undoubtedly be found when the fauna is better collected, therefore, a revision of
the subgenus at this time would be premature. The following review is primarily
a redefinition of the subgenus, a key and descriptions of five new species, and
should provide a foundation for future systematic and biological studies.

Taxonomic History

The subgenus Papuatipula was first proposed by C. P. Alexander (1935:52),
who described it as follows:

. . Antennae 13-segmented; flagellar segments with verticils that greatly

exceed the segment in length. Tibial spurs long and conspicuous; formula

1-2-2. Wings with Rs unusually short but not transverse, . . . squama naked

99

The subgenus was described in Tipula and the five originally included species
were Tipula (. Papuatipula ) divergens de Meijere, T. (P.) leucosticta Alexander, T.
(P.) meijereana Alexander, T. (P.) novaebrittaniae Alexander, and T. (P.) omis-
sinervis (de Meijere). T. (P.) novaebrittaniae was designated as the type species.
Alexander subsequently described an additional thirteen species (Oosterbroek and
Jonas, 1986): T. (P.) artifex Alexander, T. (P.) consiliosa Alexander, T. (P.) gres-

Submitted 11 November 1986.

161

162

Annals of Carnegie Museum

vol. 56

sittiana Alexander, T. (P.) lieftincki Alexander, T. (P.) melanotis Alexander, T.
(P.) nokicola Alexander, T. (P.) obediens Alexander, T. (P.) pensilis Alexander,
T. (P.) satirica Alexander, T. (P.) staryi Alexander, T. (I\) strictistyla Alexander,
T. (P.) surcularia Alexander, and T. (P.) toxopeina Alexander, He transferred two
other species, T. (P.) cyclopica Alexander and T. (P. ) pedicioides Alexander from
the subgenus Acutipula (Alexander) to Papuatipula (Alexander, 1973). Twelve
holotypes and four paratypes of the twenty described species have been examined
and compared to the study materials consisting of 7 1 specimens in The Carnegie
Museum of Natural History (CMNH) and 5 specimens from the Bernice P. Bishop
Museum.

When first established, the subgenus was characterized by venation, naked
squamae, and the male hypopygia. Some of these characters are no longer diag-
nostic for Papuatipula because of variation among the species that were subse-
quently transferred to the subgenus. Moreover, when the holotype of T. novae -
brittaniae was examined, the squamae were not naked but bore a duster of eight
setae, despite Alexander’s original description. The present study has demon-
strated that venational differences, such as the length of vein Rs compares to m-
cu, vary among the species and are of little value in delimiting the group. The
structures of the male hypopygium, on the other hand were found to be the most
informative taxonomic features and have been used to characterize this subgenus.
With the addition of the five new species described below and the removal of 7 .
nokicola and T. pedicioides , 23 species of Papuatipula are recognized at this time.

Systematics

Family Tipulidae
Subgenus Papuatipula

Description. — The terminology used in this section, in the following key to species, and in the species
descriptions, follows Byers (1961) and McAlpine (1981).

Head: Rostrum well developed, 1.5 to twice length of rest of head, nasus distinct. Frons moderately
broad with medial tubercle. Antenna 13 -segmented, subequal to length of head; two short and three
long verticillate setae arise from base of each flagellomere. Terminal segment of maxillary palp fla-
gelliform, longer than four preceding segments together.

Thorax: Wing clear with dark brown stigma and paler brown quadrate area before middle of cell
CuA, or striped with longitudinal brown and white pattern with a white band before cord and dark
pigment narrowly present along veins CuA, m-cu, and r-m. Squama setose. Tibial spur formula 1-2-
2. Male claws with two teeth; female claws simple or with single tooth.

Abdomen: Abdomen slender, in male slightly shorter than wings, in female usually longer.

Hypopygium: Tergite 9 separates from stemite 9; basistyle fused with stemite; stemite 8 unarmed.
Tergite 9 notched medially bearing small spines. Outer dististyle lobed, cylindrical, pigmented apically.
Inner dististyle variable.

Ovipositor: Tergite 9 short, not produced laterally. Cerci long and extreme slender, smooth-margined,
fused to tergite 10; hypovalves shorter than cerci and fused basally. Valvulae fused at base, slightly
sclerotized posteriorly.

Diagnosis. — The subgenus can be separated morphologically from other sub-
genera of Tipula by the following characters: squamae setose, tibial spur formula
1-2-2, male hypopygium with ninth tergite and stemite separated and ninth tergite
notched medially, and the outer dististyle sclerotized. The last two characters
appear to be apomorphic thus species of Papuatipula may form a monophyletic
group.

Three other subgenera of Tipula are sympatric with Papuatipula in New Guinea
(Alexander, 1961). They are Acutipula (Alexander), Indotipula (Edwards), and
Tipulodina (Enderlein). Species of Tipulodina have long legs that are ringed with

1987

Young —Crane Flies in the Subgenus Tipula (Papuatipula)

163

snowy white on the femora, tibiae, or basitarsi. Species of Indotipula have naked
squamae and clear, iridescent wings. The subgenus Acutipula was characterized
primarily by having the ninth tergite and stemite fused laterally and by having
unbranched weakly sclerotized outer dististyle. Species of Papuatipula bear a
strong resemblance to those of Acutipula, but can be distinguished easily from
the other subgenera.

Tipula nokicola (Alexander, 1953) and T. pedicioides (Alexander, 1948) are
excluded from Papuatipula due to several morphological characters. In T. noki-
cola, the presence of a strongly sclerotized outer lobe at the base of the inner
dististyle in the male and iridescent transparent wings suggest a relationship to
sympatric Indotipula species. However, the presence of setae on the squamae is
similar to the condition in Papuatipula. T. pedicioides was described in 1948 in
the subgenus Acutipula and was transferred to Papuatipula by Alexander (1973).
The relatively large, unbranched outer dististyle suggests an affiliation with Acu-
tipula. More specimens and study are needed to assess the systematic position of
both T. pedicioides and T. nokicola.

Included species. — Characters found useful in segregating species of Papuatipula
are as follows: (1) wing pattern; (2) shape of flagellomeres; (3) shape of the ninth
tergal lobes; (4) degree of sclerotization of the outer dististyle; (5) shape of the
inner dististyle; (6) length of the cerci relative to the hypovalves. The following
key to species is based primarily on the most easily visible genitalic characters of
males and includes only those species for which type specimens have been ex-
amined, therefore, omitting T. diver gens, T. leucosticta, T. lieftincki, T. meije-
reana, T. omissinervis, and T. surcularia.

l.

1.

2.
2.

3.

3.

4.

4.

5.

5.

6.
6.
7.

7.

8.

8.

9.

9.

10.

10.

11.

11.

12.

Key to Species of Papuatipula

Wings longitudinally striped with brown and white; dark pigment narrowly present along

vein CuA 2

Wings unmarked except for the stigmal darkening and a quadrate dark mark at middle of

cell CuA 15

Wings with dark pattern strongly contrasting, whitish crossband at cord extensive; white

pattern covering about half of cells M„ 2nd M2 and M3 3

Wings with dark pattern diffuse, not strongly contrasting, no whitish crossband at cord; white

pattern less extensive, only covering bases of cells M, and 2nd M2 7

Size large, wing length over 30 mm T. wibleae, new species

Size smaller, wing length under 25 mm 4

Cell CuA, entirely dark T. koirai, new species

Cell CuA, entirely clear or with more than apical half clear 5

Proximal half of outer dististyle dilated T. toxopeina

Outer dististyle simple, not dilated 6

Ninth tergite of male shallowly emarginated T. gressittiana

Ninth tergite of male broadly emarginated forming V-shaped notch T. staryi

Ninth tergite of male with two medial lobes and two lateral lobes 8

Ninth tergite of male without lateral lobes 10

Medial lobes on ninth tergite of male separated by a deep U-shaped notch

T. nigritus, new species

Medial lobes nearly approximate at base .... 9

Outer dististyle strongly sclerotized with two large teeth on mesal margin

T. insperata, new species

Outer dististyle with one tooth on mesal margin T. artifex

Medial lobes on ninth tergite of male separated by very narrow notch 11

Medial lobes on ninth tergite of male separated by broad U-shaped notch 14

Outer dististyle deep bifid T. satirica

Outer dististyle simple 12

Outer dististyle with two large apical teeth T. pensilis

164

Annals of Carnegie Museum

vol. 56

1 2. Outer dististyle with single dark point ............................................... 13

13. Inner dististyle with beak broadly compressed ................................ T. strictistyla

13. Inner dististyle with beak long and slender ................................... T. consiliosa

14. Outer dististyle long, straight, rod-shaped, with sharp apical spine .............. T. meianotis

14. Outer dististyle broad, compressed ........................................... T. cyclopica

15. Outer dististyle rod-shaped, apex bifurcate with dark serrate spur extending inward ......

.................................................................. T. oneiii, new species

15. Outer dististyle broad, compressed .................................................. 16

16. Outer dististyle bifurcate, one edge with a sharp spine, the other with a group of spinuloid
setae ...................................................................... T. obediens

16. Outer dististyle simple, with one apical spine ............................ T. novaebrittaniae

Tipula (Papuatipula) insperata Young, new species
(Fig. 1-3)

Description. ~ Body length: Male, 16 mm; female, 22 mm. Wing length: Male, 21 mm; female,
23 mm.

Head: Occiput and rostrum brown; palpi with first three segments yellow, terminal segment brown.
Frontal tubercle with median longitudinal groove. Antenna with scape, pedicel and first flagellomere
yellow; flagellomeres distinctly bicolored, dark brown basally, remainder yellow; all flagellomeres
except last subequal in length; five setae in verticil arising from base of each flagellomere, longest
about three times length of flagellomere.

Thorax: Pronotum reddish brown. Scutum brown with four light brown stripes, median stripes
divided anteriorly by dark brown interspace. Scutum, scutellum and mediotergite brown, pollinose.
Pleura yellowish brown, pollinose, with one narrow dark brown longitudinal stripe across cervical
sclerite, proepistemum, upper anepistemum, and base of wing to katatergite. Legs with coxae and
trochanters yellow; femora yellow with conspicuous subapical brown band; tibiae and basitarsi yellow
with apices dark; remainder of tarsi brown. Wings tinged with brown; dark areas at stigma, distal side
of anterior cord, postarcular region; a conspicuous dark brown quadrate spot before middle of cell
CuA, preceded and followed by whitish subhyaline areas; dark pigment narrowly bordering veins CuA
and base of CuA,. Halteres brown.

Abdomen: Ground color yellowish brown. First segment gray, tergites 2 to 4 brown with gray lateral
borders, basal rings glabrous, subbasal impressions conspicuous; tergites 5 to 7 dark brown. Stemite
brownish yellow.

Hypopygium: External structures as in Fig. 1, 2. Tergite 8 much shorter than tergite 9. Tergite 9
terminating in two very dark lobes medially, separated by a small rectangular notch; margins of lobes
with spines; outer lateral angle of tergite 9 produced into lower glabrous lobe. Outer dististyle dark
brown, with dark apical point posteriorly and curved inwards bearing two large teeth on mesal margin
before apex. Inner dististyle with beak slender; lower beak large, dark; outer basal lobe smaller, rounded.

Ovipositor: External structures as in Fig. 3. Cerci slightly longer than tergite 1 0. Hypovalves extending
to about one-third length of cerci.

Type material — Holotype: Male, Papua New Guinea, Northern Province, My-
ola (147°44?E, 9°9'S), lora Creek, 4 June 1984, Chen W. Young. Paratype: One
female, topotypic. Types are in the collection of the CMNH.

Etymology.— The name insperata (Latin, unexpected) was selected because of
the unexpected capture of a mating pair along the bank of lora Creek.

Tipula (Papuatipula) koiari Young, new species
(Fig. 4-7)

Description.— Body length: Males, 17-18 mm; females, 19-20 mm. Wing length: Males, 20-21 mm;
females, 9-10 mm.

Head: Occiput and rostrum pale brown, pollinose; palpi dark brown. Frontal tubercle with median
longitudinal groove. Antenna with scape and pedicel yellow; flagellomeres faintly bicolored, brownish
black basally, remainder brown; all flagellomeres except last subequal; five verticillate setae arising
from base of each flagellomere, longest subequal in length to flagellomere.

Thorax: Pronotum grayish brown. Scutum with two dark brown longitudinal median stripes and
two short indistinct lateral stripes. Pleura pale brown except laterotergites darker. Legs with coxae
and trochanters yellowish brown slightly tinged with gray; remainder of legs brown, gradually darker

1987

Young —Crane Flies in the Subgenus Tipula ( Papuatipula )

165

Fig. 1, 2.— Tipula (. Papuatipula ) insperata male hypopygium. 1. Dorsal view. 2. Lateral view. 9t—
ninth tergite, id —inner dististyle, 1 bk— lower beak of inner dististyle, o b lb— outer basal lobe of
inner dististyle, od— outer dististyle. Fig. 3.—T. (P.) insperata female ovipositor lateral view. 9t—
ninth tergite, 10s— tenth stemite, lOt— tenth tergite, cr— cercus, hv— hypovalve, vv— valvulae.

distally; apices of femora and tibiae dark. Wings tinged with grayish brown, darker along costal margin,
on veins CuA, base of CuA,; dark brown clouding at stigma and middle of cell CuA. Clear areas in
cell 1st M2, base of cells M, and M2, and apical one-third of cell R5. Female brachypterous, wings
short and narrow with all veins visible. Knobs of halteres brownish black, stems yellowish brown.

Abdomen: Ground color brown, pollinose; first segment yellowish brown; tergites 2 to 7 brown with
brownish black lateral margins forming a stripe interrupted by distinct subbasal impressions. Stemites
1 and 2 yellowish brown, other stemites gradually darker posteriorly. Female abdomen unusually
large.

Hypopygium: External structures as Fig. 4, 5. Tergite 8 short, broadly emarginate; tergite 9 rounded
at sides, with two median lobes distally, separated by shallow notch; apex of lobes with dense, black
spines. Outer dististyle large, pale brown, curved inward with one large pre-apical tooth about mid-
length; distal edge of outer dististyle heavily pigmented, serrate, with rounded apical tip. Inner dististyle
with beak broad, directed anteriorly; lower beak lobed, dark; outer basal lobe small with dark outer
margin.

Ovipositor: External structures as Fig. 6. Cerci slightly longer than tergite 10, narrowed at base,
tapered posteriorly; stemite 1 0 pale, bearing setae at posterior margin. Hypovalves extending to about
half length of cerci.

Type material — Holotype: Male, Papua New Guinea, Northern Province, My-
ola, 2 June 1984, James A. Bossert, John W. Ismay, Chen W. Young. Paratypes:
Thirty-two males, thirteen females, topotypic. The holotype and several paratypes

166

Annals of Carnegie Museum

vol, 56

Fig. 4, 5. — Tipula (. Papuatipula ) koiari male hypopygium. 4. Dorsal view. 5. Lateral view. 9t— ninth
tergite, id— inner dististyle, 1 bk— lower beak of inner dististyle, o b lb— outer basal lobe of inner
dististyle, od— outer dististyle. Fig. 6. — T. ( P .) koiari female ovipositor lateral view. 9t— ninth tergite,
10s— tenth stemite, lOt— tenth tergite, cr— cercus, hv— hypovalve, w— valvulae.

are deposited in CMNH. One male paratype is deposited in each of the following
institutions: Central Reference Insect Collection, Konedobu, Papua New Guinea;
Bernice P. Bishop Museum, Honolulu, Hawaii; National Biological Institute,
Bogor, Indonesia; Institute of Taxonomic Zoology, University of Amsterdam,
Netherlands; Snow Entomological Museum, University of Kansas, Lawrence,
Kansas; United States National Museum of Natural History, Washington, D.C.;
British Museum (Natural History), London, England.

Etymology. — This species is named after the Koiari tribesmen on whose land
the type series was collected.

Remarks. — The type series of this species was collected in a montane meadow.
The habitat was a saturated marsh dominated by Juncus. This homogeneous
habitat at high altitude (2200 m), coupled with low temperatures and frequent
windy conditions, probably explains the occurrence of female brachyptery (Byers,
1969). Mating pairs were observed at night. Copulating females grasped blades
and culms of grass with their front legs, while copulating males hung freely upside-
down by their genitalia.

A gynandromorphic specimen (Fig. 7) was collected on 4 June 1984. This
specimen has a well developed left wing, which is only slightly shorter than the

1987

Young— Crane Flies in the Subgenus Tipula (Papuatipula)

167

Fig. 7. — Tipula (. Papuatipula ) koiari. A gynandromorphic specimen showing brachypterous female
form with a well developed left wing and a vaguely defined male hypopygium.

wing of a male. The right wing has the brachypterous female form. The abdomen
terminates in an ovipositor. One cercus is present on the right side, and the left
side is a vaguely defined male hypopygium. Both hypo valves are present on the
ventral side. This specimen is in the collection of the CMNH.

Tipula {Papuatipula) nigritus Young, new species
(Fig. 8-10)

Description.— Body length: Males, 16-17 mm; female, 23 mm. Wing length: Males, 18-19 mm;
female, 21 mm.

Head: Occiput and rostrum dark brown; palpi dark brown. Frontal tubercle with median longitudinal
groove. Antenna with scape and pedicel dark brown, flagellomeres beyond first vaguely bicolored,
dark brown basally, remainder brown; flagellomeres subequal in length becoming shorter distally; five
verticillate setae arising from base of each flagellomere, longest subequal in length to flagellomere.

Thorax: Pronotum dark brown. Scutum reddish brown with four longitudinal stripes bordered by
dark brown, interspaces grayish brown; median stripes confluent anteriorly, and with pale median line
posteriorly. Scutum, scutellum and mediotergite brown, pollinose. Pleura dark brown, pollinose. Legs
dark brown, apices of femora and tibiae slightly darker. Wings patterned with pale brown and white;

168

Annals of Carnegie Museum

vol. 56

Fig. 8, 9 . — Tipula ( Papuatipula ) nigritus male hypopygium. 8. Dorsal view. 9. Lateral view. Fig. 10.—
T. (P.) nigritus female ovipositor lateral view.

dark areas include stigma, prearcular region, base of cell R4+5, base of cell CuA,, and middle of cell
CuA. Halteres dark brown.

Abdomen: Ground color brown with terga darker than sterna. Tergites 2 to 4 with pale basal rings

and subbasal impressions.

Hypopygium: External structures as Fig. 8, 9. Tergite 8 short; tergite 9 terminating in two broadly
rounded dark mesal lobes and two outer, lower lobes, mesal lobes widely separated by U-shaped
notch, margins of mesal lobes with spines. Outer dististyle narrow at base, broad at middle; apex
obliquely truncate with dark edge. Inner dististyle with beak broad, compressed; lower beak a small
dark lobe; outer basal lobe bifurcate with dark margin.

Ovipositor: External structures as Fig. 10. Cerci slightly longer than tergite 10, with round apical
expansion. Hypovalves extending to about one-third length of cerci, with basal lobe broad laterally.

Type material — Holotype: Male, Papua New Guinea, Morobe Province, Wau
Ecological Institute, 28 December 1965, J. & M. Sedlacek. Paratypes: Three males
and one female, topotypic. Holotype, 1 male paratype, and 1 female paratype are
in the collection of the Bernice P. Bishop Museum, Honolulu, Hawaii. Two male
paratypes are deposited in the CMNH.

Etymology. — The name nigritus emphasizes the relatively dark pigmentation
of this species.

Tipula ( Papuatipula ) oneili Young, new species
(Fig. 11-13)

Description. —Body length: Males, 18-20 mm; females, 28-30 mm. Wing length: Males, 23-25 mm;
females, 24-25 mm.

Head: Occiput and rostrum brown; palpi with first segment yellowish brown, remainder dark. Frontal

1987

Young— Crane Flies in the Subgenus Tipula (Papuatipula)

169

Fig. 11, 12.— Tipula ( Papuatipula ) oneili male hypopygium. 11. Dorsal view. 12. Lateral view. Fig.
13.— T. (P.) oneili female ovipositor lateral view.

tubercle with median longitudinal groove. Antenna with scape and pedicel yellow, flagellomeres 2 to
8 bicolored, brown basally with remainder yellowish brown; all except apical flagellomeres subequal
in length; five verticillate setae arising from base of each flagellomere, longest about twice length of
flagellomere.

Thorax: Pronotum gray. Scutum brown with four slightly darker brownish gray stripes bordered
narrowly with brown. Scutum, scutellum, mediotergite, and pleura evenly gray, pollinose. Legs with
coxae and trochanters gray; femora and tibiae yellow, apices dark brown; tarsi light brown, gradually
darker toward apices. Wings grayish subhyaline, with darker area at stigma and faint pattern of brown
before middle of cell CuA toward base. Halteres grayish brown.

Abdomen: First segment gray; tergites 2 to 4 yellowish brown; tergites 5 to 8 dark brown. Stemites
dark brown.

Hypopygium: External structures as in Fig. 11, 12. Tergite 8 short; tergite 9 with two low, slightly
obtuse lobes separated by a narrow notch posteriorly, lobes emarginate medially, edged with spines.
Outer dististyle long, broad at base, narrowed near mid-length, expanded apically; apex bifurcate with
a dark flange extending mesad. Inner dististyle with beak slender; lower beak a dark lobe; outer basal
lobe smaller than lower beak with dark margin.

Ovipositor: External structures as Fig. 13. Cerci long, about twice length of tergite 10. Hypo valves
broad basally, tapered posteriorly, extending to about half length of cerci.

Type material— Holotype: Male, Papua New Guinea, Morobe Province, Wau
Ecological Institute, 13 May 1984, Chen W. Young. Paratypes: Two males and

170

Annals of Carnegie Museum

vol. 56

Fig. 14, 15. Tipula ( Papuatipula ) wibleae male hypopygium. 14. Dorsal view. 15. Lateral view.

two females topotypic; two males, two females, Northern Province, near Tufi, 10
September 1982, John W. Ismay. Holotype and four paratypes are in the collection
of the CMNH. Other paratypes are deposited in Central Reference Insect Col-
lection, Konedobu, Papua New Guinea, and the Bernice P. Bishop Museum,
Honolulu, Hawaii.

Etymology. — This species is named in honor of Mr. Edward O’Neil for his
generosity in funding the field activities of the Section of Invertebrate Zoology,
CMNH.

Remark. — Specimens of this species were collected from low herbaceous plants
growing along the forest edge.

Tipula (. Papuatipula ) wibleae Young, new species
(Fig. 14, 15)

Description.— Body length: Males 25-27 mm; female, unknown. Wing length: Males, 32-35 mm.

Head: Occiput and rostrum yellowish brown; palpi dark brown. Frontal tubercle distinct, with
median longitudinal groove. Antenna with scape and pedicel yellow; flagellomeres bicolored, brownish
black basally, the remainder yellow; all flagellomeres subequal in length; five verticillate setae arising
from base of each flagellomere, longest about twice length of flagellomere.

Thorax: Pronotum yellow. Scutum yellowish brown with two reddish brown longitudinal stripes
bordered by yellow; stripes contiguous anteriorly, narrowly bordered with yellow. Scutum yellowish
brown; scutellum brown. Mediotergite quadrate, gray pollinose anteriorly, yellow posteriorly. Pleura
evenly yellow, pollinose. Legs with coxae and trochanters yellow, pollinose; remainder of legs brown
gradually darker towards apex; apices of femora and tibiae dark. Wings tinged with brown, distinctly
patterned with dark brown and white; band before cord conspicuous, white, extending to base of cell
M,; dark pigment narrowly present along veins CuA2, M1+2, m, r-m, basal half of CuA,, M„ M2, M3,
basal half and distal fourth of CuA; pattern in stigmal region of wing restricted, dark brown; dark
quadrate area before middle of cell CuA toward base. Knobs of halteres yellow, stems light brown.

Abdomen: Segment 1 brownish yellow; tergites 2 to 5 with basal one-third brownish yellow and
remainder brown with paler lateral border; basal rings pale; subbasal impressions inconspicuous;
tergite 6 brown; tergites 7 and 8 pale brown. Stemite yellowish brown.

1987

Young —Crane Flies in the Subgenus Tipula (Papuatipula)

171

Hypopygium: External structures as in Fig. 14, 1 5. Tergite 8 shorter than tergite 9, broadly emarginate
posteriorly; tergite 9 with posterior border produced into two rounded lobes distally separated by a
deep U-shaped notch; mesal margins of lobes with abundant short dense spines. Outer dististyle dark
brown, nearly parallel-sided in lateral aspect but slightly wider at base, with apex pigmented and
bifurcate; outer margin serrate with a dark ridge extending along inner ventral surface toward base.
Inner dististyle with beak long and slender; lower beak a dark lobe; outer basal lobe of inner dististyle
short, terminating in a dark, rounded lobe.

Type material. — Holotype: Male, Papua New Guinea, Northern Province, My-
ola, Iora Creek, 3 June 1984, Chen W. Young. Paratypes: Two males, topotypic.
The holotype and paratype are in the collection of the CMNH. The other paratype
is deposited in the Central Reference Insect Collection, Konedobu, Papua New
Guinea.

Etymology. — This species is named in honor of Mrs. Mary Wible, a long ded-
icated friend of The Carnegie Museum of Natural History. She has collected
specimens for the Section of Invertebrate Zoology over many years, and has
provided generous financial support for a wide variety of research projects, in-
cluding this study.

Remarks. — All specimens were collected during the day, resting on the under-
sides of tree trunks projecting over Iora Creek. The tree trunks were heavily
covered with living moss.

Acknowledgments

I would like to thank John Ismay for his help during my collecting trip in Papua New Guinea,
Adam Bossert for helping in collecting specimens, Harry Sakulas for making available the facilities
at Wau Ecology Institute, and John Rawlins for comments and suggestions on the manuscript. Ac-
knowledgments are due Wayne Mathis (United States National Museum of Natural History), Bruce
Townsend (British Museum (Natural History)), Neil Evenhuis (Bernice P. Bishop Museum), Chris
Watts (South Australia Museum), and Margaret Debenham (University of Sydney) for their generous
loan of type materials. This research was conducted with permission from the Division of Research,
Department of Morobe, Papua New Guinea, and was supported by the O’Neil Field Research Fund
and the Wible Research Fund, Section of Invertebrate Zoology, The Carnegie Museum of Natural
History.

Literature Cited

Alexander, C. P. 1935. The Diptera of the Territory of New Guinea. II. Family Tipulidae. Pro-
ceedings of the Linnean Society of New South Wales, 60:51-70.

— . 1948. New or little-known Tipulidae (Diptera). LXXXI. Oriental-Australasian species. An-

nals and Magazine of Natural History, 1 2(1):39 1—4 17.

-. 1953. New or little-known Tipulidae (Diptera). XCV. Oriental-Australasian species. Annals

and Magazine of Natural History, 1 2(1):739— 757.

- — — . 1961. New or little-known crane-flies from New Guinea (Diptera: Tipulidae). Part 1 . Pacific
Insects, 3(4):485-506.

— — -. 1973. New or insufficiently known Australian crane flies (Diptera: Tipulidae). II. Studia

Entomologica, 16(1-4):261-314.

Byers, G. W. 1961. The crane fly genus Dolichopeza in North America. University of Kansas Science
Bulletin, 46(4):665-924.

- — — . 1969. Evolution of wing reduction in crane flies (Diptera: Tipulidae). Evolution, 23(2):346-
354.

Linnaeus, C. 1758. Systema naturae per regna tria naturae. Ed. 10, Vol. 1. 824 pp.

McAlpine, J. F. 1981. Morphology and terminology —adults. Pp. 9-63, in Manual of Nearctic Diptera
(J. F. McAlpine et al., eds.), Vol 1. Agriculture Canada Monograph, No. 27. 674 pp.
Oosterbroek, P., and T. Jonas. 1986. Catalogue of the Australian-Oceanian Tipulidae (Insect,
Diptera). Including a geographical index and references to the species. Institute of Taxonomic
Zoology, Amsterdam.

INSTRUCTIONS FOR AUTHORS

ANNALS OF CARNEGIE MUSEUM consist of con-
tributions to the natural sciences and anthropology, in
30 by 46 picas format ( 1 27 by 1 95 mm or 5 by 7% inches).
All manuscripts must first be submitted for review to the
Curator in charge of the appropriate scientific Section.
Authors should give particular attention to scientific con-
tent, format, and general style for the ANNALS. Manu-
scripts that do not conform to the style of the ANNALS
will be returned to the author immediately. Every manu-
script will be peer reviewed by at least two outside per-
sons. Authors will be asked to subsidize, if possible, any
or all costs of publication (approximately $ 100/page
printed).

Manuscript style.— Articles should include the follow-
ing items in this order: title page, abstract, text (with
desired headings), acknowledgments, literature cited, ta-
bles, figure captions, and copies of illustrations. All
manuscripts must be typed double-spaced on standard
m by 1 1 inch white bond paper, with at least one inch
margins all around, and submitted in triplicate— an orig-
inal for the editors, and two review copies. All pages
should be numbered, including tables, literature cited,
and the list of figure captions. Only correspondence qual-
ity or better dot matrix printouts will be accepted; draft
mode is unacceptable.

Title Page. The title, on a separate title page should
be brief, include the animal or plant group involved, and
appear two to three inches below the top margin. Include
the authors name(s) and the affiliations of non-CMNH
authors. In the case of multiple authorship indicate the
address to which proofs should be sent.

Abstracts. Abstracts should be short, double-spaced
and substantive and included at the head of the first page
of text.

Text. Do not right justify text or break (hyphenate) a
word at the end of a line. Footnotes and acknowledgments
as footnotes are unacceptable. All text is double-spaced.

Tables and figure legends. The list of figure legends
and each table should be typed consecutively on indi-
vidual pages separate from the text. Tables must be dou-
ble-spaced throughout with no vertical lines.

Abbreviations. Refer to the CBE Style Manual for cor-
rect abbreviations. Mammalian dentition: use capital let-
ters (I, C, P, M, D, for incisor, canine, premolar, molar,
deciduous, respectively) and superscript/subscript num-
bers (M1 2, P3) to designate upper and lower teeth.

Measurements. Metric units should be used, if possi-
ble.

Literature Cited. All references in text must appear in
the Literature Cited section. The data (author, date and
page) in both citations must agree. Do not abbreviate the
titles of periodicals or serials. The following style, dou-
ble-spaced, should be used in Literature Cited:

1) Two authors in an institutional series:

Matthew, W. D., and W. Granger. 1923. The
fauna of the Houldjin Gravels. American Mu-
seum of Natural History Novitates, no. 97:
1-6.

2) Same authors repeated— use three-em dash:

. 1923. The fauna of the Ardyn Obo For-

mation. American Museum of Natural
History Novitates, no. 98.T-5.

3) Same authors plus a third author— repeat all authors:

Matthew, W. D., W. Granger, andG. G. Simpson.
1928. Paleocene multituberculates from Mon-
golia. American Museum of Natural History
Novitates, no. 331:1-4.

4) Chapter in an edited volume:

Rausch, R. L. 1963. A review of the distribution
of Holarctic mammals. Pp. 29-43, in Pacific
basin biography (J. L. Gressitt, ed.), Bishop Mu-
seum Press, Honolulu, Hawaii, xx + 450 pp.

5) Unpublished dissertation:

Smith, J. P. 1976. Review of Eocene mammals.
Unpublished Ph.D. dissert., University of Cal-
ifornia, Berkeley, 302 pp.

6) Book:

White, M. J. D. 1961. The chromosomes. Meth-
uen and Co., Ltd., London, 120 pp.

7) Journal articles with usual volume and issue number:

Anderson, W. I. 1969. Lower Mississippian con-
odonts from northern Iowa. Journal of Paleon-
tology, 43(4):9 16-928.

Illustrations. All illustrations will be called figures, and
are to be numbered in Arabic numerals. Three sets of
illustrations are required, one (original artwork) for re-
production, two for reviewers. Xerox copies of photo-
graphs for reviewers are usually not acceptable but are
adequate for line drawing review copies. All illustrations
must be reducible to a maximum of 127 by 195 mm (30
by 46 picas) without loss of clarity. Line copy should be
designed for reduction to % or xh of actual size. Type-
written illustration copy will not be accepted. Photo-
graphic figures should be submitted at actual reproduc-
tion size.

Two or more small figures should be combined for
reproduction as a unit or plate, but, if feasible, do not
combine halftones and line drawings. Rectangular half-
tone figures should be abutted, without intervening spaces.
The printer will insert narrow white spaces during the
reproduction process. Halftone figures with solid black
backgrounds will not be accepted unless the author agrees
to pay for the printing costs. All plates must have min-
imally one inch borders all around. Each plate should be
given a protective cover and identified on the back side.

Lettering and/or a magnification scale (linear metric
scale) for rectangular halftone figures should be placed
directly on the photo, not in a blank space between pho-
tos. The scale or lettering for closely cropped photos can
be placed in blank areas close to the figure.

Proof— The author should answer all queried proof
marks and check the entire proof copy. Return corrected
page proof with the edited manuscript promptly to the
editors.

If an author chooses to make extensive alterations to
a paper in proof stage the author or appropriate section
will bear the cost. Original manuscripts will not be re-
turned unless requested. Illustrations will be returned to
the author.

issn 0097-4463

THE CARNEGIE MUSEUM OF NATURAL HISTORY

4400 FORBES AVENUE • PITTSBURGH, PENNSYLVANIA 15213

VOLUME 56

28 AUGUST 1987

CONTENTS

Art. 8. Taxonomic and geographic variation of Liophis typhi
related “green” species of South America (Serperites^

ARTICLES^ 14

s^nso/v^r

SEP 8 1987

Learies

Colubridae) . James *R. Dixon 173

Art. 9.

I '15 ' ■*

I

A new species of Clinidium Kirby (Coleoptera: Carabidae or
Rhysodidae) from Mexico, and descriptions of the females of

two neotropical members of the genus

Ross T. Bell and Joyce R. Bell 193

Art. 10. Biosystematic studies in Stenanthium (Liliaceae: Veratreae) II.

Floral morphology, floral vascular anatomy, geography and
taxonomy of the Mexican S. frigidum (Schlecht. & Cham.)

Kunth Frederick H. Utech 197

Art. 1 1 . Systematics of African bats of the Genus Eptesicus (Mammalia:
Vespertilionidae). 2. Karyotypes of African species and their

generic relationships

Karen McBee, Duane A. Schlitter and R. Laurie Robbins 213

Art. 12. Revision of the Wind River faunas, early Eocene of central

Wyoming. Part 8. First fossil lizard egg (?Gekkonidae) and list

| of associated lizards

.... Karl F. Hirsch, Leonard Krishtalka and Richard K. Stucky 223

i

Art. 13. Taxonomic notes on some African warblers (Aves: Sylviinae) . .

. .Kenneth C. Parkes 231

Art. 14. Spergenaspis : a new Carboniferous trilobite genus from North

America David K. Brezinski 245

Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS:

L. Krishtalka
J. L. Carter

Office of Scientific Publications
Phone (412) 622-3287

Communications about subscriptions, purchases of individual numbers, rates,
"and changes of address should be sent to:

Publications Secretary

The Carnegie Museum of Natural History

4400 Forbes Avenue

Pittsburgh, PA 15213

ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural
History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213, by the authority of the Board of
Trustees of Carnegie Institute.

THE CARNEGIE

MUSEUM OF
NATURAL HISTORY

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 8, Pp. 1 73-- 191 28 August 1987

TAXONOMY AND GEOGRAPHIC VARIATION OF
LIOPHIS TYPHLUS AND RELATED “GREEN” SPECIES OF
SOUTH AMERICA (SERPENTES: COLUBRIDAE)

James R. Dixon1
Abstract

The green species of Liophis are reviewed. Liophis typhlus consists of three subspecies, L. t. typhlus
(Amazon), L. t. elaeoides (Chaco), and L. t. brachyurus (Cerrado). Liophis viridis consists of two
subspecies, L. v. viridis (Agreste and Atlantic rainforest), and L. v. prasinus (Caatinga). Liophis jaegeri
is shown to consist of two subspecies, L. jaegeri jaegeri, east of the Rio Parana, from Sao Paulo area
of Brazil, to Uruguay and Argentina and L. j. coralliventris, from the Rio Paraguay basin. Liophis
guentheri is a valid species from the dry central Chaco of Argentina, Bolivia, and Paraguay. The
characters of the recently described species, L. atriventer and L. maryellenae are summarized. A key
is provided for all species of “green” Liophis.

Introduction

The study of the genus Liophis, a long and difficult task, is now nearing com-
pletion. The six species presented herein (L. typhlus, L. guentheri, L. jaegeri, L.
viridis, L. maryellenae, L. atriventer) form a loosely connected complex of species
that have one feature in common, they are all some shade of “green.” The green
color may be dull to bright, chlorophyll green to leaf green, or with a deep olive
cast. Two taxa occasionally have an obscure mid-dorsal stripe of variable width
that varies in color from light brown to reddish, and occasionally secondary dark
lines as well.

When samples were adequate, each of the six species was examined for geo-
graphic variation. No additional specimens of two recently described taxa ( L .
atriventer, L. maryellenae) are known, but the species’ essential features have been
summarized. Of the remaining four species detailed discussions include, where
appropriate, the currently recognized name, its synonyms, a discussion of the
synonyms, a description of the species, geographic variation, distribution, and
comments about pertinent literature.

Systematic Account

Liophis typhlus (Linnaeus)

Coluber typhlus Linnaeus, 1758. Syntypes (not examined)— Royal Museum, Stockholm. Type-local-
ity—Indiis (in error).

Xenodon isolepis Cope, 1870. Holotype ANSP (lost). Type-locality— Pebas, Ecuador (=Peru).
Opheomorphus brachyurus Cope, 1887. Syntypes ANSP 11202-03. Type-locality— Chupada, Mato
Grosso, Brazil.

Liophis elaeoides Griffin, 1916. Holotype CM R32. Type-locality— Prov. del Sara, Bolivia.

Liophis macrops Werner, 1925. Holotype NMW 23420. Type-locality— Paramaribo, Surinam.

Nomenclatural comments. — Several names have been assigned mistakenly to
the synonomy of L. typhlus. Jan (1863a, 1863Z?) and Jan and Sordelli (1866)

1 Department of Wildlife and Fisheries Sciences, Texas A&M University, College Station, Texas
77843-2258.

Submitted 16 January 1987.

173

174

Annals of Carnegie Museum

vol. 56

suggested that L. typhlus was composed of three varieties in addition to the
nominate race, prasinus, gastrostictus, and olivaceus. I have determined that no
holotypes exist for any of these varietal names. An examination of Jan’s descrip-
tions and Jan and Sordelli’s illustrations (1866, Livr. 18, Plate IV, Fig. 3) suggest
that the name prasinus belongs to Liophis viridis. The illustration shows 69 sub-
caudals, considerably more than the maximum known number of 6 1 for L. typhlus,
but well within the known range of 63 to 84 for L. viridis. An illustration of a L.
typhlus by Jan and Sordelli (1866, Livr. 18, Plate IV, Fig. 2) from “Pernambuco,”
Milan Museum, may represent the variety gastrostictus, of Jan ( 1 863a, 18636).
The illustration shows 47 subcaudals, well within the known range of variation
in subcaudals for L. typhlus. However, the closest localities of L. typhlus to
Pernambuco are 1 300 km to the east and 800 km to the south. The illustration
shows dark smudges along the lateral edges of the ventrals, and scattered dark
scale edges on the dorsum as well. The illustration and description more closely
resemble L. poecilogyrus from Pernambuco, and I suggest gastrostictus belongs
to that species. The Jan name olivaceus is without description or illustration. It
appears in Jan’s (1863a, 18636) list and key to Liophis as a varietal name of L.
typhlus and should be considered a nomen nudum.

Another misapplied name is Wagler’s (1824) Natrix forsteri. Hoge (1964) first
applied it to L. typhlus and suggested that the Surinam race be recognized as the
nominate taxon and the Brazilian form (by inference) L. typhlus forsteri. The type-
locality of L. forsteri is Salvador, Bahia, Brazil, a locality of considerable distance
from any known sample of L. typhlus. In addition, the illustration and description
of forsteri suggests that this name is best applied to L. poecilogyrus. The description
lists 60 subcaudals for the type of L. forsteri. The greatest number of subcaudals
recorded in 73 eastern Brazilian individuals of L. typhlus is 48. The description
of the color pattern of forsteri is unlike that of L. typhlus, and more similar to
that of L. poecilogyrus. Hoogmoed and Gruber (1983) examined the holotype of
N. forsteri in the Munich Museum and gave the subcaudal number as 58/58+1.
Although they placed N. forsteri as a synonym of L. typhlus, I consider N. forsteri
to be a senior synonym of L. poecilogyrus.

Parker (1 928) on comparing one of the paratypes of L. guentheri (BM 1 946. 1 .5.69)
to samples of L. typhlus in the British Museum (Natural History), concluded that
it “failed to reveal any consequential differences between these two species” and
placed L. guentheri in the synonomy of L. typhlus. Parker erred in his analysis
because the ventrals of the specimen he examined (and L. guentheri in general)
vary from 187 to 197, while ventrals of L. typhlus vary from 133 to 172. Liophis
guentheri is a valid species that is more closely related to L. viridis than to L.
typhlus.

Distribution. —Liophis typhlus occurs in Amazonian rainforests to an elevation
of about 1500 m along the southern and eastern side of the Andes in Colombia,
Ecuador and Peru, the rainforests of southern and eastern Venezuela, and similar
forests of Guyana, Surinam and French Guiana, and Cerrado and Chaco forests
of SE Bolivia, N Paraguay, and SC Brazil (Fig. 1).

Description. — The maximum total length is 740 mm in males and 853 mm in
females. Body scales are smooth, without apical pits and typically in 19-19-15
rows. Scale rows 3 + 4 or 4 + 5 typically fuse on each side of the body between
ventrals 68 and 103 (x = 82.2). Scale rows 7 + 8 or 8 + 9 usually fuse between
ventrals 65 and 104 (x = 83.0). A sample of 48 females and 25 males from Brazil
and 49 males and 62 females from other Amazonian localities were analyzed for
sexual dimorphism in all quantitative characters and none was found. Ventrals

1987

Dixon— South American Liophis

175

Fig. 1.— Distribution of Liophis typhlus (dots) and L. guentheri (diamonds, stippled area) in South
America. Black lines enclose numbered samples used in the geographic variation analyses.

of 331 specimens numbered 133-172 (x = 157.1), the number of subcaudals of
307 specimens varied from 41-61 (. x = 49.5). Tail length divided by total length
for 105 adult males varied from 0.145-0.206 (x = 0.169), and for 160 adult
females 0.137-0.210 (x = 0.162). Eye diameter divided by snout length of 127
adults varied from 0.524-0.886 (x = 0.682). Maxillary teeth for 222 specimens
varied from 19-28 (x = 22.4). The anal plate is divided and the loreal single in
all specimens examined.

Head shields vary as follows: supralabials 4-5 (1), 7-7 (3), 7-8 (2), 8-8 (249),
8-9 (5); supralabials entering orbit 4th only (1), 4 + 5 (252), 3+4/4 + 5 (2), 4 + 5/
5 + 6 (4), 4 + 5/4 + 5 + 6 (1); infralabials 5-7 (1), 8-8 (1), 9-9 (10), 9-10 (23), 10-10
(216), 10-11 (5), 11-11 (2), 11-12 (1); preoculars 0-0 (1), 1-1 (256); postoculars
1-1 (3), 1-2 (1), 2-2 (253), 2-3 (3); temporal condition 1 + 1 (3), 1+2 (242), 1 + 1/
1+2 (12), 1 +2/1+3 (3), 0 + 2/1+2 (1).

The in situ length for 69 hemipenes varied from 6 to 13 (x = 9.0) subcaudals.
The organ is slightly bilobed, with each lobe about two subcaudals long. The
sulcus spermaticus usually bifurcates about one-third the distance from the base

176

Annals of Carnegie Museum

vol. 56

of the organ. Large spines are present on the asulcate surface of the basal two
thirds of the organ, usually becoming much smaller on the lobed part of the organ.
A smooth apical disk is present on the outer edge of each lobe and a naked basal
pocket is usually present.

Color in life.— The dorsal color of some adults is light blue on the head and
various shades of green on the body and tail. Amazon samples have reddish
brown to black chevron marks over most of the lateral (occasionally middorsal)
surfaces of the body. These dark marks are usually prominent in young and
juveniles, less so in adults. Some individuals from throughout the range have
scattered white scales mixed with the green. Other individuals are uniform green
and lack dark chevron marks and blue heads. The venter is usually white or light
yellow without darker markings. However, a few individuals have dark smudges
or distinct marks ventrolaterally, and the subcaudals may have dark spots or lines.

The juvenile dorsal pattern is variable with some juveniles having distinct black
chevrons with a wide black nuchal band. The nuchal band usually fades to an
obscure greenish black mark at a total length of 210 mm. Other juveniles from
155-255 mm in total length have a pair of blackish nuchal spots that begin on
or at the posterior edge of the parietals, and slant posteroventrally. Occasionally
a secondary pair of medium sized dark spots occur on the neck, followed by two
rows of paravertebral dark spots and two rows of lateral dark spots to above the
vent. Sometimes, the body appears to be reticulated with darker lines on a greenish
ground color. One young individual (SVL 225 mm) has distinct spots on scale
rows one, two, four, six, and seven, and occasional dark spots along the edges of
the ventrals.

Geographic variation.— There are three distinct geographic populations of L.
typhlus that correspond to forest refugia postulated by Vanzolini and Williams
(1970), and to some extent, to those of Haffer (1974). The three geographic pop-
ulations were determined by comparison of the numbers of maxillary teeth, ven-
trals, subcaudals, the ratio of tail length to total length, and the ratio of the eye
diameter to snout length. Six samples were used to determine variation. The
samples were chosen by the proximity of individuals to each other and by natural
vegetation types, as follows: Sample 1-27 specimens from Surinam and French
Guiana; Sample 2-24 specimens from Guyana and Venezuela; Sample 3-53 spec-
imens from Amazonian Colombia, Ecuador, Peru, western Brazil, and Bolivia;
Sample 4-29 individuals from Amazonian Brazil; Sample 5—45 specimens from
the Chaco of Bolivia and western Mato Grosso, Brazil; and Sample 6 — 87 spec-
imens from the Cerrado of southeastern Brazil (Fig. 1). The Student’s T test was
utilized to determine significance of pair-wise comparisons between samples. A
Student’s T test value of 6.3 14 or greater is at 95% level of significance or greater.
Pair-wise tests for differences between samples 1 and 2, 1 and 4, 2 and 3, 2 and
4 were not significant for any of the characters listed above. Samples 1 and 3 were
significantly different from each other in number of ventrals and tail/total length
ratio. However, samples 1 and 3 are separated by sample 2, and sample 2 did
not differ significantly from either 1 or 3. The significance noted between 1 and
3 is probably an artifact of distance, with sample 2 representing the middle of a
cline. Samples 3 and 4 were not significantly different in any characters except
ventrals. This may also be an artifact of distance. However, two males from
reasonably close localities in Amazonian Bolivia had ventral counts of 140 and
157. Their ventral counts fall near the average of the samples with which they
were associated, 143.5 for sample 4, and 154.2, for sample 3 (see Fig. 1). This
suggests that a hiatus may exist between samples 3 and 4 in Amazonian Bolivia.

1987

Dixon —South American Liophis

111

A statistical analysis of ventral number between samples appears to show low to
high numbers of ventrals from samples 1 to 4, 1 to 2, 2 to 4, 2 to 3. The samples
form an almost closed circle of demes (Fig. 1), with the bottom of the circle open,
representing the differences in the Bolivian samples mentioned above. It seems
clear that the Amazon samples are closely related, but with a trend to divide the
samples into eastern and western demes.

Pair-wise comparisons of samples 4 and 6 suggest a strong differentiation in all
characters examined. The differences in the numbers of ventrals, subcaudals, and
tail/total length ratios are highly significant between the two samples, with T
values of 23.8, 19.1, and 15.8 respectively. The number of maxillary teeth differed
significantly (T value = 7.0) but eye diameter/snout length ratios did not (T value =
2.14). The data suggest that the Brazilian Amazon forest and deciduous meso-
phytic forest samples represent different allopatric taxa. Comparison of samples
3 and 6 shows the same trend, except the number of maxillary teeth is not
significant (T = 4.91). Pair-wise comparison of samples 4 and 5 shows significant
differences in the number of ventrals and maxillary teeth (T = 22.03, 10.54,
respectively), but not in the number of subcaudals, tail/total length, or eye di-
ameter/snout length ratios (T = 2.66, 4.59, 4.62, respectively). Pair-wise com-
parison of samples 3 and 5 shows significant differences in the number of ventrals,
maxillary teeth, and subcaudals (T = 9.07, 8.49, 6.36, respectively), but not in
the tail/total length and eye diameter/snout length ratios (T = 2.96, 6.0). The
relative differences between the Amazon samples (1-4) and the two non- Amazon
samples (5-6) suggest that the Chaco forest sample from Bolivia is more closely
related to the Amazon forest sample than to the deciduous mesophytic forest
sample of Brazil. However, both non-Amazon forest samples are very distinct
from the Amazon forest samples.

A pair-wise comparison of samples 5 and 6 shows significant differences in the
number of subcaudals, maxillary teeth, and tail/total length ratios (T = 21.05,
6.43, 12.9, respectively), but not in the number of ventrals or eye diameter/snout
length ratio (T = 1.36, 3.59). These samples represent different taxa, even though
they are relatively close geographically (Fig. 1).

The scale count data and length ratios suggest three distinct populations of
Liophis typhlus. Samples 1, 2, 3, and 4 represent the Amazon forest form, sample
5 the northern Chaco form, and sample 6 the Cerrado form.

The color patterns of samples 1, 2, 3, and 4 are similar, and the patterns of
samples 5 and 6 are identical. The dorsum and head of adult individuals from
samples 5 and 6 are chlorophyll green. The young and juveniles have black nape
spots and some have four rows of dorsal dark spots. Adults of the Amazon samples
have a leaf green to bright green dorsum with diagonal reddish brown to blackish
chevron marks along each side of the body. The head may be greenish to bright
blue. Young and juveniles are greenish dorsally with bold dark chevron marks
and a relatively large black nuchal blotch.

Diagnoses and distributions of the three taxa follow:

Liophis typhlus typhlus (Linnaeus)

Coluber typhlus Linnaeus, 1758.

Xenodon isolepis Cope, 1870.

Liophis macrops Werner, 1925.

Diagnosis. — Dorsum usually green with distinct, reddish brown to black chev-
rons on each side, occasionally fading posteriorly. The dorsal surface of the head

178

Annals of Carnegie Museum

vol. 56

may be greenish to bright blue. The venter is white to light yellow, usually without
dark markings. Juveniles have a large black nuchal blotch and dark chevrons.
The number of ventrals varies from 1 33 to 1 63 (x = 147.3), subcaudals vary from
47 to 61 (x = 54.2), maxillary teeth vary from 20 to 28 (x = 23.6), tail/total length
ratios vary from 0.160 to 0.210 ( x = 0.184), and eye diameter/snout length ratios
vary from 0.510 to 0.861 (x = 0.736).

Distribution. — Throughout the Amazon Basin in rainforest, to about 1000 m
elevation (see specimens examined).

Liophis typhlus elaeoides Griffin
Liophis elaeoides Griffin, 1916.

Diagnosis. — Dorsally adults are uniform chlorophyll green with a light yellow
venter. Juveniles have four rows of dorsal dark spots on a green background, and
a pair of black nuchal spots. The number of ventrals varies from 158 to 172 (x =
163.5), subcaudals vary from 49 to 56 (x = 51.9), maxillary teeth vary from 17
to 23 (x = 19.9), tail/total length ratios vary from 0.160 to 0.200 (x = 0.171),
and eye diameter/snout length ratios vary from 0.520 to 0.740 (x = 0.620).

Distribution. — Liophis typhlus elaeoides is known only from the upper Rio
Paraguay Basin, which includes the mesic Chaco forests of southeastern Bolivia,
northern Paraguay, and western Mato Grosso, Brazil (see specimens examined).

Liophis typhlus brachyurus (Cope)

Opheomorphus brachyurus Cope, 1887.

Diagnosis. — Adults are uniform green dorsally with a light yellow venter. Ju-
veniles have four rows of dark dorsal spots on a green background and a pair of
black nuchal spots. The number of ventrals vary from 151 to 171 (x = 162.2),
subcaudals vary from 40 to 49 (x = 44.4), maxillary teeth vary from 18 to 24
(x = 21.5), tail/total length ratios vary from 0.140 to 0.160 (x = 0.149), and eye
diameter/snout length ratios vary from 0.55 to 0.81 (x = 0.675).

Distribution.— Liophis t. brachyurus occurs in deciduous mesophytic forests of
southeastern Brazil, and in the Campos Cerrado forests of east-central Brazil (see
specimens examined).

Comments.— Two recent papers by Miranda and Couturier (1983, 1984) com-
ment on the presence and geographic variation of L. typhlus in Argentina. I have
neither examined their specimens, nor located specimens of L. typhlus from
Argentina in other museums. I believe they may have erred in identification of
their specimens. Photographs presented by Miranda and Couturier (1984) suggest
that they may have confused L. miliaris and L. poecilogyrus with L. typhlus.

Short works on the distribution and taxonomy of L. typhlus are Hoge (1964),
Amaral (1931, 1935, 1949), Gans (1960), Parker (1928, 1935), Peters (1960),
Peters and Orejas-Miranda (1970), Roze (1966), and Prado and Hoge (1948).
Peters (1963) compared the maxillary teeth of L. typhlus with those of other
species of Liophis.

Specimens examined. — ( Liophis typhlus typhlus) BOLIVIA, Beni : Ixiamus AMNH 22457; Rio Beni
AMNH 22270. BRAZIL, Acre: Alto Purus MZUSP 2498; Porto Walter MZUSP 7390. Amapa:
Cuidade Oiapoque IB 13780, 13783; Serro do Navio KU 97873-97875. Amazonas: Barreira do
Matupiri, on the Rio Madeira MZUSP 5912; Carvoeiro AMNH 36167; Costa Altamira on the Rio
Japura MZUSP 6600; Lago Alexo MCZ 3290; Manaus MZUSP 3051, 3797; Reserva INPA MZUSP
7606, 7619; Santa Isabel on the Rio Negro USNM 83532; Sao Paulo de Olivenca AMNH 53311;

1987

Dixon— South American Liophis

179

Tapaua MZUSP 5770; no specific locality UZM 601223. Maranhao : Aldeia Aracu Igarape Gurupi-
Una MZUSP 4303, 4826. Para: As Pedras, on the Rio Cumina-Miri MZUSP 5103; Belem IB 15667;
Caninde, on the Rio Gurupi MZUSP 4267, 4285; near Maraba, Serrando Norte KU 1 24608; Oxiximina
MZUSP 4796; Urua, Parque Nac. da Amazonia, on the Rio Tapajos MZUSP 7838. Rondonia: Porto
Velho MZUSP 3690. COLOMBIA, Macanal : Rio Garagoa MCZ 27339. Meta : Villavicencio IB 8589.
Vaupes: Timbo UTA R3805; Yapima UTA R5033. ECUADOR, Napo-Pastaza : Alpayaca FMNH
4069, UMMZ 89020-22; Andoas AMNH 41949; Banos+Canelos AMNH 35892; Chamala-Nor-
mandia-Rio Bamba AMNH 35929-35930; Jarayacu AMNH 28796; Macas and Vicinity AMNH
35838-35839; 82 km ESE Macas AMNH 114616; Rio Bamba AMNH 15209, 15213, 23296, 23302,
28848; Rio Pastaza MCZ 36966; Santa Rosa, El Tigre AMNH 49165. FRENCH GUIANA (CAY-
ENNE), Mana USNM, 6172; Saul MCZ 149399; near Sophie MCZ 77510-775 1 1 . GUYANA, Bartica-
Potaro road BM 1954.1.3.64; 82 km S Bartica BM 1934.11.1.125; Berbice BM RR1964; Demerara
River, Lama Creek AMNH 36106; Essequibo, near Lethem USNM 146376; Kartabo AMNH 21335,
98197-98198; Matali AMNH 61542; Rupununi District, north of Acarai Mountains, west of New
River KU 69826-69828; Rupununi District, north of Acarai Mountains, west of New River KU
69826-69828; no specific locality MNHG 279.38. PERU, Amazonas : Caterpiza, Rio Caterpiza USNM
(RWM field series) 14995, 15033, 1 4 185-1 5 1 86. Junin: Rio Perene MCZ 42434. Loreto: BaltaLSUMZ
14584: Centro Union TCWC 44682; Cerros de Contaya, on the Rio Tapiche AMNH 53376; Con-
tamana on the Rio Ucayali AMNH 52130; Estiron on the Rio Ampiyacu MZUSP 4394; Iquitos
AMNH 52734; 53118, 53284, 53667, 53696, 53735, 53771, 53773, 53876, 53923, 53949, 54321,
54354, 54483, 54894, 56109; Moropon TCWC 38049, 44294; Orellana (Reforma) AMNH 54957,
USNM 127124; Pampa Hermosa on the Rio Cushabatay AMNH 55409, 55415, 55442, 55791, 55885,
56003; Pamya AMNH 53249; Requena (Monte Carmelo) AMNH 55600, 55626; Royaboya AMNH
52483, 53110, 53288, 55695; San Antonio on the Rio Itaya AMNH 52920, 53667, 53693, 53696,
53735; Shiriara, on the Rio Nanay AMNH 56075; Trapiche-Utuquinia AMNH 52195; Yanamono
TCWC 40542-40543. Madre de Dios: Rio Heath, 50 km S Puerto Pardo LSUMZ 36778-79; mouth
of Rio Torre on the Rio Tambopata LSUMZ 394244. Pasco: Iscozazin Valley LACM 76805. San
Martin: Moyobamba BM 74.8.4.59, 94.8.4.64. SURINAME: Charlesburg AMNH 104624; Coppen-
ame River MCZ 1 52203, 1 52205, 1 52633-1 52634; Jaraweg TCWC 60543, 60756; Paramaribo AMNH
8146, 8682, MCZ 16401; Potribo CM 44302; Sipaliwini Airstrip CM 84667, MCZ 152636; Zanderi
(Airport) MCZ 152635. VENEZUELA, Amazonas: Arabopo UMMZ 85279; Arocoima Canos MCZ
38541; Mount Duida region AMNH 36617, 36620. Bolivar: El Manteco TCWC 60168; no specific
locality MHNLS 1632.

