Citation
Ecology and behavior of the spectacled caiman (Caiman crocodilus) in the central Venezuelan llanos

Material Information

Title:
Ecology and behavior of the spectacled caiman (Caiman crocodilus) in the central Venezuelan llanos
Creator:
Thorbjarnarson, John B., 1957-
Publication Date:
Language:
English
Physical Description:
ix, 390 leaves : ill., photos ; 29 cm.

Subjects

Subjects / Keywords:
Alligators ( jstor )
Animal nesting ( jstor )
Bodies of water ( jstor )
Dry seasons ( jstor )
Eggs ( jstor )
Female animals ( jstor )
Lagoons ( jstor )
Llanos ( jstor )
Rainy seasons ( jstor )
Savannas ( jstor )
Caiman (Genus) -- Behavior -- Venezuela ( lcsh )
Caiman (Genus) -- Ecology -- Venezuela ( lcsh )
Dissertations, Academic -- Forest Resources and Conservation -- UF
Forest Resources and Conservation thesis Ph. D
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1991.
Bibliography:
Includes bibliographical references (leaves 369-389).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by John B. Thorbjarnarson.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
027763085 ( ALEPH )
26483347 ( OCLC )

Downloads

This item has the following downloads:


Full Text













ECOLOGY AND BEHAVIOR OF THE SPECTACLED CAIMAN (CAIMAN
CROCODILUS) IN THE CENTRAL VENEZUELAN LLANOS




















By

John B. Thorbjarnarson


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1991














ACKNOWLEDGMENTS


First and foremost I must offer my thanks to Tomds and Cecilia Blohm. Tomds and Cecilia initially made my trip to Venezuela possible, and offered me the hospitality of their home in Caracas and the opportunity to work on their ranch, Hato Masaguaral. Tomds' long standing interest in crocodilian conservation was one of the major impetuses for this work, and he supported it in every possible way. Without the support of the Blohms this study would never have been done.

During my stay in Venezuela I had the pleasure of

working for the Fundaci6n para la Defensa de la Naturaleza (FUDENA) as a research biologist. FUDENA, and its director Dra. Maria de Lourdes Acedo de Sucre, provided me with considerable logistic and financial support throughout. I would like to especially thank Glenda Medina Cuervo for her active support of all aspects of this work. Thanks also go to the entire FUDENA staff for their help and friendship.

The field work conducted on Hato Masaguaral was made possible with the assistance of a great number of individuals. In particular I would like to thank Gustavo Hernandez, Maria del Carmen Mufioz, Tibisay Escalona, Pedro Vernet, Ildemaro Gonzalez, and Jaimie Aranguen for their efforts. Gustavo spent nearly as much time on the ranch as


ii









I did, and kept the research program operational during my absences. Maria worked with me on the caiman radio telemetry project and collected all the data for the 1987 wet season. She will be my coauthor for the publication resulting from the information in chapter 4.

Prof. F. Wayne King, who chaired my graduate committee, has offered me a tremendous amount of support throughout this project, as well as with all aspects of my studies in the realms of crocodilian biology and conservation. Prof. King also kindly provided me with space in his office in the Florida Museum of Natural History and unlimited access to his computer facilities. I would also like to express my gratitude to the other members of my committee, Katherine Ewel, Lou Guillette, Mel Sunquist, and Fred Thompson, for their help and their patience in reviewing and correcting my dissertation. John Eisenberg served on my committee in every way except for reviewing the dissertation. John Robinson also served on my committee before departing from the University of Florida. I would also like to thank John for providing the impetus for my first visit to Venezuela.

During this lengthy study discussions with numerous individuals added considerably to my understanding of reptile and crocodilian ecology, as well as the many other topics relating to this dissertation. These individuals include F. Wayne King, Kent Vliet, Grahame Webb, Harry Messel, Stephan Gorzula, Donald Taphorn, Andres Eloy Seijas, Jose Ayarzaguena, Carlos Rivero Blanco, Phil Hall, Kent


iii









Redford, Lou Guillette, Peter Brazaitis, Alan Woodward, Dennis David, Bill Magnusson, Paul Moler, Lee Fitzgerald, Val Lance, Peter Crawshaw and Eduardo Cartaya. I would also like to acknowledge the assistance of the numerous individuals who helped me and offerred me hospitality during my work on caiman on ranches other than Hato Masaguaral, especially W. DeVrees, A. Branger, E. Hernandez, H. Escanona, L.E. Moser, P. Zarate and A. Carillo Garcia. I would also like to thank Mirna Quero de Pefia, Francisco Perez Perez, Gonzalo Medina, and Evaristo Martinez of the Venezuelan Ministerio del Ambiente y de los Recursos Naturales Renovables (MARNR) for granting me permission to carry out these investigations.

Financial support for this project was supplied

principally through the Smithsonian Institution, FUDENA, Wildlife Conservation International (New York Zoological Society), and Tomds Blohm. Additional support was provided by the World Wide Fund for Nature (WWF), and the World Wildlife Fund (US). For their help in obtaining and administering funds for this project I would like to thank Rudy Rudran, Dale Marcellini and Betty Howser of the Smithsonian Institution (National Zoological Park); Stuart Strahl, Archie Carr III, John Behler and Mary Pearl of Wildlife Conservation International; Curtis Freese and Ginette Hemley of WWF-US; and Hartmut Jungius, Sylvia Guignard and Aileen Ionescu-Somers of WWF-International.


iv








Funding for my graduate assistantship while writing up this dissertation was kindly provided by the IUCN Crocodile Specialist Group (H. Messel, chairman), and PROHESA (E. Hernandez, director).


v








TABLE OF CONTENTS


page

ACKNOWLEDGEMENTS.......................................11

ABSTRACT...............................................Viii

CHAPTERS

1 INTRODUCTION.....................................1

2 DESCRIPTION OF THE STUDY SITE....................7

The Llanos Habitat...............................9
Hato Masaguaral.................................11

3 CHARACTERISTICS OF THE CAIMAN POPULATION........20

Introduction....................................20
Methods.........................................22
Results.........................................26
Discussion......................................52

4 MOVEMENT PATTERNS, HOME RANGE SIZE,
AND HABITAT UTILIZATION.........................61

Introduction....................................61
Materials and Methods...........................62
Results.........................................67
Discussion.....................................119

5 DIEL ACTIVITY PATTERNS.........................130

Introduction...................................130
Methods........................................132
Results........................................134
Discussion.....................................138

6 CAIMAN DIET....................................144

Introduction...................................144
Materials and Methods..........................145
Results........................................147
Discussion.....................................160

7 FEEDING BEHAVIOR...............................170

Introduction...................................170
Methods........................................171
Results........................................172
Discussion.....................................184


vi








8 BEHAVIORAL AND SOCIAL ASPECTS
OF REPRODUCTION................................189

Introduction...................................189
Methods........................................192
Results........................................195
Discussion.....................................229

9 SEXUAL MATURITY, REPRODUCTIVE CYCLE,
AND EGG AND CLUTCH CHARACTERISTICS.............241

Introduction...................................241
Methods........................................242
Results........................................244
Discussion.....................................272

10 NESTING ECOLOGY................................288

Introduction...................................288
Methods........................................290
Results........................................293
Discussion.....................................340

11 SUMMARY AND CONCLUSIONS........................356

LITERATURE CITED.....................................369

BIOGRAPICAL SKETCH...................................390


I


vii









Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy


ECOLOGY AND BEHAVIOR OF THE SPECTACLED CAIMAN (CAIMAN
CROCODILUS) IN THE CENTRAL VENEZUELAN LLANOS By


John B. Thorbjarnarson


May 1991


Chairman: Dr. F. Wayne King
Major Department: Forest Resources and Conservation

The ecology and behavior of the spectacled caiman was studied on a ranch in the central Venezuelan llanos over a five-year period. The estimated caiman population size over the 5,000 ha principal study site decreased from 3,265 in 1985 to 2,611 in 1989. The reasons for the decline were unclear but appear to be related to mortality or emigration among juveniles. Virtually all aspects of caiman ecology were affected by the strongly seasonal pattern of rainfall and water availability. Caiman densities reached an annual high during the dry season when only a few permanent water habitats were available, and in some cases exceeded 1,500 caiman/ha. Caiman dispersed during the wet season and crude wet season density values were 0.45-0.60 caiman/ha.

Movement and habitat use were studied using radiotelemetry. Patterns of dispersal varied greatly, but most caiman occupied wet season home ranges that were partially or completely distinct from dry season habitats. The


viii









maximum dispersal distance found was 6.5 km, and no significant differences in movement patterns of macrohabitat use were noted between the sexes. Adult caiman tended to use the same wet season habitat in consecutive years, but subadults appeared to wander more.

Nesting was a wet season phenomenon, although gonadal recrudescence and some reproductive behavior began in the late dry season. The caiman had a well developed system of social displays, involving vocal and non-vocal acoustic, visual, olfactory and tactile signals. Males established breeding territories in newly flooded habitats early in the wet season, with large males occupying the preferred habitats near the dry season lagoons. Overall, 55% of the nests produced young, and the leading cause of nest failure was depredation. Females and males both reached sexual maturity at an age of seven years, and at sizes of 60-65 cm SVL and 75-80 cm SVL, respectively. Female fecundity was positively correlated with size.

Caiman fed principally on Pomacea snails, fish and crabs. The diet changed seasonally and with ontogeny. Feeding behavior to capture fish was quite complex, but overall success rate was low.


ix














CHAPTER 1

INTRODUCTION



Except for changes in the anatomy of the palate and the vertebral structure, crocodilians have changed very little since the Jurassic period, nearly 200 million years ago (Steel 1973). Although crocodilians have exploited a number of rather specialized niches throughout their long evolutionary history (e.g., the terrestrial sebecosuchians or the thalattosuchian marine crocodiles), they have retained an inherent conservatism in their general morphology. While conservative in their evolution, crocodilians are not by any means primitive animals. Instead, this conservatism argues that crocodilians are supremely adapted to their particular lifestyle. Apart from the evolutionarily stable body form, the success of crocodilians is evident from their historical abundance in tropical and semi-tropical lowland aquatic ecosystems. However, during this century two crocodilian attributes, the quality and durability of their hides, and the reflective glow of their eyeshines, contributed to a worldwide decline in numbers as crocodilians were hunted commercially on a vast scale. Today, populations of crocodilians represent only a small fraction of their former numbers, with hunting


1





2


and rampant habitat destruction being the principal causes (Groombridge 1982).

The worldwide decline of crocodilians has created a need for the development and implementation of management programs, but at the same time has made it more difficult to obtain the requisite biological data upon which such programs should be based. Field studies are hindered by the fact that crocodilians are principally nocturnal, aquatic, wary of any human activity, and oftentimes only found in remote regions that are difficult to census or far from areas of logistic support. Additionally, studies of crocodilian population ecology are confronted by problems such as the large size attained by most species, and the resulting delayed maturity and long generation times. An adequate understanding of the behavior and ecology of many species of crocodilians is going to require the establishment of long-term research programs that will provide information on critical population parameters and how they change over time, and with respect to impinging biotic and abiotic factors.

Although most species of crocodilians are considered to be threatened, a few species have managed to escape the effects of human development. Some crocodilians, particularly the genus Paleosuchus, have bony hides of no commercial value and have not been extensively hunted (King and Brazaitis 1971). Other species, in certain areas, have actually benefitted from human activities, and are now more





3


numerous than they were in historical times. This has been the case with spectacled caiman (Caiman crocodilus) in the Venezuelan llanos. The human occupation of the llanos has resulted in the virtual extirpation of one the caiman's competitors, the Orinoco crocodile (Crocodylus intermedius), as well as the creation of large amounts of new dry season habitat in the form of borrow pits or cattle ponds. The result is that in parts of the llanos where they have not been hunted, spectacled caiman are found at very high densities. Given these circumstances, the study of the spectacled caiman in the llanos provides us with the rare opportunity to study the behavior and ecology of a species at or near its environmental carrying capacity. Investigations of caiman behavior and ecology in the Venezuelan llanos are also facilitated by a number of other factors. The extreme seasonality of the habitat forces the entire caiman population to concentrate annually in a few shallow bodies of water, making census work quick, accurate, and easily repeatable. These conditions, and the relatively small size of spectacled caiman (maximum male length 2.6-2.8 m), also expedite mark-recapture studies. Female caiman reproduce at a small size (ca. 1.2 m) and mature at a relatively young age for crocodilians (ca. 7 years), which reduces generation time and facilitates long-term population analyses.

The spectacled caiman is one of the most widely

distributed and geographically variable of the 23 extant





4


species of crocodilians. The species complex contains four or five described subspecies including g.c. chiapasius, distributed along the Pacific coast from Oaxaca, Mexico to Ecuador, and along the Atlantic from Honduras to the Golfo de Uraba, Colombia; g.g. fuscus, found east of the Golfo de Uraba through northwestern Venezuela; g.g. apaporiensis, restricted to the upper Rio Apaporis in the Colombian Amazon; the nominate subspecies C.c. crocodilus occupying the Orinoco drainage, most of the Amazon River system and the coastal region between the mouths of the two rivers; and the southern subspecies C.c. vacare, in the Paraguay River drainage and the Beni/Mamore/Guapore river system of lowland Bolivia. Caiman c. vacare has been considered a separate species by some authorities, most notably King and Burke (1989), who followed Medem (1981, 1983). Throughout its range, caiman are found in virtually all the available lowland wetlands habitats (Gorzula and Seijas 1989), and the adaptability of this species is attested to by the presence of thriving introduced populations in Florida, Cuba, and Puerto Rico (Ellis 1980, Groombridge 1982).

This study, initiated in 1984 in cooperation with Tomas Blohm and the Venezuelan Fundaci6n para la Defensa de la Naturaleza (FUDENA), was a broad scope investigation of the behavior and ecology of the spectacled caiman in the central Venezuelan llanos. The principal goal was to examine a number of important aspects of caiman ecology under near carrying capacity conditions, and to determine how these are





5



influenced by the extreme seasonal variation of water availability. Apart from the empirical interest in how these large reptiles adapt to the rigorous llanos environment, this study was designed to promulgate information that will be useful for the management of the species on a sustained yield basis.

Investigations of the life histories of long-lived

organisms are exceeding difficult but extremely important for the study of general life-history strategies (Wibur 1975). Recently, some long-term studies of crocodilian population ecology have begun, but principally on large subtropical (American alligator) or tropical (Estuarine crocodile) species. While a relatively large number of studies of caiman ecology and behavior have been conducted, most studies have concentrated on only a few aspects of caiman ecology, and others, while rich in detail and broad in scope, have frequently lacked quantitative backing. Medem's treatises on the South American crocodilians (Medem 1981, 1983), and Alvarez del Toro's work on Mexican species are good examples. The situation of the spectacled caiman in the Venezuelan llanos lends itself particularly well to the study of a number of important aspects of the species ecology and life-history. This study, conducted from October 1984 through May 1989, concentrated on aspects of caiman population characteristics, reproductive ecology, feeding ecology, habitat use, and activity and movement patterns. Apart from providing a detailed description of





6


the ecology and behavior of this population, one of the principal goals of this work was to examine life-history attributes of the spectacled caiman, in particular the relationship between body size and fecundity.

However, besides biological interest, the spectacled

caiman has also been the subject of much commercial interest in the Venezuelan llanos as well. An experimental harvest program was established in 1982 (Gorzula 1987, Thorbjarnarson, in press) and it has quickly grown into one of the largest crocodilian management programs in the world. Venezuela is currently producing approximately 100,000 legal caiman hides annually, making it the world's largest producer of crocodilian skins. Interest in caiman management has grown in a commensurate fashion and has far outstripped the quantity and quality of research on caiman ecology. At present, the Venezuelan government has no caiman research programs in effect and appears to be relying on what information has been determined from past studies, as well as the activities of non-governmental investigators.














CHAPTER 2

DESCRIPTION OF THE STUDY SITE



This study was carried out principally on Hato

Masaguaral, a cattle ranch in the central Venezuelan llanos (Guarico State: 8033'N, 67*37'W), approximately 50 km south of the town of Calabozo (Fig. 2-1). Some parts of the study were also conducted on adjacent ranches, particularly Hato Flores Moradas and Hato Matadero, and observations were collected on aspects of caiman ecology and behavior from a variety of other ranches throughout the Venezuelan llanos. Tomds Blohm, the owner of Hato Masaguaral, has a keen interest in wildlife conservation and has maintained the ranch as a wildlife sanctuary and as a site for biological research since the 1960's. A significant amount of biological research has been conducted on Masaguaral over the last 30 years, and over 100 scientific publications have resulted from this work. Although a diverse range of topics has been investigated, most research has concentrated on the ecology of the vertebrate fauna, including previous studies on the spectacled caiman (Staton and Dixon 1975, 1977, Marcellini 1979). Detailed information on the mammalian and bird fauna of the ranch has also been published (Eisenberg


7






8











y 3

Caribbean

























HaBtoi Adaseguara

Me4


WA



Coomi


Figure 2-1. Map of Venezuela indicating the location of the
Hato Masaguaral study site.




9


et al. 1979, Rudran 1979, Thomas 1979, Sunquist et al. 1989).


The Llanos Habitat


The llanos is a large (252,530 km2) geosyncline located between the Caribbean Cordilleras to the north, the Andes to the west, and the Guayanan Shield to the south. This region is best described as a hyperseasonal savanna intermixed with varying amounts of deciduous or semideciduous forest. The entire area is drained by the Orinoco River, whose tributaries cross the llanos principally in a north-south, or west-east direction. Situated over pre-Cambrian basement rocks, the llanos is composed primarily of alluvial deposits from the Tertiary and Quaternary periods. Most surface sediments are quite recent, associated with the Pleistocene uplift of the llanos region and erosional deposition from the Andes and Caribbean Cordilleras (Vila 1960).

The llanos can be divided into four basic subregions: the piedmont region adjacent to the Andes, the high plains, the alluvial overflow region, and the aeolian plains (Sarmiento 1983). Mountainous foothills, fast flowing rocky rivers and streams, and large alluvial fans characterize the piedmont region. Vegetation is principally savanna, with varying amounts of semi-deciduous tropical forest. This subregion is principally located along the base of the Andes and Coastal mountain ranges. The high plains, or upper llanos region, are also characterized by a relatively





10


significant amount of vertical relief, here dominated by mesas with dissected or undulating topography. Pittier (1942) first described this region as one of low plateaus 200-300 m above sea level, crossed by Orinoco tributaries carving deep river courses. This subregion is divided between two principal areas, the eastern llanos of Monagas, Anzoategui, northern Bolivar, and eastern Guarico states, and the llanos region south of the Rio Meta in Colombia. These two units are separated by a central tectonic depression (occupied by alluvial plains), and may be remnants of a formerly continuous uplands region from the Pliocene-early Pleistocene. The typical tree savanna vegetation is dominated almost exclusively by three species: Bowdichia virgiloides, Byrsonomia crassifolia, and Curatella americana (Sarmiento 1983).

The alluvial overflow plain, or lower llanos, is a vast region almost without vertical relief that occupies a depression in the central part of the llanos. The dominant vegetational association in this region is the hyperseasonal savanna, characterized by few trees or palm (Copernicia tectorum) savannas (Sarmiento 1983). In the wet season, the rivers in this region generally overflow, and most areas flood from this and heavy rainfall combined with poor surface drainage and soils of poor permeability. The western and more northern sections of this subregion are characterized by a greater occurrence of semideciduous forest cover.




11


The aeolian plains regions of the llanos extends from the upper Meta region in Colombia to the northeast, including the Cinaruco and Capanaparo river drainages, and apparently is a remnant of the formerly arid climatic period during the Wurm glaciation (Tricart 1974, Schubert 1988). This region contains extensive dune fields oriented in a northeast-southwest direction, situated atop sandy soils. Vegetation is characterized by open, virtually treeless savannas, with thin strips of gallery forest or small palmlined morichales (small streams lined with moriche palm, Mauritia minor).

The entire llanos region is climatically hyperseasonal with a well defined wet season (May-November) during which over 75% of the rainfall occurs. Total annual rainfall in the llanos region generally ranges from 1000-2000 mm, with rainfall increasing towards the west and south. A mean annual temperature of 26-28' C, results in a Tropical Dry Forest plant association, as defined by the Holdridge system (Ewel et al. 1976).



Hato Masaquaral

Hato Masaguaral is situated in a region of transition between upper and lower llanos, generally referred to as intermediate llanos (Berroteran 1985). The ranch covers approximately 8,500 ha of mixed savanna/decidous forest habitat just west of the Guarico River at an elevation of 60-75 m above sea level. The dominant vegetation types tend





12


to grade from a low-stature deciduous forest located in the east to an open palm savanna in the western portion of the ranch. A description of the vegetation types and dominant plant species on the ranch is given by Troth (1979).



Climate

On Hato Masaguaral the wet season begins in May-June and ends in November-December. Mean monthly precipitation peaks in July and August (Fig. 2-2). Because of variation in drainage characteristics the seasonal water level regimes of different habitat types differed widely (Fig. 2-3). Peak flooding tended to coincide with periods of maximum precipitation, but in some areas maximum water levels were found one to three months after the peak in rainfall.

The extreme seasonal variation in water availability is the key factor in shaping annual changes in the abiotic and biotic environment. The dry season is characterized by the widespread occurrence of fire. During the wet season extensive flooding occurs, and the llanos soils become waterlogged and anoxic. These environmental changes create a rigorous environment for both plant and animal communities and lead to seasonal changes in the ecology and phenology of the llanos fauna and flora.

Mean annual rainfall for the region ranges from 1351 mm (Calabozo: 1961-1984), to 1418 mm (San Fernando de Apure: 1921-1984). During the study period total annual rainfall ranged from a low of 1448 mm (1984) to 1632 mm (1986). In









Rainfall (mm)


Temperature 03)


Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Cec Month


400.0


Rainfall - Ta Max - Ta Min


Figure 2-2.


Mean monthly precipitation and minimum and maximum air temperatures. Rainfall data from Masaguaral during the study period (1984-1988), numbers indicate mean monthly values. Temperature data from Troth (1979).


Water Depth (cm)
200,


150k


50


'Cr
U-


Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

Figure 2-3. Mean monthly water depth at four locations
within the study site.


A


'k 17"


13


300.0



200.0 100.0


0.0


282.0

226.4 6---- 231.821. 162.0

107.9

49.3

1.8 2.1 4.5 13.8
' '~ ' ' ' ~ '


Guacimos
- Savanna Road S--- Highway Pond
-&- San Juanera


40.0 30.0



20.0



10.0 0.0


1001


I





14


contrast to the seasonal changes in rainfall, the annual temperature variation is relatively minor (Fig. 2-2). Mean daily maximum temperature only varies 4.8' C among months, and the corresponding value for mean daily minimum temperature is 3.4* C (Troth 1979). Minimum nighttime temperatures are experienced in December. The lowest temperature recorded during the study period was 15' C. Peak temperatures during the late dry season (March-April) would regularly exceed 370C.



Habitat Types

The principal study site consisted of approximately

4,000 ha of mixed deciduous forest and palm savanna located to the west of the highway, in the western section of the ranch. Caiman were also found in the eastern part of the ranch in riverine habitats (Rio Guarico, Cano Caracol), but these areas were not included in the study area. Within the study area, eight basic habitat types were defined: palm savanna, marsh, forest, bank, sandhill, lagoon, borrow pit, and streams. Habitats were discernable based on the degree of flooding, plant community structure, and soil types.



Palm savanna

Palm savanna dominated the western rim of the ranch, and was the most abundant habitat type, comprising 53% of the main study area (4,136 ha). This habitat was composed principally of palm bajio as described by Troth (1979), and





15


consists of low-lying areas which were flooded to a moderate depth during the wet season. The predominant tree was the palm (Copernicia tectorum), but a host of other trees and shrubs was frequently associated with the palm: strangler figs (Ficus spp.), a number of spiny shrubs and small trees (Annona spp., Randia venezuelensis, Zanthoxylum culantrillo), or other small woody species (e.g. Coccoloba caracasana). The herbaceous vegetation was dominated by grasses, particularly Panicum laxum and Leersia hexandra. Palm savannas ranged from open, park-like areas to dense shrubby habitats that graded into forested forests.



Marshes

The more deeply flooded areas of the savanna that

supported extensive rooted herbaceous vegetative growth were referred to as marshes, and comprised 27% of the main study area. Many marshes followed well demarcated, curvilinear paths that passed through the surrounding palm savanna habitat. On aerial photographs these marshes resembled old river courses, and may indeed represent former river or stream drainages. During the wet season flooding was extensive in marshes, which were the first savanna habitats to flood and the last to dry up. Depths in some marshes reached 1-1.5 m, but usually did not exceed 50 cm.

Plant communities in marshes underwent an extensive change in phenology during the year. Following the first rains a carpet of grasses and herbs, including "platanico,"




16


Thalia geniculata, began growing. Thalia grew particularly well in the deeper marsh holes where it would form dense colonies that extended 1.5-1.8 m above the water level by the end of the wet season. The marshes also supported a luxuriant growth of grasses and sedges (Paspalum repens, Paratheria prostata, Oryza perennis, Eleocharis elegans, E. mutata, E. minima and Cyperus flavus), floating vegetation (Neptunia olivacea, Ludwiqia helminthowbriza, Utricularia inflata, Eichhornia azurea, Naias sp., Salvinia sp., Pistia stratoides) and other rooted vegetation (Ipomea crassicaulis, I. fistulosa, Justicia laevilincquis).



Forests

Forests were areas dominated by a moderate to dense growth of trees and shrubs of medium height (10-15 m). Forests either were large, fairly contiguous patches, or were composed of smaller discrete patches surrounded by more open savanna. Troth (1979) provided a fairly complete list of the woody plants found in the forest habitats, but some of the more dominant species were the palm (Copernicia tectorum), the saman (Pithecelobium saman), the masag-uaro (Pithecelobium guachapale, from which the ranch derives its name), and the caruto (Genipa americana). The degree of wet season flooding was variable in the forests, but most areas experienced moderate inundation. Forests comprised 13% of the main study area.


I





17


Banks

Banks were elevated areas that did not flood during the rainy season. Most banks were long, narrow topographic features located along the edges of marshes, and apparently represent coarse grain depositional features of former water courses. Some trees were found on banks (mostly Copernicia tectorum), but most sites were defined by a characteristic herbaceous flora consisting of herbs (Sida acuta, Cassia tora, Stachytarphaeta mutablis, Wissadula periplocifolia), and grasses and sedges (Scleria muhlenbergii, Panicum laxum) that did not tolerate flooded conditions. Banks were a relatively minor habitat type on Masaguaral, comprising 2% of the main study area.



Sandhills

Sandhills were elevated sandy soil habitats that appear to represent old windblown sand deposits (Schubert 1988). Sandhills never flooded, had a gently undulating topography and extensive herbaceous growth dominated by grasses and sedges, and supported a number of small trees and shrub species (Troth 1979). Extensive growths of the spiny

recumbent shrub Mimosa pigra were found around the edges of sandhill habitats (especially along ecotones with marshes).

Sandhills were only located in the eastern section of the ranch, and comprised 4% of the main study area.





18


Lagoons

Lagoons were permanent water habitats with extensive areas free of herbaceous vegetation. The Guacimos Lagoon was the largest on Hato Masaguaral, measuring approximately 18 ha during the wet season. The Guacimos Lagoon and several others (Piscina, San Juanera) represented extensive, deep sections of marshes. Other lagoons, such as Alta Venegas and Merecure, were located in depressions in sandhill areas. Most lagoons were quite shallow (maximum wet season depth: 1.5 m), and were fringed with extensive vegetation similar to that found in marshes. Both the Guacimos Lagoon and the Piscina were artificially supplied with water during the dry season by subsurface water pumps.

Lagoons comprised 1% of the main study area.



Borrow pits

Borrow pits were anthropogenic habitats dug using heavy machinery, mostly in association with road construction. The hydroperiod of borrow pits varied greatly. Some were seasonal, and would dry completely during the the course of the dry season. However, if the excavation was deep enough borrow pits would retain water throughout the year. Borrow pits were located alongside the national highway that passes through the ranch, and adjacent two elevated dirt roads within the ranch. Borrow pits comprised less than 1% of the main study area.






19



Streams

Streams were small seasonal water drainages that

provided surface water runoff for the savannas. During the peak of the rainy season some streams would contain significant amounts of water and reach depths of up to 1 m. The width of streams rarely exceeded 3 m.



Windmill Ponds

Windmill ponds were small borrow pits or natural

depressions located adjacent to windmills. During the wet season these ponds were extensively flooded open water habitats. In the dry season, subsurface water pumped by the windmill maintained a small, shallow (<50 cm deep) body of water.


-7*















CHAPTER 3

CHARACTERISTICS OF THE CAIMAN POPULATION



Introduction

Throughout its distribution, the spectacled caiman

occupies an enormous range of habitat types. Differences among these habitats may exert a strong influence on many aspects of the behavior and ecology of caiman populations including movement patterns, feeding ecology, and the timing of reproduction. The nature of the habitat should also play a significant role in shaping the structure of the caiman population itself. The rigorous seasonal fluctuations of the llanos savanna ecosystem are assumed to play an important role in determining caiman population density, size-class structure, and growth rates. We would expect populations inhabiting such a harsh hyperseasonal environment to differ in many respects from populations in more equitable surroundings.

This investigation had three principal objectives: 1) estimate important population parameters including total population size, size-class distribution, sex ratio and growth rates, 2) quantify the seasonal variation in caiman density and biomass in different habitat types, and 3) examine the effectiveness of the nocturnal census technique


20





21


by estimating the percentage of the total population counted, and by comparing nocturnal with diurnal counts.

one of the principal objectives of this study was to

quantify important aspects of the caiman population on Hato Masaguaral, and compare these results with the findings of other studies. Other investigations of caiman population characteristics (density, size-class distribution, sex ratio) in the Venezuelan llanos have been conducted by Staton and Dixon (1975) and Ayarzagdena (1983). Data are also available for the ecologically similar Venezuelan Guyana (Gorzula 1978) and the Brazilian Pantanal (Schaller and Crawshaw 1982, Crawshaw 1987), and for less seasonal habitats such as the Brazilian Amazon (Magnusson 1982) and Suriname (Glastra 1983, Ouboter and Nanhow 1984). In contrast, caiman growth rate has been examined in wild populations in only a few studies (Gorzula 1987, Ouboter and Nanhoe 1987), and they have dealt only with the growth of juveniles. one previous estimate of juvenile growth rates in the llanos was done by Ayarzaguena (1983), but was based on assigned estimated ages derived from a size-frequency histogram.

These comparisons offer us the chance to see how

population parameters may vary between populations living in similar habitats and those occupying very different environments. The previous study of caiman on Hato Masaguaral (Staton and Dixon 1975) also provides the





22


opportunity to see how a population may change over a 10-15 year period.



Methods

Caiman censuses on Hato Masaguaral were based on nocturnal spotlight counts. Three types of periodic censuses were conducted during the course of this study: a weekly or biweekly census of the main study area, a monthly census of 23 borrow pits in the southern portion of the ranch (outside the main study area) during the dry season, and an annual survey of the entire ranch west of the highway. All counts were made using 4 v miners headlamps (ca. 40,000 candle power), or 200,000 candle power spotlights with a 12 v car or motorcycle battery as the power source. During the borrow pit census and the annual surveys, the size-class composition of the population was estimated by classifying caiman into one of four sizeclasses: class I-<20 cm snout-vent length, 11-20-59.9 cm SVL, 111-60-89.9 cm SVL and IV->89.9 cm SVL (AyarzagUena 1983).

Information on the size-class distribution and sex

ratio of caiman in the main study was also collected through a mark-recapture program. Caiman were captured by hand, by noosing with a locking wire noose mounted on the end of a

1.8-3.6 m long pole, or by seining.

Main study area census. A regular census route through the main study area was established in October 1984 and run





23


through May 1989. From October 1984 through May 1986 this census was repeated on a weekly basis. After May 1986 the census was done at biweekly intervals. The census utilized a 3.5 km transect through seasonally flooded savanna, plus counts at 10 permanent or seasonal lagoons. The transect was done from a slow moving vehicle driven along an elevated road that passed through a variety of savanna habitat types. For each caiman spotted the following data were collected: size-class, position along the transect, distance from the road, and habitat type. Censuses of lagoons estimated only population size, and were made from a standardized vantage point, except for the Guacimos Lagoon during the wet season which was done from a 4 m boat poled slowly around the lake perimeter.

