l’fiESJS

lllllllllllllllllll‘llllllIllllllllllllllllllllllllll

301571 4227

This is to certify that the

thesis entitled

MORPHOLOGICAL VARIATION IN THE SPECIES Caiman crocodilus:
WITH CONSERVATION AND LEGAL IMPLICATIONS.

 

presented by

Sima S.Pandya

has been accepted towards fulfillment
of the requirements for

.MASIEL—degree in _SCIENCE_

 

Major profc sor

Date ‘l~l'3-'Cl7r

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY

Mlchlgan State
University

 

 

 

5—-

PLACE IN RETURN BOX to remove this checkout 1mm your rocord.
TO AVOID FINES return on or before data duo.

DATE DUE DATE DUE DATE DUE

       

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU loAn Afflrmatlvo Action/Emil Opportunity Institulon
Wilma-9.1

 

 

MORPHOLOGICAL VARIATION IN THE SPECIES Caiman crocodilus:
WITH CONSERVATION AND LEGAL IMPLICATIONS

By

Sima S. Pandya

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Fisheries and Wildlife

1 997

ABSTRACT

MORPHOLOGICAL VARIATION IN THE SPECIES Caiman crocodilus:
WITH CONSERVATION AND LEGAL IMPLICATIONS

By

Sima S. Pandya

Five proposed subspecies of Caiman crocodilus occur from southern Mexico to northern
Argentina. As with all modern species within the Crocodylia, overall morphology within the genus has
been highly conserved. This has made identification at the subspecific level extremely difficult. To address

subspecific identification within C. crocodilus twelve morphological characteristics were examined from

 

379 preserved museum specimens collected across the entire range of the species. I investigated two
competing hypotheses to explain the distribution of morphological variation using linear discriminant
function analysis. Two groups were discriminated based on the major river systems in South America. Of
the twelve variables, ratio of head width and snout length (HW/SL), mandible pattern (MP), nuchal cluster
rank (NCr), dorsal scale rows (DSR), total keeled and unkeeled rows (T KUR), ventral scale rows (V SR),
and ventral scale rows at mid-body (V SRM) had significant univariate t;te_st values (p < 0.05) for

discriminating between the two subspecies.

Copyright
SIMA SHIRISH PANDYA
l 997

To my mother and father, both of whom were great inspirations to me and who, if still living today,
would be truly proud.

ACKNOWLEDGMENTS

This study was made possible only through the commitment of the Department of the Interior and
the US. Fish & Wildlife Forensics Laboratory in Ashland, Oregon. I thank the director, Kenneth Goddard
for authorizing financial support to obtain specimens from US. museums and to travel to the University of
Florida for additional research. Many thanks go out to Dr. Stephen Busack for overseeing a great research
opportunity. I received assistance from the administration, illustration, and photographic departments at the
forensic laboratory. Figures 9- 19 were provided for one time use by the US. Fish & Wildlife Service.
Figures 11-13 were produced by Maud Kernan-Lonergan and Figure 18 was produced by Grace Hensen.
Figures 10, 14-17, and 19 were enhanced by Grace Hensen and Stephen D. Busack and were used by verbal
permission of F. Wayne King to Stephen D. Busack.

I would like to thank my committee members for guidance and editing advice. My committee
members were: Dr. Stephen D. Busack, Dr. John Epifanio, Dr. William Taylor, and Dr. Donald Straney.
Finally, I would like to thank Marcia Scott, Lori Letterhos, Ila P. Van Milligan, Julie Weeder, Ann Krause,
and Laurel Ramseyer, friends and family members for reading this document in its many stages over and

over again. They are just as happy as I am to see its completion.

TABLE OF CONTENTS

 

LIST OF TABLES ........................................................................................................................ vii
LIST OF FIGURES ...................................................................................................................... viii
BACKGROUND .............................................................................................................................. l
Ecology ............................................................................................................................. 1
History of Trade ................................................................................................................ 6
Evolution ........................................................................................................................... 8
MORPHOLOGICAL STUDY of Caiman crocodilus .................................................................... 22
Introduction ...................................................................................................................... 22
Methods and Materials -- .............................................................. 24
Results ............................................................................................................................. 26
Discussion ....................................................................................................................... 28
Appendix A ................................................................................................................................... 49
Appendix B .................................................................................................................................... 54
Appendix C .................................................................................................................................... 55
List of References .......................................................................................................................... 56

vi

LIST OF TABLES

Table 1 - Cross-validation summary (Andes Mountain Range) using linear discriminant
function analysis. ......................................................................................................... 44

2 - Cross-validation summary (by river systems) using linear discriminant function
analysis. ............................... .. ......... 45

 

 

3 - Results of the t-test with river data showing the mean, range, SD, SE, t, DF, and
significance (p < 0.05) of the twelve variables used in the analysis. Group 1:
=266: Group 2: N=113. . . .................................................................. 46

 

4 - Reduced model cross-validation summary by river system using linear discriminant
function analysis. Only variables mandible pattern (MP), dorsal scale rows (DSR),
total keeled and unkeeled lateral rows (TKUR), ventral scale rows (V SR), and
ventral scale rows at mid-body (V SRM) were used in the analysis. . 47

 

5 - Keeled vs. unkeeled lateral scale row proportions. Each specimen was assigned to
the corresponding number for analysis (KLr). ......................... - - . 54

 

6 - Nuchal cluster patterns recorded. Each specimen was assigned the corresponding
number for analysis (NCr). .......................................................................................... 55

vii

LIST OF FIGURES

Figurel - Annual trade in Caiman crocodilus skins by region from 1983-1993. ........................ l6

2 - Hypothesized evolutionary path of the family Crocodylidae and the sub-family
Crocodylinae showing approximate time of geologic appearance (Sill, 1969) G,
Goniopholidae; H, Hylaeochampisidae; C, Crocodilinae; T, Tomistominae. ............. 17

3 - Hypothesized evolutionary path of the sub-family Alligatorinae and the genus
Caiman showing approximate time of geologic appearance. C, Crocodilinae; A,
Alligatorinae; c, Caiman. ............................................................................................. 18

 

4 - Plate tectonic map of South America illustrating the position of the lithospheric
plates and trenches responsible for the formation of the Andes Mountains. ................ 19

5 - Map of South America illustrating the individual mountain ranges comprising the
Andes Mountains. ........................................................................................................ 20

6 - Enlarged map of Bolivia and surrounding countries illustrating the river systems that
were redirected into the Amazon River as a result of the uplift of the Central Andes. 21

7 - Map of South America illustrating the Andes Mountain ranges used to partition
specimens into Group 1 and Group 2 for testing hypothesis 1. .................................... 31

8 - Map of South America illustrating major river systems sampled for testing
hypothesis 11. River systems within Group 1 are illustrated in light gray, Group 2
systems are illustrated in dark gray. ............................................................................. 32

9 - Collection localities of the 379 Caim_an crocodilus specimens analyzed by
discriminant function analysis. Symbols may represent multiple specimens. .............. 33

10 - Head width (HW) and snout length (SL). Measurements expressed as a proportion

(HW:SL) for analysis. ................................................................................................. 34
11 - Character states 1, 2, and 3 defined within mandible pattern (MP) and lateral

pattern (LP). ................................................................................................................ 35
12 - Proportion of keeled to unkeeled lateral scale rows (KLr). ......................................... 36
13 - Length and width of the largest lateral scales. ............................................................. 37

14 - Nuchal cluster (NCr) patterns identified. Thirty-nine different character states were
identified. ..................................................................................................................... 38

15 - Number of dorsal scale rows (DSR); counted from posterior edge of the nuchal
cluster to the last scale row parallel with the rear margin of the hind legs. .................. 39
viii

16 - Number of ventral scale rows (V SR); counted from first horizontal row following
the ventral crest to the last complete row anterior to the anal opening. ....................... 40

17 - Number of transverse ventral scale rows at mid—body (VSRM). ................................. 41

18 - Number of double and single keeled scale rows on the dorsal portion of the tail
posterior to the anal opening (SRTAIL). ..................................................................... 42

19 - Number of tail inclusions (Tlr) counted on the ventral side of the tail between the
anal opening and the seventh scale row. ................................................... 43

 

20 - Distribution of mandible patterns within the major river systems of Central and
South America. Group 1 represents specimens from Central America, the Orinoco
river system, and the Amazon river system. Group 2 represents the Rivers Parana,
Paraguay, and Uruguay, and associated drainage systems. .......................................... 48

BACKGROUND

In Darwin’s Origin of Swie , he drafted the foundation for a branch of study known as
evolutionary systematics. Scientists have been investigating the relationship between taxonomic groups of
organisms based on genealogy and the amount and nature of their evolutionary change ever since. In the
1930’s and 1940’s the focus of systematics shifted from higher taxa (macrotaxonomy) to the level of species
and populations (microtaxonomy). The goal of these scientists was to classify organisms according to
similar characteristics in order to undertake comparative studies and formulate generalizations about the
organisms. Just as Hennig (1969) warned; these comparisons of taxonomic groups in evolutionary studies
can only be valid if the initial groupings are correctly defined. It is for this reason that the discipline of
classification is considered a prerequisite for sound biological work (Mayr, 1974).

The objectives of my systematic study on Caiman crocodilus are l) to determine if differentiation
at the subspecific level can be supported based on a series of morphological characters and; 2) to determine
the basis for the differences between the groups if distinct subspecific differences exist. To address these

objectives 1 evaluated available information concerning C. crocodilus ecology, trade, and evolution

ECOLOGY
Diet
Magnusson et al. (1987) observed that the diet of C. crocodilus in the Amazon shifted as the snout-
vent length (SVL) of the individual increased. The author concluded that C. crocodilus eats mainly
terrestrial invertebrates when smaller and fish and mollusks when larger in size. In this study, the stomach
contents of 105 specimens were examined for terrestrial vertebrates and invertebrates, fish, shrimp, crabs,
and mollusks. Juvenile caiman primarily consumed terrestrial invertebrates along with a small percentage

of shrimp. Terrestrial invertebrate consumption declined from over ten per individual in the smaller size

1

classes (SVL<40cm) to less than three invertebrates per individual in the larger size classes (SVL>40¢m).
Adult caiman consumed large numbers of fish and mollusks. Terrestrial vertebrates such as snakes, lizards,
mammals, and birds were found in the stomachs of all size classes in low frequency (Magnusson et al.,
1987).

During a 4.5 year study in the central Venezuela llanos, smaller C. crocodilus consumed a greater
number of invertebrates but depended on fish as a major food source when adult. Seasonal shifts in the
consumption of all food items reflected its availability. Caiman preyed upon greater numbers of mammals
and snails during the late dry season and greater quantities of crabs during the wet season. The increase in
fish consumption during the early dry season was most likely due to the greater concentration of caiman and
fish in permanent water holes (Thorbjarnarson, 1993a). Beetles, frogs, fish, crabs, and shrimp were also
found to be common food sources for C. crocodilus in the savanna lagoon system in Venezuelan Guyana
(Gorzula, 1978).

Fishing Behavior

Schaller and Crawshaw (1982) investigated the fishing behavior of the Paraguayan Caiman (Q,
crocodilus yacare) in the Pantanal of Brazil. Throughout the study period, the caiman exhibited two main
variations of sit-and—wait predation. Both fishing techniques required the animal to lie motionless on the
surface of the water or on the bottom of a shallow pond. When motion by a fish was sensed, the caiman
lunged forward or snapped sideways in an attempt to capture the fish in its mouth. In the first technique,
two caiman were observed positioned across a stream perpendicular to the shore. They were able to divert
the normal river flow towards their waiting jaws and act as a weir. In the second technique, caiman floated
on the surface of the water parallel to the shore. By slowly moving sideways towards land, the caiman
herded the fish into a smaller area. Then, with a slight arching of its body toward the shore, the fish were
trapped where the caiman could catch them more easily (Schaller and Crawshaw, 1982).

