SUCCESSION OF MAMMALIAN FAUNAS ON
TRINIDAD, WEST INDIES

By

ELIZABETH SCHWARZ WING

A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF

THE UNIVERSITY OF FLORIDA

IN PARTL\L FULFILLMENT OF THE REQUIREMENTS FOR THE

DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

February, 1962

ACKNOWLEDGMENTS

This study could not have been done without the cooperation
of many persons and organizations to whom I am deeply indebted. I
should like to thank my graduate supervisory committee, Drs. Archie
F. Carr, James R. Redmond, and John M. Goggin, Mr. Clayton E. Ray, and
particularly my chairman. Dr. James N. Layne, for their advice and en-
couragement throughout this study. I am also most grateful for the
loan of specimens from Miss Barbara Lawrence of the Museum of Com-
parative Zoology, Dr. Richard Van Gelder of the American Museum of
Natural History, Drs. H. G. Kugler and S. Schaub of the Basel Natural
History Museum, and Mr. J. A. Bullbrook of the Royal Victoria Institute.
Field work during the summer of 1959 was carried out with the assistance
of a National Science Foundation Summer Fellowship for Graduate Teach-
ing Assistants and with the help and hospitality of Texaco Trinidad, Inc.,
and Trinidad Petroleum Development. I would also like to extend my
thanks to Dr. H. G. Kugler, Dr. K. Barr, Mr. J. A. Bullbrook, Dr. C. C.
Wilson, and Dr. I. Rouse for help in connection with various phases
of the study. Finally, I wish to express my gratitude to my parents
and husband for their help and constant encouragement.

11

TABLE CF CONTENTS

ACKNOWLEDGMEHTS 11

LIST OF TABLES iv

LIST OF ILLUSTRATIONS v

INTRODUCTION 1

MATERIALS AND METHODS 3

POST-PLEISTOCENE HISTORY OF THE GEOLOGIC, CLIMATIC, AND

VBGETATIONAL FEATURES OF TRINIDAD 7

Geology 7

Climate 8

Vegetation 17

THE PRE-HUMAN MAMALIAN FAUNA OF TRINIDAD 25

Fossil Manuaals of Trialdad 25

Zoogeographic Sigaiflcance of Pre-human Fauna 32

MAMMALIAN FAUNA AS REPRESENTED IN INDIAN MIDDENS 33

Description of Sites 33

Interslte Comparisons '♦^7

HISTORIC MAMMALIAN FAUNA 53

List of Terrestrial Mammals 5^

Zoogeographic Affinities 65

DISCUSSION 67

SUMMARY AND CONCLUSIONS 68

LITERATURE CITED 73

BIOGRAPHICAL SKETCH 76

ill

LIST OF TABLES

Table Page

1 Rainfall data for selected stations on Trinidad .... 16

2 Measurements of the alveolar cheek tooth row of

various si)ecies of Zygodontcmiys 29

3 Measurements of fossil Zygodontomys 30

h Correlation of periods of occupation on Trinidad

and ceramic styles 3^

5 Relative abundance of mammals at different time

periods within the St. John site 36

6 Relative abundance of mammals recorded from nine
middens ^

7 Measurements of Sigmodon cf . hirsutus 62

iv

LIST OF ILLUSTRATIONS

Figiire Page

1 Shoreline of Trinidad, approximately 17,820

years ago 9

2 Shoreline of Trinidad, approximately 8,000

years ago 11

3 Physiographic regions of Trinidad I3

h Pleistocene vegetation of Venezuela and Trinidad .... I9

5 Pre-Columbian vegetation of Trinidad 21

mTROEUCTION

Animal remains excavated from Indian middens constitute a source
of valuable information to both the archeologist and the zoologist
(Taylor, 1957). These materials are particularly significant, since
they often can be accurately dated, in that they may provide informa-
tion about the transition between Pleistocene and Recent time, a period
in which world biotas underwent rapid changes (Burt, I961) . From the
animals represented in a midden, much about the faunal composition of
an area at a given time may be revealed. Changes in the range of cer-
tain forms may also be indicated. Such changes may in turn throw light
on the nature of past ecological conditions in a given region. Midden
remains occasionally include species that are new to science. More
often, comparison of specimens from middens with those of the same
species now inhabiting the region shows certain differences, analysis
of which may lead to a better understanding of evolutionary processes.
From an archeological standpoint, the study of animal remains from
middens contributes to a more nearly complete understanding of the
environmental milieu of the particular society under consideration.
In addition it provides data on which to base cultural hypotheses as
to the nature of hunting and butchering techniques, food taboos, and the
non-alimentary uses of animals. By integrating data on fossil and re-
cent faunas of the region a greater perspective on this type of as-
semblage is gained and thereby a better understanding of it.

Because of its dyuamic geological history and attendant climatic
and vegetational changes, Trinidad, West Indies, is of considerable zo-
oarcheological interest. It has become separated from the mainland of
South America, from which its fauna and aboriginal population were de-
rived since it beceune inhabited by man.

The present study is an attempt to apply the zooarcheological
approach in an analysis of the mammalian remains from aboriginal sites
in Trinidad. Iliese sites enccMnpass a time-span from approximately 2750
years ago to historic times. In axldition, an effort has been made to
integrate this faunal assemblage with data on both pre-human and his-
toric mammals in order to gain some knowledge of the history of the
mammalian fauna of Trinidad from the Pleistocene to the present day as
related to geologic, ecologlc, and human factors.

MATERIAI2 MD METHODS

The materials upon which this study is based ccMne from a
number of sources. Many of the fossil finds on Trinidad have been
made by Dr. H. G. Kugler in the course of his career as oil geologist
for Texaco Trinidad, Inc. In 1922, he discovered a vertebrate-fossil-
bearing stratum of oil sands while making a test pit at Apex (Trinidad)
Oilfields, Inc. near Pyzabad. The fossils that were collected are in
the Basel Natural History Museum collections. A similar deposit near-
by at the Forest Reserve of Texaco Trinidad, Inc. was found in 1957
when the site was being cleared for oil well Number IO6O. An almost
complete skeleton of Glyptodon was excavated, and shipped along with
some fragments to the American Museum of Natural History. None of
this material has yet been prepared or catalogued. A carbon-lU date
based upon wood associated with Megatherium bones was also taken at
this site.

Ify husband and I conducted field work in Trinidad during the
summer of 1959 with the objectives of collecting more fossil, midden,
and Recent vertebrate material and of getting first hand knowledge of
the ecology of the island. All specimens collected by us are deposited
in the appropriate University of Florida Collection. An attempt to re-
locate the fossil locality at Apex was unsuccessful in spite of help
from the Apex Company. At the site at Forest Reserve, several bones
of larger mammals were collected, and a stratum containing small ver-
tebrate fossils was discovered. The small fossils were extracted from
the oil sands in which they were embedded by soaking large clumps of it

for several days in gasoline, then boiling the clumps to break them
up, and finally washing the residue through screens. Initially 20,
lUO, and 230 gauge screens were used but since very few fossils went
through the 20 gauge screen, subsequent batches were washed only
through a 50 gauge screen.

The bulk of the midden materials was excavated by Dr. I. Rouse,
Yale University, and Dr. J. M. Goggln, University of Florida, in 19^^
and 1953> under the auspices of the Yale Caribbean Research Program
and Graduate School of the University of Florida. Seven sites--St.
Joseph, Mayo, St. John, Cedros, Erin, Palo Seco, and Quineun--were
represented in these collections. The excavations were made by re-
moving one or more sections two meters square. Each section was
divided vertically into a number of levels 20 cm. in depth. The bones
as well as the artifacts from each level were kept separate, so that
stratigraphic changes could subsequently be analyzed. Another col-
lection of midden bones which was available for study was made by
Dr. Kugler at Cedros. This was not an excavation, but rather a sur-
face collection. Additional material came from an excavation made at the
Erin site by Mr. J. A. Bullbrook, of the Royal Victoria Institute. Al-
though this site had been excavated in a manner similar to that employed
by Rouse and Ooggln, the bones were not kept segregated into their
stratigraphic sequence.

In the summer of 1959> I visited six known archeological sites
in Trinidad for the purposes of making ecological observations at
each site and to collect additional surficial material. The sites

studied Included Chagouaray, Palo Seco, Mayaro, Guayaguayare^ Mayo,
and St. John.

The zooarcheological material obtained from Trinidad middens
was in quite good condition compared to similar materieuL from
other areas. Only about 50 percent of the material was too frag-
mentary to permit identification. Of the identifiable component,
about 50 percent was mammal bones, and the remainder was fish, rep-
tile, and bird. The mammal elements identified totaled k^h. In this
study the'kinimum number of individuals" is used as an index of the
relative abundajice of a particular species in a sample of midden
remains. For each species this is determined by the count of the
most numerous skeletal element in a given lot. In some cases this
number may validly be increased by the addition of specimens of an
age group not represented in a series of the most frequent element.
For example, if the most abundant element of a given species in a
level was the mandibular ramus and there were three right adult rami and
one left juvenile ramus, the minimum number of individuals would be four.

Data on the modem fauna are derived largely from the literature
and from my notes on Trinidad mammals. In the course of my own field
work in Trinidad, Recent mammals were collected and their ecology
studied, with emphasis on those forms which are of Importance in mid-
dens. Smaller species were collected in traps. Since the large game
animals are very scarce, I accompanied professional hunters, and acquired
the unused parts of their catches. I also obtained many bones, parti-
cularly skulls and feet of wild mammals, in modem kitchen middens

around the hunters' houses. Further data on Recent mammals were
obtained through the analysis of owl pellets collected by the late
F. W. Urich of St. Augustine and lent to me by Dr. H. G. Kugler and
Dr. S. Schaub from the Basel Natural History Museum.

Where adequate series were available, comparisons of skeletal
measurements of midden and Recent mammals of the same species were
undertaken with the aim of detecting differences in size that might
perhaps be correlated with ecological or other factors.

POST- PLEISTOCENE HISTORY OF THE GEOLOGIC,
CLIMATIC, AND VBGETATIONAL FEATURES OF TRINIDAD

For the understanding of the histoiy of the mammal iagi fauna
of Trinidad, a knowledge of the post-Pleistocene geological events,
emd the climatic and vegetational changes attendant upon them, is
of primary importance. A brief sketch of these changes will, per-
haps, aid in placing the history of the fauna of the island in its
proper context.