During each census, water temperature, air temperature, wind speed, and water depth were measured at the Guacimos Lagoon. Water depth was also measured adjacent to permanent markers in the Highway Borrow Pit, the San Juanera Lagoon, and in a small roadside pool located along the transect route. Censuses were begun just after sunset and were usually finished by 23:00 h.

Borrow pit censuses. During the dry season a monthly census of 23 borrow pits (0.056-0.377 ha) was made along a

9.8 km road in the southern part of Hato Masaguaral. Censuses were conducted by slowly walking around each lagoon and tallying the number and size-class of caiman on a handheld microcassette recorder. In densely populated lagoons





24


three consecutive total counts were made and the mean value used. Water depth was measured at a standardized point in Borrow Pit 2.

Annual ranch census. The entire western portion of the ranch (5,500 ha) was censused on an annual basis over a three- to four-day period during the late dry season (earlymid April; 1985-1989). Total number and size-class distribution of caiman were estimated for all bodies of water. Small lagoons were censused as described above for borrow pits. The size-class distribution for the Guacimos lagoon was estimated from a small boat poled repeatedly through the lagoon. Censusing was begun shortly after sunset and was not continued past 24:00 h.

Repetitive count censuses. Repetitive counts were used to estimate the sighting fraction (the percentage of caiman visible) of caiman (Messel et al. 1981). At each body of water a total of 15 counts was made successively at two minute intervals using a 200,000 candle power spotlight. The sighting fraction (p) was estimated from the mean (m) and the standard deviation (a), using the following formula based on the binomial distribution: p=l-(a2/m).

Diurnal censuses. In the 1989 dry season two diurnal and one nocturnal censuses were made at a series of six borrow pits to examine the variation between daytime and nighttime censuses. Three of the lagoons had low caiman densities, and three had very high densities. Diurnal counts were conducted in the morning (08:00-09:30 h) and the





25


(14:10-15:20 h), and nocturnal counts were conducted from 19:10-22:10 h. Totals from the high density lagoons represent the mean of three counts.

Growth rate. Growth rate was determined based on the recapture of marked caiman. The minimum recapture interval used was 30 days. Because many caiman were missing the tips of their tails, measurements were based on snout-vent length. There was a great disparity in dry vs wet season growth, so growth rate was calculated for the wet season based on a model developed by Messel and Vorlicek (1989) where:

ASVL=a ATw + b ATd

ATw= growth interval, wet season ATd= growth interval, dry season

a= growth rate, wet season b= growth rate, dry season



Growth data from caiman recaptured during the same dry season (growth interval >30 days) were used to calculate b for each of three size-classes (10-30 cm SVL, 30-50 cm SVL, >50 cm SVL). From the date of initial capture and the date of recapture, the number of wet season (Tw) and dry season

(Td) days in the interval were calculated, and the equation could be solved for the wet season growth rate (a).





26


Results

Sighting Fraction

In 1986 repetitive counts were conducted in four lagoons and two borrow pits to estimate the sighting fraction. On 8, 9, 11 January, counts were made at four lagoons, producing a mean sighting fraction of 0.885 (Table 3-1). Counts made at one lagoon and two borrow pits throughout the night of 10 March 1986 indicate a somewhat lower value of 0.723 (Table 3-1). The overall mean value for all repetitive count censuses was 0.795 (a=0.173, N=27).

The four lagoons censused in early January varied in depth and the amount of fringing vegetation growing in shallow water. The two Piscina lagoons were shallow (<1 m deep), with very little vegetation. The Merecure lagoon was deep (>1 m) with some fringing vegetation, whereas the San Juanera lagoon was shallow with extensive fringing vegetation. The mean sighting fraction for the San Juanera lagoon was significantly lower than for the other lagoons

(LSD; p<0.05).



Table 3-1. Mean sighting fraction values for censuses in lagoons on Hato Masaguaral. January values are the means for three counts, March values are the means for five counts.
January 1986 March 1986
Lagoon P Lagoon P

Piscina East 0.900 Merecure 0.742
Piscina West 0.925 Borrow pit A 0.642
San Juanera 0.795 Borrow pit B 0 0.784
Merecure 0.918





27


The sighting fractions calculated on 10 March were

based on counts at two hour intervals throughout the night (20:00-04:00 h; Fig. 3-1). The mean sighting fraction for all lagoons remained relatively constant throughout the night except at 22:00 h when a sharp decline was found (0.435), resulting in a mean sighting fraction significantly below the other values (LSD; p<0.05). Diurnal Censuses

Far fewer caiman were counted during the day than at night (Table 3-2). Although the results were quite variable, no differences were noted between morning and afternoon counts, nor between high and low density lagoons. The overall mean correction value for diurnal/nocturnal counts was 0.48.

The estimated size-class distribution of the caiman

population also varied between diurnal and nocturnal counts. In four out of five cases, the proportion of class IV caiman was underestimated during the diurnal counts (Fig. 3-2), suggesting that class IV individuals have a greater tendency to remain hidden during the day than other size-classes.



Annual Ranch Census

The number of non-hatchling caiman (caiman >20 cm SVL) on Hato Masaguaral decreased during the study period (Fig. 3-3), with the greatest decline occurring between 1985 and 1986. Population size from 1986 to 1989 remained relatively

















1.00
0.90
0.80
0.70-


0.60-0.50-0.40-0.30-0.20-0.10
0.00 -


2 Merecure Prest. 2

0 \s.*Prest. 1


20:00 22:00 24:00 02:00 04:00 Hour


Figure 3-1. Sighting fraction of caiman and wind speed at
three lagoons recorded at two hour intervals
during the night of 10 March 1986.


I

)


28


C
0

L-

U
C

-c


20



-15
(I)
-10
(D


.53


- 0





29


Table 3-2. Diurnal (morning and afternoon) and nocturnal counts for six sites on Hato Masaguaral during January and February 1989. The last two columns present the ratio of morning and afternoon counts to nocturnal counts, respectively. High density values are the means of three counts.
Borrow Date AM PM Night AM/Night PM/Night
Pit Count Count Count

Low Density
2 Jan 0 0 3 -
Feb 7 7 8 0.88 0.88

8 Jan 1 1 5 0.20 0.20
Feb 7 5 8 0.88 0.63

17 Jan 1 1 3 0.33 0.33
Feb - - - -

18 Jan 2 2 8 0.25 0.25

Subtotal 0.57 0.51

High Density
9 Jan 47.7 42.0 121.0 0.39 0.35
Feb 48.5 73.0 118.3 0.41 0.62

18 Feb 24.3 33.3 91.3 0.27 0.36

19 Jan 61.5 55.3 98.7 0.62 0.56
Feb 41.8 105.0 123.3 0.34 0.85

Subtotal 0.41 0.55
























Figure 3-2. Comparison of observed caiman size-class
distribution in lagoons censused during the
morning, afternoon and at night on two separate
days in 1989.



















-V











AFTERNOON


1/25/89 n" PRESTAMO no

9 too


Lo0


I0 ~




PRESTAMO o

19 01)
'7

,oo



2/16/89 n*



PRESTAMO -oc

9 no

O

sI


PRESTAMO -ol

18 no.






Sao
600





PRESTAMO -"

19 .001

010


&w -Oasa


om.

no


IV'








I"O
40 0 8" 4"


o"


0


I


,oo,

ISao So






Go

@"Igas





too



soL ob
* U


0 g ~ 'm o. e e


.


no no LI


ra-osaa


IV


n*

'GGoo

* *
0,






too














40.


8 IV Inc myre


Wr Iwo
too





IVa 0 CTA
" 0U






00 .V


-


MORNING


NIGHT


m


0***


IV



















3000


_0
Q)
C


0

-0
E

z


2500+


2000-


1500-


1985


Figure 3-3.


1986


1987


1988


1989


Uncorrected annual census totals for nonhatchling caiman in the western section of Hato Masaguaral.


32


2612



2187
2127 2089
A A 015
A





33


constant. Corrected population size, based on a sighting fraction of 0.8, ranged from 3,265 in 1985 to a low of 2,519 in 1988.

The ranch population was found in four major dry-season habitat categories. Borrow pits contained 60% of the total non-hatchling population, and 32% were located in the main study area. A much smaller percentage (6%) of the population was in "other lagoons," which consisted of two small natural lagoons (Alta Venegas and Merecure). Windmill ponds contained 2% of the censused caimans. Because the main study area lagoons were artificially maintained by the pumping of subsurface water, the "other lagoons" were the only natural dry-season caiman habitat on Hato Masaguaral. Density and Biomass

The number of caiman in permanent water habitats, seasonal lagoons, and the seasonally flooded savannas followed a regular pattern associated with the annual rainfall pattern and the movement of the caiman. In seasonally inundated savannas, peak numbers of caiman were seen during the wet season (Fig. 3-4). Many of the areas bordering the savanna transect route contained pools that were among the first areas to flood and the last areas to dry up, so the observed caiman density was highest during the early and late wet season when caiman were dispersing from and returning to the permanent water lagoons. The dry



























Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec


Month


Figure 3-4.


0 0)



0


700


600-

500-

400 tI


300

200100

0


Seasonal trend in caiman census totals (mean SD) and water depth along a 3.5 km transect through seasonally flooded savannas.


200



-150 *

Q CD
-100



-50



0


Jon Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec


Month


Figure 3-5.


Seasonal trend in caiman census totals (mean SD) and water depth in the Guacimos Lagoon.


34


C

C


0

-0


0



0


30


20+


10+


0


T







0--0
III .


1 T T
T@ 0


T W

T T-00
T/


60


0


40 30

(D
20 -o 10
C)
3


T

0


. ,





35


season caiman observations along the transect were restricted to two windmill ponds.

The reverse trend in seasonal caiman density was seen

in permanent water lagoons, where peak numbers were observed during the dry season (Fig. 3-5 and 3-6). However, among permanent water lagoons, two patterns of seasonal population size were evident. In some lagoons (e.g. Guacimos Lagoon; Fig. 3-5), the peak number of caiman was reached during the mid dry season (February-March), whereas in other lagoons (e.g. Highway Borrow Pit; Fig. 3-6), the number of caiman tended to increase throughout the dry season. This variability may be attributed to differences in the hydroperiods of lagoons in surrounding areas. Seasonal lagoons such as the San Juanera (Fig. 3-7) contained large numbers of caiman during the early dry season, but as the lagoon dried caiman moved to other lagoons. The area surrounding the Guacimos lagoon tended to dry almost completely early in the dry season. However, the Highway Borrow Pit had a large number of deeply flooded marshes and seasonal lagoons in the area so caiman numbers continued increasing throughout the dry season. In actuality, the dynamics of a lagoons' dry season caiman population were directly related to the number, and the hydroperiods, of nearby lagoons.

Caiman density reached extremely high levels in some water bodies during the dry season, but tended to be inversely related to lagoon size. In the large Guacimos









C:
0
E
0
0

Q)
-Q
E
z
0


150-


100+


50+


0


IT
T e




0




S I I I
0~ 1


Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec


Month
Figure 3-6. Seasonal trend in caiman census totals (mean
SD) and water depth in the Highway Borrow Pit.


C
a
E
0
0

-0


~0
1


0


6050

40 30

20


10-


U


Jan


120 .80 i

(D


40 -o




0 1-


Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec


Figure 3-7.


Month
Seasonal trend in caiman census totals (mean SD) and water depth in the San Juanera Lagoon.


36


I


*~ 0

-T
0




T 0
1

.- -- -- - I'


200



-150 0

M
-100 -1
CD
-0

- 50 r



.0





37


Lagoon (dry season size ca. 12 ha), peak dry season density was 60.4 caiman/ha (April 1987), and the peak monthly fiveyear mean was 46.5 caiman/ha (March; Fig. 3-8). Much higher densities were observed in the 0.63 ha Highway Borrow Pit (Fig. 3-7), with a maximum density of 304.0 caiman/ha (May 1989), and a high five-year monthly average of 163.3 caiman/ha in May. These densities translated into mean biomass values of up to 408 kg/ha in the Guacimos lagoon and 1,655 kg/ha in the Highway Borrow Pit (Fig. 3-9). Peak monthly values were 530 kg/ha for the Guacimos lagoon (April 1987), and 3,082 kg/ha in the Highway Borrow Pit (May 1989).

The highest caiman densities and biomass values,

however, were observed in borrow pits in the southern part of the ranch. Density and biomass figures for 15 borrow pits (0.075-0.204 ha) were analyzed on a monthly basis during the dry season in four consecutive years. Five borrow pits were classified as high-density (mean=0.152 ha), and 10 were low-density (mean=0.102 ha). Lagoons located adjacent to natural savanna drainage systems (streams) were used by large numbers of caiman as the savanna water levels dropped. Low-density borrow pits were more isolated, and were usually occupied by a female with a pod of young, and frequently by one large adult male. Low-density borrow pits typically had densities of 20-80 caiman/ha, and biomass values of 100-300 kg/ha. High-density borrow pits reached extremely elevated density and biomass levels by the end of the dry season (Fig. 3-10 and 3-11). By April and May, mean










*-. Guacimos A-A Highway


A IA

TI _ I T T T T I


Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Month


Figure 3-8.


Mean seasonal trend ( SD) in caiman density in the Guacimos Lagoon and the Highway Borrow Pit.


AA



T1/ ] T T
ek-- -- L


0-0 Guacimos
A-A Highway










Aa
T


Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month
Figure 3-9. Mean seasonal trend ( SD) in caiman biomass in
the Guacimos Lagoon and the Highway Borrow Pit.


250


38


200-


150-


100+


C)
7


50-


0


\I
A
,f-&


2500


C





Co
0
E
0
M


2000-1500-1000-500 -


(~1


I


i





39


densities were 700-900 caiman/ha, with a high value of 1,619.7 caiman/ha in Borrow Pit 8 in May 1986. Peak mean biomass figures were in the 10,000-12,000 kg/ha range, with a maximum value of 23,379.1 kg/ha for Borrow Pit 19 in February 1986. Based on the calculated sighting fraction (0.8), and the assumption of equal sightability among sizeclasses, the estimated total density and biomass would be 20% higher in all cases.

Wet season density and biomass values were much lower, but harder to estimate. If wet season density is calculated by dividing the total number of caiman counted during the annual censuses by the western ranch surface area, I obtain values from 36.6-47.5 caiman/km2 for uncorrected counts and 45.8-59.4 caiman/km2 for counts corrected using the sighting fraction of 0.8. Biomass values are 398.7-448.5 kg/km2 (uncorrected) and 498.4-559.3 kg/km2 (corrected) (Table 33).




Monthly Borrow Pit Census

Monthly censuses of a group of 23 borrow pits located along a road in the southern section of the ranch revealed that the great majority of the caiman used only a small number of the available water bodies. Of the 23 available borrow pits, five (located adjacent to streams) were high density and supported 83-91% of the non-hatchling caiman










C)

C


C:

C
E
-


Figure 3-10.


0




0
0




c En
0

E
0



is
0
E
0
0


16

1412108

6

4

21

0


Figure 3-11.


Month
Mean seasonal trends ( SD) in caiman density in high- and low-density borrow pits.


0-0 High Density 0- Low Density








T OT
10
00



Nov Dec Jan Feb Mar Apr May Jun

Month
Mean seasonal trends ( SD) in caiman biomass in high- and low-density borrow pits.


1200


40


0- High Density 0-0 Low Density 0-0








N Jr

Nov Dec Jan Feb Mar Apr May Jun


1000+


800600

4002000-





41


Table 3-3. Estimated wet season density and biomass values for non-hatchling caiman on Hato Masaguaral, 1985-1989.
Year Total Density Biomass
(caiman/km2) (kg/kmn)

1985
Total Counted 2,612 47.5 404.3
Corrected Total 3,265 59.4 505.3

1986
Total Counted 2,187 39.8 448.5
Corrected Total 2,727 49.6 559.3

1987
Total Counted 2,127 38.7 398.7
Corrected Total 2,659 48.3 498.4

1988
Total Counted 2,015 36.6 409.7
Corrected Total 2,519 45.8 512.2

1989
Total Counted 2,089 38.0 415.6
Corrected Total 2,611 47.4 519.4





42


population (Table 3-4). The remaining 18 low-density borrow pits held a maximum of only 17% of the total non-hatchling caiman population.

The low-density borrow pits held a larger percentage of the total population (as opposed to the non-hatchling population) because these bodies of water were the principal "nursery" areas where hatchling caiman were found (Fig. 312). Typically, these low-density borrow pits were inhabited by an adult female with a pod of hatchlings. Many of these borrow pits were also shared by a single large male and a small number of juvenile caiman, particularly one- and two-year old individuals that would mix in with the pod of hatchlings.



Table 3-4. Percentage of the non-hatchling caiman population in high and low density borrow pits on Hato Masaquaral; 1985-1989.
Year High Density Low Density
1985-6 90 10
1986-7 83 17
1987-8 84 16
1988-9 91 9



Size-Class Distribution

The population size-class distribution was determined by estimating the size of caiman during the annual census (Fig. 3-13). A size-class breakdown of captured caiman is provided in Figure 3-14.





43


% Observed Population 60 M High Density MLow Denlty 50


40


30


2010


0

Calman Size-Class

Figure 3-12. Caiman size-class distribution in high- and
low-density borrow pits.








44


% Total Population 198



42.7 43.3





13.9
0878

)L
II Ill IV



819.1









17.1 17.8

8.98


1 11 Il IV


8MIS 1967










18.3
14.44 Q.2



II III IV




83.5 10"








18.7 17.9

8.88


Ill IV



1989


III


IV


Size-class


Figure 3-13. Annual caiman size-class distributions for the entire ranch; 1985-1988.


















Number of Calman


-Males E Females


> \





Size-class (cm)


IV


Figure 3-14. Size-class distribution of captured caiman
over 20 cm SVL (N=634).


45


350 300 250

200

150 100 50

0


11





46


The caiman population was dominated by a large

percentage of class III individuals (59.6% overall). Large adult males over 90 cm SVL comprised 16.3% of the population. The annual differences in size-class distribution were relatively small except in 1985, when a large percentage of class II individuals was seen.

There was a great deal of variation in the caiman sizeclass distribution among the major habitat categories (Fig. 3-15). The main study area lagoons tended to have a higher percentage of class II individuals and fewer class IV males than did the borrow pits. The study area lagoons also had very few hatchling caiman at the end of the dry season. Windmill ponds had few caiman, but a large percentage were hatchling (class I) and juvenile (class II) animals.

The size-class breakdown of captured non-hatchling caiman is significantly different from the size-class distribution of censused caiman (chi-square; p<0.001). However, as the vast majority of the captured caiman came from the main study area, it would be more appropriate to compare the size-class distribution of captured caiman with caiman censused from the main study area. A chi-square analysis, however, indicates that the two are significantly different (p<0.001). This discrepancy suggests that either the size-estimation procedure was biased (classifying class II caiman as class III), or that there was a bias in the capture procedure towards small animals. I suggest that the latter is the principal reason for the difference for two
























Figure 3-15. Mean size-class distribution of caiman in the
four major dry season habitat types. N is the
mean number of non-hatchling caiman censused
for each habitat.










% Total Population


KA 00


11 111 IV
Size-class


% Total Population Main Study Area N-709


% Total Population


% Total Population


Other Lagoons
N-121


AR~ OR


III IV


I-


40 20


III


SIze-cla3s3


60



40 20


Borrow Pits N-1323


801-


40


56.32


201


1.42

1


01


11


Il


Size -class


6OF


Iv


40 20


0


Windmills
N-34












0.8
Iv


SI zL-Glass





49


reasons. Firstly, although misclassifications of size-class were made, errors were not great. Size classification was made based on the experience of hundreds of captures of individuals. Also, it was my impression that I was just as likely to misclassify class III animals as class II, as the other way around. Secondly, a size-bias in captures is quite probable given the fact that small caiman were more easily captured than larger ones.

This suggests that the sample of captured individuals is probably biased towards smaller caiman and the census size-class distribution more accurately reflects the population size-class composition.



Sex Ratio

Based on a sample of 634 captured caiman, the sex ratio (male:female) was 1.10:1, and was not significantly different from a 1:1 ratio (chi-square; p>0.05). However, when broken down by size-class, class III was significantly biased towards females (p<0.001), and class IV was composed entirely of males. Class II caiman had a 1:1 sex ratio (Table 3-5).


Growth Rates

Growth rates of caiman were markedly affected by

season. During the dry season caiman grew very slowly or shrank. Dry season (January-May) growth rates for caiman in the 10-30 cm SVL range averaged 0.004 cm SVL/day (SD=0.005,





50


Table 3-5. Sex ratio of captured caiman by size-class. Size-class Females Males Total


II 162 167 336
(20-59.9 cm SVL)

III 134 91 225
(60-89.9 cm SVL)

IV 0 69 69
(>90 cm SVL)



N=5). For larger individuals, dry season growth was negative: 30-50 cm SVL=-0.009 cm SVL/day (SD=0.027, N=5), >50 cm SVL=-0.022 cm SVL/day (SD=0.022, N=3). These values were used as best estimates of average dry season growth rates, and were utilized to calculate wet season growth (Messel and Vorlicek 1989).

Wet season growth rates (June-December) were

significantly greater than dry season growth. Overall, growth rate did not differ between the sexes (ANCOVA, F1,1,101=2.40; p>0.05). Mean male growth rate was 0.032 cm SVL/day and mean female growth rate was 0.039 cm SVL/day. However, sexual differences in growth rate were noted within certain size-classes. Among the smaller size-classes growth did not differ between males and females (Fig. 3-16), but with increasing size, growth rate in females decreased more rapidly than in males. Female growth was significantly less in the 50-70 cm SVL class (F1,11=8.48; p<0.05) and the 70-90 cm SVL class (F1,19=41.09; p<0.001). The size-class and sex





51


>, 0.120-0 0 0-0 Males

.-- Females >1 0.100 27
)
E 0.080-0.060-- 9
o 5

C 0.040 -- 5
b~ T
0.020 -- 1 8
o 16 T 3
0 o*-o~
0.000 ,1i^
10-30 30-50 50-70 70-90 90-110 )110 Size-Class (cm SVL)
Figure 3-16. Mean growth rates ( SD) for male and female
caiman by size-class.



Snout-Vent Length (cm)
120
Males -4- Females
100


80


60


40


20


0
0 5 10 15 20 25
Age (years)

Figure 3-17. Average growth curves of male and female
caiman calculated out to 20 years. Zero growth during the dry season is assumed.





52


specific growth data were used to construct growth curves (Fig. 3-17).



Discussion

Caiman on Hato Masaguaral grew very slowly. Up to the age of approximately four years, both males and females grew at a rate of 0.04-0.06 cm/day during the wet season (JuneDecember). Very little growth occured during the dry season (January-May), a pattern that has been noted in other populations of caiman (Gorzula 1978) and other crocodilian species (Magnusson and Taylor 1981, Webb et al 1983a, Messel and Vorlicek 1989). Over the 214-day wet season, annual growth increments of juvenile caiman averaged 8.5-12.8 cm SVL, but there was a large amount of variation among individuals. Female growth rate slowed considerably when snout-vent lengths exceeded 50 cm, and may be related to the commencement of energy investment in reproduction. Male growth decreased at a slower rate so at sizes above 50 cm SVL males grew significantly faster. These growth rates suggest that, on average, females become sexually mature when they are 6-8 years old. Males may attain physiological sexual maturity at 7-8 years of age (75-80 cm SVL), but due to dominance related factors probably do not reproduce until they are 10+ years old (>90 cm SVL)(chapter 8).

Despite the large number of studies of caiman in the llanos, few data on caiman growth in the wild have been published. A growth curve for juvenile caiman (0-4 years





53


old) based on estimates of age derived from a size-frequency plot of captured juveniles on Hato El Frio in Apure state, indicates that these caiman are growing approximately 18% slower than on Masaguaral (Ayarzagdena 1983). Growth rates in the savannas of the Venezuelan Guyana (Gorzula 1978) are comparable to those from El Frio, but there was little seasonal variation in the growth of 1-2 year olds.

In Suriname juvenile caiman inhabiting a river bordered by swamp forest exhibited much faster growth (Ouboter and Nanhoe 1984). Over the first year of life, growth rates averaged 2.7 cm total length/month, equivalent to 5 cm/day over the entire year, approximately 31% faster than the annual growth rates of one year olds on Masaguaral. In Suriname seasonal variation is more complex, with a long and short dry season, but the dry season is less stressful than in the Venezuelan llanos. Growth was greatest during the long dry season and the long wet season, perhaps because food was not limited (Ouboter and Nanhoe 1984). Limited food availability has also been suggested as a causative factor in the low growth rates of American alligators in the Florida Everglades (Jacobsen and Kushlan 1989).

The negative growth rates exhibited by caiman during

the dry season are quite interesting. During the course of the wet season it was not uncommon for caiman to lose weight, but a shrinkage in length indicates that, in the absence of a reduction in skeletal elements or connective tissue, that shrinkage may be a natural result of loss of





54


condition resulting in the tighter packing of skeletal elements in the vertebral column.

Censusing crocodilians is frequently a very difficult endeavor. Visibility bias is especially evident in wetland habitats dominated by herbaceous and woody vegetation. However, many of the factors contributing to visibility bias are minimized when, as on Hato Masaguaral, virtually all the animals are restricted to a few small, shallow bodies of water. The concentration of caiman in a few bodies of water during the late dry season, in concert with the high sighting fraction, provides ideal conditions for censusing.

The mean sighting fraction (p) derived from repetitive counts (0.8) was similar to, but slightly lower than, empirical estimates made by Staton and Dixon (1975) for the same ranch (0.9). The percent of the total population visible is greatly affected by the type of habitat. In some parts of the llanos caiman occupy bodies of water completely covered by water hyacinth (Eichhornia sp.), or other floating vegetation that effectively hides a large part of the population from view. The lagoon with the largest amount of fringing vegetation on Masaguaral had the smallest sighting fraction, but values were high at all areas examined. The one other published study using repetitive counts reported somewhat lower p values (0.6-0.7 for juvenile C. porosus), and that the sighting fraction varied significantly among size-classes (Messel et al. 1981). More recent studies estimating the sighting fraction among





55


crocodiles have relied principally on mark-recapture techniques and have produced lower estimates of p: C. porosus-0.35-0.66 (Bayliss et al. 1986), C. niloticus-0.36 (Hutton and Woolhouse 1989).

The comparison of diurnal and nocturnal counts revealed that diurnal counts were poor estimators of total population size. The mean correction factor (diurnal/noctural counts) from this study was 0.48, well above the 0.30 value found by Seijas (1984) for 14 ranches in Apure, Barinas and Portuguesa states. Diurnal counts not only revealed a small percentage of the total population, but they had high sample variances (this study-0.47; Seijas (1984)-0.32), and were biased against the large size-class IV individuals. Taken together, these data strongly suggest that census work should be based on nocturnal survey work, and not diurnal surveys using correction factors.

The study on Hato Masaguaral was conducted during a period of declining population size. The total ranch population decreased 20% over the five year study period. Over 80% of this decline occurred during the interval between 1985 and 1986, and may be attributed to either mortality or emigration. After 1986 the population became relatively stable, although a slight downward trend was apparent (Fig. 3-3). From 1985 to 1986 the observed ranch population declined overall by an estimated 425 individuals. During this period the number of size-class II caiman dropped by 741, and the number of SC III caiman increased by





56


293. Even assuming that the increase in SC III caiman was entirely due to recruitment from SC II individuals, there was a loss of 448 SC II caiman. Clearly, the dramatic decline in numbers of caiman during the 1985-6 period was due to loss of SC II caiman. Because of their small size, juvenile caiman are more susceptible to predation than are larger caiman, but juvenile and subadult crocodilians may also enter a dispersal phase (Messel et al. 1981, Hutton 1989). However, based on the data available it cannot be determined which of these two factors was most important.

Caiman density on Hato Masaguaral was extremely high

during the dry season, with peak values in some borrow pits exceeding 1,000/ha. These density values reported here are comparable to figures presented by Staton and Dixon (1975) and Marcellini (1979), also for Hato Masagural, but are somewhat lower than the 3,000-4,000/ha reported by Marcellini (1979) for two ponds in Hato El Frio, in the low llanos. Extremely high dry season population densities appear to characterize many regions of the Venezuelan llanos, especially on ranches where the fauna historically has been protected.

However, these dry season density values represent temporary concentrations of caiman. When considering overall values of population density and biomass, wet season values must also be evaluated. Because of the large seasonal amplitude in density and biomass, it is difficult





57


to compare the results of this study with similar values obtained for more equitable environments.

Wet season caiman biomass figures tended to be 20-200 times less than dry season values, and were about one order of magnitude below those reported for herbivorous turtles, and half an order of magnitude below average values for carnivorous turtles (Iverson 1982). Caiman wet season biomass was five times greater than estimated values for the sympatric llanos pond turtle Podocnemis vogli (Iverson 1982). Caiman biomass is also much higher than biomass figures for any species of mammal occurring in the same habitat (west side of Hato Masaguaral: Eisenberg et al. 1979). In fact, the combined crude biomass figure for all non-volant mammals in the study area (478.8 kg/km2) was still below the mean corrected biomass for caiman (518.8 kg/km2).

Historically, caiman populations in the llanos were

much smaller. Two factors led to significant increases in caiman numbers in the 20th century. The first was the near extirpation of the Orinoco crocodile (Crocodylus intermedius), which was once a common faunal element in the rivers and streams throughout the llanos. The over exploitation of crocodiles for their hides left nearly vacant the river habitat niche, which was swiftly filled by spectacled caiman. The other factor was the alteration of the llanos landscape by humans. Activities associated with cattle ranching and road construction have greatly increased





58


the availability of dry season wetland habitats (mostly borrow pits). On Masaguaral, the two principal dry season lagoons were artificially maintained by pumping sub-surface water. Only 6% of the Masaguaral caiman population during the dry season was located in natural lagoons, suggesting that human alteration of the habitat has increased carrying capacity 10-20 times. In many parts of the llanos caiman migrate to rivers during the dry season. In the Guarico River, bordering the eastern limit of Masaguaral, wet season caiman density is as low as 1.24 caiman/km during the wet season, and as high as 24.86/km in the dry season (Thorbjarnarson and Hernandez, in press). However, prior to the 1930's this river had a large crocodile population, thus presumably depressing caiman density. Overall, caiman populations in the llanos have greatly benefited from the activities of humans during the 20th century.

The only census data that exist from the mid-1970's are for the Guacimos Lagoon (Staton and Dixon 1975), and suggest that total population size in the lagoon had increased from a peak of 280 in 1974, to a mean high value of 574 during the period 1985-1989. However, as no data are available from other lagoons in the area it cannot be stated with certainty that this doubling in population size represents a accurate estimate of population increase. Nevertheless, these data strongly suggest that the Hato Masaguaral caiman population has grown since the study of Staton and Dixon (1975). Although population size has increased, other





59


aspects of the population have remained relatively constant. The sex ratio of caiman determined in this study, 1.10:1 males to females, was exactly the same as that determined by Staton and Dixon (1975) for Hato Masaguaral 10-15 years previously.

The size-class distribution of caiman in this study is very similar to the one found by Staton and Dixon (1975) for the same ranch some 10 years earlier. Both showed the population dominated by size-class III caiman: adult females and subadult and small adult males. Size-class II caiman, or juvenile males and females approximately 2-5 years of age, constitute only about 20% of the total non-hatchling population. Adult males (SC IV) comprise approximately 15% of the population. Other studies of caiman population sizestructure have generally found larger numbers of juveniles (review in Gorzula and Seijas 1989). This is especially true in some of the less seasonal habitats such as the Brazilian Amazon (Magnusson 1982), and the Coesewijne River in northern Suriname (Ouboter and Nanhoe 1984).

The small number of juveniles on Masaguaral suggests that hatchling mortality, and hence recruitment into the juvenile size-class, is extremely low. Hatchling mortality on Masaguaral is known to be extremely high (>95%, Thorbjarnarson, unpublished data), and may be the factor limiting population size. However, analysis of the population size-structure in different parts of the ranch (Fig. 3-16) reveals that in areas where hatchling survival





60


is the highest (borrow pits), the proportion of SC II caiman is the lowest. The size-class structure of a population is an extremely dynamic parameter, and depends on size-specific mortality, growth, and dispersal rates. A more in-depth analysis of the significance of the population sizestructure will have to await further data on these topics.