(. Liophis typhlus elaeoides ) BOLIVIA: Cochambamba: Villa Tunari UMMZ 153095. Santa Cruz:
Buenavista CM R2696, R2698, R270 1 , R2704, R2869, R2886, R2930, R293 1 ; UMMZ 67967-67968,
67969(2), 67970(3); Buenavista, near Rio Colorado CM R2860, R2865, R2955; Santa Cruz LSUMZ
11825; Santa Cruz de la Sierra CM R32 (holotype), R44, R59, R91, R94, R95, R97, R98, R99, R102,
MCZ 11860, MZUSP 6474; San Jose de Chiquitos CM 34842. No specific locality: CM R2938-
R2939, FMNH 195898, HCD 2820, 2822, TCWC 55290. BRAZIL: Mato Grosso: Corumba, near
Urucum Mountains CM 3484 1 ; Fazenda Bela Vista of the Ilha Insua MZUSP 7264; Fazenda Vacurizai
of the Rio Paraguai MZUSP 7266; Generalso Ponce, Corumba IB 25954; Maleta IB 14975; Porto
Murtinho IB 26177-78.

(. Liophis typhlus brachyurus ) BRAZIL: Bahia: Mira Serra, 41 km from Morro de Chapeo MZUSP
7554. Goaiz: Araguari IB 6851; Fazenda Lucushac, on the Rio Verde IB 13060. Mato Grosso: Arapua
IB 9939; Aquidauana MZUSP 33589; Chapada ANSP 11202-11203, BM 92.420. 1 3; Taugara da Serra
IB 24543; Taunay IB 7674; Urucum, near Corumba BM 1928.1.12.3. Minas Gerais: Horto Forestal
IB 10493; Irma Badur IB 6933; Itambe do Mato Dentro MZUSP 8061; Juiz de Fora IB 25190, 26685;
Machado IB 16303; Santa Rita do Extrema IB 5540, 5562. Rio de Janeiro: Nova Friburgo IB 10516.
Sao Paulo: Araraguara IB 231; Atibaia IB 21376, 27081, 27201; Barra Assugnui IB 29042; Barretos
IB 5030; Baureri IB 7212; Bauru MZUSP 823; Boraceia MZUSP 4227; Botucatu MZUSP 2410, 2412;
Buri IB 6718, 6763, 12313; Campo Limpo IB 805, 6452, 10265-10267, 24903-24904; Capao Bonito
IB 23465; Cascavel MZUSP 824; Conde do Pinhal IB 9975; Cotia IB 5743, 6935, 7140, 9906, 19906,
23236, 23377, 24542, 32627; Curupa IB 7061; Dona Catarina IB 22598; Elias Fausto IB 232, 319,
392, 578, 791, 10461; Elihu Root IB 234; Engenheiro Cesar de Souza IB 8338; Ferraz de Vasconcelos
MZUSP 2487; Guarulhos IB 23726-23727; Horto Florestal IB 10493; Ibiuna IB 12252, 23794;
Ipanema IB 230, 235; Itagua IB 48 1 3; Itapecerica da Serra IB 7794, 1 9682, 25040, 25043; Itapetininga
IB 24565; Itapolis IB 7956; Itaquera IB 30758; Itaquaquecetuba IB 18349; Itirapina IB 6655; Jaragua
IB 21381; Jarinu IB 10292; Lauro Muller IB 10533, 10556, 10559; Leme IB 236, 547; Mairinque IB
7373, 7461; Mairipora IB 28603; Mogi das Cruzes IB 12592, 19658, 23547, 24881, 28284, MCZ
39415, MZUSP 6460; Osasco IB 12500, 27910-27911; Pedemeiras IB 6938; Pindamonhangaba IB
1257; Piraju MCZ 39416; Promissao IB 9969; Ribeiras Pires IB 19596; Rio Claro IB 7277, 10296;

180

Annals of Carnegie Museum

vol. 56

Santa Adelia IB 6909; Santa Emestina IB 27969, MZUSP 2479; Sao Carlos IB 10016; Sao Jose do
Rio Pardo IB 5444; Sao Jose dos Campos IB 21348; Sao Paulo IB 797, 7142, 12980, 12993, 16223,
16900, 18854, 25295, 27497, 28458, MCZ 17956, MZUSP 817-820, 2388, 3338-3339, 3695, USNM
69238-69239; Suzano IB 19104; Valinho IB 6719; Vinhedo IB 2915; no specific locality IB 233, 237,
797, 840, 4745, 5491, 5642, 6950, 9775, 10581, 12112, 17777.

Liophis guentheri Peracca

Liophis guentheri Peracca, 1 897. Six syntypes (Five lost ?), Mus. Comp. Anat., Torino. Type locality—
Caiza, Bolivia.

Comments. — Liophis guentheri has resided as a synonym of L. typhlus. Parker
(1928) compared one of the syntypes (now BM 1946.1.5.69) with a specimen of
L. typhlus , and suggested it “failed to reveal any consequential differences between
the two species.” However, no overlap in the number of ventrals occurs between
the two species. An unnumbered British Museum specimen from an unspecified
locality in Bolivia is similar to the numbered syntype in the collection of the
British Museum in its state of preservation, and may well be one of the five
missing syntypes that were housed in the Museum of Zoology and Comparative
Anatomy in Torino.

Distribution.— Liophis guentheri apparently is restricted to the central Chaco
of Bolivia, Argentina and Paraguay (Fig. 1).

Description. — The following description is based on six adult females and three
adult males. The maximum total length of males examined is 655 mm, of females
723 mm. Peracca (1897) gave 890 mm as the maximum total length but did not
indicate the sex. Scale rows are 19-19-15, smooth, without apical pits; a reduction
of the third and fourth scale row occurs between ventrals 98-118 (x = 106.9),
and reduction of the seventh and eight scale rows between ventrals 99-112
(x = 106.1). The number of ventrals varies from 187 to 197 ( x = 191.6), the
number of subcaudals from 53 to 57 (x = 54.6) (only five with complete tails).
The number of maxillary teeth varies from 20 to 22 (x = 20.9). Head scales are
as follows: supralabials 7-8 (1), 8-8 (7), 8-9 (1); supralabials entering orbit 4 + 5
(8), 4 + 5/5-F6 (1); infralabials 9-9 (1), 9-10 (1), 10-10 (5), 10-11 (2); preoculars
1-1, postoculars 2-2, loreals 1-1, temporals 1 + 1 (1), 1 +2/1 + 1 (1), 1+2 (7). The
anal plate is divided. The tail/total length ratio varies from 0.152-0.168 (x =
0.159); eye diameter/snout length ratios vary from 0.467 to 0.578 (X = 0.502).

The hemipenis is nine subcaudals long, with the sulcus spermaticus dividing
at the level of the third subcaudal, slightly bilobed (about one and a half subcau-
dals), and very spinose. The smooth apical disks are pronounced.

The general dorsal color is light blue in preservative, but Peracca (1897) states
that living specimens are immaculate green dorsally and yellowish white ventrally.
The ventrals and subcaudals are usually immaculate white in preservative. In two
preserved specimens there are dark flecks and/or spots on the outer edges of the
ventrals. Flecks cover all ventrals in one specimen and three fourths of the ventrals
in the other.

Specimens examined.— ARGENTINA: Salta : Hickmann IML 105, 303; Luna Muerta IML 301;
Formosa : Ingeniero Juarez IML 422, 490. BOLIVIA: no specific locality IB 1177, BM (unnumbered);
Chaco : Caiza BM 1946.1.5.69. PARAGUAY: Chaco : Loma Plata KU 73454.

Liophis viridis Gunther

Liophis viridis Gunther, 1862. Holotype, BM 1946.1.5.69. Type-locality— Brazil, Pernambuco, no
specific locality.

Liophis typhlus prasina Jan and Sordelli, 1866. Holotype (lost?). Type-locality— Brazil.

1987

Dixon— South American Liophis

181

Fig. 2.— Distribution of Liophis viridis in northeastern Brazil. Dots within the diagonal lines (Agreste
plant formation) and lightly stippled area (Atlantic rainforest formation) represent the Agreste- Atlantic
Rainforest sample. Dots within the dashed line (Caatinga plant formation) represent the Caatinga
sample. Densely stippled areas within the Caatinga formation represent enclaves of the Cerrado plant
formation.

Distribution.— Known only from the Caatinga, Agreste, and Atlantic rainforest
vegetation communities of northeastern Brazil (Fig. 2).

Description. — The following description of L. viridis is based on 1 59 individuals
from throughout the range. The maximum total length of males is 595 mm, and
of females 670 mm. The scale rows are 19-19-17, smooth, and with one apical
pit. The reduction to 1 7 rows occurs through a fusion of scale rows three and four
between ventrals 98 and 121 (x = 110.0). Sexual variation was not detectable
even in large samples, therefore all scale data were combined for intrasample
comparisons. Ventrals vary from 169 to 202 (x = 186.0); subcaudals from 63 to
84 (x — 74.1). Tail/total length ratios of adults vary from 0.193 to 0.249 (x =
0.221). The eye/snout ratios of adults vary from 0.500 to 0.684 (x = 0.562).

182

Annals of Carnegie Museum

vol„ 56

Maxillary teeth vary from 17 to 24 (x = 20.2), with the last two teeth enlarged
and separated from the remainder by a decided gap. The anal plate is divided in
all specimens examined.

Head scales vary as follows: supralabials 8-8 (156), 8-9 (2), 9-9 (1); supralabials
entering orbit 4 + 5 (155), 4 + 5/5 + 6 (3), 4 + 5 + 6/5 + 6 (1); infralabials 8-9 (2),
8-10(1), 9-9 (2), 9-10(4), 10-10(144), 10-11 (4), 11-11 (2); preoculars 1-1 (156),
1-2 (2), 2-2 (1); postoculars 1-1 (1), 2-2 (157), 2-3 (1); loreals 1-1 (159); temporals
1 + 1 (12), 1 + 1/1+2 (21), 1+2 (145), 1 +2/1+3 (1).

The hemipenis is 9 to 13 (x = 10.8) subcaudals long (in situ) and bilobed for
1.5 to 3 subcaudals. Large spines occur from the base to the proximal part of the
lobe. Spines about one half the size of the basal spines occur from the lobe to the
distal end. The sulcus spermaticus is divided near the base of the lobes. The
smooth apical disk is relatively large.

The dorsum is bright green in life. Young and small juveniles may have black
bands in the nape region followed by a series of black spots arranged in four linear
rows, primarily along scale rows three and four, and seven and eight. The number
of spots is variable in any row, varying from 36 to 60, especially in individuals
from 180 to 220 mm in total length. Some young have only a nape band, im-
mediately followed by one pair of dorsal spots. Occasionally young have a black
lateral tail stripe. Spots on juveniles from 240 to 370 mm in total length are faded
posteriorly. Normally all individuals over 400 mm in total length are uniform
green and the belly and subcaudals are immaculate white. Vanzolini et al. (1980)
give a very good color description of living individuals of this species.

Geographic variation.— A Caatinga sample and an Agreste- Atlantic rainforest
sample were formed from three coastal subsamples and two inland subsamples
of L. viridis (Fig. 2).

The Caatinga sample is significantly distinct (T = 13.4) from the Agreste-
Atlantic rainforest sample in having a higher number of ventrals, 181-202 (x =
189.8), versus 169-188 (x = 179.6), and a higher number of ventrals (T = 6.44)
at the 19-17 reduction site 102-124 (x = 1 14.6) versus 98-1 16 (x = 106.6). Other
scale and tooth characters are not significantly different between samples.

One specimen from Jeremoabo, Bahia, and five specimens from Mina de Sao
Felix do Amianto, Bahia, are known from the eastern part of the Caatinga. Their
taxonomic features more closely approximate those of the Agreste-Atlantic rain-
forest sample than others from the Caatinga. Although the samples are small I
suspect that they may represent intergrades between two well-defined subspecies.

The Caatinga sample (Fig. 2) does not have a formal name. However, Jan
(1863a, 1983/?) and Jan and Sordelli (1866) mention a varietal name associated
with L. typhlus that belongs to L. viridis. It was originally cited as a varietal name
(without description) by Jan (1863a, 1863/?) but was illustrated by Jan and Sordelli
(1866) and their iconotype represents the only description. Jan (1863a, 1863 b)
recorded two specimens of the variety, one from “Pernambuco” in the Milan
Museum, and one from “Brasile” in the Stuttgart Museum. Jan and Sordelli
illustrated the latter specimen, thus it becomes the holotype (by inference). How-
ever, neither specimen now can be located. I cannot definitely assign the name
to either of the two populations that require recognition. However, the genus is
already overburdened with excessive synonyms and I propose that Jan and Sor-
delli’s varietal name, prasinus, as proposed in the combination L. typhlus prasinus,
be recognized as the proper name for the Caatinga popultion of L. viridis.

1987

Dixon— South American Liophis

183

Liophis viridis viridis Gunther

Liophis viridis Gunther, 1862.

Diagnosis. — Dorsally adults are leaf green to bright green, and pale cream to
white ventrally. Juveniles have a black nape band followed by a various number
of paired black dorsolateral spots. The number of ventrals varies from 169 to 188
(x - 179.6), and reduction occurs between ventrals 98 and 1 16 (x = 106.6).

Distribution. —Liophis v. viridis occurs in the Atlantic rainforest and the Agreste
forest from Recife, Pernambuco, east and south to Salvador, Bahia, Brazil.

Liophis viridis prasinus Jan and Sordelli

Liophis typhlus prasina Jan and Sordelli, 1866.

Diagnosis. — Adults are bright green, with an immaculate cream to white venter.
Juveniles have bright green dorsums with a single black nape band, or with four
parallel rows of 35 to 60 black spots per row. Occasional juveniles have a pair of
posterior, lateral black stripes that extend onto the tail. The number of ventrals
varies from 1 8 1 to 202 (x = 189.8), with the reduction occurring between ventrals
102 and 124 (x = 114.6).

Distribution. —Liophis v. prasinus occurs from Joao Pereira, Maranhao in north-
eastern Brazil, south to Sao Francisco, Minas Gerais, then northeast to about
Parauagua, Rio Grande do Norte, Brazil.

Comment.— Yin (1983) found Hyla rubra, Physalaemus cuvieri, and several
unidentified leptodactylids in 14 stomachs of L. viridis. Vitt examined 16 gravid
females and determined that L. viridis produces continuous clutches throughout
the year. Ovarian clutch size varied from two to six (x = 3.69) eggs.

Short taxonomic works for L. viridis include Cordeiro and Hoge (1973), Gomes
(1918), and Schmidt and Inger (1951). Vanzolini et al. (1980) give an excellent
summary of scale data, color of young and adults, and brief biological notes on
a Caatinga sample.

Specimens examined.—" SOUTH AMERICA”: CM R281. BRAZIL: no specific locality FMNH
5694, MZUSP 815, NHMW 20739; no specific state, Parauagua ZMUC 601250. Algoas : Quebrangulo
MZUSP 3453. Bahia: No specific locality FMNH 5721, IB 3200, MCZ 2946, 147217-147218. Assu
da Torre IB 980, Brreiras UMMZ 108712-108716, Bom Jesus de Lapa Cm R344, UMMZ 108711,
Feiro de Santana KU 29478, Itapoan MZUSP 5713, Jeremoabo MZUSP 5738, Mina de Sao Feliz
do Amianto IB 28170-28172, 28215-28216, Salvador BM 90.1.27.2, 1924.9.20.26, MZUSP 2838-
2839. Gear a: Acude Amanari Maranguape MZUSP 3450, Boa Viagem IB 20200, Caxueira ZMUC
3342(3), Coluna MZUSP 53 1 9-20, Limoeiro do Norte IB 1 2776-12777, Unique IB 20142. Maranhao:
Lagoa de Joao Pereira CM R316. Minas Gerais : Januaria UMMZ 108705-108709, Sao Francisco
UMMZ 108710. Parahyba: Coremas MZUSP 3451, Joao Pessoa MZUSP 7976-7977, 7999, 8262,
UFFB 139, 141, 143, Mamanguape MZUSP 3456. Pernambuco: no specific locality MANH 4448,
BM 1946.1.2.69, 80.1 1.25.4, 80.1 1.25.8, MCZ 1447, 146945-146947, Agrestina MZUSP 4925, 4942,
4946, 4965, 4970, Exu MZUSP 6693-6699, 6700-6718, 7071-7091, 5 km E Exu MZUSP 6950, 13
km Exu MZUSP 6940-6943, 5 km N Exu MZUSP 6920-6921, 1 1 km S Exu MZUSP 7092, Ponta
de Pedras MZUSP 5177, Recife DEH 624, 517, 526, LSUMZ 36787. Piaui: Acude Peri-Peri MZUSP
3424, 3427-3429, 3433, 3435-3438, 3441, 3443, 3445, Terezina IB 360, 473, 1210-1214, Valenca
MZUSP 5814. Rio Grande do Norte: Ceara Mirim CAS 49320. Sao Paulo: Leme IB 543 (in error).
Sergipe: Aracaju IB 22458, Maruim ZMUC 601251, Santo Amaro das Brotas MZUSP 6985, 7306.

Liophis jaegeri (Gunther)

Coronella jaegeri Gunter, 1858. Syntypes—BM 1946.1.9.12, 1946.1.5.78. Type-locality— Brazil, no
specific locality.

184

Annals of Carnegie Museum

vol. 56

Fig. 3.— Distribution of Liophis jaegeri in southeastern South America. Black lines enclose samples
utilized in the analyses of geographic variation. The large black dot represents several overlapping
localities.

Liophis dorsalis Peters, 1863. Type-locality— Brazil, no specific locality.

Aporophis coralliventris Boulenger, 1894. Type-locality— island north of Concepcion, near San Sal-
vador, north Paraguay.

Rhadinaea lineata Jensen, 1900. Type-locality: Taboleiro Grande, Minas Geiras, Brazil.

Misapplied synonym. —Rhadinaea dichroa Werner (1899) has been in the syn-
onomy of L. jaegeri since 1972 (Peters and Orejas-Miranda, 1972). However, an
examination of the type description reveals that the holotype has 19 midbody
scale rows, 159 ventrals, and 50 subcaudals. The dorsum is brown with the bases
of the scales white and the posterior edges black. It also has a tail/total length
ratio of 0. 1 85. The description is identical to that for Liophis poecilogyrus caesius,
a form common to the Chaco of Bolivia, Argentina, Paraguay and Brazil, and I
assign it to that species.

Distribution. — Liophis jaegeri occurs from about 1 9°S latitude in Brazil, to about
35°S latitude in Uruguay and Argentina, west to about 61°W longitude, along the
Rio Paraguay basin in Argentina, Paraguay and Brazil (Fig. 3).

Description. — Some 415 individuals of L. jaegeri were examined and full data
taken on 160. Tail/total length ratios were recorded for over 400 individuals. The
maximum total length of males is 539 mm, of females 676 mm. Dorsal scales

1987

Dixon —South American Liophis

185

are smooth, in 17 rows, normally without reductions, and without apical pits.
When reductions are present ( 1 5 individuals), they occur with a fusion of scale
rows three and four on one or both sides of the body, reducing to 1 5 or 1 6 between
ventrals 88 and 141 (x = 121.2).

Sexual variation in the number of ventrals, subcaudals, maxillary teeth, and
tail/total length ratios is absent, therefore the samples were combined for statistical
analysis. The number of ventrals of 160 individuals varies from 146 to 169 (x =
157.5), and subcaudals from 52 to 75 (x = 61.7). The number of maxillary teeth
varies from 22 to 29 (x = 25.6). The tail/total length ratios varies from 0.181 to
0.268 (x = 0.221). The number of palatine and pterygoid teeth were recorded for
one individual and they were 10 and 25 respectively.

Head scales are as follows: supralabials 6-7 (1), 7-7 (1), 7-8 (4), 8-8 (147), 8-9(4),
9-9 (1); supralabials entering the orbit 3 + 4 (1), 3 + 4/4 + 5 (3), 3 +4 + 5/4+ 5 (1),
3+4 + 5/4 + S + 6 (1), 4 + 5 (148), 4 + 5/4 + 5 + 6 (2), 4+5 + 6 (1); infralabials 8-8
(3), 8-9 (4), 8-10 (2), 9-9 (5), 9-10 (24), 9-11 (1), 10-10 (112), 10-1 1 (5); preoculars
1-1 (155), 1-2 (2), 2-2 (1); postoculars (2-2) and loreals (1-1), temporals 1 + 1 (5),
1-1/ 1-2 (11), 1 +2/2+1 (1), 1+2 (141); eye diameter/snout length ratios of 20
adults varies from 0.533 to 0.844 (x = 0.655). The anal plate is divided in all
specimens. Hemipenial length varies from 7 to 13 (x = 10.3) subcaudals.

A hemipenis extending 12 subcaudals has the sulcus spermaticus divided at
the level of the 6th subcaudal, and the lobes begin at the level of the 9th subcaudal.
Large spines occur on the asulcate side of the hemipenis to the edge of the large,
smooth, apical disk. Calyces appear to be absent and only a weak basal naked
pocket is present.

The general dorsal color is dull green, olive green, or olive brown. The venter
normally is rose or coral red, with or without lateral dark marks on the edges of
the ventrals. In life there is a reddish brown to olive brown stripe covering scales
rows 8 through 12, and frequently parts of scale rows 7 and 13. The stripe varies
in width from 5 to 10 rows of scales. Occasionally dark brown spots occur on the
posterior ede of rows 3, 4, 5, and sometimes along scales rows 7 and 8. The upper
and lower lips, throat and anterior ventrals may be cream or yellowish. Color
descriptions also are presented by Miranda et al. (1982).

Geographic variation.— 1 60 individuals were grouped into nine geographic sam-
ples. Three samples were located in the Rio Paraguay basin in a north/south line,
and six samples were scattered from Uruguay to southeastern Brazil, more or less
in a north/south line. Using univariate statistics, the number of ventrals, maxillary
teeth, subcaudals, and tail/total length ratios were analyzed for intra and inter-
sample variation. The two numbers that follow a mean value are the standard
deviation and standard error, respectively. In addition, Student’s T Test values
were used to determine significance at the 95% level, in a pair-wise sample com-
parison. No significant differences in pair-wise comparisons exist for any of the
samples arranged in the two north to south lines. There are trends in both lines
to increased number of ventrals and subcaudals from north to south. The trend
for maxillary teeth is reversed, with numbers decreasing from north to south.
Pair-wise comparisons of the Student’s T Test values for the number of ventrals,
maxillary teeth and tail/total length ratios of eastern and western samples were
not significant. However, a significant difference between eastern and western
samples in the number of subcaudals exist. Subcaudals of the combined three
western samples (N = 18) vary from 63 to 75 (x = 68.5, 3.8, 0.9), while the
combined six eastern samples (N = 1 10) vary from 52 to 71 (x = 60.4, 3.4, 0.3).

186

Annals of Carnegie Museum

vol. 56

the tail/total length ratios of the western samples vary from 0.214 to 0.268 (x =
0.231, 0.29, 0.07) and those of the eastern samples from 0.193 to 0.248 (X =
0.221, 0.13, 0.01). This suggests that the subcaudals are smaller in the western
sample, hence more subcaudals are present on a tail about the same length as in
the eastern sample. The eastern and western samples are separated by the Rio
Uruguay and Rio Parana basins and appear to be ailopatric (Fig. 3). However,
additional collecting may reveal parapatry of the samples.

Since the number of subcaudals is significantly different between the two sam-
ples, I propose recognition of the following taxa:

Liophis jaegeri coralliventris Boulenger

Diagnosis.— Dorsum ground color leaf green, sometimes with a middorsal
brownish red stripe five to six scale rows wide, extending from the nape to the
tail, and with a dark brown line on scale row three on each side of the body.
Venter rose to red, with or without ventrolateral blackish marks on the edges of
the ventrals. Subcaudals vary from 63 to 75 (x = 68.5).

Distribution.— Known only from the Rio Paraguay Basin of Argentina, Para-
guay, and Brazil (Fig. 3).

Liophis jaegeri jaegeri (Gunther)

Liophis dorsalis Peters, 1863.

Rhadinaea lineata Jensen, 1900.

Diagnosis. — Dorsum olive green, grayish green or leaf green, often with a mid-
dorsal brownish or reddish stripe six to eight scale rows wide, sometimes with
dark dots along scale row four, and occasionally along scale rows three and five.
Venter rose or reddish, with or without ventrolateral black marks on the edges
of the ventrals. Subcaudals vary from 52 to 71 (x = 60.4).

Distribution. — Known only from the east side of the Rio Parana Basin east to
the Atlantic coast, and from Uruguay north to 1 9°S latitude in Brazil (Fig. 3).

Comment. —Little is known about the natural history of this species. I have
found 7 and 9 eggs in two specimens, and Miranda et al. (1982) indicated that
the average number of eggs for this species is about 1 4. Miranda et al. also indicated
that the species is diurnal, frequently found in humid places, and eats small frogs.
Notations on some field tags suggest that the species also occurs in brushy areas,
has been found “crossing dirt roads” and “in a puddle of a cow track in low
campo, 10:30 am, 84°F, and in full sun.”

Dixon (1985) presented evidence that L. jaegeri and L. maryellenae hybridize
in an area near Belo Horizonte, Minas Gerais, Brazil. Additional material from
the latter area has not been located, and the hybrid zone remains undefined.

Specimens examined.— (L. jaegeri coralliventris). ARGENTINA, Corrientes : Gral Paz, esteros de
la Santa Lucia IML 802, Ituzaingo IML 1 180. Rosario : Rosario MNHP 1899.6-7. Santa Fe: Banados
del Rincon CM 39130, Buenos Aires MNHP 676-50, Laguan Guadalupe CM 39128. BRAZIL, Mato
Grosso : Corumba IB 26575, Rio Apa UMMZ 108761, upper Rio Parana/Rio Paraguay UNM 68831.
PARAGUAY, Asuncion BM 1930.11.27.208, Buenavista, NE of Caaguarzu MCZ 34204, Isla Con-
cepcion BM 1946.1.5.85, Primavera BM 1960,1.2.84-85, 1960.1.3.40, 1962.81.

(. Liophis jaegeri jaegeri). BRAZIL, Minas Gerais : no specific locality MCZ 20715-20716. Belo
Horizonte UMMZ 109069, Campo Alegre IB 196, Coronel Fabriciano IB 22993, Imbuzeiro USNM
100676, Itapeva IB 41 126-41127, Juiz de Fora IB 30728, 33501, Mutum IB 3802, Nova Badem IB
192, Ouro Preto IB 897, 331 1, Passa Vinte IB 4968, Pocos de Caldas IB 4714, Santa Rita de Extrema
IB 5560, 8192, Taboleiro Grande ZMUC 601253, Toledo IB 4453 1, Uba IB 17308. Parana: Araucaria
IB 7252, 7346, 13632, Balsa Nova IB 18293, 2939-22941, 22943-22944, 23112, 23114-12115,

1987

Dixon— South American Liophis

187

24343-24345, 24466, 26226, 28785, 32538, Campo Largo AMNH 102257, Carambei IB 7403, Castro
IB 6244-6247, 7931, 10392, Curitiba UMMZ 108722, IB 197, 23006, 23439-23440, 29321, Dorizon
IB 893, Guarapuava IB 24630, 24661, 24710, 24889, 26954. Jaguariaiva IB 17842, Joaquim Murtinho
IB 8263, Joao Eugenio IB 13027, 15700, Mallet IB 18775, 19544, 19670, Paulo Frontin IB 24986-
24987, 25321, Pirai do Sul IB 12425, Rio Azul IB 21346, 22801, Tronco IB 730. Rio de Janeiro :
Friburgo MZUSP 2705-2707, Muri IB 21261, 21929, 24579, Petropolis IB 21473, 21475, Rio de
Janeiro (Mono da Urea) IB 985. Rio Grande do Sul : no specific locality. BM 82.10.4.60-82.10.4.61,
61 A, MCZ 5633, 17932-17933, MZUSP 1 109, 1 1 12-1 1 13, 1 127-1128, UMMZ 6724 1 , ZMUC 60 1 254-
601255, Alexandrinha MZUSP 5472, Cacequi IB 10025, Canela IB 13526, Caxias do Sul IB 6418,
6716, 9316, Cachoeira do Sul 23315, 23317, Cruz Alta IB 18386, Estacao Ecologica do Taim MZUSP
7524, Guaiba CAS 89672, Pareci Novo IB 7645, Passo Fundo IB 8049, 8060, 8212, Pelotas IB 41057,
Porto Alegre IB 7264, 7572, 1 70 1 5, 224 1 1 , 22509-225 1 1 , 22565, 22832, 23 1 89, 234 1 9-23420, 23543,
45935, Quebra Dente-Maran IB 23302, Restinga Seca IB 7305, 9837, Sao Joao MNHP 1894.440,
Sao Leopoldo MSUSP 4100, IB 7300, 7352, 7776, 8357. Santa Catarina : Banacas IB 13319, Canoin-
has, near Taunay IB 24612, Itaiopolis IB 16669, Lagoa do Norte IB 9700, Porto Uniao IB 25992,
Sao Joaquim IB 45187, Taunay IB 16328, Tres Banas IB 7948. Sao Paulo : no specific locality MCZ
207110-20713, UMMZ 62804, 108760, Agudos IB 22097-22100, Abernessia IB 9553, Altinopolis
IB 19562, Araraquara IB 19413, Arpui IB 831, Atibaia IB 23149, 23254, 26889, 26889, Anhanquera,
KM 24 IB 23163, Baninha IB 43133, Barueri IB 6987, Bento Quirino IB 20641, Boqueirao da Praia
Grande IB 11524, Boraceia MZUSP 4498, Botucatu MZUSP 2406, Braganca Paulista IB 30498,
Caieiras IB 41061, Caixa D’Agua IB 5459, 5682, 6716, Campo Largo 4639, Campo Limpo IB 4738,
5309, 5745, 7349, 7541, 18658, Campos do Jordao IB 8450, 30692, MZUSP 1122, Capitiuva IB
3070, Cotia IB 7504-7505, 24281, 29195-29196, Embu km 28, IB 30488, 30490-30494, 305741,
Embu Guagu IB 16617, 167796, 34299, Estacao Gabriel Piza IB 5428, Guarulhos IB 5041, 23556,
23724, 42961, MZUSP 4721, Ibiuna IB 26979, 32251, 32265, 40930, 42203, 42222, 42337, 43764,
Itapetininga IB 10287-10288, Itapecerica da Sena IB 41115, 42138, 42942, 43882-43883, 44116,
44189, 44498, Itatiba IB 5551, 5586, 5741, 24185-24186, Itaquera IB 10189, 10197-10200, Itu IB
32234, Jacarei IB 3293, 24926, 28502, Jarinu IB 12810, 27542, 30886, Jundiai IB 19686, Lobo IB
3253, macedonia IB 7026, Mairipora IB 24380, 26949, 27958, Mairinque IB 6042, 6048, 6075, 6618,
Mogi das Cruzes IB 10196, 10268, 10295, 10318, 21387, 21395, 22596, Moreiras IB 17606, Ped-
erneiras IB 5470, Peruibe IB 5532, Perus IB 3292, Pindamonhangaba IB 18820, Pinheirinhas IB
32420, Pirituba IB 908, Pilar do Sul IB 32121, 32454, Quilauna IB 6257-6258, Raiz da Sena MZUSP
1 1 14-1 1 16, Ribeirao Pires IB 10232, Ribeirao Preto IB 22831, Rodovia Dos Bandeirantes dm 60,
IB 34397, Salto IB 45 1 6, 4578-4579, Sao Bento do Sapucai IB 3283, Sao Bernardo do Campo MZUSP
3289, 3682, 4463-4465, IB 22608, 42919, Sao Lourenco da Sena IB 45789, Sao Paulo MZUSP 1 1 17,
1119, 1950, 2372, 2387, 2609, 2635, 2960-2961, 3355, 3688, 4095, 4130, 4756, IB 191, 193-95,
647, 695, 1315, 1321-1322, 1416-1417, 3252, 3494, 4643, 5411, 5573, 6189, 7646, 8042, 8575,
9561, 9842, 9879, 10194-10195, 10252, 10271, 12214, 12977, 12986, 17085, 17871, 18270, 18405,
18920, 22882, 24178, 24431, 24442, 26656, 27014, 27157, 27212, 27281, 27478, 27498, 27556,
27571, 27631, 28308, 28503-28508, 29222, 29236-20237, 29249, 29614, 29924, 30683, 32446,
32504, 32800, 32803, 32925, 32999, 34277, 41129, 42109, 42195, 44162, 45746, 45751, Sao Roque
IB 5391, 44228, Sena da Bocaina MZUSP 4912, Sorocaba IB 6865, Souzas IB 23151, Sumare IB
1332, Suzano IB 20376, Tapirai IB 28060. Tremembe IB 10368, Valentim Gentil IB 28671, Vinhedo
IB 30320. URUGUAY. Canelones : no specific locality BM 84.2.23.39, FMNH 10215, Anoyo Solis
Grande CM 38959, Banados de Carrasco CM 55422-55423, Rocha USNM 89995, Ruta Interbale-
nearia CAS 93100. Montevideo : no specific locality MNHP 171, Santiago Vazquez LSUMZ 27753,
MNMHP 168.

Liophis maryellenae Dixon

Liophis maryellenae Dixon, 1985. Holotype— AMNH 62202. Type-locality— Annapolis, Goias, Brazil.

Description. — The following description is based on 10 specimens. Maximum
total length of males 435 mm, of females 530 mm. Tail/total length ratios vary
from 0.221 to 0.262 (x = 0.240). Diameter of eye/snout length ratios vary from
0.605 to 0.844 (x = 0.686). Dorsal scale rows are 19-19-17, smooth, with an
apical scale pit. Reduction to 17 rows occurs between ventrals 74 and 95 (x =
82.8). The anal plate is divided. The number of ventrals varies from 144 to 159
(x = 150.9). The number of subcaudals varies from 63 to 82 (x = 67.9). The
number of maxillary teeth varies from 25 to 28 (x = 26.1). Head scales vary as

188

Annals of Carnegie Museum

vol. 56

Fig. 4.— Distribution of Liophis atriventer and L. maryellenae in southeastern Brazil. Dots represent
L. maryellenae ; the square, L. atriventer.

follows: supralabials 8-8 (10), supralabials entering orbit 4 + 5 (10), infralabials
10-10 (8), 11-11 (2); preoculars 1-1 (10); postoculars 2-2 (10); loreals 1-1 (10);
temporals 1+2(10).

The hemipenes vary between (in situ ) 9.5 to 12 (x = 10.5) subcaudals long.
The structure is slightly bilobed, each lobe about two subcaudals in length, and
their distal ends have a smooth apical disk. The sulcus spermaticus divides about
the middle of the shaft. The entire structure is spinose, with the spine length
decreasing gradually from the base to the apex. A basal naked pocket is present.

The dorsum is usually some shade of green, grayish to olive green, with some
indication of a brown or reddish brown mid-dorsal stripe. Other dark lines may
be present laterally; some indication of dark flecks may occur over the body. In
life, the venter is yellowish orange. For more details of color, see Dixon (1985).

Distribution. — This species apparently is restricted to the tablelands of south-
eastern Brazil (Fig. 4). For specimens examined, see Dixon (1985).

1987

Dixon— South American Li o phis

189

Liophis atriventer Dixon and Thomas

Liophis atriventer Dixon and Thomas, 1985. Holotype— MZUSP 5066. Type-locality— Boraceia (Es-
tacao Biologica), Sao Paulo, Brazil.

Description. =the following description is based on three specimens. Maximum
total length of male 505 mm in snout-vent length). Tail/total length ratios vary
from 0.195 to 0.214 (x = 0.203). Dorsal scales are in 19-19-17 rows, smooth,
without apical scale pits. The reduction to 17 rows occurs between ventrals 73
and 79 (x = 76.7). The anal plate is divided. The number of ventrals varies from
141 to 148 (x = 144.0). Number of subcaudals from 49 to 56 (x = 52.5); the
number of maxillary teeth from 23 to 24 (x = 23.2). Head scales are as follows:
supralabials 8-8 (3), supralabials entering orbit 4+5 (3), infralabials 10-10 (3),
preoculars 1-1 (3), postoculars 2-2 (3), temporals 1+2 (3).

The hemipenis of the only known male is short, 5.5 subcaudals long (in situ),
and bilobed for the distal half. The sulcus spermaticus divides about one third
of the distance from the base. The distal end of the lobes contain smooth apical
disks. Five large spines are present along the outer surface of each lobe, near its
base, while the inner surface of the lobe is bare. In addition to the larger spines,
many smaller ones are scattered over the shaft.

In preservative, the dorsum is uniform olive brown. The throat and first seven
ventrals are grayish white, and the remainder become progressively darker until
the posterior two-thirds of the venter is black. The subcaudals are grayish with
occasional dark smudges.

Distribution. — This taxon is known only from the type locality (Fig. 4). See
Dixon and Thomas (1985) for specimens examined.

Key to the “Green” Liophis of South America

1. Scale rows 19-19-15 or 19-19-17 2

Scale rows 17-17-17 or 17-17-15 jaegeri

2. Scale rows 19-19-15 3

Scale rows 19-19-17 4

3. Ventrals 133-172, one scale pit typhlus

Ventrals 187-199, no scale pit guentheri

4. Ventrals 141-159, maxillary teeth 23-26 5

Ventrals 169-202, maxillary teeth 17-24 viridis

5. Subcaudals 62-82, belly yellowish orange maryellenae

Subcaudals 49-56, belly black atriventer

Acknowledgments

I thank the following curators who have indulged my requests for specimens over many years R.
L. Bezy and J. W. Wright, Los Angeles County Museum of Natural History (LACM); J. A. Campbell,
University of Texas, Arlington (UTA); C. L. S. Cordeiro, Instituto Butantan (IB); H.C. Dressauer,
(HCD); R. C. Drewes and A. E. Leviton, California Academy of Science (CAS); W. E. Duellman,
University of Kansas Museum of Natural History (KU); W. R. Heyer, R. McDiarmid and G. Zug,
United States National Museum of Natural History (USNM); M. S. Hoogmoed, Rijksmuseum van
Natuurlijke Historic, (RMNH); A. G. Kluge and R. A. Nussbaum, University of Michigan Museum
of Zoology (UMMZ); W. W. Lamar (WWL); V. Mahnert, Museum d’Histoire Naturelle, Geneve
(MNHG); E. Malnate and T. Uzzell, Academy of Natural Sciences, Philadelphia (ANSP); H. Marx,
Field Museum of Natural History, Chicago (FMNH); C. J. McCoy, Carnegie Museum (CM); C. W.
Myers and R. G. Zweifel, American Museum of Natural History (AMNH); J. B. Rasmussen, Zoologisk
Museum Universitetsparken, Kobenhavn (UZM); D. A. Rossman, Louisiana State University Museum
of Zoology (LSUMZ); R. Roux, Museum National d’Histoire Naturelle, Paris (MHNP); A. F. Stimson,
British Museum (Natural History) (BM); P. E. Vanzolini, Museo de Zoologia, Universidade Sao Paulo,
(MZUSP, DEH); S. E. Williams, Museum of Comparative Zoology (MCZ).

190

Annals of Carnegie Museum

vol. 56

I express my appreciation to my wife Mary, who has shared my pursuit of Liophis and herpetology
in general, for more than 30 years.

Literature Cited

Amaral, A. do. 1931. Studies of neotropical ophidia. XXIII. Additional notes on Colombian snakes.
Bulletin Antivenin Institute of America, 4:85-89.

. 1935a. Estudos sobre ophidios neotropicos XXXII. Apontamentos sobre a fauna da Co-
lombia. Memorias do Instituto Butantan, 9:208-216.

. 1935 b. Collecta herpetologica no centro do Brasil. Memorias do Instituto Butantan, 9:235-

246.

Boulenger, A. G. 1894. List of reptiles and batrachians collected by Dr. J. Bohls near Asuncion,
Paraguay. Annals and Magazine of Natural History, (6)13:342-348.

Cope, E. D. 1870. Seventh contribution to the herpetology of Tropical America. Proceedings Amer-
ican Philosophical Society, 1 1:147-169.

. 1887. Synopsis of the Batrachia and Reptilia obtained by H. H. Smith in the province of

Mato Grosso, Brazil. Proceedings American Philosophical Society, 24:44-60.

Cordeiro, C. L., and A. R. Hoge. 1974. Contribuicao ao conhecimento das serpentes do estado
de Pernambuco. Memorias do Instituto Butantan, 37(1973):26 1-290.

Cunha, A. R., and F. P. Nacimento. 1976. Ofidios da Amazonia. X. As cobras da ragiao leste do
Para. Publica^oes Avulsas do Museu Goeldi, 31:1-217.

Dixon, J. R. 1985. A new species of the neotropical snake genus Liophis from Brazil. Proceedings
Biological Society Washington, 98:295-302.

Dixon, J. R., and R. A. Thomas. 1985. A new species of South American water snake (genus
Liophis) from southeastern Brazil. Herpetologica, 41:259-262.

Gans, C. 1 960. Notes on a herpetological collecting trip through the southeastern lowlands of Bolivia.
Annals Carnegie Museum, 35:283-314.

Gomes, J. F. 1918. Contribuicao para o conhecimento dos ophidios do Brasil. II. Ophidios do Museu
Rocha (Ceara). Revista do Museu Paulista, 10:505-527.

Griffin, L. E. 1916. A catalog of the Ophidia from South America at present (June 1916) contained
in the Carnegie Museum, with descriptions of some new species. Memoirs of the Carnegie Museum,
7:163-228.

Gunther, A. 1858. Catalogue of the colubrine snakes in the collection of the British Museum.
London, 281 pp.

. 1 862. On new species of snakes in the collection of the British Museum. Annals and Magazine

of Natural History, 9:52-59, 124-132.

Haffer, J. 1974. Avian speciation in tropical South America. Publication of the Nuttall Ornitho-
logical Club, 14:1-390.

Hoge, A. R. 1964. Serpentes da Fundacao “Surinaam Museum.” Memorias do Instituto Butantan,
30:51-64.

Hoogmoed, M. S., and U. Gruber. 1983. Spix and Wagler type specimens of reptiles and amphibians
in the Natural Musea in Munich (Germany) and Leiden (The Netherlands). Spixiana, Suppl. 9:
319-415.

Jan, G. 1863a. Enumerazione sistematica degli ofidi appartenenti al gruppo Coronellidae. Archivio
per la Zoologia, l’Anatomia e la Fisiologia, 2:213-330.

. 1 863 b. Elenco sistematico degli ofidi descritti e disegnati per ITconografia Generale, Milano,

143 pp.

Jan, G., and F. Sordelli. 1866. Iconographie generale des ophidiens. Tome Second Volume, Livr.
18. Milan.

Jensen, A. S. 1900. Lagoa Santa Egnens Slanger. Et Bidrag til det indre Brasiliens Herpetologi.
Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening i Kjobenhavn, 1900:99-1 1 1.

Linnaeus, C. 1758. Systema Naturae per regna tria naturae secundum classes, ordines, genera,
species, cum characteribus, differentiis, synonymis, locis, Tomus I. Editio Decima, reformata,
Holmiae, Laurentii Salvii, 824 pp.

Miranda, M. E., and G. A. Couturier. 1983. Consideraciones taxonomicas en tomo a Liophis
typhlus (Linnaeus), 1758 (Ophidia: Colubridae) en la Argentina. Boletin Asociacion Herpetologica
Argentina, 1:10.

. 1984. Consideraciones taxonomicas en torno a Liophis typhlus (Linnaeus), 1758 (Serpentes:

Colubridae) en la Argentina. Historia Natural, 4:225-228.

Miranda, M. E., G. A. Couturier, and J. D. Williams. 1982. Guia de los ofidios Bonaerenses.
Asociacion Cooperadora Jardin Zoologico de La Plata, La Plata, 71 pp.

1987

Dixon— South American Liophis

191

Parker, H. W. 1928. Notes on reptiles and batrachians from Matto Grosso and E. Bolivia. Annals
and Magazine of Natural History, (10)2:96-99.

. 1935. The frogs, lizards and snakes of British Guiana. Proceedings of the Zoological Society

of London, 1935:505-530.

Peracca, M. G. 1897. Viaggio del Dott. Alfredo Borelli nel Chaco Boliviano e nella Repubblica
Argentina. Rettili ed anfibi. Bollettino Musei di Zoologia ed Anatomia Comparata, Torino, 12:
1-19.

Peters, J. A. 1960. The snakes of Ecuador. A checklist and key. Bulletin Museum of Comparative
Zoology, 122:491-541.

. 1963. Taxonomic notes on Ecuadorian snakes. Beitrage zur Neotropischen Fauna, 3:57-67.

Peters, J. A., and B. Orejas-Miranda. 1970. Catalogue of the neotropical Squamata. Part 1 . Snakes.
Bulletin, United States National Museum, 297:1-347.

Peters, W. 1863. Uber einige neue oder weniger bekannte schlangenarten des zoologischen Museums
zu Berlin. Monatsberichte der deutschen Akademie der Wissenschaften, Berlin, 1863:272-289.

Prado, A., and A. R. Hoge. 1948. Notas ofiologicas XXI. Obervagoes sobre serpentes do Peru.
Memorias do Institute Butantan, 20:283-296.

Roze, J. A. 1966. La taxonomia y zoogeografia de los ofidios en Venezuela. Universidad Central
de Venezuela, Caracas, 362 pp.

Schmidt, K. P., and R. F. Inger. 1951. Amphibians and reptiles of the Hopkins-Banner expedition
to Brazil. Fieldiana: Zoology, 31:439-465.

Vanzolini, P. E., and E. E. Williams. 1970. South American anoles: the geographic differentiation
and evolution of the Anolis chrysolepis group (Sauria, Iguanidae). Arquivos de Zoologia, Sao
Paulo, 19:1-298.

Vanzolini, P. E., A. M. M. Ramos-Costa, and L. J. Vitt. 1980. Repteis das Caatingas. Academia
Brasileira de Ciencias, Rio de Janiero, 161 pp.

Vitt, L. J. 1983. Ecology of an anuran-eating guild of terrestrial tropical snakes. Herpetologica, 39:
52-66.

Wagler, J. 1 824. Serpentum brasiliensium species novae ou histoire naturelle des especes nouvelles
de serpens, recueillies et observees pendant le voyage dans 1’interieur du Bresil dans les annees
1817, 1818, 1819, 1820, execute par ordre de sa Majeste le roi de Baviere, publiee par Jean de
Spix, . . . ecrite d’spres les notes du voyageur par Jean Wagler, Monachii. Franc. Seraph. Hubsch-
mann, 75 pp.

Werner, F. 1 899. Beschreibung einiger neuer Schlangen und Batrachier. Zoologischer Anzeiger, 22:
114-117.

. 1 925. Neue oder wenig bekannte schlangen aus dem wiener naturhistorischen staatsmuseum.

Sitzungsberichte der Akademie der Wissenschaften in Wien, 134:45-66.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 9, Pp. 193-196 28 August 1987

A NEW SPECIES OF CLINIDIUM KIRBY
(COLEOPTERA: CARABIDAE OR RHYSODIDAE) FROM
MEXICO, AND DESCRIPTIONS OF THE FEMALES OF
TWO NEOTROPICAL MEMBERS OF THE GENUS

Ross T. Bell

Research Associate, Section of Invertebrate Zoology
Joyce R. Bell1
Abstract

Clinidium (Mexiclinidium) reyesi is described from Volcan Tacana in the Sierra Madre de Chiapas
in Chiapas State, southern Mexico. It is intermediate between known species of Subgenus Mexiclin-
idium and Clinidium ( Protainoa ) extrarium, so Protainoa is suppressed and C. extrarium is included
in Subgenus Mexiclinidium. Females are also described for C. {Mexiclinidium) newtoni Bell and Bell,
from the same region, and C. {sensu stricto) mathani Grouvelle, from Brazil.

Systematics

Clinidium (Mexiclinidium) reyesi , new species

Type material — HOLOTYPE male, labelled “Volcan Tacana, Talquian, Mpio.
Union Juarez, Edo. de Chiapas, alt. 1940 m, 26-VIII-81, P. Reyes, G. Quintero,
M.L.Y.C. Castillo.” PARATYPES, 2 females, same place and date, one collected
by P. Reyes and J. Valenzuela, the other by P. Reyes and G. Quintero. (All
specimens in the Museo de Historia Natural de la Ciudad de Mexico.)

Description. —Length 7. 0-7. 4 mm. Antennae with tufts of minor setae on seg-
ments 6-10; basal setae present on segments 8-11; head longer than broad; median
lobe relatively long, ending just anterior to posterior margin of compound eye;
postorbital, suborbital tubercles absent.

Pronotum relatively short; length/greatest width 1.40; lateral margins curved;
apex strongly narrowed; base moderately narrowed; basal impression large, about
0.3 of length of pronotum, open posteriorly; pollinosity continues across anterior
margin; inner, outer marginal grooves equally deep, outer one barely visible in
dorsal view; angular seta present; precoxal seta absent (Fig. 1).

Intercalary stria ending blindly posteriorly; sutural, parasutural, intercalary, and
intratubercular striae deeply impressed, separated by narrow cariniform intervals;
female with slight concavity toward apex of intercalary stria (Fig. 2, 3); supra-
marginal stria (see Bell and Bell, 1985:6) incomplete, not impressed, represented
by row of minute punctures; marginal stria effaced in basal 0.25, remainder im-
pressed, becoming deeper toward apex; sutural stria with 1-2 setae near apex;
parasutural without setae; intercalary with 1 seta at apex; intratubercular with 2-
3 near apex; marginal with about 4 near apex; transverse sulci of abdominal sterna
narrowly interrupted at midline, medial end of each sulcus with enlarged pit;
sternum VI with 4-8 very large, irregular punctures (Fig. 4); male with indefinite
lateral pit on sternum IV, pollinosity of transverse groove extended slightly into

1 Zoology Department, University of Vermont, Burlington, Vermont 05405-0086.

Submitted 26 September 1986.

193

194

Annals of Carnegie Museum

vol. 56

Fig. 1-7 . — Clinidium reyesi new species. 1. Head and pronotum, dorsal view. 2. Elytron, left half,
dorsal view, male. 3. Elytron, right half, dorsal view, female. 4. Metastemum and abdomen, ventral
view, female. 5. Anterior leg (excluding tarsus), male. 6. Middle leg (excluding tarsus), male. 7. Hind
leg (excluding tarsus), male.

Fig. 8-10 . — Clinidium newtoni Bell and Bell, female. 8. Metastemum and abdomen, ventral view. 9.
Sternum VI, ventral view. 10. Sternum VI, lateral view.

Fig. 11-13 .—Clinidium mathani Grouvelle, female. 11. Metastemum and abdomen, ventral view.
12. Sternum VI, ventral view. 13. Sternum VI, lateral view.

1987

Bell and Bell— New Species of Clinidium

195

pit; female with lateral pit of sternum IV enlarged, transverse groove of sternum
IV with pollinosity not extending into pit; anterior trochanter of male without
tooth; anterior femur of male with large, sharp tooth; anterior tibia of male with
proximal tooth (Fig. 5); female without femoral or proximal tibial tooth; middle
calcar of male acute (Fig. 6); hind calcar falcate (Fig. 7).

This species differs from all previously described species of Mexiclinidium in
having a tuft of minor hairs on antennal segment 6. This alone will identify it.
In our last key to the subgenus (Bell and Bell, 1985), it would be hard to work
through the key, since the position of the marginal grooves of the pronotum, used
in couplet 1, is intermediate between 1 and F. The enlarged pits at the medial
ends of the transverse sulci of the abdomen are similar to those of C. guatema-
lenum Sharp; however the complete strip of pollinosity across the anterior margin
of the pronotum, the absence of precoxal setae, and the reduced marginal stria
will distinguish the new species.

Clinidium reyesi also resembles C. (. Protainoa ) extrarium Bell and Bell, 1985.
In fact, the latter appears less isolated than when we created the subgenus for it.
We hereby synonymize Protainoa and incorporate C. extrarium into Mexiclini-
dium. Couplet 2 of the key to subgenera (Bell and Bell, 1985:59) should be altered
as follows:

2(1) Parasutural stria complete anteriorly; marginal setae of pronotum absent;

angular seta present or absent Mexiclinidium Bell and Bell

2' Parasutural stria restricted to posterior 0.5 to 0.25 of elytron; marginal and
angular setae present Tainoa Bell and Bell

The description of Mexiclinidium should be altered to read “antennae with
tufts of minor setae on VII-XI or VI-XI.”

The key to species should be altered by the insertion of two couplets at the very
beginning, as follows:

0. 1 Antenna with tufts of minor setae on segments VI-XI .......... 0.2

0.T Antenna with tufts of minor setae on segments VII-XI 1

0.2(0. 1) Intercalary stria ending blindly posteriorly; pronotum with two complete
marginal grooves; basal impressions open posteriorly; hind angle of

pronotum obtuse C. reyesi new species

0.2' Intercalary stria complete; pronotum with 1 marginal groove which is
effaced posteriorly; basal impressions closed posteriorly; hind angle of
pronotum denticulate C. extrarium Bell and Bell

Clinidium ( Mexiclinidium ) newtoni Bell and Bell 1985

Description of female. — Similar to male, except that calcars are lacking, anterior
trochanter and femur not dentate; transverse sulci of abdomen very narrowly
interrupted at midline, midline smooth, shining; sulcus of sternum IV forming
large glabrous lateral pit (Fig. 8); sternum VI impressed near tip, in some specimens
forming small median tubercle (Fig. 9, 10). (In male, sulci slightly more separated,
area between finely microsculptured.)

We have studied the following additional specimens of C. newtoni : 4 males, 3
females, from Mexico: Chiapas State, Municipio Angel Albino Carzo, Camino al
Triunfo, alt. 1410-1880 m, several dates and collectors, April and August, 1983
(all in the Museo de Historia Natural de la Ciudad de Mexico). They vary in
length from 7.0-8. 5 mm.

196

Annals of Carnegie Museum

vol. 56

Clinidium ( sensu stricto) mathani Grouvelle

Description of female. —Length 7.5 mm. Similar to male in most respects, but
without calcars; small lateral pit present on sternum IV, only slightly more de-
veloped than in male; very small lateral pit on sternum III (Fig. 1 1-13).

We have studied one specimen labelled “Brasil: Amapa, Porto Platon, IX- 1957,
K. Lenko” (The Carnegie Museum of Natural History).

Acknowledgments

We thank Dr. Pedro Reyes-Castillo and Biol. Gemma Quintero G. of the Museo de Historia Natural
de la Ciudad de Mexico, and Mr. Robert L. Davidson of The Carnegie Museum of Natural History
for the loan of the specimens used in this study.

Literature Cited

Bell, R. T., and J. R. Bell. 1985. Rhysodini of the world. Part IV. Revisions of Rhyzodiastes
Fairmaire and Clinidium Kirby, with new species in other genera. (Coleoptera: Carabidae or
Rhysodidae). Quaestiones Entomologicae, 21(1): 1—172.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 10, Pp. 197-212 28 August 1987

BIOSYSTEMATIC STUDIES IN STENANTHIUM
(LILIACEAE: VERATREAE) II. FLORAL MORPHOLOGY,
FLORAL VASCULAR ANATOMY, GEOGRAPHY AND
TAXONOMY OF THE MEXICAN S. FRIGIDUM
(SCHLECHT. & CHAM.) KUNTH

Frederick H. Utech

Associate Curator, Section of Botany

Abstract

The floral morphology, continuous pedicel to stigma floral vascularization, distribution and tax-
onomy of Stenanthium frigidum, a bulbous lily confined to the Trans-Mexican Volcanic Belt, are
presented. The perfect, protandrous flowers with bilobed or v-shaped tepal nectaries are arranged in
andromonoecious panicles. Raphide ibdioblasts were observed in the upper gynoecium and recurved
stylar arms. Total floral vascularization is derived from three, compound, lower pedicel bundles. Both
outer and inner tepallary and staminal bundles as well as dorsal vasculature are formed above the
level at which the ventrals and ovular supplies are established due to an inferior gynoecium with a
central capellary hole and epigynous perianth. Within each septal arm, a pair of septal laterals fuse
with a pair of ventrals from which they had a common origin. As these septal laterals re-associate with
the ventrals, septal indentations occur. The bitegmic, camplyotropous ovules develop into flat, winged
seeds.

Introduction

Four Stenanthium species are currently recognized: S. occidentale (western North
America), S. gramineum (eastern North America), S. sachalinense (eastern Asia)
and S. frigidum (central Mexico). Of these, the tall, showy, purplish-black flowered
S. frigidum (Schlecht. & Cham.) Kunth has the most southern distribution, along
with representatives of Schoenocaulon ( =Sabadilla ), of the wide-ranging, northern
hemispheric Veratreae (Baker, 1879; Engler, 1889; Krause, 1930; Zimmerman,
1958; Melchior, 1964; Hutchinson, 1973) with nearly 80 species world-wide
(Dahlgren et al., 1985).

Though poisonous alkaloids are commonly known in the Veratreae (Kupchan
et al., 1961; Williaman and Schubert, 1961; Hegnauer, 1963), the only notes
concerning such occurrences in S. frigidum are “that natives of Mount Orizaba
know it to be dangerous to the horses that bite it” (Schiede, 1 829), “folia pecoribus
lethalia” (Bentham, 1839) and “it may be supposed to furnish a part of the
venomous sabadilla seeds of commerce, from which veratria is prepared” (Lind-
ley, 1846). Several common names have been used— “sevoeja” (Schiede, 1829;
Lindley, 1 849 a), “cebadilla de tierra fria” (Lindley, 1 849 6), “cebolleta” ( Hinton
4612 MO), “cebolleja” ( Balls 4886 UC), “flores angostas” (Conzatti, 1947) and
“cebadilla” (Sanchez, 1980). During the Mexican botanical exploration of the
middle of the last century, the same illustration was used three times by Lindley
(1846, 1849a, 18496) to promote the plant’s economic and horticultural value.
The species has been figured recently in Sanchez (1980). Chromosome counts are
unknown for the species.