Geology
Although Trinidad is commonly thought of as one of the Antilles,
and is politically a member of the West Indian Federation, it is geo-
logically and biologically closely associated with the South American
mainland. The island is composed of sediments probably derived prin-
cipally from the Guiana shield which extends roughly as far south as
the present frontier of Brazil and includes Trinidad, and possibly
Tobago, at the northern limits of its influence. According to Dr.
Kugler (personal communication) several Pleistocene marine terraces
are traceable, and he has recognized at least seven of these, rang-
ing in elevation from five to 330 feet, which can be correlated
closely with similar marine terraces in Venezuela. After the last
glaciation, there was one continuous land mass encompassing what is
now Trinidad, the Gulf of Paria, the Serpent's Mouth, and Venezuela.
Thereafter a general trend of rising sea level is observable. Kolde-
wijn (1958) states that the shoreline at an estimated 17,820 years
ago was at the present 2'+- fathom line (Figure l). By about 8,000

years ago (Flgiu-e 2), the shoreline had risen to the present 12-fathom
mark (Nota, I958). Roughly TOO years ago, the drainage of the Manamo
Branch of the Orinoco River into the Gulf of Paria increased in volume,
thereby further enlarging the Gulf, and completing the separation of
Trinidad as an island from the mainland (Koldewijn, 1958). At the
present time, the Islemd is subsiding, as is evidenced by the coastal
erosion indicated by the landslips onto the beaches. Many instances
can be seen of roads that have been completely undermined by wave ac-
tion and of coconut palms lying uprooted on the beach. The Chagonaray
site on the south coast is being washed into the sea.

Trinidad may be divided into five physiographic regions (Beard,
19^6) . The first of these comprises the mountains of the North Range
(Figure 3) . Although there are three ranges in Trinidad only the
North Range, which is an extension of the coastal cordillera of
Venezuela, may be considered to be actually mountainous, with eleva-
tions over 3,000 feet. The Central and South ranges constitute the
second region. The elevation of these hills scarcely exceeds 1,000
feet. Just south of the North Range lies the third region, one of
dissected alluvial terraces, formed of material from the mountains.
The fourth and most extensive region is the dissected peneplain.
This division includes all of the rest of the land except for a few
coastal swamps which comprise the fifth and final region.

Climate
As a result of its Insular nature and varied physiographic
features, Trinidad has a somewhat varied climate. The most notable

Figure 1: The shoreline of Trinidad about 17,820 years ago.
The solid line is the present 2U -fathom mark which
corresponds to the former shoreline, and the dotted
line denotes the present geographic features of the
area.

10

\

/

_,

1

"\ 1

\\

\ N

V r

^ /-::<ii

)

;
)
I

o

■■i

or "■
13 ..'•,

•r

-s.'-^/v^

Figure 2: Kie shoreline of Trinidad about 8,000 years ago.
The solid line is the present 12 -fathom mark
which corresponds to the former shoreline, and the
dotted line denotes the present geograj^ic features
of the area.

z

u
o

12

Figure 3: The physiographic regions of Trinidad,
(after Beard)

lU

15

climatic zaae is the montane region of the North Range, which is
characterized by heavy rainfall of at least 200 inches per year and
an almost continual cloud cover. The remainder of the island experi-
ences a seasonal climate, vith two dry seasons from January through
April and Seprtember throu^ October, and two rainy seasons, as the
result of shifts in the tradewinds from northeast to southeast. Ttxe
windward coastaJL strip is affected by the wind and salt spray, and is
therefore somewhat drier than the rest of the island. Plainfall data
for selected stations, provided by the West Indies Meteorological
Service, are presented in Table 1.

Evidence indicates that temperatures in the area of Trinidad
underwent an increase, beginning about l6,500 years ago, reaching a
maximiim at 6,000 years ago, and thereafter declining slightly
(aBlliani, 1955)' Tlie enlargement of the Gulf of Paria may have had
a moistening influence on Trinidad. Present day temperatures vary
in the different parts of the island, but on the whole are generally
equable. The temperature ranges at Port -of -Spain, representative of
the seasonal climate, vary from the average monthly maximum of 86.4°
F. in January to 89.8° F. in May, and from the average monthly mini-
mum of 67.6° F. in February to 71.8° F. in September. Daily tempera-
ture fluctuation is almost constant, ranging from a maximum of l8.9°
F. in February to a minlmxun of l6° F. in September. Humidity always
exceeds 50 percent and often approaches 100 percent at night (Beard,
1946).

16

u

• V

t u

i)

0) o
CO o

^ <

(U 0)
3 <M

B o

^

^

9»

00
lA

ON

CO

)§

"S

^ ^

00

co

OJ

0) 0)

•v

^^

4)

s

CO

0]

•rt

<u

■P

o w

t^

C0

•H -

+J

H d

CO

(g^

<v

•-D

f3

u

(U

m

vH

J .

^

H

o S

0)

§18

£

<^§

0) 4)
(1| O

s

en

C\J

CO

CO

CO
C7\

^

£

2
^

S

8

>^

0)

+i "^

CD (U

-P H

>

to m

U B

<u 3

•tJ -i --^

O r-i H

m S: W

17

Vegetation

Striking changes in the vegetation of Trinidad have accompanied
the geologic and climatic events that have influenced the island.
The following discussion of the vegetation is based upon Beard's
monograph (19^6) The Natural Vegetation of Trinidad. At the time
when there was a continuous land mass including Trinidad and the
mainland, there was also an extension of the vegetation types of
Venezuela into Trinidad (Figure k) . The North Range was forested,
as were some of the higher parts of the Central and South ranges,
while the remainder of what is now Trinidad and the Gulf of Paria
were occupied by savannah. The savannah was dissected by streams
bordered by gallery forest. The savannah formation which is now wide-
spread in Venezuela as the "llanos," reaches to within 50 miles of
the Gulf of Paria and persists as relics in a few small areas on
Trinidad itself. The distribution of savannahs on Trinidad is con-
trolled by edaphic factors. They occio" in flat areas where the soil
is poorly dratied and the subsoil is impermeable. The soils became
rejuvenated, accompanying geological changes in the area, and this
resulted in the spread of forests from the streams and hills, eventu-
ally to cover nearly the entire island.

Particular climatic and edaphic factors influenced the develop-
ment of several vegetational types. Beard (19^+6) recognizes six major
plant formations (Figure 5): seasonal forest; dry evergreen forest;
montane forest] intermediate forest; sweunp; and marsh. Four of these
are further divisible into associations. The seasonal formation

18

includes evergreen, semi-evergreen, and deciduous forest depending
upon the availability of moisture.

The evergreen seasonal forest requires 70 inches or more of
available moisture, and is characterized by a discontinuous emergent
stratum more than 100 feet in height, and "... an almost continu-
ous canopy layer at UO-90 feet, and almost continuous lower story at
10-30 feet" (Beard, 19^^ P- 38). Lianas and epiphytes are common.
Conspicuous species of trees are Spondias mambin (hogplum), Carapa
guianensis (crappo), Eschveilera subglandulosa (guatecare). Pant a -
clethra macroloba (bois mulatra or fineleaf), Maximiliana elegans
(cocorite palm), and Mora excelsa (mora). The laxgest trees are
buttressed, and many are Important sources of food for wild animals.

The semi-evergreen seasonal forest exi::t3 where 50-70 inches
of available moisture occurs. It is an open forest characterized by
"... a discontinuous emergent layer at 6O-8O feet, and an almost
continuous layer at 20-UO feet" (Beard, 1946, p. 38). Lianas are
veiy abundant, but epiphytes are not. Buttressed trees are rea*e.
Common species are Bravas ia Integerrima (jigger-wood), and Hura crepi-
tans (sandbox).

The third form of seasonal forest, the deciduous, occurs in
areas of 30-50 inches of available moisture. It io characterized
by having a continuous canopy at IO-3O feet, and a few emergent trees
up to 60 feet in height.

The dry evergreen or littoral woodland formation occurs along
the coastal strip which is exposed to continual wind off the ocean.

I

Figure k: The Pleistocene vegetation of Venezuela and Trinidad,
(after Beard)

20

J^^.

% 0

•^^

-~~<K

Ul UJ

> o-

jr <

Ui

v>

Ul
UJ

a.
o

S I

Ui

(r
o

X

z
zr

>

to

Figure 5= The pre-Columbian vegetation of Trinidad,
(after Beard)

22

(/^
liJ

or
O

J

o
<

UJ

m

Bmm mn

z

1-

UJ

(n

I

UJ

UJ

<T

13

?•

cc

;^

UJ

>

o

UJ
UJ

o
u.

UJ

1-

Z"

■>

UJ

o

<

i

o

cc

UJ

UJ

o

(/^

Ul

o

>

UJ

<

UJ

5

a.

T

K

a:

5^

cy

-z

UJ

<

rr

>-

or

o

1-

^

<i

Q

5

7

Ol

2

23

Two associations of this formation exist in Trinidad: the Coccoloba
uv i f e ra - HI ppomane mancinella (sea grape-manchineel), and Roystonea
oleracea-Manilkara bidentata (palmiste-balata) .

Beard divides the montane formation into three association:
lower montane rain forest, montane rain forest, and elfin woodland.
These replace one another at successively higher altitudes. The
height of the canopy decreases with increasing altitude, from 70-100
feet in the lower montane rain forest, to 60 feet in the montane
forest, to 15-25 feet in the elfin woodland.

The intermediate or seasonal montane formation is intermediate
in characteristics between the seasonal and the montajae formations.

The swamp formation is composed of four types: swamp forest,
palm swamp, herbaceous swamp, and mangrove woodland. The greatest
area of swamp forest is the Oropouche swamp. In it the principal
tree species is Pterocarpus officinalis (swamp blackwood), which form
a closed canopy at 60 feet and is conspicuously buttressed. The
species of palm occurring in the palm swamps near the sea is Roy-
stonea oleracea. The marsh formation consists of marsh forest, palm
marsh, and savannah. The swamp and the marsh formations develop
under the influence of particular edaphlc factors. The marsh forma-
tion is characterized by only seasonal inundation and the swamp
formation by permanent inundation. Savannahs of the marsh forma-
tion are characterized by a dominant short bunch-grass, and a scat-
tering of gnarled shrubs attaining a height of up to 12 feet. Several
species of grasses occur, and the chief shrubs are Byrsonlma crasslfolia

2k

(cavanna serratta) and Curatella amerlcana (roughleaf) or Chryso-
balanus pellocarpus_(fat pork) . These fire-resistant shrubs have
in 3ome areas replaced the former Myrcia-Roupala association which
still occurs widely in the Venezuelan llanos.

The natural condition of the vegetation as sketched above
has been markedly modified in recent times by the agricultural
activities of man. About half of the Island is still forested. The
land that i? under cultivation was formerly principally seasonal
forest. Cocoa and coffee are grown mainly in the North and Central
Ranges, whereas sugar cane Is grown in the flatter area between
these ranges. Rice is cultivated in the swamp areas and coconuts
along the coast, particularly the east coast. There has been some
reforestation with teak in the southern watershed.

THE PRE-HUMM MAMMALIAN FAUNA OF TRINIDAD

The samples of the manimalian fauna from the three different
time periods to be treated will be considered separately. The old-
est period, represented by fossils, may be designated the pre-human.
The Recent fauna, representing two periods, is composed of Indian
midden bones and of specimens frcan a period that may be tenned the
historic. The time range for the pre-human period includes the
Pleistocene epoch and probably the early post -Pleistocene as well.
A carbon-l^i- date of greater than 3^>000 years has been determined
from wood associated with Megatherium remains at a depth of 12 feet
at Trinidad's most important fossil locality at Forest Reserve, the
so-called Glyptodon site. During this time Trinidad was still broad-
ly connected with the South American mainland, and savannah dissected
by streams edged with gallery forest covered the area.