Nevertheless, the observed difference between the population size-class distribution derived from the population censuses, and the size-class distribution of captured caiman is worth noting. This difference is attributed to capture techniques that are biased toward smaller animals. This indicates that unless unbiased capture techniques are used (e.g. a large seine net), care should be taken in interpreting the size-class distribution of captured individuals as a reflection of the true population size-class distribution, which has been a common approach used in studies of caiman (Staton and Dixon 1975, Gorzula 1978, Ayarzaguena 1983, Ouboter and Nanhoe 1984).















CHAPTER 4

MOVEMENT PATTERNS, HOME RANGE SIZE, AND HABITAT UTILIZATION



Introduction

Movement patterns and habitat selection, and their

variability among size-classes and between the sexes, are important components of a species' life history. Despite the importance of these topics, relatively little detailed research on crocodilians has examined these factors. By far the best studied species has been the American alligator (Alligator mississippiensis; Joanen and McNease 1970, 1972, McNease and Joanen 1974, Taylor et al. 1976, Goodwin and Marion 1979, Taylor 1984), although recent work has been done on Crocodylus niloticus (Hutton 1989), C. porosus (Webb and Messel 1978, Webb et al. 1983), and C. iohnsoni (Webb et al. 1983b).

The goal of this investigation was to study broad

aspects of movement patterns of a large number of adult and subadult caiman on a relatively long term basis. Specifically, the effects of caiman size and sex on seasonal movement patterns, home range size, and macro and microhabitat selection were examined. Radio-telemetry is an excellent method for efficiently following large numbers of animals on a long term basis, but among crocodilians this


61





62


technique has only been used on American alligators (Joanen and McNease 1970, 1972, McNease and Joanen 1974, Taylor et al. 1976, Goodwin and Marion 1979, Taylor 1984) and Nile crocodiles (Hutton 1989). With spectacled caimans, habitat use and movement have been investigated principally based on mark-recapture studies (Gorzula 1978, Schaller and Crawshaw 1982, Ouboter and Nanhoe 1988), although the study of Ouboter and Nanhoe (1988) did include some information from radio telemetry. This is the first time that a large scale telemetry study of Caiman crocodilus has been done, and is also the first for any crocodilian inhabiting a environment with extreme hydric fluctuations.



Materials and Methods

The movement patterns and habitat utilization of 39 adult and subadult caiman were studied using radio telemetry. Initially, commercially purchased 3.9 v Lonner Modules (AVM Instrument Co., LTD.) in the 164-165 MHz frequency range were used. These modules contained an SB2 transmitter and a C or D-cell lithium battery power source with a 30 cm flexible whip antenna. The larger D-cell modules (9.6 cm x 4.3 cm; 220 g; N=7) were used for adult males, and the smaller C-cell modules (8.9 cm x 3.0 cm; 110 g; N=15) for adult females and sub-adult males. In 1988 caiman were fitted with two-stage oscillating crystal controlled transmitters (165-166 MHz) made in Gainesville, Florida, by Debbie Wright, a University of Florida graduate





63


student. Ten transmitters utilized a C-cell lithium battery (8 cm x 3 cm; 100 g), and five were powered by AA-cell lithium batteries (8 cm x 2 cm; 50 g). All transmitters had 30 cm whip antennae.

Transmitters were attached to the dorsal surface of the tail immediately anterior to the junction of the double row of caudal crests with the single row (Fig. 4-1). The transmitters were fixed to the tail with 3-5 segments of 150 lb test monofilament fishing line which were sewn through the cartilage of the double caudal crests, and then wrapped repeatedly around the transmitter. The transmitter was completely encased in a wrapped layer of monofilament whose free ends were fused by partially melting the exposed monofilament strands. The entire transmitter/monofilament module was then encased in a thin protective layer of dental acrylic. The antenna was run posteriorly along the row of single caudal crests through a series of short monofilament loops sewn into the tail cartilage.

Caiman fitted with transmitters were tracked using a Telonics TR-2 receiver and an RA-2A "H" antenna. Due to signal attenuation caused by water or dense vegetation, in many cases the signal could only be heard initially from an elevated location (i.e. by climbing a tree or windmill). Tracking was done on foot, horseback, or from a vehicle. During the wet season all caiman not in permanent water habitats were located by following the signal to its source. When the caiman was located, a series of environmental






64


Figure 4-1. Position of radio transmitter on the dorsal
surface of the tail of a caiman, prior to being
covered with dental acrylic.





65


measurements were made: air (shade) temperature, water temperature (approx. 5 cm depth), water clarity (depth to which the tip of a white ruler was visible), and water depth. The location of the caiman was classified into one of seven macrohabitat types (bank, palm savanna, forest, marsh, sandhill, lagoon, borrow pit, and stream; see habitat descriptions in chapter 2). The extent of macrohabitats in the study area was determined by weighing the different habitat types cut from a habitat map drawn from aerial photographs.

Additionally, the microhabitat of the caiman was classified into one of 8 categories:



1) On land in open (OLO): caiman out of the water

in the open, not hidden in vegetation.

2) On land in vegetation (OLV): out of the water

and partially or wholly hidden under vegetation.

3) Buried (B): fully or partially buried in the

substrate.

4) In open water (OW): in open water without

vegetation in the immediate vicinity.

5) In water in vegetation (IWIV): in water among

rooted herbaceous vegetation.

6) Under floating vegetation (UFV): in water in or

under a layer of floating vegetation.

7) Among trunks/branches (ATB): in the water in or





66


among live or dead tree/shrub trunks or

branches.

8) In Thalia colony (DTC): in the water, in a dense

colonial growth of Thalia geniculata, a tall,

rooted aquatic plant restricted to deep areas of

marshes.



The position of caiman located in permanent water bodies was fixed by triangulation. Normally, no environmental data were collected at these sites because caiman would usually move away before their original location could be determined.

The position of all "fixes" (both direct localizations and triangulations) was marked on maps of the study site. For each wet season fix a "dispersal distance" was calculated by measuring the straight-line distance from the caiman to the permanent water body from which the animal dispersed. The term "dispersal" is used throughout this study to refer to the seasonal movement of caiman away from the dry season lagoons. Similarly, a "previous distance" was measured between successive fixes. All measurements were made using 1:11,000 scale aerial photographs of the study site.

Home range size was calculated using a program for microcomputers (McPaal) developed by M. Stuwe and C.E. Blohowiak at the National Zoological Park, Smithsonian Institution. Caiman locations were assigned x and y





67


coordinates using a transparent grid overlay. Analyses were performed for dry season and wet season home ranges. Only data from animals tracked over an entire dry/wet season period were used. Calculation of home range size was based on the minimum convex polygon method (Eddy 1977).



Results

Due to the large number of animals that had to be

located, and the difficulty of tracking the animals at night (as well as other project demands that required nocturnal work), most tracking was done during the day (Fig. 4-2). Although this method was not able to record nocturnal movements that returned to the same point, diurnal fixes did provide a good overall quantification of directional movements (Hutton 1989).



Patterns of Movement

Twenty-two transmitters were attached to caiman between 27 November 1986 and 28 July 1987 (Table 4-1). In June-July 1988 another 15 caiman were equipped with transmitters. Two additional caiman were fitted with transmitters that had fallen off other caiman, bringing the total number of animals in the study to 39. Two caiman were not included in the data analyses due to equipment malfunction within 30 days of release. Of the remaining 37 caiman, 16 were males and 21 were females. Males ranged from 58.4 cm to 127.8 cm SVL; females from 61.0 cm to 82.5 cm SVL (Fig. 4-3).






68


Table 4-1. Caiman equipped with radio transmitters during the course of the study. An * indicates individuals not included in the data analyses.
Caiman Sex SVL Dates Followed # Days # Fixes

008 M 127.8 19 Mar 87-5 Jun 88 443 127
036 F 80.5 17 May 87-14 Jan 89 607 99
060 F 80.1 17 May 87-15 Feb 89 636 104
094 M 99.1 23 May 87-30 Nov 89 191 31
095 M 96.2 9 May 87-10 Oct 88 519 85
117 M 102.1 28 Aug 87-17 Jul 88 323 58
118 M 107.2 6 Apr 87-19 Jan 88 288 49
163 M 58.4 4 Jun 88-20 Jul 88 46 9
254 F 74.5 19 Mar 87-19 Jan 88 306 85
271 F 81.5 21 May 87-10 May 89 719 139
298 M 113.5 27 Nov 86-10 Aug 87 255 48
312 F 78.2 31 Jan 87-10 Jul 87 161 36
358 F 75.0 6 Apr 87-19 Nov 87 227 44
359 F 79.0 6 Apr 87-10 Nov 87 218 38
361 F 75.1 7 Apr 87-17 Aug 88 497 105
362 F 74.0 6 Apr 87-31 Oct 87 208 32
376 M 115.4 9 May 87-6 Jan 89 605 100
377 F 70.1 9 May 87-13 Nov 87 188 32
378 F 82.5 9 May 87-5 Aug 88 453 83
379 M 77.0 17 May 87-17 Aug 88 457 71
382 M 79.5 21 May 87-10 May 89 719 111
384 F 74.5 23 May 87-3 Aug 88 437 76
386 F 61.0 3 Jun 88-9 Sept 88 98 33
395 M 83.4 28 Jul 87-27 Nov 88 487 78
479 M 94.5 5 Jun 88-10 May 89 339 49
480 M 102.3 6 Jun 88-3 May 89 331 62
484 F 66.5 8 Jun 88-2 Aug 88 55 10
488 F 70.0 9 Jun 88-29 Nov 88 173 21
489 M 59.0 9 Jun 88-5 July 88 26* 3
490 F 74.5 10 Jun 88-20 Jun 88 10* 3
491 M 106.3 10 Jun 88-10 May 89 334 55
492 F 75.0 11 Jun 88-13 Jan 89 216 31
493 F 80.0 13 Jun 88-23 Jan 89 224 30
495 M 119.5 13 Jun 88-10 May 89 331 63
497 F 61.7 19 Jun 88-20 July 88 31 9
500 F 81.0 26 Jun 88-10 May 89 318 60
505 F 75.0 18 July 88-5 Nov 88 110 21
508 M 121.0 11 Aug 88-28 May 89 290 61
513 F 79.4 17 Sept 88-28 May 89 253 44










No, Fixes 350 N-2146 fixes 3001-


0 1 2 3 4 5 6


15 16 17 18 19 20 21 2223


7 8 9 0 u1 12 13 14
Hour


Figure 4-2.


Time of radio fixes for caiman throughout the study period (Nov 1986-May 1989).


Number of Caiman


= Males O Females


50


75


I ..INI


100


Figure 4-3.


Snout-Vent Length Class (cm)


Size-class distribution of caiman equipped with radio transmitters.


69


250

200150100-


50

0


8



6



4



2-


U


I


125


_


_


10


I .i.





70


Complete wet-dry season cycles were followed during the 1987-8 and the 1988-9 seasons. The 37 caiman were monitored for a mean period of 327 days each (range 31-719 days). Of the 39 transmitters fitted, 22 dropped off and were recovered, 10 were lost, presumably after transmitter malfunction, and seven were still operational at the end of the study. It is not believed that any of the lost signals were caused by dispersal of caiman from the study site. An attempt was made to locate lost caiman by searching a radius of approximately 10 km around the study site from a fixed-wing aircraft (Cessna 182) in August 1988.

Weak, intermittent signals from two transmitters were received over the study site but could not be accurately located. The one failed transmitter from 1987 was later recovered when the caiman was recaptured near its original point of capture.


Dispersal from dry-season lagoons

Caiman typically remained in one body of water

throughout the dry season, although in some cases movement between lagoons occurred in response to dropping water levels. Dispersal from the dry season lagoons was associated with the annual rains and the flooding of the savannas. The late onset of the rains in 1988 delayed dispersal that year by a little over one month (Table 4-2). No significant difference in the mean dispersal date between sexes was found for either year (F1,34=0.12; p>0.05).





71


Table 4-2. Mean dates of dispersal from dry-season lagoons.
1987 1988
Mean SD Mean SD
(Days) (Days)

Females 24 May 7.8 30 June 12.2

Males 20 May 5.9 2 July 9.0


The great majority of caiman dispersed only after

savanna flooding had reached an adequate level. However, certain individuals dispersed from the dry season lagoons before significant flooding of the savannas occurred. In these instances the caiman sought cover under low-hanging vegetation, especially under small palms (Copernicia tectorum) near the marsh habitats that are the first areas to flood. Most early dispersers did not move more than 0.5 km from the body of water, but in 1987 caiman #95 (male; 95.8 cm SVL) dispersed 4 km from the Guacimos Lagoon and spent one week hiding in vegetation until the flooding of a nearby marsh created wetland habitat.

No significant effects of size on date of dispersal

were found for either males (F3,2=0.49; p=0.72) or females (F1,8=3.10; p=0.11).



Dispersal distance

Annual Dispersal Pattern. Dispersal distance is a

measure of the straight-line distance between the caiman and the water body where that individual spent the previous dry season, and is indicative of how far the caiman disperse during the course of the wet season. Despite the late





72


dispersal in 1988, the peak dispersal distances for both years were achieved from July to September (Fig. 4.4). Mean peak dispersal distance during this time was 0.8-1.3 km. The maximum straight-line dispersal distance was 4.4 km in 1987, and 6.5 km in 1988 (both females).

There is an annual dry-wet season pattern of dispersal in relation to permanent water sites (Table 4-3). Even during peak periods of dispersion (July-September for both years) a large fraction of the caiman remained within one kilometer of their dry-season refuge. Only in October 1987 does the percentage of caiman located within 0.5 km of their dry-season lagoon drop below 30%. In most peak wet-season months 40-50% of the caiman have dispersed a straight-line distance less than 0.5 km.

In the latter half of the wet season caiman began moving back towards the permanent water sites. Some individuals returned to the lagoons as early as OctoberNovember, but others remained in drying marsh habitats until January, or even early February.

Factors affecting dispersal distance. Differences in dry season dispersal distance (January-April) principally stem from a few caiman that did not return to the same dryseason body of water as in the previous year. A two-way ANCOVA analysis (using month as the covariate) of the 1987/8 wet season dispersal distance found no significant differences between males and females, but a significant year effect (F1,1521=6.68, p<0.01), indicating that caiman




73


2000 1


Q)





(9 Cn

C)
a





Cn
-)











n
-n






.5




0
(9
-
C


1750 1500 1250 1000750500-

250-


0




2000


1750-1500-1250--


1000750500250

0


Figure 4-4.


I


I
0







/


I II\ I


.0 j


0 0*T
T 1 0T
9 0 0


JFMAMJJ ASONDJ FMAMJ JASONDJ FMAM
1987 1988 1989
Month


J FMAMJJ ASONDJ FMAMJ JASONDJ FMAM
1987 1988 1989
Month

Mean dispersal distance ( SD) by month 19871989.


Females


00




TI T
, 99T
0 0.


MC, es





74


Table 4-3. Mean, maximum and percent of caiman with dispersal distances less than one kilometer, by month. N refers to the number of caiman in the sample.
Month Mean Maximum % <1 km N
(m) (m)


January
1988 1989
February
1988 1989
March
1988 1989
April
1988 1989
May
1987 1988 1988
June
1987 1988 July
1987 1988
August
1987 1988
September
1987 1988
October
1987 1988
November
1987 1988
December
1987 1988


241 496

367 229

274
44

172 70

275 157 81

1,060
294

1,043
994

1,407
983

1,184 1,150

954 940

898 1,001

1,343 547


1,810 3,110

1,810 1,770

1,940
260

1,940
300

4,181
1,460
260

4,441 4,490

4,418 6,410

4,441 6,480

4,441 6,300

4,441 6,300

4,441 6,220

4,441 3,110


86.7 75.0

76.9 88.9

84.6 100.0

84.6 100.0

93.8
92.3 100.0

62.5 87.5

50.0 57.7

60.0 66.7

64.3 63.2

71.4 70.6

61.5 62.5

80.0 78.6


15
9

13
9

13
8

13
8

15 13
7

16
24

16 26

14 21

14 19

14 17

13 16

10
14





75


dispersed further in 1987 (mean=1003 m) than in 1988 (mean=811 m), which appears to be related to the delayed flooding of the savannas in 1988.

For females, the effects of year, female size, and reproductive state (nesters vs non-nesters) on dispersal distance were examined in a 3-way ANCOVA. Both year and reproductive state accounted for significant amounts of variation. Dispersal distance was significantly greater in 1987 (adjusted mean=990.1 m) than in 1988 (474.4 m), and non-nesting females (adjusted mean=960.9 m) dispersed farther than nesters (503.7 m). Comparing large (>74.99 cm SVL) with small (<74.99 cm SVL) females revealed no effect of size on dispersal distance.

Among males the effects of year and size-class were examined in a 2-way ANCOVA. In contrast with the female caiman, no significant year effect was found for males. However, a significant size-class effect was noted (F4,694=11.1; p<0.001). Males in the 90.0-99.9 cm SVL sizeclass dispersed farther from their dry season habitats (Fig. 4-5) than did larger or smaller males.



Distance per day

Annual Dispersal Pattern. A mean daily movement index (DMI) was calculated by dividing the distance between two successive locations by the number of days between the localizations. During the dry season, movements within a single body of water were not included in these







































70-79.9 80-89.9 90-99.9 100-109.9 )1 10

Size Class (cm)


Figure 4-5.


Mean wet season (June-December) dispersal distance ( SD) of male caiman by size-class. Combined 1987/8 data.


76


5000


4000 3000


2000 1000


E
Q)




C


U)
0


N=1 15







N=153



-=-7 N = 57 N = 253

0 1


0





77


calculations. The DMI is not a true measure of the total distance moved during the interval, but serves as an index that can be used to compare patterns of movement. Peak periods of movement in 1987 were associated with the early wet season (June) dispersal, and again in the late wet season (October) during a late wet season peak in rainfall (Fig. 4-6). A similar pattern was evident in 1988 but the movement peak for males was in July/August and the late season movement peak for both sexes was one month later than in 1987. The offset peaks in DMI in 1988 may be attributed to the delay in the caiman dispersal in 1988 due to the late onset of the rains in that year. The November 1988 peak in movements was associated with the return of caiman to the dry season lagoons. The 1987 dispersal pattern may be considered to be more typical because rainfall in that year was more representative of the long-term average for the region.

Factors affecting DMI. Adjusting for differences in mean monthly DMI, no significant differences in DMI were found between sexes (ANCOVA Fl,1,2000=0.63; p=0.43) or years (F1,1,2000=0.05; p=0.80). Among females, no effect of year, size-class, or reproductive state was noted. However, a significant size-class effect was seen among males (wet and dry season data; F4,1050=2.71; p<0.05), following the same pattern of differences among size classes as was seen in dispersal distance. An analysis of wet season data found significant differences among male size classes only in 1988


























Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec



0-0 -Males *- -Females


; ; ' \W ,


1988


K


04-


'~0


5/


V
0


/


150


0


Jan Feb Mar Apr May Jun Jul Auq Sep Oct Nov Dec


Figure 4-6. Mean daily movement index (DMI) by month and
sex with monthly rainfall data.


1 1987 1


78


1004-


0

-o

E


50+


450

- 400

-350
-300 2
-250 0
-200
-150 B
-100 3


0
I'


0
07 X _


0









150


100+


0

-E


50-4-


0


0-









-


0-


50
0









450 400 350 300 250 Q
200 150 100B 50

0


0


T1
0
1





79


(F4,387=2.88; p<0.05), and not in 1987 (F4,297=1-96; p=0.10). Greatest DMI values were noted among intermediatesized males (SVL 90.0-99.9 cm) (Fig. 4-7).



Annual reuse of dry-season lacgoons

At the end of the wet season, the majority of caiman returned to the same lagoon in which they had spent the previous dry season. In only two of 23 cases (9%) did caiman spend the entire dry season in a different body of water in successive years, both were adult females in 1987. One of these females returned to her original dry-season lagoon the following year. In two other cases, caiman spent the first part of the dry season in a different lagoon, but returned to the original lagoon during the course of the dry-season (20 February 1988, 4 March 1989). In Table 4-3 the maximum dry season dispersal distance values greater than 0.5 km are individuals that did not return to the same dry season lagoon. These results indicate that adult caiman tend to return to the same dry season refuges year after year.



Home Range

Home range size

Three types of home ranges were identified based on the spatial relationship between dry and wet season home ranges. Caiman with dry and wet season home ranges that overlapped completely were designated as Type 1. Type 2 home ranges






















N=163





N=223

N=194 N=398
N=78







70-79.9 80-89.9 90-99.9 100-100.9 )110

Snout-Vent Length (cm)


Figure 4-7.


Mean wet season (June-December) daily movement index (DMI) by male size-class ( SD). Combined 1987/8 data.


80


300


E


250-


200-


150-

100-


50. 0-





81


were characterized by partial or extensive overlap between dry and wet season home ranges, but at least part of the wet season was spent outside of the dry season permanent water lagoons. Dry and wet season home ranges that were completely non-overlapping in space were classified as Type

3 home ranges.

In cases where the wet and dry season home ranges were well separated, dispersing or returning caiman occupied intermediate locations on a short term basis (usually <3 days). These points were considered to be temporary stopover points between the dry and wet season home ranges and were not included in the calculation of home range size.

Dry season home ranges varied in size from near 0 ha to a maximum of 84.6 ha, and were limited by the amount of permanent water habitat. Two principal areas were utilized by caiman during the dry season in the study site on Hato Masaguaral: the Piscina (an association of lagoons and borrow pits located near the house), and the Guacimos Lagoon. Caiman in both areas tended to be relatively sedentary, and this is reflected in the small dry season home range sizes (Table 4-4, Fig. 4-8). In the large Guacimos lagoon (ca. 15 ha in the dry season), most caiman would concentrate in the deeper northwestern section of the lagoon, particularly during the day.

However, some large dry season home ranges were noted (Fig. 4-8), the largest of which were a result of animals moving between lagoons. This movement resulted from





82


dropping water levels, and produced artificially high home range values which included the non-wetlands habitats between the lagoons.



Table 4-4. Summary of minimum convex polygon home range (ha) data by sex and season.
Males Females Total

Mean N STD Mean N STD Mean N STD

Wet
1987 31.1 7 14.4 42.6 10 44.1 37.9 17 35.6
1988 34.6 8 46.1 46.2 7 62.2 40.0 12 22.0
Total 33.0 15 35.1 44.1 17 52.3 38.9 32 45.4

Dry
1987/8 8.1 6 7.5 21.3 6 28.8 14.7 12 22.0
1988/9 8.8 6 13.9 2.6 3 2.6 6.7 9 11.8
Total 8.5 12 11.2 15.1 9 25.1 11.3 21 18.8



During the wet season, caiman had much more available wetland habitat, and mean wet season home ranges (38.9 ha; N=32, SD=45.4 ha) were significantly larger than dry season home ranges (11.3 ha: N=21, SD=18.8 ha; F1,49; p<0.05; Table 4-4). Male and female caiman demonstrated no significant differences in dry or wet season home range size (F1,49=0.67; p=0.42), and home range type played no significant role in determining wet season home range size (F2,29=0.41; p=0.67). Among females, no difference was found in mean wet season home range size of nesting (46.6 ha) and non-nesting (39.4 ha) individuals (F1,15=0.07; p=0.79).







83


Number of Caimen
S
1987 wet Season I
5 -- Ns17


4 --


3


2 -2 -- - - -


0

10 20 30 40 50 60 70 80 90 100 110 120130140150160170180190200 Number of Calman
7
1987-8 Dry Season S --N-12

5

4 ~

3 - - -_______2





10 20 30 40 50 60 70 80 90100110120130140150160170180190200 Number of Caiman
7
1988 Wet Season N-15



4
3 ----- ---- - - - - -_ __ __2 ------___ - - - ---

31 -- - -




0
10 20 30 40 50 80 70 80 90 100110 120130140150180170180190200 Number of Calman


1988-9 Dry Season
8 --- - - -
N-9



40 - - -___ __-- --------- -- - -
4 -2 - - - -- --0
10 20 30 40 50 00 70 80 90 100 110 120 130 140 150 160 170 180 190 200 Home Range (ha)


Figure 4-8. Frequency distribution of minimum home range size by season, both sexes combined.





84


Body size, however, had a significant effect on wet

season home range size. Among females, small caiman (<75.1 cm SVL) had significantly larger wet season home ranges (mean=85.1 ha; SD=73.2) than did large females (mean=21.7 ha; SD=20.3 ha)(F1,15=7.56; p<0.05). Among males, differences in wet season home range size among the five size classes were observed (Fig. 4-9), but these differences were not significant (F4,10=1.83; p=0.20). Considering that the small sample size may have hindered the analysis, males were reclassified as large (>100 cm SVL) or small (<100 cm SVL). Using these new size classes the ANOVA revealed that large males had smaller wet season home ranges (mean 13.4 ha, N=7, SD=11.6) than small adult males (mean 52.5 ha, N=7, SD=45.3)(F1,12=4.89; p<0.05).

The type of home range used did not differ

significantly between males and females (Kruskal-Wallis; p=0.44). However, home range type did differ among size classes for both males and females (Kruskal-Wallis; p<0.05). Male caiman (Fig. 4-10) always dispersed to some extent from the dry season lagoons (no Type 1 home ranges). Small males were invariably restricted to Type 2 home ranges; that is, they did not disperse long distances or have separate dry and wet season home ranges. Large males (>100 cm SVL) had almost equal tendencies to have Type 2 and Type 3 home ranges. The intermediate size males predominantly occupied separate wet and dry season home ranges (Type 3).


































80-89.9 90-99.9 100-109.9 )110 Snout-Vent Length (cm)


Figure 4-9.


Mean wet season home range size ( SD) by size class.


of males


85


150


12090.


6030.


C




I


N=3 Males


N=3





N=3 N=5 N=1 T


0


)80





86


A similar pattern was noted for females (Fig. 4-11).

Small females predominantly established Type 3 home ranges, and larger females Type 2 home ranges. The only individuals that did not disperse from the dry season habitats (Type 1 home ranges) were large females (Fig. 4-11). No significant differences in home range type were noted between nesting and non-nesting females.


Individual home ranges in consecutive years

Caiman were consistent in their use of wet season areas during consecutive years. Five caiman were followed during the entire 1987 and 1988 wet seasons. Although the size of home ranges varied between years for most of the individuals (#60, #36, and #271), this was mostly due to outlying points. Caiman #36 and #60 had much larger home ranges in 1988 due to the use of different dry season lagoons during the 1987-8 season (both had moved from the Piscina Lagoon due to low water levels). However, the main activity areas for both caiman were similar during both years. Caiman 376, a large male, was consistently found in a marsh southeast of the Guacimos lagoon and had virtually identical home ranges in 1987 and 1988. With the exception of one outlying point to the north of the Guacimos lagoon, the home ranges of caiman 271, an adult female, were also nearly identical in both years. The one caiman that differed slightly from this pattern was #382, a subadult male. Although his 1987 and









Number of Caiman 1n


8-


<90 cm SVL M 90-100 cm SVL = >100 cm SVL


Males


6 -


4


2 F


0


Figure 4-10.


14


4


1 2 3
Home Range Type

Home range type of male caiman by size class.


Number of Calman


= <75 cm SVL M >75 cm SVL


Females


2 h


U.


U
2 3
Home Range Type

Figure 4-11. Home range type of female caiman by size
class.


87


i ~


! nI\


12101-


8 6 -





88


1988 home ranges were similar, he ranged over a wider area in 1988.

An additional five caiman provided incomplete data for the two consecutive years. All these caiman were tracked during the 1987 wet season, and through part of the 1988 wet season before the radio signal was lost. Of the five caiman, in 1988 four (two adult males and two adult females) dispersed to the same areas used during the previous wet season. This included two individuals that made long distance dispersals (ca. 4 km) from the Guacimos lagoon. One individual, a subadult male (#379) had different home ranges. In 1987 this male remained in the Piscina area, but at the beginning of the 1988 wet season dispersed first to the Guacimos lagoon, then to the San Juanera lagoon.

These results indicate that, on a year to year basis,

adult caiman are consistent in their use of both dry and wet season home ranges. Subadult males appear to have home ranges that are less fixed and change more on a short-term (year to year) basis.



Habitat Utilization

Macrohabitat utilization

The use of dry season macrohabitats was mostly

restricted to permanent water areas, mostly lagoons and borrow pits (Fig. 4-12a-c). However, during the 1989 dry season the extensive use of marsh habitats (Fig. 4-12c) was a result of the flooding of these areas by pumped subsurface






89


Percent Utilization of Habitat 1987
100%........


40% 1


60%



40%


20%

Jan Feb Mar Apr May Jun Jul Aug Sep Cc' Nov Dec


Percent Utilization of Habitat 1988


100% 80%


190% ( '/"~ff %


40% 20% 0% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec


Percent Utilization of Habitat 1989


100%




B No Data June-December 40%


20%--


0%
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month

Lagoon M Borrow Pit Stream Marsh
" Palm Savanna = Forest Bank


Figure 4-12. Annual trends in macrohabitat utilization by caiman 1987-1989.





90


water. Some caiman remained in the lagoon and borrow pit habitats throughout the year, but the majority occupied newly flooded areas, particularly marsh, palm savanna, and forest habitats (Fig. 4-12a-c). Use of other areas such as stream or bank habitats was quite minor. The small number of caiman in streams reflects the limited availability of this habitat within the study site. Likewise, little flooded habitat was available in bank areas and many of the caiman in this habitat were found hiding under terrestrial vegetation (see Microhabitat, below).

Wet season habitat use (1987/8, males and females) was compared to the measured availability of habitat over a 4,136 ha area encompassing the home ranges of all radio tracked caiman (Fig. 4-13). The stream habitat was omitted from this analysis because it was not possible to estimate its total area accurately. Over the study area, palm savanna was the dominant habitat type, covering 52% of the area, followed by marsh (27%) and forest (13%). Habitat use differed significantly from availability (chi square; p<0.001). Palm savanna, sandhill and bank habitats were underutilized, whereas lagoons and borrow pits were disproportionately used based on the null hypothesis of random habitat selection. Marshes and forests were used approximately as would be expected.

Effect of sex and size-class. No significant

differences in wet season macrohabitat utilization were noted between males and females (Kruskal-Wallis; p>0.05).





91


% Fixes


- % Av 50


4030


2010 0'
Palm Savanna Marsn Forest Sandhill
Habitat T


allable Habitat M%


Bank /pe


Lagoon


Figure 4-13. Percent total fixes (locations) of caiman
during the wet season and the total
availability of macrohabitat types: 1987-1988.