This second paper in a series on the genus Stenanthium (Utech, 1987) reports
Submitted 26 March 1987.

197

198

Annals of Carnegie Museum

vol. 56

Fig. 1.— Herbarium specimen distribution of S. frigidum in the Trans-Mexican Volcanic Belt (Eje
Volcanico Transversal) (shaded area). The projection which summarizes the maximum elevation
between 18° and 22°N latitude is based in part on Clausen (1959).

on the Mexican species, S. frigidum , with emphasis on the inflorescence, flower
and the origins of the tepallary, staminal and dorsal bundles for comparisons
within the genus and the tribe Veratreae.

Materials and Methods

Floral material of Stenanthium frigidum fixed in FAA from Monte Rio Frio pass near Llano Grande,
Mexico, was made available for this study by Dr. Hugh H. litis, University of Wisconsin-Madison
(litis et al. 1085). Samples for standardized transverse (15) and longitudinal (5) paraffin sectioning
( 1 4-1 6 u) of complete flowers of varying ages were stained in safranin-methylene blue (Johansen, 1 940;
Sass, 1958). Whole flowers were also cleared and stained in a NaOH- 1% fuchsin mixture (Fuchs, 1 963)
as a check on the prepared sections. Vascular description, presentation and bundle coding follow Utech
(1987). Herbarium collections used for distribution (Fig. 1) and floral morphometries (Fig. 3) are cited
in the taxonomic review.

Observations

General morphology. —Stenanthium frigidum, a robust, glabrous perennial herb,
(4.5) 6.0-1 1.0 (15.0) dm tall, occurs in scattered clumps from large, 8.5-10.5 cm
long by 3. 5-5. 5 cm wide, bulbs which are covered by a dense, coarse brown tunica.

1987

Utech= Stenanthium frigidum

199

Fig. 2. —Flowering habit of S. frigidum. A. Bracteated panicle with flowers in bud. The progressive
reduction in the elongate bracts which subtends the lateral branches is evident. B. Panicle at anthesis
showing several nodding terminal segments in which each flower is subtended by a bract as well as
the lateral branches. Each flower within a lateral branch is subtended by a bracteole. (Photos A and
B from negatives 163 and 159 respectively, courtesy WIS herbarium and vouchered by litis et al.
902.)

The 35-60 cm long by 1 .0-1 .5 cm wide, linear, bifacial, basal leaves are sheathing
at their bases and gradually narrowed at both ends. The abaxial surfaces have
anomocytic stomatal patterns which are generally lacking on the adaxial surfaces.
The fusiform epidermal cells are six-sided. The few stem leaves are progressively
reduced apically and intergrade into bracts (Fig. 2).

The simple, andromonoecious panicle occupies the upper 35-45% of the plant’s
height. Within the panicle, the (35) 45-75 (85) bisexual flowers are distributed
over (3) 5-14 nodding lateral branches (racemes) and a dominant terminal segment
(Fig. 3). Each branch is subtended by a bract. The lower bracts are always longer
than the upper. Within each branch, each flower’s pedicel is subtended by a
scarious, linear bracteole. The lowest bracteole on a branch, like the lower bracts
on the main axis, is always larger. Each of the (10) 1 5-25 flowers on the dominant
terminal segment is subtended by a bract. Flowering begins at the base of the
inflorescence and proceeds outwards in the lateral branches and upwards in the
terminal segment. Maximum anthesis is between July and August.

Floral morphology. — The pendulous, deep fuscous brown to dark blackish ma-
roon flowers on short, glabrous pedicels have long, narrowly acuminate tepals.
The bases of the three outer tepals are wider than the inner tepals, though their
lengths, 13-16 mm, are similar (Fig. 3). Adaxial, v-shaped tepal nectaries occur
basally in both outer and inner tepals (Fig. 8C, D). At anthesis the flowers have

200

Annals of Carnegie Museum

vol. 56

Length of Freed Gynoecium cmm)

Number of Lateral Racemes from Inflorescence Base

Fig. 3. —Floral and inflorescence patterns in S. frigidum. A. Comparison between the length (L) of
freed tepals to height (H) of the flowering epigynous zone (mm). B. Comparison between the length
(L) of the freed gynoecium (mm) to height (H) of the epigynous zone (mm). C. Inflorescence com-
parisons showing the total number of flowers and the degree of branching within the panicle. The
connecting line indicates the cumulative flower number from one branched increment to the next with
the terminal segment marked with an open circle.

a campanulate shape due to tepal spreading from the epigynous zone, but the tips
are never reflexed. The six, subulate, glabrous filaments are 3. 3-3. 6 mm long,
with little length differences between the outer and inner sets. The basifixed,
extrorse anthers have a valvular dehiscence between the confluent thecae that
opens into a reniform, peltate disc. Protandry occurs in S. frigidum, since both
outer and inner anthers dehisce at the same time, and this is earlier than gynoecial
maturity. The free styles and upper areas of the tri-carpellate gynoecium surpass
the anther zone at maturity. The pedicels of the fruiting septicidal capsules are
erect and elongated compared to anthesis. The ovoid capsules with persistent,
withered tepals are 13.5-16.5 cm long and contain numerous winged seeds.

Floral vascular anatomy.— Pedicel vascularization: In transverse section, the
flowering pedicels above their subtending bracts (bracteoles) are circular and have
three large, centrally arranged bundles. Frequently several phloem caps per bundle
are observed, attesting to the compound nature of these bundles. It is from these
three bundles that the complete floral vasculature is established. The centers of
these bundles are located on radii 120° apart and are designated the outer tepal
radii.

Near the pedicel’s mid-length, each of the three outer tepal radii bundles under-

1987

Utech— Stenanthium frigidum

201

goes a tri-partite radial subdivision, which results in three bundles. These divisions
are not co-planar, but occur at different levels in a spiral pattern. Gaps along the
outer tepal radii are created by the outward departure of the central bundle prod-
ucts from each of the three, tri-partite divisions (Fig. 4A). These departing bundles
with normally arranged xylem (adaxial) and phloem (abaxial) remain free of other
vasculature and establish the outer tepal medians (OTM). At this level, the pedicels
have a triangular shape in transverse section with the comers on the outer tepal
radii (Fig. 4B).

Immediately following the OTM bundle formation the two adjacent parental
bundles undergo additional radial divisions. The two products closest to the gap
depart outward with the OTM. There are three such sets of three bundles which
correspond to the dorsal-composite bundle of Sterling (1982), and eventually
establish the outer tepallary laterals (OTL), staminal (OS) and dorsal (D) vascu-
lature (Fig. 4B-D).

The two remaining lateral products adjoining a gap following the radial divisions
fuse with similar adjacent laterals along the inner tepal radii, which are located
halfway between the outer tepal radii (Fig. 4B-D). These fusion bundles are formed
at progressively higher levels in a spiral pattern. Each of these three, fusion bundles
undergoes a tri-partite subdivision and radial division similar to that observed
at a lower level among the three original bundles. The central bundle products of
this second series of divisions depart along the inner tepal radii and become the
inner tepal medians (ITM). The ITM and the two radial division products which
depart with the ITM correspond to the “zwischenbimdel” of Sterling (1982) and
eventually establish the inner tepallary laterals (ITL) and staminal (IS) vasculature
(Fig. 4D, E).

Six sets of three bundles each remain in peripheral positions until the upper
epigynous zone. Tepallary, staminal and dorsal vasculature formation will be
discussed later in a separate section.

Ventral vascularization: Following the formation and departure of the inner
tepal medians (ITM) and associated lateral pairs, three bundle pairs remain along
the inner tepal radii in the central area. It is from these bundle pairs that the
complete ventral supply is established (Fig. 4C-F, 6A, B, 8A, B).

An inward rotation towards the inner tepal radii among the central bundle pairs
occurs as well as a radial subdivision among these remaining bundles. As the
locules open, the six inner most division products with reversed conducting ele-
ments depart inwards as a pair to establish the ventrals (V) (Fig. 4D, E, 5A, B).
In the epigynous zone, two such ventrals (V) are associated with each locule. The
remaining parental products may fuse along the inner tepal radii in a septal axial
position. These bundles also divide if they fused and their resulting products have
conducting elements which are reversed. There are three pairs of such bundles
that could be called septal laterals (L) (Fig. 4E, F, 5A-C). They move inward
following the same course as the ventrals. A given septal lateral (L) is re-associated
with the same ventral (V) of common origin in the inner septal margins. Ovule
supply occurs initially from the ventral (V) bundles in the lower epigynous zone,
but in the upper freed gynoecium, the ventrals and laterals which fused in the
ventral position continue to supply ovules. These continuing ventrals pass into
the stylar region without fusion to other ventrals or dorsals. Four rows of 16-24,
bitegmic, campylotropous ovules are associated with each carpel. Ovule supply
occurs in the lower epigynous zone below the level of dorsal formation. Raphide
idioblasts are common in the upper gynoecium and the free spreading stylar arms.

As the locules open perpendicular to the outer tepal radii in a spiral pattern,

202

Annals of Carnegie Museum

vol. 56

Fig. 4. —Photomicrographs of transverse sections from pedicel to mid-epigynous zone in S.frigidum.
A. Lower pedicel (circular outline) and subtending bract (br). Three outer tepal medians (OTM) are
established at this level. B. Upper pedicel (triangular outline) showing laterals pairs associated with
each OTM and the formation of the ITM bundles and their associated lateral pairs. C. Lower epigynous
zone showing one open locule and further formation of the ITM bundles, their lateral pairs and
ventrals. D. Epigynous gynoecium with three open locules with two ventrals (V) each and the formation
of septal laterals. Dorsals are not present at this level. E. Gynoecium with central carpellary hole,
inner septal margins subdivided along the inner tepal radii, ovule placentation, septal lateral (L) pairs
in the septal arms and dorsal (D) bundles. Outer stamen bundles are not present at this level. F. Re-
association of ventrals (V) and laterals (L) in the septal margins. Outer stamen (OS) bundles present.

1987

Utech— Stenanthium frigidum

203

Fig. 5. — Matched ventral (A-F) and dorsal (A'-F) projections of the same section (central column)
showing the ventral vascular supply in S. frigidum. The ventral projection is along the IS radii, the
dorsal along the OS radii. Views A-E are through the epigynous zone, while F is from the freed
carpellary zone. View B shows the central carpellary hole, C the subdivided inner septal margins and
inter-locular connections, and D the septal indentations within the septal arms.

three septal arms are established within the lower epigynous zone. A central
carpellary hole or opening (Fig. 4D-F, 5B), which is continuous with the open
stylar canal, appears in the central area before this region is subdivided along the
outer tepal radii. This central subdivision creates three inner septal margins or

204

Annals of Carnegie Museum

vol. 56

Fig. 6. —Transverse sections through the outer epigynous zone of S. frigidum showing the origin of
outer tepal, stamen and dorsal vasculature (dorsal-composite bundle). A. Section showing an outer
tepal median (OTM) in a distal position and two adjoining, outer tepal laterals (OTL) with laterally
orientated phloem caps (arrows). B. Section above A showing xylary fusion and plexus formation
between two laterals (OTL). C. Section above B showing a free dorsal with normally arranged con-
ducting elements which was derived from the fusion plexus and the formation of an outer stamen
(OS) bundle (arrow). D. Section above C showing an OS bundle (arrow) with reversed conducting
elements. E. Section above D showing the outward departure and division of the outer tepal laterals
(OTL) and an OS bundle with two lateral phloem caps (arrows). The OTM and D bundles are correctly

1987

Utech— Stenanthium frigidum

205

wing tips. As the ventral and lateral bundles re-associate in the inner septal
margins, septal indentations or slits open within the septal arms along the inner
tepal radii. These indentations occur within the upper epigynous zone, and fur-
thermore, are continuous with the zones which separate the gynoecium from the
surrounding stamens and tepals (Fig. 5D-T\ 8A-C). The three free carpels are
only weakly appressed along the central floral axis.

Tepallary, stamina! and dorsal vascularization: Established in the lower pedicel,
the three OTM bundles and their pair of associated laterals plus the three ITM
bundles and their lateral pairs have remained in a peripheral position through
the epigynous zone. The OTM and ITM medians which are along the outer and
inner tepal radii, respectively, are not involved in any further division or fusion.
Fig. 6 details the origin of the outer tepal laterals (OTL), outer stamen (OS) and
dorsal (D) bundles, while Fig. 7 presents the similar, but less complex pattern for
the inner tepal laterals (ITL) and inner stamen (IS) bundles.

From the three bundles in an outer group, the two laterals or the pair associated
with an OTM undergo an inward rotation and fusion along the outer tepal radii.
From this outer vascular plexus and subsequent subdivision, a dorsal (D) with
normally arranged conducting elements results. This is followed by an outer sta-
men (OS) bundle with reversed conducting elements that shifts to a normal ar-
rangement in the upper epigynous zone. It should also be noted that the dorsals
are formed before the outer stamen bundles. The dorsals follow an undivided and
unfused course into the upper stylar arms.

The remaining parental laterals depart as the outer tepal laterals (OTL) flanking
the OTM. The laterals undergo additional radial divisions such that when the
tepal is freed from the epigynous zone, each outer tepal has as many as seven
bundles, that is, three OTL + one OTM + three OTL. Additional divisions occur
in the freed outer tepal, such that a maximum of 1 1 bundles can be observed—
five OTL + one OTM + five OTL. All of these bundles were established from
three at a lower level. There is no terminal fusion between any of the laterals or
medians and all end in the upper tepal margins.

The formation of inner tepal lateral (ITL) and stamen (IS) bundles parallels
that for the OTL and OS bundles, except no dorsal is associated with the inner
members. At the base of each freed inner tepal, a seven bundled condition exists,
that is a median (ITM) and two sets of three laterals (ITL). The freed inner tepals,
like the outer tepals, can have a maximum of 1 1 bundles, that is five ITL + one
ITM + five ITL, which were derived from three lower bundles.

Each of the six stamens receives a single bundle or staminal trace. Epitepally
occurs in both the outer and inner stamens.

Geography and taxonomy. —Stenanthium frigidum occurs in a most specialized
ecological zone— the Trans-Mexican Volcanic Belt (Clausen, 1959) or “Eje Vol-
canico Transversal” (Rzedowski, 1978) (Fig. 1). Scattered clumps are frequently
encountered in open, rocky, grassy, alpine meadows associated with Pinus hart-
wegii forests (Beaman, 1962, 1965) between 18° to 22°N latitude at an elevation
of 9000-12,000 ft (ca. 2700-3700 m). Stenanthium frigidum was first described
by Schlechtendal and Chamisso (1831) as Veratrum frigidum from a C. J. W.

positioned. F. Section above E showing a normal element arrangement for the OS bundle (arrow) and
further outward departure of the laterals (OTL).

206

Annals of Carnegie Museum

vol. 56

Fig. 7.— Transverse sections through the outer epigynous zone of S. frigidum showing the origin of
inner tepal and stamen vasculature (zwischenbiindel). A. Section showing an inner tepal median (ITM)
in a distal position and two inner tepal laterals (ITL) with laterally oriented phloem caps (arrows).
Fusion has occurred between the xylar elements. B. Section above A showing laterals (ITL) departing
and an inner stamen (IS) bundle with reversed conducting elements (arrow). C. Section above B
showing an inner stamen (IS) bundle with two lateral phloem caps (arrows). D. Section above C
showing normal element arrangement of the IS bundle and division of both inner tepal laterals (ITL),
and the carpellary wall separated from the epitepalous stamen (dotted lines).

Schiede and F. Deppe collection from the alpine regions of Mt. Orizaba (Fig. 9).
Schiede and Deppe left Bordeaux in May 1828, arrived at Veracruz in July 1828
and began collecting in the Jalapa and Orizaba areas (Schiede, 1829). Their col-
lections, including type material, are known from the following herbaria (Holm-
gren et ah, 1981): “B (original set destroyed; set ex herbarium Baschant extant),

1987

Utech — Stenanthium frigidum

207

Fig. 8. — Photomicrographs of transverse sections from the upper epigynous zone to the upper floral
zone in S. frigidum. A. Upper epigynous zone showing the septal indentations along the inner tepal
radii within the septal arms. Tepallary and staminal bundles established. B. Separation of gynoecium
from surrounding stamens and tepals. Epitepally among both the outer and inner stamens and the
formation of additional tepal laterals (OTL and ITL) evident. C. Adaxial surface of freed tepals
nectiferous (arrows). Outer (OS) and inner (IS) filaments with a single bundle evident. D. Adaxial
tepal nectaries (arrows) at the same level as the reniform, peltate anthers. The adaxial nectiferous
surface at higher levels occurs marginally, thus creating a basal V-shaped nectary.

HAL (sent to Schlechtendal); further material BM, BR, C, CGE, F, FI, G, GH,
GOET, K, KIEL, LE, LZ (destroyed), M, MO, OXF, P, S, U, US, W, XAL
(photos)” (Stafleu and Cowan, 1985). Type material exists for S. frigidum (Fig.
9) and it is very probable that additional material is available for lectotypification.
Subsequent taxonomic reviews have placed this species in Zigadenus (Don, 1832),
Asagraea (Lyons, 1 907) and Stenanthella (Gates, 1918). The transfer to Stenanthi-
um was made in 1842 by Kunth.

Stenanthium frigidum (Schlechtendal & Chamisso) Kunth

Stenanthium frigidum (Schlechtendal & Chamisso) Kunth, Enum. Plant., 4:190-191. 1842.

Type. — Mexico, Veracruz, Mt. Orizaba, alpine region, 1831, Schiede & Deppe 983 (MO!; fragment
GH!; Photos CM) -Fig. 9.

Veratrum frigidum Schlecht. & Cham., Linnaea, 6:46. 1831.

Zigadenus frigidus (Schlecht. & Cham.) D. Don, Edinb. New Phil. J., 233. 1832.

Asagraea frigida (Schlecht. & Cham.) Lyons, PI. Names Scientif. Pop., 2nd ed., 508. 1907.
Stenanthella frigida (Schlecht. & Cham.) Gates, J. Linn. Soc., London, Bot., 44:152. 1918.

Representative specimens examined.— MEXICO: DISTRITO FEDERAL: Cima station, hills, 9800
ft, 30 AUG 1905, Pringle 13,622 (CAS, GH, LL, TEX, US); Ajusco, Morelos, 3000-3700 m, SEP

208

Annals of Carnegie Museum

vol. 56

D

jUr- * r£*p o ti&_ u£yti bi $ j, ks
0 * i zcJbeue. t uMt , t oueyy,

LAsy>Z<JUo W C,^ie_

4»7Wi* ie^uu^-
Scfc'eik + Depp? , Ht*ico.

^o,-bl7i

MISSOURI
BOTANICAL
- HERBARIUM

University of Wisconsin Herbarium, Madison.

(C t

STENANTHELLA F-IGIUA Gates
Dot. James H. Zimmerman 1956 - 60

Missouri Botanical Garden.

U ZJiaJZsvei (Li q%<{,ivni S.ZJldr +■

CUi. VL, %%we*^lt-i94,-03l

: //U™ hg&zu

{pdz A 'zoo / >^lsy.

Fig. 9.— Type of S. frigidum from Mt. Orizaba, Mexico, collected in 1828 by Schiede and Deppe
(983). The MO specimen is from the Bemhardi herbarium. (Courtesy of Missouri Botanical Garden.)

1987

UtECH— StENANTHIUM FRIG1DUM

209

1925, Lyonnet 22 (GH, MO, NY, US); Mt. Gauzin (Guauhztin), between rocks in alpine meadow,
10,900 ft, 26 AUG 1928, Antipovitch 96 (CM); La Portrera, 9000 ft, 22 AUG-19 SEP 1930, Russell
& Souviron 150 (US); S slope of Serjana de Ajusco, ca lA km E of Zampoalo on road to Huitzilac,
3000 m, 1 5 JUL 1960, litis, Koeppen & litis 253 (WIS); 2 km N of La Cima, SE slope of Cerro Pelado,
1-2 km N and W of La Cima station, 3050-3300 m, 14 AUG 1960, litis, Koeppen, litis & Rzedowski
902 (UC, WIS— 8 sheets); Cerro de San Miquel, near summit in grassy area, 31 JUL 1960, Penalosa
813 (CAS); Cerro Pelado, Serrania del Ajusca, 3100 m, 14 AUG 1960, Rzedowski 12,630 (MSC).
MEXICO: Zinantecatl Mts., moister places of pine forests of Nevado de Toluca, 12,000 ft, 2 SEP
1892, Pringle 4257 (GH, MO, NY, UC, US); Ixtaccihuatl, meadows near timber line, NOV 1905,
Purpus 1699 (GH, NY, UC, US); Temascaltepec, Cajones, 2480 m, 1 1 NOV 1932, Hinton 2396 (GH),
Crucero Aqua Blanca, 3170 m, 29 AUG 1933, Hinton 4612 (CAS, MO); Cierrita, 13 AUG 1935,
Hinton 8353 (GH, NY, UC, US); pine forest, 18 DEC 1935, Hinton 8792 (GH, LL, NY, UC, US);
Apitza, Mt. Ixtaccihuatl, 12,000 ft, said by natives to be poison to animals, 27 JUL 1938, Balls 5136
(UC, US); Valle de Mexico, C. de Llano Grande, 2800 m, 30 JUL 1950, Matuda 19,002 (UC);
Ixtaccihuatl, 3500 m, 7 JUN 1953, Matuda 28,389 (UNC); Telapon, N of Ixtaccihuatl, 3450-3500
m, grassy meadow under Pinus hartwegii forest, 4 SEP 1958, Beaman 2443 (GH); Ascenso al Nevado
de Toluca, en bosque de P. hartwegii, 3789 m, 5 SEP 1958, Hernandez s.n. (CAS); S of Pan-Arn
highway at pass of Monte Rio Frio (5 km WNW of Rio Frio) at Llano Grande, 3100-3200 m, 18
AUG 1960, litis, Koeppen & litis 1085 (UC, WIS -4 sheets); 3100-3200 m, 24 SEP 1962, Ugent
2299 (WIS— 14 sheets); Papayotl (El Papayo), 2 OCT 1966, Boege 292 (CAS); Ixtapaluca, Estacion
Experimental de Investigacion y Ensenanza de Zoquiapan, 8 km S de Rio Frio, 3190 m, 5 SEP 1975,
Koch & Magana 75,496 (CAS, F, NY); sotobosque de Pinus hartwegii, 3300 m, JUL 1976, Obieta 75
(CAS); Ixtapaluca, Llano de San Miquel, 3350 m, 26 AUG 1978, Garcia s.n. (TEX). MICHOACAN :
Mt. Tancitaro, 10,000 ft, rocky ridge in open pine forest, 4th Hoogstraal Biological Expedition, 22
JUL 1941, Leavenworth & Hoogstraal 1135 (GH, MO, NY); Zitacuaro-Cerro Pelon, 3600 m, 14 AUG
1938, Hinton 13,238 (GH, LL, NY, UC, US); Vicinity of Morelia, Cerros San Miquel, 2200 m, SEP
1910, Arsene 5249 (GH, MO, NY, US); Uruapan, Tancitaro, 3450 m, llano in pine forest, flower and
root smell like fungus, said to poison cattle, 23 OCT 1940, Hinton 15,572 (GH, LL, MO, NY, TEX).
PUEBLA : Lower slopes of Pico de Orizaba, W of Chalachicomula (Ciudad Serdan), 21 SEP 1957,
Beaman 1788 (GH). TLAXLALA : Mt. Malinche, pine woods, 12,000 ft, 22 JUN 1938, Balls 4886
(UC, US); N side Mt. Malinche, in open grassy Pinus hartwegii forest, 3770 m, 10 AUG 1958, Beaman
2244 (GH). VERACRUZ : Mt. Orizaba, 12,000 ft, foliage poisonous to cattle and horses, 6 AUG 1891,
Seaton 203 (GH, NY); Vaquena del Jacal, 10,000 ft. 1841-43, Liebmann 14,681 (US); Mt. Orizaba,
10,000 ft, 1841-43, Liebmann 14,682 (GH, UC), 14,683 (GH, MO); 25-26 JUL 1901, Rose & Hay
5748 (US); Vertiente del Volcan Pico de Orizaba, 3600 m, 27 JUL 1971, Nevling & Gomez-Pompa
2071 (GH, NY).

Discussion

The tall and robust Stenanthium frigidum occupies a most unusual geographic
and ecological zone, in that it occurs between 9000-1 2,000 ft in the Pinus hartwegii
forest of the Trans-Mexican Volcanic Belt and is the southern most and highest
occurring member of the tribe Veratreae. The highest elevations attained by S.
occidentale are near 7500 ft in the Pacific Northwest (Utech, 1987). Numerous
other differences also occur between S. frigidum and S. occidentale.

The branched andromonoecious, bracteated panicle of S. frigidum has numer-
ous bisexual, protandrous flowers, in marked contrast to the simple raceme of S.
occidentale (Utech, 1987) and the highly branched, bracteated panicle of S. gra-
mineum. Bracteoles, though small, are present in S. gramineum, although Gates
(1918) reported their absence. An inflorescence parallel in both S. gramineum
and S. frigidum needs further investigation. Both species have extremes in panicle
patterns: one type is lax, open and the other is tight, more compact. The flowers
of S. occidentale (Utech, 1987) and S. frigidum have a characteristic nodding
appearance, while in fruit their capsules are erect. A sclerenchymatous sheath
surrounds the fruiting pedicel bundles in S. occidentale (Utech, 1987), and this
also appears true of the erect pedicels in S. frigidum based on dried fruiting
material.

210

Annals of Carnegie Museum

vol. 56

The dark maroon to purplish-black tepals of S. frigidum are not suffused with
greenish yellow as in S. occidentale, which gives the flowers of the latter a bronze
cast. The freed tepals in S. frigidum average 14.5 mm in length and appear longer
than those of S. occidentale (Utech, 1987). This difference is only apparent, since
the tepals in S. frigidum are not reflexed and those in S. occidentale are. Epitepally
and similar basal, v-shaped, adaxial, tepal nectaries are found in both species.

Both species have epigynous gynoecium, but their flowering and fruiting pro-
portions differ. In S. frigidum, there is little difference in epigynous zone height,
which averages 4.3 mm in flower and 4.6 mm in fruit. This zone nearly doubles
in S. occidentale from a flowering height of 2.3 mm to 4.0 mm in fruit (Utech,
1987). The mature septicidal capsules also differ in shape. In S. occidentale, they
are rather narrowly elongated to a tapering apex of ascending stylar arms, whereas
in S. frigidum they are more ovoid apically with recurved stylar arms. In fruit,
the epigynous zone of S. frigidum accounts for 28.5% of the capsule’s length,
while in S. occidentale, the same zone averages 16.5%.

The general floral and carpel morphology as well as the basic floral vascular
anatomy of S. frigidum are similar to that reported for various species within the
tribe Veratreae (Anderson, 1940; El-Hamidi, 1952; Ambrose, 1975, 1980; Ster-
ling, 1982; Utech, 1986, 1987). However, there are some specific differences
between S. frigidum and S. occidentale (Utech, 1987). The septal indentations
within the septal arms of the upper epigynous zone of S. frigidum were not
observed in S. occidentale, though both species have a central carpellary hole and
spirally inserted floral parts. The equal filaments in S. frigidum are considerably
shorter than either of the dimorphic ones in S. occidentale. Raphide idioblasts
were commonly seen in the upper gynoecium of S. frigidum, but are lacking in
S. occidentale (Sterling, 1982; Utech, 1987). Such raphide distribution in S. fri-
gidum may be a possible defense against chewing insects. Both species have
bitegmic, camplyotropous ovules and flat, oblong, winged seeds.

The major differences in floral vascular patterns of S. frigidum and S. occidentale
involve tepal vascularization that arises from different kinds of compound bundles
and the fate of septal-ventral bundles. In the lower areas of the short flowering
pedicels of S. frigidum, the outer tepal medians (OTM) are established first,
followed by a pair of associated laterals from which the outer tepallary, staminal
and dorsal bundles are derived. On the other hand, in S. occidentale, which has
significantly longer flowering pedicels, a compound outer tepal (OT; Utech, 1987)
bundle is established first, but in the same position as the OTM bundle in S.
frigidum. The compound OT bundles in S. occidentale divide in the mid-pedicel
to form the outer tepallary, staminal and dorsal bundles. A similar pattern is
observed in the direct formation of the inner tepal medians (ITM) in S. frigidum
versus the formation of a compound inner tepal (IT; Utech, 1987) bundle in S.
occidentale which later divides to form an inner tepal median (ITM), tepallary
laterals (ITL) and staminal (IS) bundles.

In the upper epigynous zone of S. frigidum, the outer lateral pairs surrounding
each outer tepal median (OTM) rotate inwards and fuse along the OT radii. From
this fusion plexus, a dorsal is derived first, followed by an outer stamen (OS)
bundle which has reversed conducting elements. At their filament bases, the OS
bundles have normally arranged conducting bundles. Except for dorsal formation,
a similar set of events is associated with the formation of the inner stamen (IS)
bundles, including their reversal. This pattern corresponds to that reported for
the Veratreae in general (Sterling, 1982), Amianthium muscaetoxicum (Utech,

1987

Utech — Stenanthium frigidum

211

1986) and Stenanthium occidental (Utech, 1987). The OTM and its lateral pair
or the compound OT correspond to the dorsal-composite bundle reported within
the Veratreae (Sterling, 1982). The ITM and its lateral pair or the compound IT,
similarly, correspond to the “zwischenbimdel” (Sterling, 1982).

The total outer and inner tepal vascularization in both S. frigidum and S.
occidentale are derived from three bundles, a median and two tepal laterals, which
were established in the upper epigynous zone. Medians do not divide or fuse, but
follow a direct course to the tepal tips. The laterals, on the other hand, undergo
several divisions to innervate the tepal surface. Characteristically, the tepals of
S. frigidum have a maximum of 1 1 bundles; those of S. occidentale have seven.
Though withered, the tepals are persistent in both species.

There is also a significant difference in ventral vascularization between S. fri-
gidum and S. occidentale. In the latter, following the formation of the ventrals, a
fusion septal axial (SA; Utech, 1987) with reversed conducting elements is formed
and ends abruptly without fusion or further division in the lower septal arms. In
S. frigidum, a similar bundle with reversed elements occurs, but it divides radially
forming septal laterals which follow a course similar to that of the ventral with
which they fuse in the upper freed gynoecium. Below this level of fusion, the
ventrals supply the ovules directly, while above, the fusion products in the ventral
position do. The formation of septal laterals in S. frigidum which later fuse with
ventrals of a common origin occurs in the septal arms at the same time as the
septal indentations appear.

Conclusions

The Mexican species Stenanthium frigidum has more outer and inner tepal
bundles than S. occidentale, equal, not dimorphic filaments, septal gynoecial
indentations, different patterns of ventral vascularization and raphide idioblasts
in the upper carpellary zones. Both S. frigidum and S. occidentale have v-shaped
tepal nectaries, while no nectaries are reported for S. gramineum. The chromo-
some number and karyotype of S. frigidum are unknown; neither is the relationship
to S. occidentale with 2n = 16 and S. gramineum with In = 20.

Acknowledgments

The author gratefully acknowledges the loan and use of specimens from the following herbaria (CAS,
CM, F, GH, LL, MO, MSC, NY, TEX, UC, UNC, US, and WIS) whose acronyms are cited in
Holmgren et al., 1981. Without the kind gift of fixed floral material and use of photographic negatives
by Dr. Hugh H. litis, long a student of the Veratreae and to whom the author is indebted, this study
would have been impossible. Mr. William W. Brown deserves special thanks for his artistic aid in
figure production.

Literature Cited

Ambrose, J. D. 1975. Comparative anatomy and morphology of the Melanthioideae (Liliaceae).
Unpublished Ph.D. dissertation, Cornell University, Ithaca, vii + 240 pp.

. 1980. A re-evaluation of the Melanthioideae (Liliaceae) using numerical analyses. Pp. 65-

81, in Petaloid monocotyledons (C. D. Brickell, D. F. Cutler, and M. Gregory, eds.), Academic
Press, London, ix + 222 pp.

Anderson, C. E. 1940. Some studies on the floral anatomy of the Liliales. Unpublished Ph.D.

dissertation, Cornell University, Ithaca, iv + 142 pp.

Baker, J. G. 1879. Synopsis of the Colchicaceae and aberrant tribes of the Liliaceae. Journal of the
Linnean Society, Botany, 17:405-413.

Beaman, J. H. 1962. The timberline of Izatacchuatl and Popocatepetl, Mexico. Ecology, 43: 377-
385.

212

Annals of Carnegie Museum

vol. 56

. 1965. A preliminary ecological study of the alpine flora of Popocatepetl and Izatacchuatl.

Sociedad Botanica de Mexico, Boletin, 29:63-65.

Bentham, G. 1839. Plantae Hartwegianae. 402. Veratrum frigidum. Gulielmus Pamplin, London,
p. 53.

Clausen, R. T. 1959. Sedum of the Trans-Mexican Volvanic Belt. An exposition of taxonomic
methods. Cornell University Press, Ithaca, x + 380 pp.

Conzatti, C. 1947. Flora Taxonomica Mexicana (Plantas Vasculares). Sociedad Mexicana de His-
toria Natural, Mexico, D.F., 11:62.

Dahlgren, R., H. T. Clifford, and P. F. Yeo. 1985. The families of monocotyledons. Structure,
evolution and taxonomy. Springer- Verlag, New York, xii + 520 pp.

Don, D. 1832. On the characters and affinities of certain genera, chiefly belonging to the Flora
Peruviana. Edinburgh New Philosophical Journal, p. 233.

El-Hamidi, A. 1952. Vergleichend-morphologische Untersuchungen am Gynoeceum der Unterfami-
lien Melanthioideae und Asphodelioideae der Liliaceae. Arbeiten aus dem Institut fur allgemeine
Botanik, Universitat Zurich, ser. A, 4:1-50.

Engler, A. 1889. Liliaceae. Pp. 10-30, in Die natiirlichen Pflanzenfamilien (A. Engler, and K.
Prantl, eds.), Engelmann Verlag, Leipzig, 2:10-91.

Fuchs, C. 1963. Fuchsin staining with NaOH clearing for lignified elements of whole plants or plant
organs. Stain Technology, 38:141-144.

Gates, R. R. 1918. A systematic study of the North American Melanthaceae from a genetic stand-
point. Journal of the Linnean Society, Botany, 44:131-172.

Hegnauer, R. 1963. Chemotaxonomie der Pflanzen. Birkhauser Verlag, Basel, 2:269-359.

Holmgren, P. K., W. Keuken, and E. K. Schofield. 1981. Index herbariorum. Part I: The herbaria
of the world. Ed. 7. Regnum Vegetabile, 106:1-452.

Hutchinson, J. 1973. The families of flowering plants. Vol. II: Monocotyledons. Clarendon Press,
Oxford, Ed. 3, 268 pp.

Johansen, D. A. 1940. Plant microtechnique. McGraw-Hill Book Company, New York, xi + 523 pp.

Krause, K. 1930. Liliaceae. Pp. 227-276, in Die natiirlichen Pflanzenfamilien. (A. Engler and K.
Prantl, eds.), Engelmann Verlag, Leipzig, 2(15a):227-390

Kunth, C. S. 1842. Melanthaceae. Enumeratio Plantarum. Hinrichs, Tubingen, 4:189-191.

Kupchan, S. M., J. H. Zimmerman, and A. Alfonso. 1961. The alkaloids and taxonomy of Veratrum
and related genera. Lloydia, 24:1-26.

Lindley, J. 1846. 50: Stenanthium frigidum. Journal Horticultural Society, London. 1:302.

. 1849 a. Stenanthium. Medical and oeconomical botany. Bradbury and Evans, London, 55-

56 pp., Fig. 91.

. 1849 b. 214: Stenanthium frigidum Kunth. Flore des Serres et des Jardins de 1’Europe, ser.

1, 5:460-468.

Lyons, A. B. 1907. Plant names scientific and popular, 2nd ed., Nelson, Baker and Company,
Detroit, p. 508.

Melchior, H. 1964. Syllabus der Pflanzenfamilien. Vol. II: Angiospermen. Gebruder Bomtraeger,
Berlin, vi + 666 pp.

Rzedowski, J. 1978. Vegetacion de Mexico. Editorial Limusa, Mexico, D.F., 432 pp.

Sanchez, O. 1980. La flora del Valle de Mexico. 6 ed. Editorial Herrero, Mexico, D.F., p. 97, fig. 51.

Sass, J. E. 1958. Botanical microtechnique. Iowa State University Press, Ames, xi + 228 pp.

Schiede, C. J. W. 1829. Botanische Berichte aus Mexico. Linnaea, 4(2):205-236.

Schlechtendal, D. F. L. von, and A. von Chamisso. 1831. 983. Veratrum frigidum n. sp. Linnaea,
6:46-49.

Stafleu, F. A., and R. S. Cowan. 1985. Taxonomic literature. Volume V: Sal-Ste. 2nd Edition.
Regnum Vegetabile, 112:146-147.

Sterling, C. 1982. Comparative morphology of the carpel in the Liliaceae: Veratreae. Journal of
the Linnean Society, Botany, 84:57-77.

Utech, F. H. 1986. Floral morphology and vascular anatomy of Amianthium muscaetoxicum (Wal-
ter) A. Gray (Liliaceae- Veratreae) with notes on distribution and taxonomy. Annals of Carnegie
Museum, 55:481-504.

. 1 987. Biosystematic studies in Stenanthium (Liliaceae- Veratreae) I. Floral morphology, floral

vascular anatomy, geography and taxonomy of S. occidentale A. Gray. Annals of Carnegie Mu-
seum, 56:1 13-135.

Williaman, J. J., and B. G. Schubert. 1961. Alkaloid-bearing plants and their contained alkaloids.
United States Department of Agriculture Technical Bulletin, no. 1234.

Zimmerman, J. H. 1958. A monograph of Veratrum. Unpublished Ph.D. dissert., University of
Wisconsin, Madison, iii + 321 pp.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 11, Pp. 213-222 28 August 1987

SYSTEMATICS OF AFRICAN BATS OF THE GENUS
EPTESICUS (MAMMALIA: VESPERTILIONIDAE).
2. KAROTYPES OF AFRICAN SPECIES AND THEIR
GENERIC RELATIONSHIPS

Karen McBee1

Rea Postdoctoral Fellow, Section of Mammals

Duane A. Schlitter

Curator, Section of Mammals

R. Laurie Robbins2
Abstract

Bats of the genus Eptesicus are characterized by extreme consistency in karyotypes. All species of
Eptesicus for which karyotypes have been reported have 2n = 50, FN = 48 (Baker and Patton, 1967;
Williams, 1 978; Bickham, 1 919a) except the African E . capensis, which has 2n = 32, FN = 50 (Peterson
and Nagorsen, 1975). We report standard karyotypes for four more species of African Eptesicus, none
of which possess the 2n = 50 karyotype considered typical of the genus. Eptesicus rendalli (2n = 38),
E. brunneus (2n = 36), E. tenuipinnis (2n = 34), and E. somalicus (2n = 26) all apparently have
experienced numerous Robertsonian events and do not mirror the karyotypic conservatism of their
New World and Palearctic relatives. These data are discussed as they apply to the currently recognized
systematics of the genus Eptesicus and the proposed primitive karyotype of the family Vespertilionidae.

Introduction

The cosmopolitan bat genus Eptesicus has been a source of systematic problems
since its description. Variously, the genus has been considered as a part of the
genus Vespertilio along with Pipistrellus (Miller, 1897), and has been divided into
the genera Eptesicus and Rhinopterus (Allen, 1939). Currently, there are 34 rec-
ognized species of Eptesicus including Rhinopterus (Koopman, 1984). Approxi-
mately half of these occur in Africa, where the complex interrelationships of this
genus with the genus Pipistrellus present one of the most tangled problems of
chiropteran systematics. Koopman (1975) arranged African Eptesicus into four
groups: E. serotinus group including serotinus, platyops, loveni, hottentotus, and
bottae ; E. capensis group including melckorum, brunneus, capensis, somalicus,
and guineensis\ E. tenuipinnis group including flavescens, rendalli, and tenuipinnis ;
and E. {Rhinopterus) flowed group including E. flowed. He quickly commented,
though, that the existing Pipistrellus-Eptesicus generic separation was almost cer-
tainly wrong. He stated that “true phyletic relationships” probably run across
Pipistrellus and Eptesicus “generic lines.”

The one character regularly used to distinguish the two genera is the presence
{Pipistrellus) or absence {Eptesicus) of the first upper premolar (Fig. 1), which,
however, repeatedly has been found to be inconsistent (Koopman, 1975; Heller

1 Department of Zoology, Oklahoma State University, Stillwater, Oklahoma 74078.

2 Department of Biology, Allegheny College, Meadville, Pennsylvania 16335.
Submitted 9 February 1987.

213

214

Annals of Carnegie Museum

vol. 56

EPTESICUS-LIKE

PIPISTRELLUS - LIKE

Fig. 1.— The typical character state of the first upper premolar in members of the genera Eptesicus
and Pipistrellus.

and Volleth, 1984). No other character or suite of characters has yet proven capable
of clearly separating the two genera (Koopman, 1975).

In several instances karyology has provided a powerful tool for clarifying sys-
tematic relationships (Baker et al., 1987). The karyotypes of thirteen species of
Eptesicus have been reported (Table 1). With only the one exception of the African
species E. capensis, all other species have a diploid number (2n) of 50 and a
fundamental number (FN) of 48 (Fig. 2). Heller and Volleth (1984) considered
the departure of E. capensis from the typical Eptesicus karyological characteristics

1987

McBee et al. — Karyotypes of African Eptesicus

215

Table 1 . —A summary of known standard karyotypic data for the genus Eptesicus. SM—Submetacentric,
A —Acrocentric, M—Metacentric, ST—Subtelocentric.

Species

2n

FN

X

Y

Source

Eptesicus andinus

50

48

SM

A

Baker and Patton, 1967

Eptesicus brasiliensis

50

48

SM

A

Baker and Patton, 1967

50

48

SM

A

Baker et al., 1982

Eptesicus brunneus

36

50

SM

A

This report

Eptesicus capensis

32

50

SM

A

Peterson and Nagorsen, 1975

Eptesicus circumdatus

50

48

SM

—

Heller and Volleth, 1984

Eptesicus diminutus

50

48

SM

A

Williams, 1978

Eptesicus furinalis

50

48

SM

A

Baker and Patton, 1967

50

48

SM

A

Williams, 1978

Eptesicus fuscus

50

48

SM

A

Baker and Patton, 1967

50

48

SM

A

Bickham, 1979a

Eptesicus guadeloupensis

50

48

SM

A

Genoways and Baker, 1975

Eptesicus hottentotus

50

48

SM

—

Peterson and Nagorsen, 1975

Eptesicus japonensis

50

48

SM

SM

Ando et al., 1977

Eptesicus lynni

50

48

SM

A

Bickham, 1979a

Eptesicus nilssoni

50

48

—

—

Ando et al., 1977

50

50

M

A

Zima, 1978

50

48

M

—

Zima, 1982

Eptesicus rendalli

38

50

SM

—

This report

Eptesicus serotinus

50

48

SM

A

Baker and Patton, 1967

50

48

SM

A

Vorontsov et al., 1969

50

52

SM

SM

Fedyk and Fedyk, 1970

50

48

SM

—

Baker et al., 1974

50

48

SM

A

Bickham, 1979a

50

48

SM

—

Baker and Bickham, 1980

Eptesicus somalicus

26

48

ST

A

This report

Eptesicus tenuipinnis

36

52

SM

A

This report

as sufficient reason to

remove

the

species

from the genus Eptesicus, and the

occurrence of the typical 2n = 50 karyotype in Pipistrellus circumdatus as reason
to reclassify this species as a member of Eptesicus . They encouraged deemphasis
of the premolar as a defining characteristic and suggested that the conservative
karyological traits of Eptesicus were more reliable in differentiating between the
two genera. Hill and Francis (1984) questioned the karyologically based arrange-
ment of Heller and Volleth (1984), however, because chromosomal data were not
yet available for many species of these nominal genera. We examined standard
karyotypes prepared from bone marrow (Fig. 3-5) of four additional species of
Eptesicus from Somalia and Cameroon (Fig. 6) and found none to fit the presumed
chromosomal characteristics of the genus.

Methods

Karyotypes were obtained from 34 specimens of the genus Eptesicus using humeral bone marrow
preparations made in the field by the in vivo incubation method as described by Robbins and Baker
(1978). Nomenclature of chromosome morphology is that of Patton (1967). The presence or absence
of the first upper premolar was determined for 1 9 of the karyotyped specimens and 22 other specimens
of the same species. All specimens examined are housed in The Carnegie Museum of Natural History
(CM). Slides are housed at The Carnegie Museum of Natural History (SP) or Texas A&M University
(TK).

Specimens Examined

Eptesicus brunneus. Cameroon: 16 km S, 2 km E Yaounde, 1 F; 9 km S, 10 km W Yaounde, 4 F;
7 km S, 8 km W Yaounde, 2 F; 20 km S, 9 km E Ambam, 1 F, 1 M.

216

Annals of Carnegie Museum

vol. 56

I1*

II

II

II

II II

II

II

II

•1

II A!

1 fti

• *

i i

•»

** ft*

1 **

ft*

A*

* •

* * A 4

X

X Y

Fig. 2.— Standard karyotype of Eptesicus lynni (TK 8312), 2n = 50, FN = 48.

Eptesicus rendalli. Somalia: S.N.A.I. Sugar Plantation, IV2 km S, V2 km E Giohar, 4 F, 1 M.
Eptesicus somalicus. Somalia: S.N.A.I. Sugar Plantation, IV2 km S, V2 km E Giohar, 8 F, 5 M.
Eptesicus tenuipinnis. Cameroon: 23 km S, 1 km E Yaounde, 1 F, 1 M; 9 km S, 10 km W Yaounde,
2 F, 2 M; 20 km E Ambam, 1 F.

Results

Only two individuals, members of E. tenuipinnis, showed the presence of the
first upper premolar. In one it was present only on the left side of the jaw. In the
other it was present only on the right side of the jaw.

The standard karyotype of E. tenuipinnis is 2n = 36, FN = 52 (Fig. 3A). There
is a graded series of seven pairs of large to medium-sized metacentric and sub-
metacentric autosomes and two pairs of small submetacentric to subtelocentric
chromosomes (Fig. 3A, open arrows). The remaining eight pairs of autosomes are
acrocentric. The X is a medium-sized submetacentric chromosome and the Y is
a small acrocentric chromosome.

The standard karyotype of E. brunneus is 2n = 36, FN = 50 (Fig. 3B). This
karyotype consists of seven pairs of large to medium-sized metacentric and sub-
metacentric chromosomes but has only one small pair of subtelocentric chro-
mosomes (Fig. 3B, open arrow). There are nine pairs of acrocentric autosomes,
one of which bears a secondary constriction near the centromere (Fig. 3B, solid
arrow). The X is a medium-sized submetacentric chromosome and the Y is a
small acrocentric chromosome. The additional pair of sex chromosomes in this
figure is from SP 0318.

The standard karyotype of E. rendalli is 2n = 38, FN = 50 (Fig. 4). The

1987

McBee et al. — Karyotypes of African Eptesicus

217

»

n

n

« 4%

A*

H *

* *

o n

X.

X Y

>?

90

AA

V

n

RE

/u

o-

* m

M A

t *

OA

ft*

« *

A,

4b

X X

t.

X Y

Fig. 3.— A. Standard karyotype of Eptesicus tenuipinnis (CM 58531), 2n = 36, FN = 52. B. Standard
karyotype of Eptesicus brunneus (CM 58510), 2n = 36, FN = 50.

autosomal complement includes five pairs of large metacentric to submetacentric
chromosomes, one pair of medium-sized metacentric chromosomes, and one pair
of small subtelocentric chromosomes (Fig. 4, open arrow). The remaining 1 1 pairs
of autosomes are acrocentric, grading in size from medium to very small. A
secondary constriction is prominently apparent near the centromere of one of the
larger pairs of acrocentric chromosomes (Fig. 4, solid arrow). The X is a sub-
metacentric chromosome.

The standard karyotype of E. somalicus is 2n = 26, FN = 48 (Fig. 5). The

218

Annals of Carnegie Museum

vol. 56

M u n u n

a* Aft A IV Aft

« t

X X

■HHi

Fig. 4.— Standard karyotype of Eptesicus rendalli (CM 85307), 2n = 38, FN = 50.

autosomal complement is composed entirely of 12 pairs of metacentric to sub-
metacentric chromosomes grading from large to small. The secondary constriction
apparent on acrocentric chromosomes in other species is found near the centro-
mere of a medium-sized pair of submetacentric chromosomes in this species (Fig.
5, solid arrow). The X is a medium-sized subtelocentric chromosome and the Y
is a small acrocentric chromosome.

The standard karyotype of E. lynni is 2n = 50, FN = 48 (Fig. 2). Autosomes
consist of a graded series of 24 pairs of acrocentric chromosomes. A prominent
secondary constriction is apparent near the centromere of one medium-sized pair
(Fig. 2, solid arrow). The X chromosome is submetacentric and the Y chromosome
is minute and acrocentric. This karyotype is exemplary of chromosomal comple-
ments of 12 other species of this genus from throughout its range (Table 1).

Discussion

If karyological characteristics can be considered valid criteria for delineation
of systematic relationships (Baker, 1984; White, 1978), then serious consideration
must be given to redrawing the systematic lines defining the genera Pipistrellus
and Eptesicus. At least two interpretations are available for the karyotypic trends
observed in African Eptesicus. Of all the mammals karyologically studied, Ep-
tesicus has been considered among the most conservative (Baker and Bickham,
1980; Bickham, 1 979a). Species from North and South America, Europe, Africa,
and Asia all have an essentially identical karyotype. Prior to this study, the only
species which did not possess the typical karyotype was the African E. capensis.

1987

McBee et al. — Karyotypes of African Eptesicus

219

U '«• \S i

it it* *M\

ftp iiR <

* '« ‘A fV

fX a

*

X Y

Fig. 5.— Standard karyotype of Eptesicus somalicus (CM 85320), 2n = 26, FN = 48.

We add four more species of African Eptesicus which do not follow the conser-
vative chromosomal pattern considered typical for the genus.

The genus Eptesicus might be much more karyotypically variable than initially
thought, with the center of variability located in Africa. If this is the case, species
such as E. capensis, which Heller and Volleth (1984) would remove from the
genus on a karyotypic basis, would have to be reevaluated. Also, if Eptesicus is,
indeed, among the karyotypically more variable genera of Vespertilionidae, ad-
ditional questions are raised regarding patterns of chromosomal evolution within
the family. The typical Eptesicus karyotype has been postulated as being like the
ancestral karyotype for the family (Stock, 1983). This karyotype is found in no
other genus of the family except Histiotus. The other proposed ancestral karyotype
of 2n = 44, FN = 50, occurring in the genus Myotis (Bickham, 1979&), is found
in several other genera of the subfamily Vespertilioninae and in two of the re-
maining four subfamilies (McBee et al., 1986). Bickham (1979a) proposed a
derivation of the Eptesicus karyotype from the Myotis- like ancestral karyotype
through a process of six centric fissions. Assuming that the Myotis-like karyotype
is primitive for the family and that these chromosomally more variable forms
are indeed Eptesicus , chromosomal evolution within the genus may have pro-
ceeded first by a series of centric fissions and then by a series of centric fusions
independent of other vespertilionine lines of chromosomal evolution. G-banding
analysis of these taxa will allow identification of patterns of fusion among the
different chromosome arms and may elucidate the track of chromosomal evolution
within the family.

Alternatively, if the genus Eptesicus can be defined by the possession of 2n =
50 and FN = 48, at least five species currently recognized as members of the
genus must be reclassified. The genus Pipistrellus is characterized by broad chro-
mosomal variability (Bickham, 1979a; McBee et al., 1 986) rather than one distinct
karyotypic pattern. All five atypical Eptesicus karyotypes fit within the spectrum
of known Pipistrellus karyotypes. It is tempting to allocate these five species to

220

Annals of Carnegie Museum

vol. 56

Fig. 6. —Range maps of four species of African Eptesicus. Shaded areas represent currently recognized
ranges. Black dots represent collecting localities of individuals examined in this study.

the genus Pipistrellus, but such a move without more extensive analysis of both
standard and G-banded karyotypes and morphology would be premature. At this
point, neither interpretation of these karyotypic data is more robust than the
other. Both stress the great need for extensive redefinition and revision of the
genus Eptesicus , however, and emphasize that no single trait, whether morpho-
logical or karyological, is powerful enough to untangle the systematic lines between
Pipistrellus and Eptesicus.

Of the species in sub-Saharan Africa presently allocated to Eptesicus, E. hot ~

1987

McBee et al. — Karyotypes of African Eptesicus

221

tentotus from southern and eastern Africa (Schlitter and Aggundey, 1986) and E.
platyops from western Africa (Ib’anez and Yalverde, 1985) are typical Eptesicus
and related to E. serotinus and E.fuscus. Only E. hottentotus has been karyotyped
(McBee et al., 1986) and it agrees with the typical Eptesicus pattern. The remaining
species fall into two groups. One species in Sudan, E. floweri, is placed in the
subgenus Rhinopterus (Koopman, 1975) but a karyotype has not been reported
for this species. For the remaining species with karyotypes showing an affinity
with Pipistrellus, the subgenus Neoromicia proposed by Roberts (1926) for the
species Eptesicus zuluensis is available.

Acknowledgments

We thank M. J. Smolen, Omar Hagi, Abdulwahab Josuf, and Mohamed Ali for field assistance in
Somalia. Mohamet Abdi Nur, Minister of Agriculture; Mohamed Abikar, Director General, Ministry
of Agriculture; Abdulcadir Nur, Director, Department of Plant Protection and Locust Control in the
Ministry of Agriculture; Abdullahi Ahmed Karani, National Range Agency; John and Jonquil Ash;
Bill and Sally Srriythe; and Tony and Lynette Johnston all offered invaluable help in various capacities
in Somalia. We were ably aided in Cameroon work by S. L. Williams, L. W. Robbins, Mr. V. Belinga,
Director of Forestry Services, and Mr. C. Njiti, Ministry of Agriculture.

Field work was supported in Somalia by grants from the M. Graham Netting Research Fund,
Cordelia Scaife May Charitable Trust, and the Hays Fund of the American Philosophical Society. In
Cameroon, field work was supported by grants from the M. Graham Netting Research Fund, Cordelia
Scaife May Charitable Trust, The Loyalhanna Foundation, and National Geographic Society. Labo-
ratory work was supported by a Rea Postdoctoral Fellowship. Karl Koopman and Robert C. Dowler
reviewed an earlier version of this manuscript. Tom Mahoney assisted with photomicrography, John
W. Bickham kindly provided the karyotype of Eptesicus lynni, and Karl Koopman continues to be a
source of informative discussion on bats.

Literature Cited

Allen, G. M. 1939. A checklist of African mammals. Bulletin of the Museum of Comparative
Zoology, Harvard University, 83:1-763.

Ando, K., T. Tagawa, and T. A. Uchida. 1977. Considerations of karyotypic evolution within
Vespertilionidae. Experientia, 33:877-879.

Baker, R. J. 1984. A sympatric cryptic species of mammal: a new species of Rhogeessa (Chiroptera:

Vespertilionidae). Systematic Zoology, 33:178-183.

Baker, R. J., and J. W. Bickham. 1980. Karyotypic evolution in bats: evidence of extensive and
conservative chromosomal evolution in closely related taxa. Systematic Zoology, 29:239-253.
Baker, R. J., and J. L. Patton. 1967. Karyotypes and karyotypic variation of North American
vespertilionid bats. Journal of Mammalogy, 48:270-286.

Baker, R. J., B. L. Davis, R. G. Jordan, and A. Binous. 1974. Karyotypic and morphometric
studies of Tunisian mammals: bats. Mammalia, 38:695-710.

Baker, R. J., M. W. Haiduk, L. W. Robbins, A. Cadena, and B. F. Koop. 1982. Chromosomal
studies of South American bats and their systematic implications. Special Publications, Pyma-
tuning Laboratory of Ecology, 6:303-327.

Baker, R. J., M. B. Qumsiyeh, and C. S. Hood. 1987. Chromosomal evolution in mammals: role
of differentially-stained chromosomes. Pp. 67-96 in Current mammalogy (Hugh H. Genoways,
ed.) Plenum Press, New York.

Bickham, J. W. 1979a. Chromosomal variation and evolutionary relationships of vespertilionid
bats. Journal of Mammalogy, 60:350-363.

. 1979 b. Banded karyotypes of 1 1 species of American bats (Genus Myotis ). Cytologia, 44:

789-797.

Fedyk, A., and D. Fedyk. 1970. Karyotypes of some species of vespertilionid bats from Poland.
Acta Theriologica, 20:295-302.

Genoways, H. H., and R. J. Baker. 1975. A new species of Eptesicus from Guadeloupe, Lesser
Antilles (Chiroptera: Vespertilionidae). Occasional Papers of The Museum, Texas Tech Univer-
sity, 34:1-7.

Heller, K.-G., and M. Volleth. 1984. Taxonomic position of “Pipistrellus societatis ” Hill, 1972
and the karyological characteristics of the genus Eptesicus (Chiroptera: Vespertilionidae). Zeit-
schrift fur Zoologische Systematik und Evolutionsforschung, 22:65-77.

222

Annals of Carnegie Museum

vol. 56

Hill, J. E., and C. M. Francis. 1984. New bats (Mammalia: Chiroptera) and new records of bats
from Borneo and Malaya. Bulletin of the British Museum (Natural History), London, 47:305-
329.

Ib’anez, C., and J. A. Valverde. 1985. Taxonomic status of Eptesicus platyops (Thomas, 1891)
(Chiroptera, Vespertilionidae). Zeitschrift fur Saugetierkunde, 50:241-242.

Koopman, K. F. 1975. Bats of the Sudan. Bulletin of the American Museum of Natural History,
154:354-444.

-. 1984. Bats. Pp. 145-186, in Orders and families of Recent mammals of the world (S.