The pre-human fauna of Trinidad known at present is small,
reflecting the limited amount of collecting carried out in thl field.
The fauna is represented by three orders: Edentata, Rodentia, and
Proboscidea. The edentates are represented by the families Mega-
theriidae, Mylodontidae, and Glyptodontidae, the rodents by the
Cricetidae, and the proboscidiEins by the Gomphotherildae. All of
these ffiunilies, excepting the Cricetidae, are now extinct.

Fossil Mammals of Trinidad
Megatheriidae
This family is represented by Megatherium americanum Blumb . ,

25

26

which has been found at several localities. An astragalus was
found at Karamat mud volcano and identified by Schaub (l935)' Parts
of a skeleton were found at Los Bajos. The site at Forest Reserve
also yielded Megatherium remains. Part of a tooth found at the time
the Glyptodon skeleton was excavated (195?)^ and along with the other
materials, was sent to the American Museum of Natural History. In
1959 I found many dermal bones, a complete fourth tarsal, a complete
tibia and fibula, part of a mandibular ramus, a rib, and three
thoracic vertebrae. This fossil site is by far the most productive
on Trinidad. It is a river bed cut to a depth of 15 feet into
Miocene sands and clays, and filled with oil impregnated silt, sand,
clay stone pebbles, plant remains ranging from large logs to twigs
and stems, and vertebrate and insect fossils. This site is known
locally as the "Glyptodon site."

^^lodontidae
A single tooth of Mylodon sp. was excavated along with other
materials at the "Glyptodon site" and has been identified by me. It
is now deposited in the collection of the American Museum of Natural
History.

Glyptodontidae
An almost complete Glyptodon skeleton, from which the "Glypto-
don site" takes its name, was found while the oil well site was be-
ing prepared in 1957. Almost the entire skeleton was found in place,
and is now, unprepared, at the American Museum of Natural History.

27

I found abundant dermal bone but no other skeletal parts while working
at this site.

Crlcetidae

At the "Glyptodon site" I found a layer with a concentration
of small vertebrate remains two feet above the layers containing the
large bones and logs. A single mammal was represented which is refer-
able to the genus Zygodontanys . Associated with the Zygodontcmys are
unidentified remains of a few birds, a turtle, a frog, and many in-
sects. Another site containing small vertebrate fossils is at Apex
near the "Glyptodon site" and geologically similar to it was found
by Dr. Kugler. Zygodontomys sp. , the only manmal represented, is
associated with bird and insect remains (Blair, I927) .

Named species of the genus Zygodontomys axe distributed in
northern South America and Gouthem Central America. Forms of Zygo-
dontomys have been described from south of the Amazon River, from
northeastern Brazil and the Matto Grosso, but these are included cor-
rectly in the genus Akodon (Tate, 1939). Tate (1939) divides the
Zygodontomys into two primary groups of small and large size. In-
cluded in the group of large-sized species is Zygodontomys brevi-
cauda, which occurs abundantly on Trinidad, and other species dis-
tributed in Panama and northern Colombia. Itie group of small-sized
species include ones distributed in Central America, Colombia, Vene-
zuela, and the Guianas. Tate (1939) also reports that among the
specimens referable to Zygodontomys microtinus stellae from the Es-
merelda savannahs at the foot of Mt. Duida there are some individuals

28

with a normal upper molar tooth row of i+.2 mm. and there are others
with the shortest known tooth row of 3.7 to 3.8 mm. for this genus.
Gyldenstolpe (1932) notes that Zygodontomys thc»aasi distributed in the
Cumana district of Venezuela is allied to ZyRodontomys brevicauda but
Is much smaller. The fossil Zy|godontomys from Trinidad is also similar
to Zygodontomys brevicauda but a great deal smaller.

Since the fossil material is largely fragmentary only a few
measurements could be taken. The most useful ones were the alveolar
length of the upper and lower cheek tooth rows. A comparable set of
measurements was taken on Zygodontomys brevicauda from owl pellets
found by F. W. Urich on Trinidad, Zygodontcxnys microtlnus stellae from
Esmeralda Savannah, Venezuela, and Zygodontomys thomasi frcan Cristobal
Colon, Venezuela. These were taken to the nearest one tenth of a mm.
and appear in Table 2.

The fossil Zygodontomys differs from the other species of Zygodon«
tomys available for comparison in several respects. The most outstand-
ing characteristic of the fossil form is its small size (Table 3)- Ad-
ditional characteristics that distinguish the fossil are exceptionally
long anterior palatal foramina and deeply concave zygomatic plate. In
the younger individuals (specimens with unworn teeth) the foramina extend
posterior to the level of the posterior lingual root of the first molar.
In the oldest individuals of the fossil (with heavily worn teeth) and
in young Individuals of other species the foramina extend posteriorad
only to the anterior root of the first molar. Diagnostic dental charac-
teristics include a deep anterior median fold in IT, a broad U-shaped

29

TABLE 2

MEASUREMENTS (IN MILLIMETERS) OF THE ALVEOLAR CHEEK TOOTH ROW
OF VARIOUS SPECIES OF ZYGODOKTCMfS

Number

Mean

Range

Zygodontomys sp.

Z. thomasi

Z. brevicauda

Z. microtinus stellae

upper

7

3.7

3-3-^.1

lower

2k

3.6

3.1-i^.3

upper

5

U.6

1+.2-U.9

lower

6

5.1

U.9-5.2

upper

13

5.1

U. 7-5.6

lower

--

--

--

upper

6

U.l

3.8-U.2

lower

6

U.l .

k.O-k.3

iu

<tH

o

0)

r-J

•p

^

-a

•H

d

<M

0

(U

H

Xi

^

^

•H

a

a

^1

O 4^

Si S <D

+5 0, -P

-O «) 03

:^ ^^

s t^ Pk

^

00
-it

CO

OJ

CO

o

CO

31

first primary fold, and presence of a mescstyle, anterior labial style,
anterioi* lingual style, or enterostyle in the M"^ and Vr . The anterior
median fold is exceedingly well developed and is present in all speci-
mens except those in which it is obliterated by excessive wear. This
fold is otherwise present in only a few specimens of Zygodontomys brevi-
Cauda and in these it is merely a vestige. The first primaiy fold is
broad in the fossil specimens and narrow in the other species. In the
M"^ the mesostyle is present in all of the specimens and the other three
styles occur in between 6h percent and 73 percent of the l6 specimens.
In the NT the enterostyle is present in 78 percent of the nine specimens
and the mesostyle in 67 percent. No styles are found in any of the other
species. In the lower cheek tooth series the mesostylid is present on
the M in 9U percent of the 17 specimens. The articular process of the
mandible is also distinctive, being short and broad, whereas in the
other species it is long and narrow.

According to Tate (1939), "The South American Zygodontomys ap-
pear to be a grassland- inhabiting group of mice. I have little doubt
they extend southward from Sucre across the llanos to the Orinoco and
continue into Guiana wherever savannas exist." It appears, therefore,
that when the llanos extended into Trinidad there was opportunity for
the spread of this genus into Trinidad from South America.

Gomphotheriidae
Cuvieronlus hyodon (Fischer) is represented by two teeth found
at Los Bajos and identified by Schaub (1935), part of a tooth found
oridinally at the "Glyptodon site," and another fragment of a tooth
found by me at the same site.

32

Zoogeographlc Significance of Pre-human Fauna

The known fossil fauna of Trinidad is sparce in numbers of
species. Further collecting will probably increase this number con-
siderably. The Edentates are of South American origin but the three
genera, Megatherium, Mylodon, and Glyptodon. found in Trinidad were
distributed throu^out much of both North and South America in the
Pleistocene. Megatherium ranged from South America to southeastern
United States in the Pleistocene. Mylodon ranged from South America
to southwestern United States and was contemporaneous with man. Glyp-
todon was apparently very numerous throughout South America and
southern United States in the Pleistocene. Of the Proboscidea, the
genus Cuvleronius probably originated in southwestern United States
(Osbom, 1936) and was distrihuted widely in South America and in
southern North America in the Pleistocene. The rodent Zygodontomys
evolved in South America and exists now in South America north of the
Amazon River and in Centred. America as far north as Costa Rica. Certain
forms south of the Amazon River which have been referred to the genus
are probably referable instead to the genus Akodon (Tate, 1939) • All
known representatives of the fossil fauna of Trinidad are herbivores
and are characteristic of grassland habitats, which were widespread in
Trinidad during the Pleistocene.

MAMMALIAN FAUNA AS REPRESENTED IN INDIAN MIDDENS

The present knowledge of the pre-Columbian mammalian fauna of
Trinidad is based on skeletal material excavated from nine Indian
sites. These sites represent all the cultural levels known on Trini-
dad. The sites, like all those known from the island, are located near
the coast, indicating an orientation of the people toward the sea.
The pre-ceramic sites are, of course, the oldest, dating here from
2,750 (i 130) years ago. These are followed by sites exhibiting
four distinct ceramic styles (Bullbrook, 1953) that span the time
from the pre-ceramic period to contact with the Spanish after 1532.
The first of these periods is characterized by the Cedros style,
the second by the Palo Seco style, the third by the Erin style, and
the last by the Bontour style. Pottery of the last period is some-
times associated with Spanish ceramic work (Table k) .

Description of Sites
1. St. John

The St. John site is located on a high bluff south of the
Godineau River above the Oropouche Swamp, and at the end of the St.
John Road. This location takes advantage of the evergreen seasonal
forest on the hill and the mangrove swamp habitats. This site was ex-
cavated in 1953 by Drs. Rouse and Goggin (Rouse, 1953)- The soils
are light sands that are freely drained. The shell midden at this
site is extensive, reaching four feet in depth and 125 feet in dia-
meter. Material of two sucdtssive occupations has been obtained

33

■5^

TABLE k
CORRELATION OF PERIODS OF OCCUPATION ON TRINIDAD AND CERAMIC STYLES
(after Cjruxent and Rouse)

Venezuelan Psriods of Trinidad Ceramic

Occupation Styles

V

St. Joseph

IV

Bontour

III

Erin

Palo Seco

II

Cedros

Preceramic

I

?

'35

here. The first of these is pre-ceramlc aad is comparable to the
Ortoire site, for which there is a carbon-li*' date of about 2,750
years. The secoad period of occupation is characterized by pottery
of the Bontour style, with . ome European ware. This excavation con-
si-.ted of 13 level units, four of which were pre-ceramic, five proto-
historic, and four a mechanical transition between the two. Each of
these three groups of levels has approximately the same proportion
of each mammal species in them, but the pre-ceramic group contains
more bones (Table 5). This is probably the result of differences in
the econcany of the two cultures, the pre-ceramic population depend-
ing exclusively on hunting and fishing, while agriculture also con-
tributed to the food economy of the protohistorlc population. An-
other factor that might explain the difference in the relative quan-
tities of bones and artifacts could be that in the ceramic sites a
larger proportion of the bulkier artifacts would correspondingly re-
duce the number of bones found in a given area and level.