Thorbjarnarson (1993b) conducted an intensive study of caiman ecology and behavior in the
central Venezuelan llanos. Caiman were observed snapping at prey from stationary positions, herding fish
against the shore, and acting as a weir across narrow rivers or canals. Peak fishing activity occurred early in

the dry season and coincided with reduced water and increased fish concentration. Fishing behavior

declined during the late dry season. Throughout the study, caiman were observed using a variety of foraging
techniques to capture prey (Thorbjarnarson, 1993b).

Individual size of caiman studied was the leading explanation for the consumption of different prey
items. In general, larger caiman primarily ate fish while smaller caiman foraged on terrestrial invertebrates
and shrimp. Frogs, crabs, and other food items were consumed by all size classes depending on availability
in various habitats. Therefore, no major differences in habitat selection, feeding or breeding behavior have
been reported among populations of caiman that could account for the subspecific differences.

Habitat Selection

In the savanna lagoons of Venezuelan Guayana, C. c. crocodilus were observed near the edges of
temporary lagoons during the wet season (July to December) and crowded in the center of permanent
lagoons during the dry season (January to June). Sub-adults and adults migrated from these two habitats
during the wet season but appeared to return to established territories. Vegetation was used during the dry
season as additional protection (Gorzula, 1978). Additional studies on habitat selection by caiman in
savanna areas have been completed by Alvarez del Tom (1974), Rivero Blanco (1974), Ayarzaguena
(1983), Stanton and Dixon (1975, 1977), and Magnusson (1985).

Ouboter and Nanhoe ( 1988) investigated the habitat selection of C. crocodilus in the swamp and
swamp-forest habitat of northern Suriname. Observations were conducted in the Coesewijne river area
from 1982 to 1984. Caiman were counted at night, captured and marked, and equipped with radio-
transmitters in order to track their movement. Habitat selection was determined by the density of caiman
measured in each area type. Many caiman were positioned on mudbanks at the mouth tributary of the rivers
and appeared to prefer shallow water near gradually sloping shore. Higher caiman densities were recorded
near aquatic vegetation and sections of the river bordered by swamps. Swamps and aquatic vegetation may
provide additional protection for the animals (Ouboter and Nanhoe, 1988).

Herron (1994) investigated habitat selection of C. crocodilus in Cocha Cashu lake, located in the
Amazon drainage basin of Peru. Water depth, temperature, and vegetation types were recorded across the
lake and linear regression was used to relate these factors to caiman density. Habitat selection depended

upon the body size of the individual and vegetation cover. Smaller caiman were found in areas with more

shallow slopes while larger individuals inhabited areas with steeper sloped bottoms. Vegetation provided
additional protection for all sizes of caiman from predators such as jaguar and giant otters (Herron, 1994).

Caiman are generalist feeders and appear adaptive to new habitats. They inhabit every major type
of wetland in the low altitude neotropics. Caiman have replaced the Orinoco crocodile in many rivers and
streams of the Venezuelan llanos as the crocodile populations were decimated by the demand for hides.
Caiman even have been reported living in roadside ditches and cattle watering troughs (Thorbjarnarson,
1991).

Breeding Behavior

Caiman breeding ecology has been well studied in the llanos of Venezuela and Colombia, a
savanna located in the northern and western region of the Orinoco river drainage. The llanos is
characterized by two well-defined seasons; a wet season from May to November and a dry season from
December to April. During the dry season, C. crocodilus are forced to conjugate in permanent water holes
and this results in an increase in stress and mortality in many age classes. Caiman dispersed again to
temporary water following the first significant rainfall in May. Males secured breeding territories and
mated from June to July. Since widespread flooding occurred in the area from June to October annually,
females began building nests (late July) and laying eggs (from the middle to late August) after the initial
heavy rainfall. Nests were constructed from live and dead vegetation on elevated areas to minimize
flooding. Twenty to thirty eggs were laid in the nest cavity to incubate for 70 to 90 days, depending on
temperature. The females then opened the nest, allowing the hatchlings to escape, and led them to
permanent water before the upcoming dry season. Dry season conditions such as crowding and dehydration
led to high mortality of the young caiman during the first few years (Thorbjarnarson, 1991).

Cintra (1988) studied nesting behavior of C. crocodilus yacare in the Pantanal of Brazil over a
period of three breeding seasons. In the Pantanal, Paraguayan caiman mate during the dry season (August
to November). Females build nests and lay eggs beginning in January (wet season) and incubation of the
eggs continues for 100 days (Cintra, 1988). Crawshaw and Schaller (1980) observed that females lay eggs
after the first peak in water level, suggesting that rainfall may trigger this behavior. Nests consisted of

twigs, leaves, and bark piled into a small mound and increased the ambient temperature surrounding the

eggs to more than 29°C. A circular chamber in the center of the nests held from 18 to 38 eggs incubated at
30°C-35°C for approximately 70 days. Nests were located near a water source (approximately three meters)
and females were observed guarding the nests. Flooding of nests and predation by monkeys, foxes,
raccoons, jaguars, opossums, armadillos, and lizards accounted for approximately 48.5% and 42% mortality
for the 1984 and 1985 breeding seasons, respectively. The hatching period ran from March to April with
the peak activity occurring from March 30 to April 10 (Cintra, 1988).

In Costa Rica, C. crocodilus built their nests in dense forest adjacent to rivers or canals rather than
in open grasslands that may flood quickly. Mound nests were positioned against tree trunks on thick vines
at the highest level of the river banks. In Costa Rica, caiman nested in the early wet season (mid to late
July) and began to lay eggs between early September and October. Wet season nesting increased the
availability of resources to the hatchlings. Internal nest temperature ranged from 305°C - 328°C, and eggs
incubated for about 80 days. During the 1988 and 1989 breeding seasons, 13 of 32 nests found were
destroyed by flooding and four additional nests had been subjected to some degree of predation. Females
were frequently observed guarding the nests and displayed defensive behaviors to ward off danger
(Allsteadt, 1994).

A small study in the Mapurite river system of the Venezuelan Guayana observed five nests over a
period of three years. Hatching in these nests occurred between October and November. An additional nest
made out of leaves and twigs was examined and found to contain 24 eggs incubated at 31.4°C. The young
caiman remained together for 18 months and were defended by various adult caiman that responded to the
distress calls (Gorzula, 1978).

Maternal Care

C. crocodilus is one of many crocodilian species that exhibit maternal care. A clutch of 22 young
C. crocodilus yacare were observed with their mother for six months after hatching. The mother remained
in a forest-surrounded bay in an area where maximum water depth was approximately 40 cm. The water in
the 70 X 12 meter study area was clear and supported many small aquatic plant species. During the day
(0530 hrs to 1700 hrs) the hatchlings remained within one meter from the female adult and were assumed to

be resting. After 1700 hrs the family moved up the inlet and foraged for insects at the water’s surface and

to a depth of 15 cm. Hatchlings ventured up to 100 m from their mother during the night, but returned to
their resting site between 0530 hrs and 0550 hrs. The mother displayed attack behavior against any intruder
attempting to approach the hatchling area. Maternal care is assumed to prevent predation and increase the

survivability of hatchlings (Cintra, 1989).

HISTORY OF TRADE
The Crocodilian Skin Trade

Historically, crocodiles have been used for food, medicine, and religious worship. The skin trade
began as a result of European expansion into previously remote areas after the Industrial Revolution.
Commercial use of the American Alligator (Alligator mississippiensis) began in the late 1700's and
increased during the Civil war (1861-1865). Alligator leather was manufactured into boots, travel bags,
belts, and card cases. Some trading firms sold up to 60,000 skins to manufacturers in New York annually.
By 1900, the demand for Alligator products was too great to be met by US. supplies and the industry began
exploring additional sources in Mexico and South America. Large tanneries, open in Europe, increased the
demand for skins. Caiman skins were still considered inferior to Alligator and Crocodile skins at this time
and were not commercially harvested. As the demand continued to grow, the industry secured a new supply
of crocodile leather from African and Asian sources. Eventually, the demand for American Alligator skins
pushed the population number near extinction and, in 1943, only 6,800 skins were traded from Florida. In
1943, Florida passed legislation to prohibit alligator hunting during breeding season and taking individuals
smaller than 1.2 meters. Commercial trade in Alligator skins from Florida increased again to 25,000 in 1947
as the population started to rise. By 1969, the United States placed the American Alligator on the
Endangered Species List (Ross, 1989).

The Caiman Skin Trade

Commercial hunting for caimans began in South America in the late 1950's. The black caiman
(Melanosuchus niger) was the most sought after species because its skin lacked osteoderms (bony plates in
the scales) and more closely resembled Alligator and Crocodile skin than the other caiman species. Q,

crocodilus hunting in the Amazon began in Venezuela and rapidly spread over much of the continent.

During the 1950's, approximately 5-10 million caiman hides were harvested annually; in the 1960’s the
annual harvest was 6-8 million. In the 1980’s, skins from C. crocodilus comprised over 60% of the world
wide skin market (Brazaitis, 1989).

The decimation of crocodilian populations during the 1960’s forced many tanneries and brokers
out of business. In response, countries began to pass legislation to protect the remaining populations. In
1960 and 1968, Colombia and Bolivia, respectively, banned export of the black caiman and many or all
subspecies of C. crocodilus (Ross, 1989). In Venezuela, caiman harvest greatly reduced population
numbers and a total ban on commercial trade was established in 1972. Populations quickly rebounded and
caiman became locally abundant by the 1980’s (Thorbjarnarson, 1991). In 1978 and 1979, unregulated
caiman hunting in Suriname quickly reduced that population, but the hunting laws enacted in the 1980’s
proved hard to enforce (Ouboter and Nanhoe, 1989). Commercial trade of caiman skins in Brazil and
Paraguay in the late 1970’s and early 1980’s was banned in response to widespread illegal harvesting that
was decimating populations across South America. Effective enforcement of the legislation was a major
problem for many South American countries and as a result the illegal trade increased to sustain the supply
of raw skins to the remaining tanners (Ross, 1989).

The annual trade of C. crocodilus skins from the Convention on International Trades in
Endangered Species (CITES) annual reports (1983 to 1991) is summarized in Figure 1. World trade in Q
crocodilus skins from 1980-1984 totaled over 2,380,566 (V illalba-Macias, 1986). Trade in caiman skins
from countries in its natural range totaled over 239,802 skins in 1987 and 178,510 skins in 1988 (Luxmoore,
1990; Luxmoore and Collins, 1994; Collins and Luxmoore, 1996).

Current Caiman crocodilus Skin Trade (1992-1993)

The current trade in C. crocodilus skins is vastly different from what it was between 1960 and
1980. While problems still exist, the trade is now largely regulated and sustainable. Net export of hides,
primarily from Colombia and Venezuela, reached 410,375 in 1992 and 648,847 in 1993. Brazilian
production is also increasing as 75 registered ranches work to become operational. Historically, caiman

skin harvest has been three to four times larger than the “classic” skin trade (skins that lack osteoderms such

as Alligator and many species of Crocodile). This ratio is predicted to increase by the end of the century.
Trade in C. crocodilus is now greater than one million skins per year (Collins and Luxmoore, 1996).