Fixes Borrow Pit


- -


A0 f




Full Text

PAGE 1

ECOLOGY AND BEHAVIOR OF THE SPECTACLED CAIMAN (CAIMAN CROCODILUS ) IN THE CENTRAL VENEZUELAN LLANOS By John B. Thorb j arnarson A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1991

PAGE 2

ACKNOWLE DGMENTS First and foremost I must offer my thanks to Tomas and Cecilia Blohm. Tomas and Cecilia initially made my trip to Venezuela possible, and offered me the hospitality of their home in Caracas and the opportunity to work on their ranch, Hato Masaguaral. Tomas' long standing interest in crocodilian conservation was one of the major impetuses for this work, and he supported it in every possible way. Without the support of the Blohms this study would never have been done. During my stay in Venezuela I had the pleasure of working for the Fundacion para la Defensa de la Naturaleza (FUDENA) as a research biologist. FUDENA, and its director Dra. Maria de Lourdes Acedo de Sucre, provided me with considerable logistic and financial support throughout. I would like to especially thank Glenda Medina Cuervo for her active support of all aspects of this work. Thanks also go to the entire FUDENA staff for their help and friendship. The field work conducted on Hato Masaguaral was made possible with the assistance of a great number of individuals. In particular I would like to thank Gustavo Hernandez, Maria del Carmen Munoz, Tibisay Escalona, Pedro Vernet, Ildemaro Gonzalez, and Jaimie Aranguen for their efforts. Gustavo spent nearly as much time on the ranch as ii

PAGE 3

I did, and kept the research program operational during my absences. Maria worked with me on the caiman radio telemetry project and collected all the data for the 1987 wet season. She will be my coauthor for the publication resulting from the information in chapter 4. Prof. F. Wayne King, who chaired my graduate committee, has offered me a tremendous amount of support throughout this project, as well as with all aspects of my studies in the realms of crocodilian biology and conservation. Prof. King also kindly provided me with space in his office in the Florida Museum of Natural History and unlimited access to his computer facilities. I would also like to express my gratitude to the other members of my committee, Katherine Ewel, Lou Guillette, Mel Sunquist, and Fred Thompson, for their help and their patience in reviewing and correcting my dissertation. John Eisenberg served on my committee in every way except for reviewing the dissertation. John Robinson also served on my committee before departing from the University of Florida. I would also like to thank John for providing the impetus for my first visit to Venezuela. During this lengthy study discussions with numerous individuals added considerably to my understanding of reptile and crocodilian ecology, as well as the many other topics relating to this dissertation. These individuals include F. Wayne King, Kent Vliet, Grahame Webb, Harry Messel, Stephan Gorzula, Donald Taphorn, Andres Eloy Seijas, Jose Ayarzagiiena, Carlos Rivero Blanco, Phil Hall, Kent iii

PAGE 4

Redford, Lou Guillette, Peter Brazaitis, Alan Woodward, Dennis David, Bill Magnusson, Paul Moler, Lee Fitzgerald, Val Lance, Peter Crawshaw and Eduardo Cartaya. I would also like to acknowledge the assistance of the numerous individuals who helped me and off erred me hospitality during my work on caiman on ranches other than Hato Masaguaral, especially W. DeVrees, A. Branger, E. Hernandez, H. Escanona, L.E. Moser, P. Zarate and A. Carillo Garcia. I would also like to thank Mima Quero de Peha, Francisco Perez Perez, Gonzalo Medina, and Evaristo Martinez of the Venezuelan Ministerio del Ambiente y de los Recursos Naturales Renovables (MARNR) for granting me permission to carry out these investigations. Financial support for this project was supplied principally through the Smithsonian Institution, FUDENA, Wildlife Conservation International (New York Zoological Society) , and Tomas Blohm. Additional support was provided by the World Wide Fund for Nature (WWF) , and the World Wildlife Fund (US) . For their help in obtaining and administering funds for this project I would like to thank Rudy Rudran, Dale Marcel lini and Betty Howser of the Smithsonian Institution (National Zoological Park) ; Stuart Strahl, Archie Carr III, John Behler and Mary Pearl of Wildlife Conservation International; Curtis Freese and Ginette Hemley of WWF-US; and Hartmut Jungius, Sylvia Guignard and Aileen Ionescu-Somers of WWF-International . iv

PAGE 5

Funding for my graduate assistantship while writing up this dissertation was kindly provided by the IUCN Crocodile Specialist Group (H. Messel, chairman), and PROHESA (E. Hernandez, director).

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGEMENTS ii ABSTRACT V111 CHAPTERS 1 INTRODUCTION 1 2 DESCRIPTION OF THE STUDY SITE 7 The Llanos Habitat 9 Hato Masaguaral 11 3 CHARACTERISTICS OF THE CAIMAN POPULATION 20 Introduction 20 Methods 22 Results 26 Discussion 52 4 MOVEMENT PATTERNS, HOME RANGE SIZE, AND HABITAT UTILIZATION 61 Introduction 61 Materials and Methods 62 Results 67 Discussion 119 5 DIEL ACTIVITY PATTERNS 13 0 Introduction 130 Methods 132 Results 134 Discussion 138 6 CAIMAN DIET 144 Introduction 144 Materials and Methods 145 Results 147 Discussion 160 7 FEEDING BEHAVIOR 170 Introduction I 70 Methods 171 Results 172 Discussion I 84 vi

PAGE 7

8 BEHAVIORAL AND SOCIAL ASPECTS OF REPRODUCTION 189 Introduction Methods Results Discussion 9 SEXUAL MATURITY, REPRODUCTIVE CYCLE , AND EGG AND CLUTCH CHARACTERISTICS.. Introduction Methods Results Discussion 10 NESTING ECOLOGY Introduction Methods Results Discussion 11 SUMMARY AND CONCLUSIONS LITERATURE CITED BIOGRAPICAL SKETCH vii 189 192 195 229 241 241 242 244 272 288 288 290 293 340 356 369 390

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ECOLOGY AND BEHAVIOR OF THE SPECTACLED CAIMAN (CAIMAN CROCODILUS ) IN THE CENTRAL VENEZUELAN LLANOS By John B. Thorbjarnarson May 1991 Chairman: Dr. F. Wayne King Major Department: Forest Resources and Conservation The ecology and behavior of the spectacled caiman was studied on a ranch in the central Venezuelan llanos over a five-year period. The estimated caiman population size over the 5,000 ha principal study site decreased from 3,265 in 1985 to 2,611 in 1989. The reasons for the decline were unclear but appear to be related to mortality or emigration among juveniles. Virtually all aspects of caiman ecology were affected by the strongly seasonal pattern of rainfall and water availability. Caiman densities reached an annual high during the dry season when only a few permanent water habitats were available, and in some cases exceeded 1,500 caiman/ha. Caiman dispersed during the wet season and crude wet season density values were 0.45-0.60 caiman/ha. Movement and habitat use were studied using radiotelemetry. Patterns of dispersal varied greatly, but most caiman occupied wet season home ranges that were partially or completely distinct from dry season habitats. The viii

PAGE 9

maximum dispersal distance found was 6.5 km, and no significant differences in movement patterns of macrohabitat use were noted between the sexes. Adult caiman tended to use the same wet season habitat in consecutive years, but subadults appeared to wander more. Nesting was a wet season phenomenon, although gonadal recrudescence and some reproductive behavior began in the late dry season. The caiman had a well developed system of social displays, involving vocal and non-vocal acoustic, visual, olfactory and tactile signals. Males established breeding territories in newly flooded habitats early in the wet season, with large males occupying the preferred habitats near the dry season lagoons. Overall, 55% of the nests produced young, and the leading cause of nest failure was depredation. Females and males both reached sexual maturity at an age of seven years, and at sizes of 60-65 cm SVL and 75-80 cm SVL, respectively. Female fecundity was positively correlated with size. Caiman fed principally on Pomacea snails, fish and crabs. The diet changed seasonally and with ontogeny. Feeding behavior to capture fish was quite complex, but overall success rate was low. ix

PAGE 10

CHAPTER 1 INTRODUCTION Except for changes in the anatomy of the palate and the vertebral structure, crocodilians have changed very little since the Jurassic period, nearly 200 million years ago (Steel 1973) . Although crocodilians have exploited a number of rather specialized niches throughout their long evolutionary history (e.g., the terrestrial sebecosuchians or the thalattosuchian marine crocodiles) , they have retained an inherent conservatism in their general morphology. While conservative in their evolution, crocodilians are not by any means primitive animals. Instead, this conservatism argues that crocodilians are supremely adapted to their particular lifestyle. Apart from the evolutionarily stable body form, the success of crocodilians is evident from their historical abundance in tropical and semi-tropical lowland aguatic ecosystems. However, during this century two crocodilian attributes, the guality and durability of their hides, and the reflective glow of their eyeshines, contributed to a worldwide decline in numbers as crocodilians were hunted commercially on a vast scale. Today, populations of crocodilians represent only a small fraction of their former numbers, with hunting

PAGE 11

and rampant habitat destruction being the principal causes (Groombridge 1982) . The worldwide decline of crocodilians has created a need for the development and implementation of management programs, but at the same time has made it more difficult to obtain the reguisite biological data upon which such programs should be based. Field studies are hindered by the fact that crocodilians are principally nocturnal, aguatic, wary of any human activity, and oftentimes only found in remote regions that are difficult to census or far from areas of logistic support. Additionally, studies of crocodilian population ecology are confronted by problems such as the large size attained by most species, and the resulting delayed maturity and long generation times. An adeguate understanding of the behavior and ecology of many species of crocodilians is going to reguire the establishment of long-term research programs that will provide information on critical population parameters and how they change over time, and with respect to impinging biotic and abiotic factors. Although most species of crocodilians are considered to be threatened, a few species have managed to escape the effects of human development. Some crocodilians, particularly the genus Paleosuchus . have bony hides of no commercial value and have not been extensively hunted (King and Brazaitis 1971). Other species, in certain areas, have actually benefitted from human activities, and are now more

PAGE 12

numerous than they were in historical times. This has been the case with spectacled caiman ( Caiman crocodilus ) in the Venezuelan llanos. The human occupation of the llanos has resulted in the virtual extirpation of one the caiman's competitors, the Orinoco crocodile ( Crocodvlus intermedius ) , as well as the creation of large amounts of new dry season habitat in the form of borrow pits or cattle ponds. The result is that in parts of the llanos where they have not been hunted, spectacled caiman are found at very high densities. Given these circumstances, the study of the spectacled caiman in the llanos provides us with the rare opportunity to study the behavior and ecology of a species at or near its environmental carrying capacity. Investigations of caiman behavior and ecology in the Venezuelan llanos are also facilitated by a number of other factors. The extreme seasonality of the habitat forces the entire caiman population to concentrate annually in a few shallow bodies of water, making census work quick, accurate, and easily repeatable. These conditions, and the relatively small size of spectacled caiman (maximum male length 2.6-2.8 m) , also expedite mark-recapture studies. Female caiman reproduce at a small size (ca. 1.2 m) and mature at a relatively young age for crocodilians (ca. 7 years) , which reduces generation time and facilitates long-term population analyses . The spectacled caiman is one of the most widely distributed and geographically variable of the 23 extant

PAGE 13

4 species of crocodilians. The species complex contains four or five described subspecies including C.c. chiapasius . distributed along the Pacific coast from Oaxaca, Mexico to Ecuador, and along the Atlantic from Honduras to the Golfo de Uraba, Colombia; C.c. fuscus, found east of the Golfo de Uraba through northwestern Venezuela; C.c. apaporiensis , restricted to the upper Rio Apaporis in the Colombian Amazon; the nominate subspecies C.c. crocodilus occupying the Orinoco drainage, most of the Amazon River system and the coastal region between the mouths of the two rivers; and the southern subspecies C.c. vacare , in the Paraguay River drainage and the Beni/Mamore/Guapore river system of lowland Bolivia. Caiman c. yacare has been considered a separate species by some authorities, most notably King and Burke (1989), who followed Medem (1981, 1983). Throughout its range, caiman are found in virtually all the available lowland wetlands habitats (Gorzula and Seijas 1989) , and the adaptability of this species is attested to by the presence of thriving introduced populations in Florida, Cuba, and Puerto Rico (Ellis 1980, Groombridge 1982). This study, initiated in 1984 in cooperation with Tomas Blohm and the Venezuelan Fundacion para la Defensa de la Naturaleza (FUDENA) , was a broad scope investigation of the behavior and ecology of the spectacled caiman in the central Venezuelan llanos. The principal goal was to examine a number of important aspects of caiman ecology under near carrying capacity conditions, and to determine how these are

PAGE 14

5 influenced by the extreme seasonal variation of water availability. Apart from the empirical interest in how these large reptiles adapt to the rigorous llanos environment, this study was designed to promulgate information that will be useful for the management of the species on a sustained yield basis. Investigations of the life histories of long-lived organisms are exceeding difficult but extremely important for the study of general life-history strategies (Wibur 1975) . Recently, some long-term studies of crocodilian population ecology have begun, but principally on large subtropical (American alligator) or tropical (Estuarine crocodile) species. While a relatively large number of studies of caiman ecology and behavior have been conducted, most studies have concentrated on only a few aspects of caiman ecology, and others, while rich in detail and broad in scope, have frequently lacked quantitative backing. Medem's treatises on the South American crocodilians (Medem 1981, 1983), and Alvarez del Toro's work on Mexican species are good examples. The situation of the spectacled caiman in the Venezuelan llanos lends itself particularly well to the study of a number of important aspects of the species ecology and life-history. This study, conducted from October 1984 through May 1989, concentrated on aspects of caiman population characteristics, reproductive ecology, feeding ecology, habitat use, and activity and movement patterns. Apart from providing a detailed description of

PAGE 15

6 the ecology and behavior of this population, one of the principal goals of this work was to examine life-history attributes of the spectacled caiman, in particular the relationship between body size and fecundity. However, besides biological interest, the spectacled caiman has also been the subject of much commercial interest in the Venezuelan llanos as well. An experimental harvest program was established in 1982 (Gorzula 1987, Thorbjarnarson, in press) and it has guickly grown into one of the largest crocodilian management programs in the world. Venezuela is currently producing approximately 100,000 legal caiman hides annually, making it the world's largest producer of crocodilian skins. Interest in caiman management has grown in a commensurate fashion and has far outstripped the guantity and guality of research on caiman ecology. At present, the Venezuelan government has no caiman research programs in effect and appears to be relying on what information has been determined from past studies, as well as the activities of non-governmental investigators.

PAGE 16

CHAPTER 2 DESCRIPTION OF THE STUDY SITE This study was carried out principally on Hato Masaguaral, a cattle ranch in the central Venezuelan llanos (Guarico State: 8°33'N, 67 0 37'W), approximately 50 km south of the town of Calabozo (Fig. 2-1) . Some parts of the study were also conducted on adjacent ranches, particularly Hato Flores Moradas and Hato Matadero, and observations were collected on aspects of caiman ecology and behavior from a variety of other ranches throughout the Venezuelan llanos. Tomas Blohm, the owner of Hato Masaguaral, has a keen interest in wildlife conservation and has maintained the ranch as a wildlife sanctuary and as a site for biological research since the 1960's. A significant amount of biological research has been conducted on Masaguaral over the last 30 years, and over 100 scientific publications have resulted from this work. Although a diverse range of topics has been investigated, most research has concentrated on the ecology of the vertebrate fauna, including previous studies on the spectacled caiman (Staton and Dixon 1975, 1977, Marcellini 1979) . Detailed information on the mammalian and bird fauna of the ranch has also been published (Eisenberg 7

PAGE 17

Figure 2-1. Map of Venezuela indicating the location of the Hato Masaguaral study site.

PAGE 18

et al. 1979, Rudran 1979, Thomas 1979, Sunquist et al. 1989) . The Llanos Habitat The llanos is a large (252,530 km 2 ) geosyncline located between the Caribbean Cordilleras to the north, the Andes to the west, and the Guayanan Shield to the south. This region is best described as a hyperseasonal savanna intermixed with varying amounts of deciduous or semideciduous forest. The entire area is drained by the Orinoco River, whose tributaries cross the llanos principally in a north-south, or west-east direction. Situated over pre-Cambrian basement rocks, the llanos is composed primarily of alluvial deposits from the Tertiary and Quaternary periods. Most surface sediments are quite recent, associated with the Pleistocene uplift of the llanos region and erosional deposition from the Andes and Caribbean Cordilleras (Vila 1960) . The llanos can be divided into four basic subregions: J the piedmont region adjacent to the Andes, the high plains, the alluvial overflow region, and the aeolian plains (Sarmiento 1983). Mountainous foothills, fast flowing rocky rivers and streams, and large alluvial fans characterize the piedmont region. Vegetation is principally savanna, with varying amounts of semi-deciduous tropical forest. This subregion is principally located along the base of the Andes and Coastal mountain ranges. The high plains, or upper llanos region, are also characterized by a relatively

PAGE 19

10 significant amount of vertical relief, here dominated by mesas with dissected or undulating topography. Pittier (1942) first described this region as one of low plateaus 200-300 m above sea level, crossed by Orinoco tributaries carving deep river courses. This subregion is divided between two principal areas, the eastern llanos of Monagas, Anzoategui, northern Bolivar, and eastern Guarico states, and the llanos region south of the Rio Meta in Colombia. These two units are separated by a central tectonic depression (occupied by alluvial plains) , and may be remnants of a formerly continuous uplands region from the Pliocene-early Pleistocene. The typical tree savanna vegetation is dominated almost exclusively by three species: Bowdichia virailoides . Byrsonomia crassif olia . and Curatella americana (Sarmiento 1983) . The alluvial overflow plain, or lower llanos, is a vast region almost without vertical relief that occupies a depression in the central part of the llanos. The dominant vegetational association in this region is the hyperseasonal savanna, characterized by few trees or palm ( Copernicia tectorum ) savannas (Sarmiento 1983) . In the wet season, the rivers in this region generally overflow, and most areas flood from this and heavy rainfall combined with poor surface drainage and soils of poor permeability. The western and more northern sections of this subregion are characterized by a greater occurrence of semideciduous forest cover.

PAGE 20

The aeolian plains regions of the llanos extends from the upper Meta region in Colombia to the northeast, including the Cinaruco and Capanaparo river drainages, and apparently is a remnant of the formerly arid climatic period during the Wurm glaciation (Tricart 1974, Schubert 1988). This region contains extensive dune fields oriented in a northeast-southwest direction, situated atop sandy soils. Vegetation is characterized by open, virtually treeless savannas, with thin strips of gallery forest or small palmlined morichales (small streams lined with moriche palm, Mauritia minor ) . The entire llanos region is climatically hyperseasonal with a well defined wet season (May -November) during which over 75% of the rainfall occurs. Total annual rainfall in the llanos region generally ranges from 1000-2000 mm, with rainfall increasing towards the west and south. A mean annual temperature of 26-28° C, results in a Tropical Dry Forest plant association, as defined by the Holdridge system (Ewel et al. 1976) . Hato Masaauaral Hato Masaguaral is situated in a region of transition between upper and lower llanos, generally referred to as intermediate llanos (Berroteran 1985) . The ranch covers approximately 8,500 ha of mixed savanna/decidous forest habitat just west of the Guarico River at an elevation of 60-75 m above sea level. The dominant vegetation types tend

PAGE 21

12 to grade from a low-stature deciduous forest located in the east to an open palm savanna in the western portion of the ranch. A description of the vegetation types and dominant plant species on the ranch is given by Troth (1979). Climate On Hato Masaguaral the wet season begins in MayJune and ends in NovemberDecember . Mean monthly precipitation peaks in July and August (Fig. 2-2) . Because of variation in drainage characteristics the seasonal water level regimes of different habitat types differed widely (Fig. 2-3) . Peak flooding tended to coincide with periods of maximum precipitation, but in some areas maximum water levels were found one to three months after the peak in rainfall. The extreme seasonal variation in water availability is the key factor in shaping annual changes in the abiotic and biotic environment. The dry season is characterized by the widespread occurrence of fire. During the wet season extensive flooding occurs, and the llanos soils become waterlogged and anoxic. These environmental changes create a rigorous environment for both plant and animal communities and lead to seasonal changes in the ecology and phenology of the llanos fauna and flora. Mean annual rainfall for the region ranges from 1351 mm (Calabozo: 1961-1984) , to 1418 mm (San Fernando de Apure: 1921-1984) . During the study period total annual rainfall ranged from a low of 1448 mm (1984) to 1632 mm (1986) . In

PAGE 22

13 400.0 300.0 200.0 100.0 0.0 Rainfall (mm) Temperature A K -i + 282.0 -* * HK226.4 246.2 162.0 49.3 1.8 2.1 4. 5 231.8 211.9 107.9 13.8 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Ta Min 40.0 30.0 20.0 10.0 0.0 Rainfall Ta Max Figure 2-2. Mean monthly precipitation and minimum and maximum air temperatures. Rainfall data from Masaguaral during the study period (1984-1988), numbers indicate mean monthly values. Temperature data from Troth (1979) . 200 150 100 Water Depth (cm) Guacimos Savanna Road Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nlov Dec Month Figure 2-3. Mean monthly water depth at four locations within the study site.

PAGE 23

14 contrast to the seasonal changes in rainfall, the annual temperature variation is relatively minor (Fig. 2-2) . Mean daily maximum temperature only varies 4.8" C among months, and the corresponding value for mean daily minimum temperature is 3.4° C (Troth 1979). Minimum nighttime temperatures are experienced in December. The lowest temperature recorded during the study period was 15° C. Peak temperatures during the late dry season (March-April) would regularly exceed 37'C. Habitat Types The principal study site consisted of approximately 4,000 ha of mixed deciduous forest and palm savanna located to the west of the highway, in the western section of the ranch. Caiman were also found in the eastern part of the ranch in riverine habitats (Rio Guarico, Caho Caracol) , but these areas were not included in the study area. Within the study area, eight basic habitat types were defined: palm savanna, marsh, forest, bank, sandhill, lagoon, borrow pit, and streams. Habitats were discernable based on the degree of flooding, plant community structure, and soil types. Palm savanna Palm savanna dominated the western rim of the ranch, and was the most abundant habitat type, comprising 53% of the main study area (4,136 ha). This habitat was composed principally of palm bajio as described by Troth (1979), and

PAGE 24

15 consists of low-lying areas which were flooded to a moderate depth during the wet season. The predominant tree was the palm f Copernicia tectorum) , but a host of other trees and shrubs was freguently associated with the palm: strangler figs (Ficus spp.)/ a number of spiny shrubs and small trees (Annona spp., Randia venezuelensis , Zanthoxvlum culantrillo ) , or other small woody species (e.g. Coccoloba caracasana ) . The herbaceous vegetation was dominated by grasses, particularly Panicum laxum and Leersia hexandra . Palm savannas ranged from open, park-like areas to dense shrubby habitats that graded into forested forests. Marshes The more deeply flooded areas of the savanna that supported extensive rooted herbaceous vegetative growth were referred to as marshes, and comprised 27% of the main study area. Many marshes followed well demarcated, curvilinear paths that passed through the surrounding palm savanna habitat. On aerial photographs these marshes resembled old river courses, and may indeed represent former river or stream drainages. During the wet season flooding was extensive in marshes, which were the first savanna habitats to flood and the last to dry up. Depths in some marshes reached 1-1.5 m, but usually did not exceed 50 cm. Plant communities in marshes underwent an extensive change in phenology during the year. Following the first rains a carpet of grasses and herbs, including "platanico, "

PAGE 25

16 Thalia aeniculata . began growing. Thalia grew particularly well in the deeper marsh holes where it would form dense colonies that extended 1.5-1.8 m above the water level by the end of the wet season. The marshes also supported a luxuriant growth of grasses and sedges fPaspalum repens, Paratheria prostata . Oryza perennis , Eleocharis eleaans. E. mutata, E. minima and Cvoerus flavus ) , floating vegetation ( Neptunia olivacea , Ludwigia helminthowbriza . Utricularia inflata, Eichhornia azurea , Naias sp. , Salvinia sp. , Pistia stratoides ) and other rooted vegetation ( Ipomea crassicaulis . I. fistulosa, Justicia laevilincrais ) . Forests Forests were areas dominated by a moderate to dense growth of trees and shrubs of medium height (10-15 m) . Forests either were large, fairly contiguous patches, or were composed of smaller discrete patches surrounded by more open savanna. Troth (1979) provided a fairly complete list of the woody plants found in the forest habitats, but some of the more dominant species were the palm ( Copernicia tectorum ) , the saman ( Pithecelobium saman) , the masaguaro ( Pithecelobium quachapale . from which the ranch derives its name) , and the caruto ( Genipa americana) . The degree of wet season flooding was variable in the forests, but most areas experienced moderate inundation. Forests comprised 13% of the main study area.

PAGE 26

17 Banks Banks were elevated areas that did not flood during the rainy season. Most banks were long, narrow topographic features located along the edges of marshes, and apparently represent coarse grain depositional features of former water courses. Some trees were found on banks (mostly Copernicia tectorum ) , but most sites were defined by a characteristic herbaceous flora consisting of herbs (Sida acuta . Cassia tora . Stachvtarphaeta mutablis , Wissadula periplocif olia ) , and grasses and sedges ( Scleria muhlenberqii . Panicum laxum) that did not tolerate flooded conditions. Banks were a relatively minor habitat type on Masaguaral, comprising 2% of the main study area. Sandhills Sandhills were elevated sandy soil habitats that appear to represent old windblown sand deposits (Schubert 1988) . Sandhills never flooded, had a gently undulating topography and extensive herbaceous growth dominated by grasses and sedges, and supported a number of small trees and shrub species (Troth 1979) . Extensive growths of the spiny recumbent shrub Mimosa pigra were found around the edges of sandhill habitats (especially along ecotones with marshes) . Sandhills were only located in the eastern section of the ranch, and comprised 4% of the main study area.

PAGE 27

18 Lagoons Lagoons were permanent water habitats with extensive areas free of herbaceous vegetation. The Guacimos Lagoon was the largest on Hato Masaguaral, measuring approximately 18 ha during the wet season. The Guacimos Lagoon and several others (Piscina, San Juanera) represented extensive, deep sections of marshes. Other lagoons, such as Alta Venegas and Merecure, were located in depressions in sandhill areas. Most lagoons were guite shallow (maximum wet season depth: 1.5 m) , and were fringed with extensive vegetation similar to that found in marshes. Both the Guacimos Lagoon and the Piscina were artificially supplied with water during the dry season by subsurface water pumps. Lagoons comprised 1% of the main study area. Borrow pits Borrow pits were anthropogenic habitats dug using heavy machinery, mostly in association with road construction. The hydroperiod of borrow pits varied greatly. Some were seasonal, and would dry completely during the the course of the dry season. However, if the excavation was deep enough borrow pits would retain water throughout the year. Borrow pits were located alongside the national highway that passes through the ranch, and adjacent two elevated dirt roads within the ranch. Borrow pits comprised less than 1% of the main study area.

PAGE 28

19 Streams Streams were small seasonal water drainages that provided surface water runoff for the savannas. During the peak of the rainy season some streams would contain significant amounts of water and reach depths of up to 1 m. The width of streams rarely exceeded 3 m. Windmill Ponds Windmill ponds were small borrow pits or natural depressions located adjacent to windmills. During the wet season these ponds were extensively flooded open water habitats. In the dry season, subsurface water pumped by the windmill maintained a small, shallow (<50 cm deep) body of water.

PAGE 29

CHAPTER 3 CHARACTERISTICS OF THE CAIMAN POPULATION Introduction Throughout its distribution, the spectacled caiman occupies an enormous range of habitat types. Differences among these habitats may exert a strong influence on many aspects of the behavior and ecology of caiman populations including movement patterns, feeding ecology, and the timing of reproduction. The nature of the habitat should also play a significant role in shaping the structure of the caiman population itself. The rigorous seasonal fluctuations of the llanos savanna ecosystem are assumed to play an important role in determining caiman population density, size-class structure, and growth rates. We would expect populations inhabiting such a harsh hyperseasonal environment to differ in many respects from populations in more equitable surroundings. This investigation had three principal objectives: l) estimate important population parameters including total population size, size-class distribution, sex ratio and growth rates, 2) guantify the seasonal variation in caiman density and biomass in different habitat types, and 3) examine the effectiveness of the nocturnal census technique 20

PAGE 30

21 by estimating the percentage of the total population counted, and by comparing nocturnal with diurnal counts. One of the principal objectives of this study was to guantify important aspects of the caiman population on Hato Masaguaral, and compare these results with the findings of other studies. Other investigations of caiman population characteristics (density, size-class distribution, sex ratio) in the Venezuelan llanos have been conducted by Staton and Dixon (1975) and Ayarzagviena (1983). Data are also available for the ecologically similar Venezuelan Guyana (Gorzula 1978) and the Brazilian Pantanal (Schaller and Crawshaw 1982, Crawshaw 1987), and for less seasonal habitats such as the Brazilian Amazon (Magnusson 1982) and Suriname (Glastra 1983, Ouboter and Nanhow 1984). In contrast, caiman growth rate has been examined in wild populations in only a few studies (Gorzula 1987, Ouboter and Nanhoe 1987), and they have dealt only with the growth of juveniles. One previous estimate of juvenile growth rates in the llanos was done by Ayarzagiiena (1983), but was based on assigned estimated ages derived from a size-f reguency histogram. These comparisons offer us the chance to see how population parameters may vary between populations living in similar habitats and those occupying very different environments. The previous study of caiman on Hato Masaguaral (Staton and Dixon 1975) also provides the

PAGE 31

22 opportunity to see how a population may change over a 10-15 year period. Methods Caiman censuses on Hato Masaguaral were based on nocturnal spotlight counts. Three types of periodic censuses were conducted during the course of this study: a weekly or biweekly census of the main study area, a monthly census of 23 borrow pits in the southern portion of the ranch (outside the main study area) during the dry season, and an annual survey of the entire ranch west of the highway. All counts were made using 4 v miners headlamps (ca. 40,000 candle power), or 200,000 candle power spotlights with a 12 v car or motorcycle battery as the power source. During the borrow pit census and the annual surveys, the size-class composition of the population was estimated by classifying caiman into one of four sizeclasses: class I-<20 cm snout-vent length, 11-20-59. 9 cm SVL, III-60-89.9 cm SVL and IV->89.9 cm SVL (Ayarzaguena 1983) . Information on the size-class distribution and sex ratio of caiman in the main study was also collected through a mark-recapture program. Caiman were captured by hand, by noosing with a locking wire noose mounted on the end of a 1.8-3.6 m long pole, or by seining. Main study area census . A regular census route through the main study area was established in October 1984 and run

PAGE 32

through May 1989. From October 1984 through May 1986 this census was repeated on a weekly basis. After May 1986 the census was done at biweekly intervals. The census utilized a 3.5 km transect through seasonally flooded savanna, plus counts at 10 permanent or seasonal lagoons. The transect was done from a slow moving vehicle driven along an elevated road that passed through a variety of savanna habitat types. For each caiman spotted the following data were collected: size-class, position along the transect, distance from the road, and habitat type. Censuses of lagoons estimated only population size, and were made from a standardized vantage point, except for the Guacimos Lagoon during the wet season which was done from a 4 m boat poled slowly around the lake perimeter. During each census, water temperature, air temperature, wind speed, and water depth were measured at the Guacimos Lagoon. Water depth was also measured adjacent to permanent markers in the Highway Borrow Pit, the San Juanera Lagoon, and in a small roadside pool located along the transect route. Censuses were begun just after sunset and were usually finished by 23:00 h. Borrow pit censuses . During the dry season a monthly census of 23 borrow pits (0.056-0.377 ha) was made along a 9.8 km road in the southern part of Hato Masaguaral. Censuses were conducted by slowly walking around each lagoon and tallying the number and size-class of caiman on a handheld microcassette recorder. In densely populated lagoons

PAGE 33

24 three consecutive total counts were made and the mean value used. Water depth was measured at a standardized point in Borrow Pit 2. Annual ranch census . The entire western portion of the ranch (5,500 ha) was censused on an annual basis over a threeto four-day period during the late dry season (earlymid April; 1985-1989). Total number and size-class distribution of caiman were estimated for all bodies of water. Small lagoons were censused as described above for borrow pits. The size-class distribution for the Guacimos lagoon was estimated from a small boat poled repeatedly through the lagoon. Censusing was begun shortly after sunset and was not continued past 24:00 h. Repetitive count censuses . Repetitive counts were used to estimate the sighting fraction (the percentage of caiman visible) of caiman (Messel et al. 1981) . At each body of water a total of 15 counts was made successively at two minute intervals using a 200,000 candle power spotlight. The sighting fraction (p) was estimated from the mean (m) and the standard deviation (a) , using the following formula • . • • 2 based on the binomial distribution: p=l-(a /m) . Diurnal censuses . In the 1989 dry season two diurnal and one nocturnal censuses were made at a series of six borrow pits to examine the variation between daytime and nighttime censuses. Three of the lagoons had low caiman densities, and three had very high densities. Diurnal counts were conducted in the morning (08:00-09:30 h) and the

PAGE 34

25 (14:10-15:20 h), and nocturnal counts were conducted from 19:10-22:10 h. Totals from the high density lagoons represent the mean of three counts. Growth rate . Growth rate was determined based on the recapture of marked caiman. The minimum recapture interval used was 30 days. Because many caiman were missing the tips of their tails, measurements were based on snout-vent length. There was a great disparity in dry vs wet season growth, so growth rate was calculated for the wet season based on a model developed by Messel and Vorlicek (1989) where : ASVL=a A T w + b AT30 days) were used to calculate b for each of three size-classes (10-30 cm SVL, 30-50 cm SVL, >50 cm SVL) . From the date of initial capture and the date of recapture, the number of wet season (T w ) and dry season (Td) days in the interval were calculated, and the equation could be solved for the wet season growth rate (a) .