List of Species Identified

Dldelphis marsupialis, opossum: 10 individuals.

Dasypus novemclnctus, nine-banded armadillo: 27 individuals. As in
all other sites thejre is a far higher proportion of bone fragments
to minimum number of individuals for the armadillo than for any
any other mammal because of the numerous bones in the shell. On
the average for every 1+ to 5 individuals, there were over 75
armadillo bones identified, but fewer than 50 bones identified
for the same number of any of the other mammals.

3b

TABLE 5

RELATIVE ABUNDANCE OF MAMMALS AT DIFFERENT TIME PERIODS
WITHIN THE ST. JOHN SITE

Preceramic

Mechanical
Transition

Protohistoric

Didelphis

Dasypus

Alouatta

Rhipldomys

Coendu

Agouti

Dasyprocta

Echlmys

Proechirays

Procyon

Lutra

Fells

Pecari

Mazaoa

3

15
5
1
1

11
k
2
9
1

1
27
10

18
7

12
7

37

Alouatta seniculus^ red howler monkey: 10 indlvldueils.

Rhlpldomys couesi: 1 individual.

Coendu prehensilis, porcupine : k Individuals

Agouti paca^ paca: 26 individuals in all; 8 in the proto-historic
group of levels, and 11 in the pre-cei-amic group. The remaining
7 were found in the levels of mechanical transition.

Dasyprocta aguti, agouti: 13 individuals in all', 6 in the proto-
historic group of levels, and k in the pre-ceramic group; 3 in
the transitional group of levels. Tlie paca and agouti were repre-
sented numerously enough in all sites to allow useful comparison.

Echimys armatus, spiny rat: 4 individuals.

Proechimys guyannensia, spiny rat: 12 individuals.

Procyon cancrlvorus, crab-eating raccoon: 3 individuals.

Lutra enudris, otter: 1 individual.

Felis pardalis, ocelot: 1 individual.

Pecari tajacu, collared peccary: 57 individuals in all; 12 in the
protohistoric group of levels, and 27 in the pre-ceramic group;
l8 in the transitional group of levels.

Mazama amerlcana, brocket: 24 individuals in all; 7 in the proto-
historic group of levels, and 10 in the pre-ceramic group; 7 in
the transitional group of levels. As with the paca and agouti,
the peccary and brocket were sufficiently well represented at
each site to allow comparison.

Other vertebrates: As at all other sites there were very few birds
represented, Cairina moshata (muskovy duck) being identified.

38

There were a few land turtles, but no sea turtle remains at this
site. Fish remains, particularly cat fioh, were abundant here as
at all other sites.
2. Cedros

The Cedros site is situated on the United States Army's
Green Hill Reservation in the extreme southwest corner of the is-
land and is located on a slight elevation within one fourth mile of
the sea. The area is now a coconut plantation and i3 inhabited; it
was, however, formerly palm swamp surrounded by evergreen seasonal
forest. The sandy soil in this area is classified as having impeded
drainage, since it is close to sea level and the water table. Cedros
is a shall mound yielding purely Cedros style pottery. The materials
studied were from an excavation made by Dr. Rouse and Mr. Bullbrook
(Rouse, 1953), consisting of 25 level units, and from tvo random
surface collections made by Dr. Kugler and by me. The economy of
this culture, as i;; true for all the ceramic cultures known on Trini-
dad, was probably based on hunting, fishing, gathering, and agricul-
ture.

List of Species Identified
Didelphis marsupialis , opossum: 5 individuals.
Dasypus novemcinctus, nine-banded armadillo: 15 Individuals.
Tamandua longicaudata, ant eater: 2 individuals.
Alouatta seniculus, red howler monkey: 1 individual.
Agouti paca, paca: I6 individuals.
Dasyprocta aguti, agouti: 1^4^ individuals.

39

Proechlmys guyannensls, spiny rat: 1 individual.

Canis cf. famlllarlSj dog: 3 Individuals. These were represented

only in the surface collection, and were undoubtedly added in

recent times.
Procyon cancrivorus, crab-eating raccoon: 2 individuals.
Herpestes auropunctatus, mongoose: h Individuals. This species,

like the dog, was found only in the surface collection and was

deposited in recent times.
Pecarl tajacu, collared peccary: 1^ individuals.
Mazama americana, brocket: 23 individuals.
Bovid: 1 Indlvidml. This was found only in surface collections

and was deposited in recent times.
Trlchechus manatus, manatee: 2 individuals.

3- Palo Seco

The Palo Seco midden is on the south coast of Trinidad. It
is located on a ridge at an elevation of 90 feet and is surrounded
by hills. The shoreline is approximately 200 yards distance. The
Palo Seco site is now at the center of the Trinidad Petroleum De-
velopment camp, but it was formerly in semi -evergreen forest habitat.
The soils of thi.. area are predominantly clays and silts with Impeded
drainage. The excavation, made by Dr. Rouse (Rouse, 1953) at this
site consisted of 57 level units, and yielded two pottery types,
Cedros and Palo Seco.

kQ

List of Species Identified
Dldelphls marsuplalis^ opossum: Ik individuals.
Dasypus novemclnctus^ nine-banded armadillo: l8 individuals.
Tamandua longlcaudata^ ant eater: 2 individuals.
Alouatta seniculus, red howler monkey: 3 individuals.
Coendu prehensilis, porcupine: 6 individuals.
Agouti paca, paca: kl individuals.
Dasyprocta aguti, agouti: hj individuals.
Proechimys guyannensla^ spiny rat: 2 individuals.
Fells pardalis, ocelot: 1 individual.
Pecarl tajacu, collared peccary: 26 individuals.
Mazama americana, brocket: 76 individuals.
Bovid: 1 individual. This was found at the surface, presumably

from recent occupation.
Tapirus sp. , tapir: 1 individual. A right lower second pre-molar

was found. This animal occurs on the mainland of South j^merica.
Trlchechus manatus, manatee: 1 individual.
Cetacean: 1 individual.
Other vertebrates: Sea turtle remains were abundant at this site

and at the other south coast site. There were few or no remains of

land turtles.

h. Erin

The Erin midden is Ju t west of Palo Seco in the center of the
town of Erin under the present police station. It is on a hill that
was covered with semi -evergreen forest, about one-half mile from the

kl

beach and the Erin River. The soils here, as at Cedros and Palo Seco,
are poorly drained clays and silts. The site yielded Palo Seco and
Erin style pottery. The materials that were examined came from an
excavation made by Dr. Rouse (Rouse, 1953) that consisted of 69
level units, and one excavation made by Mr. Bullbrook, in which, how-
ever, the specimens from different levels were not kept separate.

List of Species Identified
Didelphis marsupialis, opossum: I8 individuals.
Caluromys philander, woolly opossum: 2 individuals.
Dasypus novemcinctus, nine-banded armadillo: 10 individuals.
Tamandua longicaudata, ant eater: 2 individuals.
Alouatta seniculu3, red howler monkey: 3 individuals.
Sciurus granatensis, squirrel: 1 individual.
Nectomys squamipes : 1 individual.
Coendu prehensilis, porcupine: k individuals.
Agouti paca, paca: l^i individuals.
Dasyprocta aguti, agouti: 59 individuals.
Echimys armatus, spiny rat: 2 individuals.
Proechimys grjyannensis, spiny rat: 2 individuals.
Canis cf. familiaris, dog: 1 individual. This came from a surface

level .
Procyon cancrivorus, crab-eating raccoon: 1 individual.
Lutra enudris, otter: 1 individual.
Felis pardalis, ocelot: 2 individuals.

h2

Herpestes auropunctatus, mongoose: 1 individual. This came from
Mr. Bullbrook's collection and presumably was collected from the
surface.

Pecari taj acu^ collared peccary: 11 individuals.

Mazama americana, brocket: 6l individuals.

Trichechus manatus, manatee: 1 individual.

5. Mayaro

This site is Just back of the Atlantic coast shore in a coco-
nut grove belonging to the St. Bernard Estate, a short distance di-
rectly south of Cape Mayaro. The ceramic style at this site is pre-
dominantly Palo Seco. The bones available for study consisted of
two small surface collections, one made by Dr. J. Hill of Port-of-
Spain and the other by me, both in 1959 •

List of Species Identified
Dasypus novemclnctus, nine-banded armadillo: 2 individuals.
Agouti paca, paca: 1 individual.
Dasyprocta aguti, agouti: 1 individual.
Proechimys guyannenais, spiny rat: 1 individual.
Pecari tajacu, collared peccary: 3 individuals.
Mazama americana, brocket: 2 individuals.

6. Quineun

The Qulnam midden is located in semi -evergreen forest east
of the Palo Seco site, near the beach and the Palmiste River. The
soil is similar to that of Erin and Palo Seco. The excavations made

k3

by Dr. Rouse (Rouse, 1953) > consisting of 58 level units, produced
pottery of Palo Seco, Erin, and Bontour styles.

List of Species Identified
Didelphis marsupial is, opossum: k individuals.
Dasypus novemcinctus, nine-tanded armadillo: 10 individuals.
Tamandua longicaudata, ant eater: 2 individuals.
Alouatta seni cuius, red hovler monkey: 2 individuals.
Agouti paca, paca: 16 individuals.
Dasyprocta aguti, agouti: 26 individuals.
Lutra enudris, otter: 1 individual.
Felis pardalis, ocelot: 2 individuals.
Pecari tajacu, collared peccary: 23 individuals.
Tayassu pecari, white-lipped peccary: 5 individuals. This was the

only site where remain? of this species were found.
Mazama americana, brocket: 77 individuals.

7. Chagonaray

This site is half way between Palo Seco and Quinam on
Chagonaray Point. It has quite the same ecological conditions as
Quinam, but is undergoing coastal erosion. The pottery found was
principally of Erin and Bontour style, but also included sane of
the Barrancas style typical of the Lower Orinoco region (Rouse,
personal communication). The macmal material examined was from a
surface collection made by me.

kh

List of Species Identified
Dasypus novemcinctus, alne-banded armadillo: 1 individual.
Agouti paca^ paca: 1 Individual.
Dasyprocta aguti, agouti: 1 Individual.
Proechiniys guyannensis, spiny rat: 1 individual.
Pecari tajacu, collared peccary: 3 individuals.
Mazajaa americana^ brocket: 2 individuals.

8. Mayo

Mayo is a Spanish mission site located in the Montserrat
Hills of the Central Range, seven miles from the Gulf of Paria. A
Catholic church has been built over the site. Formerly this area
was covered by evergreen seasonal foireat. The soil is li^t and
freely drained. On the basis of the Bontour style pottery and the
European ware found, this site has been dated (Goggin, personal commu-
nication) from the late seventeenth or early eighteenth centuries.
The material exsunined came from an excavation of 8 level units made
by Dr. Goggin (Rouse, 1953) and from my surface collection.