Both the South American economy and the species have benefited from the shift to legal and
regulated trade in the 1990’s. Today, the caiman skin industry is at the same level of organization that the
classic skin trade was only 10 years ago. Management programs are being developed in many countries and
uniform conservation and trade regulations are being implemented throughout the natural range. Public
awareness of the benefits of sustainable use is increasing to a level that will aid both conservation and
industry. The major exporting countries (Colombia, Venezuela, Nicaragua, and Brazil) and the major
importing countries (France, Italy, Switzerland, Japan, and the United States) are leading the way toward

improving the management of this natural resource (Collins and Luxmoore, 1996).

EVOLUTION
Crocodilian Evolution And Distribution

Reptiles first appeared, in the late Paleozoic era, three hundred million years ago. The oldest
crocodilian fossils found in Europe, South America, and Africa date back nearly 230 million years.
Crocodilians are the only living representatives from the Archosaurian complex, “The Ruling Reptiles,”
which includes early crocodilians, dinosaurs, flying reptiles [Pterosaurs] and an assortment of early
Mesozoic forms [thecodontians]. The Mesozoic era (245-265 mybp) is called the “Age of Reptiles.”
During this era reptiles were dominant and extremely diverse (Ross, 1989).

The evolution of all crocodilians can be traced through five suborders: Archeosuchia, Protosuchia,
Sebecosuchia, Mesosuchia, and Eusuchia. The suborder Archeosuchia is composed of two families, the
Proterocharnpsidae and the Notochampsidae. Proterochampsa barrionuevoi is the first known fossil
crocodile, and it is from the Upper Middle Triassic period of what is now western Argentina. Notochampsa
and Egthrochampsa are known from the Upper Triassic of South Africa. The genus Proterochampsa has
many features linking it to early thecodontians and it shares a flattened head, fused external nares, and long
snout with modern crocodilians. This genus is believed to an important link in the evolution from early

crocodilians (Sill, 1968).

The suborder Protosuchia flourished in the late Triassic and early Jurassic. It includes one genus,
mtosuchu , which was a deep-skulled, short-snouted crocodilian that probably originated in South
America during the middle Triassic (Sill, 1968). The Protosuchia were small (1 meter) and lizard-like in
appearance. Most Protosuchia were terrestrial and all possessed bony armor along the back and tail. This
characteristic has been conserved by modern crocodilians (Ross, 1989). Bernard Krebs, a Swiss
paleontologist, linked crocodilians to ancestral thecodontians because of their shared unique ankle joint
structure permitting the foot to swivel or twist sharply during a form of locomotion called the “high walk."
The peg-and-socket structure of the crocodilian ankle allows the animal to move quickly over land. During
the high walk, the body is lifted completely off the ground and the limbs are bent outward from the body
(Ross, 1989).

The suborder Sebecosuchia consists of two families, the Baurusuchidae and the Sebecidae. The
Baurusuchidae is found in the late Cretaceous of Brazil and the Sebecidae probably evolved from it. The
Sebecidae most likely originated in the middle or late Jurassic from primitive members of the suborder
Mesosuchia. Evidence suggests that the Sebecosuchia originated in the western hemisphere, expanded
worldwide by late Cretaceous, and disappeared by the end of the Miocene. The Sebecosuchia were
characterized by a deep, narrow skulls and may even have descended from the Protosuchia (Sill, 1968).

The suborder Mesosuchia is the next major line of crocodilian evolution from the Protosuchia.
Fossil records from the early and middle Jurassic age are unavailable and little is known of this lineage.
Only two of the eight families from the suborder Mesosuchia are represented by fossil remains. Both the
families, Metriorhynchidae and Teleosauridae, were specialized forms adapted to marine life and
distributed around the world (Sill, 1968).

The remaining families of Mesosuchia, found in late Jurassic and Cretaceous deposits, already
possessed many characteristics common to modern crocodilians (Sill, 1968). Adaptation to aquatic life and
evolution from herbivore to piscivore led to a trend toward snout elongation in crocodilians. Second, an
increase in the size of the bony palate has provided crocodilians extra strength and protection while
capturing larger prey. The secondary bony palate separates the mouth from the respiratory structures,

allowing the animal to breathe while the mouth is open under water (Ross, 1989). Development of large

10

pterygoid flanges, closure of the otic notch, and loss of antorbital fenestra accompanied the evolution of the
bony palate (Sill, 1968). Finally, the evolution of vertebral shape is used to determine crocodilian lineage.
Early crocodilian vertebrae were biconcave at both the anterior and posterior surfaces and the vertebrae
were amphicoelous. Vertebrae in living crocodilians are concave on the anterior surface and convex on the
posterior surface. This modern procoelous vertebral form provides increased flexibility of the spinal
column (Ross, 1989).

Modern characteristics found in the family Goniopholidae suggest that it was either in the central
line of crocodilian evolution or closely related to it. The Goniopholidae appeared in North America,
Europe and Asia during late Jurassic (Figure 2). Two additional families, Theriosuchidae and
Atoposauridae, lived in Europe and Asia during late Jurassic, but were probably minor offshoots of the
Goniopholidae. Fossils from another family, the Pholidosauridae, appear in upper Jurassic sediments of
Europe and Asia, but the family is thought to have become extinct by middle Eocene. The remaining two
families from the suborder Mesosuchia (Libycosuchidae and Notosuchidae) existed for only a short period
of time and were found only in North and South America, respectively. They represented minor offshoots
from late Cretaceous forms (Sill, 1968).

The earth’s land masses were united as one land mass (Pangaea) through the end of the Paleozoic
era when changes in the polarity of the earth’s magnetic field, plate tectonics, and sea-floor spreading began
to shape the modern continents. The initial breakup of Pangaea most likely began in middle to late Triassic
but Briden (1967) presents evidence suggesting that the extreme South Atlantic may have opened slightly
during the Permian. As a result of North Atlantic and Pan-Atlantic rifts, the continents of Laurasia and
Gondwana were created by the end of the Jurassic; Laurasia consisted of North America, Europe and Asia
and was positioned north of the second major plate, Gondwana. Gondwana, divided by water, was
composed of South America, Africa, Antarctica, and Australia. A third plate, supporting India, separated
from Gondwana by the Afro-Indian rift, was located west of the African continent (Dietz and Holden,
1970).

Beginning in late Jurassic and continuing until the Miocene period, Europe had an abundance of

crocodilians. Most of the evidence points to Europe as the site of the origin of modern crocodilians.

11

Eusuchia, the modern crocodilian suborder, evolved during the Cretaceous as continents migrated into their
present locations. The first of five families, the Hylaeochampsidae, appeared in Europe in the early
Cretaceous period and served as a precursor for the families Gavialidae and Crocodylidae (Figure 2). The
Crocodylidae represents the main line of crocodilian evolution since the late Cretaceous (Sill, 1968).

Within the family Crocodylidae there are three subfamilies: the Gavialinae, the Crocodylinae, and
the Alligatorinae. The subfamily Gavialinae is represented by just one species, Gavialis gangeticus, that is
known from as far back as the Miocene. Its range once included North and South America, Africa, and
Asia, but it is now restricted to the northern part of India (Ross, 1989). The subfamily Crocodylinae
consists of the genera Crocodylus, Osteolaemus, and Tomistoma and has thrived since it first appeared in
the fossil records of the late Cretaceous. The subfamily originated in Europe or Eurasia and spread
worldwide by early Tertiary (Figure 2) (Sill, 1968).

The subfamily Alligatorinae includes the genera Alligator, Caiman Paleosuchus, and

 

Melanosuchus (Ross, 1989). The subfamily arose in North America or Central Asia and expanded outward
through the connection between Asia and North America. By Eocene, Alligatorinae reached Europe and
were present there until the atmosphere began to cool during the Oligocene. By early Tertiary, alligatorines
reached South America from North America and from there underwent a secondary radiation to occupy
much of their present distribution (Figure 3) (Sill, 1968).
951W Evolution

Knowledge of plate tectonic activity during the Cenozoic era helps explain the emergence of _C_.
crocodilus in Central and South America. North and South America were separated by a large body of
water (“Panamanian portal”) throughout the Mesozoic and there was no migration of the Alligatorinae into
South America before this time (Savage, 1966). During the Cretaceous, tectonic plates supporting North
and South America moved westward and closed the Caribbean region slightly (Dietz and Holden, 1970). In
the Paleocene, a land bridge (the “Isthmian Link”) was formed by volcanic activity that raised a broad area
of central Nicaragua to northern Colombia above sea level for approximately 12 million years. This bridge
permitted the first radiation of the Alligatorinae and the Crocodylinae into South America. During this

time, the area from the central United States to South America is believed to have been tropical, having

12

higher average temperature and precipitation levels than today. By the end of the Paleocene, the Isthmian
Link submerged again and isolation of South America continued through the Eocene to the Pliocene.
During this period, a second radiation of the Alligatorinae occurred in South America (Savage, 1966).
There were two known genera of the Alli gatorinae in South America by Oligocene and there were four
known genera of the Alligatorinae by Miocene. Alligatorine representation in the Pliocene fauna of South

America included present-day Caiman, Melanosuchus, and Paleosuchus (Sill, 1968). The extensive

 

network of islands available from late Miocene forward facilitated the northern migration of C. crocgilus
into its northern range of southern Mexico. The Isthrnian Link connecting North and South America was
fully restored again by early Pliocene (Savage, 1966).

The Geology of South America

A thorough investigation of the evolution and distribution of C. crocodilus in South America
requires knowledge of the geologic history of the continent. Major geologic regions were established
through tectonic activity prior to the Paleozoic. These regions are: the Guiana shield, the Brazilian shield
and the Coastal Brazilian shield; the Amazon, Parnaiba, and Parana river basins; the Andean Belt, Pampean
ranges, Patagonian’ region, the Llanos region, and the Chaco. Much of continental South America was
engulfed by sea water during the Paleozoic. By the end of the Permian, the sea had recessed from all but
the northwestern and southern regions of South America (Harrington, 1962).

Like the Sierra Nevada in western North America and the Appalachian mountains in eastern North
America, the Andes Mountains were created by movement of large lithospheric plates across the earth’s
surface. The South American continent is suspended on the western edge of the South American Plate and
is adjacent to the Nazca Plate to the West. The Peru-Chile trench lies at the junction between the South
American plate and Nazca plate. West of the Nazca Plate lies the Pacific Plate (Figure 4). Sedimentary
rocks, which form the foundation of the continental edge, were deposited during the Paleozoic. In the
Cretaceous, the Nazca Plate moved eastward away from the East Pacific Ridge and plunged into the Peru-
Chile trench, submerging itself under the South American plate. The rising temperature near the earth’s
core melted the edge of the undersliding Nazca Plate and magma rock surfaced to form a volcanic arc.

Large amounts of magma, from intense volcanic activity, pierced through the surface of existing

l3

sedimentary rock beds and formed enormous batholiths (regions of igneous rock) which remain exposed in
the western ridges today (James, 1973).

Triassic orogenies uplifted many areas of South America forcing the sea water to recede further.
Invasions by sea water in localized areas of the Venezuelan-Peruvian, Chilean, and Patagonian basins
continued through the Cretaceous as a result of fluctuating elevations in topography. Marine sediments
were deposited in these basins throughout the Mesozoic. The Colombian-Peruvian basin was finally filled
in during late Cretaceous by continental deposits from the uplifting mountains and the sea was forced to
leave this area for the last time. The Ecuadorian portal, the connection from the Pacific Ocean inland, was
preserved throughout the Cretaceous. The Caribbean and Patagonian basins began to rise above sea level,
restricting marine water to narrow stretches west of the rising Coast Cordillera (Harrington, 1962).