Anderson and J. K. Jones, Jr., eds.), John Wiley and Sons, New York, 686 pp.

McBee, K., J. W. Bickham, S. Yenbutra, J. Nabhitabhata, and D. A. Schlitter. 1986. Standard
karyology of nine species of vespertilionid bats (Chiroptera: Vespertilionidae) from Thailand.
Annals of Carnegie Museum, 55:95-1 16.

Miller, G. S., Jr. 1 897. Revision of the North American bats of the family Vespertilionidae. North
American Fauna, 13:1-135, 3 plates.

Patton, J. L. 1967. Chromosome studies of certain pocket mice, genus Perognathus (Rodentia:
Heteromyidae). Journal of Mammalogy, 48:27-37.

Peterson, R. L., and D. W. Nagorsen. 1975. Chromosomes of fifteen species of bats (Chiroptera)
from Kenya and Rhodesia. Life Science Occasional Papers, Royal Ontario Museum, 27:1-14.

Robbins, L. W., and R. J. Baker. 1978. Karyotypic data for African mammals, with a description
of an in vivo bone marrow technique. Pp. 188-210, in Ecology and taxonomy of African small
mammals (D. A. Schlitter, ed.). Bulletin of the Carnegie Museum of Natural History, 6:1-214.

Roberts, A. 1926. Some new South African mammals and some changes in nomenclature. Annals
of the Transvaal Museum, 1 1:245-263.

Schlitter, D. A., and I. R. Aggundey. 1986. Systematics of African bats of the genus Eptesicus
(Mammalia: Vespertilionidae). 1. Taxonomic status of the large serotines of eastern and southern
Africa. Cimbebasia, series A, 8(20): 167-1 74.

Stock, A. D. 1983. Chromosomal homologies and phylogenetic relationships of the vespertilionid
bat genera Euderma, Idionycteris, and Plecotus. Cytogenetics and Cell Genetics, 35:135-140.

Vorontsov, N. N., S. I. Radjabli, and V. T. Volobuev. 1969. The comparative karyology of the
vespertilionid bats, Vespertilionidae (Chiroptera). Pp. 16-21, in The mammals (N. N. Vorontsov,
ed.), 2nd All Union Mammalogical Conference, USSR Academy of Sciences, Moscow.

White, M. J. D. 1978. Chain processes in chromosomal speciation. Systematic Zoology, 27:285-
298.

Williams, D. F. 1978. Taxonomic and karyologic comments on small brown bats, Genus Eptesicus,
from South America. Annals of Carnegie Museum 47:361-383.

Zima, J. 1978. Chromosome characteristics of Vespertilionidae from Czechoslovakia. Acta Scien-
tiarum Naturalium Academiae Scientiarum Bohemoslovacae, Brno, 12:1-38.

. 1982. Chromosomal homology in the complements of bats of the family Vespertilionidae.

II. G-banded karyotypes of some Myotis, Eptesicus and Pipistrellus species. Folia Zoologica, 31:
31-36.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 12, Pp. 223-230 28 August 1987

REVISION OF THE WIND RIVER FAUNAS, EARLY
EOCENE OF CENTRAL WYOMING. PART 8. FIRST FOSSIL
LIZARD EGG (7GEKKONIDAE) AND LIST OF
ASSOCIATED LIZARDS

Karl F. Hirsch1

Leonard Krishtalka
Associate Curator, Section of Vertebrate Fossils

Richard K. Stucky

Assistant Curator, Section of Vertebrate Fossils

Abstract

A small fossil egg from the Lostcabinian (early Eocene, CM loc. 1040) of the Wind River Formation,
Wyoming, is the first unequivocal record of a fossil lizard egg. It has a shell structure similar to that
of rigid-shelled eggs of modem geckos, especially Tarentola delalandii. The unique eggshell structure
of Recent geckos warrants recognition of a fourth type of rigid eggshell in addition to those of chelonians,
crocodilians and birds. At least 22 species of lizards are represented by skeletal material at the fossil
egg locality, including anguids, necrosaurids, agamids, varanids, xantusiids, xenosaurids, iguanids,
and teiids.

Introduction

In 1984, paleontologists from The Carnegie Museum of Natural History col-
lected several fossil eggs from early Eocene (late Wasatchian, Lostcabinian) ho-
rizons of the Wind River Formation in central Wyoming. One of these eggs is
small (8-9 mm diameter), has a rigid eggshell, and represents the first unequivocal
record of a fossil lizard egg. It resembles eggs of living geckos, the only extant
lizards with rigid eggshells. Other fossil squamate eggs have been mentioned in
the literature (Meyer, 1867) and identified in some collections on the basis of
gross features of the specimens, which may not be diagnostic.

The eggshell structure of only a few species of squamates has been studied in
detail; that of geckos has been mentioned only briefly (Schmidt, 1957; Hirsch,
1983, 1985; Packard et al., 1982). However, studies of the shell microstructure
(Erben and Newesely, 1972), and the biominerals of the shell membrane and
calcitic layer (Krampitz et al., 1972, 1974) established that gecko eggshells differ
from those of birds, crocodilians and chelonians, and represent a fourth type of
rigid eggshell.

Fossil geckos are first known from the late Paleocene of Brazil and the late
Eocene of Europe (Estes, 1983). In North America, the earliest geckos are known
from the early Miocene of Florida, but have been identified with question from
the late Middle Eocene of the Mission Valley Formation, California.

1 University of Colorado Museum, Campus Box 218, Boulder, Colorado 80309.
Submitted 6 April 1987.

223

224

Annals of Carnegie Museum

vol. 56

Table 1 . —Fossil lizards known from skeletal material from the B-2 horizon at Buck Spring (CM loc.
1040), Lost Cabin Member, Wind River Formation.

Taxon

No. specimens

Anguidae

Diploglossinae sp.

18

Xestops

14

lOphisaurus sp.

3

Machaerosaurus sp.

2

Glyptosaurus donohoei

1

cf. Gerrhonotus sp.

1

Anguinae sp.

1

Xantusiidae

Xantusiidae sp. A

9

Xantusiidae sp. B

3

Xantusiidae sp. C

3

Xenosauridae

Exostinus sp.

6

Agamidae

Tinosaurus n. sp.

4

Iguanidae

Iguanidae sp. A

7

Iguanidae sp. B

4

Iguanidae sp. C

3

Parasauromalus sp.

1

Varanidae

Varanidae sp.

3

Necrosauridae

Necrosaurus sp.

2

Teidae

?Teidae sp.

1

Sauria

Oligodontosaurus sp.

2

“Scincomorpha” sp.

3

Sauria n. gen.

2

Sauria undetermined

100+

Locality and Associated Fauna

The fossil egg was recovered from the Buck Spring Quarries (Quarry 1 , Horizon
B-2, CM loc. 1040) in the type area of the Lost Cabin Member of the Wind River
Formation. As described elsewhere (Stucky and Krishtalka, 1987; in press), the
geology of the locality and mode of preservation of the fossil vertebrates are unique
among known Wind River Formation localities. The eggshell-bearing horizon
consists of alternating thinly-laminated mudstone and limestone couplets, which
were apparently deposited in freshwater ponds and/or well-drained swamps within
250 m of the nearest permanent stream.

At least 22 species of lizards are known from dental and skeletal remains from
the B-2 horizon (Table 1). This is the highest diversity recorded among known
Eocene lizard faunas. However, no gekkonid skeletal material has been positively
identified from this horizon. One of the lizards, Glyptosaurus, is here considered
a senior synonym of Eoglyptosaurus. Study of the type species of the latter, E.

1987

Hirsch et al.— Eocene Lizard Egg

225

Table 2.— Size, shell thickness and status of modern gecko and fossil egg.

Specimens

Egg size
(mm)

Shell thickness
(mm)

Status of egg

Eocene egg

CM 46668/HEC 321

8x9

0.04-0.05

Tarentola delalandii

UCM OS1 142/HEC 357

9 x 10

0.06-0.07

hatched

T. mauritanica

UCM OS1 129/HEC 408

11 x 12

0.09-0.1

?infertile

Phelsuma madagascariensis

UCM OS1 130/HEC 125

13 x 15

0.14-0.16

hatched

Gekko gecko

UCM OS1 143-1/HEC 391-1

21 x 19

0.27-0.34

?infertile

G. gecko

UCM OS1 143-2/HEC 391-2

21 x 19

0.27-0.34

?infertile

G. gecko

UCM OS1 143-3/HEC 391-2

21 x 19

0.27-0.34

embryo/early stage

Lepidodactylus lugubris

Erben and Newesely, 1972

_

?0.9

Hemidactylus turcicus

Packard et al., 1982

—

?0.06

—

donohoei (including the holotype, USNM 18316), reveals that it has a linear band
of pterygoid teeth, a feature also found in G. sylvestris, the type species of Glyp-
tosaurus (Sullivan, 1986). Sullivan erred in separating Eoglyptosaurus from Glyp-
tosaurus on this characteristic, as well as on the presence of raised subconical
osteoderms, which appears to vary allometrically in G. donohoei (Stucky, unpub-
lished data).

In addition to the lizard eggshell and skeletal material, relatively complete and
well preserved specimens of fossil mammals, turtles, birds, snakes, amphibians,
crocodilians and fish have been recovered from the Buck Spring Quarries. The
B-2 horizon also preserves fossil roots and stems of plants (carbonate-replaced
and lignified), avian eggshells, algal skeletal grains, and rarely, gastropods and
decapod exoskeletons.

Material and Techniques

Specimens of modem and fossil eggshell (Table 2) were studied using polarizing light microscopy
(PLM), scanning electron microscopy (SEM) and X-ray diffraction (XRD). Methods described in
Hirsch (1979, 1983) and Packard et al. (1984) were followed as far as the very delicate and fragile
nature of the specimen allowed. Three small shell fragments were detached from the fossil egg, the
remainder of which is still preserved in the surrounding block of matrix (CM 46668, Fig. 1). These
fragments were too fragile to be treated with chemicals or cleaned in an ultrasonic bath. Thus the
outer and the inner surfaces are somewhat contaminated by secondary deposits or preservatives.
However, fresh fractures of the shell permitted examination of radial (edge) sections. Tangential
sections of the fossil specimen could not be prepared. The status of the egg (fertilized or unfertilized,
stage of incubation) was known in three cases for the modem eggshells.

Abbreviations: CM, The Carnegie Museum of Natural History; HEC, Hirsch Egg Collection; UCM,
University of Colorado Museum; USNM, U.S. National Museum.

Results

Modern Rigid-Shelled Gecko Eggs

The rigid eggshell, one of the three main kinds of amniote eggshells (Hirsch,
1983, 1985; Packard and Packard, 1980), is the most likely to become fossilized.
The calcareous layer of these eggshells is very thick in comparison to the underlying
membrane and is composed of well-defined, interlocking, more or less spherulitic

226

Annals of Carnegie Museum

vol. 56

Chelonian Avian Crocodilian Gekko

shell

h- unit -H

fi

c

1 - — (-shell unit

1ll|

blumn

layer

^--Organic

cover

i

H — shell unit — H

Zone of
spiculor
aragonite
aggregates

\

/ 'oyer 1 MS } Tny™

— f , » Membrane .. t

Membrane / ^ \ 1 Mem|>rone Membrane

Primary spherite'" Mammilla cap Basic plates

(Central core)

2

Fig. 1, 2.— 1. Fossil egg embedded in matrix. Scale = 1 mm. 2. Terminology and structure of the four

types of modem rigid eggshells.

shell units that are nucleated on the membrane. Among amniotes with rigid-
shelled eggs, a typical shell structure has been recognized for birds, crocodilians
and turtles (Schmidt, 1943, 1957; Schmidt and Schoenwetter, 1943; Erben and
Newesely, 1972; Hirsch, 1983, 1985). The characteristic structure of rigid-shelled
gecko eggs differs from these types and warrants classification as a fourth type
(Fig. 2; see also Erben and Newesely, 1972; Krampitz, 1972).

Gecko eggshells are composed of a thin inner membrane overlain by a com-
paratively thick calcareous layer (0.06-0.34 mm, see Table 2), which is covered
by a more or less fibrous, organic cuticle. The surface of the cuticle can be an
open network of fibers, or relatively smooth, or bumpy as a result of the nodes
of the underlying calcareous layer.

The calcareous layer, as in the avian eggshell, is made up of tightly abutted
calcite columns extending from the membrane to the outside of the shell. In edge
view, the columns interlock along uneven, jagged surfaces (Fig. 4, 5) and thereby
differ from the spherulitic, even, interlocking crystal pattern found in avian egg-
shells. The jagged columnar structure is typical for the first four species examined
in Table 2 and has also been observed in Lepidodactylus and Hemidactylus (Erben
and Newesely, 1972; Packard et al., 1982). In radial view the columns show
horizontal and sometimes almost lamellar layering. Under PLM, faint columnar
extinction patterns can be observed. In tangential thin sections, the pattern of the
columns is polygonal, but randomly interlocking. As in all other shell types, the
gecko eggshell varies from one point to another in columnar size and shell thick-
ness.

1987

Hirsch et al.— Eocene Lizard Egg

227

Fig. 3-5. — Tarentola delalandii eggshell (scale = 10 /*). Outside of shell is up. 3. Radial thin section
viewed under polarized light. Note that columnar structures are somewhat narrower than in the fossil
egg (see Fig. 6). 4. Photomicrograph of radial view (edge) of eggshell. 5. Enlargement of Fig. 4. Note
the minute crystals on mammillary column faces (arrow).

228

Annals of Carnegie Museum

vol. 56

Fig. 6-8.— Fossil eggshell (scale = 10 n). Outside of shell is up. 6. Radial thin section viewed under
polarized light. Note faint columnar structures. 7. Micrograph of radial view (edge) of eggshell. Note
similarity to radial view in Fig. 4. 8. Enlargment of Fig. 7. Note jagged columnar structure and
horizontal layering.

1987

Hirsch et al.— Eocene Lizard Egg

229

A shell unit structure (see Fig. 2) has not yet been observed in gecko eggshells,
either with SEM or PLM. It also has not been established how the calcite columns
are nucleated or how they are fastened to the membrane. Central cores, such as
occur in avian and chelonian eggshells, and basal plate groups, as in crocodilian
eggshells, have not been observed. In unfertilized and fertilized but unhatched
eggs the basal tips of the columns are tightly fastened to the membrane; in hatched
eggs this membrane is much less securely fastened.

The shape of the basal tip or face of the columns in gecko eggs varies inter-
specifically. Variation within a species, observed in several eggs with different
incubational histories, may be partly a result of dissolution of columns caused by
the withdrawal of calcium from the eggshell by the developing embryo (Packard
et al., 1984). In hatched eggs, the basal faces are studded with minute crystals
pointing toward the inside of the egg. In eggs of unknown hatching status, round
or lumpy tips can be observed.

The Wind River Fossil Egg

The Eocene egg (CM 46668) was compressed during fossilization, but its original
shape was apparently spheroidal and measured approximately 8-9 mm in di-
ameter. Eggshell thickness is 0.04-0.05 mm. Like the eggshell of living gekkonids,
the fossil eggshell is rigid and does not have the unit structures typical of chelonian,
crocodilian or avian eggshells (Fig. 2).

In size and shell structure, the fossil egg is most similar to that of Tarentola
delalandii (Fig. 4, 5, 7, 8). In both, the columns are wide and strongly jagged, and
the horizontal layering (growth lines) is very distinct (Fig. 5, 8). These features
are not as pronounced in other living species of geckos that have been studied.
Both specimens also exhibit a faint columnar extinction pattern when rotated
under crossed nicols (Fig. 3, 6).

The outer surface of the fossil egg is relatively smooth, with no indication of
nodes or other sculpturing. Under high magnification, some crystal structure of
the column heads is visible.

It is not clear whether the details of the inner surface of the fossil shell reflects
original morphology or diagenetic alteration. Low magnification (SEM) reveals a
very faint indication of a polygonal pattern, which may represent the faces of
dissolved column tips as described above for some hatched gecko eggs. However,
unlike Tarentola (Fig. 5), high magnification does not reveal distinct small crystals,
only small crystalline nodes, perhaps a result of diagenetic recrystallization.

Microprobe and X-ray fluorescence studies determined that the fossil eggshell
is composed of calcium carbonate. X-ray diffraction analysis could not distinguish
between aragonite and calcite because of the limited amount of fossil material;
however, the crystalline structure strongly suggests calcite.

Conclusions

A preliminary study of modern rigid-shelled gecko eggs indicates that the jagged,
columnar structure of the calcareous layer distinguishes them from recognized
types of avian, crocodilian and chelonian eggshells. Based on this diagnostic
feature, recognition of a fourth type of rigid eggshell is warranted, although in-
terspecific variation in the microstructure of the jagged columns and cuticle occurs
among geckos.

The Early Eocene egg from the Wind River Formation is the first such record
for fossil lizards. Two of its features— the rigid shell and jagged, columnar structure

230

Annals of Carnegie Museum

vol. 56

of the calcareous layer— are diagnostic. They are known only among extant geckos,
and closely resemble the condition in Tarentola.

Fossil gecko bones are unknown in the Wind River Formation, although some
of the lizard material is provisionally identified and may represent gekkonids.
Nevertheless, the fossil egg may have belonged to one of the other lizards recovered
from the B-2 horizon in the Wind River Formation at Buck Spring.

Acknowledgments

We thank Mercedes Martin and Vidal M. Mantel (Canary Islands, Spain), J. P. Bacon (San Diego
Zoo) and Dale Marcellini (National Zoological Park) for modem gecko eggshell material; Judith Harris
and Mary Dawson for their comments on the manuscript; John Drechsler and Richard Harding for
assistance with the SEM work; and Elizabeth A. Hill for wordprocessing.

The study was supported in part by grants from the University of Colorado, Boulder, and Rockwell
International (to UCM), the National Science Foundation (BSR-840205 1 to Krishtalka and Stucky)
and the M. Graham Netting Research Fund (to Krishtalka and Stucky).

Literature Cited

Erben, H. K., and H. Newesely. 1972. Kristalline Bausteine und Mineralbestand von kalkigen
Eierschalen. Biomineralisation, 6:32-48.

Estes, R. 1983. Sauria Terrestria, Amphisbenia. Part 1 0A. Encyclopedia of paleoherpetology. Gustav
Fischer Verlag, New York, 249 pp.

Hirsch, K. F. 1979. The oldest vertebrate egg? Journal of Paleontology, 53:1068-1084.

. 1983. Contemporary and fossil chelonian eggshells. Copeia, 1 983(2):382— 397.

. 1985. Fossil crocodilian eggs from the Eocene of Colorado. Journal of Paleontology, 59:

531-542.

Krampitz, G., H. K. Erben, and K. Kriesten. 1972. On the amino acid composition and structure
of eggshells. Biomineralization, 4:87-99.

Krampitz, G., W. Boehme, K. Kriesten, and W. Hardebeck. 1974. Die Aminosaeurenzusam-
mensetzung von Reptilien-Eischalen in biochemischer und evolutiver Sicht. Zeitschrift fur Zoolo-
gische Systematik und Evolutionsforschung, 12:1-22.

Meyer, H. von. 1867. Ueber fossile Eier und Fedem. Palaeontographica, 15:223-252.

Packard, G. C., and M. J. Packard. 1980. Evolution of cleidoic egg among reptilian antecedents
of birds. American Zoologist, 20:351-362.

Packard, M. J., G. C. Packard, and T. J. Boardman. 1982. Structure of eggshells and water
relations of reptilian eggs. Herpetologica, 38:136-155.

Packard, M. J., G. C. Packard, and W. H. N. Gutzke. 1984. Calcium metabolism in embryos
of the oviparous snake Coluber constrictor. Journal of Experimental Biology, 1 10:99-1 12.
Schmidt, W. J. 1943. Ueber den Aufbau der Kalkschale bei den Schildkroeteneiem. Zeitschrift fur
Morphologic und Oekologie der Tiere, 40:1-16.

. 1957. Ueber den Aufbau der Schale des Vogeleies nebst Bemerkungen ueber kalkige Eischalen

anderer Tiere. Berichte der Oberhessischen Gesellschaft fur Natur- und Heilkunde, Giessen, 28:
81-108.

Schmidt, W. J., and M. Schoenwetter. 1 943. Beitraege zur Kenntnis der Krokodileier, inbesondere
ihrer Kalkschale. Zeitschrift fur Morphologic und Oekologie der Tiere, 40:17-36.

Stucky, R. K., and L. Krishtalka. 1987. The Buck Spring Quarries: exceptional preservation of
early Eocene (Wasatchian) vertebrates from the Wind River Basin, Wyoming. Abstracts with
Programs, Geological Society of America, 19(5):337.

. In press. The Buck Spring Quarries: exceptional preservation of early Eocene (Wasatchian)

vertebrates from the Wind River Basin, Wyoming. Geological Society America, Special Paper.
Sullivan, R. M. 1986. The skull of Glyptosaurus sylvestris Marsh, 1871 (Lacertilia: Anguidae).
Journal of Vertebrate Paleontology, 6:28-37.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 13, Pp. 231-243 28 August 1987

TAXONOMIC NOTES ON SOME AFRICAN WARBLERS
(AVES: SYLVIINAE)

Kenneth C. Parkes

Senior Curator, Section of Birds
Abstract

Notes on geographic variation are presented for six species of African warblers: Chloropeta similis,
Cisticola chiniana, Apalis jacksoni, Apalis porphyrolaema, Apalis cinerea, and Sylvietta leucophrys.
Two new subspecies of C. chiniana are described from Tanzania, and one new subspecies of S.

leucophrys from Uganda.

Introduction

During the course of incorporating the Twomey collections of African birds
(see Parkes, 1980 for history of these collections) into the main series at The
Carnegie Museum of Natural History, I encountered several taxonomic problems
that could not be solved with the material at hand. The pertinent specimens were
therefore taken to the American Museum of Natural History for further study.
As a followup, critical specimens of certain species were borrowed from other
museums (see Acknowledgments). This paper presents the results of these studies.
Species discussions are given in the sequence of Traylor (1986). The present paper
was essentially completed before the publication of that book, but minor modi-
fications have been made in the manuscript to cite Traylor where appropriate.
Distances are given metrically except when cited from original labels or publi-
cations.

Abbreviations for museums: AMNH, American Museum of Natural History;
ANSP, Academy of Natural Sciences of Philadelphia; BM(NH), British Museum
(Natural History); CM, Carnegie Museum of Natural History; FMNH, Field Mu-
seum of Natural History; LACo, Los Angeles County Museum; MCZ, Museum
of Comparative Zoology, Harvard University; UMMZ, University of Michigan
Museum of Zoology; USNM, United States National Museum of Natural History;
YPM, Peabody Museum of Natural History, Yale University.

Chloropeta similis

Direct comparison of a series of 9 topotypes of similis from Mt. Kilimanjaro
with 1 1 topotypes of kenya Sharpe from Mt. Kenya indicates that Friedmann
(1937:234) was correct in synonymizing the latter with similis. Sharpe (1901), in
describing kenya, compared it only with natalensis, now considered a separate
species, and apparently overlooked the description of similis by Richmond in
1897. In series, the Mt. Kilimanjaro birds average faintly browner, less greenish
on the dorsum, but the series overlap broadly. There appears to be no significant
change in dorsal coloration owing to museum age (“foxing”); all of the Mt. Kil-
imanjaro series were collected in 1 920-192 1 , but Mt. Kenya specimens were taken
in 1926, 1936, and 1963, and show no age-related differences.

Submitted 16 January 1987.

231

232

Annals of Carnegie Museum

vol. 56

Friedmann (1937:234) did not examine specimens of Chloropeta schuhotzi
Reichenow, described from the Rugege Forest, east of Lake Kivu (now in Rwanda),
but synonymized it with similis on the basis of notes by other authors. Reichenow
(1908) compared his schubotzi only with “C7*. major ” (= Chloropeta natalensis
major of Angola) and not with either similis or kenya.

Comparison of one topotype and a series of 1 9 specimens from Kivu, Baraka,
and Ruwenzori assigned to “ schubotzi ” at the AMNH with the 20 specimens of
similis from Kenya mentioned above failed to demonstrate the darker back at-
tributed by some authors to this supposed western race. I concur with Friedmann
and others who have synonymized schubotzi with similis.

Within Kenya, comparisons were made between the Mts. Kilimanjaro/Kenya
series and a series from the Aberdare Range and the vicinity of the Mau Escarp-
ment, just to the west. The Aberdare/Mau series, which triggered this investiga-
tion, averages noticeably brighter (yellower) above, especially on the forehead,
but the differences are subtle and individual specimens would be difficult to allot.
The best course seems to be to continue to consider this species to be monotypic,
as Traylor (1986) has done, admitting that there are some incipient tendencies
toward geographic differentiation.

Cisticola chiniana

Before beginning the discussion of this species, a note on the plumages of
Cisticola is in order. It is well known that some populations of this genus have
two distinctive seasonal plumages, whereas others have the same aspect all year
round. It is fairly clear from the account by Lynes in his classical monograph of
this genus (1930:41) that those forms that do not change in appearance do not
nevertheless have two molts annually: . . a perennial dress which is renewed
only once a year during a brief off-season.” As the term “off-season” is sometimes
used in older literature in connection with a “non-breeding” or “eclipse” plumage,
it should be made clear that Lynes used it merely to indicate a time of year when
no breeding takes place, i.e., during the molt. Lynes used the terms “perennial
mode” and “seasonal mode” to differentiate the one-plumage and two-plumage
annual cycles in Cisticola. His description of the plumage sequence in the perennial
mode (p. 44) makes it clear that his “immature dress” is the First Basic Plumage
of Humphrey and Parkes (1959), which is highly variable, among species and
undoubtedly among populations of Cisticola, as to duration and appearance. Even
though the terms may seem inappropriate for Africa, Lynes uses “winter dress”
and “summer dress” for the seasonal plumages of Cisticola ; these correspond
functionally to the non-nuptial and nuptial plumages of many authors, and from
a homology viewpoint, to the Basic and Alternate plumages of Humphrey and
Parkes (1959). The Alternate Plumage has thus evolved in some but not all
populations of Cisticola ; it is particularly interesting to note that the Basic Plumage
of some of the forms with the perennial mode resembles in color the Alternate
Plumage and in others the Basic Plumage of those forms with a two-plumage
cycle. One of the few species in which both modes can be found, with implications
for geographic variation, is Cisticola chiniana.

Geographic variation in this species in Tanzania is complex. Lynes (1930)
attributes four races to this country: procera Peters in the southeast, heterophrys
Oberholser in the east, fischeri Reichenow in the central and northern interior,
and victoria Lynes in the Victoria Nyanza basin in the north. Vincent (1944)
divided the range attributed to procera by Lynes into northern and southern

1987

Parkes— African Warblers

233

components, describing the more northern subspecies (the one that would reach
southeastern Tanzania) as emendata, and stating that this action had the support
of Lynes, who had been dissatisfied with his own 1930 concept of procera. White
(1962) substantially altered the arrangement of Tanzanian races of C. chiniana.
He synonymized emendata with procera, thus returning to the treatment of Lynes
for the southeastern area. He synonymized victoria with fischeri, thus admitting
only one subspecies across northern and north-central Tanzania, from the eastern
and western shores of Lake Victoria south to Tabora. However, he added another
subspecies to the Tanzanian list. He identified the populations of the northeastern
highlands, along the Kenya border, as ukamba Lynes, described from the eastern
plateau region of Kenya, and those of Lake Rukwa, in the southwest, as fortis
Lynes, described from interior Angola. Britton (1 980) essentially followed White’s
treatment of subspecific ranges, although without mentioning Tanzania in the
range of ukamba.

Lynes characterized his race victoria, with a type locality on the east side of
Lake Victoria, as larger and more boldly marked than fischeri, with “the markings
giv[ing] the upper side as much a striped as mottled appearance.” In synonymizing
victoria with fischeri, White (1962) attributed the large size of the specimens from
along the south and east sides of Lake Victoria to intergradation with humilis
Maderasz of the West Kenya highlands. He did not address the subject of color.
I have compared 5 near-topotypical males of victoria with five males of typical
fischeri. The two series differ in size as stated by both Lynes and White; in addition,
the black streaks of the mantle are distinctly broader in victoria, and the crown
is darker brown, with more conspicuous black centers to the feathers; in fischeri
the crown feathers are only slightly darker in the center than on the edges. The
difference in appearance of the crown is well expressed by Lynes’s comparison
between “striped” ( victoria ) and “mottled” {fischeri ). There is no difference in
underparts color. A series from the Serronea River, in the Serengeti region (ANSP)
is more or less intermediate between victoria and fischeri, and a rather worn male
from 10 miles S of Mwanza, on the south shore of Lake Victoria (BM[NH]), also
appears intermediate. I am skeptical about White’s identification of victoria as a
population of fischeri whose large size is attributable to intergradation with hu-
milis. The latter is conspicuously paler dorsally than fischeri with more contrasting
edges to the tertials, and there is no sign of approach to the color of humilis among
specimens of victoria. Furthermore, the dorsal streaks of humilis, although more
conspicuous than those of fischeri, are as narrow as those of that race, so that the
broader streaks of victoria cannot be attributed to intergradation with humilis. It
would appear that victoria Lynes is worthy of acceptance, but its exact range and
relationships with bordering races will require examination of better material than
is now available. I have not seen specimens that would permit me to evaluate
the validity of emendata Vincent, so I follow White in using the name procera
for the southeastern Tanzanian populations, without prejudice. I do agree with
White that the populations of the Kilimanjaro region, northwest to Lake Natron,
are indistinguishable from ukamba. Specimens from Loliondo, northwest of Lake
Natron, however, are nearest the other Kenyan race, humilis. These two races
differ more in size than in color; I am unable to appreciate most of the differences
suggested by figs. 38 and 39, plate 10, of Lynes (1930), except the more buffy,
less grayish edgings of the tertials of humilis ; the ground color of the crown averages
duller in humilis, and, contrary to the figures cited above, the streaks on the crown
are more distinct in humilis than in ukamba.

234

Annals of Carnegie Museum

vol. 56

Fig. 1.— Principal localities mentioned in text account of Cisticola chiniana. 1. Arusha. 2. Chimala.
3. Dar-es-Salaam. 4. Dodoma. 5. Iringa. 6. Kidodi. 7. Kilimanjaro. 8. Kilosa. 9. Kipangere Range.
10. Lake Manyara. 11. Lake Natron. 12. Lake Nyasa. 13. Lake Rukwa. 14. Lake Tanganyika. 15.
Lake Victoria. 16. Loliondo. 17. Mbeya. 18. Mombasa. 19. Mwanza. 20. Same. 21. Serronea River.
22. Shinyanga. 23. Tabora. 24. Tura. 25. Uluguru Mts. 26. Usambara Mts.

A careful reading of the ranges attributed to Tanzanian races of C. chiniana by
White (1962) and Britton (1980) reveals a large area of the interior of the country
missing from any of the range descriptions. Judging from the map of the species’
range in Hall and Moreau (1970), part of this gap in range descriptions may reflect
an actual lack of records of occurrence, but neither White nor Britton assigns any
subspecies to the area southeast of Tabora, where the species does indeed occur.
Britton’s account seems to suggest that chiniana is replaced in the Iringa District
of south-central Tanzania by Cisticola njombe, but this is true only in the south-
ernmost part of the District, where njombe is confined to higher elevations than
those reached by chiniana in this area. It appears that njombe is limited to a

1987

Parkes — African Warblers

235

rather restricted habitat, and its relationships are uncertain. Although other recent
authors had considered it nearest to C. lais, Hall and Moreau (1970) tentatively
listed it as forming a superspecies with C. chiniana. Britton (1980) stated that
njomhe forms “a superspecies with the allopatric C. chiniana ,” but his concept
of superspecies must differ from the usual one, as he placed 12 other species of
Cisticola between chiniana and njombe. In the most recent treatment of the genus,
Traylor (1986) reverts to placing njombe near lais rather than chiniana. It is
doubtful that any sort of competitive exclusion is involved in the relative distri-
butions of chiniana and njombe in south-central Tanzania.

The ranges given by Lynes are not precisely outlined, but it appears as if the
birds from all of central Tanzania should be fischeri. However, a series of 27
specimens from the vicinities of Dodoma and Iringa (BM[NH], YPM, CM) in-
dicates that there is a population occupying an area in central Tanzania that differs
from fischeri in having seasonal plumages. I am indebted to M. A. Traylor, Jr.,
for calling my attention to an easily overlooked statement by Lynes (1934 b).
Referring to a series taken at Iringa (BM[NH]), Lynes stated only “The Cisticola
Review requires correction in respect of fischeri’ s mode of dress”; by this he meant
that some of the Iringa specimens, taken in November, were molting, demon-
strating that this population does in fact have seasonal plumages, whereas he had
attributed the perennial mode to fischeri in his 1930 monograph. However, Lynes
was mistaken in considering this population to be typical of fischeri. I have ex-
amined specimens of fischeri from elsewhere in its range taken during the months
of December through August, and find no seasonal differences other than those
attributable to normal wear and fading. Although the “summer” plumage of the
birds in the Dodoma/Iringa series is indistinguishable from that of fischeri, the
“winter” plumage is highly distinctive, and matched in no other population of
the species. The BM(NH) series clearly shows that the prealternate molt into
“summer” plumage begins in mid-November. Unfortunately no specimens are
available from this area taken later than 6 March or earlier than 20 August, the
period during which the prebasic molt must take place (see beyond for mention
of a worn 20 May specimen from the putative eastern edge of the range of the
central Tanzanian race). August and September specimens in the type series are
relatively lightly worn.

It may seem to some that a population recognizable during only part of the
year should not be given a subspecific name, but a little contemplation will aid
in recalling that there is ample precedent among migratory temperate species; the
population of Cisticola chiniana differs strikingly from these in that it is the Basic
rather than the Alternate Plumage that is distinctive. This population of central
Tanzania may be called:

Cisticola chiniana keithi , new subspecies

Holotype. — CM 146864, adult male, from 50 miles N of Dodoma, Dodoma
(formerly Central) Province, Tanzania, collected 10 September 1960 by A. C.
Twomey and John Williams (field no. 2559).

Diagnosis. — Inseparable from C. c. fischeri (sensu Lynes, 1930) in Alternate
(“summer”) Plumage. Basic Plumage unlike Alternate Plumage, nearest to the
“perennial” (Basic) plumage of C. c. ukamba of the highlands of Kenya and
northeastern Tanzania, but decidedly paler. The edges of the mantle feathers,
tertials, and wing coverts are not merely paler but huffier, less gray. The crown
is paler, with black streaks more sharply defined. The pale buff* superciliary line

236

Annals of Carnegie Museum

vol. 56

is distinct (obsolete or absent in ukamba). The tail is paler and of a more reddish
brown, and the flanks and a pale wash across the lower breast are of a yellower,
less grayish buff. In the distinctness of the crown streaking, keithi resembles humilis
rather than ukamba , but the crown is otherwise much paler than in humilis.

C. c. keithi differs most distinctly from the coastal race heterophrys, which is a
dark, plain-backed form with no streaks on the crown and, in some individuals
only, barely discernible blurry streaks in the centers of the dark gray-brown mantle
feathers.

There is some noticeable individual variation among the 9 relatively unworn
August-September specimens. CM 146844 has the mantle edgings (but not those
of the wing feathers including the tertials) more rufescent, approaching those of
procera, and also has the broadest black mantle streaks. The crown is darkest in
YPM 93258 (which also has the grayest edges to the dorsal feathers), palest in
CM 146754. The holotype and CM 146844 have the greatest development of the
narrow black crown streaks; in YPM 93260 they are almost invisible. Collectively,
however, these specimens do not differ inter se any more than series of other
subspecies, from which they are clearly differentiated.

Range. — Central Tanzania; see Remarks for intergradation with adjacent sub-
species.

Etymology. — It is a pleasure to dedicate this new subspecies to G. Stuart Keith
of the American Museum of Natural History, my mentor in African ornithology.

Remarks. — Eastern Tanzania is occupied by the distinctive C. c. heterophrys.
A coastal series from Mombasa (the type series in CM) and Diani, Kenya, to Dar-
es-Salaam, Tanzania, is quite uniform. The subspecies is not confined to the coastal
plain, however. Two specimens from 6200 ft at Lukwangule, southern end of the
Uluguru mountains (BM[NH]) and seven specimens (MCZ, BM[NH], USNM)
from Kilosa, just west of the Ulugurus and about 225 km SE of the type locality
of keithi, differ only slightly from the coastal series. They are slightly paler and
brighter on the crown, paler and less grayish (more rufescent) on the back, and
paler and more reddish brown on the tail. This can be interpreted as a tendency
toward keithi, but the Kilosa series shows no indication of the marked dorsal
streaking so characteristic of keithi. On the other hand, two specimens (UMMZ
157561-2) from Kidodi (ca. 160 km E of Iringa, ca. 95 km S of Kilosa) appear
to illustrate intergradation with keithi. Both Traylor and Storer have annotated
the label of UMMZ 157561 as heterophrys ; it is indeed closest to this race as
typified by the type series from Mombasa in CM, with dorsal streaking only
slightly better defined than in extreme specimens of heterophrys. UMMZ 1 57562,
however, was annotated by Storer as “too streaked for heterophrys ” and identified
as fischeri by Traylor. It is a very worn 20 May bird and does, indeed, resemble
that race, but from its southern locality it is more plausibly interpreted as rep-
resenting the eastern limit of keithi, which, at that time of year, would be expected
to be in worn Alternate Plumage and inseparable from the perennial plumage of
fischeri.

Farther north, a series from Same (YPM), about the same distance inland as
Kilosa and also separated from the coast by mountains (the Usambaras), is ukam-
ba, with no suggestion of introgression from heterophrys ; the coastal plain at this
latitude is much narrower than that east of Kilosa.

A single 3 August specimen from the north end of Lake Manyara (UMMZ
94961) is particularly interesting. It is just completing a molt into a plumage in
which the crown and back are much like that of the late August specimens of

1987

Parkes— African Warblers

237

keithi. The tertials, however, instead of being black with sharply contrasting buff
edges (varying in degree of rufescence) as in the Basic Plumage of keithi , have
less intensely black centers, and the edgings are darker and less contrasting. In
this respect they resemble those of the perennial plumage of fischeri and Alternate
Plumage of keithi , in that they are not intensely black and the edgings show little
contrast, but the latter are distinctly warmer, less gray in color in the Manyara
bird than in fischeri . Lake Manyara is an area where intergradation between keithi
and fischeri might be expected. Two 26 January LACo specimens are labeled “E.
side Lake Manyara, 60 mi So. of Arusha.” This locality is difficult to interpret,
as Lake Manyara is approximately 60 miles west, not south of Arusha. These two
specimens are typical of fischeri in appearance. Both are moderately worn, and
both are initiating a molt on the throat. There is no molt at this time of year in
keithi (specimens still in worn alternate plumage as late as early March), and Lynes
(1930) stated that, at Tabora (ca. 120 km E of Tura, the type locality of fischeri ),
the months of June through September “seem to more or less cover the general
moulting period.” On the other hand, BM(NH) 1933.6.1 1.1 19, from Ndembezi,
Sinyanga [=Shinyanga], ca. 210 km N of Tura, 23 December 1933, was annotated
by Lynes on 10 October 1934 as “(nearly ad.) S completing m.[oult] to first
S.[ummer plumage]”. Clearly there is much yet to be learned about the molts of
Cisticola chiniana, especially in areas where subspecies typified by seasonal plum-
ages, such as keithi, may intergrade with those having only a perennial plumage,
such as fischeri.

I have seen no specimens from areas of Tanzania where intergradation between
keithi and procera to the south or southeast might be expected.

A series of 8 specimens from the vicinity of Mbeya, southwestemmost Tanzania
(between Lake Rukwa and the Kipengere Range at the north end of Lake Nyasa,
where the map in Hall and Moreau [1970] shows a gap in the range of chiniana;
see map in Ripley and Heinrich 1966, where the lake is spelled “Rukwe”) cannot
be assigned to any known race of Cisticola chiniana. White (1962) included
“South-west Tanganyika (Rukwa)” in the range of C. c. fords Lynes; this is based
on his earlier brief footnote (White, 1960), which reads, in its entirety: “C. c.
fords occurs at rukwa [sic]; one specimen examined since writing this note.” I
have been unable to discover where this specimen is housed. Britton (1980)
assigned “SW Tanzania between Lakes Rukwa and Tanganyika” to fords. It is
not clear whether this was taken on White’s authority (based on one specimen),
or additional specimens seen by Britton or one of his collaborators. In any case,
the Mbeya series is utterly unlike fords specimens taken at the same time of year
in inland Angola, where the type locality of fords lies. The Mbeya series is most
closely related to keithi and fischeri. Lake Rukwa is so close to the collecting
localities of the YPM series as to make it difficult to believe that two such different
races of chiniana can be found so close together. There are two possible expla-
nations. One is that White erred in extending the range of fords as far east as
Lake Rukwa. I cannot believe, however, that White would have identified as fords
a Rukwa specimen that resembled the Mbeya series. An alternative explanation
is that fords is, as originally described by Lynes (1930), a full species, overlapping
in range with C. chiniana subsp. in the vicinity of Lake Rukwa. Certainly fords
is distinct in appearance; it differs from all races of chiniana in having the white
of the underparts greatly reduced, with the gray area of the sides and flanks more
extensive medially, even meeting or almost so across the breast (suggested in
Lynes’s plate 1 1). It completely lacks streaks on the back, resembling in this respect

238

Annals of Carnegie Museum

vol. 56

the geographically distant C. c. heterophrys. The latter, however, has the typical
chiniana amount of white on the underparts, and also has a fairly distinct su-
perciliary line, lacking in fords. The differences between fords and the possibly
sympatric population of chiniana are much greater than between fords and het-
erophrys. Against the theory that fords is not a race of chiniana is the later paper
of Lynes (1934^:24), in which, after extensive experience with fords in the field,
he stated . . now that it is known in life, [it] proves to be no more than a race
of chiniana. There is no mistaking the chiniana behavior of the bird . . . Writing
of the two plain-backed forms fords and heterophrys in relation to streaked-backed
races of chiniana , Traylor (letter of 6 November 1986) stated that “people who
know them in the field consider them identical behaviorally.” It is obvious that
White’s (1960) statement about the gap in southwestern Tanzania of our knowl-
edge of the forms of Cisdcola chiniana is still largely true; clearly more collecting
needs to be done. Meanwhile, the distinctive population represented by the YPM
series may be called:

Cisticola chiniana mbeya , new subspecies

Holotype.— YPM 93268, adult male just completing prealtemate molt. Col-
lected ca. 20 mi ENE of Mbeya, Southern Highlands Province, Tanzania (elevation
1800 m), 28 December 1962, by Gerd Heinrich (field no. 36151).

Diagnosis. — “Summer” (= Alternate) plumage similar to those of keithi and
fischeri dorsally, but averaging slightly darker, especially on the crown. Underparts
in general purer white, without buffy wash. Sides and flanks much grayer, with
buff tinge confined to posteriormost flanks. “Pectoral patch” of Lynes (1930) (a
darker pigmented area near the anterior end of the pigmented sides, extending
farther medially) more prominent. Differs from procera, the next race to the
southeast, in being much less rufescent dorsally, including crown, back, wing
edgings, and tail; dorsal streaks much heavier, flanks grayer, and pectoral patch
more prominent. All of the series of mbeya is in late stages of the molt, and
several, including the holotype, have a few worn feathers remaining that indicate
that this race has a bright “winter” plumage like that of keithi, and does not
exhibit the “perennial mode” of Lynes.

Range. — Kno wn only from the vicinity of Mbeya and Chimala, Southern High-
lands Province, southwestern Tanzania (elevation 1400-1800 m).

Etymology. —Named for the type locality.

Apalis jacksoni

White (1962) recognized only two subspecies of this species, the nominate race
and the distinctive A. j. bambuluensis Serle from the Bamenda highlands of
Cameroon. Traylor (1986) also admitted A. j. minor Ogilvie-Grant for the isolated
population from the lowlands of southern Cameroons; this race was also recog-
nized by Bannerman (1953) and by Mackworth-Praed and Grant (1973), although
the latter authors apparently did not examine specimens; Louette (1981) followed
White in considering the Cameroon lowland population inseparable from nom-
inate jacksoni . I have not been able to examine specimens to verify the alleged
narrower black throat patch of minor, but the measurements presented by Ogilvie-
Grant (1917) in his original description certainly suggest that minor is indeed
significantly smaller than jacksoni, with (in small series) no overlap in either wing
or tail measurements.

The population of northern Angola was described by Meise (1958) as A. j.
albimentalis, characterized as having more white on the chins of females and

1987

Parkes African Warblers

239

generally lighter color (but with deeper golden underparts). Traylor (1962) com-
pared two adult males, an adult female, and a juvenile from Angola with four
males and seven females from Kenya, and found the only difference between the
two series to be slightly darker, more greenish (not deeper gold) underparts in the
Angola males. He later (Traylor, 1986) formally synonymized albimentalis with
nominate jacksoni . I have examined the two females from N’Dalla Tando, Angola,
in the AMNH collection, the same specimens mentioned by Chapin (1953:297);
three of the four FMNH Angola specimens seen by Traylor (1962); and a series
of three adult males and four adult females from YPM; see Ripley and Heinrich
(1966) for localities and map. This is a larger series of Angola specimens than
has been assembled by any previous worker. Comparison of these specimens with
an ample series from East Africa indicates that albimentalis is a recognizable
subspecies, although the alleged character that gave it its name is not valid. In
both sexes (especially in males), the underparts of Angola birds are brighter yellow
(< contra Traylor) than in those of East Africa, with less invasion toward the midline
of the duskier dolor of the flanks. The throat patch (black in males, gray in females)
extends slightly farther caudad, especially in females; depending on preparation,
this gives the appearance of the yellow of the breast extending slightly farther
anteriorly on each side of the throat patch. The green of the dorsum of Angola
specimens is distinctly brighter and yellower (less grayish) in both sexes. The gray
crown of Angola males is paler than in Kenya jacksoni, but little different from
that of Uganda males (see beyond); the crown color of females is paler than in
either Kenya or Uganda specimens. The extent of the black or blackish face mask
is variable in both sexes.

Like so many other African highland birds, Apalis jacksoni has isolated pop-
ulations in Kenya and in western Uganda and adjacent Rwanda and Burundi (see
Hall and Moreau, 1970: map 216). The western birds differ in series from true
jacksoni of Kenya and eastern Uganda (type locality Mt. Elgon) as follows:

Males: crown, wing coverts and tertials paler gray, less blackish; black facial
mask averaging less extensive, generally not extending above eye, whereas the eye
is completely surrounded by black in Kenya males; dorsum slightly brighter,
yellower green.

Females: difference in color of crown, wings and back less noticeable but present;
the area of the male’s mask (ear coverts to lores) tending to be uniform with the
crown rather than contrastingly blacker.

Color of the underparts is too variable to permit its use in identifying individual
specimens, but there is a tendency in Kenya males for the duskier color of the
flanks to expand toward the midline, reducing the amount of pure yellow.

The characteristics of the western Uganda specimens strongly suggest an ap-
proach to albimentalis of Angola. However, a series of eight males and one female
from Baraka, west of northern Lake Tanganyika, Zaire (AMNH), is not the same
as the form from western Uganda, but is, if anything, even darker than Kenya
jacksoni. Chapin (1953:297) has already commented briefly on this series.

It is apparent that Traylor’s and especially White’s treatment of Apalis jacksoni
is oversimplified, and that a full scale revision, bringing together all available
material from throughout the range of the species, is highly desirable.

Apalis porphyrolaema

White (1962) and Traylor (1986) synonymized Apalis affinis Ogilvie-Grant,
from Ruwenzori, with the nominate race from the Kenya highlands, but it was
recognized as a valid race of porphyrolaema by Jackson (1938) and Chapin (1953:

240

Annals of Carnegie Museum

vol. 56

287). The only color character mentioned for affinis has been its darker reddish-
brown throat patch. Comparison of the composite AMNH/CM series of three
adults and one juvenile from Ruwenzori with 19 adults and two juveniles of
porphyrolaema from Kenya supports recognition of affinis\ an even better char-
acter than the throat patch is the color of the gray underparts— much paler (whiter)
in affinis, thus accentuating the slightly darker throat. There is no discemable
difference in the color of the upperparts.

CM 145902, an adult female, was collected on 3 August 1960 by A. C. Twomey
and John Williams in the Impenetrable Forest, Kigezi, Uganda. It is a typical
example of affinis, extending the known range of that subspecies slightly to the
south.

Apalis cinerea

Several Fernando Poo endemic subspecies are also resident on Cameroon
Mountain, on the adjacent mainland. Bannerman (1939) specifically stated that
this was not true of Apalis cinerea sclateri (Alexander). He considered specimens
from Cameroon Mountain to be the same as A. c. cinerea (Sharpe) from East
Africa (type locality Mt. Elgon). Those from elsewhere in Cameroon he identified
as A. c. funebris, a supposedly dark race that he had described himself three years
earlier. Serle (1950) compared a series of 30 skins from “British Cameroons” with
the BM(NH) material, including types, of cinerea, funebris, and sclateri. He showed
that “ funebris ” was based on a worn, dirty holotype, and that his new series of
topotypes was not separable from cinerea. He called all Cameroon birds A. c.
cinerea, and in this he was followed by White (1962), Louette (1981), and Traylor
(1986). Of Searle’s 30 specimens, however, only one was from Cameroon Moun-
tain. As he stated that there was considerable variation in his series in the amount
of buff on the underparts and in the shade of brown on the head, it is highly likely
that the distinctiveness of his solitary non-inland specimen escaped him.

The Carnegie Museum of Natural History has two specimens of Apalis cinerea
(CM 106345, 106362) from Buea, on the slope of Cameroon Mountain. These
are completely outside the range of variation in the combined CM and AMNH
East African series of 87 A. c. cinerea. On the other hand, they agree exactly with
AMNH 598672 from Fernando Poo (sclateri) in the intensity of buff color on the
underparts, and in having the crown much grayer, less brown, resulting in less
contrast between the crown and the slate-gray black. In most East African spec-
imens, the area below the eye from the ear coverts to the base of the bill is
noticeably blacker than the crown, resulting in a masklike appearance. This is
much less conspicuous in the two Buea and one Fernando Poo specimens.

After the above comparisons were made at AMNH, I examined two specimens
from the FMNH, taken at 5900 ft elevation on Cameroon Mountain. These are
somewhat intermediate, but are clearly closer to sclateri than to cinerea . FMNH
95803 has the heavily pigmented underparts typical of sclateri, but in the contrast
of brown cap with gray back it approaches cinerea. FMNH 95802 is less heavily
pigmented below, matching a few extreme specimens of cinerea, but has the
grayish, non-contrasting cap of sclateri. As in the Buea and Fernando Poo spec-
imens, there is little or no indication of a dark facial mask at the lower edge of
the brown cap.

On the basis of the specimens examined, I have no hesitation in assigning the

1987

Parkes— African Warblers

241

Cameroon Mountain population of Apalis cinerea to A. c. sclateri and not to the
nominate race.

Sylvietta leucophrys

Chapin (1953) and other authors have generally recognized two subspecies of
this little warbler of montane forests. The populations of the Kenya highlands
and easternmost Uganda (type locality Mt. Elgon, on the Kenya/Uganda border)
and that of the Ruwenzori Mountains of the Uganda/Zaire border are assigned
to the nominate race. The name S. 1. chloronota Hartert (type locality forest NW
of Baraka, near Lake Tanganyika, Zaire) is used for the populations from “Moun-
tains west of Lake Edward, the Kivu Volcanoes, and highlands of Kigezi District,
south to the vicinity of Baraka and to the Kungwe-Mahare highlands east of Lake
Tanganyika” (Chapin, 1953:252). In addition, White (1962) and Traylor (1986)
considered the well-marked S. chapini Schouteden of the Lendu Plateau, Zaire,
which lacks the distinctive white superciliary of S. leucophrys, as a subspecies of
the latter species. Study of the combined AMNH/CM series of 62 skins of Sylvietta
leucophrys (excluding chapini) indicates that some modification needs to be made
in the range statements.

Specimens of this species from the areas attributed to S. 1. chloronota by Chapin
(1953:252) are by no means uniform. The subspecies is generally said to be
characterized by a greenish rather than brown dorsum, and by having the brown
of the postocular line extending downward on the cheeks. The latter character is
not as consistent as is back color. In Kenya specimens of leucophrys the postocular
line is variable, but usually rather broad. In about one-third of the specimens,
the adjacent cheeks are slightly stained with brown. Among six topotypes of
chloronota from Baraka, the postocular line is broad, with a variable extension
of brown onto the cheeks. The two specimens with the least development of this
character state could be matched with extreme examples in the Kenya series.
There is an additional, previously unmentioned character for distinguishing chlo-
ronota and leucophrys. In the latter, the posterior flanks are pale gray, contrasting
abruptly with the greenish yellow under tail coverts. In chloronota the gray of the
flanks is much mixed with greenish on the posterior half of the body, leading to
the color of the under tail coverts.

Typical chloronota occurs from Baraka north, west of Lake Tanganyika, the
Ruzizi River, and Lake Kivu. As mentioned by Chapin (1953:252), a specimen
he collected at 8 1 00 feet on Mt. Nyemilima, northwest of Lake Edward, is referable
to chloronota. However, two AMNH specimens from “Kagera, Kivu, 2400 m”
and one from “NW slope Mt. Millens, Kivu Distr., 7900',” appear to be inter-
grades between the two subspecies. The highly variable series from Kigezi (4
AMNH, 4 CM), an area attributed to chloronota by Chapin, must also be con-
sidered to be intergrades, a status already suggested by the fact that Stuart Keith
and John Williams independently labeled their specimens as “ chloronota ” in
quotation marks.

A series of nine specimens collected by W. Doherty from October 1900 to
March 1901 (AMNH) at “Escarpment, Kikuyu Mts., 8000-9000',” appears to
differ from other Kenya material of leucophrys in being very heavily instead of
lightly washed with brown on the underparts. However, they are poorly made,
understuffed skins and are apparently stained, so the significance of this color

242

Annals of Carnegie Museum

vol. 56

difference is dubious. Specimens from nearby Molo do not differ from those from
other localities in Kenya.

As mentioned above, the population of this species in the rather isolated Ru-
wenzori range has been attributed to the nominate race. A series of nine specimens
from Ruwenzori and the nearby Mpanga Forest shows that this population, not
unexpectedly, is separable from leucophrys. It may be called:

Sylvietta leucophrys arileuca , new subspecies

Holotype. — CM 145398, adult female, from the Mpanga Forest, Fort Portal,
Toro, Uganda, elevation 5000 feet, collected 20 July 1960 by A. C. Twomey and
John Williams (field no. 1094).

Diagnosis. — Differs from both nominate leucophrys and chloronota in having
a much broader superciliary line, which is pure white rather than slightly stained
with brownish (especially at the posterior end), and extends farther posteriorly.
The underparts are paler and purer gray than in any other population, and the
midabdominal area is pure white or nearly so rather than the same gray as the
rest of the underparts. The back color is somewhat variable, but tends to be closer
to the greenish of chloronota than to the brown of leucophrys. The postocular line
is narrow; all but one of the eight adult specimens have this line narrower than
in any other specimens examined.

The CM series includes 7 specimens of leucophrys, 4 of leucophrys-chloronota
intergrades, and 3 of arileuca that have soft-part color annotations on the labels.
The bills of the leucophrys series are annotated as “flesh-brown” (J. Williams) or
“grey horn” (A. Forbes-Watson). The intergrades are annotated as “dark horn,
lower mandible flesh” or “flesh-horn, lower mandible paler” (Williams). The
arileuca specimens are labeled “bill horn [or dark horn], lower mandible flesh-
white” (Williams). The dried bills of the leucophrys series support the lack of
mention on their labels of a paler lower mandible, and the arileuca series has the
palest bills of all; even the one specimen whose upper mandible is described as
“dark horn” has both mandibles paler than those of any of the leucophrys or
leucophrys-chloronota intergrade series. It appears likely that a paler bill is another
subspecific character of arileuca.

Range.— Known only from the Ruwenzori range on the Zaire/Uganda border
and the nearby Mpanga Forest, south of Fort Portal, Toro, Uganda.

Etymology. — From the Greek, ari-, meaning very, and leukos, meaning white,
referring to the broad white superciliary line and white midabdomen that char-
acterize this subspecies.

Acknowledgments

I am grateful to the staff of the Department of Ornithology, American Museum of Natural History,
for their hospitality during this study. I am especially indebted to G. Stuart Keith for providing
assistance, advice, and reference works. He also gave me access to his collection from East Africa,
housed at the AMNH but as yet uncatalogued. Melvin A. Traylor, Jr., supplied advice and key
references on Cisticola chiniana. Specimens were borrowed from the British Museum (Natural History)
through the courtesy of Graham S. Cowles and Peter R. Colston; Field Museum of Natural History
(Melvin A. Traylor, Jr.); University of Michigan Museum of Zoology (R. B. Payne); Peabody Museum
of Natural History (Eleanor Stickney); Museum of Comparative Zoology (Raymond A. Paynter, Jr.);
Academy of Natural Sciences, Philadelphia (Frank B. Gill); Los Angeles County Museum (Ralph W.
Schreiber); and United States Museum of Natural History (James Dean). The manuscript was critically
read by Melvin A. Traylor, Jr., and Emil K. Urban.

1987

Parkes— African Warblers

243

Literature Cited

Bannerman, D A. 1939. The birds of tropical West Africa, voh 5. Crown Agents for the Colonies,
London, xliii + 485 pp.

— — — . 1953. The birds of West and Equatorial Africa, vol 2. Oliver and Boyd, Edinburgh, pp. i-
viii + 797-1526.

Britton, P. L. (ed ). 1980. Birds of East Africa: their habitat, status and distribution. East Africa
Natural History Society, Nairobi, xiv + 27 1 pp.

Chapin, J. P. 1953. The birds of the Belgian Congo, part 3. Bulletin American Museum of Natural
History, 75A: 1-821.

Friedmann, H. 1937. Birds collected by the Childs Frick Expedition to Ethiopia and Kenya Colony.
Part 2.— Passeres. Bulletin United States National Museum, 153, xii + 506 pp.