List of Species Identified
Dasypus novemcinctus, nine-banded armadillo: Ik individuals.
Tamandua longlcaudata, ant eater: 3 individuals.
Alouatta seniculus, red howler monkey: 6 individuals.
Coendu prehensills, porcupine: 2 individusLLs.
Agouti paca, paca: 12 individuals.
Dasyprocta agutl, agouti: 5 individuals.

1*5

Canls cf . familiarlSf dog: 1 individual.

Pecarl tajacu> collared peccary: 17 Individuals.

Mazama americana, brocket: 8 individuals.

9. St. Joseph

St. Joseph was the first Spanish capital of Trinidad, and
the site is located in the town of St. Joseph on a branch of the
Caroni River on the south side of the North Range. This area was
semi -evergreen forest. The excavation made by Dr. Rouse and Dr. and
Mrs. Groggin (Rouse, 1953)* consisting of 13 level units, yielded Bon-
tour style pottery and European ware.

List of Species Identified
Coendu prehensilis, porcupine: 1 individual.
Agouti paca, paca: 1 individual.
Dasyprocta aguti, agouti: 3 individuals.
Mazama americana, brocket: 2 individuals.
Bovid: 2 individuals.
Cetaceaji: 1 individual.

46

Of

o >>

6^

o o

H O

t^

H H H

\OV0
H <M

H OJ OOLTN t-

CVJ Oi

VO V£» rH iH H

00 CM O OJ m ,
H H

LTv U% OJ H
H

HJ'-4-O\OJ0JiHH<HOJrHr-| H

r-i H rH CM

O C~- O H -d- VO m_d- CM
rH CVl H W H r-1

u% cy

D— 00

W <U >H

rH h O

4) 4) d)

Pm a Ph

(d 3

^1

CO o d

ft -H +J

0) ^1 0)

e5 M O

^7

Inters Ite Comparisons

The mammals that have been Identified from Indian middens com-
prise a surprisingly complete list of the larger animals found in Trini-
dad today. One group of animals that is, however, entirely missing Is
the bats, although there are a great number of bats (Goodwin and
Greenhall, I961) now living on Trinidad. Another group, the cricetid
rodents, is represented in the middens only by the two largest genera,
Rhipidomys and Nectcanys. Of the four species of marsupials now occurring
on Trinidad, two species of the genus Marmosa were not found in the mid-
denb. Presumably the bulk of the animals represented were used by the
Indians for food. This autcmatically excludes from the middens a cer-
tain portion of the natural fauna of the area, either because the Indi-
ans would make no effort to catch animals they considered useless, or
because they might not have the technological ability to catch other-
wise edible forms. Thus, for example, although there seems no reason
to suppose that the bat population on Trinidad was markedly inferior to
that at the present time, it Is not wholly surprising that no bat re-
mains are to be found in the middens.

The smallest animal represented in the middens is Rhipldonys .
With three exceptions, all larger animals now known from Trinidad occur
in the middens. One of those not represented is the small arboreal ant-
eater, Cyclopes didactylus. This animal, aside from being relatively
small, emits a very mournful wail, which may have caused it to be
avoided. The capuchin monkey, Cebus albifrons, was also not found.
This species has a large gap in its distribution on the mainland, and

kQ

the closest ally to the Trlnldadian form occurs around the Maracalbo
Basin in eastern Venezuela and nowhere in the area closer to Trinidad.
Hershkovitz (19U9) has suggested two possible explanations for its ap-
pearemce on Trinidad: that Cebus albtfrons may at an earlier time have
had a wider distribution than it does at present, or that it was
brought to Trinidad by early Inhabitants as a pet. In regard to the
second of these suggestions, it seems highly doubtful that, had the
monkey been kept about the camps, no remains of it would be found. The
absence of this species frata the midden material does not, however,
seem necessarily to invalidate the first suggestion that it was in-
digenous to Trinidad. PoBsibly the explanation lies in the fact that
the monkey is of relatively small size, and may have been undesirable
as an article of food. The third animal which might have been expected
but was not recorded is the tayra, Eira barbara. However, remains of it
have been recorded by Bullbrook (1953) from his excavations at Palo Seco.

Remains of two other large animals which may exist on Trinidad —
the coati-mundi, Hasua, and the water opossum, Chironecte3--were also
absent from the middens. Probably Individuals of these species find
their way over from the mainland from time to time. Mr. L. Wehekind
(personal communication) has informed me that the coati-mundi possibly
occurs in the North Range. There is a somewhat better record of the
water opossxim (GreenhaLLl, 1956) based on a photograph, and the descrip-
tion of a strange opossum shot by a hunter in Guayaguayare at the south-
eastern tip of Trinidad. The absence of remains of these animals in
the middens, taken together with their present status on Trinidad, would

h9

seem to indicate that in the past as now they were either merely oc-
casional visitors to the island, or, if native, so rare as to be of
no consequence to the fauna.

Only one mammal not previously reported from Trinidad was repre-
sented in the midden samples. This is the tapir, Tapirus , of which the
remains found consisted of a single premolar. The tapir now exists on
the mainland, and it is quite conceivable that the tooth found at Palo
Seco was brought there by man. Ttxe white-lipped peccary, Tayassu pecari,
has been reported from Trinidad, but it is questionable whether it still
exists there or ever did in considerable numbers. Remains of it were
found at Quinam. Another mammal, the savannah deer, Odocoileus gymnotis,
which does not at present occur on the island, was identified from the
Erin midden by Dr. S. Schaub, according to a personal communication by
Dr. I&igler. I found no remains of this animal, however.

The surface levels at the Cedros, Palo Seco, and Erin sites con-
tained remains of introduced animals that were probably deposited at
the site subsequent to occupation by the Indians. Each of these sites
is in a populated area, making this supposition even more likely. At
Mayo and St. Joseph, as a result of colonial contact, these introduced
forms were incorporated in the middens. Bovid remains appear in the
Cedros, Palo Seco, and St. Joseph Middens. The mongoose, Herpestes a.
auropunctatus , that was introduced into Trinidad in I87O, occurred in
the Cedros and Erin sites. Remains of a small dog, Caais familiaris,
were found at Mayo and Cedros.

50

In Trinidad, considering all sites together, 25 species of
meunmals are represented (Table 6). Much can be learned from those
species which were caught in sufficient quantities to make there, on
the basis of the evidence of the remains found in the middens, the
chief sources of meat in the Indian's diet. In every site, the five
most abundant species were the armadillo, Dasypus novemcinctus; paca,
Agouti paca; agouti, Dasyprocta agutij collared peccary, Pecari tajacu;
and brocket, Mazama americana. These five fonus make up more than 75
percent, in number of individuals, of the animeds at each of the sites.
The incidence of these forms at the six largest sites was analyzed
statistically for randomness by a contingency table. The results
(x a 115.09, h degrees of freedcan) show that the observed differences
are highly significant. There are two possibilities to explain this
difference; either food preference by the Indians or the local animal
population. Since inhere is no cultural evidence to indicate a differ-
ence in food preference, the conclusion that is drawn is that the three
smaller ones, the armadillo, paca and agouti, and the two larger ones,
the peccary and the brocket, were probably equally sought. Therefore,
the relative abundance of each at the different sites probably reveals
their actual abundance in each area rather than selection by the Indians.

At the three south coast sites of Quinam, Palo Seco, and Erin,
the percentage of armadillo among the mammals represented within each
site is one half or one third as great as within the sites of Mayo, St.
John, and Cedros (see Table 6). This distribution may well be correlated
with the soil best suited to the burrowing activity of the armadillo.

51

At the three south coast sites, the soils are predominantly clays and
silts with relatively poor drainage. Thus soil conditions would not
he as favorable for burrowing as light soils with free drainage or ex-
cessive drainage found in the aj-ea of the Mayo and St. John site. The
sandy soil at the Cedros site, although classified as having impeded
djrainage because it 1 . close to sea level and the water table, would,
however, provide suitable substrate for burrows. Ihe evidence suggests,
therefore, that the relative abundance of these animals is related to
the character of the soil as Indicated by midden remains.

The relative abundance of agouti and paca at the various sites
would also appear to be correlated with the ecology around each site.
At the two sites in the denser evergreen seasonal forest — Mayo and St.
John — paca were found to be approximately twice as abundant as agouti,
whereas the reverse was found to be true at the three sites — Erin, Palo
Seco, and Qulnam — located in more open semi -evergreen foi*est. At Cedros
near evergreen forest, but actually in palm swamp, paca and agouti re-
mains are almost equally abundant. Kie paca is characteristicaJLly a
forest ajaimal, feeding on fruits and eggs laid on the ground, eilthough
it does venture into the edge of clearings as well. The agouti is usu-
ally found in more open forest.

A similar pattern is seen when the relative abundance of deer and
peccary remains are compared. At the Mayo and St. John sites, the re-
mains of peccary are more than twice as abundant as those of deer, where-
as at the Ebrin, Palo Seco, and Qulnam sites, deer are almost three times
as abundant as peccaxy. At the Cedros site also, more deer thaxi peccary

52

remains were found. The Qulnam site revealed a few bones referable to
the white-lipped peccary, Tayassu pecari, as well as remains of the col-
lared peccary. Peccary are often found in forested areas, and parti-
ciilarly along river bottoms, where their feeding habits are similar to
those of pigs (Seton, 1929, quoting Audubon and Bachman). Ideal habitat
of this kind is present at the St. John site. This habitat was also found
at Mayo with the Mayo River providing the river bottom conditions. Deer,
although numerous in densely forested areas, prefer the open forests
found along the south coast.

With the exception of the St. John site, which is a special case,
no significant trends in the numbers of individuals of a given species
at different levels of excavation in the other sites could be discovered.
Moreover, when the faunas from the various Indian middens are compared
on the basis of the total fauna rather than by levels, no striking dif-
ferences are found in their composition. Animals that are found only in
a few sites appear to be incidental, and are represented by only a bone
or two. As has been noted, differences in the faunas can be observed
in the relative abundance of certain forms which were locally numerous
and presumably equally sought after by the Indians.