PAGE 35

Results Sighting Fraction In 1986 repetitive counts were conducted in four lagoons and two borrow pits to estimate the sighting fraction. On 8, 9, 11 January, counts were made at four lagoons, producing a mean sighting fraction of 0.885 (Table 3-1) . Counts made at one lagoon and two borrow pits throughout the night of 10 March 1986 indicate a somewhat lower value of 0.723 (Table 3-1). The overall mean value for all repetitive count censuses was 0.795 (a=0.173, N=27) . The four lagoons censused in early January varied in depth and the amount of fringing vegetation growing in shallow water. The two Piscina lagoons were shallow (<1 m deep) , with very little vegetation. The Merecure lagoon was deep (>1 m) with some fringing vegetation, whereas the San Juanera lagoon was shallow with extensive fringing vegetation. The mean sighting fraction for the San Juanera lagoon was significantly lower than for the other lagoons (LSD; p<0.05) . Table 3-1. Mean sighting fraction values for censuses in lagoons on Hato Masaguaral. January values are the means for three counts, March values are the means for five counts . January 1986 Lagoon P March Lagoon 1986 P Piscina East 0. 900 Merecure 0. 742 Piscina West 0. 925 Borrow pit A 0. 642 San Juanera 0. 795 Borrow pit B 0. 784 Merecure 0. 918

PAGE 36

27 The sighting fractions calculated on 10 March were based on counts at two hour intervals throughout the night (20:00-04:00 h; Fig. 3-1). The mean sighting fraction for all lagoons remained relatively constant throughout the night except at 22:00 h when a sharp decline was found (0.435), resulting in a mean sighting fraction significantly below the other values (LSD; p<0.05). Diurnal Censuses Far fewer caiman were counted during the day than at night (Table 3-2) . Although the results were quite variable, no differences were noted between morning and afternoon counts, nor between high and low density lagoons. The overall mean correction value for diurnal/nocturnal counts was 0.48. The estimated size-class distribution of the caiman population also varied between diurnal and nocturnal counts. In four out of five cases, the proportion of class IV caiman was underestimated during the diurnal counts (Fig. 3-2), suggesting that class IV individuals have a greater tendency to remain hidden during the day than other size-classes. Annual Ranch Census The number of non-hatchling caiman (caiman >20 cm SVL) on Hato Masaguaral decreased during the study period (Fig. 3-3) , with the greatest decline occurring between 1985 and 1986. Population size from 1986 to 1989 remained relatively

PAGE 37

28 1.00-r 0.90c 0.80-o u 0.70-o 0.60-Ll 0.50-c 0.40-_c 0.30cn 0.20-if) 0.10-0.00SMerecure Prest. 2 Prest. 1 20:00 22:00 24:00 02:00 04:00 Hour 20 15 g_ if) 10 CD --5 "O Figure 3-1. Sighting fraction of caiman and wind speed at three lagoons recorded at two hour intervals during the night of 10 March 1986.

PAGE 38

29 Table 3-2. Diurnal (morning and afternoon) and nocturnal counts for six sites on Hato Masaguaral during January and February 1989. The last two columns present the ratio of morning and afternoon counts to nocturnal counts, respectively. High density values are the means of three Borrow Date Pit AM Count PM Count Night Count AM/Night PM/Night Low Densitv 2 Jan Feb 0 7 0 7 3 8 0.88 0.88 8 Jan Feb 1 7 1 5 5 8 0.20 0.88 0.20 0.63 17 Jan Feb 1 1 3 0.33 0.33 18 Jan 2 2 8 Subtotal 0.25 0.57 0.25 0.51 Hiah 9 Densitv Jan Feb 47.7 48.5 42.0 73.0 121.0 118.3 0.39 0.41 0.35 0.62 18 Feb 24.3 33.3 91.3 0.27 0.36 19 Jan Feb 61.5 41.8 55.3 105.0 98.7 123.3 Subtotal 0.62 0.34 0.41 0.56 0.85 0.55

PAGE 39

si p o i o N •H Tj c id s -H d o "0 ts a) w XI o ro c o V) (0 P. -H a-p c e w p. o-h o 2 (0 0< Q) W o -p c o -p ja 0> •H c p id c • P cn
PAGE 40

31

PAGE 41

3000 2500 -2000 -1500 1985 1986 1987 1988 1989 Figure 3-3. Uncorrected annual census totals for nonhatchling caiman in the western section of Hato Masaguaral .

PAGE 42

constant. Corrected population size, based on a sighting fraction of 0.8, ranged from 3,265 in 1985 to a low of 2,519 in 1988. The ranch population was found in four major dry-season habitat categories. Borrow pits contained 60% of the total non-hatchling population, and 3 2% were located in the main study area. A much smaller percentage (6%) of the population was in "other lagoons," which consisted of two small natural lagoons (Alta Venegas and Merecure) . Windmill ponds contained 2% of the censused caimans. Because the main study area lagoons were artificially maintained by the pumping of subsurface water, the "other lagoons" were the only natural dry-season caiman habitat on Hato Masaguaral. Density and Biomass The number of caiman in permanent water habitats, seasonal lagoons, and the seasonally flooded savannas followed a regular pattern associated with the annual rainfall pattern and the movement of the caiman. In seasonally inundated savannas, peak numbers of caiman were seen during the wet season (Fig. 3-4) . Many of the areas bordering the savanna transect route contained pools that were among the first areas to flood and the last areas to dry up, so the observed caiman density was highest during the early and late wet season when caiman were dispersing from and returning to the permanent water lagoons. The dry

PAGE 43

34 c o o CD -O O O Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 3-4 Month Seasonal trend in caiman census totals (mean ± SD) and water depth along a 3.5 km transect through seasonally flooded savannas. 0 —i — i — i — i 1 — Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 200 -150 00 + 50 0 Q CD o CD o Figure 3-5. Month Seasonal trend in caiman census totals (mean ± SD) and water depth in the Guacimos Lagoon.

PAGE 44

season caiman observations along the transect were restricted to two windmill ponds. The reverse trend in seasonal caiman density was seen in permanent water lagoons, where peak numbers were observed during the dry season (Fig. 3-5 and 3-6) . However, among permanent water lagoons, two patterns of seasonal population size were evident. In some lagoons (e.g. Guacimos Lagoon; Fig. 3-5) , the peak number of caiman was reached during the mid dry season (February-March) , whereas in other lagoons (e.g. Highway Borrow Pit; Fig. 3-6) , the number of caiman tended to increase throughout the dry season. This variability may be attributed to differences in the hydroperiods of lagoons in surrounding areas. Seasonal lagoons such as the San Juanera (Fig. 3-7) contained large numbers of caiman during the early dry season, but as the lagoon dried caiman moved to other lagoons. The area surrounding the Guacimos lagoon tended to dry almost completely early in the dry season. However, the Highway Borrow Pit had a large number of deeply flooded marshes and seasonal lagoons in the area so caiman numbers continued increasing throughout the dry season. In actuality, the dynamics of a lagoons 1 dry season caiman population were directly related to the number, and the hydroperiods, of nearby lagoons. Caiman density reached extremely high levels in some water bodies during the dry season, but tended to be inversely related to lagoon size. In the large Guacimos

PAGE 45

200 c o E o O 0) E 3 O o 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 3-6. Seasonal trend in caiman census totals (Man ± SD) and water depth in the Highway Borrow Pit. 120 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 3-7. Seasonal trend in caiman census totals (mean SD) and water depth in the San Juanera Lagoon

PAGE 46

37 Lagoon (dry season size ca. 12 ha), peak dry season density was 60.4 caiman/ha (April 1987), and the peak monthly fiveyear mean was 46.5 caiman/ha (March; Fig. 3-8). Much higher densities were observed in the 0.63 ha Highway Borrow Pit (Fig. 3-7), with a maximum density of 304.0 caiman/ha (May 1989), and a high five-year monthly average of 163.3 caiman/ha in May. These densities translated into mean biomass values of up to 408 kg/ha in the Guacimos lagoon and 1,655 kg/ha in the Highway Borrow Pit (Fig. 3-9). Peak monthly values were 530 kg/ha for the Guacimos lagoon (April 1987), and 3,082 kg/ha in the Highway Borrow Pit (May 1989). The highest caiman densities and biomass values, however, were observed in borrow pits in the southern part of the ranch. Density and biomass figures for 15 borrow pits (0.075-0.204 ha) were analyzed on a monthly basis during the dry season in four consecutive years. Five borrow pits were classified as high-density (mean=0.152 ha), and 10 were low-density (mean=0.102 ha). Lagoons located adjacent to natural savanna drainage systems (streams) were used by large numbers of caiman as the savanna water levels dropped. Low-density borrow pits were more isolated, and were usually occupied by a female with a pod of young, and freguently by one large adult male. Low-density borrow pits typically had densities of 20-80 caiman/ha, and biomass values of 100-300 kg/ha. High-density borrow pits reached extremely elevated density and biomass levels by the end of the dry season (Fig. 3-10 and 3-11) . By April and May, mean

PAGE 47

38 250 200 -150-10050-Guacimos Highway Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 3-8. Mean seasonal trend (± SD) in caiman density in the Guacimos Lagoon and the Highway Borrow Pit. 2500 Jon Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 3-9. Mean seasonal trend (± SD) in caiman biomass in the Guacimos Lagoon and the Highway Borrow Pit.

PAGE 48

39 densities were 700-900 caiman/ha, with a high value of 1,619.7 caiman/ha in Borrow Pit 8 in May 1986. Peak mean biomass figures were in the 10,000-12,000 kg/ha range, with a maximum value of 23,379.1 kg/ha for Borrow Pit 19 in February 1986. Based on the calculated sighting fraction (0.8), and the assumption of equal sightability among sizeclasses, the estimated total density and biomass would be 20% higher in all cases. Wet season density and biomass values were much lower, but harder to estimate. If wet season density is calculated by dividing the total number of caiman counted during the annual censuses by the western ranch surface area, I obtain values from 36.6-47.5 caiman/km 2 for uncorrected counts and 45.8-59.4 caiman/km 2 for counts corrected using the sighting fraction of 0.8. Biomass values are 398.7-448.5 kg/km 2 (uncorrected) and 498.4-559.3 kg/km 2 (corrected) (Table 33). Monthly Borrow Pit Census Monthly censuses of a group of 23 borrow pits located along a road in the southern section of the ranch revealed that the great majority of the caiman used only a small number of the available water bodies. Of the 2 3 available borrow pits, five (located adjacent to streams) were high density and supported 83-91% of the non-hatchling caiman

PAGE 49

'o c O C o I o u 1 200 -j1000-800 600 400-200 -0-Figure O — O High Density # — # Low Density o — o o o T o* Nov Dec Jan Feb Mar Apr May Jun Month 3-10. Mean seasonal trends (± SD) in caiman density in highand low-density borrow pits. o JZ \ cn o o o (A w o E o in c o E o o 16 14 12 10 8 6 4 2 0 Figure 3-11, O — O High Density # — • Low Densitw I O' o•o P— o o Nov Dec Jan Feb Mar Apr May Jun Month Mean seasonal trends (± SD) in caiman biomass in highand low-density borrow pits.

PAGE 50

41 Table 3-3. Estimated wet season density and biomass values for non-hatchling caiman on Hato Masaquaral, 1985-1989. Year Total Density _ (caiman/ km ) Biomass (kg/km^) 1985 Total Counted Corrected Total 2, 612 3,265 47.5 59.4 404 . 3 505.3 1986 Total Counted Corrected Total 2,187 2,727 39.8 49.6 448.5 559.3 1987 Total Counted Corrected Total 2,127 2,659 38.7 48.3 398.7 498.4 1988 Total Counted Corrected Total 2,015 2,519 36.6 45.8 38.0 47.4 409.7 512.2 1989 Total Counted Corrected Total 2,089 2,611 415.6 519.4

PAGE 51

population (Table 3-4) . The remaining 18 low-density borrow pits held a maximum of only 17% of the total non-hatchling caiman population. The low-density borrow pits held a larger percentage of the total population (as opposed to the non-hatchling population) because these bodies of water were the principal "nursery" areas where hatchling caiman were found (Fig. 312). Typically, these low-density borrow pits were inhabited by an adult female with a pod of hatchlings. Many of these borrow pits were also shared by a single large male and a small number of juvenile caiman, particularly oneand two-year old individuals that would mix in with the pod of hatchlings. Table 3-4. Percentage of the non-hatchling caiman population in high and low density borrow pits on Hato Masaquaral; 1985-1989. Year Hiqh Density Low Density 1985-6 90 10 1986-7 83 17 1987-8 84 16 1988-9 91 9 Size-Class Distribution The population size-class distribution was determined by estimating the size of caiman during the annual census (Fig. 3-13) . A size-class breakdown of captured caiman is provided in Figure 3-14.

PAGE 52

43 % Observed Population 60 1 " Caiman Size-Class Figure 3-12. Caiman size-class distribution in highand low-density borrow pits.

PAGE 53

Figure 3-13. Annual caiman size-class distributions for the entire ranch; 1985-1988.

PAGE 54

Number of Caiman 350 1 Size-class (cm) 3-14. size-class distribution of captured over 20 cm SVL (N=634) .

PAGE 55

46 The caiman population was dominated by a large percentage of class III individuals (59.6% overall). Large adult males over 90 cm SVL comprised 16.3% of the population. The annual differences in size-class distribution were relatively small except in 1985, when a large percentage of class II individuals was seen. There was a great deal of variation in the caiman sizeclass distribution among the major habitat categories (Fig. 3-15) . The main study area lagoons tended to have a higher percentage of class II individuals and fewer class IV males than did the borrow pits. The study area lagoons also had very few hatchling caiman at the end of the dry season. Windmill ponds had few caiman, but a large percentage were hatchling (class I) and juvenile (class II) animals. The size-class breakdown of captured non-hatchling caiman is significantly different from the size-class distribution of censused caiman (chi-square; p<0.001). However, as the vast majority of the captured caiman came from the main study area, it would be more appropriate to compare the size-class distribution of captured caiman with caiman censused from the main study area. A chi-square analysis, however, indicates that the two are significantly different (p<0.001). This discrepancy suggests that either the size-estimation procedure was biased (classifying class II caiman as class III) , or that there was a bias in the capture procedure towards small animals. I suggest that the latter is the principal reason for the difference for two

PAGE 56

in H i

PAGE 57

48

PAGE 58

49 reasons. Firstly, although misclassif ications of size-class were made, errors were not great. Size classification was made based on the experience of hundreds of captures of individuals. Also, it was my impression that I was just as likely to misclassif y class III animals as class II, as the other way around. Secondly, a size-bias in captures is quite probable given the fact that small caiman were more easily captured than larger ones. This suggests that the sample of captured individuals is probably biased towards smaller caiman and the census size-class distribution more accurately reflects the population size-class composition. Sex Ratio Based on a sample of 634 captured caiman, the sex ratio (male: female) was 1.10:1, and was not significantly different from a 1:1 ratio (chi-square ; p>0.05). However, when broken down by size-class, class III was significantly biased towards females (p<0.001), and class IV was composed entirely of males. Class II caiman had a 1:1 sex ratio (Table 3-5) . Growth Rates Growth rates of caiman were markedly affected by season. During the dry season caiman grew very slowly or shrank. Dry season (January-May) growth rates for caiman in the 10-30 cm SVL range averaged 0.004 cm SVL/day (SD=0.005,

PAGE 59

50 Table 3-5. Sex ratio of captured caiman by size-class . Size-class Females Males Total II (20-59.9 cm SVL) 162 167 336 III (60-89.9 cm SVL) 134 91 225 IV (>90 cm SVL) 0 69 69 N=5) . For larger individuals, dry season growth was negative: 30-50 cm SVL=-0.009 cm SVL/day (SD=0.027, N=5) , >50 cm SVL=-0.022 cm SVL/day (SD=0.022, N=3). These values were used as best estimates of average dry season growth rates, and were utilized to calculate wet season growth (Messel and Vorlicek 1989) . Wet season growth rates (June-December) were significantly greater than dry season growth. Overall, growth rate did not differ between the sexes (ANCOVA, F l, 1, 101 =2 • 40 ' P>0.05). Mean male growth rate was 0.032 cm SVL/day and mean female growth rate was 0.039 cm SVL/day. However, sexual differences in growth rate were noted within certain size-classes. Among the smaller size-classes growth did not differ between males and females (Fig. 3-16) , but with increasing size, growth rate in females decreased more rapidly than in males. Female growth was significantly less in the 50-70 cm SVL class (F 1/1:L =8.48; p<0.05) and the 70-90 cm SVL class (F 1 f ig=41 . 09 ; p<0.001). The size-class and sex

PAGE 60

0.1200.000 4 O O Males # • Females 10-30 30-50 50-70 70-90 90-110 >110 Size-Class (cm SVLl Figure 3-16. Mean growth rates (±SD) for male and female caiman by size-class. 10 15 Age (years) 20 Figure 3-17. Average growth curves of male and female caiman calculated out to 20 years. Zero growth during the dry season is assumed.

PAGE 61

52 specific growth data were used to construct growth curves (Fig. 3-17) . Discussion Caiman on Hato Masaguaral grew very slowly. Up to the age of approximately four years, both males and females grew at a rate of 0.04-0.06 cm/day during the wet season (JuneDecember) . Very little growth occured during the dry season (January-May) , a pattern that has been noted in other populations of caiman (Gorzula 1978) and other crocodilian species (Magnusson and Taylor 1981, Webb et al 1983a, Messel and Vorlicek 1989) . Over the 214-day wet season, annual growth increments of juvenile caiman averaged 8.5-12.8 cm SVL, but there was a large amount of variation among individuals. Female growth rate slowed considerably when snout -vent lengths exceeded 50 cm, and may be related to the commencement of energy investment in reproduction. Male growth decreased at a slower rate so at sizes above 50 cm SVL males grew significantly faster. These growth rates suggest that, on average, females become sexually mature when they are 6-8 years old. Males may attain physiological sexual maturity at 7-8 years of age (75-80 cm SVL) , but due to dominance related factors probably do not reproduce until they are 10+ years old (>90 cm SVL) (chapter 8) . Despite the large number of studies of caiman in the llanos, few data on caiman growth in the wild have been published. A growth curve for juvenile caiman (0-4 years

PAGE 62

53 old) based on estimates of age derived from a size-frequency plot of captured juveniles on Hato El Frio in Apure state, indicates that these caiman are growing approximately 18% slower than on Masaguaral (Ayarzagviena 1983) . Growth rates in the savannas of the Venezuelan Guyana (Gorzula 1978) are comparable to those from El Frio, but there was little seasonal variation in the growth of 1-2 year olds. In Suriname juvenile caiman inhabiting a river bordered by swamp forest exhibited much faster growth (Ouboter and Nanhoe 1984). Over the first year of life, growth rates averaged 2.7 cm total length/month, equivalent to 5 cm/day over the entire year, approximately 31% faster than the annual growth rates of one year olds on Masaguaral. In Suriname seasonal variation is more complex, with a long and short dry season, but the dry season is less stressful than in the Venezuelan llanos. Growth was greatest during the long dry season and the long wet season, perhaps because food was not limited (Ouboter and Nanhoe 1984) . Limited food availability has also been suggested as a causative factor in the low growth rates of American alligators in the Florida Everglades (Jacobsen and Kushlan 1989) . The negative growth rates exhibited by caiman during the dry season are quite interesting. During the course of the wet season it was not uncommon for caiman to lose weight, but a shrinkage in length indicates that, in the absence of a reduction in skeletal elements or connective tissue, that shrinkage may be a natural result of loss of

PAGE 63

54 condition resulting in the tighter packing of skeletal elements in the vertebral column. Censusing crocodilians is freguently a very difficult endeavor. Visibility bias is especially evident in wetland habitats dominated by herbaceous and woody vegetation. However, many of the factors contributing to visibility bias are minimized when, as on Hato Masaguaral, virtually all the animals are restricted to a few small, shallow bodies of water. The concentration of caiman in a few bodies of water during the late dry season, in concert with the high sighting fraction, provides ideal conditions for censusing. The mean sighting fraction (p) derived from repetitive counts (0.8) was similar to, but slightly lower than, empirical estimates made by Staton and Dixon (1975) for the same ranch (0.9) . The percent of the total population visible is greatly affected by the type of habitat. In some parts of the llanos caiman occupy bodies of water completely covered by water hyacinth ( Eichhornia sp.), or other floating vegetation that effectively hides a large part of the population from view. The lagoon with the largest amount of fringing vegetation on Masaguaral had the smallest sighting fraction, but values were high at all areas examined. The one other published study using repetitive counts reported somewhat lower p values (0.6-0.7 for juvenile C. porosus ) , and that the sighting fraction varied significantly among size-classes (Messel et al. 1981). More recent studies estimating the sighting fraction among

PAGE 64

55 crocodiles have relied principally on mark-recapture techniques and have produced lower estimates of p: C. porosus -0 .35-0 . 66 (Bayliss et al. 1986), C. niloticus-0 . 36 (Hutton and Woolhouse 1989) . The comparison of diurnal and nocturnal counts revealed that diurnal counts were poor estimators of total population size. The mean correction factor (diurnal/noctural counts) from this study was 0.48, well above the 0.3 0 value found by Seijas (1984) for 14 ranches in Apure, Barinas and Portuguesa states. Diurnal counts not only revealed a small percentage of the total population, but they had high sample variances (this study-0.47; Seijas (1984) -0.32) , and were biased against the large size-class IV individuals. Taken together, these data strongly suggest that census work should be based on nocturnal survey work, and not diurnal surveys using correction factors. The study on Hato Masaguaral was conducted during a period of declining population size. The total ranch population decreased 20% over the five year study period. Over 80% of this decline occurred during the interval between 1985 and 1986, and may be attributed to either mortality or emigration. After 1986 the population became relatively stable, although a slight downward trend was apparent (Fig. 3-3). From 1985 to 1986 the observed ranch population declined overall by an estimated 425 individuals. During this period the number of size-class II caiman dropped by 741, and the number of SC III caiman increased by

PAGE 65

56 293. Even assuming that the increase in SC III caiman was entirely due to recruitment from SC II individuals, there was a loss of 448 SC II caiman. Clearly, the dramatic decline in numbers of caiman during the 1985-6 period was due to loss of SC II caiman. Because of their small size, juvenile caiman are more susceptible to predation than are larger caiman, but juvenile and subadult crocodilians may also enter a dispersal phase (Messel et al. 1981, Hutton 1989) . However, based on the data available it cannot be determined which of these two factors was most important. Caiman density on Hato Masaguaral was extremely high during the dry season, with peak values in some borrow pits exceeding 1,000/ha. These density values reported here are comparable to figures presented by Staton and Dixon (1975) and Marcellini (1979), also for Hato Masagural, but are somewhat lower than the 3 , 000-4 , 000/ha reported by Marcellini (1979) for two ponds in Hato El Frio, in the low llanos. Extremely high dry season population densities appear to characterize many regions of the Venezuelan llanos, especially on ranches where the fauna historically has been protected. However, these dry season density values represent temporary concentrations of caiman. When considering overall values of population density and biomass, wet season values must also be evaluated. Because of the large seasonal amplitude in density and biomass, it is difficult

PAGE 66

57 to compare the results of this study with similar values obtained for more equitable environments. Wet season caiman biomass figures tended to be 20-200 times less than dry season values, and were about one order of magnitude below those reported for herbivorous turtles, and half an order of magnitude below average values for carnivorous turtles (Iverson 1982). Caiman wet season biomass was five times greater than estimated values for the sympatric llanos pond turtle Podocnemis vogli (Iverson 1982) . Caiman biomass is also much higher than biomass figures for any species of mammal occurring in the same habitat (west side of Hato Masaguaral: Eisenberg et al. 1979) . In fact, the combined crude biomass figure for all non-volant mammals in the study area (478.8 kg/km 2 ) was still below the mean corrected biomass for caiman (518.8 kg/km 2 ). Historically, caiman populations in the llanos were much smaller. Two factors led to significant increases in caiman numbers in the 20th century. The first was the near extirpation of the Orinoco crocodile ( Crocodvlus intermedius ) , which was once a common faunal element in the rivers and streams throughout the llanos. The over exploitation of crocodiles for their hides left nearly vacant the river habitat niche, which was swiftly filled by spectacled caiman. The other factor was the alteration of the llanos landscape by humans. Activities associated with cattle ranching and road construction have greatly increased

PAGE 67

the availability of dry season wetland habitats (mostly borrow pits). On Masaguaral, the two principal dry season lagoons were artificially maintained by pumping sub-surface water. Only 6% of the Masaguaral caiman population during the dry season was located in natural lagoons, suggesting that human alteration of the habitat has increased carrying capacity 10-20 times. In many parts of the llanos caiman migrate to rivers during the dry season. In the Guarico River, bordering the eastern limit of Masaguaral, wet season caiman density is as low as 1.24 caiman/km during the wet season, and as high as 24.86/km in the dry season (Thorbjarnarson and Hernandez, in press). However, prior to the 1930' s this river had a large crocodile population, thus presumably depressing caiman density. Overall, caiman populations in the llanos have greatly benefited from the activities of humans during the 20th century. The only census data that exist from the mid-1970* s are for the Guacimos Lagoon (Staton and Dixon 1975) , and suggest that total population size in the lagoon had increased from a peak of 280 in 1974, to a mean high value of 574 during the period 1985-1989. However, as no data are available from other lagoons in the area it cannot be stated with certainty that this doubling in population size represents a accurate estimate of population increase. Nevertheless, these data strongly suggest that the Hato Masaguaral caiman population has grown since the study of Staton and Dixon (1975) . Although population size has increased, other

PAGE 68

aspects of the population have remained relatively constant. The sex ratio of caiman determined in this study, 1.10:1 males to females, was exactly the same as that determined by Staton and Dixon (1975) for Hato Masaguaral 10-15 years previously. The size-class distribution of caiman in this study is very similar to the one found by Staton and Dixon (1975) for the same ranch some 10 years earlier. Both showed the population dominated by size-class III caiman: adult females and subadult and small adult males. Size-class II caiman, or juvenile males and females approximately 2-5 years of age, constitute only about 20% of the total non-hatchling population. Adult males (SC IV) comprise approximately 15% of the population. Other studies of caiman population sizestructure have generally found larger numbers of juveniles (review in Gorzula and Seijas 1989) . This is especially true in some of the less seasonal habitats such as the Brazilian Amazon (Magnusson 1982) , and the Coesewijne River in northern Suriname (Ouboter and Nanhoe 1984) . The small number of juveniles on Masaguaral suggests that hatchling mortality, and hence recruitment into the juvenile size-class, is extremely low. Hatchling mortality on Masaguaral is known to be extremely high (>95%, Thorbjarnarson, unpublished data) , and may be the factor limiting population size. However, analysis of the population size-structure in different parts of the ranch (Fig. 3-16) reveals that in areas where hatchling survival

PAGE 69

60 is the highest (borrow pits) , the proportion of SC II caiman is the lowest. The size-class structure of a population is an extremely dynamic parameter, and depends on size-specific mortality, growth, and dispersal rates. A more in-depth analysis of the significance of the population sizestructure will have to await further data on these topics. Nevertheless, the observed difference between the population size-class distribution derived from the population censuses, and the size-class distribution of captured caiman is worth noting. This difference is attributed to capture techniques that are biased toward smaller animals. This indicates that unless unbiased capture techniques are used (e.g. a large seine net) , care » should be taken in interpreting the size-class distribution of captured individuals as a reflection of the true population size-class distribution, which has been a common approach used in studies of caiman (Staton and Dixon 1975, Gorzula 1978, Ayarzagiiena 1983, Ouboter and Nanhoe 1984).

PAGE 70

CHAPTER 4 MOVEMENT PATTERNS, HOME RANGE SIZE, AND HABITAT UTILIZATION Introduction Movement patterns and habitat selection, and their variability among size-classes and between the sexes, are important components of a species' life history. Despite the importance of these topics, relatively little detailed research on crocodilians has examined these factors. By far the best studied species has been the American alligator ( Alligator mississippiensis ; Joanen and McNease 1970, 1972, McNease and Joanen 1974, Taylor et al. 1976, Goodwin and Marion 1979, Taylor 1984), although recent work has been done on Crocodvlus niloticus (Hutton 1989) , C. porosus (Webb and Messel 1978, Webb et al. 1983), and C. iohnsoni (Webb et al. 1983b). The goal of this investigation was to study broad aspects of movement patterns of a large number of adult and subadult caiman on a relatively long term basis. Specifically, the effects of caiman size and sex on seasonal movement patterns, home range size, and macro and microhabitat selection were examined. Radio-telemetry is an excellent method for efficiently following large numbers of animals on a long term basis, but among crocodilians this 61

PAGE 71

62 technique has only been used on American alligators (Joanen and McNease 1970, 1972, McNease and Joanen 1974, Taylor et al. 1976, Goodwin and Marion 1979, Taylor 1984) and Nile crocodiles (Hutton 1989) . With spectacled caimans, habitat use and movement have been investigated principally based on mark-recapture studies (Gorzula 1978, Schaller and Crawshaw 1982, Ouboter and Nanhoe 1988), although the study of Ouboter and Nanhoe (1988) did include some information from radio telemetry. This is the first time that a large scale telemetry study of Caiman crocodilus has been done, and is also the first for any crocodilian inhabiting a environment with extreme hydric fluctuations. Materials and Methods The movement patterns and habitat utilization of 39 adult and subadult caiman were studied using radio telemetry. Initially, commercially purchased 3.9 v Lonner Modules (AVM Instrument Co., LTD.) in the 164-165 MHz frequency range were used. These modules contained an SB2 transmitter and a C or D-cell lithium battery power source with a 30 cm flexible whip antenna. The larger D-cell modules (9.6 cm x 4.3 cm; 220 g; N=7) were used for adult males, and the smaller C-cell modules (8.9 cm x 3.0 cm; 110 g; N=15) for adult females and sub-adult males. In 1988 caiman were fitted with two-stage oscillating crystal controlled transmitters (165-166 MHz) made in Gainesville, Florida, by Debbie Wright, a University of Florida graduate

PAGE 72

63 student. Ten transmitters utilized a C-cell lithium battery (8 cm x 3 cm; 100 g) , and five were powered by AA-cell lithium batteries (8 cm x 2 cm; 50 g) . All transmitters had 30 cm whip antennae. Transmitters were attached to the dorsal surface of the tail immediately anterior to the junction of the double row of caudal crests with the single row (Fig. 4-1) . The transmitters were fixed to the tail with 3-5 segments of 150 lb test monofilament fishing line which were sewn through the cartilage of the double caudal crests, and then wrapped repeatedly around the transmitter. The transmitter was completely encased in a wrapped layer of monofilament whose free ends were fused by partially melting the exposed monofilament strands. The entire transmitter/monof ilament module was then encased in a thin protective layer of dental acrylic. The antenna was run posteriorly along the row of single caudal crests through a series of short monofilament loops sewn into the tail cartilage. Caiman fitted with transmitters were tracked using a Telonics TR-2 receiver and an RA-2A "H" antenna. Due to signal attenuation caused by water or dense vegetation, in many cases the signal could only be heard initially from an elevated location (i.e. by climbing a tree or windmill) . Tracking was done on foot, horseback, or from a vehicle. During the wet season all caiman not in permanent water habitats were located by following the signal to its source. When the caiman was located, a series of environmental

PAGE 73

64 Figure 4-1. Position of radio transmitter on the dorsal surface of the tail of a caiman, prior to being covered with dental acrylic.