Hall, B. F., and R. E. Moreau. 1970. An atlas of speciation in African passerine birds. British
Museum (Natural History), London, xv + 423 pp.

Humphrey, P. S., and K. C. Parkes. 1959. An approach to the study of molts and plumages. Auk,
76:1-31.

Jackson, F. J. I. 1938. The birds of Kenya Colony and the Uganda. Protectorate, vol. 2. Gumey
and Jackson, London, pp. i-viii, 545-1134.

Louette, M. 1981. The birds of Cameroon. An annotated checklist. Verhandelingen van de Kon-
inklijke Academic voor Wetenschappen, Letteren en Schone Kunsten van Belgie, 43, no. 163:1—
295.

Lynes, H. 1930. Review of the genus Cisticola. Ibis, 1930, supplement: ii + 673 pp.

. 1934a. Lynes -Vi ii cent Tour in Central and West Africa in 1930-31. —Part II. Ibis, 1934:
1-51.

. 1 934Z?. Contribution to the ornithology of southern Tanganyika Territory. Journal fur Or-
nithologie, 82, Sonderheft: 1-147.

Mackworth-Praed, C. W., and C. H. B. Grant. 1973. Birds of West Central and Western Africa,
vol. 2. Longman, London, 818 pp.

Meise, W. 1958. Tiber neue Hiihner-, Specht- und Singvogelrassen von Angola. Abhandlungen und
Verhandlungen des Naturwissenchaftlichen Vereins in Hamburg, N. F. (1957), 2:63-83.

Ogilvie-Grant, W. R. 1917. Remarks on some recent collections of birds made by Mr. G. L. Bates
in Cameroon. Ibis, 1917:72-90.

Parkes, K. C. 1 980. Collections of African birds in Carnegie Museum of Natural History. Proceedings
IV Pan African Ornithological Congress, pp. 51-56.

Keichenow, A. 1908. Neues aus Ostafrika. Omithologische Monatsberichte, 16:1 19.

Richmond, C. W. 1897. Descriptions of ten new species of birds discovered by Dr. W. L. Abbott
in the Kilimanjaro region of East Africa. Auk, 14:154-164.

Ripley, S. D., and G. H. Heinrich. 1966. Additions to the avifauna of northern Angola II. Postilla,
No. 95:1-29.

Serle, W. 1950. A contribution to the ornithology of the British Cameroons. Ibis, 92:602-638.

Sharpe, R. B. 1901. New birds from East Africa. Bulletin of the British Ornithologists'’ Club, 12:
35-36.

Traylor, M. A. 1962. Notes on the birds of Angola, passeres, Publicafoes Culturais da Companhia
de Diamantes de Angola, Lisboa, 58:53-142.

. 1986. African Sylviidae. In Check-list of birds of the world, vol. 1 1 (E. Mayr and G. W.
Cottrell, eds.), Museum of Comparative Zoology, Cambridge, Mass., xii + 638 pp.

Vincent, J. 1944. A new race of Cisticola from Portuguese East Africa. Bulletin of the British
Ornithologists5 Club, 64:63-64.

White, C. M. N. 1960. Notes on some savanna species of the genus Cisticola. Bulletin of the British
Ornithologists5 Club, 80:128-130.

■ 1962. A check list of the Ethiopian Muscicapidae (Sylviinae). Parts II and III. Occasional

Papers National Museum of Southern Rhodesia, no. 26B:653-738.

'

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 14, Pp. 245-251 28 August 1987

SPERGENASPIS: A NEW CARBONIFEROUS TRILOBITE
GENUS FROM NORTH AMERICA

David K. Brezinski1
Research Associate, Section of Invertebrate Fossils

Abstract

A new trilobite genus, Spergenaspis, is erected based upon three Mississippian trilobite species from
the midcontinent of the United States. Two of the species, S. easleyi, new species, from the Chouteau
Formation of Missouri, and S. salemi, new species, from the Salem Limestone of Indiana are previously
undescribed. The third species, S. mauvaisensis (Hessler) is reassigned from the genus Richterella and
is known from the Upper Osagean or Lower Meramecian of Illinois.

Known occurrences of Spergenaspis suggest that species of this genus inhabited similar environ-
mental settings. Lithologies from which Spergenaspis is recovered vary from sandy lime packstone
to oolitic to crinoidal lime grainstone. Interpreted depositional environments for these lithologies
include transgressive shallow subtidal and sand shoal.

Introduction

Hessler (1965) erected the Mississippian trilobite genus Richterella and based
it on three species known from the midcontinent region of North America. One
of these species, Richterella mauvaisensis Hessler, differs from the diagnosis of
the genus in that it exhibits considerably less vaulting with a bell-shaped glabella,
possesses a broad concave-up frontal area, lacks a straight section in the facial
sutures from e to f, and possesses a distinct subdivision of the pygidial ribs into
anterior and posterior bands. Recent recognition of two previously undescribed
Mississippian trilobite species bearing characters similar to R. mauvaisensis Hess-
ler justifies their being grouped together as a distinct genus, Spergenaspis. These
features, which characterize R. mauvaisensis Hessler, can best be grouped under
the subfamily Linguaphillipsiinae as outlined by Hahn and Hahn (1972).

Members of the trilobite subfamily Linguaphillipsiinae are not known from the
Lower Carboniferous of North America even though they are relatively wide-
spread geographically and stratigraphically in Europe and Asia (Hahn and Hahn,
1973, 1982; Engel and Morris, 1975; Brauckmann, 1978; Kobayashi and Hamada,
1980). Assignment of the new genus Spergenaspis to the subfamily Linguaphil-
lipsiinae marks the first documentation of this subfamily in North America.
Moreover, recovery of a species of Spergenaspis from the base of the Chouteau
Formation (Middle Kinderhookian) indicates that immigration into North Amer-
ica was concomitant with the Kaskaskia onlap (Brezinski, 1986a). That these
representatives have previously gone unnoticed may be the result of either their
scarcity or their paleoecology. The latter of these two possibilities will be discussed
below.

Terminology utilized in this paper is similar to that proposed by Harrington
(1959) and Richter and Richter (1949). Specimens illustrated in this study are
reposited at The Carnegie Museum of Natural History (CMNH). Type specimens

1 The Maryland Geological Survey, 2300 St. Paul Street, Baltimore, Maryland 21218.
Submitted 9 February 1987.

245

246

Annals of Carnegie Museum

vol. 56

of Spergenaspis mauvaisensis (Hessler) examined for this study are reposited with
the Illinois Geological Survey.

Stratigraphic Distribution

The known species of Spergenaspis are found in Middle Kinderhookian through
Lower Meramecian strata in the midcontinent region of the United States. The
oldest species, Spergenaspis easleyi, new species, was recovered from the basal
strata of the Chouteau Formation (Middle Kinderhookian) of central Missouri
(locality 1 of Brezinski, 1986a). Canis (1968) found that these strata are assignable
to the Siphonodella quadruplicata-S. crenulata conodont zone, which is roughly
equivalent to the Cu IIa ammonoid zone of Toumaisian rocks of Europe. The
next youngest species, Spergenaspis mauvaisensis (Hessler), was found in Upper
Osagean or Lower Meramecian strata of Scott County, Illinois (see Hessler, 1965,
for locality information). The youngest species, Spergenaspis salemi, new species,
was collected from several localities in the Salem Limestone (Meramecian) of
southwestern Indiana. Additionally, two pygidia, from the Keokuk Limestone
(Osagean) of St. Genevieve County, Missouri, tentatively assigned to the genus
Waribole by Brezinski (1986Z?), are here reassigned to the new genus.

Paleoecology

Although Spergenaspis exhibits a considerable stratigraphic range, its restricted
geographic distribution parallels many other trilobite genera known from this time
interval. As a result, Kobayashi and Hamada (1980) were able to discern several
distinct trilobite realms composed of numerous provinces. One of the most distinct
of these faunal provinces is the midcontinent area of North America, where no
less than 1 0 of 1 8 known genera are endemic. Studies by Brezinski (1986a, 1 98 6b)
indicate that this endemism is the result of trilobite genera becoming narrowly
adapted to specific environmental conditions, thus precluding emigration to other
areas which might not provide the precise conditions necessary for habitation.
This niche-specialization is indicative of stenotopic species (Valentine, 1972).

All presently known occurrences of Spergenaspis appear to be found in rocks
that represent similar environmental settings. S. easleyi from the Chouteau was
recovered from a sandy wackestone lithology which is interpreted as having been
deposited in a shallow subtidal transgressive environment (King, 1980; Brezinski,
1986a). Coarse tuberculate ornament on several of the associated trilobite species
may indicate agitated conditions (McNamara and Fordham, 1981). The lithology
in which S. mauvaisensis (Hessler) occurs, is a coarse-grained crinoidal packstone
with lenses of crinoidal grainstone, also suggestive of high-energy conditions of
deposition. S. salemi was recovered from a poorly cemented, coarse-grained,
oolitic, gastropod grainstone which Donahue (1967) interpreted as having been
formed in a sand-bar shoal environment. The two pygidia of the undetermined
species of Spergenaspis from the Keokuk Limestone of Missouri were collected
from well-sorted crinoidal lime grainstones associated with oolitic grainstone
strata. These occurrences of Spergenaspis indicate that members of this genus
probably inhabited high-energy environmental settings such as shoal sands or
nearshore tidal sands, although some transport is probable. The low relief and
lack of surface ornamentation shown by members of this genus may be the result,
in a large part, of their ecological affinities. Similar features are exhibited by
trilobite species from the Gilmore City Limestone of Iowa (Kinderhookian-Os-
agean). The Gilmore City Limestone, like the examples discussed above, consists

1987

Brezinski— Spergenaspis

247

Fig. 1 . —Comparative reconstructions of, cranidia of a, Spergenaspis easleyi new genus and new species;
b, Spergenaspis salemi new genus and new species; c, Spergenaspis mauvaisensis (Hessler), all ap-
proximately x3.5.

of an oolitic grainstone of shoal-water origin. The lack of surface features, and
low profile to the exoskeletons of these species are interpreted as adaptations to
reduce friction created by passing currents, thus making them more hydrodyn-
amically stable.

Systematic Paleontology

Family Proetidae Salter, 1864
Subfamily Linguaphillipsiinae Hahn and Hahn, 1972
Genus Spergenaspis , new genus

Type species.— Spergenaspis salemi, new genus and new species.

Other species assigned. —Spergenaspis easleyi, new genus and new species, Sper-
genaspis mauvaisensis (Hessler, 1965).

Diagnosis. —Smooth isopygous genus with cranidium possessing a broad con-
cave-up fixigenae, bell-shaped glabella, broader behind y than in front of it, and
large palpebral lobes, located at the base of the glabella. Pygidium smooth with
wide border flange; ribs divided by pleural furrow into bands of equal width.

Comparisons. — Genera similar to Spergenaspis and with which it might be
confused are Warihole Richter and Richter, 1926, Gitarra Gandl, 1968, Lingua-
phillipsia Stubblefield, 1948, Pseudowarihole Hahn and Hahn, 1967, and Griffi-
thidella Hessler, 1965. Waribole can be distinguished from Spergenaspis in that
the former possesses a conical glabella lacking a constriction at y, more anteriorly
located eyes, and narrower palpebral lobes. A somewhat similar genus Lingua-
phillipsia can be differentiated from Spergenaspis by the greater anterior elon-
gation of the glabella, more anterior location of the palpebral lobes, by the much
smaller width to length ratio of the pygidium, by the commonly well-developed
surface sculpture, and structure of the pygidial ribs on the former. The genus
Gitarra differs from Spergenaspis in that the former has a more strongly con-
stricted, guitar-shaped glabella, very large eyes, and a shorter, broader pygidium
with well-defined segmentation and ribs that extend to the margin. Spergenaspis
differs from Pseudowarihole (. Pseudowarihole ) in that Spergenaspis is isopygous,
lacks a well-defined border furrow to the cephalon, has more posteriorly located
(with respect to the glabella), larger, palpebral lobes, stronger constriction of the
glabella at 7, and a longer and broader pygidium possessing a well-defined border.
Spergenaspis differs from Pseudowarihole (Geigibole) in that the latter is not

248

Annals of Carnegie Museum

vol. 56

isopygous, exhibits a very short and broad, subtriangular pygidium, a broad cra-
nidium, much larger, more anteriorly located palpebral lobes and eyes, and a
narrower anterior fixigenae. Pygidia of Griffithidella depressus var. depressus (Gir-
ty) (see Hessler, 1965, pi. 38, figs. 1-11, 14) are quite similar in many character-
istics to those of Spergenaspis . Inasmuch as no cranidia of G. depressus var.
depressus are known, it is questionable whether these pygidia are truly assignable
to the genus Griffithidella.

Spergenaspis salemi , new genus and new species
Fig. 2A~I, M

Holotype.— An incomplete cephalon from the Salem Limestone at a railroad
cut near Spergen Hill, Washington County, Indiana. NW lA, SE lA, sec. 24, T2N,
R4E, Salem Quadrangle. Collected by R. T. Garney. CMNH 34433.

Paratypes. — An external mold of a cranidium and two pygidia from the Cleve-
land Quarry, Monroe County, Indiana. SE Vi, NW lA, sec. 20, T7N, R4E, Clear
Creek Quadrangle. Collected by H. B. Rollins. 1 cranidium and 6 pygidia from
the Spergen Hill railroad cut collected by R. Garney. CMNH 34434-34437.

Diagnosis. — Species of Spergenaspis with a subcylindrical glabella, slightly wid-
er at the base than at 7, palpebral lobes very large, anterior facial sutures recurved
from 7 to (3. Pygidium semicircular in outline.

Description. — Cephalon semicircular in outline, very low in relief and vaulting. Glabella approxi-
mately 1.50 times as long as the maximum width, subcylindrical and rounded anteriorly, widest
between the eyes, tapering slightly to the 3 p glabellar furrow, then nearly parallel sided to faintly
diverging anteriorly. 1 p furrow broad, shallow, and nearly obsolete, all other glabellar furrows very
faint to obsolete on the dorsal exoskeleton. Occipital lobe of nearly even width only slightly wider at
axis. Occipital furrow narrow and arcuate, becoming shallower near the axial furrow. Palpebral lobes
large and crescentic in outline, located very near the base, with the 1 p furrow intersecting the dorsal
furrow anterior to the palpebral midline (longitudinal). Facial sutures strongly divergent from 7 to 0,
broadly rounded through ft. Preglabellar field gently concave-up (sagittal) to flat (exsagittal). Eyes
narrow and near vertical, reniform. Border becomes more concave-up anteriorly, extends posteriorly
onto a short sharp genal spine which is 0.33 times the total librigenae length.

Pygidium low in relief and semicircular in outline, 1.48 to 1.61 times as wide as long. Axis 0.37
times as wide as maximum pygidial width, composed of 1 2 to 13 smooth rings. Ring furrows become
increasingly obscure and shallow near the posterior terminus. Pleural fields mildly convex, composed
of 6 to 9 ribs, the posterior two of which are difficult to discern on the dorsal surface, but are evident
on the internal surface. Each rib exhibits an anterior and posterior moiety of nearly equal width.
Pleural furrows terminate at the border. Border from 0.20 to 0.28 the total pygidial length (sagittal),
broader near the posterior terminus and becoming narrower anteriorly.

Discussion. — Similarity in morphology suggests that Spergenaspis salemi is
closely related to Spergenaspis mauvaisensis (Hessler). This is especially evident
in the overall plan to the cephalon and pygidium, but differs in several respects.
Firstly, S. mauvaisensis (Hessler) has a straight diverging section on the anterior
facial sutures from 7 to 0, more acutely rounded facial suture at (3, a bell-shaped
glabella which is relatively wider between the eyes and more constricted at the
base and 7. The pygidium of S. mauvaisensis is more parabolic in outline and
possesses a smaller width to length ratio. Griffithidella depressus var. depressus
(Girty) exhibits similar numbers of rings and ribs as S. salemi. However, it can
be differentiated from the latter by the slightly shorter, broader pygidial outline
and the poorly defined border.

Age and distribution. — Known only from the Salem Limestone, Meramecian,
of southern Indiana.

Etymology. — This species is named for the Salem Limestone from which the
type material was collected.

1987

Brezinski — Spergenaspis

249

Fig. 2.-A-I, M, Spergenaspis salemi new genus and new species. A-C, dorsal, oblique, and lateral
views of partially exfoliated holotype cephalon, CMNH 34433, x4; D, external mold of paratype
cranidium, CMNH 34434, x3; E, F, dorsal and oblique views of unexfoliated paratype cephalon,
CMNH 34435, x2.5; G, H, M, dorsal, posterior, and lateral views of paratype pygidium, CMNH
34436, x 3.5; I, partially exfoliated paratype pygidium, CMNH 34437, x3. J-L, Spergenaspis easleyi
new genus and new species, dorsal, oblique, and anterior views of unexfoliated holotype cephalon,
CMNH 34438, x 5.

250

Annals of Carnegie Museum

vol. 56

Spergenaspis easleyU new genus and new species
Fig. 2 J-L

RichterelW. sp. BREZINSKI, J. Paleont. 1986, 60:873, table 1, fig. 5.

Holotype.— A partial cephalon from the basal Chouteau Formation 1.75 km
northwest of Easley, Missouri along the bluffs of the Missouri River, collected by
the author, 1980; CMNH 34438.

Other material. —Only the holotype is known.

Diagnosis. — Species of Spergenaspis with a narrow (transverse) cranidium, gla-
bella narrower at 7 and the base than between the eyes. Frontal glabellar lobe
semicircular in outline, anterior facial sutures only slightly divergent, rounded
from 7 to ($, 7 narrower than e.

Description.— Profile of cephalon low. Glabella smooth and elongate, 1 .76 times as long as maximum
width, constricted at 7, frontal glabellar lobe broadly rounded. Glabellar furrows, except 1 p, very
faint, nearly obsolete, 1 p furrow shallow and faint, intersects dorsal furrow at the anterior half of the
palpebrae, extends posteriorly at an oblique angle to the axis, then forms an obtuse angle and intersects
the occipital furrow at nearly a right angle. Frontal glabellar lobe descends gently into a concave
preglabellar field which is upturned to the margin (sagittal) and becomes flattened to downsloping
(exsagittal). Palpebral lobes relatively large and crescentic, located at the base of the glabella. Occipital
lobe broad and smooth, widest at the axis, very gently downsloping into a narrow, sharp occipital
furrow with an arcuate trace. Eyes not well preserved on holotype, but appear to be vertical and
reniform. Librigenae gently downsloping from the eyes becoming more steeply inclined onto the border
then descending less steeply to the margin. Facial sutures moderately divergent from 7 to 0, broadly
rounded from 7 to a.

Discussion. — Although it is generally inadvisable to propose new species on the
basis of a single specimen, S. easleyi is so markedly differentiable from either S.
salemi or S. mauvaisensis the erection of a species in this case seems warranted.
The narrow, anteriorly elongate glabella, which is constricted at 7, appears to be
a synapomorphic character connecting this genus with ancestral Linguaphillip-
siinae such as Pseudowaribole.

Age and distribution. — Known only from the type locality in the basal Chouteau
Formation (Middle Kinderhookian) of Boone County, Missouri.

Etymology. — Named for the town of Easley in Boone County, Missouri near
where the holotype was collected.

Conclusions

The subgenus Spergenaspis arose during or immigrated into North America
concurrently with the Kaskaskia transgression. Representatives have probably
previously gone unnoticed primarily because: 1) they are not common where they
do occur, 2) they are highly restrictive with respect to the type of environment
in which they are found. This type of habitat restriction may also be the main
reason for the strong endemism exhibited by this genus, inasmuch as members
may have been unable to emigrate owing to unfavorable conditions outside their
local habitat.

Acknowledgments

The holotype and several of the paratype specimens of Spergenaspis salemi were collected by Ronald
T. Gamey from the Spergen Hill locality. Jack Donahue and H. B. Rollins made available additional
bulk samples from the Cleveland Quarry. Lois Kent of the Illinois Geological Survey loaned the type
material of Spergenaspis mauvaisensis (Messier). J. H. Stitt, J. F. Taylor and C. A. Kertis made many
helpful suggestions to improve earlier drafts of this manuscript. Dr. C. Brauckmann provided infor-
mation as to the possible taxonomic assignment of the three species herein assigned to Spergenaspis.

1987

Brezinski-- Spergenaspis

251

Literature Cited

Brauckmann, C. 1 978. Pseudowaribole ( Geigibole ) Gandl 1 968. Senckenbergiana Lethaea, 59: 1-27.
Brezinski, D. K. 1986a. Trilobite associations from the Chouteau Formation (Kinderhookian) of
central Missouri. Journal of Paleontology, 60:870-881.

. 1986 b. Trilobites from the Keokuk Limestone (Mississippian) of Missouri. Annals of Car-
negie Museum, 55:137-143.

Canis, W. 1968. Conodonts and biostratigraphy of the lower Mississippian of Missouri. Journal of
Paleontology, 42:525-555.

Donahue, J. 1967. Depositional environments of the Salem Limestone (Mississippian, Meramec)
of south-central Indiana. Unpublished Ph.D. dissertation. Columbia University, New York,

108 pp.

Engel, B, A., and N. Morris. 1975. Linguaphillipsia (Trilobita) in the Carboniferous of eastern
Australia. Senckenbergiana Lethaea, 56:147-189.

Gandl, J. 1968. Die Trilobiten in Unterkarbon des Frankenwaldes. Senckenbergiana Lethaea, 49:
39-117.

Hahn, G., and R. Hahn. 1967. Zur phylogenie der Proetidae (Trilobita) des Karbons und Perms.

Zoologie Beitrage, no. 13, 303-349.

. 1972. Trilobitae carbonici et permici III. Pp. 336-531, in Fossilium Catalogus 1. Animalia

(F. Westphal, ed.), s’Gravenhage (Dr. W. Junk N.V.), no. 120, 336-531.

— . 1973. Zur Evolution von Linguaphillipsia (Trilobita; Unter-Karbon). Senckenbergiana Le-
thaea, 53:479-515.

1982. Die Trilobiten des belgischen Kolenkalkes (Unter-Karbon). Geologica et Palaeonto-
logica, 15:1 15-124.

Harrington, H. J. 1959. General description of trilobita. Pp. 38-117, in Treatise of invertebrate
paleontology, Part O, Arthropoda 1 (R. C. Moore, ed.), University of Kansas Press, Lawrence,

506 pp.

Hessler, R. R. 1965. Lower Mississippian trilobites of the family Proetidae in the United States,
Part II. Journal of Paleontology, 39: 248-264.

King, D. T. 1 980. Genetic stratigraphy of the Mississippian System in central Missouri. Unpublished
Ph.D. dissertation, University of Missouri, Columbia, 121 pp.

Kobayashi, T., and T. Hamada. 1980. Carboniferous trilobites of Japan in comparison with Asian,
Pacific and other faunas. Palaeontological Society of Japan Special Paper No. 23, 123 pp.
McNamara, K. J., and B. G. Fordham. 1981. Mid-Cautleyan (Ashgill Series) trilobites and facies
in the English Lake District. Palaeogeography, Palaeoclimatology, Palaeoecology, 34:137-161.
Richter, R., and E. Richter. 1949. Die Trilobiten der Erdbach-Zone (Kulm) im Rheinischen
Schiefergebirge und im Harz. 1. Die Gattung Phillibole. Senckenbergiana, 30:63-94.

Valentine, J. W. 1972. Conceptual models of ecosystem evolution. Pp. 192-215, in Models in
paleobiology (T. Schopf, ed.), W. H. Freeman and Company, San Francisco.

INSTRUCTIONS FOR AUTHORS

ANNALS OF CARNEGIE MUSEUM consist of con-
tributions to the natural sciences and anthropology, in
30 by 46 picas format ( 1 27 by 1 95 mm or 5 by 7% inches).
All manuscripts must first be submitted for review to the
Curator in charge of the appropriate scientific Section.
Authors should give particular attention to scientific con-
tent, format, and general style for the ANNALS. Manu-
scripts that do not conform to the style of the ANNALS
will be returned to the author immediately. Every manu-
script will be peer reviewed by at least two outside per-
sons. Authors will be asked to subsidize, if possible, any
or all costs of publication (approximately $ 100/page
printed).

Manuscript style.— Articles should include the follow-
ing items in this order: title page, abstract, text (with
desired headings), acknowledgments, literature cited, ta-
bles, figure captions, and copies of illustrations. All
manuscripts must be typed double-spaced on standard
8I/2 by 1 1 inch white bond paper, with at least one inch
margins all around, and submitted in triplicate— an orig-
inal for the editors, and two review copies. All pages
should be numbered, including tables, literature cited,
and the list of figure captions. Only correspondence qual-
ity or better dot matrix printouts will be accepted; draft
mode is unacceptable.

Title Page. The title, on a separate title page should
be brief, include the animal or plant group involved, and
appear two to three inches below the top margin. Include
the authors name(s) and the affiliations of non-CMNH
authors. In the case of multiple authorship indicate the
address to which proofs should be sent.

Abstracts. Abstracts should be short, double-spaced
and substantive and included at the head of the first page
of text.

Text. Do not right justify text or break (hyphenate) a
word at the end of a line. Footnotes and acknowledgments
as footnotes are unacceptable. All text is double-spaced.

Tables and figure legends. The list of figure legends
and each table should be typed consecutively on indi-
vidual pages separate from the text. Tables must be dou-
ble-spaced throughout with no vertical lines.

Abbreviations. Refer to the CBE Style Manual for cor-
rect abbreviations. Mammalian dentition: use capital let-
ters (I, C, P, M, D, for incisor, canine, premolar, molar,
deciduous, respectively) and superscript/subscript num-
bers (M2, P3) to designate upper and lower teeth.

Measurements. Metric units should be used, if possi-
ble.

Literature Cited. All references in text must appear in
the Literature Cited section. The data (author, date and
page) in both citations must agree. Do not abbreviate the
titles of periodicals or serials. The following style, dou-
ble-spaced, should be used in Literature Cited:

1) Two authors in an institutional series:

Matthew, W. D., and W. Granger. 1923. The
fauna of the Houldjin Gravels. American Mu-
seum of Natural History Novitates, no. 97:
1-6.

2) Same authors repeated— use three-em dash:

. 1923. The fauna of the Ardyn Obo For-

mation. American Museum of Natural
History Novitates, no. 98:1-5.

3) Same authors plus a third author— repeat all authors:

Matthew, W. D. , W. Granger, and G. G. Simpson.
1 928. Paleocene multituberculates from Mon-
golia. American Museum of Natural History
Novitates, no. 331:1-4.

4) Chapter in an edited volume:

Rausch, R. L. 1 963. A review of the distribution
of Holarctic mammals. Pp. 29-43, in Pacific
basin biography (J. L. Gressitt, ed.). Bishop Mu-
seum Press, Honolulu, Hawaii, xx + 450 pp.

5) Unpublished dissertation:

Smith, J. P. 1976. Review of Eocene mammals.
Unpublished Ph.D. dissert., University of Cal-
ifornia, Berkeley, 302 pp.

6) Book:

White, M. J. D. 1961. The chromosomes. Meth-
uen and Co., Ltd., London, 120 pp.

7) Journal articles with usual volume and issue number:

Anderson, W. I. 1969. Lower Mississippian con-
odonts from northern Iowa. Journal of Paleon-
tology, 43(4):9 16-928.

Illustrations. All illustrations will be called figures, and
are to be numbered in Arabic numerals. Three sets of
illustrations are required, one (original artwork) for re-
production, two for reviewers. Xerox copies of photo-
graphs for reviewers are usually not acceptable but are
adequate for line drawing review copies. All illustrations
must be reducible to a maximum of 127 by 195 mm (30
by 46 picas) without loss of clarity. Line copy should be
designed for reduction to % or lh of actual size. Type-
written illustration copy will not be accepted. Photo-
graphic figures should be submitted at actual reproduc-
tion size.

Two or more small figures should be combined for
reproduction as a unit or plate, but, if feasible, do not
combine halftones and line drawings. Rectangular half-
tone figures should be abutted, without intervening spaces.
The printer will insert narrow white spaces during the
reproduction process. Halftone figures with solid black
backgrounds will not be accepted unless the author agrees
to pay for the printing costs. All plates must have min-
imally one inch borders all around. Each plate should be
given a protective cover and identified on the back side.

Lettering and/or a magnification scale (linear metric
scale) for rectangular halftone figures should be placed
directly on the photo, not in a blank space between pho-
tos. The scale or lettering for closely cropped photos can
be placed in blank areas close to the figure.

Proof— The author should answer all queried proof
marks and check the entire proof copy. Return corrected
page proof with the edited manuscript promptly to the
editors.

If an author chooses to make extensive alterations to
a paper in proof stage the author or appropriate section
will bear the cost. Original manuscripts will not be re-
turned unless requested. Illustrations will be returned to
the author.

ANNALS

of CARNEGIE MUSEUM

THE CARNEGIE MUSEUM OF NATURAL HISTORY

4400 FORBES AVENUE © PITTSBURGH, PENNSYLVANIA 15213

VOLUME 56 20 NOVEMBER 1987 ARTICLES 15-19

CONTENTS

Art. 15. Decompression syndrome in fossil marine turtles

Bruce M. Rothschild 253

Art. 16. Towards a postglacial history of the northern Great Plains: A

review of the paleoecologic problems

Cathy W. Barnosky, Eric C. Grimm and H. E. Wright, Jr. 259

Art. 17. Fossil crab (Decapoda: Brachyura) fauna from the Late

Cretaceous (Campanian-Maastrichtian) Pierre Shale in Bowman

County, North Dakota

Annette B. Tucker, Rodney M. Feldmann, F. D. Holland, Jr.

and Kenneth F. Brinster 275

Art. 1 8. Excavations at the Harney Site slave cemetery, Montserrat,

West Indies David R. Watters 289

Art. 19. Description of skeletal remains from a Black slave cemetery

from Montserrat, West Indies Robert W. Mann,

Lee Meadows, William M. Bass and David R. Watters 3 1 9

Editors, ANNALS, BULLETIN and SPECIAL PUBLICATIONS:

L. Krishtalka
J. L. Carter

Office of Scientific Publications
Phone (412) 622-3287

Communications about subscriptions, purchases of individual numbers, rates,
and changes of address should be sent to:

Publications Secretary

The Carnegie Museum of Natural History

4400 Forbes Avenue

Pittsburgh, PA 15213

ANNALS OF CARNEGIE MUSEUM is published quarterly by The Carnegie Museum of Natural
History, 4400 Forbes Avenue, Pittsburgh, Pennsylvania 15213, by the authority of the Board of
Trustees of Carnegie Institute.

THE CARNEGIE

MUSEUM OF
NATURAL HISTORY

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 15, Pp. 253-258 20 November 1987

DECOMPRESSION SYNDROME
MARINE TURTLES

Bruce M. Rothschild1
Research Associate, Section of Vertebrate Fossils

Abstract

Avascular necrosis of bone occurs in Tertiary and Quaternary marine turtles, as it does in Cretaceous
mosasaurs. Assessment for potential underlying pathophysiology suggests decompression syndrome
as the etiology of this bone pathology. The pathophysiology of decompression syndrome, and pre-
disposing and protective factors are reviewed.

Introduction

Avascular necrosis of the humeral head results in loss of viability (death) of
bone. The devitalized bone often becomes necrotic and liquefied, subsequent to
the loss of vascular supply, so that the stresses of normal use result in fracture of
the surviving subchondral bone (Resnick et al., 1981; Feldman et al., 1981). The
result of such fractures is discrete collapse of the articular surface. This compli-
cation of decompression syndrome, which has been observed in man, was recently
discovered in Cretaceous mosasaurs (Rothschild and Martin, 1987). The present
report describes avascular necrosis in Cretaceous marine turtles and Cenozoic
marine and freshwater Chelonians.

Methods

Cretaceous marine turtle specimens in the collection of the University of Kansas Museum of Natural
History were examined for avascular necrosis of the humeral head, as evidenced by focal subsidence
(e.g. collapse) of the articular surface. Specimens were also examined in the Institut Royal des Sciences
Naturelle des Belgique (Brussels, Belgium), the Red Mountain Museum (Birmingham, Alabama) and
the Auburn University Museum (Auburn, Alabama). Specimens found to have focal subsidence of
the humeral articular surface were sectioned for confirmation of avascular necrosis. A simple coronal
cross-section was adequate for examination. The sections were subjected to radiologic and electron
probe analysis and to bright field, polarizing, fluorescent, and scanning electron microscopic exami-
nation. Polarizing and fluorescent microscopy of the sections revealed trabecular patterns, the dis-
ruption of which characterize avascular necrosis. Scanning electron microscopy revealed ultrastructure
alterations, and permitted precise placement of the electron probe utilized for bismuth analysis.
Standard ionization chamber technique was utilized for assessment of radiation levels.

Results

Examination of Cretaceous marine turtles contemporary with the mosasaurs
revealed a specimen of Desmatocheles (Desmatochelyidae) manifesting avascular
necrosis (Table 1) of the humeral heads. This was seen as focal subsidence within
the articular surface of the intact specimen (Fig. 1 A) and by collapse of the necrotic
zone as seen in cross section (Fig. IB). Evidence of avascular necrosis was also
found in both Eocene ( Eosphargis : Dermochelyidae and Cheloniidae, species
indet.) and Oligocene ( Oligochelone and Cheloniidae, species indet.) turtles. Avas-
cular necrosis was also present in Porthochelys (Toxochelyidae), and Protostega

1 Address: St. Elizabeth Hospital Medical Center, Youngstown, Ohio 44501-1790.

253

254

Annals of Carnegie Museum

vol. 56

Fig. 1 . — Desmatochelys humerus showing A, focal subsidence on articular surface of head (arrows)
and B, collapsed necrotic zone in cross-section of head.

dixie (Protostegidae) the Cretaceous. Avascular necrosis was not observed in
Psephophorus, Lytoloma, and Trionyx, although only limited numbers of humeri
were available for examination. Microscopic examination of sectioned material
revealed only focal linear loss of definition of trabeculae and collapse of sub-
chondral bone. Trabecular architecture was intact and normal extrinsic to the
zone of involved bone. Electron probe analysis revealed no evidence of significant
bismuth exposure. Radiation levels in affected and unaffected specimens were
similar.

Avascular Necrosis in Turtles

Avascular necrosis is an easily recognized pathology, characterized by the ap-
pearance of articular surface collapse. Its presence is confirmed in 1 1% of fossil
Cheloniidae examined and in four other marine turtle families (Table 1). The
differential diagnosis of avascular necrosis in turtles is as previously reported for
mosasaurs: radiation poisoning, bismuth poisoning, and decompression sickness
(Rothschild and Martin, 1987). Lack of variation in radiation levels and absence
of bismuth as determined by electron probe analysis suggest decompression sick-
ness as the etiologic factor. As Protostega shared a common habitat with the
mosasaurs, (evidenced by mosasaur bite lesions in a Protostega carapace (RMM
2255) in the Red Mountain Museum collection), it is possible that Protostega was
subjected to the same pathophysiologic mechanism in development of avascular
necrosis as the mosasaurs. Sudden changes in depth (e.g. rapid surfacing), possibly

1987

Rothschild— Decompression Syndrome in Fossil Turtles

255

Table 1.— Incidence of avascular necrosis of turtle humeral heads.

Taxon

Epoch

Specimens

examined

Frequency %

Cheloniidae

Caretta caretta

Holocene

8

0

Oligochelone rupeliensis

Oligocene

9

11

species indet.

Oligocene

3

33

species indet.

Eocene

15

7

Toxochelidae

Porthochelys laticeps

Cretaceous

1

100

Ctenochelys cf. tenuitesta

Cretaceous

1

0

species indet.

Cretaceous

3

0

Lytoloma gosseleti

Eocene

2

0

Protostegidae

Protostega dixie

Cretaceous

2

50

Desmatochelyidae

Desmatochelys sp.

Cretaceous

4

50

Dermochelyidae

Psephophorus rupeliensis

Oligocene

2

0

Eosphargis gigas

Eocene

4

50

Testudinidae

species indet.

Pleistocene

9

0

Trionychidae

Trionyx levalensis

Eocene

3

0

related to escape from mosasaurs or other predators, may have played a major
role in the development of the lesions.

Decompression Syndrome in Non-Turtle Vertebrates

Decompression syndrome is commonly thought of as a phenomenon dependent
upon breathing an exogenous air supply, but this requirement is not absolute.
Decompression illness has also been noted in man subsequent to breath-holding
dives (Pauley, 1965; Kooyman et al., 1973; Strauss, 1970). Repeated two minute
dives of 20 to 40 m have resulted in the decompression syndrome (Kooyman et
al., 1973). Pauley (1965) reported neurologic manifestations of decompression
syndrome in man after sixty 1 5 to 20 m dives performed over a five hour period.
The decompression syndrome etiology was confirmed by the “dramatic response”
of symptomatology to recompression. Pauley (1965) also reported three similar
cases occurring after 20 “bottom drops” during Norwegian escape-tank training.
Polynesian (Tuamoto) pearl divers manifest an intriguing dichotomy of suscep-
tibility to decompression illness. Those who performed 40 to 50 two minute dives
to greater than 30 m, with three to four minutes between dives, or who participated
in vigorous 4 to 5 hr diving sessions to 50 m developed decompression illness
(Pauley, 1965; Strauss, 1970). Those who maintained at least ten minute intervals
between dives (a different “diving habit” pursued at an adjacent island) did not
develop decompression syndrome. The Ama (Japanese diving women), who dive
to more shallow depths or maintain a longer surface interval, are apparently
unaffected by decompression disease (Anderson, 1966; Strauss, 1970).

Decompression syndrome has also been documented in other vertebrates. Berk-

256

Annals of Carnegie Museum

vol. 56

Table 2.= Physiological mechanisms promoting vertebrate protection from decompression syndrome
(derived from Anderson, 1966; Denison et al., 1971; Strauss, 1970).

Decreased lung surface area for gas exchange
—Small lung volume
—Thoracic wall collapse
—Exhalation prior to descent

—Cartilage/muscle airway support (minimizes air trapping)
Augmented, targeted circulation
—Intracardiac blood shunting
—Increased intravascular volume
—Increased hematocrit
—Decreased red blood cell size
—Elasticity of blood vessels and sinuses
—Anuria- water conservation
—Vasoconstriction (e.g. of muscle beds)

Protection from hypothermia
—Decreased ratio of surface area to mass
Metabolic

—Decreased activity of
—histamine
—serotonin
— bradykinin

—smooth muscle acting factor
—Altered complement activation

son (1967) reported the death of a seal from decompression syndrome after forced
surfacing from 300 m. Gas bubbles have been noted in blood samples obtained
from diving birds at various depths. After four dives to 4 atm of pressure, nitrogen
tension is increased to two to four times the original; after four dives to 7.8 atm
nitrogen pressure increased to five times the basal value. After 2.5 minutes of
submersion nitrogen increased from 82.4 to 89.5% of total gases (Kooyman et
al., 1973). Most vertebrates avoid decompression syndrome only because of the
briefness of their dives (Kooyman et al, 1973), but decompression syndrome can
and does occur with prolonged submergence in a variety of vertebrate taxa.

The pathophysiology of decompression syndrome is more complex than simple
vascular occlusion by nitrogen bubble emboli. Major causative factors may be
hemoconcentration and red blood cell agglutination, resulting in decreased tissue
perfusion (Strauss, 1970). Risk factors (derived from Strauss, 1970) for devel-
opment of decompression syndrome include dehydration, fatigue, increased he-
moconcentration, stress, impaired circulation, hypothermia, systemic disease, in-
jury, and obesity.

Physiological protective mechanisms are listed in Table 2. However, the impact
of the protective mechanisms is not straightforward. For example, the polycy-
themic (increased packed red blood cell content of blood) response may in some
instances actually reduce circulation due to its effect on blood viscosity (Lutz and
Bentley, 1985).

Decompression Syndrome in Turtles

Berkson (1967) first identified decompression syndrome in turtles. He reported
that blood drawn from turtles under pressure forms bubbles in the syringe and

1987

Rothschild— Decompression Syndrome in Fossil Turtles

257

further noted the gas emboli death of a turtle after rapid decompression from
18.7 atm. Berkson also noted that turtle blood was supersaturated with respect
to nitrogen at a given atmospheric pressure and that stepwise decompression also
produced caisson disease (decompression sickness). The susceptibility of turtles
to decompression syndrome is further substantiated by the observation of avas-
cular necrosis, a phenomenon previously noted in mosasaurs and attributed to
decompression syndrome (Rothschild and Martin, 1987).

While diving habits of some turtles are known, the impact on turtles of these
risk and protective factors during diving has been the subject of limited study.
The leatherback turtle ( Dermochelys coriacea) is known to dive almost continu-
ously with short surface intervals and can dive to 475 m. Lepidochelys olivacea,
the olive ridley sea turtle, can dive to 300 m (Eckert et al., 1986). The loggerhead
turtle ( Caretta caretta) utilizes its lungs as a major oxygen storage system and
spends only 1% of its time on the surface (Lutz and Bentley, 1985). Some turtles,
such as Pseudemys scripta (Glass and Wood, 1983), have significant cutaneous
respiration which could conceivably protect them from decompression sickness;
65% of the total C02 elimination in Trionyx mucita is non-pulmonary.

There is evidence that some of the physiological protective mechanisms may
be non-operative in turtles. Lutz and Bentley (1985) noted that turtles inhale prior
to dives, enhancing their risk. Lung collapse does occur in turtles at 80-160 m
(Berkson, 1967), which leaves a substantial range of depth in which gas exchange
continues.

Avascular necrosis has been identified in Desmatocheles, Prostostega, Porth-
ochelys, Oligochelone, Eosphargis, and several indeterminate Cheloniidae. Factors
which might affect susceptibility to this phenomenon, apparently related to de-
compression syndrome, include depth of dives, time at depth, surface time, lung
type, cutaneous respiration, complement responses to micro-bubble formation,
and perhaps even nitrogen excretion in the form of ammonium carbonate (Strauss
and Sampson, 1986).

A general survey of contemporary turtle humeri would allow identification of
susceptible species. Correlation of occurrence of avascular necrosis with diving
habits, risk factors, and protective mechanisms in extant turtles may allow as-
sessment of the relationship between risk factors and protective mechanisms.

Acknowledgments

I wish to express my appreciation to Drs. Paul Sartenaer, Larry Martin, Gordon Bell, James Lamb,
and Orville Bonner for opening their collections for my study and for their assistance in sectioning
specimens. Appreciation is also expressed to Drs. Karen and Scott Eckert for their constructive
criticisms during the development of this manuscript and to David Berman for his editorial assistance.

Literature Cited

Anderson, H. T. 1966. Physiological adaptations in diving vertebrates. Physiologic Review, 46:
212-243.

Berkson, H. 1967. Physiological adjustments to deep diving in the Pacific green turtle ( Chelonia
mydas agassizii ). Comparative Biochemistry and Physiology, 21:507-524.

Dennison, D. M., D. A. Warrell, and J. B. West. 1971. Airway structure and alveolar emptying
in the lungs of sea lions and dogs. Respiratory Physiology, 13:253-263.

Eckert, S. A., D. W. Nellis, K. L. Eckert, and G. L. Kooyman. 1986. Diving pattern of two
leatherback sea turtles (Dermatochelys coriacea) during intemesting intervals at Sandy Point, St.
Croix, U.S. Virgin Islands. Herpetologica, 42 (3):38 1-388.

Feldman, J. L., C. J. Menkes, B. Amor, A. Chevrot, and F. Delbvarre. 1981. L’osteonecrose
vertebrale de Tadulte. Revue de Rhumatisme, 48:773-780.

258

Annals of Carnegie Museum

vol. 56

Glass, M. L., and S. C. Wood. 1983. Gas exchange and control of breathing in reptiles. Physiologic
Review, 63:232=260.

Kooyman, G. L., J. P. Schroeder, D. G. Greene, and V. A. Smith. 1 973. Gas exchange in penguins
during simulated dives to 30 and 68 meters. American Journal of Physiology, 225:1467-1471.
Lutz, P. L., and T. B. Bentley. 1985. Respiratory physiology of diving in the sea turtle. Copeia,
1985(3):67 1-679.

Pauley, P. 1965. Decompression sickness following repeated breath-hold dives. Journal of Applied
Physiology, 20:1028-1031.

Resnick, D., G. Niwayama, J. Guerra Jr., V. Vint, and J. Usselman. 1981. Spinal vacuum
phenomena: Anatomical study and review. Radiology, 139:241-248.

Rothschild, B.M., and L. Martin. 1987. Avascular necrosis: Occurrence in diving Cretaceous

mosasaurs. Science, 236:75-77.

Strauss, M. B. 1970. Physiological aspects of mammalian breath-hold diving: A review. Aerospace
Medicine, 41:1362-1381.

Strauss, M. B., and R. L. Sampson. 1986. Decompression syndrome: An update. Physician and
Sports Medicine, 14(3): 1-9.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 16, Pp. 259-273

20 November 1987

TOWARDS A POSTGLACIAL HISTORY OF THE NORTHERN
GREAT PLAINS: A REVIEW OF THE
PALEOECOLOGIC PROBLEMS

Cathy W. Barnosky

Assistant Curator, Section of Paleobotany

Eric C. Grimm12
H. E. Wright, Jr.1 2

Abstract

Central to the late Quaternary ecological and cultural history of the northern Great Plains is an
understanding of (1) the nature and magnitude of vegetational and climatic change during the last
glaciation, when the Pleistocene megafauna went extinct, and (2) the timing and extent of aridity in
the early or middle Holocene, when populations of bison and humans apparently were reduced. Limited
pollen data suggest that during full glacial time the eastern half was covered with spruce parkland,
while the western half was probably treeless. The difference persisted through late glacial time, when
an open forest of spruce and hardwoods developed in the east and grassland in the west. During the
Holocene, the timing of Altithermal warmth and aridity is poorly documented, as is the response of
the biota, yet the east/west contrast is well demarcated. Sites in the eastern sector commonly show
middle Holocene warming and dryness typical of the Midwest, whereas western sites register maximum
drought in the early Holocene. The variable response is best explained as a metasynchronous shift
from a glacial climate influenced by the ice sheet to one governed by the early Holocene maximum
in summer radiation. The protracted presence of Laurentide ice delayed the radiation response in the
east, whereas “upwind” of the ice sheet, warming and drought were registered early.

Introduction

Two major trends have dominated the environmental history of the northern
Great Plains since the late Pleistocene. The first occurred between 18,000 and
10,000 yr B.P. at the close of the last glaciation, when the amelioration of climate
significantly altered the biota and landscape. With the retreat of the Laurentide
and Cordilleran ice sheets, a vast region was available for colonization by plants,
animals, and humans. Upon this landscape new communities formed, composed
of those species that had survived outside the ice border. The composition of late
glacial communities is not well known, but it undoubtedly resulted from a com-
bination of factors, including the prevailing climate, the distribution of suitable
substrates, the ability of species to take advantage of freshly exposed surfaces,
and the rate at which plant populations were able to migrate from unknown
refuges. The environment was clearly favorable for large herbivores, which were
abundant on the plains at this time. The heyday of the megafauna, however, ended
abruptly, either as a result of environmental and climatic change or at the hands
of big-game hunters of the Clovis culture (see Martin and Klein, 1984).

The second paleoenvironmental trend was towards increased dryness and warmth
in the early and middle Holocene (the Altithermal), followed by greater effective

1 Limnological Research Center, University of Minnesota, Minneapolis, Minnesota 55455.

2 Present address: Illinois State Museum, Springfield, Illinois, 62706.

Submitted 29 April 1987.

259

260

Annals of Carnegie Museum

vol. 56

moisture and cooling in the late Holocene. This climatic change is well docu-
mented in the Far West and Midwest, but on the Great Plains the details of the
climatic signal and vegetational response are poorly known, as are the attendant
adjustments in human and animal populations.

Understanding the kind, timing, and magnitude of the late Pleistocene and
Holocene vegetational and climatic changes of the northern Great Plains is im-
portant in solving some current problems in archeology, paleoclimatology, and
biogeography. Among these are the controversy over the extinction of the Pleis-
tocene megafauna as well as the nature of Holocene cultural history. The purpose
of this paper is to present the available paleoecologic data, to compare the post-
glacial environment in different regions, and finally to clarify the paleovegetational
and paleoclimatic questions that still need to be answered. The northern Great
Plains is defined here as the region bounded by the Rocky Mountains on the west,
the Nebraska Sandhills on the south, the Prairie Coteau on the east, and (arbi-
trarily) the Canadian border on the north.

The Pleistocene-Holocene Transition

For the vast area from the eastern Dakotas to the western edge of the plains in
Montana and Wyoming— a distance of 1500 km— almost no published paleo-
botanical data exist to indicate whether forest, tundra, or steppe prevailed during
the glacial period and when and how grassland came into existence. Because of
this absence of data, inferences must be drawn from sites peripheral to the northern
Great Plains. The only information on full glacial vegetation is the record of Picea
forest in northeastern Kansas at Muskotah Marsh and Arrington Marsh (see Fig.
1 for location of sites), which lie southeast of the Nebraska Sandhills and 250 km
southwest of the ice margin (Griiger, 1973). This full glacial forest differed from
the modern boreal forest of central Canada by the absence of Pinus banksiana.

By the time ice lobes in Iowa and the Dakotas had reached their maxima about
14,000 yr B.P. (Clayton and Moran, 1982), Picea had begun to spread its range
northward into the Des Moines area (Baker and Wain, 1985). By 12,000 yr B.P.,
spruce forest was replaced along its southern margin by prairie in southern South
Dakota. This transition is recorded at the Rosebud site, an interdunal depression
at the northern edge of the Nebraska Sandhills, 175 km southwest of the ice
margin (Watts and Wright, 1966). About 1 1,000 yr B.P., Quercus, Populus, Frax-
inus, and other hardwoods, which were probably confined to the central United
States in glacial time (Jacobsen et al., in press), expanded their ranges northward
and mixed with Picea in the eastern part of the northern Great Plains and the
Midwest. This admixture of trees has no close analogue in the present-day, but
the vegetation is presumed to have been open and dominated by spruce, with
some hardwoods but no pine. The western limit of the forest is not known. At
two sites in the glaciated region of northeastern South Dakota (Pickerel Lake:
Watts and Bright, 1968; Medicine Lake: Radle, 1981), deciduous trees replaced
Picea about 1 1,000 yr B.P. before prairie developed at 10,000 yr B.P. Pollen data
from east-central North Dakota (Woodworth Pond: McAndrews et al, 1967;
Seibold site: Cvancara et al., 1971) indicate a similar sequence. Farther north in
Canada, Picea forest was succeeded directly by prairie about 10,000 yr B.P. at
the Hafichuk, Riding Mountain, Herbert, Scrimbit, Glenboro, Sewell, and Russell
sites (see Ritchie, 1976). At Cottonwood Lake in central South Dakota, late glacial
Picea forest gave way to a parkland of Populus and Picea, with Fraxinus nigra,
minor other hardwoods, Artemisia, Poaceae, and other herbs about 12,000 yr

1987

Barnosky et al.— Northern Great Plains Postglacial History

261

Fig. 1.— Location of sites discussed in text. Circles indicate sites with fossil pollen records; squares
show the location of archeological/paleontological sites.

B.P. Parkland, in turn, was replaced by prairie at about 10,000 yr B.P. (E. C.
Grimm, unpublished data; Fig. 2).

Thus, for the eastern Great Plains the pollen records suggest that open spruce
forest with some deciduous trees existed in late glacial time. This forest followed
the ice retreat northward into southern Manitoba and Saskatchewan. Farther
south, forest was replaced by prairie and parkland in southwestern and central
South Dakota and by deciduous forest in the eastern Dakotas. A wedge of decid-
uous forest apparently extended westward as far as Woodworth in central North
Dakota, bounded by prairie on the southwest and spruce forest on the north and
northeast. This wedge seems to resemble the zone of deciduous forest that now
occurs between prairie and conifer forest in northwestern Minnesota, but unfor-
tunately only a preliminary pollen diagram is available from the Woodworth site
(McAndrews et al., 1 967), and it lacks radiocarbon dates. With continued warming
and drying, the band of deciduous forest disintegrated, and spruce forest gave
way directly to prairie as far north as southern Canada.

To the west of the Great Plains, late glacial vegetation is inferred mainly from
high-altitude sites of glacial origin. Upper treeline in the northern Rocky Moun-

262

Annals of Carnegie Museum

vol. 56

COTTONWOOD LAKE
Sully County, South Dakota
Analyst: E.C. Grimm

% terrestrial pollen

Fig. 2.— Pollen percentage diagram for selected taxa from Cottonwood Lake, South Dakota. Core
segment 6.38-6.48 m has a radiocarbon date of 13,410 ± 120 (WIS-1626), which is judged 1000-
3000 years too old based on age determinations at other Dakota sites (see discussion in text). The
erroneous date is presently attributed to an incorporation of ancient carbon into the sediments from
glacial till.

tains was lowered by at least 1000 m during the last glaciation, and high elevations
were probably covered by alpine tundra or steppe vegetation (Baker, 1983; Bar-
nosky et al., in press). Although the location of montane conifers during glacial
time is unclear, they expanded their range rapidly throughout the Rocky Moun-
tains between 1 1,000 and 9000 yr B.P. The early appearance of present-day forest
communities and the general uniformity of the late glacial forest from one moun-
tain range to another argues against great latitudinal or elevational displacement

1987

Barnosky et al.— Northern Great Plains Postglacial History

263

of conifer forests in glacial time. A more reasonable scenario is that conifers grew
in small populations along the lower mountain slopes during glaciation, bounded
by tundra or steppe at higher and lower elevations. The geomorphic evidence of
permafrost and strong winds in the Great Plains near the ice margin and on the
high plains of Wyoming also suggests that the lowlands were largely treeless
(Mears, 1981; Pewe, 1983; Kutzbach and Wright, 1986).

In the northwestern Great Plains, Guardipee Lake in western Montana seems
to corroborate the inferences drawn from montane sites (C. W. Barnosky, un-
published data). The site lies in the area of the former Two Medicine glacial lobe,
which originated in the Rocky Mountains and advanced 65 km eastward onto
the plains, terminating only a few kilometers from the Laurentide ice sheet (Alden,
1932; Lemke et al., 1965). The pollen data indicate treeless vegetation for the
past 12,000 years or so (Fig. 3), and the ratio between percentages of Artemisia
and Poaceae suggests grassland prior to 1 1,400 yr B.P. as opposed to sagebrush
steppe. By contrast, a contemporaneous record from Marias Pass 35 km to the
west implies parkland of Picea and Abies in the mountains (Carrara et al, 1986).
The high percentages of Finns pollen in the late glacial sediments at Guardipee
are similar to present-day values, and pine pollen was probably blown from distant
sources, just as today. Unfortunately, the Guardipee record is not old enough to
reveal whether conifers were on the plains during full glacial time. If trees were
indeed present beforehand, it could be argued that their elimination occurred with
the onset of greater seasonal contrasts and windier conditions in the late glacial,
when the Cordilleran and Laurentide ice sheets parted and allowed cold polar air
to funnel down the Alberta corridor onto the Montana plains. The influence of
an ice-free corridor on the climate, however, must have been short-lived, because
pollen types diagnostic of tundra are not present at Guardipee Lake. Also, by
11,400 yr B.P. the Two Medicine lobe had retreated west of the Continental
Divide (Carrara et al., 1986).

Two other sites in the western sector suggest an unforested landscape in the
late glacial period. Antelope Playa, a 13,000-yr record from the Powder River
Basin of eastern Wyoming, features Artemisia- dominated vegetation during both
late glacial and Holocene times, with no evidence of trees (Markgraf and Lennon,
1986). Bear Butte Lake just east of the Black Hills records open parkland vege-
tation in the late glacial interval, succeeded by prairie in the Holocene (E. C.
Grimm, unpublished data). At this site, the presumed (but undated) late glacial
period of open vegetation is dominated by Artemisia and Poaceae. Picea pollen
amounts to about 10%, and Finns is absent. Later, percentages of Picea pollen
decline and Finns appears, marking the Holocene immigration of pine into the
nearby Black Hills. Coincident with this transition, herbaceous pollen types change
in character, with Ambrosia and Chenopodiaceae/Amaranthaceae becoming dom-
inant.

Extinction of the Pleistocene Megafauna

Vertebrate records of glacial age are equally sparse on the northern Great Plains,
although those that have been studied are generally consistent with the environ-
mental reconstructions based on pollen data. Assemblages from Natural Trap
Cave (Martin and Gilbert, 1978; Gilbert and Martin, 1984) and the Colby Site
(Walker and Frison, 1980) in the Bighorn Basin of north-central Wyoming show
similarities with late Pleistocene faunas of interior Alaska, which imply a steppe-
tundra vegetation. However, they also contain alpine-tundra elements not found

264

Annals of Carnegie Museum

vol. 56

in the Alaskan assemblages (Chomko and Gilbert, 1987). Graham and Mead
(in press) discuss some of the specific differences between Alaska and Wyoming
faunas. Saiga, Lemmus, and Mammuthus primigenius, while common in Alaska,
were absent in Wyoming. Conversely, the Wyoming faunas contain taxa, including
Perognathus, Cynomys, and Thomomys, that do not occur in the Alaska faunas.
Not surprisingly, the glacial faunas from Wyoming are more similar to those from
sites along the former ice sheet margin in the Dakotas, Nebraska, and western
Iowa than those from interior Alaska (Graham and Mead, in press). Considered
as a whole, the late Pleistocene faunas from these northern plains sites suggest
increasing grasslands and aridity towards the west (Graham and Mead, in press;
R. W. Graham, personal communication, 1987), as do the pollen data.