HISTORIC MA^MAL FAUNA

One of the first lists of Recent msunrnal s of Trinidad was pub-
lished by De Verteuil (l88it^). It included 2k species of large, con-
spicuous land mammals. Ihree of these — Gulo, Viverra vittata
(= Orison vittata), and Cachicamus septeiacinctus (= Dasypus septem-
cinctus) — have since been found not to occur on the island. De Ver-
teuil 's list was followed in I892 by a preliminary list of 2k species
published by Oldfield Thomas. This number of species was enlarged by
the work of J. A. Allen and F. M. Chapman which was based on material
collected in Trinidad by Chapman in I893 and I89U. In their first pub-
lication (1893), these authors listed 37 land mammals, I3 of which had
not been previously reported. New species that were described included
Hectomys palmipes (Holochilus squamipes listed by Thomas is later
synonymized with this), Tylomys couesi, Oryzomys speciosus, Oryzomys
trinitatis , Oryzomys velutinus, Oryzomys brevicauda, Loncheres castaneus,
and Echimys trinitatis. In the second paper (I897) they omitted
Hyrmecophaga Jubata, Choloepus didactylus, Cercoleptes caudi volvulus,
Loncheres castaneous ( = Lcncheres guianae = Echimys armatus) , and
Dicotyles labiatus (= Tayassu pecari) . However, recent evidence has
been obtained to indicate that the last two forms do, in fact, occur on
Trinidad. Allen and Chapman also added the following species: Oryzomys
dellcatus, Akodon frustrator, and Thy 1 amy s carri . The most recent list
of Trinidad mammals is that of Vesey-Fitzgerald (1936), while the Chirop-
tera have been dealt with in a monograph by Goodwin and Greenhall (196I).

53

54

One form included in the Vesey-Fitzgerald list, Akodon fru t rat or, evi-
dently was described from a juvenile specimen of Zygodontomys brevicauda,
and had earlier been synonymized (Gyldenstolpe, 1932). A second species
given, Oryzomys brevicauda, had previously been referred to the genus
Zygodontomys (Gyldenstolpe, 1932).

A list of the Recent land mammals of Trinidad is presented below.
Although by far the largest segment of the mammal fauna of the island
is made up of bats, 58 species having been reported, the Chiroptera are
not included in the following list, since they have been the subject of
a very recent monograph, and are of comparatively little importance
archeologically. The list is essentially that of Vesey-Fitzgerald ex-
cept for the addition of two species, and certain nomenclatural changes.
Most of the nomenclatural changes follow the work of Hershkovitz (l9^7>
19^*8, 1949, 1955, i960) and Cabrera (1957)- The synonymy for the
species conceins only the references to the occurrence of these species
on Trinidad. The list includes obsejrvations that I made during the sum-
mer of 1959.

List of Terrestrial Mammals
Didelphis marsupialis insularis Allen. Bull, Amer. Mus. Nat. Hist.,

XVI, 1902.

Didelphvs marsupialis Thomas, Jour. Trinidad Field Nat. Club, I,

1893.
Didelphis marsupialis Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1893.

55

Dldelphis karklnophaga Allen and Chapman, Bull. Amer. Mus. Nat.
Hist., DC, 1897.

The opctssum or "Manicou" appears to be quite common, particular-
ly around plantations. One male was shot on 3 September 1959 on a cocoa
estate in the North Range, Arlma Valley, Spring Hill Estate. Its
measurements were 7^4- 399- 55 -En U8. Mango, insect and rodent remains
were contained in the stomach. One skeleton from Biche was obtained
from a hunter.
Caluromys philander trinitatis (Thomas)

Dldelphis (Philander) philander Allen and Chapman, Bull. Amer.

Mus. Nat. Hist., V, I893.
Dldelphis (Philander) trinitatis Thcmas, Ann. and Mag. Nat. Hist.
(6) XIII, idok.

"Manicou gros-yeux" is also quite common around plantations.
One male was shot on 3 September 1959 at Spring Hill Estate. The
measurements were 536-322-3^-En 30- His stomach contained fruit and
insect remains.
Manaosa robinsoni chapmani Allen

Didelphis (Mlscoureus) murina Allen and Chapman, Bull. Amer. Mus.

Nat. Hist., V, 1893.
Marmosa murina Allen and Chapman, Bull. Amer. Mus. Nat. Hist., DC,

1897.
, Mannosa chapmani Allen, Bull. Amer. Mus. Nat. Hist., XIII, I9OO.

One immature male was trapped on 3I July 1959 at the base of a
large clump of bsunboo at Pointe-a-Pierre. His measurements were 192-
111-lU.

36

Marmosa carrl (Allen and Chapman)

Thylamys carri Allen and Chapman, Bull. Amer. Mu . Nat. Hist., IX,

1897.
Marmosa carrl, in Trouessart Cat. Mamm. vlv. foss. Suppl. 856.
Alouatta senlculus insularls Elliot

htycetes seniculus (Linn.), Allen and Chapman, Bull. Amer. Mus. Nat.

Hist., IX, 1897.
Alouatta seniculus insularis Elliot, Ann. and Mag. Nat. Hist. Ser.
8, V, 1910.

Hovler monkeys were often heard in the late afternoon or before
rain at Guayaguayare . They were usually in bands of 3 to 10, each band
appearing to have a fixed route through the forest. We observed a small
band, composed of several adults with at least one young of that season j
feeding on Mora flowers at Guayaguayare on 26 August 1959- One young
male of this band was shot. His measurements were 12lU-651-155-En 30-
Hog plum seeds were contained in his stomach and intestine. Two skulls
were found on the trail to La Table in Guayaguayare on 26 August 1959-
Cebus albifrons trinitatis Pusch

Cebu'-. sp., Thomas Trinidad Field Nat. Club I, I893.

Cebus apella, Vesey-Fitzgerald (nee. Linnaeus), Tropical Agr.

(Trinidad), 13(6), 1936.
Cebus albifrons trinitatis Pusch, Zeitschr. fur Sauget, I6, I9UI.
Capuchin monkeys move in large bands, 15 or more. They will come
quite close to investigate a call. One band attracted in such a way
was seen in Guayaguayare forest. One male specimen, formerly a pet.

57

was obtained from the Trinidad Regional Virus Laboratory. Its measure-
ments were 8lO-itOii-123-En 35.
Tamandua longicaudata longlcaudata Wagner

Tamandua tetradactylus (Linn.), Allen and Chapman, Bull. Amer. Mus.

Nat. Hist., V, 1893.
Tamandua longicaudata Wagner, Vesey-Fitzgerald, Tropical Agr.

(Trinidad), 13(6), I936.

One ant eater or "mataperro" was shot at Guayanguayare on 27 Aug-
ust 1959. A partial skeleton was found on the trail to La Table on 26
August 1959.
Cyclopes didactylus didactylus (Linn. )

Cyclothurus didactylus (Linn.), in Allen and Chapman, Bull. Amer.

Mus. Nat. Hist., V, I9IO.
Cyclopes didactylus didactylus (Linn.) in Vesey-Fitzgerald, Ti-opical

Agr. (Trinidad), 13(6), 1936.
Dasypus novemcinctus novemcinctus Linn.

Tatusia novemcinctus (Linn.) in Allen and Chapman, Bull. Amer. Mus.

Nat. Hist., V, 1910.

The armadillo or "tatou" is widely distributed and common. It is
the most common large mammal in the Mora forest of southeastern Trinidad
according to hunters. One male was shot on 31 August 1959 in Guayaguay-
are with measurements of 813-385-91-En 30- Three more individuals were
shot in the same area on 27 August 1959- An immature female was run
down in the cocea at Spring Hill Estate, 13 August 1959, with measure-
ments of 670-3U5-80-En 37.

58

Scluirus granatensls chapman! Allen

Sclurus aetuana hoffmanl Peters, in Allen and Chapman, Bull. Amer.

Mus. Nat. Hist., V, I9IO.
Sclurus chapman 1 Allen, Bull. Araer. Mus. Nat. Hist., XII, I899.
These are fairly common.
Nectomys squamipes palmipes Allen and Chapman

Holochilus squamipes (Brants) in Thomas, Jour. Trinidad Field Nat.

Club I, 1893.
Nectomys palmipes Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1910.

These are commonly found at the edge of ponds or streams. Two
were trapped at the edge of a shallow pond in a logged area and one
was shot along a stream in Guayaguayare forest. The two males were
both in breeding condition and were caught the 29 and 30 August 1959.
They measured 1+37-220-51-En 23, 3U6-177-U5-En 21, respectively. One
pregnant female with three embryos which averaged 27 mm. crown-rump was
caught the 3I August I959 with measurements of 3UO-l83-U5-En 20.
Oryzomys concolor speclosus Allen and Chapman

Oryzomys speciosus Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1910.

Oryzomys trinitatis Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1910.

One pregnant female with two very early embryos was trapped at
Polnte-a-Pierre on 28 July 1959- Its measurements were 2UI-I36-23.

:59

Oryzomys latlcep3 velutinus Allen and Chapman

Oryzomys velutinus Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1910.

Three were trapped "between I3 and I6 August 1959 at Spring Hill
Estate between the cocoa and the uncleared land. There were some gnawed
cocoa pods on the ground. One female was a subadult, one was in breed-
ing condition, and one was pregnant with six embryos of which one was
resorblng. The measurements were l63-75-2i|-En I6, 210-109-27-En I8,
207- injured tail 87-28-En 20, respectively.
Oryzomys (Oligoryzcmys) delicatus Allen and Chapman

Oryzomys delicatus Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

IX, 1897.

Twelve skulls and 52 rami of this species were found in owl
pellet material. This form is rare in collections.
Rhipidomys couesi (Allen and Chapman)

Tylomys couesii Allen and Chapman, Bull. Amer. Mus. Nat. Hist., V,

1893.
Rhipidomys couesi (Allen and Chapman), Bull. Amer. Mus. Nat. Hist.,

IX, 1897.

Zygodontomys brevicauda brevicauda (Allen and Chapman)

Oryzomys brevicauda Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1893.

Akodon frustrator Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

IX, 1897.

bO

Forty-three skulls are represented in the owl pellet material.
This form is commonly trapped.
Akodon urichi Allen and Chapman

Abrothrix caliginosus (Tomes), provisional reference in Allen and

Chapman, Bull. Amer. Mus. Nat. Hist., V, I893.
Akodon urichi Allen and Chapman, Bull. Amer. Mus. Nat. Hist., IX,

1897.
Sigmodon cf . hirsutus

Twenty-one skulls referable to Sigmodon are represented in the
owl pellet material. All the owl pellets studied were collected by the
late F. W. Urich at St. Augustine, to the best of everyone's knowledge.
St. Augustine is located eight miles east of Port-of-3pain and the Gulf
of Paria and 35 miles from the nearest point on the mainland. Both the
numbers of animals represented and the distance from the mainland would
suggest that, if the owl pellets were indeed collected at St. Augustine,
these rats must have been caught on Trinidad and as such represent a new
record. Only minor differences were noted when they were compared with
four skulls referred to Sigmodon hirsutus. Three of these Sigmodon
hirsutus came from Caracas and one from Rancho Grande which is the
closest record of this species to Trinidad. Sigmomys, characterized by
grooved upper incisors, has a range nearest to Trinidad. Hershkovitz
(1955) suggests that probably all species of Sigmodon and Sigmomys are
synonymous with Sigmodon hispidus. This material tends to support the
idea that at least the skulls from Trinidad, the four skulls studied
from Venezuela, and the skulls studied of Sigmodon hispidus from Florida

61

are very close. Some measurements taken on the skulls from Trinidad

are given in Table 7>

Heteromys ancMialus anomalus (Thompson), in Thomas, Trinidad Field Nat.