PAGE 74

measurements were made: air (shade) temperature, water temperature (approx. 5 cm depth) , water clarity (depth to which the tip of a white ruler was visible) , and water depth. The location of the caiman was classified into one of seven macrohabitat types (bank, palm savanna, forest, marsh, sandhill, lagoon, borrow pit, and stream; see habitat descriptions in chapter 2) . The extent of macrohabitats in the study area was determined by weighing the different habitat types cut from a habitat map drawn from aerial photographs . Additionally, the microhabitat of the caiman was classified into one of 8 categories: 1) On land in open (0L0) : caiman out of the water in the open, not hidden in vegetation. 2) On land in vegetation (OLV) : out of the water and partially or wholly hidden under vegetation. 3) Buried (B) : fully or partially buried in the substrate. 4) In open water (OW) : in open water without vegetation in the immediate vicinity. 5) In water in vegetation (IWIV) : in water among rooted herbaceous vegetation. 6) Under floating vegetation (UFV) : in water in or under a layer of floating vegetation. 7) Among trunks/branches (ATB) : in the water in or

PAGE 75

66 among live or dead tree/ shrub trunks or branches . 8) In Thalia colony (DTC) : in the water, in a dense colonial growth of Thalia aeniculata . a tall, rooted aguatic plant restricted to deep areas of marshes . The position of caiman located in permanent water bodies was fixed by triangulation. Normally, no environmental data were collected at these sites because caiman would usually move away before their original location could be determined. The position of all "fixes" (both direct localizations and triangulations) was marked on maps of the study site. For each wet season fix a "dispersal distance" was calculated by measuring the straight-line distance from the caiman to the permanent water body from which the animal dispersed. The term "dispersal" is used throughout this study to refer to the seasonal movement of caiman away from the dry season lagoons. Similarly, a "previous distance" was measured between successive fixes. All measurements were made using 1:11,000 scale aerial photographs of the study site. Home range size was calculated using a program for microcomputers (McPaal) developed by M. Stuwe and C.E. Blohowiak at the National Zoological Park, Smithsonian Institution. Caiman locations were assigned x and y

PAGE 76

67 coordinates using a transparent grid overlay. Analyses were performed for dry season and wet season home ranges. Only data from animals tracked over an entire dry/wet season period were used. Calculation of home range size was based on the minimum convex polygon method (Eddy 1977) . Results Due to the large number of animals that had to be located, and the difficulty of tracking the animals at night (as well as other project demands that required nocturnal work) , most tracking was done during the day (Fig. 4-2) . Although this method was not able to record nocturnal movements that returned to the same point, diurnal fixes did provide a good overall quantification of directional movements (Hutton 1989) . Patterns of Movement Twenty-two transmitters were attached to caiman between 27 November 1986 and 28 July 1987 (Table 4-1) . In June-July 1988 another 15 caiman were equipped with transmitters. Two additional caiman were fitted with transmitters that had fallen off other caiman, bringing the total number of animals in the study to 39. Two caiman were not included in the data analyses due to equipment malfunction within 3 0 days of release. Of the remaining 37 caiman, 16 were males and 21 were females. Males ranged from 58.4 cm to 127.8 cm SVL; females from 61.0 cm to 82.5 cm SVL (Fig. 4-3).

PAGE 77

68 Table 4-1. Caiman equipped with radio transmitters during the course of the study. An * indicates individuals not included in the data analyses. Sex SVL Dates Followed # Davs # Fixes M 127 . 8 19 Mar 87-5 Jun 88 443 127 036 F 80.5 17 May 87-14 Jan 89 607 99 060 F 80 . 1 17 May 87-15 Feb 89 636 104 094 M 99. 1 23 May 87-30 NOV 89 191 31 095 M 96.2 9 May 87-10 Oct 88 519 85 117 M 102 . 1 28 Aug 87-17 Jul 88 323 58 118 M 107 . 2 6 Apr 87-19 Jan 88 288 49 163 M 58 . 4 4 Jun 88-20 Jul 88 46 9 254 F 74 . 5 19 Mar 87-19 Jan 88 306 85 271 F 81.5 21 May 87-10 May 89 719 139 298 M 113 . 5 27 Nov 86-10 Aug 87 255 48 312 •J JM F 78 . 2 31 Jan 87-10 Jul 87 161 36 358 F 75 . 0 6 Apr 87-19 Nov 87 227 44 359 F 79 . 0 6 Apr 87-10 Nov 87 218 38 361 F 75.1 7 Apr 87-17 Aug 88 497 105 362 F 74.0 6 Apr 87-31 Oct 87 208 32 376 M 115.4 9 May 87-6 Jan 89 605 100 377 F 70. 1 9 May 87-13 Nov 87 188 32 378 F 82. 5 9 May 87-5 Aug 88 453 83 379 M 77.0 17 May 87-17 Aug 88 457 71 382 M 79.5 21 May 87-10 May 89 719 111 384 F 74.5 23 May 87-3 Aug 88 437 76 386 F 61.0 3 Jun 88-9 Sept 88 98 33 395 M 83.4 28 Jul 87-27 Nov 88 487 78 479 M 94.5 5 Jun 88-10 May 89 339 49 480 M 102.3 6 Jun 88-3 May 89 331 62 484 F 66. 5 8 Jun 88-2 Aug 88 55 10 488 F 70.0 9 Jun 88-29 NOV 88 173 21 489 M 59.0 9 Jun 88-5 July 88 26* 3 490 F 74.5 10 Jun 88-20 Jun 88 10* 3 491 M 106.3 10 Jun 88-10 May 89 334 55 492 F 75.0 11 Jun 88-13 Jan 89 216 31 493 F 80.0 13 Jun 88-23 Jan 89 224 30 495 M 119.5 13 Jun 88-10 May 89 331 63 497 F 61.7 19 Jun 88-20 July 88 31 9 500 F 81.0 26 Jun 88-10 May 89 318 60 505 F 75.0 18 July 88-5 Nov 88 110 21 508 M 121.0 11 Aug 88-28 May 89 290 61 513 F 79.4 17 Sept 88-28 May 89 253 44

PAGE 78

6 No. Fixes 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour Figure 4-2. Time of radio fixes for caiman throughout the study period (Nov 1986-May 1989) . Number of Caiman 50 75 100 Snout-Vent Length Class (cm) 125 Figure 4-3. Size-class distribution of caiman equipped with radio transmitters.

PAGE 79

70 Complete wet-dry season cycles were followed during the 1987-8 and the 1988-9 seasons. The 37 caiman were monitored for a mean period of 327 days each (range 31-719 days) . Of the 39 transmitters fitted, 22 dropped off and were recovered, 10 were lost, presumably after transmitter malfunction, and seven were still operational at the end of the study. It is not believed that any of the lost signals were caused by dispersal of caiman from the study site. An attempt was made to locate lost caiman by searching a radius of approximately 10 km around the study site from a fixed-wing aircraft (Cessna 182) in August 1988. Weak, intermittent signals from two transmitters were received over the study site but could not be accurately located. The one failed transmitter from 1987 was later recovered when the caiman was recaptured near its original point of capture. Dispersal from dry-season lagoons Caiman typically remained in one body of water throughout the dry season, although in some cases movement between lagoons occurred in response to dropping water levels. Dispersal from the dry season lagoons was associated with the annual rains and the flooding of the savannas. The late onset of the rains in 1988 delayed dispersal that year by a little over one month (Table 4-2). No significant difference in the mean dispersal date between sexes was found for either year (Fi / 34=0.12; p>0.05).

PAGE 80

71 Table 4-2. Mean dates of dispersal from dry-season lagoons. Mean 1987 SD (Days) Mean 1988 SD (Days) Females 24 May 7.8 30 June 12.2 Males 20 May 5.9 2 July 9.0 The great majority of caiman dispersed only after savanna flooding had reached an adeguate level. However, certain individuals dispersed from the dry season lagoons before significant flooding of the savannas occurred. In these instances the caiman sought cover under low-hanging vegetation, especially under small palms ( Copernicia tectorum ) near the marsh habitats that are the first areas to flood. Most early dispersers did not move more than 0.5 km from the body of water, but in 1987 caiman #95 (male; 95.8 cm SVL) dispersed 4 km from the Guacimos Lagoon and spent one week hiding in vegetation until the flooding of a nearby marsh created wetland habitat. No significant effects of size on date of dispersal were found for either males (F3 f 2 =0 49 » p=0.72) or females (F 1/8 =3.10; p=0.11) . Dispersal distance Annual Dispersal Pattern . Dispersal distance is a measure of the straight-line distance between the caiman and the water body where that individual spent the previous dry season, and is indicative of how far the caiman disperse during the course of the wet season. Despite the late

PAGE 81

72 dispersal in 1988, the peak dispersal distances for both years were achieved from July to September (Fig. 4.4). Mean peak dispersal distance during this time was 0.8-1.3 km. The maximum straight-line dispersal distance was 4.4 km in 1987, and 6.5 km in 1988 (both females). There is an annual dry-wet season pattern of dispersal in relation to permanent water sites (Table 4-3). Even during peak periods of dispersion (July-September for both years) a large fraction of the caiman remained within one kilometer of their dry-season refuge. Only in October 1987 does the percentage of caiman located within 0.5 km of their dry-season lagoon drop below 30%. In most peak wet-season months 40-50% of the caiman have dispersed a straight-line distance less than 0.5 km. In the latter half of the wet season caiman began moving back towards the permanent water sites. Some individuals returned to the lagoons as early as OctoberNovember, but others remained in drying marsh habitats until January, or even early February. Factors affecting dispersal distance . Differences in dry season dispersal distance (January-April) principally stem from a few caiman that did not return to the same dryseason body of water as in the previous year. A two-way ANCOVA analysis (using month as the covariate) of the 1987/8 wet season dispersal distance found no significant differences between males and females, but a significant year effect (F lf 1521=6 . 68 , p<0.01), indicating that caiman

PAGE 82

(D U c o co b 5 CO 0) Q_ CO b 0) O C O -t—> CO s "5 CO L_ CD COCO b J FMAMJJ ASONDJ FMAMJ JASO NDJ A M 1987 1988 1989 Month 200017501500^ 125010007505002500— h J Females i i i i i i i M J J AS 0 N 1987 i i i i i i i i 1988 1989 Month Figure 4-4. Mean dispersal distance (±SD) by month 1987 1989.

PAGE 83

Table 4-3. Mean, maximum and percent of caiman with dispersal distances less than one kilometer, by month. N refers to the number of caiman in the sample. Montn Mean (m) rldX XIUUII1 o V X Twill N January O A 1 e. 4 X i sin X , O X u 86 7 15 1989 ji o ^ 496 J , 111) *7 e> n Q February 1988 O ^ *7 3 67 1 , 8 10 "7 IS Q / o • y X J 1989 229 1 , / / 0 go Q o o • y Q y Marcn 1988 274 1 , 940 Q A C 84 . O X J 1 Q D Q 7 fin i nn n R o April 1988 1 "7 1 172 1 , 940 Q A C 84 . O X J 1QQQ "7 n "inn i nn n Q o May 1 Q O "7 1987 279 A 1 Q 1 4 , 181 y o . o XD iy o o 19/ i a An J. , 4 O U X J 1 Q O Q 8 1 o £ n i nn n xuu . u •7 / June 1 , UoU >l yl >i i 4,441 fi "5 c; 1 fi X6 1 QOO A A Q n fi 7 R 9 A juxy 1 O O "7 1 , U 4 J A A T O 4,4 18 en r\ SU . U Id 1 QHR QQ A j J *4 fi a i n ^7 7 O fi August 1Q07 A A A 1 4 , 4 4 X fin n 1 A X4 1 QOO Q Q O c a a n DO . / z X September 1987 1, 184 4,441 64.3 14 1988 1,150 6,300 63.2 19 October 1987 954 4,441 71.4 14 1988 940 6,300 70. 6 17 November 1987 898 4,441 61.5 13 1988 1,001 6,220 62.5 16 December 1987 1, 343 4,441 80. 0 10 1988 547 3, 110 78. 6 14

PAGE 84

75 dispersed further in 1987 (mean=1003 m) than in 1988 (mean=811 m) , which appears to be related to the delayed flooding of the savannas in 1988. For females, the effects of year, female size, and reproductive state (nesters vs non-nesters) on dispersal distance were examined in a 3 -way ANCOVA. Both year and reproductive state accounted for significant amounts of variation. Dispersal distance was significantly greater in 1987 (adjusted mean=990.1 m) than in 1988 (474.4 m) , and non-nesting females (adjusted mean=960.9 m) dispersed farther than nesters (503.7 m) . Comparing large (>74.99 cm SVL) with small (<74.99 cm SVL) females revealed no effect of size on dispersal distance. Among males the effects of year and size-class were examined in a 2 -way ANCOVA. In contrast with the female caiman, no significant year effect was found for males. However, a significant size-class effect was noted ( F 4,694= 11 * 1 ' P<0.001). Males in the 90.0-99.9 cm SVL sizeclass dispersed farther from their dry season habitats (Fig. 4-5) than did larger or smaller males. Distance per day Annual Dispersal Pattern . A mean daily movement index (DMI) was calculated by dividing the distance between two successive locations by the number of days between the localizations. During the dry season, movements within a single body of water were not included in these

PAGE 85

76 5000 4000 -CD u § 3000 CO b o CO i_ (D Q_ CO b 2000 115 < > N = = 153 N=127 t N=57 1 N=253 • I I ? ' I • 1 1 70-79.9 80-89.9 90-99.9 100-109.9 > 1 1 0 Size Class (cm) Figure 4-5. Mean wet season (June-December) dispersal distance (±SD) of male caiman by size-class. Combined 1987/8 data.

PAGE 86

77 calculations. The DMI is not a true measure of the total distance moved during the interval, but serves as an index that can be used to compare patterns of movement. Peak periods of movement in 1987 were associated with the early wet season (June) dispersal, and again in the late wet season (October) during a late wet season peak in rainfall (Fig. 4-6) . A similar pattern was evident in 1988 but the movement peak for males was in July/August and the late season movement peak for both sexes was one month later than in 1987. The offset peaks in DMI in 1988 may be attributed to the delay in the caiman dispersal in 1988 due to the late onset of the rains in that year. The November 1988 peak in movements was associated with the return of caiman to the dry season lagoons. The 1987 dispersal pattern may be considered to be more typical because rainfall in that year was more representative of the long-term average for the region. Factors affecting DMI . Adjusting for differences in mean monthly DMI, no significant differences in DMI were found between sexes (ANCOVA F± j, 2000 =0 « 63 > p=0.43) or years ( F l, 1, 2000 =0 ° 5 ' P=0.80) . Among females, no effect of year, size-class, or reproductive state was noted. However, a significant size-class effect was seen among males (wet and dry season data; F 4 , 1050 =2 . 71 ; p<0.05), following the same pattern of differences among size classes as was seen in dispersal distance. An analysis of wet season data found significant differences among male size classes only in 1988

PAGE 87

78 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec O — O-Males • •-Females Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 4-6. Mean daily movement index (DMI) by month and sex with monthly rainfall data.

PAGE 88

( F 4 , 387 =2 • 88 ' P<0.05), and not in 1987 (F 4f 297=1«96; p=0.10). Greatest DMI values were noted among intermediatesized males (SVL 90.0-99.9 cm) (Fig. 4-7). Annual reuse of dry-season lagoons At the end of the wet season, the majority of caiman returned to the same lagoon in which they had spent the previous dry season. In only two of 23 cases (9%) did caiman spend the entire dry season in a different body of water in successive years, both were adult females in 1987. One of these females returned to her original dry-season lagoon the following year. In two other cases, caiman spent the first part of the dry season in a different lagoon, but returned to the original lagoon during the course of the dry-season (20 February 1988, 4 March 1989). In Table 4-3 the maximum dry season dispersal distance values greater than 0.5 km are individuals that did not return to the same dry season lagoon. These results indicate that adult caiman tend to return to the same dry season refuges year after year. Home Range Home range size Three types of home ranges were identified based on the spatial relationship between dry and wet season home ranges. Caiman with dry and wet season home ranges that overlapped completely were designated as Type 1. Type 2 home ranges

PAGE 89

300 250 -O 200 -150100500 N=163 N=194 N = 78 N = 223 N = 398 Figure 4-7. 70-79.9 80-89.9 90-99.9 100-100.9 > 1 1 0 Snout-Vent Length (cm) Mean wet season (June-December) daily movement index (DMI) by male size-class (±SD) . Combined 1987/8 data.

PAGE 90

81 were characterized by partial or extensive overlap between dry and wet season home ranges, but at least part of the wet season was spent outside of the dry season permanent water lagoons. Dry and wet season home ranges that were completely non-overlapping in space were classified as Type 3 home ranges. In cases where the wet and dry season home ranges were well separated, dispersing or returning caiman occupied intermediate locations on a short term basis (usually <3 days) . These points were considered to be temporary stopover points between the dry and wet season home ranges and were not included in the calculation of home range size. Dry season home ranges varied in size from near 0 ha to a maximum of 84.6 ha, and were limited by the amount of permanent water habitat. Two principal areas were utilized by caiman during the dry season in the study site on Hato Masaguaral: the Piscina (an association of lagoons and borrow pits located near the house) , and the Guacimos Lagoon. Caiman in both areas tended to be relatively sedentary, and this is reflected in the small dry season home range sizes (Table 4-4, Fig. 4-8). In the large Guacimos lagoon (ca. 15 ha in the dry season) , most caiman would concentrate in the deeper northwestern section of the lagoon, particularly during the day. However, some large dry season home ranges were noted (Fig. 4-8), the largest of which were a result of animals moving between lagoons. This movement resulted from

PAGE 91

82 dropping water levels, and produced artificially high home range values which included the non-wetlands habitats between the lagoons. Table 4-4. Summary of minimum convex polygon home range (ha) data by sex and season. Males Females Total Mean N STD Mean N STD Mean N STD Wet 1987 31.1 7 14.4 42.6 10 44. 1 37.9 17 35.6 1988 34.6 8 46.1 46.2 7 62.2 40.0 12 22.0 Total 33.0 15 35.1 44.1 17 52.3 38.9 32 45.4 Dry 1987/8 8.1 6 7.5 21.3 6 28.8 14.7 12 22.0 1988/9 8.8 6 13.9 2.6 3 2.6 6.7 9 11.8 Total 8.5 12 11.2 15.1 9 25. 1 11.3 21 18.8 During the wet season, caiman had much more available wetland habitat, and mean wet season home ranges (38.9 ha; N=32, SD=45.4 ha) were significantly larger than dry season home ranges (11.3 ha: N=21, SD=18.8 ha; F^^g; p<0.05; Table 4-4) . Male and female caiman demonstrated no significant differences in dry or wet season home range size (Fi / 49=0.67; p=0.42), and home range type played no significant role in determining wet season home range size ( F 2,29 =0 * 41 ' p=0.67). Among females, no difference was found in mean wet season home range size of nesting (46.6 ha) and non-nesting (39.4 ha) individuals (F 1/15 =0.07; p=0.79) .

PAGE 92

Number ol Caiman 10 20 30 40 50 60 70 80 90 100110120130140150160170180190200 Number ol Caiman 10 20 30 40 50 60 70 80 90100110120130140150160170180190200 Number ol Caiman 10 8 6 10 20 30 40 50 60 70 80 90100110120130140150160170180190200 Number ol Caiman 1988-9 Dry Season N-9 _J I I i_j I 1 L_ 10 20 30 40 50 80 70 BO 00 100 110 120 130 140 150 160 170 180 190 200 Home Range (ha) Figure 4-8. Frequency distribution of minimum home range size by season, both sexes combined.

PAGE 93

84 Body size, however, had a significant effect on wet season home range size. Among females, small caiman (<75.1 cm SVL) had significantly larger wet season home ranges (mean=85.1 ha; SD=73.2) than did large females (mean=21.7 ha; SD=20.3 ha) (Fi f 15=7.56; p<0.05). Among males, differences in wet season home range size among the five size classes were observed (Fig. 4-9) , but these differences were not significant (F4 io=1.83; p=0.20). Considering that the small sample size may have hindered the analysis, males were reclassified as large (>100 cm SVL) or small (<100 cm SVL) . Using these new size classes the ANOVA revealed that large males had smaller wet season home ranges (mean 13.4 ha, N=7, SD=11.6) than small adult males (mean 52.5 ha, N=7, SD=45.3) (Fi /12 =4.89; p<0.05). The type of home range used did not differ significantly between males and females (Kruskal-Wallis ; p=0.44). However, home range type did differ among size classes for both males and females (Kruskal-Wallis; p<0.05). Male caiman (Fig. 4-10) always dispersed to some extent from the dry season lagoons (no Type 1 home ranges) . Small males were invariably restricted to Type 2 home ranges; that is, they did not disperse long distances or have separate dry and wet season home ranges. Large males (>100 cm SVL) had almost egual tendencies to have Type 2 and Type 3 home ranges. The intermediate size males predominantly occupied separate wet and dry season home ranges (Type 3).

PAGE 94

85 O JO 80 80-89.9 90-99.9 100-109.9 > 1 1 0 Snout-Vent Length (cm) Figure 4-9. Mean wet season home range size (±SD) of males by size class.

PAGE 95

86 A similar pattern was noted for females (Fig. 4-11) . Small females predominantly established Type 3 home ranges, and larger females Type 2 home ranges. The only individuals that did not disperse from the dry season habitats (Type 1 home ranges) were large females (Fig. 4-11) . No significant differences in home range type were noted between nesting and non-nesting females. Individual home ranges in consecutive years Caiman were consistent in their use of wet season areas during consecutive years. Five caiman were followed during the entire 1987 and 1988 wet seasons. Although the size of home ranges varied between years for most of the individuals (#60, #36, and #271), this was mostly due to outlying points. Caiman #36 and #60 had much larger home ranges in 1988 due to the use of different dry season lagoons during the 1987-8 season (both had moved from the Piscina Lagoon due to low water levels) . However, the main activity areas for both caiman were similar during both years. Caiman 376, a large male, was consistently found in a marsh southeast of the Guacimos lagoon and had virtually identical home ranges in 1987 and 1988. With the exception of one outlying point to the north of the Guacimos lagoon, the home ranges of caiman 271, an adult female, were also nearly identical in both years. The one caiman that differed slightly from this pattern was #382, a subadult male. Although his 1987 and

PAGE 96

10 Number of Caiman <90 cm SVL MS 90-100 cm SVL CD MOO cm SVL Males II 1 2 3 Home Range Type Figure 4-10. Home range type of male caiman by size class 14 12 10 8 Number of Caiman <75 cm SVL >75 cm SVL Females Home Range Type Figure 4-11. Home range type of female caiman by size class.

PAGE 97

88 1988 home ranges were similar, he ranged over a wider area in 1988. An additional five caiman provided incomplete data for the two consecutive years. All these caiman were tracked during the 1987 wet season, and through part of the 1988 wet season before the radio signal was lost. Of the five caiman, in 1988 four (two adult males and two adult females) dispersed to the same areas used during the previous wet season. This included two individuals that made long distance dispersals (ca. 4 km) from the Guacimos lagoon. One individual, a subadult male (#379) had different home ranges. In 1987 this male remained in the Piscina area, but at the beginning of the 1988 wet season dispersed first to the Guacimos lagoon, then to the San Juanera lagoon. These results indicate that, on a year to year basis, adult caiman are consistent in their use of both dry and wet season home ranges. Subadult males appear to have home ranges that are less fixed and change more on a short-term (year to year) basis. Habitat Utilization Macrohabitat utilization The use of dry season macrohabitats was mostly restricted to permanent water areas, mostly lagoons and borrow pits (Fig. 4-12a-c) . However, during the 1989 dry season the extensive use of marsh habitats (Fig. 4-12c) was a result of the flooding of these areas by pumped subsurface

PAGE 98

100% 80% Percent Utilization of Habitat 1987 60% i 40% 20% | Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 100% 80% 60% Percent Utilization of Habitat 1988 40% 20% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 100% Percent Utilization of Habitat 1989 No Data June-December Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month i Lagoon M Borrow Pit EH Stream H Marsh LSI Palm Savanna I I Forest Hi Bank Figure 4-12. Annual trends in macrohabitat utilization by caiman 1987-1989.

PAGE 99

90 water. Some caiman remained in the lagoon and borrow pit habitats throughout the year, but the majority occupied newly flooded areas, particularly marsh, palm savanna, and forest habitats (Fig. 4-12a-c) . Use of other areas such as stream or bank habitats was guite minor. The small number of caiman in streams reflects the limited availability of this habitat within the study site. Likewise, little flooded habitat was available in bank areas and many of the caiman in this habitat were found hiding under terrestrial vegetation (see Microhabitat , below) . Wet season habitat use (1987/8, males and females) was compared to the measured availability of habitat over a 4,136 ha area encompassing the home ranges of all radio tracked caiman (Fig. 4-13) . The stream habitat was omitted from this analysis because it was not possible to estimate its total area accurately. Over the study area, palm savanna was the dominant habitat type, covering 52% of the area, followed by marsh (27%) and forest (13%) . Habitat use differed significantly from availability (chi sguare; p<0.001). Palm savanna, sandhill and bank habitats were underutilized, whereas lagoons and borrow pits were disproportionately used based on the null hypothesis of random habitat selection. Marshes and forests were used approximately as would be expected. Effect of sex and size-class . No significant differences in wet season macrohabitat utilization were noted between males and females (Kruskal-Wallis; p>0.05).

PAGE 100

Figure 4-13. Percent total fixes (locations) of caiman during the wet season and the total availability of macrohabitat types: 1987-19

PAGE 101

92 However, significant differences among size-classes of males and females were found (Kruskal-Wallis ; p<0.001). Large males (>109.9 cm SVL) tended to use deep water habitats (borrow pits/marshes/ lagoons) during the early wet season (Fig. 4-14). Other habitat types (forest, palm savanna, and stream) were inhabited mostly in the middle to late wet season (August-December). Medium size males (90.0109.9 cm SVL) followed a similar pattern in 1988, when marshes and borrow pits were the principal habitats throughout the wet season, but in 1987 there was a significant use of forest and palm savanna habitats during the early to middle wet season (May-August) . Marshes were not used to any significant degree until July-August. Among small males (<90.0 cm SVL) it is difficult to discern any consistent trends. During 1987 forest and palm savanna dominated the wet season habitats, and in 1988 most small males remained in either borrow pits or lagoons. Among females (Fig. 4-15) , the use of forest and palm savanna habitats was greatest among small individuals. Large females principally selected marsh (1987) , or marsh and lagoon habitats (1988) , but palm savanna was also heavily utilized in the mid wet season of 1987. As with the larger males, habitat use during the early wet season was restricted mostly to marsh, borrow pit or lagoon areas. Effect of reproductive status . Significant differences in wet season macrohabitat use were found between nesting and non-nesting females (Kruskal-Wallis; p<0.05). Over the

PAGE 102

Figure 4-14. Percent of locations by macrohabitat type and month for male caiman. a) small males (<90 cm SVL) ; 1987 b) small males; 1988 c) medium males (90-100 cm SVL) ; 1987 d) medium males; 1988 e) large males (>100 cm SVL) ; 1987 f) large males; 1988

PAGE 103

Percent Utilization of Habitat Small Males in 1987 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month H Lagoon Borrow Pit EH Stream B Marsh nHHD Palm Savanna 1 I Forest Hi Bank Percent Utilization of Habitat Small Males in 1988 BB Palm Savanna I I Forest Hi Bank

PAGE 104

95 Percent Utilization of Habitat Medium Males m 1987 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Hi Lagoon ESSS Borrow Pit I I Stream Marsh f3B p a |m Savanna I I Forest Hi Bank Percent Utilization of Habitat Medium Males in 1988 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Hi Lagoon B8H Borrow Pit I I Stream Marsh LiUii] Palm Savanna I I Forest HI Bank Figure 4-14 — Continued.

PAGE 105

Figure 4-15. Percent of locations by macrohabitat type and month for female caiman. a) small females (<75 cm SVL) ; 1987 b) small females: 1988 c) large females (>75 cm SVL) ; 1987 d) large females: 1988

PAGE 106

Percent Utilization of Habitat Small Females m 1987 Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Lagoon ) Borrow Pit CZ3 Stream HI Marsh tiiiiiJ Palm Savanna Forest BH Bank Percent Utilization of Habitat Small Females in 1988 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month HI Lagoon BH Borrow Pit I I Stream eM2 Marsh Liiiiii Palm Savanna I I Forest Hi Bank

PAGE 107

Percent Utilization of Habitat Large Females n 1987 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month i Lagoon ! Borrow Pit CZJ Stream Marsn LUiiil Palm Savanna Forest HH Bank Figure 4-15 — continued.

PAGE 108

100 two-year period, non-nesting females demonstrated a greater use of marsh habitats than did nesting females. This was true both during the early wet season (June-July) and the nesting period (August-October) . Nesting females were found principally in lagoon and marsh (early wet season) , or borrow pit and marsh (nesting period) , habitats (Fig. 4-16) . Seasonal differences in habitat selection . The use of wet-season habitats primarily reflected the availability of suitably flooded habitats. Water depths selected by caiman varied consistently among habitats (Fig. 4-17) , suggesting that the availability of deep water sites in seasonally flooded areas was greatest in marshes, lowest in forests, and intermediate in palm savanna. The principal difference in wet-season habitat utilization between 1987 and 1988 was a decrease in the use of seasonally flooded areas (marshes, palm savanna and forests) , and an increase in utilization of the permanent water areas (lagoons, borrow pits; Fig. 4-18) . This shift in habitat usage between the years was undoubtedly related to the drier conditions in 1988, and the less extensive flooding of the savannas during the wet season (Fig. 4-19) . An analysis of covariance examining the effect of year, habitat type and sex on the water depth selected by caiman (month used as the covariate) found a significant year effect, with mean adjusted water depth in 1987 (44.1 cm) being deeper than the 1988 value (35.6 cm).

PAGE 109

101 % Fixes June-July 1987/8 Lagoon Borrow Pit Stream Marsh Palm Savanna Forest Habitat Type % Fixes August-October 1987/8 50 i 1 Lagoon Borrow Pit Stream Marsh Palm Savanna Forest Habitat Type Figure 4-16. Macrohabitat use of nesting vs non-nesting females; 1987-1988 data. Based on the percent of locations (fixes) of radio-tagged females. a) June-July b) August-October

PAGE 110

102 CL CD Q CD o 100 8060-4020N=73 N=91 N=322 N=167 N=97 Lagoon Borrow Pit Marsh Palm Sav. Forest Habitat Type Figure 4-17 Mean water depth (±SD) where caiman were encountered for five different macrohabitat types. Mean Percent Utilization Lagoon Borrow Pit Stream Marsh Palm Savanna Forest Habitat Type M 1987 1988 Figure 4-18. Mean percent utilization of five macrohabitat types during the wet season; 1987 vs 1988.

PAGE 111

Percent Utilization of Habitat Large Males in 1987 100% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Lagoon BB Borrow Pit ED Stream Marsn HO Palm Savanna I I Forest Ml Bank 100% Percent Utilization of Habitat Large Males in 1988 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Lagoon Ma Borrow Pit Palm Savanna I I Forest Stream Bank Marsn Figure 4-14 — Continued.

PAGE 112

103 125 £ 100 U £ 75-->r-> 0) O 50-CD O 25-Figure 4-19 T 0 1 I o 1 O -1987 A -1988 T o 1 Lagoon Marsh Palm Savanna Habitat Type Mean wet season water depth at three locations; 1987 vs 1988.

PAGE 113

104 However, an examination of the seasonal variation in water depth between 1987 and 1988 indicates that savanna water levels were only lower during the early part of the wet season (June-July; Fig. 4-20a-c) . This suggests that the initial low water levels of 1988 changed the entire seasonal pattern of macrohabitat usage, shifting it towards increased utilization of permanent water sites in the year with delayed flooding of the savannas. Water depth As was noted in the previous section, there was a significant difference in the water depths of areas used by caiman in the two years, as well as among habitats. There was also a significant difference between the sexes ( F l,728 =5 « 15 ' P<0.05), with males selecting deeper areas (adjusted wet season mean 42.1 cm) than females (37.7 cm) (Fig. 4-21) . Males were consistently found in deeper water areas in 1988, but in 1987 the difference was principally in the late wet season. Selection of deeper water by males may be a function of sex, or may simply be correlated with caiman size. An analysis of covariance among male sizeclasses indicates that there is indeed significant variation ( F l , 4 , 377 =5 • * 3 » p<0.001) in wet season water depths among size-classes (Fig. 4-22), with the deepest water areas being selected by large males (>109.9 cm SVL; mean depth 47.0 cm), and by smaller males in the 80.0-89.9 cm SVL range (mean 56.9 cm). However, this latter group may be biased due to

PAGE 114

105 Water Depth (cm) Aug Sep Month Dec Water Depth (cm) Jun Aug Sep Month Dec 130 110 90 70 50 30 10 Water Depth (cm) Quacimo* Road Jun Jul Aug Sep Month Oct Nov Dec 1987 1988 Figure 4-20. Monthly mean water depth values (±SD) : 1988. a) Guacimos Lagoon b) San Juanera Harsh c) Guacimos Road palm savanna 1987 VS

PAGE 115

106 U Q. CD Q i_ CD -t— ' b May Jun Jul Aug Sep Oct Nov Dec Month Q_ 0 Q -t— » o 120 100-806040 20 + 0 1988 # — • Males A. — A. Females Figure 4-21. H H 1 1 1 1 1 1 May Jun Jul Aug Sep Oct Nov Dec Month Mean monthly selected water depth (±SD) of male and female caiman during the wet season a) 1987 b) 1988

PAGE 116

107 100 E Q_ CD Q ^_ CD +' o goso -70-6050-40-30201004N=60 N=9 N=156 N=80 N = 78 70-79.9 80-89.9 90-99.9 100-109.9 >1 10 Size-Class (cm SVL) Figure 4-22. Mean wet season water depth (±SD) selected by male caiman, by size class.