This reconstruction does not seem to apply farther south. From the occurrence
of typically montane species in the fauna of the Hell Gap site, Roberts (1970)
argues for a late glacial forest on the plains of southeastern Wyoming. Hoffman
and Jones (1970) believe that the present occurrence of montane species in the
Black Hills indicates a once forested connection with the mountains to the west.
The early Holocene fauna at the Agate Basin site, in the short-grass prairie just
southwest of the Black Hills, contains species that today occur in boreal and
montane forests, including those of high altitudes in the Black Hills, as well as
species typical of steppe (Walker, 1982). Thus a parkland of forest and grassland
is indicated. Paleosols at the Agate Basin site, the nearby Sheaman site, and the
Sister’s Hill archeological site in the Powder River Basin in northeastern Wyoming
all indicate conditions wetter than present in the early Holocene (Reider, 1982a,
1982 b, 1983). Faunas from the Selby and Dutton sites in northeastern Colorado
indicate a grassland environment prior to 12,000 yr B.P. and a subsequent trend
towards increased moisture (Graham, 1981). From a survey of vertebrate fossil
sites in the western plains, Graham (1981) concludes that grassland prevailed
until about 1 1,000 yr B.P., but from then until 8000 yr B.P. the area was probably
forested.

A study of the Waubonsie local fauna from southwestern Iowa suggests that
extinction of the grazing megafauna may have come with the collapse of the mixed
grassland and open forest in the eastern Great Plains (Rhodes, 1984). This site,
dated at about 14,600 yr B.P., yielded 23 taxa of small mammals that together
suggests greater microhabitat diversity in the area than at present. Cold-steppe
taxa predominate, but the presence of temperate/mesic mammals implies isolated
areas of conifer and deciduous trees (Rhodes, 1984).

Proponents of the climatic hypothesis of megafaunal extinction argue that the
late Pleistocene spruce forest was a greater food resource than the modem closed
boreal spruce forest of Canada, because openings provided a nutritious ground
flora that grazing and browsing animals could utilize. In addition, at the end of
the Pleistocene the rapidity of climatic change and the resultant disruption of
plant communities throughout the continent prevented large mammals from
adapting their physiological and behavioral patterns to new conditions (Guthrie,
1984). Many animals became extinct, and others were much reduced in range
(muskox). Some that survived developed smaller body size (bison). Other sur-
vivors had been minor components before but were favored by the environmental
change and increased in relative numbers (moose). It is well established that late
Pleistocene faunas in the Great Plains, as elsewhere on the continent, were com-
posed of assemblages that have no analogues today; the local faunas contained
species that are now allopatric (Graham and Lundelius, 1984). One inference is

1987

Barnosky et al.— Northern Great Plains Postglacial History

265

that the vegetation was not so strongly zoned as it is today but rather was char-
acterized by a mosaic pattern that allowed animals to utilize a variety of vegetation
types without travelling long distances. Guthrie (1984) postulates that such a
mosaic pattern was particularly favorable for the non-ruminant herbivores, which
had to rely on a varied diet in order to obtain proper nutrition. Their relatively
simple digestive system could not handle toxins that accumulate in plants during
the growing season. Non-ruminants thrive in a climate with low seasonal contrasts,
that is, one with a long period of spring plant growth, when nutrient contents of
the plants are high and toxins low. It was these non-ruminant mammals that
became extinct (for example, horse, mammoth, mastodon, ground sloth, rhinoc-
eros, camelids), while the ruminants survived (caribou, deer, moose, wapiti, bi-
son). This scenario for megafaunal extinctions puts much of the burden of support
on the change in vegetational patterns at the end of the Pleistocene. It is possible,
however, that the sympatry in animals was not widely shared by the plants—
making human predation a more likely cause of extinction. The data base is not
adequate to choose between these hypotheses.

The Middle Holocene Dry Interval

The second environmental trend of interest is the early or middle Holocene
interval that was drier and possibly warmer than today, the Altithermal of Antevs
(1948). The case for such a climatic interval is strong at the eastern prairie/forest
border, which had moved well east of its present position between about 8000
and 6000 yr B.P. (Webb et al., 1983). Using pollen/climate transfer functions,
Bartlein et al. (1984) estimated that precipitation in the Minnesota area was about
20% less than it is today, but that temperature was only slightly higher. In the
Great Plains, an extreme view of middle Holocene environmental change is that
the entire Nebraska Sandhills was a vast area of desert dunes as a result of
Altithermal drought (Ahlbrandt et al., 1983). This reconstruction is countered by
direct evidence that the major dunes date to the late Pleistocene and not the
Holocene (Wright et al., 1985).

Archeological evidence implies that the middle Holocene climate was dry enough
in the heart of the Great Plains to reduce the density of the vegetation cover,
which in turn diminished the bison herds as well as the human populations
dependent on them (Frison, 1975, 1978). A tally of bison fossils in dated arche-
ological sites in the southern plains shows a paucity of bison kill sites during the
interval from 8000 to 4500 yr B.P. (Dillehay, 1974). Others, however, consider
that the archeological evidence is inadequate to postulate significant reductions
in human populations (Johnson and Holliday, 1986; Lynott, 1978; Reeves, 1973;
Davis, 1987; Graham, 1987), and the issue remains unresolved (Stoltman and
Baerreis, 1983).

Pollen studies carry information in addition to that implied by the transfer
functions in the Midwest. For example, the middle Holocene dominance of herb
pollen at sites in the present-day hardwood forest of west-central Minnesota, which
is a record of the eastward retreat of the prairie/forest border, contains more
Artemisia than presettlement samples from the eastern edge of the modern prairie
(Jacobson and Grimm, 1986). Artemisia percentages this high (in comparison to
Ambrosia and other herbs) are more characteristic of the mid-grass or short-grass
prairie today than of the tail-grass prairie, which occupied western Minnesota at
the time of settlement. The implication is that the prairie zones moved eastward
during the middle Holocene, just as did the prairie/forest border. Near the western

266

Annals of Carnegie Museum

vol. 56

edge of the tail-grass prairie, pollen diagrams from eastern South Dakota (Pickerel
Lake: Watts and Bright, 1968; Medicine Lake: Radle, 1981) and northwestern
Iowa (Lake West Okoboji: Van Zant, 1979) support this interpretation. In ad-
dition, studies in northwestern Minnesota suggest that the “prairie period” is
more complex than originally thought. Sedimentologic and paleoecologic data
from a varved lake record show at least two dry phases separated by a moister
interval during the period from 8500 to 4000 yr B.P. (Dean et al., 1984). The
macrofossil data from Kirchner Marsh in south-central Minnesota also imply
major climatic fluctuations during the Altithermal (Watts and Winter, 1966).

The timing of the Holocene dry /warm interval appears to vary geographically.
In Minnesota the maximum of Altithermal warmth and dryness occurred between
about 8000 and 4000 yr B.P., peaking at 7200 yr B.P. (Wright, 1976). In the
northwestern U.S. most sites register greatest drought in the early Holocene, about
10,500 to 7000 yr B.P. (Barnosky, 1984, 1985a, 1985Z?; Barnosky et al, in press;
Mathewes, 1973; Alley, 1976; Davis et al., 1986; Beiswenger, 1984; Mack et al,
1978a; Mack, Rutter, and Valastro, 1978), although at some sites it was delayed
until the middle Holocene, concurrent with the Midwest (Baker, 1976; Mehringer
et al., 1977; Mack, Rutter, Valastro, and Bryant, 1978; Mack et al., 1978&, 1979,
1983; Nickmann, 1979). The differences in timing may be more apparent than
real in the northwestern United States. Many of the sites featuring middle Ho-
locene drought are missing the early Holocene record, presumably because the
lakes dried or shallowed (Barnosky et al., in press).

The general tendency for an early Holocene warm/dry period in northwestern
U.S. has been explained as a response to increased summer radiation between
12,000 and 9000 yr B.P. controlled by the Milankovitch earth/sun orbital vari-
ations (Barnosky et al., in press). Enhanced summer insolation amplified the
subtropical high over the Northeast Pacific Ocean and thus intensified summer
warmth and aridity across the Pacific Northwest and northern Rocky Mountains.
In the early Holocene, the climate of the Midwest and northeastern United States
was presumably still under the influence of the ice sheet and thus the Altithermal
is expressed later, after the withdrawal of Laurentide ice well into Canada (Webb
et al., 1983).

A few pollen records from the northern Great Plains provide clues to the
Holocene climate in that region, but whether they are representative of the vast
area is uncertain. A record from Swan Lake in the southwestern part of the
Nebraska Sandhills indicates that an interdune depression contained a marsh
from its inception about 9000 yr ago until about 3700 yr ago, when a lake de-
veloped because of a climatically controlled rise in water table (Wright et al.,
1985). At Cottonwood Lake in central South Dakota, high percentages of Che-
nopodiaceae/Amaranthaceae pollen (Fig. 2), as well as degraded pollen, mark an
interval of increased aridity, coinciding with lake drying in the middle Holocene.
To the northwest, Guardipee Lake shows high ratios of Artemisia to Poaceae
through the early Holocene (Fig. 3). Because pollen surface samples in the West
indicate the general prevalence of Artemisia over Poaceae in drier regions (Mack
and Bryant, 1974; McAndrews and Wright, 1969), the climate near Guardipee
must have been more arid during the early Holocene than the late glacial or late
Holocene. At Antelope Playa in northwestern Wyoming, the percentage data show
trends similar to the Guardipee record, but Markgraf and Lennon (1986) reject
these in favor of interpretations based on pollen accumulation rates and ostracods.
Their conclusion of drier conditions after 5000 yr B.P. contradicts not only the

1987

Barnosky et al.— Northern Great Plains Postglacial History

267

GUARDIPEE LAKE
Glacier County, Montana
Analyst: C.W. Barnosky

% Terrestrial pollen

Fig. 3. —Pollen percentage curve for Pinus and the ratio of Artemisia to Poaceae pollen at Guardipee
Lake, Montana. Pinus pollen is inferred to be from distant sources throughout the record. Dashed
lines show the stratigraphic position of volcanic ash layers and their approximate age based on
radiocarbon dates from other sites (Mehringer et al., 1 984; P. Carrara, personal communication, 1986).
The Artemisia/ Poaceae ratio is generally highest between 2.20 and 5.70 m and provides evidence of
warmer and/or drier conditions in the early postglacial period, (ka = thousand years before present)

interpretation from Guardipee and other western pollen sites, but also evidence
of conditions moister than present in the Laramie Basin of southeastern Wyoming,
about 100 km to the south. From that area Wells (1970) reports late Holocene
snags of Juniperus scopulorum and fossil woodrat middens with needles of juniper
and ponderosa pine. The midden record includes 15 separately dated samples
evenly covering the time range from about 200 to 2300 yr ago, with three addi-
tional dates to about 5600 yr ago. Because of the nature of the fossil occurrences,
Wells (1970) is not able to detect a climatic trend within the late Holocene, only
that conditions were probably wetter 200 to 5600 yr ago than they are at present.

268

Annals of Carnegie Museum

vol. 56

Cooler/moister conditions in the late Holocene are also inferred from glacial
advances during the Little Ice Age and Neoglaciation in the Rocky Mountains
(Burke and Birkeiand, 1983).

Conclusions and Implications for Future Research

There are several reasons why the postglacial history of the northern Great
Plains is among the least well known in North America. First is the scarcity of
natural lakes and wetlands suitable for paleoecologic study. Existing wetlands are
often large shallow basins that contain mineral sediments with little organic matter
and cannot easily be cored by the standard square-rod piston corer (Wright et al.,
1983). The sites frequently yield an incomplete sedimentary record as a conse-
quence of intermittent desiccation and deflation. Another problem is that in the
glaciated areas of the northern Great Plains, the till contains carbonaceous shale,
lignite, and limestone, all of which make standard radiocarbon dates of lake
sediments unreliable. The more expensive accelerator method for dating plant
remains must be used instead. Analysis of the pollen record poses additional
problems that constrain the resolution of the vegetational reconstruction. The
fossil grains from these sites are frequently degraded and difficult to identify.
Furthermore, the common pollen types (for example, Gramineae, Artemisia,
Chenopodiaceae/Amaranthaceae) include several species with great ecological
amplitude.

Nonetheless, the available pollen studies with associated radiocarbon control
permit a preliminary analysis of regional climate during glacial and postglacial
time. At the glacial maximum the effect of the vast Laurentide ice sheet on the
Great Plains climate was particularly great. The physical evidence for periglacial
conditions in the northern Great Plains suggests a cold and windy climate. Ac-
cording to paleoclimatic model simulations, these winds were easterly or north-
easterly because of the anticyclonic circulation over the ice sheet (Kutzbach and
Guetter, 1986; Kutzbach and Wright, 1986). Their effect on full glacial vegetation
to the south is not known in the absence of old sites, but in late glacial time,
southerly winds generated by the diminished anticyclone are used to explain the
rapid dispersal of Picea glauca into western interior Canada (Ritchie and
MacDonald, 1986).

During the glacial maximum it is likely that open forest or parkland covered
much of the northeastern Great Plains. The late glacial records seem to reflect
the decline of periglacial conditions as the ice sheet thinned and retreated. By
contrast, the western part was probably treeless in glacial time, with a low diversity
of pollen types and low pollen accumulation rates, although this interpretation
comes from records no older than 12,000 yr B.P. The opening of the Alberta
corridor, although caused by increased temperature, may have brought colder,
windier winter conditions to the south by letting previously trapped Arctic air
escape from the polar region, but this effect was short-lived, as warming and ice
retreat proceeded.

During the late glacial period summer insolation in the Northern Hemisphere
was at a maximum, and seasonal contrasts were greatest. In the northern Great
Plains, it seems likely that the western sector was dominantly under the influence
of an insolation-forced climate, while the eastern sector experienced a periglacial
climate. The vegetational expression of this boundary may be the western limit
of the late Pleistocene spruce forest. Several eastern sites indicate late glacial
sympatry of plants that are presently allopatric. The overlapping of taxa with

1987

Barnosky et al.-— Northern Great Plains Postglacial History

269

different tolerances of maximum and minimum temperatures implies reduced
seasonality when the climate was under the influence of the ice sheet.

Today the Great Plains is the meeting place for Arctic, Caribbean, and Pacific
air masses. Changes in their relative importance during the Holocene resulted in
shifts in the location of major formational ecotones bordering the Great Plains
grassland (Bryson, 1966; Bryson et ah, 1970; Bartlein et al., 1984; Ritchie, 1976).
Although the movement of formational ecotones peripheral to the northern Great
Plains has been related to air mass regimes, this exercise has not been undertaken
for the plains region itself. The Holocene climatic history in the northern Great
Plains shows conditions that are intermediate between those in the Far West and
Midwest. Just as in late glacial time, early Holocene warmth/aridity in the western
part suggests a direct response to a climate driven by insolation changes. The
eastern sites, on the other hand, imply a climate controlled by the retreating ice
sheet until the middle Holocene. Model simulations show this contrast very nicely
(Kutzbach and Guetter, 1986), but the interpretations from the few existing sites
disagree. The South Dakota and North Dakota sites show middle Holocene aridity
as predicted by the model. To the west, percentages of xerophytic taxa at Guardipee
in northwestern Montana imply drought in the early Holocene, followed by greater
effective moisture. Wood from southeastern Wyoming also suggests moist con-
ditions in the late Holocene (Wells, 1970). But, at Antelope Playa in northeastern
Wyoming, relatively moist conditions are inferred in the early Holocene from
pollen accumulation rates and ostracod data (Markgraf and Lennon, 1986). The
explanation for these apparent discrepancies lies in more well-dated sites.

The controversy surrounding the Pleistocene megafaunal extinction requires
additional information about the past structure and composition of Great Plains
vegetation. The environmental change hypothesis (Guthrie, 1984; Graham and
Lundelius, 1984) states that during the Pleistocene the vegetation was structured
as a mosaic rather than in the strongly zonal pattern of today. The documentation
of parkland across much of the eastern plains would make this a compelling
argument for the region as a whole (although it does not cause rejection of the
overkill hypothesis). The environmental change hypothesis also assumes de-
creased seasonality during the Pleistocene, which can be tested by the late Pleis-
tocene sympatry of plants now separated because of different seasonal tolerances.

Holocene vegetation and climate need better characterization to determine
whether human and bison populations were reduced on the Great Plains during
the Altithermal. Based primarily on the paucity of archeological sites in the middle
Holocene, this hypothesis assumes that populations were low because extreme
aridity limited food and water resources (Frison, 1975, 1978). Alternatively, per-
sistent hunting by Indians may have kept bison populations below carrying ca-
pacity at all times (McDonald, 1984), in which case environmental changes had
little if any impact on bison population size. Another possibility is that the paucity
of middle Holocene archeological sites reflects the fact that these sites are deeply
buried and not readily apparent at the surface (R. W. Graham, personal com-
munication, 1987). Clearly further paleoecological and archeological as well as
geomorphological research in this area is needed.

Acknowledgments

This study is based upon work supported by the National Science Foundation under grants No.
BSR-831 1473 and BSR-86 16746. We thank R. W. Graham and R. G. Baker for their helpful reviews
of the manuscript.

270

Annals of Carnegie Museum

vol. 56

Literature Cited

Ahlbrandt, T. S., J. B. Swinehart, and D. G. Maroney. 1 983. The dynamic Holocene dune fields
of the Great Plains and Rocky Mountain basins, U.S.A. Pp. 379-409, in Eolian sediments and
processes (M. E. Brookfield and T. S. Albrandt, eds.), Elsevier, Amsterdam.

Alley, N. F. 1976. The palynology and paleoclimatic significance of a dated core of Holocene peat,
Okanogan Valley, southern British Columbia. Canadian Journal ofEarth Sciences, 13:1 131-1 141.

Alden, W. C. 1932. Physiography and glacial geology of eastern Montana and adjacent areas. United
States Geological Survey Professional Paper 174, 133 pp.

Antevs, E. 1948. The Great Basin, with emphasis on glacial and postglacial times. III. Climatic
changes and pre-white man. Bulletin of the University of Utah, 38:168-191.

Baker, R. G. 1976. Late Quaternary vegetation history of the Yellowstone Lake Basin, Wyoming.
United States Geological Survey Professional Paper 729-E, 48 pp.

. 1983. Holocene vegetational history of the western United States. Pp. 109-127, in Late

Quaternary environments of the United States, v. 2 (H. E. Wright, Jr., ed.), University of Min-
nesota Press, Minneapolis.

Baker, R. G., and K. A. Waln. 1985. Quaternary pollen records from the Great Plains and central
United States. Pp. 191-203, in Pollen Records of Late-Quaternary North American sediments
(V. M. Bryant, Jr. and R. G. Holloway, eds.), American Association of Stratigraphic Palynologists
Foundation, Dallas.

Barnosky, C. W. 1 984. Late Pleistocene and early Holocene environmental history of southwestern
Washington State, U.S.A. Canadian Journal ofEarth Sciences, 21:619-629.

. 1985 a. Late Quaternary vegetation near Battle Ground Lake, southern Puget Trough, Wash-
ington. Geological Society of America Bulletin, 96:263-271.

. 19856. Late Quaternary vegetation in the southwestern Columbia Basin. Quaternary Re-
search, 23:109-122.

Barnosky, C. W., P. M. Anderson, and P. J. Bartlein. In press. The northwestern U.S. during
deglaciation: vegetational history and paleoclimatic implications. In North America and adjacent
oceans during the last deglaciation (W. F. Ruddiman and H. E. Wright, Jr., eds.), Geology of
North America, v. K-3, Geological Society of America, Boulder, Colorado.

Bartlein, P. J., T. Webb III, and E. C. Fleri. 1984. Holocene climatic change in the northern
Midwest: pollen-derived estimates. Quaternary Research, 22:361-374.

Beiswenger, J. M. 1984. Late Quaternary vegetational history of Grays Lake Basin, southeastern
Idaho [Abstract]. American Quaternary Association Eighth Biennial Meeting, Boulder, Colorado,

p. 8.

Bryson, R. A. 1966. Air masses, streamlines, and the boreal forest. Geographical Bulletin, 8:228-

269.

Bryson, R. A., D. A. Baerreis, and W. M. Wendland. 1970. The character of late-glacial and
postglacial climatic changes. Pp. 53-74, in Pleistocene and recent environments of the central
Great Plains (W. Dort, Jr. and J. Knox Jones, Jr., eds.), The University Press of Kansas, Special
Publication 3, Lawrence.

Burke, R. M., and P. W. Birkeland. 1983. Holocene glaciation in the mountain ranges of the
western U.S. Pp. 3-1 1, in Late-Quaternary environments of the United States, v. 2 (H. E. Wright,
Jr., ed.), University of Minnesota Press, Minneapolis.

Carrara, P. E., R. E. Wilcox, and S. K. Short. 1986. Deglaciation of the mountainous region of
northwestern Montana east of the Continental Divide. Arctic and Alpine Research, 18:317-325.

Chomko, S. A., and B. M. Gilbert. 1987. The late Pleistocene/Holocene faunal record in the
northern Bighorn mountains, Wyoming. Pp. 394-408, in Late Quaternary mammalian biogeog-
raphy and environments of the Great Plains and prairies (R. W. Graham, H. A. Semken, and M.
A. Graham, eds.), Scientific Papers of the Illinois State Museum, 22.

Clayton, L., and S. R. Moran. 1982. Chronology of late Wisconsinan glaciation in middle North
America. Quaternary Science Reviews, 1:55-82.

Cvancara, A. M., L. Clayton, W. B. Bickley, Jr., A. F. Jacob, A. C. Ashworth, J. A. Brophy, C.
T. Shay, L. D. Delorme, and G. E. Lammers. 1971. Paleolimnology of late Quaternary deposits:
Seibold site, North Dakota. Science, 171:172-174.

Davis, L. C. 1987. Late Pleistocene/Holocene environmental changes in the central plains of the
United States: the mammalian record. Pp. 88-143, in Late Quaternary mammalian biogeography
and environments of the Great Plains and prairies (R. W. Graham, H. A. Semken, and M. A.
Graham, eds.), Scientific Papers of the Illinois State Museum, 22.

Davis, O. K., J. C. Sheppard, and S. Robertson. 1 986. Contrasting climatic histories for the Snake
River Plain, Idaho, resulting from multiple thermal maxima. Quaternary Research, 26:321-339.

1987

Barnosky et al. — Northern Great Plains Postglacial History

271

Dean, W. E., J. P. Bradbury, R. Y. Anderson, and C. W. Barnosky. 1984. The variability of
Holocene climatic change: evidence from varved lake sediments. Science, 226:1 191-1 194.

Dillehay, T. C. 1974. Late Quaternary bison population changes on the Southern Plains. Plains
Anthropologist, 19:180-196.

Prison, G. C. 1975. Man’s interactions with Holocene environments on the Plains. Quaternary
Research, 5:289-300.

. 1978. Prehistoric hunters of the High Plains. Academic Press, New York, 457 pp.

Gilbert, B. M., and L. D. Martin. 1984. Late Pleistocene fossils of Natural Trap Cave, Wyoming,
and the climatic model of extinction. Pp. 259-289, in Quaternary extinctions: a prehistoric
revolution (P. S. Martin and R. G. Klein, eds.), University of Arizona Press, Tucson.

Graham, R. W. 1981. Preliminary report on late Pleistocene vertebrates from the Selby and Dutton
archeological/paleontological sites, Yuma County, Colorado. Contributions to Geology, Univer-
sity of Wyoming, 20:33-56.

. 1987. The late Pleistocene and Holocene mammalian fauna the southwestern Plains of the

United States. Pp. 24-86, in Late Quaternary mammalian biogeography and environments of the
Great Plains and prairies (R. W. Graham, H. A. Semken, and M. A. Graham, eds.), Scientific
Papers of the Illinois State Museum, 22.

Graham, R. W., and J. I. Mead. In press. Environmental fluctuations and evolution of vertebrate
faunas during the last deglaciation in North America. In North America and adjacent oceans
during the last deglaciation (W. F. Ruddiman and H. E. Wright, Jr., eds.), Geology of North
America, v. K-3, Geological Society of America, Boulder, Colorado.

Graham, R. W., and E. L. Lundelius, Jr. 1984. Coevolutionary disequilibrium and Pleistocene
extinctions. Pp. 223-249, in Quaternary extinctions: a prehistoric revolution (P. S. Martin and
R. G. Klein, eds.), University of Arizona Press, Tucson.

Gruger, J. 1973. Studies on the late Quaternary vegetation history of northeastern Kansas. Geo-
logical Society of America Bulletin, 84:239-250.

Guthrie, R. D. 1984. Mosaics, allelochemics, and nutrients: an ecological theory of late Pleistocene
megafaunal extinction. Pp. 259-298, in Quaternary extinctions: a prehistoric revolution (P. S.
Martin and R. G. Klein, eds.), University of Arizona Press, Tucson.

Hoffman, R. S., and J. K. Jones, Jr. 1970. Influence of late-glacial and postglacial events on the
distribution of recent mammals on the northern Great Plains. Pp. 355-394, in Pleistocene and
recent environments of the Central Great Plains (W. Dort, Jr. and J. K. Jones, Jr., eds.), University
of Kansas Press, Lawrence.

Jacobson, G. L., and E. C. Grimm. 1986. A numerical analysis of Holocene forest and prairie
vegetation near Billy’s Lake in west-central Minnesota. Ecology, 67:958-966.

Jacobson, G. L., E. C. Grimm, and T. Webb, III. In press. Patterns and rates of vegetation change
in eastern North America from full-glacial to mid-Holocene time. In North America and adjacent
oceans during the Last Deglaciation (W. F. Ruddiman and H. E. Wright, Jr., eds.), Geology of
North America, v. K-3, Geological Society of America, Boulder, Colorado.

Johnson, E., and V. T. Holliday. 1 986. The Archaic record at Lubbock Lake. Plains Anthropologist
Memoir, 21:7-54.

Kutzbach, J. E., and P. J. Guetter. 1 986. The influence of changing orbital parameters and surface
boundary conditions on climate simulations for the past 1 8,000 years. Journal of the Atmospheric
Sciences, 43:1726-1759.

Kutzbach, J. E., and H. E. Wright, Jr. 1986. Simulation of the climate of 18,000 yr B.P.: Results
for the North American/North Atlantic/European sector. Quaternary Science Reviews, 4:147-
187.

Lemke, R. W., W. M. Laird, M. J. Tipton, and R. M. Lindvall. 1965. Quaternary geology of the
northern Great Plains. Pp. 15-28, in The Quaternary of the United States (H. E. Wright and D.
G. Frey, eds.), Princeton University Press, Princeton, New Jersey.

Lynott, M. J. 1978. An archaeological assessment of Bear Creek Shelter, Lake Whitney, Texas.
Southern Methodist University Archeology Research Program, Research Report, 1 15:1-109.

Mack, R. N., and V. M. Bryant, Jr. 1974. Modern pollen spectra from the Columbia Basin,
Washington. Northwest Science, 48:183-194.

Mack, R. N., N. W. Rutter, V. M. Bryant, Jr., and S. Valastro. 1978a. Late Quaternary pollen
record from Big Meadow, Pend Oreille Co., Washington. Ecology, 59:956-965.

. 1978&. Reexamination of postglacial vegetation history in northern Idaho: Hager Pond, Bonner

Co. Quaternary Research, 10:241-255.

Mack, R. N., N. W. Rutter, and S. Valastro. 1978. Late Quaternary pollen record from the
Sanpoil River Valley, Washington. Canadian Journal of Botany, 56:1642-1650.

272

Annals of Carnegie Museum

vol. 56

. 1979. Holocene vegetation history of the Okanogan Valley, Washington. Quaternary Re-
search, 12:177-193.

. 1983. Holocene vegetational history of the Kootenai River Valley, Montana. Quaternary

Research, 20:177-193.

Mack, R. N., N. W. Rutter, S. Valastro, and V. M. Bryant, Jr. 1978. Late Quaternary vegetation
history at Waits Lake, Colville River Valley, Washington. Botanical Gazette, 139:499-506.

Markgraf, V., and T. Lennon. 1986. Paleoenvironmental history of the last 13,000 years of the
eastern Powder River Basin, Wyoming, and its implication for prehistoric cultural patterns. Plains
Anthropologist, 31:1-12.

Martin, L. D., and B. M. Gilbert. 1978. Excavations at Natural Trap Cave. Transactions of the
Nebraska Academy of Sciences, 6:107-1 16.

Martin, P. S., and R. G. Klein, Eds. 1984. Quaternary extinctions: a prehistoric revolution.
University of Arizona Press, Tucson, 802 pp.

Mathewes, R. W. 1973. A palynological study of postglacial vegetation changes in the University
Research Forest, southwestern British Columbia. Canadian Journal of Botany, 1 1:2085-2103.

McAndrews, J. H., and H. E. Wright, Jr. 1969. Modem pollen rain across the Wyoming basins
and the northern Great Plains (U.S.A.). Review of Palaeobotany and Palynology, 9:17-43.

McAndrews, J. H., R. E. Stewart, Jr., and R. C. Bright. 1967. Paleoecology of a prairie pothole:
a preliminary report. North Dakota Geological Survey Miscellaneous Series, 30:101-1 13.

McDonald, J. N. 1984. The reordered North American selection regime and late Quaternary
megafaunal extinctions. Pp. 404-439, in Quaternary extinctions: a prehistoric revolution (P. S.
Martin and R. G. Klein, eds.), University of Arizona Press, Tucson.

Mears, B., Jr. 1981. Periglacial wedges and the late Pleistocene environment of Wyoming’s inter-
montane basins. Quaternary Research, 15:171-198.

Mehringer, P. J., Jr., S. F. Arno, and K. L. Petersen. 1977. Postglacial history of Lost Trail Pass
Bog, Bitterroot Mountains, Montana. Arctic and Alpine Research, 9:345-368.

Mehringer, P. J., Jr., J. C. Sheppard, and F. F. Foit, Jr. 1984. The age of Glacier Peak tephra in
west-central Montana. Quaternary Research, 21:36-41.

Nickmann, R. J. 1979. The palynology of Williams Lake Fen, Spokane County, Washington [M.S.
thesis]. Eastern Washington State University, Cheney. 57 pp.

Pewe, T. L. 1983. The periglacial environment in North America during Wisconsin time. Pp. 1 57—
189, in Late-Quaternary environments of the United States, v. 1 (S. C. Porter, ed.), University
of Minnesota Press, Minneapolis.

Radle, N. J. 1981. Vegetation history and lake-level changes at a saline lake in northeastern South
Dakota [M.S. thesis]. University of Minnesota, Minneapolis, 126 pp.

Reeves, B. 1973. The concept of an Altithermal hiatus in Northern Plains prehistory. American
Anthropologist, 75:1221-1253.

Reider, R. G. 1982 a. Soil development and paleoenvironments. Pp. 331-344, in The Agate Basin
Site: a record of the Paleoindian occupation of the Northwestern High Plains (G. C. Frison and
D. J. Sanford, eds.). Academic Press, New York.

. 1982 b. The soil of Clovis age at the Sheaman archaeological site, eastern Wyoming. Con-
tributions to Geology, University of Wyoming, 21:195-200.

. 1983. Soils and late Pleistocene-Holocene environments of the Sister’s Hill archeological

site near Buffalo, Wyoming. Contributions to Geology, University of Wyoming, 22:1 17-127.

Rhodes, R. S., II. 1984. Paleoecology and regional paleoclimatic implications of the Farmdalian
Craigmile and Woodfordian Waubonsie mammalian local faunas, southwestern Iowa. Illinois
State Museum Reports of Investigations, 40, 5 1 pp.

Ritchie, J. C. 1976. The late-Quatemary vegetational history of the western interior of Canada.
Canadian Journal of Botany, 54:1793-1818.

Ritchie, J. C., and G. M. MacDonald. 1986. The patterns of post-glacial spread of white spruce.
Journal of Biogeography, 13:527-540.

Roberts, M. F. 1970. Late glacial and postglacial environments in southeastern Wyoming. Palaeo-
geography, Palaeoclimatology, Palaeoecology, 8:5-17.

Stoltman, J. B., and D. A. Baerreis. 1983. The evolution of human ecosystems in the eastern
United States. Pp. 252-270, in Late-Quatemary environments of the United States, v. 2 (H. E.
Wright, Jr., ed.), University of Minnesota Press, Minneapolis.

Van Zant, K. L. 1979. Late-glacial and postglacial pollen and plant macrofossils from Lake West
Okoboji, northwestern Iowa. Quaternary Research, 12:358-380.

Walker, D. N. 1982. Early Holocene vertebrate fauna. Pp. 274-308, in The Agate Basin Site: a
record of the Paleoindian occupation of the Northwestern High Plains (G. C. Frison and D. J.
Stanford, eds.), Academic Press, New York.

1987

Barnqsky et al.— Northern Great Plains Postglacial History

273

Walker, D. N„, and G. C. Frison. 1980. The late Pleistocene mammalian fauna from the Colby
mammoth kill site, Wyoming. Contributions to Geology, University of Wyoming, 19:69=79.

Watts, W. A., and R. C. Bright. 1968. Pollen, seed, and mollusk analysis of a sediment core from
Pickerel Lake, northeastern South Dakota. Geological Society of America Bulletin, 79:855=876.

Watts, W. A., and T. C. Winter. 1966. Plant macrofossils from Kirchner Marsh— -a paleoecologic
study. Geological Society of America Bulletin, 77:1339=1360.

Watts, W. A., and H. E. Wright, Jr. 1966. Late-Wisconsin pollen and seed analysis from the
Nebraska Sandhills. Ecology, 47:202=210.

Webb, T , III, E. J. Cushing, and H. E. Wright, Jr. 1983. Holocene changes in the vegetation of
the Midwest. Pp. 142=165, in Late-Quatemary environments of the United States, v. 2 (H. E.
Wright, Jr., ed.), University of Minnesota Press, Minneapolis.

Wells, P. V. 1970. Postglacial vegetational history of the Great Plains. Science, 167:1574=1582.

Wright, H. E . Jr. 1976. The dynamic nature of Holocene vegetation— a problem in paleoclima-
tology, biogeography, and stratigraphic nomenclature. Quaternary Research, 6:581=596.

Wright, H. E., Jr., D. H. Mann, and P. H. Glaser. 1983. Piston corers for peat and lake sediments.
Ecology, 65:657-659.

Wright, H. E., Jr., J. C. Almendinger, and J. Gruger. 1985. Pollen diagram from the Nebraska
Sandhills, and the age of the dunes. Quaternary Research, 24:1 15=120.

■

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 17, Pp. 275-288

20 November 1987

FOSSIL CRAB (DECAPODA: BRACHYURA) FAUNA FROM THE
LATE CRETACEOUS (CAMP ANI AN-M A ASTRICHTI AN)

PIERRE SHALE IN BOWMAN COUNTY, NORTH DAKOTA

Annette B. Tucker1

Rodney M. Feldmann1

F. D. Holland, Jr.2

Kenneth F. Brinster3

Abstract

Twenty-seven specimens of bracfayuran decapods from the Late Cretaceous Pierre Shale in Bowman
County, North Dakota, have been referred to five species in five genera. Three specimens were assigned
to Ekaiakia lamberti Bishop, two to Dakoticancer overanus Rathbun, two to Homolopsis punctata
Rathbun, 19 to Raninella oaheensis Bishop, and one to Dioratiopus dawsonensis (Bishop). These
assignments extend the geographic ranges of all five genera into North Dakota and extend the strati-
graphic ranges of E. lamberti and D. dawsonensis . The specimens were collected from three localities
near the Little Missouri River in western Bowman County; all have been preserved as molds of the
interior with only small fragments of integument.

Introduction

Several specimens of decapod crustaceans were collected by one of us (KFB)
during a study of molluscan paleontology of the Late Cretaceous Pierre Shale in
Bowman County, North Dakota (Fig. 1). The decapod specimens were found in
fossiliferous concretions collected as float from the upper 16 m of the Pierre. The
dominant organic remains within the concretions consisted of gastropods, bi-
valves, scaphopods and cephalopods, with a relatively small number of decapod
crustaceans. The mineralogy of the concretions is primarily calcite with some
quartz and some clay minerals as determined by X-ray diffraction (Brinster, 1970).

The ages of these concretions were determined using baculitids within the
various concretions. The Baculites reesidei and B, grandis biozones were recog-
nized, while the intervening biozones of B. jenseni , B. eliasi , and B, baculus were
not recognized by Brinster (1970).

Subsequently, W. A. Cobban (personal communication to FDH, 22 March
1979) identified one of Brinster’ s baculites (Loc. A 1235) as B eliasi and other
ammonites (Loc. A 1238) as Didyrnoceras cheyennense (Meek and Hayden) and
Oxyheloceras meekanum (Whitfield) of late Campanian age. The apparent absence
of intervening biozones may simply reflect the lack of recognition of them in
Bowman County, to date. In the widely used ammonite zonation scheme for Late

1 Department of Geology, Kent State University, Kent, Ohio 44242.

2 Department of Geology and Geological Engineering, University of North Dakota, Grand Forks,
North Dakota 58202.

3 Science Applications International Corp , 2109 Air Park Road SE, Albuquerque, New Mexico 87111.
Submitted 13 April 1987.

275

276

Annals of Carnegie Museum

vol. 56

Fig. 1. — Map of contact between the Pierre Shale and the Fox Hills Formation showing localities in western Bowman County, North Dakota, from which
fossil crabs were collected. Shaded area consists of Fox Hills and younger strata; dotted line indicates inferred contact. Geology by KFB.

Table l.— Decapod faunas from the Pierre Shale and the Bearpaw Shale.

1987

Tucker et al.— Late Cretaceous Fossil Crabs

277

js 8

13 c
u

a

o

s

IK

&

~c>

S

,g

g

Cd

cd

s

1 3

II

•a ^

If

§ g
<3 £»

S, y

3 ^3
^ £ fi

■C § S

III
o e^2
IS, STo

§* K Uh

^ 3

a

o

T3 V5

I a

o

'3.-S

K

ISS

s J $

g< K -Si

gag

-S a

o

.1 §.§
c-^ s

S.-S S.
E-c e

cd ^ Cd

u u

s

o

3

8*8 8
£ « o
s g a

3 fe-a

CQ 3

cd

-C

C/3

£

&

cd

<u

m

^ g °

8^ .5
•3-3 |

g o g

Q. Co Q<

c <*=<

.2 c
jEs cd £

0 -s .2
'C ^
to & o

1 |’S
So

"g

I

«o °9
"g ■--
c g
g o

oq cq

cd cd
-C X

C/3 C/3

fc fc

<u (U

ffi ffi

S 2

- 5 ni £

* o

•g o

=5

U Q

3 +-

•2 ,3 »

<5 ^ a .a

^ ‘55 a

a I o |
§ .fc Q .1

g oj ex ^»:

s <« <
•2 K g -

a g % §

& g a a

I s! §

o 8 ~a

tq ^ O

S*

^ ? -a
3| I

life

| a's.

° aa;

^ g a
o

evSJ -c

V5 co ««

a rfS

c

e

.2

.2

'2

‘2

cd

cd

a

a

£

£

cd

cd

O

U

dj

to

C

dj

SC

K

<”S 5^.

S\

5>

K S3

s? a

Cj

I

$

stevensoni

nebrascense
xiteloceras je,
iculites rugos

ju T3

l igD

Q

C/3

1)Q
00 ,

£ d

C o

X> £ —

O

Su

>; o ^

" Ss
•*— *
C/5

8

u

Cd 03

“ s

1'i

c ^

Q q tq cq oq

ffl

a .
S Q

O rrC

<u n

oo .
X! C/3

•c «

rj a>
0.2

oaheensis, lobsters

Table \.— Continued.

278

Annals of Carnegie Museum

vol. 56

a

o

43

I

K

8

N

o

a-H
o ^

43

C/5 "

CO

R

I

cS 8
'I, c
(2 w

m

so»

gS

oa

Is

N ^

a

o

43

05

s

'3

£

8

•3

3 $3,

q &

o

o 2 3

&sl

II

r

a

43 .«??

ffi

o „ 5

&&J.a §

tt: j=j= m aT*
^ .2 .2 ‘a § p
a CQ « ^ &

c .a r o t- a

« 5 o "c S ■

«

gISAt
•5 1.3 l-S §

O «5

« 'i2, § «

- 151 <3 15 5?
a E ^ e Si
^ a -S « c
q Q ^

3

r

- s

ts

Cu >
-Cj O
R ^
,Q
1 R

•3 3
^ •«
-2 o

-S

^ 5

q q

2 ®
R <3

1 ^

O

* 3

s ^ a
^H ’'a -V
.S3 o -

2 =

! § - -

^ q q q

3-3

3 |

o p

a

.S

"S

03

a

6

03

u

o 5

a>

m

g fl

g o
.2 *C
S

I s
|S
u

o 3

+j 4 >
Cfl

B a

j3 .2

g o

.2 -a

a ■

- 3

|SS

o

"a o
8 §
5 §

q

«

'd

« c

jg -2

go

.2 -C

|1

is

U

Co

"3

R

I

q

•N 1

o

Q

§ J

q

d U

jo

'o

U

m

d

U

d

U IH

>>B

<o +Z

73®

cality

arimer.

2

M

O

&

£3

a

o

>

s

P3

0©

Os

o

m

PQ

References: 1) Bishop, 1973; 2) Bishop, 1976; 3) Feldmann, Bishop, and Kammer,
1983; 8) Bishop, 1985a; 9) Bishop, 19856; 10) this study; 11) Bishop, 1987.

1987

Tucker et al.— Late Cretaceous Fossil Crabs

279

Cretaceous strata of the Western Interior (Gill and Cobban, 1966:A28-A35; Gill
and Cobban, 1973:4; Kennedy and Cobban, 1978:70; Scott and Cobban, 1986),
the exposed part of the Pierre Shale in Bowman County, therefore, ranges from
late Campanian to early Maastrichtian age.

The Pierre Shale extends over a broad area of the Western Interior of North
America. Most exposures are located in north-central Montana and in central
South Dakota. Outcrops of the Pierre also occur in the Black Hills in South Dakota
and Wyoming, along the flanks of other uplifts in the Rocky Mountain region,
and in south-central and extreme southwestern North Dakota. Much of the Pierre
Shale is concealed by younger rocks or by glacial debris (Tourtelot, 1962). Lateral
equivalents of the Pierre Shale include the Claggett Shale, the Judith River For-
mation, and the Bearpaw Shale in Montana. The Pierre is conformably overlain
by the Fox Hills Formation with a gradational change from a fine, silty, clay shale
of the upper Pierre Shale to a sand, silty clay or shaly silt of the lower Fox Hills
Formation (Waage, 1961).

Exposures of the Pierre Shale in Bowman County result from erosion of the
eastern flank of the Cedar Creek anticline (Fig. 1). The axis of the asymmetrical
anticline is oriented northwest-southeast; beds dip steeply on its southwest flank
and gently on its northeast flank (Bishop, 1965). The Little Missouri River cuts
the southeastern tip of the structure. Little Beaver Creek which empties into the
Little Missouri River cuts the anticline normal to its axis.

The purpose of this paper is to describe the decapods from Bowman County,
North Dakota, and to compare them to those described by Bishop and others
from South Dakota, Montana, and Colorado (Table 1). Descriptive treatment of
the different taxa varies in length and detail depending upon the need to augment
previous descriptions. Extensions in the geographic and stratigraphic ranges of
the decapod taxa are noted, as well.

The decapod specimens were found in concretions from the three localities
given below. Measured specimens, their catalog numbers in The Carnegie Museum
of Natural History paleontological collection, and their University of North Da-
kota accession numbers are listed in Table 2. The localities are:

A 1 235 Float below east bank of Little Beaver Creek, NEVA SWVA sec. 24, T 1 32N, R 1 07 W, Webster

NE Quadrangle, Montana-North Dakota, Bowman County, about 6 miles southwest of
Marmarth, Slope County, North Dakota (KFB field no. 69-9c).

A 1238 Float in creek bottom, SEVA NW'A sec. 6, T129N, R106W, Cedar Ridge Quadrangle,

Bowman County, about 20 miles south of Marmarth, Slope County, North Dakota (KFB
field no. 69-12).

A 1 24 1 Float below west bank of Little Missouri River, SEVA NWVA sec. 3, T 1 30N, R 1 06 W, Cedar

Ridge Quadrangle, Bowman County, about 1 2 miles south of Marmarth, Slope County,
North Dakota (KFB field no. 69-17).

Systematic Paleontology

Order Decapoda Latreille, 1803
Suborder Pleocyemata Burkenroad, 1963
Infraorder Brachyura Latreille, 1803
Section Podotremata Guinot, 1977
Subsection Dromiacea Guinot, 1977
Superfamily Dromioidea de Haan, 1833
Family Prosopidae von Meyer, 1860

280

Annals of Carnegie Museum

vol. 56

Subfamily Prosopinae von Meyer, 1860
Genus Ekalakia Bishop, 1976

Type species. —Ekalakia lamberti Bishop, 1976.

Ekalakia lamberti Bishop, 1976
Fig. 2. 1-2.3

Preservation.— Three specimens can be referred to Ekalakia lamberti Bishop.
The largest is a well-preserved, complete mold of the interior; pitting, caused by
dissolution of the calcitic mud filling, occurs over the entire specimen. A smaller
specimen is a well-preserved, complete mold of the interior with the anterior
portion of the rostrum absent. The thoracic region of the smaller, complete spec-
imen exhibits slight exfoliation. A third specimen is a poorly preserved mold of
the interior with the anterior one-third fractured and somewhat displaced. The
appendages, abdomen, and sternum are not preserved on any of the specimens.

Remarks. — The specimens agree in most respects with Ekalakia lamberti Bish-
op. The carapace is longer than wide, strongly convex transversely, and convex
longitudinally. The rostrum is downturned and has a median sulcus which splits
into fainter grooves posteriorly on both sides of the mesogastric and metagastric
regions. There is a U-shaped, subparallel depression between the protogastric and
epigastric regions. There are also cervical and branchiocardiac furrows which are
U-shaped, subparallel and equally deep. The metabranchial regions are inflated
and the cardiac region is raised well above the rest of the carapace.

The specimens differ from the original description of E. lamberti in that a node
is situated on the posterior cardiac region of the largest specimen (Fig. 2.2). In
addition, two well-developed tubercles occur on the mesogastric region and one
tubercle is developed on the metagastric region of the two complete specimens.
The occurrence of this node and the tubercles, however, does not warrant naming
a new species. The branchial region is pustulose and the branchial region of the
largest specimen is more inflated than that of the smaller, complete specimen
making the larger specimen appear to be more vaulted (Fig. 2. 2-2. 3). The posterior
portion of the broken specimen is well preserved and the morphology of this,
together with various length-width ratios similar to those of the other two spec-
imens, strongly suggests that it belongs in E. lamberti (Fig. 2.1). Although the
inflated branchial region, as well as the rapidly narrowing anterolateral margin,

Table 2 .—Decapod Fauna, Pierre Shale, Bowman County, North Dakota. An additional sixteen spec-
imens of Raninella oaheensis are from UND locality number A 1238.

Species

Carnegie Museum Paleon-
tology Catalog No.

UND loc. no.

Length

width

Ekalakia lamberti

CM 34558

A1238

13.6 mm

12.2 mm

CM 34559

A1235

8.8 mm

7.8 mm

CM 34560

A1238

10.0 mm

8.3 mm

Dakoticancer overanus

CM 34561

A 1241

14.4 mm

15.4 mm

CM 34562

A1235

16.7 mm

broken

Homolopsis punctata

CM 34563

A 1241

10.0 mm

8.0 mm

CM 34564

A 1241

6.7 mm

5.0 mm

Raninella oaheensis

CM 34565

A1238

17.9 mm

10.2 mm

CM 34566

A1238

16.0 mm

9.8 mm

CM 34567

A1238

16.2 mm

10.1 mm

Dioratiopus dawsonensis

CM 34582

A1238

16.2 mm

14.8 mm

1987

Tucker et al.— Late Cretaceous Fossil Crabs

281

Fig. 2.—Ekalakia lamberti Bishop; dorsal view of carapaces. 1, CM34560. 2, CM34558. 3, CM34559.
Scale bars equal 1 cm.

causes this specimen to resemble the majids; conservatively, it should be referred
to E. lamberti.

Superfamily Dakoticancroidea Rathbun, 1917
Family Dakoticancridae Rathbun, 1917
Genus Dakoticancer Rathbun, 1917

Type species. — Dakoticancer over anus Rathbun, 1917.

Dakoticancer overanus Rathbun, 1917
Fig. 3.1, 3.2

Preservation. —Two specimens have been identified as Dakoticancer overanus
Rathbun. The larger specimen is a partially broken mold of the interior; the right
side is complete while the left side is incomplete. The smaller specimen is a
complete, very well-preserved mold of the interior with some adherent, exfoliating
integument.

Emendation to description.— C arapace thick, deeply grooved, and slightly wider than long. Dorsal
surface convex longitudinally, front severely downtumed anteriad the cervical groove; slightly convex
transversely. Surface of carapace finely granulate on raised portions; intermediate areas smooth to
finely punctate.

Width of fronto-orbital region approximately one-half greatest width. Orbits circular in outline;
margins flare outward and upward; outer orbital margin bears one dorsal and one ventral tooth.
Proximal portion of rostrum sulcate with raised rims; length undetermined; downtumed.

Anterolateral margin convex in outline with a nearly vertical sidewall. Margins below hepatic region
marked by well-defined rim and shallow furrow which extends backward and upward to form groove
separating mesobranchial and metabranchial regions. Posterolateral margin weakly sigmoid. Posterior
margin slightly convex with marginal rim and furrow.

Epigastric regions on either side of posterior rostral sulcus broadly swollen centrally; regions poorly
defined by broad shallow grooves. Interior groove of epigastric region is posteriad continuation of
rostral sulcus; grooves separating epigastric from protogastric regions converge posteriorly from frontal
margin. Protogastric regions defined by shallow grooves; two interior shallow longitudinal furrows
diverge posteriorly from rostral sulcus to separate protogastric and mesogastric regions; shallow fur-
rows, parallel to diverging interior furrows, separate protogastric and hepatic regions. Hepatic regions
swollen and pustulose. Transverse, deeply-etched cervical groove defines posterior margins of hepatic
and protogastric regions; a deep pit marks juncture of posterior protogastric-hepatic furrow and cervical
furrow. Protogastric regions swollen posteriorly. Transverse cervical groove not crossing mesogastric-
cardiac region. Longitudinal furrows converge gradually until they almost meet at posterior margin
of carapace; furrows interrupted by shallow pits on either side of anterior intestinal region. Another
transverse groove separates mesobranchial and metabranchial regions; groove interrupted at midpoint
so that incision is incomplete between cardiac and intestinal regions. Mesobranchial region bears
inflated boss at midpoint; boss tapers into ridge which extends forward and down along side-wall.

282

Annals of Carnegie Museum

vol. 56

Fig. 3 .—Dakoticancer overanus Rathbun; dorsal view of carapaces. 1, CM34561. 2, CM34562. Scale
bars equal 1 cm.

Metabranchial region with central ridge which extends gently foward to lateral margin where it abruptly
terminates. Cardiac region gently swollen and marked centrally by two arcuate depressions on either
side of midline; anterior portion of depressions marked by pits.

A series of nodes ornaments dorsal carapace. One small node occurs just above each orbit; series
of three nodes forms an inverted triangle on posterior-most portion of intestinal surface.

Remarks. —These specimens conform, in most respects, to the description of
Dakoticancer overanus Rathbun. The carapace is thick and deeply grooved; the
front, between the eyes, is small; the orbits are well defined; the carapace is covered
with small granules.

The specimens differ from the description by Rathbun of D. overanus in that
the mesobranchial regions bear inflated bosses (Fig. 3. 1-3.2). Several of Bishop’s
(1981) illustrations of D. overanus also clearly show bosses on the mesobranchial
regions of those specimens. However, the recognition of these bosses does not
warrant the naming of a new species. The original description of the mesobranchial
regions of D. overanus was unclear.

Superfamily Homoloidea White, 1847
Family Homolidae White, 1847
Genus Homolopsis Bell, 1863

Type species. —Homolopsis edwardsii Bell, 1863.

Homolopsis punctata Rathbun, 1917
Figs. 4. 1-4.3, 5

Remarks. — Two specimens of Homolopsis punctata Rathbun have been iden-
tified. The larger specimen is a well-preserved, partially exfoliated mold of the
interior of a carapace broken longitudinally along the linea homolica ; the rostrum
is absent (Fig. 4.2).

The smaller specimen is a very well-preserved juvenile that is also a mold of
the interior of a carapace broken along the linea homolica (Fig. 4.3). The rostrum,
separated from the carapace, is preserved as a mold of the exterior.

The specimens agree in most respects with Homolopsis punctata Rathbun. The
anterior one-third of the carapace curves strongly downward. The imprint of the
rostrum of the smaller specimen confirms that the rostrum is sharply downturned.
The elevations form regularly placed bosses: one boss on each epigastric lobe,

1987

Tucker et al.— Late Cretaceous Fossil Crabs

283

Fig. A. —Homolopsis punctata Rathbun; dorsal view of carapaces. 1, Holotype USNM32058. 2,
CM34563. 3, CM34564. Scale bars 1 & 2 equal 1 cm; scale bar 3 equals 1 mm.

three larger bosses and one small boss on each protogastric lobe, one anterior
node and a posterior boss on the mesogastric region, two kidney-shaped nodes
on the metagastric region, and one small boss and two large bosses on each
epibranchial lobe. The cardiac region is urn-shaped and is broadly swollen into
a dome in the posterior portion; the domed region is not nodose. Deep grooves
delineate the hepatic, gastric, mesobranchial, and cardiac regions. The surface of
the carapace is finely punctate. The posterior margin is bordered by a thin raised
rim.

These characters strongly resemble those of Homolopsis dispar Roberts (1962)
from the Late Cretaceous Merchantville Formation in New Jersey. In addition,
there is a pronounced similarity in the preserved right lateral margins of the
holotype of H. punctata (Fig. 4.1) and H. dispar (Fig. 5); each specimen bears a
well-preserved epibranchial lateral spine which is aligned in a similar manner
with an epibranchial boss just within the line a homolica. These similarities confirm
the synonymy proposed by Wright and Collins (1972:46) of H. dispar with H.
punctata. The absence of the anterior mesogastric node on H. dispar and the
dissimilarities in the size of some of the nodes when compared to H. punctata
serve to illustrate the variation among individuals within this species. In fact, the

Fig. 5.— Homolopsis punctata Rathbun {=H. dispar Roberts). Dorsal view of carapace of the holotype
of H. dispar, ANSP20030. Scale bar equals 1 cm.

284

Annals of Carnegie Museum

vol. 56

Fig. 6.—Raninella oaheensis Bishop. 1, Left lateral view of carapace, CM34565. 2, Oblique anterior
view showing right orbital ventral fissures, CM34566. 3, Dorsal view of carapace, paratype USNM
173581, exhibiting bite marks. 4, Dorsal view of carapace, CM34565. 5, Dorsal view of carapace,
holotype USNM 173589. 6, Dorsal view of carapace, CM34567. 7, Dorsal view of carapace, CM34566.
Scale bar equals 1 cm.

range of variation between specimens from the Pierre Formation is as great, or
greater, than that between the New Jersey material and the specimens from the
Pierre.

Subsection Archaeobrachyura Guinot, 1977
Superfamily Raninoidea de Haan, 1833
Family Raninidae de Haan, 1833
Subfamily Ranininae Serene and Umali, 1972
Genus Raninella A. Milne Edwards, 1862

Type species. —Raninella trigeri A. Milne Edwards, 1862.

Raninella oaheensis Bishop, 1978
Fig. 6. 1-6.7

Preservation. — Nineteen specimens have been referred to Raninella oaheensis
Bishop. Most of the specimens are broken; however, three specimens used for
description are nearly complete molds of the interior. One specimen exhibits a
well-preserved frontal region; another is a mold of the interior, with adhering
portions of the integument that illustrate the distribution of the dorsal nodes. A
third specimen has preserved the ventral orbital fissure.

Emendation to description. — Carapace ovate in outline, widest in anterior one-third, greatest width
about 60% of total length. Surface finely punctate. Carapace slightly convex longitudinally and very
convex transversely. Lateral margins turned under carapace (Fig. 6.1).

Width of fronto-orbital region 63% of greatest width. Fronto-orbital region widest anteriorly, tapering
posteriorly; orbits directed slightly away from long axis, ovate, height approximately one-half width;
margins of orbits thicken and flare outward and upward. Dorsal margin of each orbit marked by two
deeply grooved, open fissures, each about twice as deep as wide that approximately parallel lateral
margins of orbits; axes of fissures converge posteriorly toward midline of carapace; median fissures
about twice as deep and twice as wide as lateral fissures (Fig. 6.4). Each ventral orbital margin with
one open fissure near lateral margin (Fig. 6.2). Frontal margin of carapace produced to form rostrum

1987

Tucker et al.—Late Cretaceous Fossil Crabs

285

that extends well beyond orbits, not downtumed. Rostrum long, triangular in outline with straight
margins, keeled medially; keel subtle, extending onto carapace a distance greater than length of rostrum;
keel bounded laterally by shallow sulci (Fig. 6.4-6. 5).

Anterolateral margin of carapace convex in outline, bearing two pairs of short spines; pair of hepatic
spines point outward and slightly foward, pair of lateral spines point forward and slightly outward.
Posterior lateral spines define greatest width of carapace. Posterolateral margin weakly sigmoid, with
carapace tapering to narrow, blunted, posterior margin; posterolateral margin with narrow, well-
defined, marginal rim and furrow; margin turned under carapace. Posterior margin smoothly concave
with marginal rim and furrow.

Midline of carapace slightly keeled from rostrum to posterior mesogastric region where keel merges
into cardiac region; cardiac region gently and broadly swollen, flattening to posterior margin. Urn-
shaped cardiac region defined by pair of shallow grooves; anterior two-fifths deeply etched as arcuate
impressions curving toward midline from a position at midlength of carapace, becoming less prominent
posteriorly; remainder of groove subtle (Fig. 6. 3-6. 7).

Cardiac region with four nodes; a pair of nodes on either side of midline on a transverse line passing
through posterior end of deeply impressed portion of cardiac grooves and another pair on either side
of midline on transverse line just anteriad of cardiac grooves. A pair of gastric pits occurs on either
side of midline on transverse line between lateral spines; gastric pits with raised medial margins form
nodes and begin series of small nodes forming curved lines that extend forward and diverge from
midline axis. Just anterior to gastric nodes, three nodes form a triangle with its base on transverse
line between lateral spines and its apex directed anteriorad. Two larger gastric nodes occur anteriad
apex of nodal triplet. All nodes most obvious where carapace has been exfoliated; subtle on surface
of integument. No other grooves present. Hepatic region bears pustulose prosopon (Fig. 6.1).