Club I, 1892.

These are commonly trapped.
Rat t us i^ttus (lilnn. )

Five were trapped in areas under cultivation, inhabited areas,
and in the deep forest. One was from Pointe-a-Pierre, two were frcxn
Spring Hill Estate, and two were from Guayaguayare.
Rat t us norvegicus (Berkenhovit)
Mus mus cuius Waterhouse
Coendu prehensilis (Linn. )

Synetheres prehensilis (Linn.), in Allen and Chapman, Bull. Amer.

Mus. Nat. Hist., V, I893.

A skeleton was obtained from a specimen killed about 10 August
1959 in Gruayaguayaxe and a Jaw of another was found. Another skeleton
was obtained from a hunter from Biche.
Agouti paca (Linn. )

Coelogenys paca (Linn.), in Allen and Chapaan, Bull. Amer. Mus. Nat.

Hist., V, 1893-

Paca or "Lappe" is a delicacy and as a result have become very
rare. It and the peccary are probably the most difficult game animal to
obtain in Trinidad. A skeleton of one was procured from a hunter from
Biche.

TABLE 7
MEASUREMENTS (IN MILLIMETERS) OF SIGMODOM CF. mRSUTUS

Measurement Number Mean Range

Skull length

1

36.6

Zygomatic breadth

3

20.0

19.5-20.6

Interorbltal breadth

18

5.9

5.5- 6.3

Length of palate

16

^•9

6.3- 7.7

Alveolar length of

upper molar row I6 6.3 5.8-6.9

Length of Incisive

foramen I6 7' 5 6.6-8.5

Breadth of rostrum lU 9-3 8.2-10.1

62

63

Dasyprocta aguti (Linn. )

Dasyprocta rubrata Thomas, Ann. and Mag. Nat. Hist. {l)ll, I898.

Dasyprocta agutl (Linn.), in Allen and Chapman, Bull. Amer. Muso
Nat. Histo, V, 1893.

In Guayaguayajre the "guti" Is thought by hunters to be next in
abundance after the armadillo. One was shot at Guayaguayare on 20
August 1959- One vas obtained from Blche.
Echimys armatus castaneous (Allen and Chapman)

Loncheres guianae Ifeomas, Trinidad Field Nat. Club !(?), I892.

Loncheres castaneus Allen and Chapman, Bull. Amer. Mus. Nat. Hist.,

V, 1893.

Echimys annatus Geoff.
Proechimys guyannensia trtnitatis (Allen and Chapman)

Echimys trinitatis Allen and Chapmsua, Bull. Amer. Mus. Nat. Hist.,

V, 1893.

The "Pilori" inhabit the stream banks. One male in breeding
condition was shot in Guayaguayare on 29 August 1959' Its measurements
were U78-212-56-ai 30. One partial skull was found at the edge of the
Aripo River.
Procyon cancrivorus cancrivoinis (Cuvier) in Thomas, Trinidad Field Nat.

Club 1(7), 1892.
Eira barbara trinitatis (Thomas)

Gallctis barbaja (Linn.) in Allen and Chapman, Bull. Amer. Mus. Nat.

Hist., V, 1893.
Tayra barbara trinitatis Thomas, Ann. and Mag. Nat. Hist. (?) V, I9OO.

6h

Lutra enudri* eaudrls (Cuvier)

Lutra insularls Thomas , Ann. and Mag. Nat. Hist. (8) I, I908
Herpestes auropiinctatus auropunctatus
Fells pardalis Linn.

Fells tlgrina Erxl. , Allen and Chapman, Biill. Amer. Mus. Nat.

Hist., IX, 1897
Leopard us pardalis Linn. , in Vesey-Fltzgerald, Tropical Agr.
(Trinidad), 13(6), 1936.
Pecari tajacu (Linn. )

Dicotyles tajacu (Linn.), in Allen and Chapman, Bull. Amer. Mus.
Nat. Hist., V, 1893.
Tayassu Pecari Fischer

Dicotyles labiatus Cuvler, in Allen and Chapman, Bull. Amer. Mus.
Nat. Hist., V, 1893-
Mazama amerlcana trinltatis (Allen)

Carlacus (Coassus) nemorivagus (Cuvieri in Allen and Chapmaji, Bull.

Amer. Mus. Nat. Hist., V, l393- s

Mazama rufa ( Cuvler)

Mazama trinltatis Allen, Bull. Amer. Mus. Nat. Hist., XXXIV, 1915-
Deer are less abundant than Dasyprocta and must be hunted in-
creasingly further away from the population center, Port-of-Spaln. They
have one or two young, dropping them mainly in the summer. One hunter
killed a deer with a well formed fetus in the middle of August. The
skull and leg of one specimen was obtained which was killed during the
spring of 1959 in Siparla Forest Reserve. A hunting party with dogs

65

killed a lactatlng doe with a single corpus luteum ii the ovary at Taba-
quit on 26 June 1959- One buck was shot on 23 August 1959 at Guaya-
guayare that dressed out at 35 pounds. One deer was seen in the Guay-
aguayare Forest.

Zoogeographic Affinities
Trinidad Is part of the savannah region of South America as
defined by Cabrera (l9^). In addition to savannah, however, the
region also includes the foirested coeistal range which is the mountain-
ous extension of the Andes. Trinidad Is Included in this forested sec-
tion, and may be considered as the extreme northeastern end of the
coastal range. This geographic character is reflected in the present
mammal fauna which is composed to a large extent (38 percent) of arboreal
forms. The savannah faunas which existed on the island in the Pleis-
tocene is represented by only one relict form, Oryzomys delicatus, in
the present fauna, and by one other, Odocolleus ^ypmotis, of the pre-
Columbian fauna,

Westermann (1953) iias stated that "The flora and fauna of
Trinidad and Tobeigo have strong affinities to those of the neighboring
South American continent, but quite a number of species common to the
Gulanas are absent here. Their absence is due to natural circumstances
and not to extermination by man. " When the present land mammals of
Trinidad are compared with the mammals of the Guiana Region and Rancho
Grande, Venezuela, as reported by Tate (1939, 19^7), and those of
Northern Colombia as reported by Hershkovitz (l9^7> 19^8, 19^9, i960),
it is found that more than twice as many (28 percent) have affinities

e(>

to the west of Trinidad in Venezuela, Colombia, and the Andes than have
affinities to the south in the Guiana region. Those forms with allies
to the west are: Cebus albifrons, Alouatta senlculus, Sciurus gjrana-
tensisj Oryzomys delicatus, Rhipidanys couesi, Akodon urlchl, Heteromys
anomalus, and Marmosa carri, whereas thoae with allies to the south are
Oryzomys concolor, Lutra enudris, and Caluromys philander. The re-
maining fonnB (59 percent) have relatively extensive ranges. This
aBsociaticm again reflects the geographic ties which Trinidad has had
with Venezuela and its recent ecological relationship with the forested
areas. It also suggests that the Orinoco River constitutes an important
barrier to many widely ranging mammal species.

DISCUSSION

Since a large part of this study is based on animal tones as-
sociated with archeological remains it is appropriate that some of the
problems involved In the use of such materials for zoologicsLl Interpre-
tation be discussed. It is important to understand the nature of the
deposit in which the bones are found, the method of sampling, and the
cultural background of the human population represented. The nature
of the deposit may be ceremonial where anlmsLls are burled as part of a
ritual, but more often animal remains are associated with kitchen middens.
The number of different forms represented in a midden may often be cor-
related with the economy of the aborigines that made the midden. Usu-
ally gathering, hunting, and fishing peoples will depend on a greater
variety of animal foods than agricultural people. Middens are sampled
by the archeologist either randomly by a surface collection that may
be preliminary to an excavation or by an excavation that systematically
samples all parts of the midden. Each level, the smallest unit of the
sample, is of known volume and all that is fovind in each level is kept
separate for analysis. The zooarcheologist bases his analysis on material
sampled in the same meumer.

The archeologist ' 8 goal is to reconstruct the life of the abori-
ginal group he Is studying whereas the zooarcheologist ' s objective is to
reconstruct as far as possible the faunal characteristic of that period.
Such data contribute to the archeologist 's interpretation of the rela-
tionship of the human population to its natural environment. The

67

68

analysis of partlcvaar faunal associations must rest upon a foundation
of accurate identifications. For this purpose as large series of skele-
tons as possible must be obtained and prepared. Usually between 25 per-
cent and 75 percent of the material will be too fragmentary to identify.
Fran the identifiable portion, the relative number of individuals of
each species may be determined by calculating the minimum number of
individuals. This must be daxe to give equal consideration to each form
and not favor those with more bones. With these data a better picture
of the former ecology and various changes in the fauna may be revealed o
Certain changes in the fauna within one site may become apparent when
the materials are analyzed stratlgraphically, or between the fauna of
the site and the more distant past or present fauna of the same area.
Work with the individual species may reveal new forms, certain varia-
tions of recognized forma, or osteological properties of a species or
genus. This material may provide much valuable information when ana-
lysed with Its origin in mind.

In this study of samples of past and present mammalian faunas of
Trinidad an effort was made to determine the composition of the faunas
and to gain some insight into the possible factors that might have
played a role in the formation of the faunas. It is obvious that the
fauna will be composed of such forms as have had geographic access to
the area and whose habitats are represented. The fossil fauna existed
at a time vhea what is now Trinidad was connected to the mainland, and
savannahs covered most of the area. Forests grew at higher elevations
and alwig streams. The known fossil fauna reflects these conditions in

69

that it is composed principally of large herbivores of grassland
habitats. The geological events that resulted in the isolation of
Trinidad as an island also resulted in the rejuvenation of the soils
and thereby the spread of forest vegetation. Isolated in this habitat
is a recent fauna of entirely different nature. Aa many as 38 percent
of these mammals are arboreal and a great many more are forest dwell-
ing forms.

As would be expected, accompanying the changes in the composi-
tion and isolation of the faunas are changes in its affinities and per-
centage of endemism. The fossil fauna was composed largely of species
with very wide ranges. There is no endemism among the fossil species
(with the possible exception of Zygodontomys) since there was no iso-
lation in the uninterrupted expanse of savanneih that spread across
Venezuela and Trinidad. The Recent fauna is composed to a large ex-
tent of widespread forms that originated principally from the west of
Trinidad. Evidently the Orinoco River presented a barrier to the move-
ment of certain species from the south to Trinidad. Due to the gradual
separation of the island from the mainland, species became isolated.
This is reflected in the extent of endemism now found in Trinidad. Of
the 29 land mammals now known from Trinidad, two Marmosa carri, and Ako-
don urichi, are endemic. Twelve endemic subspecies are also recognized.