PAGE 117

108 the small sample size. Small males (<90 cm SVL) selected deeper water than small females (Fi f 703=10.60; p<0.01), suggesting the male preference for deeper water is not simply related to caiman size. On a seasonal basis, in 1987 large males were located in the deepest water early and late in the wet season (June, October-December) , and small males selected deep water during the late wet season (November-December) (Fig. 4-23). Few discernable seasonal trends were evident among sizeclasses in 1988 as most small males remained in the deep, permanent water habitats due to the late start of the rains. No significant difference was noted in water depths selected by large (>75 cm SVL; adjusted mean=35.7 cm) and small females (<75 cm SVL; adjusted mean=38.8 cm), and few consistent seasonal trends were evident in selected water depth (Fig. 4-24). Air and water temperature The air and water temperatures recorded at the time caiman were located varied on a diel basis, reflecting environmental temperature changes. The mean water temperature recorded was 29.4°C (N=257) . No difference in water temperature was found between males and females locations, among microhabitat categories (see below) , or among size-classes (ANCOVA; p>0.05, adjusted for hour of day) . However, a significant difference in both air and water temperatures was found among habitat types (ANCOVA;

PAGE 118

Water Depth (cm) 1987 0 May Jun Jul Lagoon Small Males Water Depth (cm) Aug Sep Month Palm Savanna Medium Males Oct Nov Marsh Large Males Lagoon Small Males Aug Sep Month Palm Savanna Medium Males Marsh Large Males Dec 1988 Dec Figure 4-23. Mean water depths at which male caiman were encountered during the wet season, by size class. Values for lagoons and marshes were measured at deep water areas infrequently used by caiman. a) 1987 data b) 1988 data

PAGE 119

120 100 Water Depth (cm) 1987 May 120 100 80 60 40 1 20 Jun Jul Lagoon Small Females Water Depth (cm) Aug Sep Month Palm Savanna Large Females Oct Nov Dec Marsh 1988 May Jun Jul Lagoon ~*~ Small Females Aug Sep Month Palm Savanna Large Females Oct ~*~ Marsh Nov Dec Figure 4-24. Mean water depths at which female caiman were encountered during the wet season; by size class. Lagoon and marsh values as in Fig. 4-23. a) 1987 data b) 1988 data

PAGE 120

Ill p<0.01). Water temperatures tended to be the highest in palm savanna, marsh, and borrow pit habitats. (Fig. 4-25) . Air and water temperatures at the one stream site (in a densely forested region) were consistently low. Microhabitat Data on the use of microhabitats was collected during the wet season of 1987 and 1988 (N=1053) . Overall, caiman were most frequently encountered in open water, or among herbaceous vegetation in the water (Fig. 4-2 6; OW and IWIV categories) . Other frequently used microhabitat categories were among woody trunks or branches (ATB) , or in dense growths of the aquatic Thalia oeniculata (DTC) . The use of microhabitats varied on a seasonal basis, as well as between sexes (Kruskal-Wallis; p<0.05). No significant differences in microhabitat use were noted between 1987 and 1988 (Kruskal-Wallis; p>0.05). Much of the seasonal use of microhabitats reflected the changing savanna environment during the wet season. The utilization of open water areas tended to be high throughout the wet season, but the highest values were found during the months of peak dispersal from the dry season lagoons: June 1987 and July 1988 (Fig. 4-27). At this time, early in the wet season, rooted or floating aquatic vegetation was relatively sparse and more open water microhabitats were available. However, some aquatic macrophytes were present throughout the year and so a large percentage of the caiman

PAGE 121

O 33-Z3 O Cl CD Lagoon Borrow Stream Marsh Palm Forest Pit Savanna Habitat Type Figure 4-25. Mean air and water temperature (±SD) at caiman locations; by macrohabitat type.

PAGE 122

113 Frequency of Use (%) 0L0 OLV BUR OW IWIV UFV Physical Situation I Males EM Females Figure 4-26. Frequency of utilization of microhabitat types by sex. 1987/1988 data. OLO=on land in open; OLV=on land in vegetation; BUR=buried in substrate; OW=open water; IWIV=in water in vegetation; UFV=under floating vegetation; ATB=among trunks and branches; DTC=in dense colony of Thalia .

PAGE 123

Figure 4-27. Annual trends in microhabitat utilization by male and female caiman. Microhabitat types in Fig. 4-26. a) 1987 b) 1988

PAGE 124

115 was found in rooted herbaceous vegetation (IVIW) , even during the early wet season. Caiman were regularly found in the water among woody vegetation (Fig. 4-27; ATB) . This microhabitat category included caiman found among shrubs or small trees, partially submerged vines, or adjacent fallen logs. The peak use of this microhabitat was during the middle of the wet season when savanna water levels were at their highest and the wooded savanna habitats were flooded more extensively. One microhabitat of importance during the late wet season was Thalia colonies. Thalia aeniculata is a tall, rooted, aguatic plant (Marantaceae) which grows principally in the deep water sections of marshes, or in other localized depressions. Thalia sprouts soon after the beginning of the rains and grows throughout the wet season, reaching a stature in excess of 1.5-2 m by the beginning of the dry season. Many of the savanna areas that retain water during the late wet or early dry season are dominated by dense colonies of Thalia, providing excellent caiman habitat. Many of the caiman that disperse long distances from the dry season lagoons will utilize these microhabitats during the late wet season before returning to the permanent water sites. The utilization of Thalia colonies was guite extensive during the late wet season of 1987, but less so in 1988. The Thalia die during the ensuing dry season but with the return of the rains these same depressions are usually the first areas to flood. Caiman in the early wet season

PAGE 125

make extensive use of these holes and the cover offered by the dense accumulation of dead Thalia stalks. As with Thalia , the extent of floating aquatic vegetation ( Salvinia . Eichornia . Pistia) increases throughout the wet season and is utilized by caiman mostly during the late wet season. This was most notable during the 1988 wet season (Fig. 4-27b) . These areas provide excellent cover for the caiman and may also harbor a greater concentration of food resources than is found in open water areas. A small percentage of caiman was encountered on land, either in the open (0L0) , or hiding under vegetation (OLV) . Caiman on land and in the open were generally either basking, or moving overland. Caiman hiding in vegetation were most frequently encountered at the very beginning of the dispersal period (Fig. 4-27) , and this was especially evident in 1988. Individuals that dispersed prior to significant flooding were forced to hide under low-lying vegetation (typically small palms) until the marshes flooded. A somewhat similar behavior was observed when caiman would bury themselves, either partially or almost entirely, in the substrate. These individuals were typically found in muddy terrestrial areas, frequently well shaded, throughout the wet season (Fig. 4-27) . Caiman were also observed partially or completely buried, with just the tip of the nostrils exposed, along the fringes of wetland habitat during the dry season. The significance of burial

PAGE 126

117 and the use of terrestrial hiding sites is unclear, but may result from agonistic social encounters. The mean size of male caiman found in these two microhabitat categories during the wet season was significantly smaller than all other categories (F7 ^ 435=3 . 65 ; p<0.01). Because the outcome of agonistic encounters usually is highly correlated with caiman size, these males were probably subordinate individuals. The early dispersal from lagoons in the dry season by some small males may be a result of a low social dominance status, and the burial behavior during the dry or wet season likewise may serve to remove the individual from social contact with other, larger, caiman. The use of microhabitats differed significantly between males and females (Kruskal-Wallis; p<0.05; Fig. 4-26). Males tended to be encountered more frequently in open water habitats, and females were more prevalent among flooded woody vegetation, or in the Thalia colonies. These data suggest that males overall may prefer more open microhabitats than do females. The reproductive status of females also had a significant effect on the use of microhabitat (KruskalWallis; p<0.05), although this may be related primarily to differences in macrohabitat selection. During the early wet season, nesting females were largely found in open water habitats, whereas non-nesting females were mostly in herbaceous or woody vegetation (Fig. 4 -2 8a) . During the

PAGE 127

Microhabitat % Fixes iNesters Non-nesters Aug-Oct 87/8 Microhabitat Figure 4-28. Comparison of microhabitat use between nesting and non-nesting females. Microhabitats as in Fig. 4-26. a) June-July; 1987-1988 data b) August-October; 1987-1988 data

PAGE 128

actual nesting period, differences between nesting and nonnesting females were less evident (Fig. 4-28b) . The utilization of microhabitats was strongly influenced by macrohabitat (Fig. 4-29) . Open water sites were dominant in the lagoon, borrow pit and stream habitats. In palm savanna a wide selection of microhabitats was used, with open water (OW) and herbaceous vegetation (IVIW) being co-dominant. Similarly, in marshes a variety of microhabitats was utilized, but caiman were found most freguently among light growths of rooted herbaceous vegetation (IWIV) or, during the late wet season, in the dense stands of Thalia (DTC) . In the forested forest habitats, caiman utilized open water microhabitats primarily, or were among woody vegetation (ATB) . The few observations in bank habitats were mostly of caiman on land hiding under vegetation. Discussion The patterns of movement and home range found for caiman in this study are difficult to compare with the few other radio telemetry studies of crocodilians . Most of these studies have involved the American alligator (Alligator mississippiensis ; Joanen and McNease 1970, 1972, McNease and Joanen 1974, Taylor et al. 1976, Goodwin and Marion 1979, Taylor 1984), a larger, and more temperate crocodilian. The study of Hutton (1989) was conducted in a lake and the extreme differences in habitat obviate any

PAGE 129

120 Frequency ol Use (S| Frequency ol Use <%> OLO OLV BUR OW IW1V UFV ATB DTC OLO OLV BUR OW IW1V UFV ATB DTC Frequency ol Use (%) Frequency ol Use (*) Marsh l.ll OLO OLV BUR OW IWIV UFV ATB OTC OLO OLV BUR OW I WW UFV ATB DTC 80 80 40 20 Frequency ol Use (*) Palm twniM 80 80 40 20 Frequency ol Use (%) Font OLO OLV BUR OW IWIV UFV ATB DTC OLO OLV BUR OW IWIV UFV ATB OTC 80 1 1 Bank 80 PI ' — i 1 — i *m__BH_ OLO OLV BUR OW IWIV UFV ATB DTC Physical Situation Figure 4-29. Frequency distribution of microhabitat use by macrohabitat type. Microhabitat types as in Fig. 4-26.

PAGE 130

121 direct comparisons. Seasonal variation in alligator movement is principally related to annual temperature fluctuations, and not to extreme variation in the hydric environment. Alligators have a sexually dimorphic pattern of movement, with adult males having much larger home ranges than females or juveniles. Most studies have found adult female alligators have annual home ranges similar to, or slightly smaller than, those reported here for caiman (Joanen and McNease 1970; 8.5 ha, N=3 , Goodwin and Marion 1979; spring home range=15.6 ha, N=5, Taylor 1984; 56 ha, N=9) . However, adult male alligators have much larger home ranges than caiman, with reported values in the Louisiana coastal marsh averaging 2,623 ha (calculated from Joanen and McNease 1972 for three males followed for an entire year) . In a Florida lake, male alligator home ranges were somewhat smaller, but reached a seasonal high of 256.7 ha (N=4) during the spring. Movements of juvenile alligators are intermediate between those for adult males and females and typically were in the 100-300 ha range (McNease and Joanen 1974, Taylor et al. 1976). No differences in male or female movement patterns were found among juveniles (McNease and Joanen 1974) . Caiman on Hato Masaguaral followed a regular annual pattern of movements that was intimately associated with the savanna flooding regime. During the dry season the available wetlands habitats were reduced to a few lagoons or borrow pits. Caiman would concentrate in these bodies of

PAGE 131

122 water throughout the dry season, and disperse out into the flooding savanna habitats at the beginning of the rainy season. The annual pattern of movements on Masaguaral is similar to the qualitative descriptions of caiman movement in other parts of the llanos (Ayarzaguena 1983) , in the seasonal savannas of the Venezuelan Guyanan region (Gorzula 1978) , and in the Brazilian Pantanal (Crawshaw and Schaller 1980) , and appears to be a regular feature of caiman ecology in seasonally flooded savanna habitats. A similar pattern of seasonal migration was also noted by Ouboter and Nanhoe (1988) among caiman inhabiting a seasonal swamp habitat in Suriname. However, several differences were noted. Ayarzaguena (1983) estimated that only 30-60% of the adult males disperse from the principal water bodies on a ranch in the low llanos region of Venezuela. Ouboter and Nanhoe (1988) found dry season movements of two adult females in a river were fairly extensive (home ranges of 11 ha and 35 ha) , but although the caiman seasonally dispersed from the river into the surrounding flooded swamp, the authors implied that the wet season home ranges were relatively small. Contrary to the implication that caiman return to the same dry season lagoons, Schaller and Crawshaw (1982) found that 25 of 89 (28.1%) marked caiman tagged one year were sighted in different lagoons (borrow pits), up to 9.4 km away, during the following dry season. These reports suggest that although the seasonal dispersal and concentration are widely

PAGE 132

123 found among caiman, local biotic or abiotic factors may strongly influence movement patterns. This study found no evidence of the dispersal of caiman away from their natal population. All study animals remained in the vicinity of the study site over the 2.5 year period of investigation. The long-distance dispersal of animals may be infrequent among adults, possibly associated with extreme hydric conditions (very wet or very dry) . If long distance dispersal does take place regularly, it is likely to be done mostly by juveniles (Messel et al. 1981, Hutton 1989) , a size-class that was not included in this study. The exact timing of wet season dispersal varies according to the annual flooding regime. During years of early rains caiman would be able to abandon the dry season lagoons early, but conversely during unusually long dry seasons (as in 1988) the caiman are forced to delay dispersal. This delayed dispersal can alter the behavioral timetable of caiman as was seen in 1988 when courtship and mating activity began in the dry season lagoons prior to dispersal (chapter 8) . Delayed dispersal also appears to have a significant effect on the entire wet season pattern of habitat usage, restricting caiman to more permanent water sites and reducing the average dispersal distance. Following dispersal, a large percentage of the caiman remained within one kilometer of the dry-season refuge. No male-female differences were seen in dispersal distance or

PAGE 133

124 mean daily distance moved, nor were there any size-related differences in movement among the females. However, the data from this study suggest that wet season dispersal and movement patterns of male caiman are primarily regulated by the size, and resulting social rank of the individual caiman. Immediately following dispersal, males began to establish territories, principally in the deep water marsh habitats (see chapter 8) . From these territories males bellowed to attract females, and aggressively excluded other males. The results of these agonistic encounters were usually highly correlated with size, larger caiman being behaviorally dominant over smaller ones. These territories presumably served as the principal sites for mating and copulation. On Masaguaral, large males preferentially established territories closer to the dry-season lagoons. Because relatively few females dispersed long distances, as the distance from the male caiman's territory to the dry season lagoon increased, the probability of attracting females and mating diminished, so deep water areas adjacent to permanent water bodies would be preferred by breeding males. The size-related pattern of movement among males suggests that the larger males are occupying the preferred marsh habitats near the dry-season lagoons, and forcing the smaller males to disperse longer distances. The largest male from the Guacimos lagoon (115.4 cm SVL) tracked during this study established a territory in a marsh adjacent to

PAGE 134

125 the south end of the Guacimos lagoon. Conversely, the two males that dispersed approximately 4 km to the Matadero marsh were intermediate in size (95.8 cm, 94.5 cm SVL) . The two largest males from the Piscina population also remained in or around the Piscina virtually throughout the entire wet season. Overall, large males had small home ranges located near the dry season lagoons, Type 2 or Type 3 home ranges and small daily movement indexes (Table 4-5) . Table 4-5. Summary of movement patterns among size classes of male and female caiman on Hato Masaquaral. Size Class (cm SVL) Dispersal Distance (m) DMI (m/day) Home Range Type Home Range Size (ha) Males <90. 0 317 28.7 2 55.1 90.0-99.9 2,074 62.8 3 46.7 >99.9 452 38.4 2-3 19.1 Females <74.9 >74.9 807.1 729.2 61.5 43.2 3 2 65.4 1.9 Intermediate size males (90-100 cm SVL) are just entering the breeding population. Because of their relatively small size, and subordinate social rank they are largely excluded from areas near the dry season lagoons. Males in this size class tended to disperse much further

PAGE 135

than other adult males, had large Type 3 home ranges, and had large daily movement indexes (Table 4-5) . Small males (<90 cm SVL) , like the large males, did not disperse far from the dry season lagoons, but probably for different reasons. Males in the 80-90 cm SVL size range are physiologically mature (chapter 9) . However, because of their small size these males may not be capable of establishing and defending territories. It may be that it is not feasible for these small males to attempt to establish breeding territories until they reach a size of at least 90 cm SVL. If they do not attempt to establish territories, they may escape from the agonistic encounters with larger males that force the intermediate sized males to disperse. For this reason the smaller males may remain in the flooded habitats near the dry season lagoons. However, these small males appeared to avoid the deep water areas and spent much of their time in marginal areas, such as the forest habitat in 1987 (Fig. 4-14a,d). Intermediate sized males also appeared to be less fixed in their pattern of annual movements and home ranges varied more on a year-toyear basis. Overall, small adult males did not disperse far from dry season lagoons, had small daily movement indexes, but had large Type 2 home ranges (Table 4-5) . These data suggest that male caiman dispersal patterns (and activity ranges) may remain similar year after year on a short or medium-term basis, but that they may change

PAGE 136

127 significantly with the size, and the resulting social rank of the individual. Among females, there was no corresponding effect of size on dispersal distance or the daily movement index. However, small females did have larger home ranges than large females (Table 4-5) . A somewhat similar result was found among Nile crocodiles in a lake in Zimbabwe (Hutton 1989) , but the smaller home ranges of large females was apparently a result of attaining sexual maturity. On Hato Masaguaral reproductive state influenced dispersal distance, with nesting females tending to remain near the dry season lagoons and non-nesters dispersing farther. This may suggest that nesting females prefer the larger males located near the lagoons, but nevertheless some nesting females did disperse long distances. In fact, the two longest dispersal distances were by nesting females in 1988. Why then do some females disperse so far for nesting? If the nest was successful, the female would be forced to bring her pod of young back over the same distance to reach permanent water for the dry season, increasing the chances of high hatchling mortality during the move (chapter 10; fox depredation on moving hatchlings) . Also, long distance dispersal, and mating with males in distant territories, would result in pairing with smaller and perhaps less fit males. It is not known for certain how female dispersal patterns relate to courtship and copulatory behavior, but all the long-distance dispersing females in 1987 and 1988 made initially large

PAGE 137

128 displacements out of the dry season lagoons. This suggests that these females were not mating with males near the dry season lagoons, and then moving to distant nest sites. One advantage of long-distance dispersal for both male and female caiman is the lower caiman density and reduced competition for food or physical space. Another potential reason for long-distance dispersal by nesting females is to reduce the potential of predation on nests. Depredation levels are high in the areas near permanent water, and tend to drop off farther away. However, this relationship is only evident out to a distance of 3 . 0-3 . 5 km from permanent water (chapter 10) . Beyond this distance nest depredation rates appear to increase but the sample size is very small and few conclusions can be made about depredation intensity. Another consideration is that females that nest at great distances from their dry season lagoon may subsequently choose a different dry season body of water if the nest is successful. Instead of returning to one of the large dry season bodies of water, the female may seek out a small, isolated body of water (e.g., borrow pit, windmill pond) where survivorship of hatchling caiman is much higher (pers. obs.). However, none of the long-distance dispersing females in this study had a successful nest, so this hypothesis could not be evaluated. Another interesting finding of this study was the lack of any significant differences in wet season home range size or daily movement index between nesting and non-nesting

PAGE 138

129 females. Nesting has been tacitly assumed to restrict a female's movements (Ayarzagiiena 1983, Crawshaw 1987), thereby incurring opportunity costs in the form of lost opportunities to forage. However, nesting females on Hato Masaguaral exhibited no such restrictions and tended to move just as much as non-nesting females. The real "behavioral" costs of reproduction appear not to be associated so much with nesting as with the post-natal care of hatchlings (see chapter 10) . The results of this study have shown that seasonal variation in water level is the overriding factor controlling the movement patterns of spectacled caiman in the llanos. However, body size also plays a critical role, especially among males, and indicates that movement patterns reflect a complex interaction between the caiman's social and reproductive behavior (polygny) , the physical structure of the habitat, and annual changes in water level. Home range strategies may be important components of crocodilian life-histories, affecting reproductive fitness, feeding, growth and survivorship. The interrelations between social rank, sex, reproductive state and movement patterns in caiman and other crocodilians certainly warrant further study, both for a more complete understanding of crocodilain population dynamics, as well as for the development of conservation and management programs. Of particular interest would be a study of spectacled caiman inhabiting riverine habitats.

PAGE 139

CHAPTER 5 DIEL ACTIVITY PATTERNS Introduction The activity patterns of reptiles have been a topic of active investigation for quite some time. Most work has dealt with the daily activity cycles of terrestrial reptiles, particularly lizards (Brattstrom 1965, Heatwole 1976) , and how these are influenced by environmental variables. Daily activity cycles are known to vary seasonally (Mayhew 1964) , and temperature is one of the most important factors governing activity, but other environmental and internal factors are equally important. Lang (1976) found that juvenile alligators adjusted their daily activity to synchronize with an altered light cycle. The parietal eye has been shown to play an important role in regulating daily activity cycles in lizards (Stebbins and Wilhoft 1966) . It is likely that diel activity patterns are controlled ultimately by endogenous mechanisms (an "internal biological clock") responding to variations in external stimuli (temperature/ light cycles) (Heatwole 1976) . Thermoregulatory behavior has been the best studied reptile activity. Reptiles exhibit some ability to regulate their body temperature physiologically (Batholomew 1982), but most 130

PAGE 140

131 thermoregulation is done by behavioral means. Basking in many species of reptiles involves a complex series of behaviors which function to regulate body temperature (Huey 1982) . Because of their large size, crocodilians present a number of interesting guestions regarding behavioral thermoregulation (Spotila et al. 1972, Terpin et al. 1979). Recent work also indicates that there exists a dichotomy in the thermoregulatory strategies between the temperate and subtropical heat seeking species and the more tropical species that may avoid high temperatures (Lang 1987) . Investigations of diel activity cycles of crocodilians in the wild have dealt principally with daily movements between aguatic and terrestrial habitats. Diurnal patterns of basking behavior have been documented for a number of species of crocodilians (Cott 1961, Lang 1987) . The only investigation that has addressed the non-thermoregulatory activity pattern of a crocodilian, and nocturnal activity, was that of Rodda (1984) , with juvenile Crocodylus acutus in a lacustrine habitat. However, this last study was based on the interpretation of the signal strength from radios attached to the crocodiles, and not on direct observation, so was limited in its interpretation of activity. Previous investigations on the basking cycle of Caiman crocodilus have been conducted by Staton and Dixon (1975), Marcellini (1979) and Ayarzagiiena (1983). However, these studies only presented information on the number of animals observed on shore, and did not address other activities.

PAGE 141

132 Furthermore, few interpretations could be made regarding the proportion of the total population that was basking, or was engaged in any other activity, throughout the day. Some evidence exists suggesting that spectacled caiman do avoid high temperatures (Lang 1987) , but these observations have not been put into the context of daily activity patterns. On a daily basis, the activity patterns of crocodilians are influenced by other factors such as social interactions and feeding. Behavioral studies on crocodilians have examined diel variation in social behavior (Vliet 1987), but to date no work has been done on wild populations. The seasonal concentration of large numbers of caiman into lagoons and borrow pits in the llanos provides an ideal opportunity to examine diel variation in dry season activity. The objectives of this study were to guantify the pattern of land-water movements including the use of intermediate shallow water areas, observe levels of nocturnal activity, and examine diel variation in the occurrence of social behaviors. Daily variation in feeding behavior is addressed in chapter 7. Methods Diurnal behavior was studied at one borrow pit (18 m x 71 m) on the southern half of the ranch. Observations were made over 12 hr periods (06:00-18:00 h) using a vehicle as a blind. The vehicle was parked approximately 30 m from the lagoon approximately one-half hour prior to beginning

PAGE 142

133 observations. After an initial period caiman seemed to ignore the vehicle and individuals would even haul out and bask next to it. At 15-min intervals a count of all visible caiman was made, caiman were placed in one of three categories: basking (on land) , in the water (at the surface) , or on the edge (caiman resting in shallow water adjacent to the shore or partially hauled out) . The number of caiman underwater was calculated by subtracting the total number of caiman visible from the maximum total count made (usually during the 06:00 h count). Caiman in the water were also classified in terms of their body profile in the water: head only, head-back, head-back-tail, and headelevated tail arched (see chapter 8 for description of body postures) . Notes on social interactions and fishing behaviors were also kept continuously. Air temperature (Omega 871 hand-held digital thermometer) and wind speed (Dwyer wind speed indicator) were recorded hourly adjacent to the side of the vehicle facing away from the caiman. Nocturnal counts were made also from a vehicle at the same borrow-pit used for diurnal counts, as well as at a small natural lagoon (Merecure lagoon) . The number of caiman visible was the average of three counts made at halfhour intervals using a 200,000 cp spotlight powered with a 12 v car battery. Wind speed and air temperature were recorded hourly. At the Merecure lagoon water temperature (10 cm deep) was also recorded using an Omega 871 hand-held

PAGE 143

134 digital thermometer and a 10-m-long chromel-alumel thermocouple. Results Diurnal Activity Twelve-hour diurnal observations were made on six separate days (5 Jan, 16 Feb, 23 Mar, and 2 May 1986, 18 Jan 1987 and 18 Dec 1988) . The weather was very similar on all days. Mean temperatures during the observation periods ranged from 29.3°C (5 Jan 1986) to 31.4°C (23 Mar 1986). Most days were a mixture of sunny or partly cloudy weather. Overcast weather was only experienced from 12:00-14:00 h on 18 Jan 1987, and from 06:00-09:15 h on 2 May 1986. A brief drizzle fell at 08:32 h on this latter date. Diurnal activity patterns were guite variable. Basking was most commonly observed during the late afternoon. Only on 5 Jan 1986 was a strong morning basking peak seen. Typically a small number of caiman emerged to bask during the morning, and some would remain out during the middle of the day (Fig. 5-1). A peak in "edge" animals, either basking in shallow water, or partially hauled out on shore, was seen during the hottest part of the day (Fig. 5-2). Shallow water basking was more freguently observed than caiman partially hauled out on shore. The largest percentage of the population observed basking (completely out of the water) was 44.9% in the late afternoon of 18 Dec 1988. The corresponding maximum values for the other days

PAGE 144

135 100 Percent Maximum Count 6:00 8:00 10:00 12:00 Hour 14:00 16:00 18:00 — Submerged -+— in water at surface -*Basking -B On edge Figure 5-1. Diurnal activity cycle of caiman in Borrow Pit #9, all data combined. Percent of Maximum Count 6:00 7:00 8:00 9:0010:0011:0012:0013:0014:0015:0016:0017:0018:00 Hour Figure 5-2. Freguency of shallow water basking and partially hauled-out (PHO) caiman at Borrow Pit #9, all data combined.

PAGE 145

136 were: 5 Jan 1986-36.4%, 16 Feb 1986-17.7%, 23 Mar 198613.7%, 2 May 1986-13.5%, and 18 Jan 1987-32.6%. These values suggest that the number of caiman basking declines during the dry season. The high percentage of animals in the water at the surface early in the morning may have been an artifact caused by submerged caiman coming to the surface when the vehicle was driven near the lagoon, and remaining at the surface. During nocturnal observations a smaller percentage of caiman was seen at the surface in the early morning (see next section) . However, it is apparent that a large percentage of caiman remain submerged during the day (Fig. 5-1) . Frequently 60-80% of the observed population was submerged. Peak values were usually reached during the mid to late morning hours. The great majority of caiman in the water were observed with only the head at the surface ("head" posture). The percentage of caiman in elevated "head-back" or "head-backtail" postures, which are frequently associated with social encounters, was always relatively small but showed a peak during the late morning hours (Fig. 5-3), in association with increased movement around the lagoon. Few social interactions were seen during these observations. However, the largest number of observed chases and bellows occurred in the early morning during the first hour of observations (Fig. 5-4). This agrees with the timing of bellowing and

PAGE 146

137 Percent of Caiman in Water 1 00 80 60 40 20 0 7:00 ll 3k U 9:00 11:00 13:00 Hour 15:00 17:00 Head Head-back I I Head-back-tail HETA Figure 5-3. Diurnal variation in body posture of caiman at water's surface at Borrow Pit #9. Number Bellows Chases 6:00 8:00 10:00 12:00 Hour 14:00 16:00 18:00 Figure 5-4. Timing of observed social interactions in Borrow Pit # 9.

PAGE 147

138 chasing noted in the Casa area throughout the year (Fig. 55, and chapter 8) . Caiman feeding behavior also showed a daily rhythm and this is discussed in chapter 7 . Nocturnal Activity Nocturnal activity was guantified principally in terms of the number of individuals observed on the surface throughout the night. Counts were made throughout the night once at each of the two lagoons. In both cases the number of caiman at the surface declined throughout the night (Fig. 5-6, 5-7). Discussion Basking on land, typically associated with heat seeking behavior, was found to have a bimodal distribution with a peak in the mid-morning and again in the late afternoon and evening. Although this result agrees with previous studies of caiman basking cycles (Staton and Dixon 1975, Marcellini 1979, Ayarzaguena 1983) the data indicated that only a small proportion of the total population, less than 25%, was out basking at any one time. The small number of caiman that bask suggests that one of the principal functions of caiman thermoregulatory behavior is not heat gain, but heat avoidance. Many of the late afternoon baskers would remain on shore well into the night. During nocturnal counts in the

PAGE 148

139 Number up i A i r 1 M . MlLLU 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour i Chases EM Fights EH Patrols Adult Distress Calls Figure 5-5. Daily variation in frequency of observed social behaviors in the Casa area (see chapter 8) .

PAGE 149

140 Number Visible mph or °C 20:00 21:00 22:00 23:00 0:00 1:00 2:00 3:00 4:00 5:00 6:00 Hour — Caiman visible Ta EZD Wind Speed Figure 5-6. Nocturnal census of caiman in Borrow Pit #9 on 24 March 1989. 40 35 30 25 20 15 10 5 Number Visible mph or C —* % — — ^ — — % — — * — — * — — * — — * — \r — i — — i — . — i — — i — . § JUL H , i 1 35 30 25 20 15 10 5 0 19:30 20:30 21:30 22:3023:30 0:30 1:30 2:30 3:30 4:30 5:30 6:30 Hour Caiman visible Ta Tw Wind Speed Figure 5-7. Nocturnal census of caiman in the Merecure Lagoon on 30 Jan 1989.

PAGE 150

141 dry season, over one-third of the caiman were seen on shore at some bodies of water. Nocturnal counts on the Capanaparo River in southern Apure state found 22.6% of observed caiman were on shore (Thorbjarnarson and Hernandez, in press). Similar patterns of temperature regulation have been reported for C. porosus (Lang 1987) and C. niloticus (Loveridge 1984) . Caiman may come ashore in the evening in order to avoid social interactions with other caiman, but this behavior may also be thermoregulatory in nature and involve convective and radiative heat loss. The large percentage of animals submerged during the day may also reflect avoidance of high ambient temperatures (Smith 1979) . Submerged caiman may be trying to remain cool, or to avoid the social interactions caused by the high density conditions in the dry season lagoons. Other reasons for submergence are subsurface foraging and avoidance of wave action. In shallow water lagoons, submergence by itself may not be an effective heat avoidance behavior, but may be accompanied by burial in bottom sediments (Thorbjarnarson, personal observation) . The significance of heat-avoidance behavior is not clearly understood. Reptiles typically regulate their body temperature within an optimal range (Huey 1982) , and heatavoidance may simply reflect body temperatures exceeding that range. Alternatively, caiman may be deliberately seeking lower body temperatures to reduce their rates of metabolism and energy demands during the stressful dry

PAGE 151

142 season. More work needs to be done on caiman thermal relations and energetics before this question can be properly addressed. Nevertheless, some caiman do exhibit heat seeking behavior. Roughly 10% of the population emerged to bask in the morning, presumably because of low body temperature. Basking in shallow water peaked during the hottest part of the day when few animals were hauled out on shore. Previous studies of caiman thermoregulation have not distinguished between basking on land, and basking in shallow water, but this difference may be important. Peak edge basking, with at least part of the caiman in the water, occurs during the hottest part of the day, and may accelerate heat gain due to the superior heat transport properties of water. However, this may be counteracted to some extent by evaporative cooling. Late morning was also when the largest number of caiman were observed in elevated float posture. It has been suggested that these elevated postures may serve a thermoregulatory function (Smith 1979) as a means of raising body temperature. The results of this study suggest that this may be true as increased high profile postures at this time were not related to peaks in social or feeding behavior (chapter 7) . Why some caiman may exhibit heat seeking behavior when the majority of the population is trying to avoid heat is uncertain. Some of these individuals may have low body temperatures resulting from being submerged in deep water.