Remarks. —Raninella trigeri A. Milne Edwards, type species of the genus, has
both pairs of marginal spines pointing forward; the spines are much more robust
than in R. oaheensis. The posterolateral margin is convex in R. trigeri and not
sigmoid as in R. oaheensis. Raninella eocenica Rathbun (1935) is widest a little
behind the midpoint and possesses a long, cylindrical spine on the outer margin
of each orbit. R. oaheensis has a longer, more narrow rostrum, not downtumed,
than that in Raninella tridens Roberts (1962); the rostrum of R. oaheensis is
keeled medially, not excavated as in R. tridens. The orbits of R. tridens have
closed fissures on both upper and lower margins. Raninella mucronata Rathbun
(1935) is widest at the midpoint and the posterolateral margins converge rapidly
posterior of the midpoint. Raninella carlilensis Feldmann and Maxey (1980) has
nearly circular orbits with closed fissures on the rims; the rostrum is shorter than
in R. oaheensis ; the deeply impressed gastric grooves observed in R. oaheensis
are fainter in R. carlilensis.

Superfamily Tymoloidea Alcock, 1896

Family Torynommidae Glaessner, 1980
Genus Dioratiopus Woods, 1953

Type species.— Dioratiopus salebrosus Woods, 1953.

Dioratiopus dawsonensis (Bishop) 1973

Preservation. — This species is represented by one complete, very well-preserved,
exfoliated mold of the interior. The distal portion of the rostrum is absent, and
the proximal portion of the rostrum has been cracked transversely and is displaced
slightly. The appendages, abdomen, and sternum are not preserved.

Emendations to description. —Frontal region with slightly downtumed, sulcate rostrum with raised
proximal rims and one proximal spine (Fig. 7). Another spine on dorsal, interior rim of orbit. Outer
extension of orbit marked by forwardly directed hepatic spine. Protogastric region divided posteriorly
by shallow, longitudinal furrow parallel to groove between mesogastric and protogastric regions into
a swollen adaxial area and an outer area bearing two nodes.

Remarks. —Glaessner (1980: 182) considered Glaessnerella Wright and Collins,

286

Annals of Carnegie Museum

vol. 56

1975 ( =Glaessneria Wright and Collins, 1972) to be the junior, subjective syn-
onym of Dioratiopus. Although considered closely allied by Wright and Collins
(1972:33), Glaessnerella was distinguished from Dioratiopus by being much less
vaulted in both transverse and longitudinal section. Glaessnerella also lacks an
inner, oblique sulcus on the mesobranchial area and has a more strongly spinose
frontal area than Dioratiopus (Wright and Collins, 1972). However, Glaessner
(1980) did not consider these differences to be reliable generic characters. He
assigned Dioratiopus Woods to his new family, Torynommidae in the Tymoloidea.
He considered the Torynommidae to be a natural taxon since it separates the
Cretaceous genera of widely-distributed, shallow-water crabs from some of their
deep-water descendants. Glaessner (1980) also reassigned Homolopsis dawsonen-
sis Bishop to Dioratiopus because of the absence of a linea homolica and the
presence of weaker ornamentation than that found on Homolopsis. Bishop (1985/?:
6 1 6) concurred with this reassignment.

Discussion. — The specimen exhibits the following family characters: the cara-
pace is pentagonal but only slightly longer than wide; it is convex transversely
and slightly convex longitudinally; the front is spatulate, projecting, and not sharp;
the sidewalls are steep; the orbits are shallow. The remainder of the family char-
acters are not preserved on this specimen.

The specimen agrees with the generic characters of Dioratiopus Woods, 1953.
The carapace is pentagonal and longer than broad. The cervical and branchial
grooves are equally well-developed and the sidewalls are nearly vertical. The
rostrum is produced to form a short, anteriorly directed triangle. The orbits are
shallow and form wide depressions.

The specimen conforms in every respect to the original description of this species
(Bishop, 1973). With the preserved frontal region including the rostral area, the
shape of the carapace is pentagonal.

Conclusions

This study represents the first examination of brachyurans from the upper Pierre
Shale in Bowman County, North Dakota and, as such, expands the geographic

1987

Tucker et al.— Late Cretaceous Fossil Crabs

287

range of the five genera represented. Ekalakia lamberti Bishop, 1976 has had no
previously definitive biostratigraphic association. As a result of the association
with Baculites reesidei and B. grandis, a range of late Campanian to early Maas-
trichtian is confirmed for this species.

The material of this study represents an assemblage dominated by Raninella
oaheensis Bishop, 1978. Bishop has suggested a Campanian-Maastrichtian range
for this species that is confirmed by this study, as is the Campanian-Maastrichtian
range of Homolopsis punctata and Dakoticancer overanus . Dioratiopus dawso-
nensis (Bishop) was originally reported from the zones of Baculites baculus and
B. grandis suggesting a Maastrichtian age for this species (Bishop, 1973:19). Al-
though Bishop (1985^:6 1 8) extended the North American range of the genus back
into the Campanian with the description of D. hearttailensis, this study suggests
that the range of D. dawsonensis also be extended to include late Campanian.

Acknowledgments

We wish to thank several people who aided us in this study. Frederick Collier, U.S. National Museum,
Washington, D.C., and George M. Davis, Academy ofNatural Sciences of Philadelphia, made available
type material used for comparative work; Luke Savoy, North Dakota Geological Survey, Grand Forks,
drafted Fig. 1; and Barry B. Miller, Department of Geology, Kent State University, read portions of
the manuscript. The paper was reviewed and substantially improved by Gale Bishop, Georgia Southern
College, and an anonymous reviewer. Contribution 350, Department of Geology, Kent State Uni-
versity, Kent, Ohio 44242.

Literature Cited

Bishop, G. A. 1965. The paleontology and stratigraphy of the Upper Cretaceous upper Pierre Shale
on Cedar Creek anticline, southeastern Montana. South Dakota Academy of Science Proceedings,
44:91-96.

. 1973. Homolopsis dawsonensis : a new crab (Crustacea, Decapoda) from the Pierre Shale

(Upper Cretaceous, Maastrichtian) of Cedar Creek anticline, eastern Montana. Journal of Pa-
leontology, 47:19-20.

. 1976. Ekalakia lamberti n. gen., n. sp. (Crustacea, Decapoda) from the Upper Cretaceous

Pierre Shale of eastern Montana. Journal of Paleontology, 50:398-401.

. 1978. Two new crabs, Sodakus tatankayotankaensis n. gen., n. sp. and Raninella oaheensis

n. sp. (Crustacea, Decapoda), from the upper Cretaceous Pierre Shale of South Dakota. Journal
of Paleontology, 52:608-617.

— . 1981. Occurrence and fossilization of the Dakoticancer assemblage, Upper Cretaceous Pierre

Shale, South Dakota. Pp. 383-413, in Communities of the past (J. Gray, A. J. Boucot, and W.
B. N. Berry, eds.), Hutchinson Ross Publishing Co., Stroudsburg, Pennsylvania.

— — . 1982. Homolopsis mendryki : a new fossil crab (Crustacea, Decapoda) from the Late Cre-
taceous Dakotacancer assemblage, Pierre Shale (Maastrichtian) of South Dakota. Journal of Pa-
leontology, 56:221-225.

— . 1983. Two new crabs, Notopocorystes ( Eucorystes ) eichhorni and Zygastrocarcinus griesi

(Decapoda: Brachyura) from the Bearpaw Shale (Campanian) of north-central Montana. Journal
of Paleontology, 57:900-910.

— . 1985a. A new crab, Eomunidopsis cobbani n. sp. (Crustacea, Decapoda), from the Pierre

Shale (Early Maastrichtian) of Colorado. Journal of Paleontology, 59:601-604.

. 1985&. Fossil decapod crustaceans from the Gammon Ferruginous Member, Pierre Shale

(Early Campanian), Black Hills, South Dakota. Journal of Paleontology, 59:605-624.

. 1987. Dromiopsis kimberlyae, a new Late Cretaceous crab from the Pierre Shale of South

Dakota. Proceedings of the Biological Society of Washington, 100(1):35— 39.

Brinster, K. F. 1970. Molluscan paleontology of the Pierre Shale (Upper Cretaceous), Bowman
County, North Dakota. Unpublished M.S. thesis, University of North Dakota, 136 p.

Feldmann, R. M., G. A. Bishop, and T. W. Kammer. 1977. Macrurous decapods from the Bearpaw
Shale (Cretaceous: Campanian) of northeastern Montana. Journal of Paleontology, 51:1 16 1—1 1 80.

Feldmann, R. M., and M. Maxey. 1980. Raninella carlilensis, a new raninid crab from the Carlile
Shale (Turonian) of Kansas. Journal of Paleontology, 54:858-861.

288

Annals of Carnegie Museum

vol. 56

Gill, J. R., and W. A. Cobban. 1966. The Red Bird section of the Upper Cretaceous Pierre Shale
in Wyoming. U.S. Geological Survey Professional Paper, 393-A:l-73, 12 pis.

. 1973. Stratigraphy and geologic history of the Montana Group and equivalent rocks. Mon-
tana, Wyoming, and North and South Dakota. U.S. Geological Survey Professional Paper, 776:
1-37.

Glaessner, M. F. 1969. Part R, Arthropoda 4(2), Decapoda. Pp. R399-R533, in Treatise on in-
vertebrate paleontology (R. C. Moore, ed.), Geological Society of America and University of
Kansas Press, Lawrence.

— . 1980. New Cretaceous and Tertiary crabs (Crustacea: Brachyura) from Australia and New

Zealand. Transactions of the Royal Society of South Australia, 104:171-192.

Kennedy, W. J., and W. A. Cobban. 1976. Aspects of ammonite biology, biogeography, and bio-
stratigraphy. Palaeontological Association Special Papers in Palaeontology, 17:1-94.

Rathbun, M. J. 1935. Fossil Crustacea of the Atlantic and Gulf Coastal Plain. Geological Society
of America Special Papers, 2:1-160.

Roberts, H. B. 1962. The Upper Cretaceous decapod crustaceans of New Jersey and Delaware, Pp.
163-237, in The Cretaceous fossils of New Jersey (H. G. Richards et al), New Jersey Bureau of
Geology and Topography, Bulletin 61, Part II.

Scott, G. R., and W. A. Cobban. 1986. Geologic and biostratigraphic map of the Pierre Shale in
the Colorado Springs-Pueblo area, Colorado. U.S. Geological Survey Miscellaneous Investigations
Series, Map 1-1627.

Tourtelot, H. A. 1 962. Preliminary investigation of the geologic setting and chemical composition
of the Pierre Shale, Great Plains Region. U.S. Geological Survey Professional Paper, 390:1-74.

Waage, K. M. 1961. The Fox Hills Formation in its type area, central South Dakota: Wyoming
Geological Association, 16th Annual Field Guidebook: 229-240.

Woods, J. T. 1953. Brachyura from the Cretaceous of Central Queensland. Memoirs of the Queens-
land Museum, 13:50-57.

Wright, C. W., and J. S. H. Collins. 1972. British Cretaceous crabs. Palaeontographical Society
Monographs, 126:1-114, 22 pis

. 1975. Glaessnerella (Crustacea Decapoda, Cymonomidae), a replacement name for Glaess-

neria Wright and Collins, 1972, non Takeda and Miyake, 1969. Palaeontology, 18:441.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 18, Pp. 289-318

20 November 1987

EXCAVATIONS AT THE HARNEY SITE SLAVE CEMETERY,
MONTSERRAT, WEST INDIES

David R. Watters

Assistant Curator, Section of Anthropology

Abstract

An archaeological salvage project at the Harney site, Montserrat, recovered parts of ten human
skeletons from burials exposed in trenches dug by construction workers during excavation of a house
foundation. Osteological analysis of these skeletons and additional bones previously removed by
construction workers identified 17 individuals. Several burial patterns were discernible despite the
disturbed nature of the Hamey site and incomplete condition of its skeletons. The patterns include
placement of bodies in graves dug to a rock stratum underlying the site, interment in a west-headed
direction, and arrangement of the corpses on their backs with legs extended. The presence of nails in
some graves provides suggestive but not conclusive evidence for the use of coffins or some other form
of burial apparatus. Recovered artifacts indicate this unmarked and unrecorded cemetery was used
at least in the latter eighteenth century and possibly longer. The most unusual artifact was an intact
Turlington Balsam of Life phial dated OCT 29 1751.

Introduction

When human skeletons were exposed by a construction crew digging a house
foundation near Bransby Point, the Montserrat National Trust asked the author,
who was on the island conducting prehistoric archaeological research, to examine
the remains. It soon became evident the construction crew had encountered an
unmarked and unrecorded cemetery site, a finding that led to a six-day salvage
excavation project between 29 June and 4 July 1979.

The burial ground was named the Harney site (MS-A-H4) in recognition of St.
Clair Hamey, the contractor who permitted the salvage project to take place
despite the disruption it caused in the construction schedule. Bones were first
uncovered by the construction crew about ten days before the author was scheduled
to leave Montserrat, after having completed his dissertation field research on
prehistoric sites (Watters, 1980).

This article treats the burial excavations and artifacts found at the Hamey site
salvage project. A second article (Mann et al., 1987) deals with the detailed
osteological analyses of the skeletons.

Site Description

Montserrat, with an area of 102 km2 (39.5 mi2), is one of the smaller islands
of the northern Lesser Antilles (Leeward Islands) in the eastern Caribbean Sea.
It is located on the inner or volcanic arc of the Lesser Antilles. Three major
volcanic centers trending north to south, Silver Hill, Centre Hills, and Soufriere
Hills, dominate the island’s landscape (Fig. 1). West of Centre Hills near the
leeward coast are two smaller volcanic centers, St. George’s Hill and Garibaldi
Hill. Watters (1980:103-136) and Steadman et al. (1984) provide additional in-
formation about Montserrat’s physiography.

The triangular rocky headland of Bransby Point, the westernmost point of land

Submitted 5 May 1987.

289

290

Annals of Carnegie Museum

vol. 56

62° 10'w

MONTSERRAT, WEST INDIES

Fig. 1.— Location of the Harney site in relation to Bransby Point on the west coast of Montserrat.

on Montserrat, is about 1.8 km NW of Plymouth, capital of the island (Fig. 1).
Bransby Point is formed by a horizontal bed of coarse but well-consolidated
agglomerate overlaying fine-grained tuff (MacGregor, 1939:39). Martin-Kaye (1 959:
fig. 12) labels the latter material “Bransby tuff.” A valley begins north of Bransby
Point at Fox’s Bay and trends east to the base of St. George’s Hill. The valley is
flanked by Garibaldi Hill on the north and the much lower Richmond Hill on
the south. Along the southern border of the valley, a ridge runs SSW to the leeward
coast.

The Harney site is located on the lower slope of that ridge (UTM coordinates
NP8 18484) near the 90-ft contour line. The site is separated from Richmond Hill
by Sandy Ghaut, a generally dry watercourse (Fig. 1), and it lies some 50 m inland
from the coastal cliff reaching a height of about 15 m above a narrow cobble
beach.

Construction work had disturbed the Harney site before the author was apprised
that bones had been found. Later interviews with the contractor established that
as much as two meters of soil had been removed from parts of the site by bulldozers
while leveling the landscape. Conspicuous and numerous backdirt piles attested
to the amount of overburden removed. It was while the construction crew, using
picks and shovels, was digging footers across the site that the skeletal materials
were first revealed.

Fragmented bones found in these trenches were initially regarded by the con-
struction crew as being “old cattle bones.” However, the foreman, Jackie Dangler,
who had been a military medic, thought the bones were human. The workers
concurred with Dangler’s evaluation after a jaw, undoubtedly human, was re-
covered from one trench. Dangler put the bones in a paper sack and buried it in
an isolated part of the site, away from the main digging area. Several days later,
the author was notified of the finds and went to the site to investigate the remains.

1987

Watters —Harney Site Slave Cemetery

291

Fig. 2.— View looking NW of the grid pattern of platforms (lower case letters) and trenches (upper
case) excavated by the construction workers (see Fig. 3).

Methodology

By the time the author arrived, the workers already had dug many footers and
produced a grid pattern of trenches bordering a series of raised platforms (Fig.
2). The tops of these platforms were artificial surfaces (not the original land surface)
created when the overburden was removed by bulldozers. Soil dug from the
trenches had been piled atop the platforms. Beyond the wood frame outlining the
building site were the backdirt piles left by the bulldozers.

Dangler indicated the areas in the footers where bones had been exposed be-
forehand. Fragmented bones protruding from the sides of various platforms in-
dicated that pick and shovel activity already had damaged some skeletons. Footers
had been dug through the soil layer into a rock stratum underlying the site.

Dangler also pointed out the location of the buried paper sack. When dug up,
it was found to contain many broken bones. The construction workers confirmed
that the breaks, clearly recent in origin, resulted from their pick and shovel activity.
These bones, which had been found in widely separated areas of the site, came
from more than one individual and, therefore, were termed the “multiple burial.”
The sack also yielded historic artifacts (a glass phial, three clay pipe stems, one
pipe bowl, and seven rusted nail fragments). These artifacts confirmed the skel-
etons to be historic rather than prehistoric burials.

The author was examining the items in the sack when the construction crew,
working in a trench in another part of the site, uncovered yet another burial. By
that time, it was evident the house foundation was being dug in a cemetery site.

It was decided that an archeological salvage project was the logical strategy to

292

Annals of Carnegie Museum

vol. 56

MN VGN

1-10 Skeletons

■ Reburied sack with
bones of 1,9, 10
^ Isolated tooth on
* surface
Q- 1 Platforms
A~0 Trenches
■ | Trenches dug before

L- J salvage project

kstowk] Trenches dug during
salvage project

K

M

O

(in east bank of
swimming pool eg. 10m.)

N

o

0 2 4 6 8 10 feet

HARNEY SITE (MS-A-H4 )
MONTSERRAT, WEST INDIES

Fig. 3.— Plan view of the Harney site showing locations of the in situ skeletons in relation to the grid.

adopt at the Harney site, given the previous disturbance, limited time for exca-
vation, and few personnel available. For six days, the author and four volunteers,
Cathy Watters, Marilyn Townsend, Gayle Baumgardner, and Joan Margolin,
excavated skeletons as they were exposed by the construction crew. Upon un-
covering a burial, the construction workers would suspend operations at that
location until the skeleton could be removed by the archaeologists.

Throughout the six-day project, the volunteers concentrated on removing the
bones while the author recorded excavation information and dug when possible.
Some standard excavation procedures were not used because of the previous
disturbance and limited time available. All horizontal control was based on the
grid pattern on the contractor’s blueprint rather than a site datum. Depth below
surface measurements were not recorded since original land surface no longer
could be determined. No attempt was made to screen soil from around the skel-
etons. However, several bulk samples of soil were taken from the graves. Pho-
tographs and field sketches were made; artifacts were recorded by individual burial
when possible; and a burial orientation measurement was taken for each skeleton
that was complete enough to warrant it.

Skeletons

Ten skeletons were identified during the salvage excavation at the Harney site.
Their locations are shown in Fig. 3 in relation to the grid pattern of footers and
platforms. Numbers applied to the skeletons generally correspond to the sequence
in which the burials were exposed. In Fig. 3, footers dug prior to the salvage
project are distinguished from those dug during that project. Four skeletons (1,

1987

Watters —Harney Site Slave Cemetery

293

2, 9, 10) were found beside previously dug trendies; five (3-7) were found in or
adjacent to trenches dug during the salvage project; one (8) occurred outside the
site grid.

Based on bones found to be articulated and in primary context when excavated,
the minimum number of individual (MNI) skeletons identified during the salvage
project was nine. Articulated remains were observed for one complete skeleton
and eight incomplete ones. Some articulated bones from each of the eight incom-
plete skeletons were seen in situ , and it was apparent that the missing bones had
been removed previously by bulldozers or picks and shovels.

In one case the archaeologists found bones that were in situ but were not
articulated. These disarticulated bones, which were assumed to be from one in-
dividual (designated as skeleton 9), raised the MNI to ten skeletons recovered by
the archaeologists during the salvage project. Identification of the ten skeletons
was based only on those bones observed in context in the undisturbed areas of
the site excavated by the archaeologists. Other bones from areas disturbed by the
construction workers or bulldozers were not observed in context and thus did not
count toward the MNI determined in the field.

Skeleton 1. — When first observed by the author, fragments of this skeleton were
protruding into trench stM” from the west comer of platform “h” (Fig. 3). When
excavated, the skeleton consisted of part of the innominate, some hand bones
(resting on the hips), and the lower extremities, all articulated. The right tibia and
the innominate were broken. Other bones, which were disarticulated, were found
under and on both sides of skeleton L Fig, 4 shows the bones of skeleton 1 (#5,
6, 7) and the disarticulated bones (#1-4, 8-9) of at least one other individual
(termed skeleton 9). The position of bone 4 of skeleton 9 is particularly illustrative
because it underlies the right femur and extends beneath the pelvis of skeleton 1
(Fig. 4). The emplacement of skeleton 1 caused the disarticulation of skeleton 9
(see discussion under Skeleton 9).

Skeleton 2.— This skeleton was exposed by the construction crew at the same
time the author was examining the materials from the paper sack. Part of the
skull and broken long bones were lying in trench “I” east of platform “d” (Fig.
3). A mandible was observed in situ in the east wall of that trench, near where
the footer ended. Most of skeleton 2 had been previously removed by the workers.

Skeleton 3. — Major damage had occurred to skeleton 3 before the author arrived
at the Harney site. The skeleton was located in trench t4G” between platforms
‘V’ and “f,” where the previously dug footer ended (Fig. 3). These bones were
well preserved, but all that remained were the lower extremities, from the distal
ends of the femora (Fig, 5).

Skeleton 4. —This is the most complete skeleton excavated. It was found near
the northeast comer of platform t6e” in trench “I” (Fig. 3), in the part of the trench
that was being dug while the archeologists were on the site. The skull was exposed
in trench “I” and the rest of the body extended eastward, outside of the grid
pattern. The cranium, which was collapsed and fragmented from the crashing
weight of overlying soil (Fig. 6), was removed first because it was in the trench
and liable to be damaged. When the post-cranial skeleton was uncovered, the
vertebral column was seen to be slightly bowed and not in line with the lower
extremities (Fig. 7). Skeleton 4 was fully articulated but the bones were flaking
and rather friable. The skeleton had been placed in a pit dug into the rock stratum
and displaced chunks of that rock were found resting on the skeleton and in the
soil above.

294

Annals of Carnegie Museum

vol. 56

Fig. 4.— Skeletons 1 and 9 in the west comer of platform “h” at the intersection of trenches “D” and
“M”. Bones 5-7 are from skeleton 1; bones 1-4 and 8-9 are from disarticulated skeleton 9. Skeleton
10’s lower leg bones were recovered later from the soil deposit in the upper left comer (under the
pencil).

1987

Watters— Harney Site Slave Cemetery

295

Fig. 5.— The lower extremities of skeleton 3. The white rock was positioned to support the left femur
for the photograph.

Skeleton 5. — Skeleton 5, another partially damaged one, was found in a narrow
pit east of platform “e” in trench “I,” not far from skeletons 3 and 4 (Fig. 3). It
contained part of the vertebral column and upper extremities as well as the
innominate and lower extremities (Fig. 8). The poor condition of the skeleton was

296

Annals of Carnegie Museum

vol. 56

Fig. 6. —Collapsed skull of skeleton 4 found in trench “I.”

initially perplexing. The damage could not have been caused by picks and shovels
because that part of trench “I” had not yet been dug by the construction crew.
In this case, the damage probably was caused by a bulldozer blade dipping deeper
at this spot while leveling the site’s surface.

Skeleton 6. —This skeleton was located near the intersection of trenches “C”

1987

Watters ==Harney Site Slave Cemetery

297

Fig. 7.— Post-cranial part of skeleton 4, with a bowed vertebral column that is not aligned with the
lower extremities. This skeleton, the most complete one excavated, was placed in a shallow pit in the
rock stratum.

298

Annals of Carnegie Museum

vol. 56

and “J” at the east corner of platform “b” (Fig. 3). Most of the bones had been
removed by construction workers when trench “C” was dug before the salvage
project began. When trench “J” was dug later, all that remained were fragments
of various long bones that seemed to be articulated.

Skeleton 7. — Skeleton 7 was found at the intersection of a previously dug trench

1987

Watters— Harney Site Slave Cemetery

299

(“C”) and a newly dug one (“L”). It was found on the south edge of platform “b”
about 3 m away from skeleton 6 (Fig. 3). Again, part of the skeleton was removed
during earlier construction work. Only the cranium was observed in situ. It was
intact when first observed but broke into pieces when excavated.

Skeleton 8. —This partial skeleton was positioned about 10 meters outside of
the grid pattern (Fig. 3). Lower leg bone fragments extended from the east wall
of the pit dug for the swimming pool. The rest of the bones had been bulldozed.

Skeleton 9. —Bones of skeleton 9 and skeleton 1 were found in the same soil
level of platform “h” beside trench “M” (Fig. 3). Skeleton 9 was not articulated;
its bones were scattered on both sides of the lower part of skeleton 1 (which was
articulated). Thus, the same undisturbed soil layer yielded part of an articulated
skeleton and bones from a disarticulated one (Fig. 4). Although the scattered
condition of skeleton 9’s bones indicated the skeleton had been disturbed at some
time, the disarray could not have been caused by the construction workers because
their digging had been limited to the trenches and did not intrude into the platform.
The soil layer in the platform was undisturbed when it was dug by the archae-
ologists.

Skeleton 9’s bones were disarticulated and scattered in the historic past rather
than recently. Skeleton 9 had been interred before skeleton 1. Gravediggers for
skeleton 1, upon finding skeleton 9 already in place, pushed aside skeleton 9’s
bones to make room for the new burial. In doing so, they created the disordered
condition of skeleton 9 that was observed upon excavation.

It is likely that some of skeleton 9’s scattered bones were removed by the
construction workers when they dug out the upper part of skeleton 1 in trench
“M.” An incongruous finding of mismatched bones in the opposite wall of trench
“M” (in platform “1”) supports this view. Fig. 9 shows part of a jaw (with an
adhering clay pipe stem) found adjacent to a long bone.

Skeleton 10. —After skeletons 1 and 9 had been excavated, part of another
skeleton was found nearby (Fig. 4). Skeleton 10 was located in the soil profile
above skeletons 1 and 9. Only the distal ends of the lower leg bones of skeleton
10 were recovered. Most of the skeleton, which had been positioned in trench
“D” beside platform “h” (Fig. 3), was removed by the construction crew before
the archaeologists arrived.

Osteological Summary

Subsequent osteological analysis (Mann et al., 1987) has established that 17
individuals are represented in the Harney site skeletal material. The discrepancy
between ten individuals found during the salvage project and 17 skeletons iden-
tified during laboratory osteological analysis is due to the construction activity
that occurred at the cemetery before the arrival of the archaeologists. The seven
additional individuals were identified from the “multiple burial” bones dug by
construction workers as well as other bones found by the author on bulldozer
backdirt pile surfaces. None of these bones was observed in situ by the archae-
ologists. The osteological study also determined age, sex, and race estimates for
some of the skeletons. Several individuals could reliably be raced as Black (Mann
et al., 1987), and it was this determination, in combination with the artifact dates,
that indicated the Harney site was a cemetery for Black slaves.

Burial Patterns

Ubelaker (1978: 13-18) indicates information should be recorded for a skeleton’s
location, deposition, position, orientation, and depth. Although some burial in-

300

Annals of Carnegie Museum

vol. 56

Fig. 9. —The mismatched bones (long bone and jaw fragment) possibly belonging to disarticulated
skeleton 9. A clay pipe stem fragment (see arrow) rests beneath the jaw.

formation was secured for most skeletons, it was not possible to record complete
information for all skeletons at the Harney site because of disturbance at the
cemetery, the partial condition of most skeletons, and limited time available.
Information on deposition, position, and orientation is presented below; location

1987

Watters —Harney Site Slave Cemetery

301

Table l,— Orientations of articulated skeletons at the Harney site . All are " west-headed ” interments.

Skeleton

Azimuth (magnetic)

Measurement taken at

1

243°

femora

3

250°

femora

4*

270°

vertebral column

281°

femora

5

256°

femora

8

263°

tibiae

* Legs and vertebral column were not aligned.

information has been presented above; no measurements for depth below original
land surface were possible because of previous removal of the overburden.

Deposition and Position

Six of the nine articulated skeletons (1, 3, 4, 5, 8, 10) were confirmed to have
been interred lying on their backs. Whether the other three articulated skeletons
(2, 6, 7) or the disarticulated one (9) were emplaced on their backs could not be
established because too few bones were observed in situ.

Position of the legs in relationship to the trunk could be determined with
certainty in only three skeletons (1, 4, 5), each of which retained at least part of
its pelvis. Legs fully extended from the pelvis in each instance. For skeleton 3,
the junction of the remaining distal, parts of the femora with the tibiae suggests
its legs probably were fully extended as well (Fig. 5).

Position of arms with respect to the trunk was observed for the same skeletons
(1, 4, 5) as the legs. The hands of skeletons 1 and 5 rested on their respective hip
(Figs. 4, 8). The right hand of skeleton 4 rested on its right hip but the left hand
was centered on the pelvis (Fig. 7). In no instance did hands actually cross over
one another on the pelvis.

Orientation

The direction in which the body was interred was established for eight artic-
ulated skeletons excavated at the Harney site. Direction was based on placement
of the legs alone in six cases (1, 3, 5, 6, 8, 10), of the skull alone in one case (7),
and of both legs and skull in one case (4). Legs consistently pointed toward the
east, while the top of the skull, in the two observed cases, was positioned toward
the west. Handler and Lange (1978:161) refer to this as the “west-headed” di-
rection.

Azimuths were taken for five skeletons that were complete enough to allow for
measurement. Orientations of those five skeletons are shown in Table 1. Most
measurements were taken between femora or tibiae because the upper parts of
the skeletons had been removed. Two azimuths were taken for burial 4 because
the vertebral column and legs were not aligned (Fig. 7). Azimuths range from
243° to 281° with most being slightly south of west.

Neither direction nor azimuth could be determined for skeleton 9, the disar-
ticulated one, or for skeleton 2, an articulated one having few bones observed in
situ.

Burial on “Shoal”

The author observed bone fragments protruding from the walls of some plat-
forms during his first visit to the Harney site. These fragments occurred at the
top of the rock stratum at the base of the soil layer. When the articulated skeletons

302

Annals of Carnegie Museum

vol. 56

Fig. 10. —The marked contrast between the white rock (“shoal”) below and the grey soil above is
evident near skeleton 2. Skeletons found in situ had been consistently buried at the interface of these
strata.

1987

Watters —Harney Site Slave Cemetery

303

were excavated, bones were again observed to rest atop the rock. A pattern of
interment at the interface of these two strata was consistent across the site. Persons
who originally dug the graves made it a practice to dig through the soil layer to
the rock stratum before interring the body. In a few instances, narrow and shallow
burial pits were dug into the top of the rock stratum. Chunks of this displaced
rock were found adjacent to and resting on the skeletons in several burials as well
as in the fill above.

Fig. 10 shows that the grey, friable soil contrasts markedly with the white,
compacted rock stratum, which is locally called “shoal.” This term has been used
in several ways in the past. Hardy (1922:191) uses shoal rock to refer to fragmental
volcanic material that has been consolidated into agglomerates, breccias, and tuffs
and then exposed on the surface of the land. His shoal soils “are typical sedentary
soils, for they have been derived from consolidated fragmental rocks which they
usually directly overlie” (Hardy, 1922:193). Later (Hardy et ah, 1949:9-10), the
terms geological shoal and pedological shoal are substituted, with the latter classed
as a hydromorphic soil that is subject to drainage impedence, hard cementation
(“resembles concrete”), and development of a characteristic clay type of soil. A
more recent study of Montserrat’s soils places shoal soils in the widespread Smec-
toid Clays category (Lang, 1967:15-17).

Conversations with the contractor and construction crew established the fact
that their use of the term “shoal” referred to the rock stratum upon which the
burials were found. Samples of this consolidated material and the overlying soil
near skeleton 2 were submitted for X-ray diffraction analysis and microscopic
observation. The report, “Analysis of soil and rock from Montserrat” (by G. A.
Cooke, 1986; unpublished report on file at Division of Anthropology, The Car-
negie Museum of Natural History) identifies the rock sample as “. . . a partially
welded tuff composed of trydimite/cristobalite, anorthite, magnetite, amorphous
silica and quartz” (p. 3). This finding confirms the tuff identifications of MacGregor
(1939:39) and Martin-Kaye (1959: fig. 12) for the Bransby Point area. The report
also states (p. 2): “The mineralogy of this soil is entirely consistent with a weath-
ered ash. The lack of a well developed smectite indicates a Late Pleistocene to
Recent age.” Mineralogical similarities between samples of soil and tuff indicate
this tuff, or one very similar to it, is the parent material for the soil.

The Coffin Issue

The presence of nails (see discussion under Artifacts) suggested that interment
in coffins may have been another burial pattern at the Harney site. The possibility
of coffin interment was recognized early in the project when seven nail fragments
were found in the paper sack.

Efforts to clarify the coffin issue took two forms. Initially the author examined
areas around the bones that protruded from the platform walls for evidence of
coffins. During subsequent excavation of the in situ burials, the volunteers were
especially attentive to any indicators of coffin remains.

Nails found in the graves continued to be the only indicator of possible coffin
interment at the Harney site. Nails were found adjacent to and in the fill above
five skeletons (3 females, 2 males; Mann et ah, 1987: table 1). Despite careful
observation of the platform walls and graves, no fragments of coffin wood, no
discolored or stained soil from decomposed coffins, and no coffin hardware such
as handles were found.

The issue of coffin burial at the Harney site remains a perplexing one. If one

304

Annals of Carnegie Museum

vol. 56

Table 2.— Distribution of artifacts by excavated burial.

Imported

Burial3

Nailsb

Pipe stem

Colonoware

pottery

Metal disc

Total

1 (+9?)

33(17)

1

8

_

1

43

3

A

13 (9)

1

1

—

—

14

1

32

1

5

q.

5

£

32(12)

1

-

--

0

8

4 (0)

_

1

__

10

_7J2)

—

—

—

—

7

89 (40)

2

10

1

1

103

No artifacts in burials 2 and 7.

b Fragments first; nail heads (MNI) in parentheses.

accepts the occurrence of nails as suitable evidence for the presence of coffins,
then coffin burial took place in at least five of the excavated graves. Occurrence
of coffins would be a reasonable explanation for the presence of the nails in those
five graves. It is difficult to account for why else nails would be found.

Barring a reasonable alternative explanation for the presence of nails, it seems
logical to conclude that coffins, or some kind of burial apparatus (e.g., planks)
requiring nails, were used in some graves at the Harney site. Handler and Lange
(1978:150), in their work at Newton Plantation, Barbados, used the presence of
nails to indicate occurrence of coffins when other evidence was absent. Yet, in
some graves they did find additional hardware, such as coffin handles, that verified
use of coffins at Newton Plantation. At the Harney site, the presence of coffins or
other burial apparatus cannot be independently corroborated because we found
no coffin wood, other coffin hardware, or even discolored soil.

Artifacts

A total of 1 34 artifacts were recovered from three locations at the Harney site—
the graves, paper sack, and backdirt piles. The range of artifacts includes nails,
other metal items, clay pipes, “colonoware” pottery, imported earthenware and
stoneware ceramics, and glass.

Artifacts by Burial

Only those artifacts found in excavated graves can be related to the ten burials.
Of the 134 artifacts recovered, 103 (77%) are from graves. Although some artifacts
occurred in proximity to the bones, others were recovered from soil above the
skeletons and may have been unintentionally included when the graves were being
filled. It is not certain that any of these artifacts were purposefully deposited as
grave goods to accompany the deceased.

Table 2 presents the distribution of artifacts by burial. Seven of the ten burials
had artifacts; two burials with incomplete skeletons (2, 7) had no artifacts; if there
were artifacts from burial 9 (disarticulated skeleton), they could not be distin-
guished from those for burial 1. Of the seven burials with artifacts, two (4, 6) had
only one artifact apiece. The fact burial 4 has but one artifact is of special note
because that grave contained the skeleton that was most complete when excavated.

Nails

Iron nails found in the excavated burials are all heavily corroded. No attempt
was made to measure the nails because of the corrosion, but it is worth noting

1987

Watters —Harney Site Slave Cemetery

305

that there does appear to be variation in both length and diameter among the
whole nails. No effort was made to determine the shape of a nail head because it
is a rusted lump without clear form. These seem to be wrought rather than cut
nails.

The 89 nail fragments are by far the most abundant artifact (86%) from the
burials (Table 2). There are 40 nail heads, each representing an individual nail,
while the other fragments are broken shanks.

Nails were found in five of the excavated graves (1, 3, 5, 8, 10). In Table 2,
some of the 1 7 nails attributed to skeleton 1 may have been originally associated
with disarticulated skeleton 9. However, two nails found immediately adjacent
to the right knee and lower tibia of skeleton 1 can be reliably associated with that
skeleton. The partial skeletons from the other four graves, all of which were
individual interments, yielded a wide range of nails (Table 2). Mann et al. (1987)
found one nail adhering to the left femur of skeleton 10.

Nails were not found with three of the incomplete skeletons (2, 6, 7). Their
absence may be due to the fact that very few bones remained in situ when the
archaeologists arrived because the construction crew already had removed most
of the skeletons. However, the same logic does not apply to burial 4, which also
yielded no nails. Absence of nails in that grave cannot be attributed to the con-
struction workers’ action since the skeleton was complete when excavated by the
archaeologists. In the case of skeleton 4, the lack of nails implies that this indi-
vidual originally was interred without a coffin.

Pipe Stems

Pieces of clay pipe stem were recovered from two excavated burials (1,4) both
of which were females (Mann et al., 1987: table 1). One tapering piece (length =
57.1 mm; maximum diameter = 6.3 mm) from the rear of the stem was found
with skeletons 1 and 9. It has a stem hole diameter of 4/64ths of an inch. A second
stem fragment, shorter but thicker (length = 41.8 mm; maximum diameter = 9.3
mm), was the sole artifact found with skeleton 4. Its stem hole diameter is 5/64ths
of an inch. Both stem hole diameters predominated during the eighteenth century
(Noel Hume, 1969c:298). No pipe bowls were found in the excavated graves.

"Colonoware Pottery”

Ten small sherds from three burials at the Harney site have been labeled “co-
lonoware,” a term that implies the pottery was made locally and probably by
slaves, although that is not certain.

Three distinct vessels have been identified. One vessel, based on one body
sherd, occurred in burial 6. Another vessel, also based on a single body sherd,
was found in burial 1 . The third vessel is of special interest because seven of its
sherds (one rim, six body) were found with burial 1 while one body sherd occurred
with burial 3. Colonoware sherds were found with male and female skeletons.

The vessels are coil constructed with generally well finished surfaces but without
decoration. Sherds are predominately gray or brown but there is considerable
variation in color even for a single sherd. This reflects a reduced firing environment
with typically fast cooling. Temper is principally derived from a sort of quartz
dominated composite. Hornblende is also present as are limestone and grog in
lesser amounts.

It is unlikely these sherds are the broken remnants of purposefully placed grave
goods because most were found in the fill overlying the skeletons. Also, sherds
from the same vessel were found in two separate burials (1 and 3) located some

306

Annals of Carnegie Museum

vol. 56

Table 3.— Distribution of disturbed artifacts.

Location

Nails2

Pipe stem

Pipe bowl

Colonoware

Imported

pottery

Glass

Other metal

Total

Sack

7 (3)

4b

1

—

....

lc

_

13

Backdirt

-

-

1

9

6

l

18

7 (3)

4

2

9

6

2

1

31

a Fragments first; nail heads (MNI) in parentheses.
b Includes pipe stem found with jaw and long bone.
c Turlington phial.

10 m apart. A more plausible explanation is that the sherds were unintentional
introductions into the graves when they were filled.

Imported Ceramics

A single unglazed earthenware body sherd was found beside skeleton 8 in the
wall of the pit dug for the swimming pool. This sherd has been separately classed
because it is distinctly different from the “colonoware” pottery. The earthenware
sherd is red (2.5YR5/8), thick (about 13 mm), and has a uniformly fine-grained
paste. Utilitarian earthenwares of this sort were imported into the West Indies in
great quantity. Such sherds are commonplace in plantation sites.

Metal Disc

An unusual object found in the fill above skeletons 1 and 9 is a metal disc
somewhat ovoid in shape. The maximum diameter is 23.8 mm; the minimum
diameter is 21.4 mm; the thickness is 0.6 mm. Green patination on its surface
indicates the disc is a copper alloy. The item’s shape at first suggested it might
be a coin, but subsequent microscopic examination revealed there was no in-
scription visible. Its sheared edge indicates the disc was cut and its surface appears
to be hammered. It does not appear to be a coin or a button. The function of the
artifact remains uncertain.

Handler and Lange (1978:201) note that a Barbadian document refers to the
practice of putting coins over the eyes of the deceased. They did not find any
examples of this practice among the Newton Plantation skeletons. Although it is
possible that the Harney site disc was meant to function as an eye covering, this
seems unlikely since the item was not found with skeleton 1 or 9 but instead in
the fill above.

Disturbed Artifacts

Thirty-one artifacts, which had been removed from primary context by the
construction crew or bulldozer, were collected by the author during the project.
The paper sack contained 1 3 of these artifacts. Discussion with the crew revealed
that these artifacts had been found with various bones in different trenches. Another
1 8 artifacts were collected by the author from backdirt piles created when the site
surface was leveled by bulldozers. Table 3 shows the distribution of artifacts from
the sack and backdirt piles. Discussion about these disturbed artifacts is limited
because they cannot be definitely associated with particular skeletons.

Three clay pipe stem fragments and one bowl were found in the sack. A fourth
stem fragment, found with the mismatched jaw and long bone (see discussion
under Skeleton 9), fits with one stem from the sack. All pipe stem hole diameters
are 4/64ths of an inch. The pipe bowl from the sack has no heel but does have

1987

Watters— Harney Site Slave Cemetery

307

raised relief on the front and back of the bowl in the form of a branching stemmed
device. The relief vaguely resembles that on a pipe bowl found at Newton Plan-
tation (Handler and Lange, 1978: fig. 37). The bowl is scorched very little from
smoking. A heavily charred pipe bowl from the backdirt pile is plain and lacks a
heel There are no markings on any of the pipe fragments indicating the name or
location of the maker.

The one rim and eight colonoware sherds from the backdirt piles, which are
very similar to the ones excavated, have somewhat more variation in color but
the surface and temper are almost identical The nine sherds come from at least
four vessels. These are four different vessels from the ones identified from ex-
cavated colonoware sherds.

Six sherds of imported ceramics came from bulldozed backdirt piles. The paste
(color and texture) in one body sherd is very similar (apart from a gray core) to
the earthenware sherd found with skeleton 8, but the exterior is black glazed. It
resembles the Buckley pottery of North Wales described by Noel Hume (1969c:
133). The second earthenware piece is an unglazed reddish yellow (5YR6/6) rim
with a gray core. Two body sherds of tin-enamelled decorated delft. ware, one with
the remnant of a cobalt blue design and the other with manganese purple and
cobalt blue (Noel Hume, 1969c: 106), were found in the backdirt piles. The fifth
earthenware piece is a thin, plain creamware body sherd. The sixth sherd is
stoneware with a light red (2.5YR6/6) and ribbed interior, gray to cream paste,
and brown-mottled exterior. The shiny exterior, which lacks the pitting associated
with salt-glazed ceramics, suggests this may be Nottingham stoneware or a similar
variety (Noel Hume, 1 969c: 1 14). All of these ceramics were in production for at
least part of the eighteenth century (South, 1977: table 31).

One intact glass phial was found in the paper sack. One straight sided, thick
(7.9 mm) fragment of green glass and a corroded, wedge-shaped metal object
(possibly a shim) were recovered from the backdirt piles.

“ Missing ” Artifacts

One category of artifacts that might be expected to be present in the burials
was not found. No buttons, pins, or other clothing accoutrements were recovered.
It is possible that some of these materials, which might have originally accom-
panied the corpse to the grave, decomposed over the years. Mann et al (1987)
found some evidence to support this notion for two female skeletons (2, 4) in the
form of green copper salts stains that suggest use of a pin or ornament. It is also
possible that the corpses were interred in shrouds, which do not necessarily require
pins or other fastening devices, or naked.

A second category of absent artifacts includes those of African origin or deri-
vation. Several artifacts of this kind were found in the Newton Plantation graves
on Barbados (Handler, 1981; Handler and Lange, 1979). No such artifacts were
found at the Harney site.

The Turlington Balsam Of Life Phial

Preliminary information on the Turlington Balsam of Life phial from the Har-
ney site has been presented previously (Watters, 1981). That paper described the
phial and compared it with a similar bottle found in the “abortive well” near
Wetherbum’s Tavern in Williamsburg, Virginia (Noel Hume, 196 97?: 3 4). The
author had not seen the Williamsburg bottle before publication of the 198 1 paper
and, therefore, was uncertain whether the Montserrat artifact was an authentic

Fig. 11.— Four views (actual size) of the “cello-shaped” Turlington Balsam of Life phial (clockwise
rotation sequence).

308

1987

Watters “-Harney Site Slave Cemetery

309

Turlington Balsam of Life phial or one of the many counterfeit bottles produced
in the mid 1700s. In 1982, the author had the opportunity to observe the Weth-
erburn’s Tavern bottle and compare it to the Montserrat phial. It was evident the
two bottles basically were very similar, apart from the date, and that both were
well crafted in comparison to counterfeits. For these reasons, the Montserrat phial
is now regarded as another genuine Turlington bottle.

Location and Recovery

The Montserrat phial was first observed by the author as one of the artifacts
contained in the paper sack that had been buried in another part of the Harney
site by the construction workers. Interviews with the workers revealed that the
bottle had been found near the intersection of trenches “D” and “M” at the west
comer of platform “h” (Fig. 3). They recalled the location where they found the
phial because it was intact and therefore noteworthy, which aroused their curiosity.
The bottle was found in the soil above the shoal stratum but its exact location in
the soil profile was not remembered. It is remarkable the bottle was not shattered
by picks or shovels as the soil and shoal were being dug from the trenches.

Portions of three skeletons (1, 9, 10) eventually were excavated at the west
corner of platform “h.” Part of each skeleton had been removed when the two
trenches were dug by construction workers; in addition, the bones of skeleton 9
already had been disarticulated in the historic past. Although it is logical to
conclude that the Turlington bottle originally was emplaced, whether intentionally
or not, with one of these three skeletons, it is unlikely we will ever know with
which one in particular.

Turlington’s Balsam of Life

Robert Turlington patented his “Balsam of Life” in Great Britain in 1744 and
the patent medicine, which contained no less than 27 ingredients, soon became
quite popular (Griffenhagen and Young, 1959:160). Its success quickly attracted
some unscrupulous competitors, whose actions caused numerous problems for
Turlington. These competitors bought used Turlington bottles and filled them
with spurious material for resale; eventually they began to produce counterfeit
bottles bearing Turlington’s name (Noel Hume, 1969c:74). Turlington’s efforts to
thwart competitors appear to have been unsuccessful for the most part.

The Williamsburg bottle, which is dated MARCH 25 or 26 1750 (a crack
obliterates the questionable digit), is the earliest known example of Turlington’s
proprietary bottle. It is the “cello shaped” variety (Noel Hume, 1969a: fig. 38;
19696: fig. 24). The Harney site phial, dated OCT 29 1751, is also cello shaped
(Fig. 1 1). The dates and other raised inscriptions on the bottles were imparted by
the molds into which the glass was blown. The fact that these two almost identically
shaped bottles were made in different molds and have dates that are only twenty
months apart may represent an effort by Turlington to authenticate his product
by changing production dates on his phials. If so, the practice apparently was not
successful because Turlington eventually altered the shape of the bottle itself in
1754 to one with a distinctly angular form (Jones and Smith, 1985: Fig. 115).
Notwithstanding Turlington’s efforts, his competitors were soon producing an-
gular-shaped counterfeit bottles (Noel Hume, 1969a: fig. 39).

Comparison

The Wetherbum’s Tavern and Harney site phials are similar in many ways.
Both are of swirled, clear glass. The “. . . crude rendering of the British royal arms

310

Annals of Carnegie Museum

vol. 56

Fig. 12. —Details (enlarged) of the British royal arms (corresponds to Fig. 1 Id).

and their supporters . . .” (Noel Hume, 1969 b: fig. 24 caption) above the words
BY THE KINGS PATENT occurs also on the Montserrat phial (Fig. 12).

An embellishment on the reverse side of the Montserrat phial is difficult to
discern clearly, but it may correspond to the shield with three pear-shaped devices
described by Noel Hume (1969Z?: fig. 24 caption) for the Wctherb urn’s Tavern
bottle. The embellishment seems to be a shield having three projections arranged
horizontally across its center (Fig. 1 3). Above the shield and articulating with it
is a rounded projection. A raised, curved device encloses the shield on each side
and joins the projection above and almost touches yet another projection below
the shield. Branch-like emblems join the curved device on each side. Beneath the
shield are the words ROBT TURLINGTON.

Both phials have the words BALSAM OF LIFE arranged along the curves on
one edge. On the Harney site bottle the letters BAL are somewhat larger than
SAM and the latter appear to be somewhat compressed (Fig. 12, 14). The Wil-
liamsburg bottle has letters of more equal size.

The other edge of each bottle has its date. Although the dates are different, the
arrangement of letters and numbers, with the month, day, and year fitting into
the curves of the bottle’s edge, is the same. The numeral “ 1 ” is decidedly J-shaped
and numerals “9” and “7” have tails on the Montserrat phial (Fig. 15).

Apart from the date, there are a few other differences between the two bottles
including the rendering of the lip, symmetry of the shoulders, and arrangement

1987

Watters —Harney Site Slave Cemetery

311

Fig. 13.— Details (enlarged) of the shield with three pear-shaped devices (corresponds to Fig. 1 lb).

of various projections around specific words (Watters, 1981:107). All of these are
variations specific to a mold.

There is a slight variation in height between the Williamsburg and Montserrat
bottles. The lip to base measurement for the Montserrat phial is 2ll/32 inches. The
Williamsburg bottle (excluding the slightly projecting pontil) measures 2% inches
(based on a remeasurement made in 1980 by Noel Hume to whom I am indebted;
see Watters, 1 98 1 : 1 07, for a discussion of other heights given for the Wetherbum’s
Tavern bottle).

There is some variation in details of the Montserrat phial depicted in the
illustrations accompanying this paper. Certain parts of the phial showed consid-
erable wear and that tended to obliterate some of the details or make them difficult
to discern. Also, Fig. 1 1 was drawn from the bottle itself by one illustrator, whereas
Fig. 12-15, drawn by another person, were based on prior illustrations and pho-
tographs since the bottle itself already had been returned to the Montserrat Na-
tional Trust, where it has been displayed in the museum since 1983.

The Montserrat and Williamsburg phials are examples of authentic Turlington
bottles that are noteworthy because of their overall similarity in form and in-
scription. The minor differences between these two bottles can be attributed to
variation in the molds. The difference that is most intriguing is the twenty month
separation in mold dates, especially if it really indicates Turlington’s effort, albeit
apparently unsuccessful, to block the actions of his unscrupulous competitors.

312

Annals of Carnegie Museum

vol. 56

Fig. 14.— Details (enlarged) of the BALSAM OF LIFE lettering (corresponds to Fig. 1 la).

Discussion
The Cemetery Site

The existence of a cemetery at the Harney site was first revealed by the exca-
vation work for the house foundation. According to the contractor and construc-
tion workers, the surface of the site contained no indications of this unmarked
and unrecorded cemetery. It was only after they began to dig the trenches and
encountered bones that they suspected the presence of human burials.

Historic structures are not found in the immediate vicinity of the cemetery.
The closest historic buildings are located about 500 m north of the Harney site.
A recent map of Montserrat (Directorate of Overseas Surveys, 1978) shows these
structures as “building remnants” (UTM coordinates NP8 19488). It is possible
they represent the remnants of the Bransby Plantation, noted on earlier historical
maps of Montserrat, although that is uncertain. It also is uncertain that the cem-
etery is related to those buildings. Other historic structures may occur in the
general vicinity of the Harney site, but, because of time constraints, no systematic
survey was conducted to investigate their occurrence.

One intriguing development occurred about four months after the Harney site
excavations. In October 1979, the Montserrat National Trust found the remains
of a gun battery including stone fortifications and the guns themselves at the tip

1987

Watters —Harney Site Slave Cemetery

313

NUP'84

Fig. 15.— Details (enlarged) of the OCT 29 1751 date (corresponds to Fig. 1 1c).

of Bransby Point (Fig. 1) about 1.2 km WNW of the Hamey site. This discovery
raised the possibility that the Harney site skeletons may have been those of soldiers
or militiamen killed during one of the attacks against Montserrat, perhaps one of
several raids made by the French.

The possibility of the skeletons being soldiers added to the list of other possi-
bilities that included Black slaves, White indentured servants, and freedmen. All
were present on Montserrat during the colonial era. Indentured servants, especially
Irish, were present in considerable numbers but later were largely replaced by
Black slaves (Pulsipher, 1 977). The lack of burial accoutrements and the unmarked
nature of the graves suggested that the burial ground probably was not for plan-
tation owners or managers. However, if the deceased had died during an epidemic
and been hurriedly buried, then it is possible that the skeletons were of those of
plantation managers. Given the range of possibilities, an analysis by physical
anthropologists was needed to answer these questions.

Osteological study (Mann et al., 1987: table 1) has now shown that the two
skulls recovered (4, 7) from the in situ burials are definitely from Black individuals.
Two bones from the disturbed “multiple burial” are identified as Black (one is
provisionally identified). Race was not determined for post-cranial skeletons from
in situ burials or for the other disturbed bones. They are simply listed as “race
indeterminate” (Mann et al., 1987: table 1). Although it is likely that all the

314

Annals of Carnegie Museum

vol. 56

skeletons from the Harney site are Black individuals, that is not certain. The
possibility remains that at least some are White individuals.

The designation of these individuals as slaves rather than freedmen is primarily
based on the most likely time of use of the Harney site cemetery in the latter
eighteenth century, before slavery was abolished in the British West Indies. The
1751 date inscribed on the Turlington Balsam of Life phial provides the most
secure date for use of the cemetery. Although the bottle obviously could not have
been placed in the grave prior to its manufacture in 1751, it could have been
emplaced any time thereafter. It most likely was deposited quite some time after
1751 because of the lag time between production in England and shipment to the
West Indies. How the Turlington bottle ultimately came to be placed in a slave’s
grave is an intriguing question that probably never will be answered.

The Turlington phial is the only artifact bearing a date. However, there are two
other categories of artifacts that support the view that the cemetery was used in
the latter eighteenth century. One pipe stem from an in situ burial has a bore
diameter of 5/64ths of an inch; the second in situ pipe stem diameter is 4/64ths
of an inch. The former size predominated in the first half of the eighteenth century
and the latter in the second half. Four pipe stems and two pipe bowls from the
sack and backdirt piles also have 4/64ths inch diameters. Since the four stems
and two bowls were found out of context, they cannot directly support the ar-
gument for a latter eighteenth century use of the cemetery. However, they are in
accord with and do not contradict the findings for the pipe stems found in situ.
The imported ceramics, for which manufacturing ranges are known, were all
produced for at least part of the eighteenth century. Again, this is only indirect
support because they were found out of context.

The six-day project was able to salvage remains from 17 individuals, but it is
unlikely that this represents the total skeletal population buried at the Harney
site. The distribution of recovered skeletons (Fig. 3) tends to cluster around plat-
forms “b, h, and d-e.” However, this is because archaeological work was restricted
to the confines of the house foundation and, apart from the swimming pool, areas
outside were not investigated. The probability that additional burials still exist
at the Harney site is high because much of the potential burial area was not tested.
Since the now completed house stands on the site, the possibility of further testing
is limited.

Correlations

The Harney site skeletons are the only slave skeletons known from Montserrat.
Burials of prehistoric Amerindians have been found on occasion, the most recent
of which was within the confines of the capital of Plymouth a few years ago.

Significant work on plantation archaeology has been underway at Galways
Plantation, located in southwest Montserrat on the slope of the Soufriere Hills
(Fig. 1). Pulsipher and Goodwin (1982:83-85) mention that two burying grounds
may have been identified at Galways. One presumably is for Whites because it
has headstones. The other, which consists of mounded piles of rocks, was thought
to possibly contains slave burials. More recently, further research (Goodwin, 1987:
95-104) has been carried out in the slave village at Galways. Goodwin (personal
communication 1987) suspects the rock mounds do not contain burials but instead
were erected by small plot cultivators in later years. If that is the case, then the
location of the Galways slave burial ground remains to be discovered. Work at
Galways planned for the summer of 1987 may help resolve the issues related to
its cemeteries.

1987

Watters— Harney Site Slave Cemetery

315

Although studies of plantation archaeology have occurred elsewhere in the
Caribbean, such as Armstrong’s (1983, 1985) work at Drax Hall, Jamaica, the
research generally has not included excavations of slave cemeteries. On the other
hand, skeletons of West Indian Blacks have been exhumed inadvertently on oc-
casion. Ubelaker and Angel’s (1976) brief report on two Black skeletons from St.
Thomas, U.S. Virgin Islands, essentially falls into this category.

The most significant study on Black slave burials in the West Indies is on the
Newton Plantation site on Barbados. That study used historic documents and
archaeological field research to provide a broad-based background on slave life
in Barbados (Handler, 1972; Handler and Lange, 1978, 1979). Much of the earlier
work focused on information that could be gained from analysis of the cemetery
site, burial patterns, and artifacts in conjunction with historic documents (Lange
and Handler, 1985).

More recently, the skeletal material has been examined from physical anthro-
pological perspectives to gain information about the health, diet, and diseases of
the slaves (Corruccini et al., 1982; Handler and Corruccini, 1983). While some
studies have dealt with quite specific topics, such as tooth mutilation (Handler et
al, 1982), others have more general intent, such as evidence for weaning (Cor-
ruccini et al., 1985; Handler and Corruccini, 1986). One of the most interesting
revelations is the recent discovery of high lead content among Newton Plantation
skeletons, which apparently resulted from consumption of rum contaminated by
lead parts used in plantation distilleries (Handler et al., 1986).