Great geologic changes occurred during the Tertiary of South America. In the Eocene, the eastern
Cordillera of Colombia, the Perija Range, parts of the Caribbean range, and the Northern Range of Trinidad
were uplifted. Rising mountains to the west deposited thick layers of continental beds in the Colombian-
Peruvian basin. Uplift of the Principal Cordillera (Cordillera Occidental) in southern Peru and western
Bolivia deposited continental debris in the area now known as the Altiplano (Figure 5). In the Oligocene,
the Ecuadorian portal closed, which ended the water connection from the Pacific Ocean to the Colombian-
Peruvian basin. The Caribbean-Venezuelan mountain region was uplifted and continental sediments created
shallow areas of marine and brackish-water. The Altiplano of southern Peru and western Bolivia received
additional continental sediment as the Principal Cordillera was uplifted further (Figure 5). Finally, the
Patagonian basin was uplifted with the Coast Cordillera forcing the sea to retreat from the basin for the last
time. Additional sediments filled in the lower areas and created the extra-Andean Patagonia region
(Harrington, 1962).

The Eastern Cordillera of Colombia and the Central and Eastern Cordilleras of Bolivia were
uplifted by the beginning of the Miocene (Figure 5). Tectonic activity at the end of middle Miocene,
created the paleo-Amazon river, which fed into the paleo—Orinoco river and drained toward the Caribbean
into Maracaibo. Additional activity in late Miocene shifted the Amazon and Orinoco river flow toward the

East and established the water connection with the Atlantic Ocean (Hoom et al., 1995). The Orinoco River

l4

and its tributaries still retain a connection to the Amazon River near the Venezuelan-Brazilian border.
During peak rainfall these areas are covered by continuous water.

Tectonic activity in the Miocene altered the flow of rivers in northern Bolivia. The Guapore,
Itenes, Mamore, Beni, and Madre de Dios rivers and their tributaries flowed southward and emptied into the
Pilcomayo and Paraguay rivers prior to the Miocene. As a result of the uplift of the Bolivian Andes,
massive amounts of sediment were deposited into central Bolivia and raised the area. The increase in
elevation of the rivers redirected the water flow northward. The rivers and their tributaries mentioned
above now empty into the Amazon river basin (Figure 6) (King, correspondence).

Continental sediment continued to accumulate in the Altiplano of western Bolivia. Initial uplift of
the eastern Cordillera of Venezuela and further uplift in Colombia created the present-day Llanos region.
The western Cordillera of Bolivia, Peru, and the Chilean-Argentine Puna experienced volcanic activity.
Finally, the Bolivar geosyncline, located in northwestern Venezuela, was uplifted in the Pliocene. Sea
water was restricted to the lower Atrata and lower Magdelena rivers. The last major Andean uplift occurred
at the end of the Pliocene. At this time the Precordillera of western Argentina and Bolivia and the eastern
part of the Cordillera Real of Ecuador was uplifted (Figure 5) (Harrington, 1962).

Evolutionary Hypotheses

The Andes Mountains are the longest mountain chain in the world; spanning over 66 degrees of
latitude and responsible for the establishment of a variety of climates and environments across western
South America. The creation of the Andes mountains has affected wind and river flow, rain accumulation,
and has restricted the movement of many lower-altitude species. Tectonic activity and climatic changes
during the Cenozoic created many environments that are utilized by a rich and varied fauna across South
America (Duellman, 1979b). The present structure of the Andean mountain range, which began 10 to 15
million years ago in northern Chile and southern Peru, occurred in discrete periods until the Pliocene or
Pleistocene time (James, 1973).

The changes caused by the creation of the Andes Mountains may help explain the evolution of
subspecific differences within C. crocodilus. Major tectonic activity in the Andes Mountains has been

suspended since the Pleistocene. Prior to this time (Paleocene), the first radiation of Alligatorinae migrated

15

to South America across a temporary land bridge from the North. Isolation of South America resumed from
the end of Paleocene to early Pliocene.

Subspeciation of C. crocodilus may have occurred subsequent to the Andean uplift if the
individuals in Central America and northwestern Venezuela and Colombia were isolated from the rest of the
individuals in South America. This would imply that a subspeciation event occurred recently, probably
following the radiation of caiman into Central America in the Pliocene.

Subspeciation of C. crocodilus may have occurred following the initial radiation of Alligatorinae
into South America (Paleocene). If C. crocodilus differentiated prior to the major Central Andean uplift
this would place a hypothesized secondary radiation during the time of continental isolation (Paleocene to
Miocene). Sill (1968) reported the existence of two known genera of the Alligatorinae in South America by
Oligocene and four known genera of the Alligatorinae by Miocene. Subspecific differentiation of Q

crocodilus during this time of great speciation would certainly be plausible.

Trade Data 1983-1993

 

 

was 184 1935 1” 1W7 1988 1m 1990 1991 1992 1993

Figure 1 Annual trade in Caiman crocodilus skins by region from 1983-1993.

l7

 

 

late Jurasic

   

:iilllllllllllllllllllllll ere-owns - 5...... -
early Cretaceous Paleocene late Miocene -

Pleistocene

   

 

 

 

 

Figure 2 Hypothesized evolutionary path of the family Crocodylidae and the sub-family Crocodylinae
showing approximate time of geologic appearance (Sill, 1969). G, Goniopholidae; H, Hylaeochampisidae;
C, Crocodilinae; T, Tomistominae. .

l8

 

 

late Jurasic WWW." late Cretaceous Eocene -
early Cretaceous Paleocene E: late Miocene -

Pleistocene

 

 

 

 

 

Figure 3 Hypothesized evolutionary path of the sub-family Alligatorinae and the genus Caiman showing
approximate time of geologic appearance. C, Crocodilinae; A, Alligatorinae; c, Caiman.

 

 

<—->

 

cocos
PACIFIC PLATE
PLATE

4— ->

EAST PACIFIC

ruse NAZCA
PLATE

‘— —> CHILE
RISE

ANTARTIC
lPLATE

 

_>

MID- ATLANTIC
RIDGE

SOUTH
AMERICAN <— —>
PLATE

Figure 4 Plate tectonic map of South America illustrating the position of the lithospheric plates and

trenches responsible for the formation of the Andes Mountains.

20

  
   

"

J, Sierra Nevada de Santa Marta
- or ' lera de la Costa
‘ f f -

Cordillera Oriental
Cordillera Occidental

Cordillera Central
Cordillera Occidental

v

Figure 5 Map of South America illustrating the individual mountain ranges comprising the Andes
Mountains.

21

Amazon

Madre

de Dios

 

Figure 6 Enlarged map of Bolivia and surrounding countries illustrating the river systems that were
redirected into the Amazon River as a result of the uplift of the Central Andes.

MORPHOLOGICAL STUDY OF Caiman crocodilus

INTRODUCTION

Accurate identification of crocodilian skins and products is essential to the work of law enforcement, agents
of the skin industry, and conservationists (Brazaitis, 1986). Crocodilians provide a major source of leather
for handbags, shoes, wallets, belts, luggage, watchbands, key chains, and a variety of novelty items. During
the 1980’ s, more than 1.5 million crocodilian skins were harvested each year and it was estimated that
caiman skins comprised over 60% of the world market in crocodilian skins (Brazaitis, 1989). Due, in part,
to the increase in registered South American caiman ranches, approximately 1.2-1.5 million C. crocodilus
skins could enter the market annually before the end of the century (Collins and Luxmore, 1996).

The trade of caiman skins, for the most part, is the impetus for many population studies to date.
The demand for better management by the international community has forced countries to evaluate this
natural resource and to implement sustainable use programs. Initiating sound management requires that the
investigators collect basic information on caiman biology, ecology, and distribution.

C. crocodilus is one of the two species in the genus Caiman. Since 1758, the zoological

 

classification of this taxon has been subject to much debate. More than a dozen subspecific divisions have
been proposed between 1758 and 1965 (Kuhn, 1973). Currently there are five recognized subspecies: m
crocodilus, C. c. apamriensis, C. c. fuscus, C. c. yacare, and C. c. chiapasius. The Apaporis River caiman
(C. c. apapgriensis) is currently listed as endangered on the U. S. Endangered Species list, and on appendix
I of the Convention on the International Trade in Endangered Species (CITES) (Brazaitis, 1973). The
Paraguayan or Yacare caiman (C. c. yacare) is also considered endangered by the United States, and
importation of this species into the country is prohibited. C. c. fuscus and C. c. crocodilus are listed on

appendix II of CITES and commercial trade in their skins is also regulated by CITES (Brazaitis, 1986).

22

23

To establish legal control of commercial trade as a conservation mechanism, the international body
connected to CITES must be able to properly identify skins from the different subspecies of C. crocodilus.
CITES and the US. Endangered Species Act requires that each species or population be listed by common
and scientific nomenclature. Recovery plans, mandated by legislation, are based on distribution information
and diagnostic or taxonomic identification. Crocodilian identification is based on the morphology of skulls,
skins, and dentition. (Brazaitis, 1973). To date, no definitive method of differentiating among subspecies of
C. crocodilus has been demonstrated when only skulls or skins are available. An accurate method for
subspecific determination will lead to better understanding of present distribution and breeding habits, and
will assist in the conservation of the species.

To address subspecific identification, morphological variation across the entire range of Q
crocodilus was investigated. Former studies have examined variation across limited geographical areas, and
this is the first study of this magnitude. To determine boundaries for the possible subspecies, variation in
twelve morphological variables was recorded or measured. Two hypotheses, based on possible geographic
and geologic isolating events, were considered when assigning the data into mg groups for linear
discriminant function analysis. The null hypothesis for this study states that there are no significant
differences among populations of C. crocodilus in Central and South America.

Hypotheses Tested

Hypothesis I states that the Andes mountain range isolated the population of caiman in Central
America and northwestern South America from that in rest of the continent. To test this hypothesis,
specimens were divided into two groups according to individual sites of collection. Group 1 included all
specimens from northwest of the Cordilleras de Marida and de la Costa of Venezuela, and west of the
Cordillera Oriental of Colombia (Figure 7). Group 2 included all specimens collected southeast of these
mountain ranges.

Hypothesis 11 states that C. crocodilus subspeciation occurred following the initial radiation of the
Alligatorinae into South America (Paleocene) and prior to the Central Andean uplift (Miocene). For this
hypothesis, specimens were divided into two groups according to the former boundaries of the major river

drainages in South America. C. crocodilus are believed to have migrated into Central America from

24

northern South America after late Miocene (Savage, 1966). Group 1 for this hypothesis included all
specimens collected in areas draining into the Orinoco and Amazon rivers, and all specimens collected in
Central America. The Bolivian rivers Guapore, Itenes, Mamore, Beni, and Madre de Dios flowed South
into the Pilcomayo and Paraguay rivers prior to the Miocene. Group 2 included all specimens collected in
Bolivia and south of the Amazon River system in the Parana, Paraguay, Pilcomayo, and Uruguay Rivers

and their tributaries (Figure 8).

MATERIALS AND METHODS

General Methods

Three hundred and seventy-nine preserved specimens were obtained from every known museum
collection of C. crocodilus housed in the United States (see Specimens Examined). Only whole,
undamaged, alcohol preserved specimens with complete collection information were retained for analysis.
While this sample of specimens represents material from within the entire natural range of C. crocodilus, Q
c. apapgriensis was not represented; this taxon is thought to be extinct in the wild (Brazaitis, 1984).
Localities represented by museum specimens are illustrated in Figure 9.

Morphological Data

Twelve morphological characteristics were examined on each specimen. Characteristics were
chosen with reference to the existing literature concerning the identification of living, preserved, and tanned
specimens (Brazaitis, 1973; Brazaitis, 1989; Brazaitis and King, 1983; King and Brazaitis, 1971). The
twelve variables selected for analysis are defined below:

1) Ratio of head width to snout length (HW/SL). Head width was measured across the angle of the jaws.
Snout length was measured from the end of the snout to the bridge of the eyes (Figure 10). All
measurements were recorded to the nearest tenth of a centimeter using vemier calipers.