Since the colonization of Trinidad by Europeans the isolation of
the island fauna has in one sense broken down and as a result of man's
activities the composition of the fauna has been altered. At the
present time the forest forms are decreasing in numbers as the forest
habitat is replaced by agricultural and industrial areas. With the

70

increase in the human population from about 635,81+3 in 1950 to 825,700
in i960 a drastic change has been seen in the abundance of game animals.
This situation is particularly serious because of three principal fac-
tors. The first of these is unemployment resulting from the increase
in population. The second is the good market for wild meats that
stimulates hunting by the people. Finally the game laws, although
reasonably strict, are poorly enforced. Hunters hunt without licenses,
shoot protected animsQ-S, and animals out of season without hesitation.
The clearing of land provides habitats for rodents and particularly
Rat t us and Mus at the expense of the native forms. A few of the native
forms such as Zygodontomys adapted to habitats resulting from cleared
Isuid have been able to compete successfully with the introduced forms.
However, the persecution of game species would appear inevitably to
lead to their extermination if no reservoirs are provided. Perhaps the
most practical resolution would be to set aside sanctuaries that could
be managed and stringently protected for the replenishment of g6une in
other forested areas. The one area which is absolutely protected now
is the United States base at Chagaramus where game is said to be plenti-
ful. As Westerraann (1953^ p. 12) has said, "The relatively small size of
these islands, and their rapidly growing population, make the outlook for
the preservation of nature rather precarious, unless large forest and
nature reserves can be kept in perpetuity. "

SUMMARY AND CONCLUSIONS

Three samples of mammal bones representing different time peri-
ods were studied to see what changes have taken place since the late-
Pleistocene and to determine the cause of these changes.

The first of these samples was deposited more than 34,000 years
ago in a stream edged by gallery forest when Trinidad and Venezuela
were broadly connected. The ma.mwB.T fauna represented was composed of
large herbivorous forms as Cuvieronius, Glyptodon, %lodon^ and Mega-
therium of which all are now extinct. In addition to these large mam-
mals one small rodent referable to the genus Zygodontomys was repre-
sented. Although a species of Zygodontoays now exists on Trinidad it
is not the same species as the fossil form. The fossil Zygodontcxnys
is characterized by its very small size. The ecological situation
indicated by the fossil mammals is that the area's habitat was grass-
land and forest.

The second sample was composed of bones excavated at six Indian
middens. The mammals represented comprise almost all the larger land
mammals native to Trinidad. Those larger than the cricetid rodent
Rhipidomys cotesi that were not found are Cyclopes didactylus, Cebus al-
bifrons, and Eira barbara, although the latter has previously been re-
ported from a midden. Species that do not now occur on the island are
Odocoileus gymnotis, which probably constitutes a relict; Tayassu
pecari, which is probably now extinct on the island but did appear in
the early lists of Trinidad mammals; and Tapirus represented by one
tooth which may have been a trade object.

71

72

The principal difference between sites was not found to "be the
presence or absence of certain animals, but rather the percentage of
comparable and abundant forms. The armadillo was found to be mojre
abundant at sites where the soils are more friable. At the three
sites situated on the south coast in open forest, deer and agouti were
more abundant than peccary and paca, respectively. At the site in palm
forest surrounded by heavy forest agouti and paca were almost equally
abundant but deer were more abundant than peccary. The two sites
located in heavy forest had far more paca and peccary than agouti and
deer, respectively. These differences in the percentage of various
forms reveal more subtle differences in the ecology of an area than
are revealed by the presence or absence of an incidental form.

The last sample is composed of specimens that were collected by
us, and by mammal remains from owl pellets, and includes a compilation
of published records. !nie only unreported form found is a rodent refer-
able to Sigmodon cf . hirsutus. This record is based on 21 skulls from
owl pellets collected originally by F.W. Urich. The Recent fauna is
composed to a large extent of forest forms that are allied principally
to species from the west of Trinidad or with widespread ranges. These
affinities with the west reflect the past land connection of Trinidad
and Venezuela and the barrier of the Orinoco River to the dispersal of
certain forms. Today with the removal of forests by man many of the
native wAinmRl populations are dwindling.

LITERATURE CITED

Allen, J. A. and F. M. Chapman, l893' On a collection of mammals from
the Island of Trinidad, with descriptions of new species. Bull.
Amer. Mus. Nat. Hist., 5(13) :203-23l+.

. 1897 • On a second collection of msunmals from the Island of

Trinidad, with description of new species, and a note on some
mammals frcsn the Island of Dominica, W. I. Bull. Amer. Mus. Nat.
Hist., 9(2):13-30.

Beard, J. S. 19k6. The natural vegetation of Trinidad. Oxford Fores-
try Memoirs No. 20, pp. 152.

Blair, K. G. 192?. Remains of insects from oil sand in Trinidad. Trans.
Entom. Soc. London. 137-142.

Bullbrook, J. A. 1953' On the excavation of a shall mound at Palo Seco,
Trinidad, B. W. I. Yale Univ. Publ. Anthr. No. 50:5-11^.

Burt, W. H. 1961. A fauna from an Indiem site near Redington, Arizona.
J. Mamm. U2(l): II5-II6.

Cabrera, A. 1957- Catalogo de los mamiferos de America del Sur. Museo
Argentine de Ciencias Naturales "Bernardina Rivadavia."
4(l):1.307.

Cabrera, A. and J. Yepes. 19UO. Mamiferos Sud: Americanos. Ceunpania
Argentina de Edit'ores, Argentina, pp. 370.

Cruxent, J. M. and I. Flouse. 1959' An archeological chronology of

Venezuela. Vol. 1 and 2:1-223. Social Science Monograph 6 of
Pan American Union.

Einillani, C. 1955. Pleistocene Temperatures. Jour, of Geo., 63(6):
538-557.

Goodwin, G. G. and A. M. Greenhall. 196I. A review of the bats of

Trinidad and Tobago. Descriptions. Rabies infection, and ecology.
Bull. Amer. Mus. Nat. Hist., 122(3) : 191-301.

Greenhall, A. M. I956. Is the Yapok or water opossum found in Trinidad?
Jour. Trinidad Field Naturalist's Club, p. 27.

Gyldenstolpe, N. 1932. A manual of Neotropical Sigmodont rodents.

Kungl. Svennka Vetenskapsakademiens Handllngar. ll(3) :l-l6U.

73

7U

Her3hkovltz, P. 19^7- Mammals of Northern Colombia preliminary report

No. 1: squirrels (Sciuridae). Proc. U. S. Nat. Mus. 9T(3208)l-ii6.

. 19U8. Mammals of Northern Colombia preliminary report No. 2:
Spiny rats (Bchimyidae), with supplemental notes on related
forms. Proc. U. S. Nat. Mus. 97(321^+) :125-li^.

. 19U8. Mammals of Northern Colombia preliminary report No. 3:

water rats (genus Hectomys) . with supplemental notes on related
forms. Proc. U. S. Nat. Mus. 98(3221) : 49- 56.

. 1949. Mammals of Northern Colombia preliminary report No. k:

monkeys (Primates), with taxonomic revision of some forms. Proc.
U. S. Nat. Mus. 98(3232): 323-427.

. 1955. South American marsh rats genus Holochilus with a

summary of aigmodont rodents. Fieldiana: Zool. 37 ••639-673.

. i960. Mammals of Northern Colombia preliminary report No. 8:

Arboreal rice rats, a systematic revision of the subgenus Oecomys,
genus Qryzomys. Proc. U. S. Nat. Mus. 110(3^20) : 513-568.

Koldewijn, B. W. 1958. Sediments of the Paria-Trinidad shelf. Mouton
and Co. ' S-Gravenhage , the Hague, pp. 109 .

Nota, D. J. G. 1958. Sediments of the western Guiana shelf. Mededelin-
gen van de Landbouwhoge school Te Wageningen. Nederland, pp. 98.

Osbom, H. F. I936. Proboscidea. A monograph of the discovery, evolu-
tion, migration and extinction of the mastodonts and elephants
of the world. Amer. l-lus. Nat, Hist., l:l-802.

Rouse, I. 1953. Appendix B: Indian sites in Trinidad, in Bullbrook:

Excavation at Palo Seco, Trinidad. Yale Univ. Publ. Anthr. Ho.
50:94-111.

Schaub, S. 1935. Saugetierfunde aus Venezuela and Trinidad. Abh.
Schweizerischen Palaeontologischen Gesellschaft . 55:1-21.

Seton, E, T. 1929. Lives of game animals. Vol, 3, part 2. Doubleday,
Doran and Co., Inc., New York, pp. 78O.

Simpson, G. G. and C. de Paula Couto, 1957- The mastodonts of Brazil,
Bull. Amer. Mas. Mat. Hist. 112(2): 125-190.

Tctte, G. H. H. 1939. The mammals of the Guiana region. Bull. Amer.
Mu8. Nat. Hist. 76(5):151-229.

75

• 19^1' A list of the mammals collected at Rancho Grande,

In a montane cloud forest of northern Venezuela. Zoologica
32(7):65-66.

Taylor, W. W. (ed,) 195?. The identification of non -art if actual

archaeological materials. Nat. Acad. Sci.-Hat. Res. Council.
Publ. 565:l-6U.

Thomas, 0. I893. A preliminary list of the mammale of Trinidad.
Jour. Trinidad Field Naturalists' Club 1(7) :158-l69.

Verteuil, L. A. A. de. 1384.- Trinidad: its geography, natural re-
sources, administration, present condition, and prospects.
Cassel and Co., Ltd. London, pp. kQk.

Vesey-Fitzgerald, D. 1936. Trinidad mammals. Tropical Agriculture
13(6):l6l-l65.

Westermann, J. H. 1953* Nature preservation in the Caribbean. Publ.
of the Found, for scientific research in Surinam and the
Netherlands Antilles. No. 9, pp. IO6.

BIOGRAPHICAL SKETCH

Anne Elizabeth Schwarz was born in Cambridge, Massachusetts,
March 5, 1932. She attended the Wlnsor School in Boston, Massachusetts,
and was graduated in 1951- In 1955 she received her Bachelor of Arts
degree from Mount Holyoke College in South Hadley, Massachusetts, and
in 1957 she received her Master of Science degree from the University
of Florida.

During her graduate studies she has held graduate assistantships
in the Department of Biology and Florida State Museum. In the summer
of 1959 she was awarded a National Science Foundation Summer Fellowship
for Graduate Teaching Assistants « She is presently employed as a re-
search associate of the Florida State Museum on a zooarcheological
project supported by the National Science Foundation, Grant 17948.

She is a member of Sigma Xi and Phi Sigma societies and the
American Society of Mammalogists.

On April l8, 1957^ she was married to James E. Wing, Jr. and
has one child, Mary Elizabeth, born September 12, I96I.

76

This dlBsertation was prepared under the direction of the
chairman of the candidate's supervisory cocmittee and has been ap-
proved by all members of the committee. It was submitted to the dean
of the College of Arts and Sciences and to the Graduate Council and
was approved as partial fulfillment of the requirements for the degree
of Doctor of Philosophy.

February I, 1962

Dean^ Graduate School

SUPERVISORY CCMUTTEE:

Chairman

/-.. ^/9...

^

7

SiD<^\AA£>)AC^

7

A.L hu }i

d~^( c

A J ws