The total number of individuals recovered from the Newton Plantation site is
104 (Corriccini et al., 1985), which is 12 more than originally reported (Handler
and Lange, 1978), three more than reported after the first physical anthropology
study (Corruccini et al., 1982) and about six times greater than for the Harney
site where 1 7 individuals were identified (Mann et al., 1987). The interment period
at Newton Plantation extends from about 1660 to 1820. At the Harney site,
burials probably took place in the latter eighteenth century (possibly longer), which
is a much more restricted timespan.

There are interesting similarities between the populations. At Newton Planta-
tion, west-headed orientation predominated although skeletons with east and
rarely north orientation were also found (Handler and Lange, 1978:161). Of par-
ticular interest is the stratigraphic and artifactual evidence that indicates west-
headed interments at Newton Plantation tended to occur more recently than the
east-headed ones. In general, the Barbados data show that the more recent the
burial, the more likely it was to be oriented toward the west. The Harney site
graves, which would correspond to the later part of the Newton Plantation burial
timespan, would be expected to yield skeletons emplaced in a west-headed di-
rection. At the Harney site, the orientation was west-headed for eight burials; two
others could not be determined.

Also of interest, confirmed coffin burials at Newton Plantation were all west-
headed (Handler and Lange, 1978:163). At the Harney site, the five corpses that
may have been interred in coffins were all west-headed. Bodies were placed on
their backs with legs extended at the Harney site; most bodies from Newton
Plantation were also interred that way.

For adult skeletons at the Harney site that could be sexed, the ratio of males
to females is about even. Four subadults were identified; no fetal or infant skeletons
were found. Overall, males appear to have died at earlier ages than females. Mann
et al., (1987) discuss age, sex, race, and stature estimates as well as paleopatho-
logical conditions in detail.

316

Annals of Carnegie Museum

vol. 56

Apart from nails, relatively few artifacts were found with the in situ burials
(Table 2) at the Harney site. Although some artifacts were found near the skeletons,
others were found in the fill above. Because it is not certain that any one of these
objects was purposefully emplaced with the deceased, it is better not to regard
them as grave goods per se. Handler and Lange (1978: table 16) record a wider
variety of artifacts from Newton Plantation burials. However, in general they
note that “very few grave goods were encountered at Newton . . .” (Lange and
Handler, 1985:26), and those found were mainly glass beads and clay pipes. A
major difference was that artifacts of African origin or derivation were found at
Newton Plantation but not at the Harney site.

Burial in mounds, a mortuary pattern that was identified at Newton Plantation,
cannot be addressed at the Harney site because the surface had already been
bulldozed before the archaeologists arrived. A second Newton Plantation mor-
tuary pattern, family burial plots (Corruccini et al., 1982:446), seems related to
the existence of the mounds and, therefore, also cannot be addressed at the Harney
site.

In summary, some of the burial patterns, including west-headed orientation,
arrangement of corpses on their backs, paucity of artifacts, and possible interment
in coffins for some burials, exist at both the Harney site on Montserrat and the
Newton Plantation on Barbados. The practice of burying the corpse atop the shoal
stratum at the Harney site would be a site-specific phenomenon due to the geology
of the Bransby Point area. The Harney site was used as a cemetery at least in the
latter eighteenth century and its burial patterns seem to correspond to the later
burial timespan at Newton Plantation.

Anthropological Observations

During the course of the archaeological excavations at the Harney site, we
observed several behavioral patterns among the construction workers that are of
passing interest. None of the workers would actually touch the skeletal remains
we excavated. In the one instance when a worker did bring us a bone, he did not
actually touch the bone but instead had it suspended from a thin loop of vine.
Their decided reluctance to handle bones most likely derives from the persistence
of obeah (magic) practices on Montserrat (Dobbin, 1983:26-31). The workers’
concern was as much for our well-being as for theirs.

There is no doubt that the workers initially were uneasy about us working at
the site. Eventually, however, they displayed great interest in the work we were
doing and understood the careful excavation techniques used, once we explained
why these methods were important. Jackie Dangler was the person most respon-
sible for allaying the workers’ initial concern about our activities because he was
the foreman as well as a former medic. Although they became used to the idea
that we could work together without harm during the day, the workers expressed
considerable worry about my well-being when I stayed alone at the site several
evenings to continue the excavations. Having been advised to be wary of the
spirits, I suspect my decision to continue working after the construction crew left
at the end of the workday was viewed as foolhardy.

Undoubtedly the most interesting activity occurred the day after we completed
the excavations. I brought a bottle of rum to the site to share with the workers
in appreciation of the assistance they had given to us during the project. I was
very intrigued when, before any of us partook, Jackie Dangler walked to the house
foundation and sprinkled small amounts of rum into the trenches where we had

1987

Watters “Harney Site Slave Cemetery

317

found the skeletons. He and the construction crew expressed their satisfaction
that now since “they” (the spirits) had partaken, we could enjoy the remainder
of the rum. This activity was interesting in its own right, but I later became even
more intrigued when I read a statement (cited in Handler and Lange, 1978:199)
concerning the practice of spilling rum on a slave’s grave, based on an observation
made in Jamaica in 1707.

Acknowledgments

Field research in 1978-79 was supported by a Fulbright-Hays Doctoral Dissertation Research
Abroad Fellowship (DHEW:OE) and an Andrew Mellon Predoctoral Fellowship at the University of
Pittsburgh. We are grateful to members of the Montserrat National Trust for their assistance and
cooperation, especially Samuel McChesney, then President of the Trust, and Percy Arthurton, then
Chairman of its Archaeology Subcommittee. We also acknowledge St. Clair Harney, Jackie Dangler,
and the construction workers for their cooperation at the site, Marilyn Townsend, Gayle Baumgardner,
Joan Margolin, and Cathy Watters for volunteering their time for the excavations, and Ivor Noel
Hume for his answers to questions about the Williamsburg bottle. Figures 1 and 12-15 were drawn
by Nancy Perkins; Fig. 3 and 1 1 by Lucinda Tear; photographs taken by the author were printed by
Stanley Lantz. Funds provided in 1 983 by the Anne L. and George H. Clapp Charitable and Educational
Trust of The Carnegie Museum of Natural History allowed us to return the Turlington phial to
Montserrat in the course of other field work. Comments on drafts of this paper by Verna L. Cowin,
James B. Richardson III, Jerome S. Handler, Lydia M. Pulsipher, and Conrad M. Goodwin, were
greatly appreciated. I also thank James B. Petersen for studying the colonoware pottery, Ronald C.
Carlisle for looking at the nails, and Goodwin for information about the rock mounds at Galways.
Finally, I express my gratitude to Frederick W. Lange and David Hurst Thomas for their willingness
to review this manuscript.

Literature Cited

Armstrong, D. V. 1 983. The “Old Village” at Drax Hall plantation: an archaeological examination
of an African- Jamaican settlement. Unpublished Ph.D. dissertation, University of California, Los
Angeles, 42 1 pp.

. 1985. An Afro-Jamaican slave settlement: archaeological investigations at Drax Hall. Pp.

261-287, in The archaeology of slavery and plantation life (T. A. Singleton, ed.), Academic Press,
Orlando, 338 pp.

Corru ccini , R. S., J. S. Handler, and K. P. Jacobi. 1985. Chronological distribution of enamel
hypoplasias and weaning in a Caribbean slave population. Human Biology, 57(4):699-71 1.

Corruccini, R. S., J. S. Handler, R. J. Mutaw, and F. W. Lange. 1982. Osteology of a slave
burial population from Barbados, West Indies. American Journal of Physical Anthropology, 59:
443-459.

Directorate of Overseas Surveys (U.K.). 1978. Montserrat (5th ed.). D.O.S. 359 (Series E803),
scale 1:25,000.

Dobbin, J. D. 1 983. “Do’en dee dance”: description and analysis of the jombee dance of Montserrat.
Unpublished Ph.D. dissertation, Ohio State University, 231 pp.

Goodwin, C. M. 1 987. Sugar, time, and Englishmen: a study of management strategies on Caribbean
plantations. Unpublished Ph.D. dissertation, Boston University, 199 pp.

Griffenhagen, G. B., and J. H. Young. 1959. Old English patent medicines in America. United
States National Museum, Bulletin, 2 18(10): 15 5-1 82.

Handler, J. S. 1972. An archaeological investigation of the domestic life of plantation slaves in
Barbados. The Journal of the Barbados Museum and Historical Society, 34(2):64-72.

. 1981. A Ghanaian pipe from a slave cemetery in Barbados, West Indies. West African

Journal of Archaeology, 1 1:93-99.

Handler, J. S., A. C. Aufderheide, R. S. Corruccini, E. M. Brandon, and L. E. Wittmers, Jr.
1986. Lead contact and poisoning in Barbados slaves: historical, chemical, and biological evi-
dence. Social Science History, 10(4):399-425.

Handler, J. S., and R. S. Corruccini. 1983. Plantation slave life in Barbados: a physical anthro-
pological analysis. Journal of Interdisciplinary History, 14(1):65— 90.

— 1 986. Weaning among West Indian slaves: historical and bioanthropological evidence from

Barbados. The William and Mary Quarterly, 3rd series, 43:1 1 1-117.

318

Annals of Carnegie Museum

vol. 56

Handler, J. S., R. S. Corruccini, and R. J. Mutaw. 1982. Tooth mutilation in the Caribbean:
evidence from a slave burial population in Barbados. Journal of Human Evolution, 1 1:297-313.

Handler, J. S., and F. W. Lange. 1978. Plantation slavery in Barbados: an archaeological and
historical investigation. Harvard University Press, Cambridge, 368 pp.

. 1979. Plantation slavery on Barbados, West Indies. Archaeology, 32(4):45-52.

Hardy, F. 1 922. Studies in West Indian soils. II. The soils on Montserrat, their natural history and
chief physical properties, and the relationship of these to the problem of die-back of lime trees.
West Indian Bulletin, 19:189-213.

Hardy, F., G. Rodrigues, and W. R. E. Nanton. 1 949. The agricultural soils of Montserrat. Studies
in West Indian Soils 1 1. Imperial College of Tropical Agriculture, Trinidad, 68 pp.

Jones, Olive R., and E. Ann Smith. 1985. Glass of the British Military, ca. 1755-1820. National
Historic Parks and Sites Branch, Parks Canada, Ottawa, 1 34 pp.

Lang, D. M. 1 967. Soil and land-use survey no. 22: Montserrat. Regional Research Centre, University
of the West Indies, St. Augustine, Trinidad, 31 pp.

Lange, F. W., and J. S. Handler. 1985. The ethnohistorical approach to slavery. Pp. 15-32, in
The archaeology of slavery and plantation life (T. A. Singleton, ed.), Academic Press, Orlando,
338 pp.

MacGregor, A. G. 1939. The Royal Society expedition to Montserrat, B.W.I.: the volcanic history
and petrology of Montserrat, with observations on Mt. Pele in Martinique. Philosophical Trans-
actions of the Royal Society of London, series B, 229(B557):l-90.

Mann, R. W., L. Meadows, W. M. Bass, and D. R. Watters. 1987. Description of skeletal remains
from a black slave cemetery from Montserrat, West Indies. Annals of Carnegie Museum, 56(19):
319-336 .

Martin-Kaye, P. H. A. 1959. Reports on the geology of the Leeward and British Virgin Islands.
Voice Publishing, Castries, St. Lucia, 117 pp.

Noel Hume, I. 1969a. Glass in Colonial Williamsburg’s Archaeological Collections. Colonial Wil-
liamsburg Archaeological Series 1, 48 pp.

. 1 969 b. Archaeology and Wetherburn’s Tavern. Colonial Williamsburg Archaeological Series

3, 47 pp.

. 1969c. A Guide to artifacts of colonial America. Alfred A. Knopf, New York, 323 pp.

Pulsipher, L. M. 1977. The cultural landscape of Montserrat, West Indies, in the seventeenth century:
early environmental consequences of British colonial development. Unpublished Ph.D. disser-
tation, Southern Illinois University at Carbondale, 183 pp.

Pulsipher, L. M., and C. M. Goodwin. 1982. Galways: A Caribbean sugar plantation— a report of
the 1981 field season. University of Tennessee, Knoxville, 1 14 pp.

South, S. 1977. Method and theory in historical archaeology. Academic Press, New York, 345 pp.

Steadman, D. W., D. R. Watters, E. J. Reitz, and G. K. Pregill. 1984. Vertebrates from ar-
chaeological sites on Montserrat, West Indies. Annals of Carnegie Museum, 53(1): 1-29.

Ubelaker, D. H. 1978. Human skeletal remains: excavation, analysis, and interpretation. Taraxa-
cum, Washington, D.C., 1 16 pp.

Ubelaker, D. H., and J. L. Angel. 1976. Analysis of the Hull Bay skeletons, St. Thomas. Journal
of the Virgin Islands Archaeological Society, 3:7-9.

Watters, D. R. 1980. Transect surveying and prehistoric site locations on Barbuda and Montserrat,
Leeward Islands, West Indies. Unpublished Ph.D. dissertation, University of Pittsburgh, 416 pp.

. 1981. A Turlington Balsam phial from Montserrat, West Indies: genuine or counterfeit?

Historical Archaeology, 15(1): 105-108.

ANNALS OF CARNEGIE MUSEUM

Vol. 56, Article 19, Pp. 319-336

20 November 1987

DESCRIPTION OF SKELETAL REMAINS FROM A
BLACK SLAVE CEMETERY FROM
MONTSERRAT, WEST INDIES

Robert W. Mann1

Lee Meadows1

William M. Bass1

David R. Watters

Assistant Curator, Section of Anthropology

Abstract

The skeletons of 17 Black slaves were excavated from an unmarked 18 th century cemetery in
Montserrat, West Indies. A pharmaceutical phial found with one of the probable coffin burials bears
the inscription OCT 29 1751. The date on the phial, skeletal analysis and ethnohistorical records
suggest the skeletons are those of Black slaves, possibly from the nearby Bransby Plantation. Evidence
of pathological conditions include enthesopathies, osteoarthritis, anemia, malnutrition, a high inci-
dence of fractures and one possible case of lepromatous leprosy. Ages at death suggest that the adult
females were outliving the males.

Introduction

In 1979 a construction crew digging a house foundation and adjacent swimming
pool on the western coast of Montserrat, West Indies, unearthed fragmentary
bones in association with a historic bottle bearing a date of OCT 29 1751 (Watters,
1981). Construction halted and David R. Watters, a visiting archaeologist, was
summoned to examine the skeletons. Watters identified the bones as human and
belonging to more than one individual. A preliminary examination of the site
revealed additional clusters of human bones. Although construction resumed, by
the end of one week a total of 10 individual burials (1-10) and one multiple burial
(probably commingled individual burials removed by the construction crew) had
been excavated in an attempt to salvage as many skeletons as possible (Watters,
1987). Final analysis reveals that a minimum of 17 individuals (Table 1) were
excavated.

The Harney site, so named after the building contractor, represents an unmarked
and unknown (to the local inhabitants) Black slave cemetery in use during the
latter 18th century (inclusive dates are unknown). Due to the paucity of recovered
skeletons in the West Indies, little is known about the diet, disease, stature,
nonmetric traits and dentition of Black slaves. The Harney site, therefore, offers
researchers an opportunity to examine the skeletons of 18th century slaves of
African ancestry (Williams, 1970).

1 Department of Anthropology, University of Tennessee, Knoxville, Tennessee 37996-0720.
Submitted 5 May 1987.

319

320

Annals of Carnegie Museum

vol. 56

Table l.— Minimum number of individuals.

Burial

Sex

Age

Race

1

F

50+

Indeterminate

2

F

20-30

Indeterminate

3

M

35-45

Indeterminate

4

F

60+

Black

5

F

50+

Indeterminate

6

F

18-35

Indeterminate

7

M

25-35

Black

8

F

40 +

Indeterminate

9

M

40 +

Indeterminate

10

M

35-45

Indeterminate

11

M?

18-40

Indeterminate

12

M

18-40

Indeterminate

13

M

18-40

Indeterminate

14

?

7-8

Indeterminate

15

?

8-10

Indeterminate

16

?

10-12

Indeterminate

17

?

14-16

Indeterminate

Methods and Materials

A minimum of 17 individuals were recovered from the Harney site. The 10
individual burials each represent one adult (6 females and 4 males) while the
multiple burial (single burials commingled by the construction crew) contains
portions of at least 7 individuals (4 subadults and 3 adult males). The minimum
number of individuals present in the multiple burial is conservatively based on
a count of 4 left subadult femora and 3 left adult femora. There is, however,
evidence of one additional subadult (maxilla and teeth indicate 13-15 years) as
well as a very small female (60+ years) as evidenced by an innominate and two
femoral heads (Table 2). The femoral heads measure 36 mm in diameter and are
much too small to correspond to the 3 left femora identified as males. These two
individuals were not included in the minimum number of individuals (MNI)
reported. The multiple burial, then, contains the commingled skeletons of indi-
viduals of both sexes, ranging in ages from 7-60 years. No fetal or infant skeletons
were recovered from the site.

The skeletons were received at the University of Tennessee in three boxes.
Individual burials were packed separately from those of the multiple burial. Each
individual burial was placed on a metal tray and tagged with the appropriate
burial number assigned in the field. This facilitated examination of the skeletons
and eliminated the possibility of mixing bones of two burials. During the labo-
ratory analysis every effort was made to match elements in the multiple burial
with the individual burials (numbers 1-10). In no instance were the authors able
to crossmatch a bone in this manner. Burials 1-10 represent primary single in-
terments with no evidence of commingling with the multiple burial (although
Burial 9 was disturbed upon interment of Burial 1 [Watters, 1987]).

Bones in the disturbed multiple burial were sorted on the basis of intragroup
size sedation, sex and age. The 4 subadult left femora (burials 14-17) were aged
by comparison of size and length to an age-graded reference series. The adults
were assigned burial numbers (11-13) following a detailed sorting of all elements
that considered sex (size and robusticity), age (osteoporosis and degree of degen-

1987

Mann et al. -- Black Slave Cemetery from Montserrat

321

Table 2.— Bones of the multiple burial indicating age and sex.

Bone

Sex

Age

Race

1

F

60+

Indeterminate

2-3

F

60+

Indeterminate

4-5

M

Adult

Indeterminate

6

M

Adult

Indeterminate

7

?

20-30

Black?

8

F?

Adult

Indeterminate

9

F?

Adult

Indeterminate

10

F?

Adult

Indeterminate

11

F?

60+

Indeterminate

12

?

Adult

Indeterminate

19

M

30-40

Black

erative joint disease), morphological similarity and metrical analysis. By a process
of elimination most of the bones were matched to specific individuals to the
exclusion of others. Some elements, however, remain unmatched. For example,
skulls in the multiple burial could not be reasonably matched with any of the
postcrania, even if the bones were of similar color (soil discoloration), sex and
age. Corresponding elements of the crania and postcrania were reconstructed to
obtain measurements. In a few instances we were able to match femora with tibiae
based on intragroup seriation and metrical analysis (this did not affect the MNI).
No attempt was made to match unlike elements (e.g. arm with leg bones) unless
morphological traits strongly suggested a match. Each bone was then inventoried
on separate sheets denoting metrics, age, race (skulls), sex, completeness and
pathological conditions. Although this is a time consuming method of analysis,
it offers an appraisal of the general characteristics of each bone and individual in
the sample (for a more detailed account of this method refer to Owsley et al.,
1985a). The criteria used in estimating age, race, sex and stature consist of con-
ventional methods used by practicing anthropologists. In instances of conflicting
aging criteria the most reliable method was emphasized in the final assessment.
For example, an individual with open cranial sutures and an eroded auricular
surface (ilium) was aged by the latter. When dental attrition conflicted with de-
generative joint disease the former was chosen. Most individuals, however, did
not present such ambiguities.

Primary race estimates are based on those criteria exhibited in the skull (Bass,
1971). The shape, width and inferior border of the nasal aperture were given the
most weight. Alveolar prognathism, dental crenulations (wrinkles), vertical sub-
nasal corrugations (De Villiers, 1968) and general cranial contour were also con-
sidered. Of secondary importance were the criteria of femur curvature and the
lack of shovel-shaped incisors.

Sex was determined by the shape and depth of the sciatic notch, maximum
diameter of the femoral head, size of the mastoid processes, chin shape, browridge
flatness, sharpness of the upper margins of the eye orbits, nuchal and supramastoid
crest development, cranial size and shape, long bone seriation (intragroup) and
comparison with known-sex skeletons.

Due to the fragmentary condition of the skeletons, stature could only be de-
termined for 4 individuals (3 females and 1 male). Trotter and Gleser’s (1952)
formula for American Blacks was used. In most instances the long bones of the
legs had to be reconstructed before measurements could be taken.

322

Annals of Carnegie Museum

vol. 56

Fig. 1 (above).— -Burial 2. Frontal view of skull showing copper salts stain.

Fig. 2 (below). — Burial 3. Inferior view of left calcaneus (arrow points to secondary os calcis “notch”).

Paleopathological analysis is based on a three-category system designed to clas-
sify the bone cell response (to disease) present in the bone. The categories consist
of bone loss (e.g. resorption, porosity), bone apposition (e.g. periostitis) or both.
Each category has specific codes that further narrow and define the bone lesion
according to severity, location (joint or nonjoint) and, in some instances, specific
disease (e.g., lepromatous leprosy).

1987

Mann et al.— Black Slave Cemetery from Montserrat

323

Burial Analysis
In situ Burials

Burial I. —Female, race indeterminate, 50 + , 5' PA". This skeleton is in a fair
state of preservation with the following elements present: both femora, tibiae,
fibulae, right foot, fragmentary left radius, ulna, hand bones and innominates.
Most of the arm bones and long bone ends are missing (post mortem). This female
exhibits moderately robust long bones but a gracile radius and ulna. Female traits
consist of a wide, shallow sciatic notch and a delicate ischium. Marked osteo-
porosis combined with porosis of the acetabula suggest old age.

Pathological Condition: The left hand exhibits a swollen and disfigured meta-
carpal, eburnated (polished) carpal, arthritic destruction of one distal phalanx and
an osteochondritic pit in the greater multangular. There is an area of ossified
connective tissue (enthesopathy) on the ventral surface of the left patella and
posterior surfaces of the tibiae resulting from repeated physical stress. The right
fibula exhibits a healed fracture near its midshaft. It is probable that this female
favored her right leg following an injury, as noted by the slightly atrophied right
femoral shaft. The midshaft circumferences of the femora are 8.5 cm (left) and
8.0 cm (right).

Burial 2. —Female, race indeterminate, 20-30. This fragmentary skeleton is in
a poor state of preservation with the following elements present: most of the skull,
mandible and teeth of both jaws, and a fragmentary left humerus, left femur and
dens of the second cervical vertebra. Sex is based on the sharp orbital margins,
a pointed chin, flat browridges, small mastoids, cranial shape and gracile long
bones (Krogman, 1962; Bass, 1971; Stewart, 1979). Aging criteria consist of slight
dental attrition, open cranial sutures, thick cortical bone and lack of degenerative
joint disease. The overall size of the skeleton suggests a small female. There is a
green copper salts stain on the right frontal bone suggesting a shroud pin or other
ornament was buried with her (Fig. 1).

Pathological Condition: Slight erosion and porosity of the right mandibular
fossa. Bilateral cribra orbitalia and slight porotic hyperostosis (Ortner and Put-
schar, 1981). The teeth exhibit slight attrition and no abscesses; 3 teeth are carious
and present hypercementosis. Eleven teeth have enamel hypoplastic lines encir-
cling the crowns.

Burial 3. — Male, race indeterminate, 35-45, 5'4". Both legs and feet are present
(the upper skeleton was destroyed by bulldozers). Muscle markings, a femoral
head diameter of 46 mm and large long bones suggest this individual to be one
of the larger males recovered. The contour of the femora is relatively flat. An age
of 35-45 is based on moderately thick cortical bone and only slight arthritic
lipping of the proximal articular surface of the left tibia.

Pathological Condition: This individual presents a number of pathological le-
sions. First are the osteophytes on the articular surface of the distal right femur.
Both tibiae exhibit superiorly directed bony projections (enthesopathies) on the
tibial tuberosities. Two proximal foot phalanges exhibit severe osteoarthritis of
the distal joint surfaces.

Anomaly: Both calcanei (although the right calcaneus is damaged a trace of the
trait is still visible) exhibit evidence of what Dwight (1907) terms a secondary os
calcis (Fig. 2, 3). This true accessory bone lies within the joint spaces between the
calcaneus, talus, cuboid and navicular. The secondary os calcis is a small, roughly
semilunar-shaped bone that is slightly convex along its medial border and convex

324

Annals of Carnegie Museum

vol. 56

Fig. 3 (above). —Burial 3. Superior view of left calcaneus (close up of secondary os calcis “notch”).
Fig. 4 (below). — Burial 4. Frontal view of skull.

at its lateral margin where it articulates with a '‘notch” in the anterior talocalcaneal
facet. The presence of such a bone would not cause pain or discomfort. Although
the frequency of this accessory bone (in skeletal and extant populations) is not
known, preliminary research by one of the authors (RWM) shows it to occur in
about 1 of every 8 individuals examined (general population).

1987

Mann et al. — Black Slave Cemetery from Montserrat

325

Fig. 5 (above). — Burial 4. Left lateral view of skull.
Fig. 6 (below). — Burial 4. Superior view of skull.

326

Annals of Carnegie Museum

vol. 56

Table 3. —Burial 4 cranial measurements (in millimeters).

Glabello-occipital L

195

Nasion-occipital L

196

Basion-nasion L

115

Basion-bregma Ht

146

Maximum cranial Br

135

Minimum frontal Br

94

Biauricular Br

113

Minimum cranial Br

80*

Biasterionic Br

99

Mastoid Ht

23

Mastoid Br

10

Orbital Ht

32

Orbital Br

32

Bifrontal Br

90

Biorbital Br

91

Cheek Ht (minimum)

20

Supraorbital projection

5

Glabella projection

1

Frontal chord

122

Frontal subtense

29

Parietal chord

118

Parietal subtense

24

Occipital chord

98

Occipital subtense

23

Foramen magnum L

34

Foramen magnum Br

29*

Nasion radius

108

Bregma radius

135

Vertex radius

136

Lambda radius

119

Opisthion radius

41

Basion radius

12

* Approximate measurement.

Burial 4.— Female, Black, 60 + , S'1/*". This female is the most complete indi-
vidual recovered. The only elements missing are most of the ribs, sternum, portion
of the maxilla and right pubis. Sex is based on the size and shape of the skull
(Fig. 4-6; Table 3), flat browridges, small mastoids, pointed chin and a wide,
shallow sciatic notch; note the small, rounded eye orbits in Fig. 4. Racial traits
consist of nasal guttering and alveolar prognathism. Although the cranial sutures
are open, an age of 60+ is based on the edentulous maxilla and mandible, marked
osteoporosis, eroded symphyseal face, degree of osteoarthritis and auricular sur-
face changes. There is a green copper salts stain on the left frontal bone.

Pathological Condition: Bilateral cribra orbitalia (Fig. 7). The endocranial sur-
face of the frontal bone exhibits 3 or 4 small protuberances, often referred to as
enostoses or Hyperostosis Calvariae Interna (Perou, 1964), probably of a benign
nature. The innominates exhibit porosity in the acetabular surfaces and a small
lytic lesion superior of the left acetabular rim. The left thumb (metacarpal) exhibits
a healed fracture. The right patella and joint surface of the distal right femur are
ebumated. The corresponding articular surfaces display deep grooves and erosion
resulting from bone on bone contact (osteoarthritis). Based on the lack of long
bone atrophy it is probable that this individual remained active until or near the
time of her death. The most striking pathological condition seen in this skeleton
is the marked degree of osteoarthritic changes in the femoral heads (Fig. 8). The

1987

Mann et al. — Black Slave Cemetery from Montserrat

327

Fig. 7 (above).— Burial 4. Bilateral cribra orbitalia.

Fig. 8 (below).— Burial 4. Osteoarthritic femoral heads (posterior view).

surfaces exhibit porosity and new bone growth as well as periarticular mush-
rooming. The femoral necks appear normal in length but display small areas of
bone apposition, possibly in reaction to hyperextension of the legs (Angel, 1964)
or contact with the rim of the acetabula (Fig. 9, 10).

Burial 5. —Female, race indeterminate, 50 + , 5'1". The skeleton is represented
by both lower arms and hands, distal left humerus, 4 lumbar vertebrae, left
innominate, both femora, tibiae, fibulae and feet. Sex is based on a wide, shallow
sciatic notch, small long bones and femoral head diameter of 40 mm. Age is based
on the degree of osteoporosis and auricular surface changes (Lovejoy et al., 1985).

Burial 6. —Female, race indeterminate, 18-35. This individual is represented
only by the tibiae and a crushed right femur and fibula. Sex is based on the small

328

Annals of Carnegie Museum

vol. 56

Fig. 9 (above). — Burial 4. Posterior view of right femur (note the small osteophyte extending from the
neck).

Fig. 10 (below). — Burial 4. Close up of posterior right femur (porosity artifactual).

size of the tibiae and gracile muscle markings. Age is based on the adult size of
the tibiae and thick cortical bone.

Pathological Condition: The distal right tibia exhibits a small area of nearly
healed periostitis, possibly the result of acute trauma (fracture?).

Burial 7. —Male, Black, 25-35. The only elements present consist of a nearly
complete skull (Fig. 11, 12) and mandible. Male traits consist of a well developed
nuchal crest, intermediate mastoids, square chin and a large gonial angle with a
wide ascending ramus. Because of the fragmentary nature of the facial region race
is based on the general contour of the skull (R. L. Jantz, personal communication,
1986). The open cranial sutures, moderate dental attrition, obliterated incisive
suture (Mann et al., 1987) and thick diploe suggest an age of 25-35 years.

1987

Mann et al. — Black Slave Cemetery from Montserrat

329

Fig. 1 1 (above). — Burial 7. Right lateral view of skull (portion of frontal bone is missing).
Fig. 12 (below). — Burial 7. Superior view of skull.

330

Annals of Carnegie Museum

vol. 56

Fig. 13 (above). “-Burial 9. Robust left femur in comparison to a known large adult male.

Fig. 14 (below).— -Bones 14 and 15 (hand phalanges) of the multiple burial exhibiting enlarged nutrient
foramina.

1987

Mann et al.— Black Slave Cemetery from Montserrat

331

Pathological Condition: The right mandibular condyle displays slight erosion.
Bilateral cribra orbitalia and slight porotic hyperostosis. Active endocranial po-
rosity is visible in the frontal and parietals bordering the superior sagittal sinus.
Enamel hypoplasias. There is a pipe-stem groove in the left mandibular canine.

Anomalies: The left maxilla exhibits a small supernumerary tooth (crown miss-
ing) on the buccal surface, interproximal of M2 and M3. Another supernumerary
tooth is present in the socket of the right mandibular incisor (two-rooted). A third
small supernumerary tooth is present but the alveolus is missing. The mandibular
central incisors present bulbous lingual cingula.

Burial 8. —Female, race indeterminate, 40 +. This individual is in a very poor
state of preservation and represented only by the shafts of the femora, a few long
bone fragments and the left mastoid process. The extremely small mastoid and
gracile femur strongly suggest female. Age is based on thin cortical bone and a
rugged cortex.

Pathological Condition: The posterior surface of the right tibia displays a small
area of porous, striated bone indicating active periostitis.

Burial 9.— Male, race indeterminate, 40 + . This large male is represented by
portions of the left femur, tibia and both ulnae. The left femur is extremely robust
and measures 34 mm (anteroposterior) and 27 mm (mediolateral) at the midshaft
(Fig. 13). The development of the aspera contributes heavily to the robust ap-
pearance. The right ulna is also robust. Age is based on the presence of marginal
lipping of the olecranon fossa and thin cortical bone.

Pathological Condition: The left femur is the most robust bone in the sample
and measures within the stenomeric range (103) on the platymeric index (Bass,
1971).

Burial 10. — Male, race indeterminate, 35-45. The only bones presented are the
femora, tibiae and one left metatarsal. No articular surfaces survive. The generally
smooth, thick cortex of the leg bones suggest middle age. Muscle markings are
moderate indicating this to be a male of medium build (intragroup comparison).

Multiple Burial (Disturbed by Construction)

Burial 11. —Probable male, race indeterminate, 18-40. This individual is rep-
resented only by a portion of the left femur. This bone is intermediate in size and
robusticity but fits better within the male range. The thick cortical bone suggests
young to middle age.

Burial 72. —Male, race indeterminate, 18-40. Present only is the left femur
missing both articular ends. This bone is moderate to robust and has a well
developed linea aspera and thick cortex suggesting young to middle age.

Burial 13. —Male, race indeterminate, 18-40. Present only is the left femur
missing both articular ends. This bone is also robust and exhibits a thick, smooth
cortex suggesting a young to middle age adult.

Burial 74. — Subadult, 7-8. The only element present is a left femur, missing
both articular ends. Based on intragroup seriation and comparison to a known-
age femur this individual is 7-8 years of age.

Burial 75. — Subadult, 8-10. Elements present are both femoral shafts and tibiae.
Intragroup seriation and known-age comparison suggest an age of 8-10. The right
tibia exhibits active porosity over much of its surface.

Burial 76. —Subadult, 10-12. Bones present consist of fragmentary femora. The
right femur exhibits a portion of the billowy epiphyseal surface for attachment of
the head. Intragroup seriation places this child at 10-12 years.

Burial 17. — Subadult, 14-16. Present are fragments of both femora and tibiae.

332

Annals of Carnegie Museum

vol. 56

This is the largest subadult in the sample. The right femur and tibia exhibit active
periostitis.

Bones of the Multiple Burial Indicating Age and Sex

Bone 1. —Female, race indeterminate, 60 + . This fragment consists of the right
innominate with a wide sciatic notch and small acetabulum. The auricular surface
is eroded and macroporotic suggesting an age of 60 + .

Bones 2 and 3. —Female, race indeterminate, 60 + . Both femoral heads (left
measures 36 mm and the right 37 mm) are present and match. The small size of
the heads and obliteration of the epiphyseal lines indicate this to be a small, adult
female. The femoral heads fit the small right acetabulum listed above (Bone 1)
and belong to the same individual.

Bones 4 and 5. —Left and right femoral heads (each 45 mm in diameter); size
suggests this to be a good match and probably of the same adult male.

Bone 6. — Right femoral head (45 mm), adult male.

Bone 7. —Right maxilla and 5 teeth, adult. The shape of the nasal aperture
suggests a probable Black (admixture?). Dental development and attrition suggest

an age of 20-30 years.

Bone 8. —Portions of a skull (one individual). Active porotic hyperostosis of
the parietals and occipital. This is probably a young adult female as evidenced
by small mastoids processed and undeveloped supramastoid crests. Cranial su-
tures are open.

Bone 9. — Mandible fragment with a pointed chin, probable adult female.

Bone 10. —Mandible fragment with a pointed chin, probable adult female.

Bone 77.— Edentulous mandible, female 60+ (probably goes with the innom-
inate and 2 femoral heads noted above).

Bone 72. —Right clavicle (adult) with a green copper salts stain.

Bone 13. —Numerous cranial fragments of one individual (sex and age inde-
terminate). Green copper salts stain on the left frontal bone.

Bones of the Multiple Burial Exhibiting Pathological Conditions

Bones 14 and 15. — Two middle hand phalanges with enlarged nutrient foramina
(Fig. 14) suggesting lepromatous leprosy (Brothwell and Sandison, 1967; Stein-
bock, 1976; Donald Ortner, personal communication, 1986).

Bone 76. —One metatarsal with an oval-shaped lytic lesion in the diaphysis.

Bone 77. —One left metacarpal with a healed fracture of the diaphysis.

Bone 18. —Left ulna fragment with healed periostitis.

Bone 79.— Male, Black, 30-40 years. This maxilla (palate and 11 teeth only)
is very large and exhibits classic Black features (Fig. 15, 16): a wide nasal aperture
(27 mm), alveolar prognathism, vertical subnasal corrugations, nasal guttering,
tapered central incisors, crenulated molars and a diastema between the central
incisors. Age is based on slight dental attrition, erupted third molars and visible
incisive suture. Carabelli’s cusp is present.

Pathological Condition: Linear enamel hypoplasias and slight to moderate hy-
percementosis. Chronic osteitis of the right maxillary antrum (Fig. 1 7) due to a
periapical abscess of the second premolar. The left first molar has a large occlusal
caries. The mandibular premolars and molars exhibit large buccal wear facets.
There were a few loose mandibular teeth recovered in the multiple burial that
show corresponding wear facets on their lingual surfaces. This pattern of dental
wear suggests that the mandible may have been much smaller than the maxilla.

1987

Mann et al. — Black Slave Cemetery from Montserrat

333

Fig. 15 (above). —Bone 19 of the multiple burial (maxilla of adult Black male; missing bone is arti-
factual).

Fig. 16 (below).— Bone 19 of the multiple burial (lingual surface of maxilla).

Another possibility for the uneven wear is that pain associated with the periapical
abscess may have caused the individual to favor chewing on the left side.

Dental Summary

A total of 92 teeth (both loose and in sockets) were recovered from the site and
include the following: 1 8 incisors, 1 3 canines, 27 premolars and 34 molars. Many

334

Annals of Carnegie Museum

vol. 56

Fig. 17. — Bone 19 of the multiple burial (chronic osteitis of right antrum; note the thickened and

porous appearance of the new bone).

of the teeth exhibit caries (1 7 of 92), root hypercementosis and enamel hypoplasias
similar to those described by Corruccini et al. (1982, 1985), of a Black slave
cemetery in Barbados. The frequency of dental caries by tooth type includes the
following: 1 incisor, 1 canine, 8 premolars and 7 molars. Also present are 3
periapical and 1 periodontal abscesses. Two individuals exhibit pipe-stem grooves.
In this small sample there is no evidence of intentional tooth mutilation or shovel-
shaped incisors.

In general the teeth reflect a moderately high fiber diet resulting in a slight to
moderate dental attrition. The high frequency of dental caries (18%) may reflect
a diet high in natural sugar, possibly from chewing sugar cane (Owsley et al.,
1985Z?). Dental attrition and wear patterns of this sample do not reflect a rough,
gritty diet as seen in prehistoric Amerindian groups. The dental wear patterns of
this group represent a diet between the abrasive diet of the Amerindian and the
refined diet of the more industrialized peoples.

Discussion and Summary

Final analysis suggests that the skeletons examined are those of Blacks, possibly
slaves from the nearby Bransby sugar plantation. The lack of fetal and infant

1987

Mann et al. = Black Slave Cemetery from Montserrat

335

skeletons supports the findings of other researchers that Black slaves were selec-
tively burying their dead in particular sections of the cemeteries, if not elsewhere
(Corruccini et al., 1982). Construction deadlines at the Harney site precluded a
thorough excavation of the entire cemetery which would have allowed for a more
accurate rendering of slave mortuary practices in Montserrat. It is likely, however,
that the burial practices at the Harney site were similar to those deduced by
Corruccini et al. (1982).

The males in the Harney site sample are moderate to robust in build yet short
in stature and differ from the historical description of the “best slave” as being
5T1" (Williams, 1970). The stature of the three measurable females clustered
around 5T". A general comparison of Harney femora and tibiae with known-sex
and -race individuals suggests generally taller but less robust than the Amerindian
sample (Arikara of the Northern Plains), and shorter but more robust than the
White sample (forensic cases). Development of the linea aspera (femur) and a
greater anteroposterior diameter are distinguishing features of the Harney skele-
tons.

Due to the fragmentary nature of the skeletons only 3 individuals could be
reliably raced (Black; a fourth individual exhibits Negroid traits but not as strongly
as the others). The general contour of most femora (anteroposteriorly) is flattened,
further suggesting Black individuals.

The age distribution by sex presents an interesting pattern. Five of the 6 iden-
tifiable males were less than 45 years old at death and the sixth 40+ years. The
females, on the other hand, outlived the males in this sample; of the 6 identifiable
females 4 are 40+ and only 2 died in their thirties.

Most of the Black slaves brought to the West Indies were either bought or
captured on the coast of West Africa. Fully 16% of the slaves died aboard ships
during “middle passage” and another 33% within three years of their arrival in
the West Indies (Williams, 1970). Since there was little value in bringing older
slaves to work on the plantations, the presence of old females suggests that they
had survived as slaves for many years (not necessarily on Montserrat). Further
research to detect the presence and amount of lead present in the bone may help
clarify this question (Handler et al., 1986). The subadults, on the other hand, may
have either been brought to Montserrat from Africa or were born on the island.
The 5 subadults identified in this sample comprise nearly 30% of the individuals
recovered from the Harvey site.

The overall disease pattern reflects a high frequency of anemia as evidenced by
cribra orbitalia and porotic hyperostosis. Although present in a few individuals,
periostitis doesn’t appear to have been a major problem among these people. The
most interesting disease noted in this sample is the presence of enlarged nutrient
foramina (2 hand phalanges) suggestive of lepromatous leprosy. The lack of other
diagnostic evidence precludes a firm diagnosis of the disease.

Another point of interest is the relatively high frequency of fractures in this
group: 3 females (burials 1 , 4, and 6) as well as an unprovenienced metacarpal in
the multiple burial. It is interesting to note that fractures are identified only with
the females.

The incidence of degenerative joint disease is moderate and seems to fit the
expected scenario that joint involvement increases with age, probably commenc-
ing sometime in the late twenties. In only one individual is there evidence of
severe osteoarthritis, this being in the right knee and hip. The sacra, innominates
and few recovered vertebrae are relatively free of osteophytosis and osteoarthritis.
Muscle attachment sites are also nearly free of enthesopathies which suggests that

336

Annals of Carnegie Museum

vol. 56

these individuals were not engaged in prolonged activities that produce ossified
connective tissue in the extremities.

Examination and measurements of the hand bones suggest these individuals
had long, slender fingers in comparison to a sample of Amerindians and Whites.
The hands appear to have been subjected to arduous grasping activities, as evi-
denced by the developed muscle attachment areas in the phalanges and meta-
carpals of the 5 individuals with recovered hands or feet; 4 individuals exhibit
one or more of the following pathological conditions: osteoarthritis, healed frac-
tures and volar curvature. Speculations concerning the bony changes present in
these individuals are of little substantive value due to the small sample size.

In conclusion, the skeletons of at least 17 Black slaves consisting of 6 adult
females, 6 adult males, 1 probable male and 4 subadults were recovered from the
Harney site. Although the ratio of males to females is even, the males appear to
have died at earlier ages, possibly reflecting a harsher and more disease prone
lifestyle. Although excavation of the entire cemetery was not possible, the lack of
fetal and infant skeletons suggests a selective burial pattern. The frequent occur-
rence of fractures, cribra orbitalia, porotic hyperostosis and enamel hypoplasias
suggests that these individuals suffered the effects of periodic severe malnutrition,
common illnesses and a harsh lifestyle that resulted in the early deaths of the
plantation males.

Acknowledgments

The authors gratefully acknowledge the assistance and expertise of the following people: Dr. Richard
L. Jantz for his suggestions regarding race and sex, and his editorial comments; Dr. P. Willey’s assistance
in the osteology laboratory; Dr. Donald Ortner for his helpful discussion of some of the pathological
specimens; Dr. Ann Bass for editorial comments; Kathy Goldsmith for her analysis of the hands and
feet; Peer H. Moore-Jansen and Henry Case for their assistance with the dental analysis and Kim
Johnson for typing the manuscript. Acknowledgments for the field research appear in Watters (1987).

Literature Cited

Angel, J. L. 1964. The reaction area of the femoral neck. Clinical Orthopedics, 32:130-142.

Bass, W. M. 1971. Human osteology: a laboratory and field manual of the human skeleton. The
Missouri Archaeology Society, Inc., Columbia, 281 p.

Broth well, D., and A. T. Sandison. 1967. Diseases in antiquity. Charles C Thomas, Springfield,
Illinois, 766 pp.

Corruccini, R. S., J. S. Handler, R. J. Mutaw, and F. W. Lange. 1982. Osteology of a slave
burial population from Barbados, West Indies. American Journal of Physical Anthropology, 59:
443-459.

Corruccini, R. S., J. S. Handler, and K. P. Jacobi. 1985. Chronological distribution of enamel
hypoplasias and weaning in a Caribbean slave population. Human Biology, 57:699-71 1.

De Villiers, H. 1968. The skull of the South African Negro. Witwatersrand University Press,
Johannesburg, 342 pp.

Dwight, T. 1907. Atlas of variation of the hand and foot. J. B. Lippincott, Co., Philadelphia, 79 pp.
Handler, J. S., A. C. Aufderheide, R. S. Corruccini, E. M. Brandon, and L. E. Wittmers, Jr.
1986. Lead contact and poisoning in Barbados slaves: historical, chemical, and biological evi-
dence. Social Science History, 10(4):399-425.

Lovejoy, C. O., R. S. Meindl, T. R. Pryzbeck, and R. P. Mensforth. 1985. Chronological meta-
morphosis of the auricular surface of the ilium: a new method for the determination of adult
skeletal age at death. American Journal of Physical Anthropology, 68:15-28.

Krogman, W. M. 1962. The human skeleton in forensic medicine. Charles C Thomas, Springfield,
Illinois, 337 pp.

Mann, R. W., S. A. Symes, and W. M. Bass. 1987. Maxillary suture obliteration: aging the human
skeleton based on intact or fragmentary maxilla. Journal of Forensic Sciences, 32(1): 148-1 57.
Ortner, D. J., and W. G. J. Putschar. 1981. Identification of pathological conditions in human
skeletal remains. Smithsonian Institution Press, Washington, 479 pp.

1987

Mann et al. — Black Slave Cemetery from Montserrat

337

Owsley, D. W., C. E. Orser, Jr., R. Montgomery, and C. C. Holland. 1985<2. An archaeological
and physical anthropological study of the first cemetery in New Orleans, Louisiana. Report to
Louisiana Division of Archaeology, Office of Cultural Development, Department, Culture, Rec-
reation and Tourism, Baton Rouge, Louisiana, 242 pp.

Owsley, D. W., A. M. Mires, and G. W. Gill. 1985 b. Carious lesions in permanent dentitions of
Protohistoric Easter Islanders. Journal of the Polynesian Society, 94:415-422.

Perou, M. L. 1964. Cranial hyperostosis. Charles C Thomas, Springfield, Illinois, 157 pp.

Steinbock, R. T. 1976. Paleopathological diagnosis and interpretation: bone diseases in ancient
human populations. Charles C Thomas, Springfield, Illinois, 423 pp.

Stewart, T. D. 1979. Essentials of forensic anthropology. Charles C Thomas, Springfield, Illinois,
300 pp.

Trotter, M., and G. C. Gleser. 1952. Estimation of stature from long bones of American Whites
and Negroes. American Journal of Physical Anthropology, 10:463-514.

Watters, D. R. 1981. A Turlington Balsam phial from Montserrat, West Indies: genuine or coun-
terfeit? Historical Archaeology, 15:105-108.

. 1987. Excavations at the Harney site slave cemetery, Montserrat, West Indies. Annals of

Carnegie Museum, 56(1 8):289— 3 1 8.

Williams, E. 1970. From Columbus to Castro: the history of the Caribbean 1492-1969. Andre
Deutsch Limited, London, 576 pp.

INSTRUCTIONS FOR AUTHORS

ANNALS OF CARNEGIE MUSEUM consist of con-
tributions to the natural sciences and anthropology, in
30 by 46 picas format ( 1 27 by 1 95 mm or 5 by 75/8 inches).
All manuscripts must first be submitted for review to the
Curator in charge of the appropriate scientific Section.
Authors should give particular attention to scientific con-
tent, format, and general style for the ANNALS. Manu-
scripts that do not conform to the style of the ANNALS
will be returned to the author immediately. Every manu-
script will be peer reviewed by at least two outside per-
sons. Authors will be asked to subsidize, if possible, any
or all costs of publication (approximately $ 100/page
printed).

Manuscript style. — Articles should include the follow-
ing items in this order: title page, abstract, text (with
desired headings), acknowledgments, literature cited, ta-
bles, figure captions, and copies of illustrations. All
manuscripts must be typed double-spaced on standard
8'/2 by 1 1 inch white bond paper, with at least one inch
margins all around, and submitted in triplicate— an orig-
inal for the editors, and two review copies. All pages
should be numbered, including tables, literature cited,
and the list of figure captions. Only correspondence qual-
ity or better dot matrix printouts will be accepted; draft
mode is unacceptable.

Title Page. The title, on a separate title page should
be brief, include the animal or plant group involved, and
appear two to three inches below the top margin. Include
the authors name(s) and the affiliations of non-CMNH
authors. In the case of multiple authorship indicate the
address to which proofs should be sent.

Abstracts. Abstracts should be short, double-spaced
and substantive and included at the head of the first page
of text.

Text. Do not right justify text or break (hyphenate) a
word at the end of a line. Footnotes and acknowledgments
as footnotes are unacceptable. All text is double-spaced.

Tables and figure legends. The list of figure legends
and each table should be typed consecutively on indi-
vidual pages separate from the text. Tables must be dou-
ble-spaced throughout with no vertical lines.

Abbreviations. Refer to the CBE Style Manual for cor-
rect abbreviations. Mammalian dentition: use capital let-
ters (I, C, P, M, D, for incisor, canine, premolar, molar,
deciduous, respectively) and superscript/subscript num-
bers (M1 2, P3) to designate upper and lower teeth.

Measurements. Metric units should be used, if possi-
ble.

Literature Cited. All references in text must appear in
the Literature Cited section. The data (author, date and
page) in both citations must agree. Do not abbreviate the
titles of periodicals or serials. The following style, dou-
ble-spaced, should be used in Literature Cited:

1) Two authors in an institutional series:

Matthew, W. D., and W. Granger. 1923. The
fauna of the Houldjin Gravels. American Mu-
seum of Natural History Novitates, no. 97:
1-6.

2) Same authors repeated— use three-em dash:

. 1923. The fauna of the Ardyn Obo For-

mation. American Museum of Natural
History Novitates, no. 98:1-5.

3) Same authors plus a third author— repeat all authors:

Matthew, W. D., W. Granger, andG. G. Simpson.
1928. Paleocene multituberculates from Mon-
golia. American Museum of Natural History
Novitates, no. 331:1-4.

4) Chapter in an edited volume:

Rausch, R. L. 1963. A review of the distribution
of Holarctic mammals. Pp. 29-43, in Pacific
basin biography (J. L. Gressitt, ed.), Bishop Mu-
seum Press, Honolulu, Hawaii, xx + 450 pp.

5) Unpublished dissertation:

Smith, J. P. 1976. Review of Eocene mammals.
Unpublished Ph.D. dissert., University of Cal-
ifornia, Berkeley, 302 pp.

6) Book:

White, M. J. D. 1961. The chromosomes. Meth-
uen and Co., Ltd., London, 120 pp.

7) Journal articles with usual volume and issue number:

Anderson, W. I. 1969. Lower Mississippian con-
odonts from northern Iowa. Journal of Paleon-
tology, 43(4):9 16-928.

Illustrations. All illustrations will be called figures, and
are to be numbered in Arabic numerals. Three sets of
illustrations are required, one (original artwork) for re-
production, two for reviewers. Xerox copies of photo-
graphs for reviewers are usually not acceptable but are
adequate for line drawing review copies. All illustrations
must be reducible to a maximum of 127 by 195 mm (30
by 46 picas) without loss of clarity. Line copy should be
designed for reduction to % or of actual size. Type-
written illustration copy will not be accepted. Photo-
graphic figures should be submitted at actual reproduc-
tion size.

Two or more small figures should be combined for
reproduction as a unit or plate, but, if feasible, do not
combine halftones and line drawings. Rectangular half-
tone figures should be abutted, without intervening spaces.
The printer will insert narrow white spaces during the
reproduction process. Halftone figures with solid black
backgrounds will not be accepted unless the author agrees
to pay for the printing costs. All plates must have min-
imally one inch borders all around. Each plate should be
given a protective cover and identified on the back side.

Lettering and/or a magnification scale (linear metric
scale) for rectangular halftone figures should be placed
directly on the photo, not in a blank space between pho-
tos. The scale or lettering for closely cropped photos can
be placed in blank areas close to the figure.

Proof— The author should answer all queried proof
marks and check the entire proof copy. Return corrected
page proof with the edited manuscript promptly to the
editors.

If an author chooses to make extensive alterations to
a paper in proof stage the author or appropriate section
will bear the cost. Original manuscripts will not be re-
turned unless requested. Illustrations will be returned to
the author.

{

y&w&y — /

— W,

i ~ VN^» g,* ' Z

HSONIAN JNSTITUTI0NWN0UniliSN(_NVIN0SHiiHS</>S3ia’vaanfLIBRARIES</>SMITHS0NIA

cr

m .*

ioshxiws S3 id v^a n libraries Smithsonian institution noixoxixsni nvinoshxii

2 r- z r* 2 r-

m rn ~ xsoms*^ m

_ co ± co £ — co

[HSONIAN INSTITUTION NOIlfUllSNI NVINOSHXIWS S3! dVdai! LIBRARIES SMITHSONW

2 Ay- co Z _ co Z A *■> . to

< ^v'v- 2 4< x^I°N/X 2 , < ,, 2

- i iCj 1 1 i

^ > v'4(j0pr 2 2 ^-v >

<o k* z , z CO V z

joshiiws S3iavaan libraries Smithsonian institution NoiiniiisNi nvinoshxii

CO

— /s>\

cr

< l03

££L Vo\

co 5

_j z _j z -j z

rnsoNiAN institution Nouniiism nvinoshxms ssiavaan libraries smithsoni,

r~ Z » Z r- Z

O x<I?rm^Nv vv O ZuSTir77>v *“ O

m Ns^^osv^x x; x^vasv^x m ' \w 3^ rn

co £ co ' £ CO

sioshhims S3 lava an libraries Smithsonian institution NoiiniiisNi nvinoshxii

CO Z CO Z V CO z

< \v 2 xSstiroX <r a ifc. <

I
CO

±7 O

_ y

Ovos >’ 5 ^ ;>

rHSON IAN ^INSTITUTION NOIXnXIXSNl_ NVINOSHXIIAIS^SB I aVaan\lBRAR I ES°°SM1THS0NI)

CO _ V <0 — <0

losHxms S3 lava an libraries Smithsonian institution noixoxixsni nvinoshxii

2- r~ Z r~ Z —

m ^ m

FHSONIAN ~ INSTITUTION ^ N0IXnXIXSm~NVIN0$HXII/\l$ S3 I a va a IT~Li B RAR lES^SMJTHSONb

2 vv- CO Z CO Z

< v.. 2! « < 2!

Z ' CO z co v z

sjoshxiiais S3 lava an libraries Smithsonian institution noixoxixsni nvinoshxii

CO =; CO ^ CO

& - 3 /aaiWEX - a. 2 - /vzrm&\ z

V

2: / ’ £ x^&MWpj ± s -r s E

2 2 "W >■' 2 X&os*iX >* 2

io '*' 2 </> "'* 2 CO 2 CO

ION NOIXfUIXSlSIl NVINOSHXIWS S3IHVB8 IT LIBRARIES SMITHSONIAN INSTITUTION N

CO “2! \ CO 2 _ CO — ...... CO

Ul to V ^ m X5o^v UJ ^

811 LIBRARIES SMITHSONIAN INSTITUTION NOIlllillSNI NVINOSHXMS S3ldVBail t
r- _ 2 2 r* ... 2 r

3|)>

;o

w - vc£

/ ro p-9jX ^ ni ^ rn

riON^ NOIinilXSNI^NVINOSHlIIAIS S3 I H VH 8 n~L IBRARI ES^SMlTHSONlAN^INSTITUTlON i

CO 2 ^ 2 ^r. CO 2 *, CO

^ ~ q

o /?!?* %% x . '-M'/MA

CO [S§s£ CO X#'

X fewiiWv/ °

t Vc«P®' 2

CO

2 CO

* =§ ~ “

>

*\ X

^ CO

o
2

S >

_ _ CO V 2 _

an libraries Smithsonian institution noixhxixsni nvinoshxuais S3iavaan i

<0 5 ^ CO “ CO 5

!l " ^ (U) * w

CO
H

q

O ^.OC^ ^ ’^feP^ Q ^ q ^ O

ION NOIXnXlXSNi^NVINOSHXliAiS^SB I B VB 8 II ^Ll B R AR I ES^SMITHSONIAN^INSTITUTION f
2 r~ 2 r~ 2 £ 2

° ~ ^ O E X^OA^x O _j<sA-;. ™ O

m

0 1- ^ m O | r jn g

a n~ LI B R AR I ES SMITHSONIAN~INSTITUTIONWrtOliru.iiSNI~NVINOSHillAISW>S3 I 8Vd 8 11 “l

2 \ 02 2 CO 2 CO Z

< ^ 1 < a 2 ^ A <

>/

X
co

o

2

> s > 2 >’ £

■|ONC/)NOIinilJ.$N! NVINOSHXUAIS^ S3 I H VB 8 n\| B R AR I ES^SMITHSONIAN^INSTiTUTION ” i

CO q \ CO ___ rr to rr co

< ((& 3 <

y*J cc H <* v,w48rfV ^ V^oW-/ - >•

^ m g '^p/ ?? r 0

an libraries smithsonian jnstitution NouruiiSNi nvinoshiiws saiavaan s

ii ^ ^ ^ i »

Z /a 2 ^ V 2

m p.l^ " m r; x^uT^X ro ^

co !z co £ co st: co

noN NoiiniixsNi nvinoshxiiais saiavaan libraries smithsonian institution i

CO 2 ___ CO 2 .... . Co 2 .v CO

•S xiSSSx I
§ /je? ra?j x

co (&££ co

^ 2 tO * 2 CO k 2

an libraries smithsonian institution noixhxixsni nvinoshxiiais saiavaan i

CO q ___ CO co r;

iti v^uxvXT'v ^ i>i ^\TZTT>^, *—