2) Mandible pattern (MP) was partitioned into three states based on degree of pigmentation. Character

state 1 was no pigmentation on either upper or lower jaw, character state 2 was faint pigmentation on either

25

(upper or lower) jaw, and character state 3 was dark pigmentation or blotches present on both the upper and
lower jaws (Figure 11).

3) Lateral scale pattern (LP), partitioned into three states based on number of scale rows, included degree of
keeling, and ratio of keeled to unkeeled scale rows. Pattern state 1 was assigned to lateral rows of large
oval scales alternating with a network of irregular shaped small scales and creases. Rows of large scales
were completely and heavily keeled and organized into horizontal rows. State 2 was assigned to lateral
rows of large scales that appeared randomly keeled. Scale rows were not completely keeled but rather
resembled an inverted arched pattern. State 3 was assigned to lateral rows of scales not usually keeled and
not organized in horizontal rows (Figure 11).

4) Total number of keeled and unkeeled lateral rows (TKUR).

5) Proportion of keeled to unkeeled lateral rows (KLr; Figure 12), ranked (Table 5).

6) Length and width of the three largest lateral scales. The proportion of scale width to scale length was
calculated for each lateral scale, and the average proportion for all three scales was analyzed (AVEL/W)
(Figure 13).

7) Nuchal cluster pattern. Thirty-nine different character states were identified (Figure 14); each was
assigned a corresponding number (NCr) (Table 6).

8) Number of dorsal scale rows (DSR); counted from the posterior edge of the nuchal cluster to the rear
margin of the hind limbs (Figure 15).

9) Number of ventral scale rows (V SR); counted from the first horizontal row after the ventral crest to the
last complete row anterior to the anal opening (Figure 16).

10) Ventral scales at mid-body WSRM); defined as the number of transverse scales across the venter half
way between the insertions of front and hind legs (Figure 17).

11) Number of double and single keeled scale rows (SRTAIL) on the dorsal portion of the tail posterior to
the anal opening (Figure 18).

12) Number of tail inclusions (TIr). Counted on the ventral side of the tail between the anal opening and
the seventh scale row on the tail, inclusions are defined as smaller scales placed between rows of existing

larger scales. Many specimens had no tail inclusions, so to eliminate zeros from the data set the number of

26

tail inclusions was ranked with “1” representing zero, “2” representing one, and “3” representing 2, ect.
(Figure 19).
Statistical Analysis

Data were analyzed using linear discriminant function analysis in the package SAS (For Windows,
version 6.08). The morphometric data met assumptions for univariate normality and from this I assumed
multivariate normality. Statistical significance was predetermined at alpha equal to 0.05. Because
discriminant function analysis requires am; groups be defined for analysis, all mg groups were
defined according to hypothesis I or hypothesis II (see Hypotheses Tested).

Each variable was tested for significance within hypothesis 11 using a univariate t-test. This
additional test showed which variables were responsible for increasing the discriminating power. These

variables should be used in all future morphological identification attempts.

RESULTS
Discriminant Analysis

Hypothesis I-- The possibility that the Andes Mountain Range acted as an isolating mechanism
does not fully explain the present distribution of morphological variation within C. crocodilus. Using
twelve morphological variables, 105 of the 125 (84%) specimens assigned to the northern population
(Group 1) were correctly classified, and 221 of the 254 (87%) specimens assigned to the southern
population (Group 2) were correctly classified. The misclassification rate for Group 1 was 16%, while the
misclassification rate for Group 2 was 13%. Using the geographic location of the Andes Mountain Range
to set apri_0_ri groups produced 53 (14%) incorrectly classified specimens (Table 1).

Hypothesis II-- The same twelve morphological variables clearly partitioned the C. crocodilus data
set when considered in accordance with the geology and geography of two major river systems in South
America. Two hundred and fifty-eight of the 266 (97%) specimens collected in Central America, the
Orinoco and the Amazon river systems (Group 1) were correctly classified based on the variables. One
hundred and eleven of the 113 (98%) specimens assigned to the southern population (Group 2) were

correctly classified. The misclassification rate for Group 1 was 3%, while the misclassification rate for

27

Group 2 was 2%. Using geographic and geologic data for the river systems in South America to set M
groupings produced 10 (3%) incorrectly classified specimens (Table 2).

Of the twelve variables, seven characteristics had significant univariate t-values (p < 0.05). The
significant characteristics included: ratio of head width to snout length (HW/SL), mandible pattern (MP),
rank of nuchal cluster pattern (NCr), dorsal scale rows (DSR), total keeled and unkeeled lateral rows
(T KUR), ventral scale rows (V SR), and ventral scale rows at mid-body (VSRM) (Table 3). Using seven
morphological variables in a reduced-model discriminant function analysis produced 257 (97%) and 110
(97%) correctly classified specimens for Group 1 and Group 2, respectively.

Further investigation showed that the ratio of head width and snout length (HW/SL) and nuchal
cluster rank (NCr) did not significantly increase discriminating power above that of the other five
characteristic: the misclassification rates remained unchanged when this datum was excluded. The results
of a second reduced-model discriminant function analysis confirmed that two C. crocodilus populations
could be discriminated using the variables MP, DSR, TKUR, VSR, and VSRM. In this analysis, the
misclassification rate for both Group 1 and Group 2 was 3%; using only five variables resulted in 12 (3%)
incorrectly classified specimens (Table 4).

For mandible pattern (MP), Group 1 was dominated by specimens with character state 1 (no
pigmentation) while Group 2 was dominated by specimens with character state 3 (dark pigmentation) for
jaw pattern (Figure 20). The discriminating power decreased significantly without the inclusion of this key
characteristic. The misclassification rate for Group 1 and Group 2 increased to 15% and 21%, respectively:
subspecies determination without the mandible pattern (MP) produced 63 (17%) incorrectly classified
specimens.

The number of dorsal scale rows (DSR) ranged from 17 to 20 for Group 1 and Group 2. Two
hundred and forty-six of 266 (92%) specimens from Group 1 had either 18 or 19 dorsal scale rows; 109 of
113 (96%) specimens from Group 2 had either 19 or 20 dorsal scale rows. The total number of lateral flank
scales (TKUR). ventral scales (VSR), and ventral scales at mid-body (V SRM) are significantly greater in
Group 2 than Group 1. Identification of specimens using only the flanks, as suggested by Brazaitis (1984),

failed to produce acceptable classification rates in this study. Lateral patterns (LP) and proportion of keeled

28

to unkeeled lateral rows (KLr) were not significantly different between the groups. Total keeled and

unkeeled lateral rows (T KUR) analyzed alone produced 179 (47%) incorrectly classified specimens.

DISCUSSION

One of the goals of this study was to investigate morphological variation within the species Q
crocodilus in order to more effectively manage the present populations. Challenges of conserving Q
crocodilus are similar to many other crocodilian species and proper identification to the subspecific level is
a critical element in the legal protection and management of this species. For example, C. crocodilus is one
of 15 or more crocodilian species that are commercially harvested for the worldwide market. C. crocodilus
skins are a major source of raw and finished material. The present caiman trade has reached over one
million skins annually and is three or four times greater than the “classic" skin trade. International trade in
crocodilians is regulated by the Convention on the International Trade in Endangered Species (CITES) and
additional regulations are imposed by the United States under the Lacey Act of 1900 and the United States
Endangered Species Act of 1973 (Brazaitis, 1989). Classification and sub-species identification of Q
gocodilus is based on morphological characters and morphological variation among populations has been
used as a basis for implementing CITES and regional management plans (Brazaitis, 1984). Large
shipments of skins have been seized and confiscated based on one or more of these morphological
characters (Brazaitis, 1986).

A second goal of this study was to determine if enough morphological variation existed in Q
crocodilus to justify subspecific divisions. Historically, the number of keeled flank scales (KLR) and lateral
flank pattern (LP) have been cited as definitive characteristics for differentiating between C. c. crocodilus

and C. c. yacare. According to Brazaitis (1984), the northern subspecies of caiman, C. c. crocodilus has 2-

 

3 complete rows of alternating large and small well-keeled scales and C. c. yacare has four or more rows of
equal-sized scales that are scarcely keeled (Brazaitis, 1984). It has been stated that C. c. crocodilus has 18-
19 dorsal scale rows (DSR) (Fuchs, 1974) and 20-24 ventral scale rows (V SR) (Werrnuth and Fuchs, 197 8)
while C. Q. yacare has 17-18 dorsal scales rows (DSR) and 22-25 ventral scale rows (V SR) (Brazaitis,

1984). Distinct pigmentation on both upper and lower mandibles (MP) of C. c. yacare specimens has been

29

used as a character to differentiate it from C. c. crocodilus: C. c. crocodilus is thought to lose mandible

 

pigmentation one year following hatching (Brazaitis, 1973).

In this study, not one of the morphological characteristics considered by itself could discriminate
between populations or subspecies. There was no significant difference in flank patterns (LP) or
proportions of keeled to unkeeled lateral scale rows (KLr) in C. crocodilus when examined across the entire
range of the species. There was, however, a significant difference in the total number of keeled and
unkeeled lateral scale rows (TKUR). All states of jaw pattern (MP) were found in Group 1 and Group 2
even though there was a significant difference in frequency of occurrence between the groups. Thirty-nine
different nuchal cluster patterns (NCr) were found between the two groups and this degree of variation is
considerably greater than previously documented. The results of the discriminant analysis using five
morphological characteristics suggest that two populations or subspecies of C. crocodilus are distributed in
concert with the major river systems in South America. This number is significantly less than the number of
proposed divisions for C. crocodilus since 1758, or those suggested by Werrnuth and Fuchs (1978).

Morphological variation appears to have been present prior to the Miocene. This study supports
the theory that the Alligatorinae migrated into South America across the Isthmian Link in the Paleocene,
and underwent a secondary radiation prior to the Miocene (Savage, 1966). Between the Paleocene and the
Miocene, C. crocodilus emerged and spread to occupy its present range in South America. Morphological
variation between specimens living in the Amazon drainages and those in the Paraguay/Parana River
drainages developed during this period of continental isolation. When North America and South America
became reconnected, individuals from the northern (Amazonian) population of C. crocodilus migrated into
Central America and were founders of the population presently known as C. c. fuscus.

For purposes of management, law enforcement, and systematic study, the two northern populations
should be treated together. Regional management of the northern subspecies of C. crocodilus will require
additional cooperation among the various countries supporting northern populations. Regional management
efforts for the subspecies (C. c. yacare) is complicated by the fact that the Bolivian population, once

continuous with the southern drainages, is now reproductively connected to the Amazon basin population.

30

Additional investigation and management consideration must be implemented in Bolivia in order to
maintain variation within these specimens.

Current law enforcement practices must be changed. Subspecific identification of skins from Q
crocodilus can no longer be based on one characteristic, but rather must include mandible pattern (MP),
dorsal scale rows (DSR), total keeled and unkeeled lateral rows (TKUR), ventral scale rows (V SR), and
ventral scale rows at mid-body (V SRM). All five characteristics from a single specimen can be entered,
along with the original data set, into a linear discriminant function (LDF) analysis and reliably assigned to
either of the two populations. Membership probabilities (the probability of correctly assigning an unknown
sample to a group) can be obtained from LDF, and this probability should be used by law enforcement
agents to determine the geographic source of C. crocodilus skin shipments.

Correctly identifying subspecies of C. crocodilus by the method illustrated above is complicated by
the fact that crocodilian skins are not generally shipped with the head attached. Because mandible pattern
(MP) is one of five characteristics demonstrating discriminatory potential, discriminating power is
diminished without this datum. Differentiation to a degree acceptable to the courts cannot be made without
this characteristic. The only practical way to rectify this situation is to modify shipping procedures and
require that dried and salted heads be included when hides are subjected to the tanning process. Another
viable option is to find genetic markers that can be used as a definitive means to identify geographic origin
of the commercial trade. Genetic investigation, coupled with further morphological study, will help address

the myriad of systematic questions identified by this study.

31

 

 

 

CatfilleracblaGNa
J Cmill (h
Grdillera Carnal {I f, “3
ChdilleraOienal is»;
Carillaanzichtal

Figure 7 Map of South America illustrating the Andes Mountain ranges used to partition specimens into
Group 1 and Group 2 for testing hypothesis I.

32

 

Figure 8 Map of South America illustrating major river systems sampled for testing hypothesis 11. River
systems within Group 1 are illustrated in light gray, Group 2 systems are illustrated in dark gray.

 

 

 

 

 

 

 

    
  

,. F'- .
' A u—
" 515,.

'r-

Figure 9 Collection localities of the 379 Caiman crocodilus specimens analyzed by discriminant function
analysis. Symbols may represent multiple specimens.

H19N3‘l
rnous

I
1'11
>
U
U
.1
EF
(I)
z
0
C
_|
[-
rn
2
C)
.r
I

(SMVl‘ so 319w)
H.LCIIM ClVElH

 

Figure 10 Head width (HW) and snout length (SL). Measurements expressed as a proportion (HW:SL) for
analysis.

35

 

  
     
 

 
  
 

 

h .1 .s - .\a-
"7' ' '7"qu

 
   

 

Figure 11 Character states 1, 2, and 3 defined within mandible pattern (MP) and lateral pattern (LP).

 

(z-z) smou m3rv1 03133>1Nn SA 03133)!

 

Figure 12 Proportion of keeled to unkeeled lateral scale rows (KLr).

(saivos rsaom r :10 AM :10 HH’IVA 'rvnomroaoad sows/w)
SEl'lvOS TWELLV'I .LSBEDHV‘I

 

Figure 13 Length and width of the largest lateral scales.

NUCHAL CLUSTERS

 

Figure 14 Nuchal cluster (NCr) patterns identified. Thirty-nine different character states were identified.

E
i
I'D
'33
“to
fir:
3m
0
5.2
am
gm
a
C
I!)

.i‘
#35
'32.

6‘“.
..:.’5:

5

92'.

 

,3"

Figure 15 Number of dorsal scale rows (DSR); counted from posterior edge of the nuchal cluster to the last
scale row parallel with the rear margin of the hind legs.

a
8
5%
82
E?
3r-
>rn
50
a:
om
”m
E
C
$

 

Figure 16 Number of ventral scale rows (VSR); counted from first horizontal row following the ventral
crest to the last complete row anterior to the anal opening.

55
Z
tn
<
111
W
tn
111
to
O
>
I-
m
W
2
U}
>
.r
E
P
a:
8
.<

 

Figure 17 Number of transverse ventral scale rows at mid—body (VSRM).

TAIL INCLUSIONS
(7 ROWS POST ANAL, VENTRAL ASPECT ONLY)

 

Figure 19 Number of tail inclusions (TIr) counted on the ventral side of the tail between the anal opening
and the seventh scale row.

Number of Observations and (Percent Classified) into ANDES:

From ANDES

1

Total

Percent

Priors

1

105
(84.0)

33
(13.0)

138
36.4

0.50

Error Count Estimates for ANDES:

Rate

Priors

1

0.16

0.50

2

20
(16.0)

221
(87.0)

241
63.6

0.50

0.13

0.50

Table 1 Cross-validation summary (Andes Mountain Range) using linear discriminant function analysis.

Total

125
(100.0)

254
(100.0)

379
100.0

Total

0.14

45

Table 2 Cross-validation summary (by river systems) using linear discriminant function analysis.

Number of Observations and (Percent Classified) into GRP:

From GRP 1 2 Total
1 258 8 266
(97.0) (3.0) (100.0)
2 2 1 1 1 1 13
(1.8) (98.2) (100.0)
Total 260 1 19 379
Percent 68.6 3 l .4 100.0
Priors 0.50 0.50

Error Count Estimates for GRP:

1 2 Total

Rate 0.03 0.02 0.02

Priors 0.50 0.50

46

Table 3 Results of the t-test with river data showing the mean, range, SD, SE, 1, DP, and significance (p
<0.05) of the twelve variables used in the analysis. Group 1: N=266; Group 2: N=113.

 

Variable Mean Range SD SE t DF Prob>lTl
HW/SL 1.3 ------ 0.28 0.02 7.28 375.0 0.00 *
1.2 ------ 0.1 l 0.01 5.36 377.0 0.00 *
MP 1.3 1-3 0.51 0.03 -41.71 372.1 0.00 *
3.0 1-3 0.25 0.02 -32. 17 377.0 0.00 *
NCr 21.9 ----- 6.66 0.41 4.32 241.2 0.00 *
19.0 ----- 5.79 0.54 4.09 377.0 0.00 *
DSR 18.7 17-20 0.61 0.04 -7.36 296.6 0.00 *
19.1 17-20 0.42 0.04 -6.40 377.0 0.00 *
TKUR 3.7 2—8 0.66 0.04 -5.81 182.1 0.00 *
4.2 2-7 0.79 0.07 -6.23 377.0 0.00 *
KLr 14.0 ----- 4.09 0.25 1.29 255.5 0.20 NS
13.4 ----- 3.35 0.32 1.19 377.0 0.24 NS
LP 2.0 1-3 0.37 0.02 -1.71 330.8 0.09 NS
2.0 2-3 0.22 0.02 - l .42 337.0 0.16 NS
VSR 20.8 18-27 1.14 0.07 - 10.43 214.6 0.00 *
22.1 19-25 1.12 0.10 - 10.36 377.0 0.00 *
VSRM 13.5 10- 17 1.24 0.08 -9.24 206.8 0.00 *
14.8 12-18 1.27 0.12 -9.33 377.0 0.00 *
TIr 1.1 1-3 0.31 0.02 1.69 335.0 0.09 NS
1.0 1-2 0.18 0.02 1.39 377.0 0.16 NS
SRTAIL 8.5 5- l 1 1.29 0.08 -1.04 246.6 0.30 NS
8.6 6- 12 1.10 0.10 -0.98 377.0 0.33 NS
AVEIJW 1.9 ----- 0.57 0.04 2.01 228.8 0 04 *

1.8 ----- 0.53 0.05 1.95 377.0 0.05 NS

47

Table 4 Reduced model cross-validation summary by river system using linear discriminant function
analysis. Only variables mandible pattern (MP), dorsal scale rows (DSR), total keeled and unkeeled lateral
scale rows (TKUR), ventral scale rows (VSR), and ventral scale rows at mid-body (V SRM) were used in
the analysis.

Number of Observations and (Percent Classified) into GRP:

From GRP 1 2 Total
1 257 9 266
(96.6) (3.4) (100.0)
2 3 1 10 l 13
(2.6) (97.4) (100.0)
Total 260 1 19 379
Percent 68.6 31.4 100.0
Priors 0.50 0.50

Error Count Estimates for GRP:

1 2 Total

Rate 0.03 0.03 0.03

Priors 0.50 0.50

48

 

Distribution of Mandlble Pattern by River System Groups

 

mam

DGRP 2

NumboroISpoelm

 

 

 

 

 

 

STATE 1 STATE 2 STATE 3
Plgmerutton nah:

 

 

 

Figure 20 Distribution of mandible patterns within the major river systems of Central and South America.
Group 1 represents specimens from Central America, the Orinoco river system, and the Amazon river
system. Group 2 represents the Rivers Parana, Paraguay, and Uruguay, and associated drainage systems.

APPENDICES

49

APPENDIX A

Specimens Examined
BOLlVIA--BENT: Rio Itenez at Versalles (12 37'S, 63 22'W) (UF-A 40163-40165, 40167); 20 km ESE
Santa Ana de Yacuma (13 45'S, 65 15'W) (UF-A 40160—40161); Cercado Prov, 40 km N. Trinidad,
Estancia La Habana (14 08'S, 65 04W) (UF-A 40152, 40156, 40158-40159); Itenez Prov, Rio Blanco,
Bella Vista (13 168, 63 41'W) (UF-FGT 3310); Marban Prov, 30 km S. Trinidad on road to San Lorenzo
(15 01'S, 64 43'W) (UF-A 40150); Vaca Diez Prov, Lago San Jose, 12 km NNE Riberalta (10 51'S, 66
01W) (UF-A 40136-40138, 40140-40141, 40144-40147, 40149), Tumi Chucua, 30 km S Riberalta (10
59's, 66 06'W) (USNM 280228-280229); Yacuma Prov, 20-25 km E San Borja (14 55'S, 66 20W) (UF-
FGT 3314-3316). SANTA CRUZ: Chiquitos Prov, 8 Shore Laguna Concepcion, Estancia San Juan (17
31'S, 61 26'W) (UF-A 40050, 40052); Nuflo De Chavez Prov, San Pablo (15 53'S, 63 17'W) (UF-A 40069-
40070), Rio Zapacos (15 35'S, 63 14'W) (UF-A 40077). BRASIIr-AMAZONAS: Manaus (3 08'S, 60
OO'W) (AMNH 64855); Rio Itacuai (4 20'S, 70 12'W), 10 mi SE Rio Tachitu Junction (LACM 138238-
138239). BAHIA: (12 008, 42 OO'W) (ANSP 9725). GUANABARA: Rio de Janeiro (22 54'S, 43 14'W)
(AMNH 31747-31748). MARANHAO: 850 Joao Dos Patos (6 308, 43 42'W) (AMNH 62123). MATO
GROSSO: Aquadauna, Rio Aquadauna (20 12'S, 55 45'W) RIF-A 46178); Caceres, 80 km SE Fazenda
Nova Esperanca (16 20'S, 57 48W) (UF-A 46045, 46047-46048, UF-A 46051-46052); Caramujo (15 468,
57 45'W) (UP-46030, 46032-46033, 46036); Corumba (19 01'S, 57 39'W) (UMMZ 103216); Cuiaba (15
35'S, 56 05'W), Rio Paraguai (UF-A 46201, 46204); Descalvado (16 45'S, 57 42'W) (FMNH 10824); Rio
Paraguay below Descalvado (FMNH 9064); Rio Paraquai (15 558, 57 40'W) (UF-A 46039-46040); Rio
Sangridorazinho ( 15 558, 57 05'W) (UF-A 46063, 46065-46066); N Rio Sepatuba (15 48$, 57 39'W) (UF-
A 46043); Tapirape River (mouth) (10 41'S, 50 38'W) in Araguaya River (UMMZ 103008 [two
specimensl); Vila Bella, Rio Guapore (14 27'S, 60 06'W) (UF-A 46057, 46059-46060). MATO GROSSO
DO SUL: Coxim, Rio Taqueri (18 288, 54 45'W) (UF-A 46170); Fazenda Prarnisone, Rio Paraguai (18
27'S, 55 45'W) (UF-A 46166, 46169); Fazenda Sao Vicente, Rio Paraquai (18 45'S, 56 40'W) (UF-A

46160-46161); Nabilequa, Fazenda Piuval (19 408, 57 40'W) (UF-A 46100, 46103); Nhumini, Rio

 

50

Paraguai (14 008, 56 15'W) (UF-A 46151); Porta Murtinho, Rio Paraguay (21 25'S, 57 40'W) (UF-A
46106, 46108-46109, 46194); Region Corumba, Rio Paraguai (:9: 19 12'S, 57 49W) (UF-A 46188); Rio
Negro, Rio Negro (19 US, 55 15'W) (UF-A 46174). PARA: Altamira, 50 km (air) S near Cachoeira do
Espelho (3 488, 52 32'W) (USNM 303485—303487); Belém (1 27'S, 48 29'W) (CM 2974); Obydos
[Obidos] (1 55'S, 55 31'W) (FMNH 16073, MCZ 2596, 162011-162013); Santarém (0 39'N, 50 05'W)
(MCZ 2814); Reserva Biologica Rio Trombetas, Igarapé Abui (1 308, 56 30W) (USNM 289094-289095).
RONDONIA: Costa Marques (12 25'S, 64 10'W) (UF-A 46024, 46028-46029); Guarjara Mirim (10 55'S,
65 15'W) (UF-A 46013, 46018, 46023). COLOMBIA--AMAZONAS: Leticia (4 09's, 69 57W) (CAS
91653-91662). ANTTOQUIA: Turbo (8 06'N, 76 43'W) (MCZ 2721); Uraba [Boca] (8 03'N, 76 50'W),
Rio Currulao (FMNH 63899-63900). ATLANTICO: Barranquilla (10 59'N, 74 48'W), Rio Magdalena
(CU 12809, KU 94579). CHOCO: Tagachi (6 13'N, 76 43'W), Rio Atrato (LACM 47185-47190); Golfo de
Uraba, Unguia (8 DIN, 77 04'W) (FMNH 63901). CUNDINAMARCA: Giradot (4 18'N, 74 49W) (MCZ
17602-17608). MADGALENA: Cienaga [Grande] (11 OO'N, 74 15'W) (FMNH 69890-69891); Don Diego
(10 59’N, 73 47'W) (UMMZ 46680). META: Villavicencio (4 09'N, 73 39'W) (MVZ 71377).
PUTUMAYO: Rio Mecaya (0 28'N, 75 12'W) (FMNH 69679). VALLE: Buenaventura (3 53'N, 77 04'W),
Rio Raposo (USNM 151748). COSTA RICA--HEREDIA: Puerto Viejo (10 28'N, 84 01W) (KU 67460).
LIMON: Puerto Viejo (9 39'N, 82 46'W) (KU 43893); Tortugero (10 35'N, 83 31’W) (UF 10324, UF
10494, UP 10737, UF 21161, MCZ 54997); Zent (10 02'N, 83 17'W) (MCZ 15333-15334).
PUNTARENAS: Rincén (8 42'N, 83 29'W) (LACM 38018). CUBA--Isla de la Juventud (=Isla de Pinos [
21 45'N, 82 45'W]) (KU 61857). ECUADOR--NAPO: Santa Cecilia (0 03'N, 76 58'W) (KU 107047,
112242, 122199, 126670-126671). PASTAZA: Rio Curaray (2 22'S, 74 05'W) at Paracachi (USNM
196276-196277, 211256-211262), Teresa Mama (1 568, 77 13'W), Bobonaza Rio (USNM 196275).
GUATEMALA-JUTIAPA: Casa Grande, 4 km E at Finca La Trinidad Laguna (14 UN, 89 54'W)
(UMMZ 107839-107 843). GUYANA-EAST DEMERARA: Buxton (6 47'N, 58 02'W) (FMNH 26673-
26678, 26681-26687); 50 mi above Georgetown (6 48'N, 58 10'W) (USNM 145473). WEST
DEMERARA: Dunoon, Demerara River (6 26'N, 58 18'W) (UMMZ 46111). HONDURAS-Rio Aguan

( 15 55'N, 85 41'W) (UF-A 42519-42520); Rio Chapagua (15 58'N, 85 43'W) (UF-A 42516-42518); Rio

 

51

Platano (15 53'N, 84 41'W) (UF-A 42510). GRACIAS A DIOS: Laguna de Brus (15 45'N, 84 29'W) Canal
Amatingni (UF-A 42509); Laguna Brus (UF-A 42512); Laguna de Tansin, 15 km NW Puerto Lempira (15
19'N, 83 58'W) (LACM 48459-48461, LSUMZ 21718-21719); Palacios (15 55'N, 84 58'W), Laguna
Bacalar (UF-A 42504-42507, 42515); Rio La Criba, W of Laguna De Ibans (15 52'N, 84 46'W) (UF-A
42513- 42514); Rio Patuca (15 47'N, 84 15'W) (UF-A 42508); Rio Sico (15 58'N, 84 58'W) (UF-A 42001).
LA REGION MOSQUITIA: Rio Platano (15 53'N, 84 41'W) (UF-A 42511). MEXICO-CHIAPAS:
Soconusco, Escuintla (15 20'N, 92 38'W) (UMMZ 88262). YUCATAN: Chichén Itza (20 40'N, 88 34'W)
(MCZ 17581). NICARAGUA--RIVAS: Rio Sapoa (11 15'N, 85 36W) (LACM 38015-38016); San Jefe
del Sur, Isla de Ometepe (11 30'N, 85 35'W) (KU 86272). ZELAYA: Rama (l2 09'N, 84 13'W), 6 mi W
(TCWC 19285-19286). PANAMA--BOCAS DEL TORO: Almirante (9 18'N, 82 24'W) (KU 80219);
Almirante, 3 km W (KU 108317-108318); Bastimentos (9 21'N, 82 12'W), Isla Bastimentos (KU 96989);
Cayo de Agua (9 09'N, 82 02'W) (USNM 148276); Cayos Zapatilla [Grande] (9 16'N, 82 03W) (KU
96990); Descudo de Veraguas Id (9 15'N, 81 40'W) (USNM 148249-148250); Fish Creek (9 00'N, 82
14'W) (KU 96991-96992). CHIRIQU : Guabala (8 13'N, 81 44'W), 2 km S (KU 107863). COLON: Fort
Sherman (9 22'N, 79 57W) (MCZ 26760); Gamboa (9 07'N, 79 42'W) (UMMZ 98381); Gatuncillo (9 MN,
79 39'W) (CU 3967 [2 specimensD; Gutrin, Esclusas (9 16’N, 79 55W) (LACM 74403); Toro Point (MCZ
22332-22333). DARIEN: Boca de Cupe (8 03'N, 77 35'W), 12 km SSW at Rio Cupe (KU 96987); Cana
(7 56'N, 77 43'W) (USNM 50090-50092); El Real de Santa Maria (8 08'N, 77 43'W) (KU 80605, 107879;
USNM 140610-140611, 211264); Isthmus (9 OO'N, 80 00'W) (UMMZ 65917); Ortiga Camp (8 45'N, 78
00'W) along Rio Chucunaque (FMNH 170162); Rio Chucunaque at Rio Ilagandi, 8 km above Rio Morti
(8 51'N, 77 59'W) (KU 107870); Rio Chucunaque near mouth of Rio Canglén (8 20'N, 77 40'W) (UMMZ
124209-124212); Rio Tuira at Rio Monos (9 16'N, 78 08'W) (KU 96993); Rio Ucurganti (8 27'N, 77
48'W), 7 km above mouth (KU 107871-107878); Rio Urusela, 2 mi above El Real de Santa Maria (8 08'N,
77 43'W) (USNM 167225); Yaviza (8 UN, 77 41'W) (MVZ 83217). LOS SANTOS: Guanico Abajo (7
18'N, 80 26'W) Rio Guanico (KU 107864). PANAMA: Arraijan (7 46'N, 80 52'W), 8 km WSW (KU
96988); Barro Colorado Isla (9 09N, 79 51'W) (FMNH 179366, 13329, 13402; USNM 165968); Barro

Colorado Isla, Wheeler Trail (UMMZ 63729); Panama Bay (9 DIN, 79 29’W) (USNM 72384-72385);

52

Pedro Miguel (9 01'N, 79 37'W) (AMNH 67201); Rio Chagres (9 07'N, 79 42'W) (KU 107865-107869).
PARAGUAY--ALTO PARAGUAY: 2 km N Bahia Negra (20 158, 58 12'W) (UF-C 108, 111, 117);
Estancia General Diaz, Laguna General Diaz (21 108, 58 32'W) (UF-C 129); Frog Ponds N Laguna
General Diaz (UF-C 131); 3 km E Estancia General Diaz (UF-C 133); 7 km ESE Estancia General Diaz,
Tajamar 9 (UF-C 137); Puerto Ramos 7 km S Bahia Negra (20 17'S, 58 10'W) (UF-C 72, 77-78, 86, 92).
AMAMBAY: Arroyo Chacalalina (=A Chacal) 4 km above mouth, 60 km SSW Bellavista (22 08'S, 56
31'W) (UF-C 142). CENTRAL: Asuncion (25 168, 57 40'W) (FMNH 9497-9498). PRESIDENTE
HAYES: Estancia Golondrina, 30 km W. Villa Hayes, Potrero Doiia Cynthia (25 068, 57 34'W) (UF-C 63-
66, 68); Estancia Juan De Zalazar, Laguna Vicente, 1 km W headquarters (23 068, 59 18'W) (UF-C 16,
18); Estancia Juan De Zalazar, Laguna 8 km NE Headquarters Canadon Cacique, on map on south eastern-
most Laguna 7.2 km NE Trans Chaco Hwy (UF-C 5-6); Estancia Juan De Zalazar, Juan De Zalazar Tuya, 3
km SW Headquarters (UF-C 22, 24); Estancia Tingunque, Rio Monte Lindo at Retiro Pozo Arias, 18 km S
Fortin Avalos Sénchez (23 288, 60 07'W) (UF-C 46, 49). MISIONES: Estancia Sarandy, 15 km SSW
Santiago (27 09'S, 56 47'W) (UF-C 27-28, 31-33). NEEMBUCU: Est. 26, 8 km (air) NE Tacuara (25 59'S,
57 31'W) (UF-C 147). PERU--LORETO: 10 mi E Leticia, Colombia ( 4 09'S, 69 57W) (UF 35084); near
Yaguas River, 90 mi NW Leticia, Colombia ( 3 008, 71 00'W) (AMNH 73156); Nauta (4 328, 73 33'W)
Rio Samiyia, Santa Elena (FMNH 109805); Nazareth (3 54'S, 70 46'W), 20 mi above mouth Yavari River
opposite Remate De Males Brazilian Village (FMNH 5671); Iquitos (3 468, 73 15'W) on Nanay Road
(TCWC 48675). MADRE DE DIOS: Madre De Dios ( 12 365, 70 05'W) (FMNH 40004). UCAYALI:
Pucallpa (8 23'8, 74 32'W), 28 mi ESE (TCWC 24129-24130); Rio Ucayali, Yarina-Cocha (Lake) (8 15‘S,
74 43’W) (FMNH 45639, 45641-45649, 56188); Upper Ucayali River (4 308, 73 27'W) (AMNH 71090-
71092); Yarina-Cocha near Pucallpa (FMNH 45293). SURINAM--PARAMARIBO: Charlesburg
Paramaribo (5 SUN, 55 10'W) (AMNH 8149, CM 44424-44425, TCWC 23567-23570, UF 99371); 25 km
from Paramaribo, on road to Nickerie (MCZ 154922). TOBAGO--ST. MARY: SE of Goldsborough (11
15'N, 60 40'W) in Goldsborough River (USNM 228138). TRINIDAD--(10 20'N, 61 30W) (AMNH
72878; FMNH 41676, 53650-53652; USNM 17718). VENEZUELA--APURE: San Fernando de Apure (7

54'N, 67 28'W), 4 km W at Hato La Guanota (T CWC 45392); San Fernando de Apure, 6 km W (TCWC

 

53

46238, 47411); Estado Aragua (9 19'N, 64 54'W) (UMMZ 135258). BOLIVAR: Cerro Guanaco (4 40'N,
63 51'W) (AMNH 25197, 27191, 27375). FALCON: Palma Sola (10 36'N, 68 33'W), Rio Yaracuy
(UMMZ 55954). MONAGAS: Barrancas, 140 km NE (=Delta Amacuro Territory, 9 48'N, 61 08'W)
(LACM 31500); Caripito (10 08'N, 63 06'W) (AMNH 65559); La Guanota (10 13'N, 63 32'W), Rio Apure
(TCWC-LE 15, 19, 21-23, 25-31, 33, 37, 39-40, 42-49, 51); Maturin, 60 km SE( 9 19'N, 62 48'W)
(LACM 31499). PORTUGUESA: Mun. Nueva Florida (8 57’N, 69 01'W), Santo Domingo (UF 99370).

SUCRE: Guaratinos (10 33'N, 63 07'W), 7 km S, Rio Sabacual (KU 167616).

 

54

APPENDIX B

Table 5 Keeled vs unkeeled lateral scale row proportions. Each specimen was assigned the corresponding
number for analysis (KLr).

 

Proportion of Keeled To Number
Unkeeled Lateral Rows Assigned
0:2 1
0:3 2
0:4 3
0:8 4
1:1 5
1:2 6
1:3 7
1:4 8
1:5 9
1:6 10
2:0 11
2:1 12
2:2 13
2:3 14
2:4 15
3:0 16
3:1 17
3:2 18
3:3 19
3:4 20
4:0 21
4:1 22
4:2 23
4:3 24
4:4 25

55

APPENDIX C

Table 6 Nuchal cluster patterns recorded. Each specimen was assigned the corresponding number for
analysis (NCr).

 

 

N uchal Cluster Pattern Number
Assigned

4-4-4-4-2 1
4-4-4-2-2 2
4-4-2-2-2 3
3-4-4-2-2 4
2-4-4-4-2 5
2-4-4-3-2 6
2-4-4-2-2 7
2-4-2-2-2 8
2-2-4-2-2 9
1-5-4-4-2 10
1-4-4-4-2 1 1
1-4-4-2-2 12
1-4-3-2-2 13
1-2-4-4-2 l4
5-5-2-2 15
5-4-4-2 16
5-4-2-2 17
4-4-4-2 1 8
4-4-2-3 19
4-4-2-2 20
4-4-1-2 21
4-3-2-2 22
4-2-2-2 23
4-2-2-1 24
3-4-4-2 25
3-4-3-2 26
3-4-2-2 27
3-2-2-2 28
2-4-4-2 29
2-4-2-2 30
2-3-2-2 3 1
2-2-2-2 32
1-4-4-2 33
1-4-2-2 34
1-2-2-2 35
4-4-4 36
4-4-2 37
4-3-2 38
4-2-2 39

 

LIST OF REFERENCES

56

LIST OF REFERENCES

Allsteadt, John. 1994. Nesting ecology of Caiman crocodilus in Cano-Negro, Costa Rica. Journal of
Herpetology 28(1): 12-19.

Alvarez del Toro, M. 1974. Los Crocodylia de Mexico. A. C., Mexico, D. F., Institutio Mexicano de
Recursos Naturales Renovables.

Ayarzaguena, J. 1983. Ecologia del caiman de anteojos o baba Caiman crocodilus L. en los Llanos de
Apure (Venezuela). ACTA Vertebrata Donana 10, 1-16.

Brazaitis, P. 1973. The identification of living crocodilians. Zoologica 58:59-101.
Brazaitis, P., and F. W. King. 1983. A Guide to the Identification and forensic examination of the caiman,

Caiman crocodilus crocodilus and the Yacare caiman, Caiman crocodilus yacare. U. S. Fish and
Wildlife Service and Florida State Museum Presentation. (Transcript).

 

Brazaitis, P. 1984. Problems in the identification of commercial crocodilian hides and products, and the
effect on law enforcement. pp. 110-116, I_n: Proc. 6th Working Meet, IUCN/SSC Crocodile
Specialist Group, Victoria Falls, Zimbabwe and St. Lucia Estuary, Repub. South Africa, Sept. 19-
30, 1982.

Brazaitis, P. 1986. Reptile leather trade: the forensic science examiner’s role in litigation and wildlife law
enforcement. Journal of Forensic Science 31:621-629.

Brazaitis, P. 1989. The forensic identification of crocodilian hides and products. IUCN Pub1., New Ser.
Gland, Switzerland. pp. 17-43.

Briden, J. 1967. Recurrent continental drift of Gondwanaland. Nature 215: 1334-1339.

Cintra, R. 1988. Nesting ecology of the Paraguayan caiman (Caiman yacare) in the Brazilian Pantanal.
Journal of Herpetology 22:219-222.

Cintra, R. 1989. Maternal care and daily pattern of behavior in a family of caimans, Caiman yacare, in the
Brazilian Pantanal. Journal of Herpetology 23:320-322.

Collins, L. and R. Luxmoore. 1996. World Trade in Crocodilian Skins, 1992-1993. World Conservation
Monitoring Centre. International Alligator and Crocodile Trade Study (IACI‘ S) Report, pp. 1-40.

Crawshaw, P. G. and G. B. Schaller. 1980. Nesting of Paraguayan caiman (C. yacare) in Brazil. Pap.
Avulsos 2001., S. Paulo. 33:283-293.

Dietz, R. S., and J. C. Holden. 1970. Reconstruction of Pangaea: breakup and dispersion of continents,
Permian to present. Journal of Geophys. Res. 75:4939-4956.

Duellman, W. E. 1979a. The South American Herpetofauna: A panoramic view. pp 1-28. I_n: W. E.
Duellman (ed.). The South American Herpetofauna: Its Origin, Evolution, and Dispersal. Mon.
1. Museum of Natural History, Univ. of Kansas, Lawrence, Ka.

57

Duellman, W. E. 1979b. The Herpetofauna of the Andes: Patterns of Distribution, Origin, Differentiation,
and Present Communities. pp 371-459. I_n: W. E. Duellman (ed.). The South American
Herpetofauna: Its Origin, Evolution, and Dispersal. Museum of Natural History. Mon. 15. Univ.
of Kansas, Lawrence, Ka.

Fuchs, K. 1974. Die Krokodilhaut. 183 S.; Darmstadt (Verlag Eduard Roether). Translated for the
USFWS by Christopher W. Spears.

Gorzula, S. J. 1978. An ecological study of (Caiman crocodilus crocodilus) inhabiting savanna lagoons in
the Venezuelan Guayana. Oecologia 35:21-34.

Harrington, H. J. 1962. Paleogeographic development of South America. Bulletin. American Association
of Petroleum Geology 46:1773-1814.

Hennig, W. 1969. Die Starnmesgeschichte der Insenkten. Senckenberg-Bush 49. Frankfurt/M.

Herron, J. C. 1994. Body size, spatial distribution, and microhabitat use in the Caimans, Melanosuchus
niger and Caiman crocodilus, in a Peruvian lake. Journal of Herpetology 28:508-513.

 

Hoom, C., Guerrero, J., Sarmiento, G. A,. and M. A. Lorente. 1995 Andean tectonics as a cause for
changing drainage patterns in Miocene northern South America. Geology 23:237-240.

James, D. E. 1971. Plate tectonics model for the evolution of the central Andes. Bulletin. Geologic
Society of America 82:3325-3346.

James, D. E. 1973. The Evolution of the Andes. Scientific America 229(2):60-70.

King, F. W. and P. Brazaitis. 1971. Species identification of commercial crocodilian skins. Zoologica
56:15-70.

Kuhn, Oskar. 1973. Handbuch der Palaoherpetologie. Encyclopedia of paleoherpetology: Crocodylia.
part 16. Gustav Ficher Verlag. Stuttgart. pp. 116.

Luxmoore, R. 1990. International Crocodilian Skin Trade: Trends and Sources. In Crocodiles.
Proceedings of the 10th Working Meeting of the Crocodile Specialist Group of the Species
Survival Commission of the International Union for Conservation of Nature and Natural Resources
convened at Gainesville, Florida. 23 to 27 April 1990.

Luxmoore, R. and L. Collins. 1994. World Trade in Crocodilian Skins, 1990-1991. World Conservation
Monitoring Centre. International Alligator and Crocodile Trade Study (IACT S) Report. pp. 1-24.

Magnusson, W. E. 1985. Habitat selection, parasites and injuries in Amazonian crocodilians. Amazoniana
9: 193-204.

Magnusson, W. E., E. Vieira da Silva, and A. P. Lima. 1987. Diets of Amazonian Crocodilians. Journal of
Herpetology 21:85-95.

Mayr, E. 1974. Cladistic analysis or cladistic classification?. Z. 2001. syst. Evolution Forsch. 12:94-128.

Ouboter, P. E., and L. M. R. Nanhoe. 1988. Habitat selection and migration of Caiman crocodilus in a
swamp and swamp-forest habitat in northern Suriname. Journal of Herpetology 22:283-294.

 

58

Outobter, Paul E., and Lurly M. R. Nanhoe. 1989. Notes on the dynamics of a population of Caiman
crocodilus crocodilus in northern Suriname and its implications for management. Biological
Conservation 48:243-264.

 

Rivero Blanco, C. 1974. Habitos reproductivos de la baba en los Llanos de Venezuela. Natura 52:24-29.
Ross, C. A. and Garnett, S. (Eds.). 1989. Crocodiles and Alligators. Facts on File, Inc., New York, NY.
Savage, J. M. 1966. The origins and history of the Central American herpetofauna. Copeia 1966:719-766.

Schaller, G. B. and P. G. Crawshaw. 1982. Fishing behavior of Paraguayan caiman (Caiman crocodilus).
Copeia 1982:66-72.

Sill, W. D. 1968. The zoogeography of the crocodilians. Copeia 1968:76-88.

Stanton, M. A. & J. R. Dixon. 1975. Studies on the dry season biology of Caiman crocodilius crocodilus
from Venezuelan Llanos. Mems Soc. Cien. Nat. la Salle 35:237-265.

Stanton, M. A. & J. R. Dixon. 1977. The breeding biology of the spectacled caiman, Caiman crocodilus
crocodilus, from Venezuelan Llanos. Fish and Wild. Service. Wildlife. Resource Report, No. 5.

Thorbjarnarson, J. B. 1991. An analysis of the spectacled caiman (Caiman crocodilus) harvest program in
Venezuela. pp. 217-235. In: Robinson J. G. and K. H. Redford (Ed.). Neotropical wildlife use
and conservation. xvii+50p.

Thorbjarnarson, J. B. 1993a. Diet of the Spectacled caiman (Caiman crocodilus) in the central Venezuelan
llanos. Herpetologica 49: 108-1 17.

Thorbjarnarson, J. B. 1993b. Fishing behavior of Spectacled caiman in the Venezuelan llanos. Copeia
1993:1166-1171.

Villalba-Macias, J .S. 1986. South American exports of crocodile skins. In: Crocodiles. Proceedings of
the 8th Working Meeting of the Crocodile Specialist Group of the Species Survival Commission of
the International Union for the Conservation of Nature and Natural Resources convened at Quito,
Ecuador, 13 to 18 October, 1986.

Wermuth, H., and K. Fuchs. 1978. Bastarde zwischen sudostasiatischen Krodilen. Stuttgart 1 (8)21-17.
Translated for the USFWS.by Christopher W. Spears.

 

MICHIGAN STATE UNIV. LIBRARIES
llllllllllll lllllllll||l|l|lllllllllllllllll11111
31293015714227