PAGE 152

143 Some individuals may also be sick or digesting meals, as it has been demonstrated experimentally that crocodilians will seek higher temperatures after feeding and when infected with pathogens (Lang 1987) . Caiman activity peaks during the cool early morning hours. Social behaviors such as bellowing, agonistic encounters and courtship and mating all peak in the first hour after dawn (chapter 8), when body temperatures are presumably at their lowest. A similar pattern of activity was noted by Rodda (1984) for juvenile Crocodvlus acutus in Panama. Thus, thermoregulatory behavior in crocodilians appears to be very different from that of diurnal lizards who apparently thermoregulate to maintain high body temperatures during periods of activity (Lang 1987) .

PAGE 153

CHAPTER 6 CAIMAN DIET Introduction The diet of crocodilians was first treated in a quantitative fashion by Cott (1961), who analyzed the feeding habits of Nile crocodiles ( Crocodylus niloticus ) in relation to crocodile size and location. Since that time, a number of dietary studies have been conducted on other species: e.g. Alligator mississippiensis (Fogarty and Albury 1967, Chabreck 1971, McNease and Joanen 1977, Delany and Abercrombie 1986) ; Crocodvlus porosus (Taylor 1979) ; Crocodvlus iohnsoni (Webb et al. 1983c) ; Crocodylus acutus (Seijas 1988, Thorbjarnarson 1988). Several studies have examined the diet or feeding behavior of the spectacled caiman, Caiman crocodilus (Gorzula 1978, Schaller and Crawshaw 1982, Magnusson et al. 1987), including a large amount of information for the Venezuelan llanos (Staton and Dixon 1975, Seijas and Ramos 1980, Ayarzaguena 1983, Fitzgerald 1988) . In this study, the diet of caiman in the vicinity of Hato Masaguaral is described. Emphasis is placed on quantifying dietary variation with caiman ontogeny and on a seasonal basis. Given the extreme seasonality of the llanos 144

PAGE 154

145 ecosystem, it has been implicitly assumed that annual variation in dominant prey items and the amount of prey consumed exists. Although a large body of information exists for caiman in the llanos, no studies have specifically addressed the question of seasonal variation in diet. The diet of crocodilians is also very dependent on the nature of the habitat (Magnusson et al. 1987) , implying that a species' dietary habits change on a regional basis. Comparative studies are needed in different habitat types to determine the extent of regional variation in caiman diet. Most previous studies of caiman diet in the llanos have been conducted in the low llanos region (Seijas and Ramos 1980, Ayarzagiiena 1983, Fitzgerald 1988) where wet season flooding is more extensive than on Hato Masaguaral. Although the work of Staton and Dixon (1975) was conducted on the same ranch as this study, their sample size was relatively small (40) . Furthermore, a comparison of the findings of Staton and Dixon (1975) with the results of this study provide us with the opportunity to compare the diet of one population over a 10 year interval. Materials and Methods Between October 1984 and June 1989 dead Caiman were collected from a 44 km stretch of highway between the ranch and the Rio Guarico bridge near the town of Calabozo. Only caiman DOR (=dead on road) less than two days were

PAGE 155

collected. Caiman with ruptured stomachs would infrequently have stomach contents strewn throughout the body cavity and these individuals were not included in dietary studies. In addition, from May 1986 to June 1987 a monthly sample of caiman from areas adjacent to the main Hato Masaguaral study site was killed for studies of diet and reproduction. Stomach contents of all caiman were first collected in a plastic basin, then rinsed with water in a fine sieve to remove mucous and gastric juices. Prey items were classified as either freshly ingested, partly digested or fragments. All fresh and partly digested items were weighed on an O-Haus Triple Beam balance to the nearest 0.1 g. Prey were divided into one of 10 principal taxonomic categories: Pomacea (an ampullarid snail) , arachnids, Coleoptera, other insects, crabs ( Dilocarcinus dentatus ) , fish, amphibians, reptiles, birds, or mammals. Identification of prey was made to the lowest taxon possible. Prey categories were ranked using a modified version of the technique used by Webb et al. (1982). To reduce the bias favoring the sampling of slowly digesting prey (e.g. chitinous prey; Jackson et al. 1974, Garnett 1985), fresh and partly digested prey were ranked according to weight. For each caiman, a score was calculated for each prey category by subtracting the rank of that prey category from the total number of fresh or partly digested prey categories present in that animal, plus one. The score for each prey category was the percentage of the total score for that

PAGE 156

147 animal. For example, if crabs were the second most abundant prey in a caiman containing three prey categories, the rank would be calculated as (3-2) +1=2, and the overall score would be 2/3+2+1=33.3%, providing a total score of 33.3. Because so few animals had fresh or partly digested prey, individuals with prey fragments were given a arbitrary low score of 10 for that prey category. No attempt was made to correct for loss of prey material due to digestion. The total weight of stones (gastroliths) was recorded, along with the presence of vegetation. The presence of parasites was noted consistently in the sacrificed caiman but was not always evident in the DOR individuals. Results The stomach contents of 274 caiman were examined: 72 were sacrificed and 202 found dead on the road. Caiman were classified by size into three categories: size-class II (2059.9 cm snout-vent length), size-class III (60-89.9 cm SVL) , and size-class IV (>89.9 cm SVL). Hatchling caiman (size class I; <20 cm SVL) were not included in this analysis. No significant differences were found between sexes in the utilization of any of the 10 principal prey categories (Mann-Whitney U; p>0.05). A comparison of DOR and sacrificed caiman found no significant difference in any of the prey categories except for fish (Mann-Whitney U; p<0.01), which comprised a significantly larger part of the diet of sacrificed caiman. An examination of the effects of

PAGE 157

148 size class and season on fish diet in DOR and sacrificed caiman shows a similar seasonal pattern for both samples, with dry and late wet season peaks. However, utilization by size-class differs, with sacrificed size-class III (60-89.9 cm SVL) caiman eating more fish than DOR size-class III individuals. I felt that this difference was not great enough to warrant dividing DOR and sacrificed caiman into two separate categories analysis, so all individuals were grouped together. Frequency of Prey Occurrence The most commonly encountered prey remains in caiman stomachs were Pomacea snails, freshwater crabs ( Dilocarcinus dentatus ) , and Coleoptera (Table 6-1) . Coleoptera found in the stomach were predominantly aquatic beetles of the family Hydrophilidae (Table 6-2) , with terrestrial scarab beetles being somewhat less common. Other insects, primarily aquatic Hemiptera of the family Belastomatidae , were also commonly consumed. Fish were found in 17.5% of the caiman stomachs examined, with the three most common species being Synbranchus marmoratus (an eel-like member of the Synbranchi formes) , Hoplosternum littorale (an armored catfish; Callicthyidae) , and Hoplias malabaricus (a predatory characoid) . Mammals remains were less commonly encountered and usually consisted of small, unidentified rodents. Remains of amphibians, reptiles, birds and spiders were infrequently found.

PAGE 158

149 Table 6-1. Frequency of occurrence of prey categories, vegetation, gastroliths and empty stomachs in caiman from Hato Masaguaral by size-class. No. refers to the number of caiman that contained any remains of the indicated prey Prey Cateaorv No. II % SizeIll No. % Class IV No. % Total NO. % ColeoDtera 22 34 . 9 32 19 . 6 4 8 . 3 58 21.2 Other Insects 16 25.4 30 18.4 3 6. 3 49 17.9 Crabs 13 20.6 42 25.8 10 20. 8 65 23.7 Pomacea 18 28.6 61 37.4 16 33. 3 95 34.7 Arachnids 1 1.6 3 1.8 0 0. 0 4 1.5 Fish 4 6.4 31 19.0 13 27. 1 48 17.5 Amphibians 1 1.6 7 4.3 1 2. 1 9 3.3 Reptiles 1 1.6 7 4.3 4 8. 3 12 4.4 Birds 3 4.8 6 3.7 0 0. 0 9 3.3 Mammals 2 3.2 24 14.7 4 8. 3 30 11.0 Empty 12 19.1 38 23.3 17 35. 4 67 24.5 Vegetation 9 14.3 49 30.1 15 31. 3 73 26.6 Gastroliths 24 38.1 88 54.0 39 81. 3 151 55.1 The three most commonly encountered prey (Pomacea, crabs and coleoptera) were precisely the prey categories that would be expected to be favorably biased due to slow digestibility. The ranking method reduces this bias, but still indicates that Pomacea were the most important prey for caiman on Masaguaral. Following Pomacea in importance were fish, and then crabs (Table 6-3) . Coleoptera, other insects and mammals were of intermediate importance, and reptiles, amphibians, birds, and arachnids were the least important prey. Only 85 of the 274 (31%) caiman examined contained fresh or partly digested prey. A large percentage (25%) had no food remains at all in their stomachs (Table 6-1) . The A

PAGE 159

150 Table 6-2. Prey items identified from caiman stomachs on Hato Masaguaral. Values indicate the percent of caiman with identified items from the indicated prey category that contained the identified prey item (e.g., hydrophilids were found in 75.6% of the caiman containing remains of coleopterans) . Prey Category Prey Item Percent of Total Coleootera Hydophilidae YD . 6 Scarabaeidae 12 . 2 Unidentified 14 . D Other Insects Belastomatiaae Q "3 1 Orthoptera 1U . 3 Unid. Hemiptera D . 1 Mantidae A • D Lepidoptera larvae 2.6 Unidentified 2.6 Crabs uiiocar mus i nn n Pomacea Pomacea sp. 1UU . u Arachnids Unidentified 100 . 0 Fisn t>ynDrancnus in q Hoplosternum 1 O Q 18 . 9 nopiias XO>« Licni luae o . X Unid. Siluriformes D . 4 Pimelodidae D . 4 Pyaocentrus 5.4 Unidentified 35.1 Amphibians Pseud is 4 ^ . y Liepuoaacuvius 14 . J Pleuroderma 14.3 Unidentified 42.9 Reptiles Iauana 54.5 Caiman 16.7 Chironius 9.1 Unid. Ophidia 9.1 Unidentified 9.1 Birds Porohvrula 28. 6 Dendrocyana 14.3 Icteridae 14.3 Unid. Ciconiif ormes 14.3 Unidentified 28.6 Mammals Unid. Rodentia 44.4 Dasvous 5.6 Didelohis 5.6 Unidentified 44.4

PAGE 160

mean mass of fresh and partly digested prey was 15.6 g for all caiman examined. Table 6-3. Overall prey rank by prey category. Prev Category Rank Prey Cateaorv Rank Pomacea 9.8 Mammals 3.3 Fish 7.7 Reptiles 2.4 Crabs 6.4 Amphibians 1.6 Coleoptera 3.9 Birds 1.2 Other Insects 3.6 Arachnids 0.8 Seasonal Variation The freguency of caiman with fresh prey varied somewhat throughout the year but was highest in the late dry (April) and middle wet (September) season (Fig. 6-1) . Mean prey mass did not show any obvious seasonal trends (Fig. 6-2) , but utilization of prey items exhibited seasonal shifts that were presumably related to the variations in the seasonal abundance of those prey (Fig. 6-3a,b) . Crabs were consumed throughout the year, but a sharp peak was evident in September. Pomacea snails peaked in the early wet season (July-August) , and again in October. Coleopeterans were also consumed mainly in the early (June-July) and late (November-December) wet season. Other insects, principally belastomatid water bugs, were most evident in the late dry season (April) and late wet season (November-December) . Fish were the dominant vertebrate prey and were consumed mainly in the dry season, although a smaller peak

PAGE 161

Figure 6-1. Percent of caiman examined containing fresh prey, by month and size-class. a) size-class II caiman b) size-class III caiman c) size-class IV caiman d) all size-classes

PAGE 162

Percent 100 80 80 40 20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Oec Month Percent Size -class II — fel 111. 100 80 60 40 20 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Percent Size-class iv ..l.lll. I Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Oec Month 100 80 60 40 20 Percent Overall III lllll. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

PAGE 163

Mass (g) 300 200 | Bscn ESS SC III EZDSC IV Overall 1 ; | 1 j 1 \ 11 n 1 n i 11 1 j j ij ii i Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 6-2. Mean mass of fresh prey by month and caiman size-class.

PAGE 164

155 40 35 30 25 20 15 10 5 0 Mean Rank Invertebrates MM r El 1 li ] Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Arachnids H! Other Insects HI Coleoptera 1^ Pomacea Siiiiiil Crabs 40 35 30 25 20 15 10 5 0 Mean Rank Vertebrates a ail 1 1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Fish Amphibians EH Reptiles EM Birds HM Mammals Figure 6-3. Prey rank by month, all caiman size-classes. a) invertebrate prey b) vertebrate prey

PAGE 165

156 was evident in the late wet season. Few fish were eaten during the early wet season (June-July) . Reptiles were principally a dry season prey, and birds and amphibians were consumed largely during the wet season. Mammal remains were encountered most frequently during the late dry, and early wet seasons. A comparison of prey rank among the early dry (December-February) , late dry (March-May) , early wet (JuneAugust) , and late wet (September-November) seasons revealed that fish, Pomacea and crab utilization varied significantly during the year (Kruskal-Wallis ; p<0.05). Ontogenetic Dietary Shifts As caiman size increased so did mean prey mass. Mean prey mass for size-class II individuals was 8.3 g, for sizeclass III 48.1 g, and 128.5 g for size-class IV caiman ( F 2,75= 4 56 / p<0.05). With the exception of reptiles ( F l,2 =646 8 / p<0.01), within any one prey category prey size did not increase with caiman size (Fig. 6-4) . This indicated that larger caiman did not consume larger crabs (or other prey) than did smaller caiman, but that the increase in mean prey mass was primarily due to the consumption of prey taxa of a larger mean size. Ranking prey utilization by size class for each of the 10 prey categories showed size-related trends (Fig. 6-5) . Among juvenile caiman (size-class II) Pomacea snails, Coleoptera, crabs and other insects were the most important dietary

PAGE 166

157 1000 100 Mean Prey Mass (g) 10 0.1 ISCH GnUDsCIII E3SC IV Overall Arach Other Coleopt Pom Crab Fish Amph Rept Birds Mam Prey Category Figure 6-4. Mean mass of fresh prey for each prey category, by caiman size-class.

PAGE 167

158 Pr#y Rank Pray Rank Colaoptara SCII SO III BO IV Ovwall SCIl SOW SO IV Owall Figure 6-5. Ontogentic changes in prey utilization for each of the 10 prey categories.

PAGE 168

159 components. Size-class III individuals also fed predominantly on snails, but fish replaced Coleoptera as the second principal prey item. Large caiman prey to a much greater extent on vertebrates, with fish, crabs, mammals and reptiles being the dominant prey. Significant trends for increased utilization with greater caiman size were found in fish (Kruskal-Wallis; p<0.05), and mammals (p<0.05). Increasing trends were also seen in crabs and reptiles but these were not significant. Decreasing trends in Coleoptera (p<0.01), and other insects (p<0.05) were significant. The percentage of caiman containing fresh prey did not show any clear relationship with size (SC 11=16.1%, SC 111=38.0%, SC IV=27.1%), and no relationship was found among seasons (F3 '262 =1 * 16 > P>0.05), or caiman size-classes and the frequency of empty stomachs (F2 f 262 =1 « 72 ' P>0.05). Vegetation, aastroliths and parasites Vegetation was found in the stomachs of 26.6% of the caiman examined. Grasses were the most commonly found type of vegetation (76.2% of stomachs with vegetation), followed by leaves (19.0%), woody material (14.3%), and seeds (14.3%). Vegetation was usually found in small amounts, although in some cases large boluses of grass (to 183 g) were ingested. Most vegetation was probably taken in accidentally along with food items. Stones (gastroliths) were encountered in more than one half of the caiman

PAGE 169

160 examined (Table 6-1) . The presence and total mass of gastroliths were clearly related to caiman size. The frequency of gastrolith occurrence increased from 38.1% in size-class II caiman, to 81.3% in size-class IV individuals (Table 6-1), and mean gastrolith mass increased from 0.9 g to 32.9 g (F2,271 =32 * 46 ' p<0.001). The greatest mass of stones found in one caiman was 255.2 g in a 110 cm SVL male. Ascarids (Sebekia) were encountered in 24.6% of the stomachs, and 54.9% of the body cavities of sacrificed caiman. Discussion Crocodilians are opportunistic predators. Even slender snouted species (e.g. Crocodvlus iohnsoni ) , which have been presumed to be primarily piscivorous, are catholic in their choice of prey (Webb et al. 1982) . The relatively broadsnouted Caiman crocodilus is certainly best classified as an opportunistic predator. Caiman feed on a tremendous variety of prey, aquatic and terrestrial, vertebrate and invertebrate. To a certain extent the diet of crocodilians reflects the availability of prey (Webb et al. 1982), but some selection does occur (Fitzgerald 1989) , which may reflect foraging techniques (see chapter 7) . The results of this study, as well as those of other studies on caiman and other crocodilians, suggest that diet is strongly affected by temporal and spatial variability in prey abundance. Superimposed on this is an ontogenetic shift in prey choice.

PAGE 170

Diet in Relation to Caiman Size In this and four other studies of caiman diet in the Venezuelan llanos, similar trends in prey utilization have been noted (Table 6-4) . The principal prey of juvenile caiman consists almost entirely of invertebrates. Coleoptera and belastomatids are the dominant food item in most areas, although in this study Pomacea snails were the principal prey of small caiman. Crabs were taken by all size-classes of caiman, but on Hato Masaguaral they were more important in the diet of larger individuals. As caiman size increases the importance of vertebrate prey increases. Fish were the principal prey of large caiman in all of the studies except for that of Ayarzaguena (1983) . However, because Ayarzaguena ranked prey by percent occurrence, his data overrepresented the importance of the slow digesting Pomacea and crabs. My study is the only one to find an important dietary role for reptiles and mammals. Similar ontogenetic shifts in diet have been noted in all studies of crocodilian diet (e.g., Cott 1961, Webb et al. 1982, Delany and Abercrombie 1986, Magnusson et al. 1987, Thorbjarnarson 1988, Fitzgerald 1989). Dietary shifts are evident in both size and the taxonomy of the prey (Webb et al 1982) . Ontogenetic dietary shifts presumably reflect the ability of larger individuals to capture larger prey, but habitat selection and foraging mode may also play crucial roles (Magnusson et al. 1987, Seijas 1988).

PAGE 171

162 Because Staton and Dixon (1975) used very broad sizeclass categories in their description of caiman diet on Hato Masaguaral in 1974, it is difficult to compare the results of these two studies. However, the dominant prey items were similar in both studies. Among fish, Hoplosternum and Svnbranchus . were freguently consumed in both studies. Similar trends in the consumption of snails and insects were also noted. The one big difference between the studies was in the consumption of crabs, which was high in 1985-9 but relatively small in 1974. The reason for the higher consumption of crabs in this study is unclear, and in the absence of data on relative prey abundance few conclusions can be drawn. Prey Type and Seasonal Variation On Hato Masaguaral the four most important taxonomic prey categories were Pomacea snails, fish, crabs and Coleoptera. Pomacea remains were encountered in 34.7% of the caiman stomachs examined. The durable opercula remain in the stomach long after any soft tissue had been digested. Caiman were found with up to 66 opercula in the stomach, but the length of time over which these snails were ingested is unknown . Fish were the dominant vertebrate prey taken and were the most important prey of any type for large caiman. The fish most commonly ingested were the eel-like Svnbranchus marmoratus . and the armored catfish, Hoplosternum littorale .

PAGE 172

163 Table 6-4 . Comparison of principal dietary components of Caiman from five different studies in the Venezuelan Llanos. Prey ranked in decreasing order of importance by size-class (cm SVL) . Sources are: AAyarzagiiena 1984, BSeijas and Ramos 1980, CFitzgerald 1989, Dthis study, EStaton and Dixon 1975. Location Principal Prey Caiman Size-class (cm SVL) 20-59.9 60-89.9 >90 A. £,1 Frio Insects Pomacea Pomacea (Apure) Crabs Crabs Crabs Fish Fish Fish Amphibians Shrimp Shrimp Mammals Insects is. nancecax 14-45.9 46-66.9 >67 (Apure) Coleoptera Fish Fish Shrimp Shrimp Shrimp Fish Coleoptera Belastomatidae C Pi nftro Overall I TD /™\ -y +• n <->c i t /—. \ ^ rUI LUyueSd ^ Fish Coleoptera Other Insects Crabs Pomacea D. Masaguaral 20-59.9 60-89.9 >90 (Guarico) Pomacea Pomacea Fish Coleoptera Fish Crabs Crabs Crabs Mammals Other Insects Other Insects Reptiles E. Masaguaral <20 >20 (Guarico) Coleoptera Fish Other insects Coleoptera Other Insects Pomacea

PAGE 173

Few fast-moving mid-water column species were consumed. Synbranchus were taken more commonly in the wet season whereas Hoplosternum were eaten more frequently in the dry season. The importance of fish in the caiman diet showed dramatic annual variation (Fig. 6-3) . Fish were consumed principally in the dry season, with a smaller peak in the late wet season. Few fish were consumed during the early wet season (May-Aug) . This lull in fish consumption occurred when caiman moved from their dry season concentrations and dispersed across the freshly flooded savannas. Fish biomass in diked modules in Apure state reach their lowest levels during this period (Taphorn and Lillyestrom 1984) and this likely results in few fish being consumed by caiman in the early wet season. The crab, Dilocarcinus dentatus . was an important wet season component of caiman diet on Masaguaral. Many of these crabs pass the dry season buried in the mud and emerge with the beginning of the rains. Females carrying juvenile crabs are seen during the early wet season, and these newborn crabs continue growing throughout the wet season (Taphorn and Lilyestrom 1984) . The reason for the sharp peak in crab consumption in September is unknown. Studies in the low llanos region in Apure state (Seijas and Ramos 1980, Ayarzagiiena 1984) have found shrimp, Macrobrachium yelski and M. amazonicum . to comprise a large part of the diet of caiman of all size classes. However, shrimp were

PAGE 174

165 not an important dietary item at Hato Pinero, and were never found in caiman at Masaguaral. Coleoptera comprise an important part of the juvenile caiman diet. Staton and Dixon (1975) found that beetles were the principal prey of hatchling caiman. Coleoptera remain an important dietary component among juveniles, and are consumed by even the largest caiman. Most studies have found aguatic beetles (principally Hydrophilidae) to be most freguently consumed, but on El Frio terrestrial insects were more commonly ingested (Ayarzagviena 1983) . Aguatic hemipterans (principally Belastomatids) were important components of the diet of size-class II and III caiman on Masaguaral, and among hatchling caiman at Pinero as well (Fitzgerald 1989) . Both Coleoptera and hemipterans appeared to be most commonly ingested at the very beginning and the end of the wet season. On Masaguaral, both these insect groups were very abundant following the first rains of the year, when they would disperse at night and be attracted to lights, and predation may be facilitated by this mass dispersal. The peak at the end of the wet season may be associated with dropping water levels and the concentration of these prey. Vertebrate prey other than fish were important in the diet of large caiman on Masaguaral. This differs from other llanos studies, none of which found mammals, birds or reptiles to figure significantly in the diet. Rodents and iguanas were regularly taken by adult caiman. Also, two

PAGE 175

instances of cannibalism were noted, and in both cases hatchling caiman were eaten by sub-adult males. It is also interesting to note that amphibians were a minor prey item in all studies of caiman in the llanos. Anurans were relatively abundant, especially following the first rains when explosive breeding takes place. Gorzula (1978) found anurans to be a major prey of caiman in the Venezuelan Guyana region. Amount of Food Consumed Caiman consume relatively small guantities of prey. Fewer than one third of the caiman examined in this study had more than fragments of prey in their stomachs, and the average caiman had only 15.6 g of recently ingested prey, which represented 0.2% of the caiman's body mass. One guarter of all caiman stomachs contained no food at all. Studies on passage rates of prey items in Alligator mississippiensis indicate that remains of food items such as crustaceans, snails, mammals, birds and turtles may remain in the stomach for five days or more after ingestion (Delany and Abercrombie 1986) . Some relatively undigestible items, such as snail opercula, may remain for considerably longer periods. Only 31% of the caiman examined had fresh or partially digested prey, indicating that caiman were not eating small guantities of food on a continual basis. These results are in accord with the low metabolic demands of crocodilians (Coulson and Hernandez 1983), and low-energy

PAGE 176

167 lifestyle of vertebrate ectothenns (Pough 1980) . Caiman, like all reptiles, do not maintain a high constant body temperature via the metabolic production of energy. Ectotherms have low basal metabolic rates and depend largely on the anaerobic production of energy for bursts of activity (Bennett 1982) . As a consequence the energetic demands of crocodilians are relatively low. One of the obvious benefits of this is the ability to survive periods of low food availability. Juvenile caiman feed to a large extent on aguatic invertebrates. Invertebrate prey are largely unavailable during the extended dry season. In some areas juvenile caiman may have to live through an annual period of some 3-5 months with very little food. The dry season is an extremely stressful period for juvenile caiman, but apparently few juveniles starve during this period. Their low-energy lifestyle allows them to survive through this period that would be lethal for animals with larger energetic demands. Gastroliths Stomach stones, or gastroliths, have been reported for a number of species of crocodilians such as Alligator mississippiensis (Neill 1971, Delany and Abercrombie 1986), Crocodvlus niloticus (Cott 1961) , and C. iohnsoni (Webb et al. 1982). Despite the fact that large areas of the study site are without stones of any sort, a very large percentage of caiman on Masaguaral had gastroliths. The frequency of

PAGE 177

168 gastroliths reported here was higher than found by Staton and Dixon (1975) for caiman at the same site (32.5% in adult/subadult caiman) . The acquisition of gastroliths is quite likely an active process (Cott 1961, Peaker 1969, Davenport et al. 1990) . There has been much debate over the presence of gastroliths, and two basic functions have been ascribed: a hydrostatic role and a gastric trituration role (see Webb et al. 1982, and Davenport et al. 1989 for discussion) . The hydrostatic function has usually been couched in terms of providing stability while in the water (Cott 1961) . However, perhaps a more important role of gastroliths is for increasing potential dive time (Seymour 1982). The ingestion of stones increases the caiman's specific gravity. Buoyancy in crocodilians is regulated by the amount of air in the lungs, and in order to dive a crocodilian has to expel air from its lungs and become negatively buoyant. The presence of stones in the stomach would allow the caiman to become negatively buoyant with a larger volume of air in its lungs, thereby increasing the amount of time it could spend underwater. In summary, caiman on Hato Masaguaral were opportunistic predators, taking a wide variety of prey. There was a clear ontogenetic shift in diet, with small caiman feeding largely on aquatic invertebrates, and larger individuals switching to a larger proportion of vertebrates. However, even among large caiman the consumption of invertebrates, particularly Pomacea snails and crabs, was

PAGE 178

169 important. My results have also shown that a seasonal variation in prey utilization occurs, but in the absence of information on seasonal changes in the prey base it is not possible to state whether these differences result from the variability of prey abundance, or from seasonal changes in foraging technigues. A comparison of these results with the findings of other studies in the Venezuelan llanos indicates that a certain amount of regional variation in diet exists, but the overall trends of prey utilization are similar. A comparison with a previous study conducted at the same ranch suggests that caiman diet has changed little over a 10 year interval, with the exception of an increased consumption of crabs.

PAGE 179

CHAPTER 7 FEEDING BEHAVIOR Introduction The techniques that crocodilians use to capture prey are myriad. Cott (1961) stated of the Nile crocodile, "Its methods of hunting, capture and disposal of prey are as varied as the habits of its victims." Nevertheless, in contrast to studies of diet, very few attempts have been made to describe crocodilian foraging techniques, and many of the descriptions in the literature deal principally with large crocodiles taking large prey (Cott 1961, Pooley 1982) . The prevalence of accounts of the ambush techniques used by these crocodiles for large mammalian prey gave crocodilians the reputation of sit-and-wait predators. Pooley (1982) was the first to disperse some of the myths concerning crocodilian foraging techniques, referring to crocodiles as "highly active and versatile" hunters, and reporting a number of active techniques by which they would procure prey. Judging by the diverse assemblage of prey items found in their stomachs (chapter 6) , spectacled caiman are very versatile predators. A previous study by Schaller and Crawshaw (1982) on Caiman crocodilus has been the one 170

PAGE 180

171 examination of the feeding behavior of a crocodilian in the wild. Several other studies have described feeding technigues, but all were in captivity (Fleishman et al. 1988, Fleishman and Rand 1989, Davenport et al. 1990, Thorbjarnarson 1990) . While making observations on caiman social behavior, I was provided the opportunity to observe also a number of foraging technigues. The objectives of this work were to describe some of the technigues that caiman used for prey capture and, where possible, to guantify the efficiency of these different technigues. One limitation of the study was that the majority of the observations had to be made during the day, but information on six different feeding technigues was obtained. Methods Observations on caiman feeding behavior were made at several different locations on Hato Masaguaral between October 1984 to June 1989. Most observations of trapping and surface feeding were made in Borrow Pit #9 (18 m x 71 m) located on the southern part of Hato Masaguaral (5 January, 16 February, 23 March, and 2 May 1986, 18 January 1987 and 18 December 1988) . Caiman in the borrow pit were observed continuously throughout the day from 06:00 to 18:00 h, and all feeding behaviors were recorded. The weather was very similar on all days. Mean temperatures during the observation periods ranged from a low of 29.3 °C (5 January

PAGE 181

1986) to a high of 31.4°C (23 March 1986). Other observations were made at the Guacimos Lagoon in association with studies of social behavior (chapter 8) . Nocturnal observations of other feeding techniques were made opportunistically during the course of other caiman-related work. Results Spectacled caiman have a large repertoire of feeding behaviors. Observations of feeding behavior on Hato Masaguaral were placed in six categories: surface feeding, weir fishing, ambush feeding, jumping, underwater feeding, and carrion feeding. Surface Feeding Surface feeding was an easily observed technique because it frequently took place during the day. Surface feeding utilized three different techniques: stationary snaps, float-fishing, and trapping. Caiman at the surface of the water were frequently observed making rapid side swipes with their jaws in an attempt to capture prey just under the surface of the water. Side swipes were made when the caiman was immobile at the surface, usually with just the head visible above the water (stationary snaps) , or while slowing swimming in a semiemergent posture (float-fishing; Fig. 7-1). While floatfishing, caiman would typically extend one or both forelimbs

PAGE 182

173

PAGE 183

174 and hold them rigidly at a 90° angle to the body, frequently with the fingertips extended upwards. The extended forelimbs were usually just under the surface of the water, but the extended fingertips could frequently be seen protruding out of the water. In this posture the caiman would slowly swim by sculling the tail, making occasional sideswipes to capture fish. The rapid sideswipes occasionally were accompanied by a forward sweeping motion of the tail. Observations of successful predation using this technique indicate that the caiman are feeding on small fish (probably characoids) at or near the surface of the water. Many species of characoids school close to the surface of the water where they feed on organic debris blown in from surrounding terrestrial areas (D. Taphorn, pers. comm.). The float-fishing technique appears to be one that herds these fish into the region around the caiman's head where they can be captured with a rapid side swipe of the jaws. Caiman would also make quick, snapping motions from a stationary position at the surface of the water as well. These snaps, from a stationary position, or while swimming were classified into two broad categories: midwater snaps (>1 m from the shoreline) , and edge snaps. Midwater snaps were more commonly observed (Fig. 7-2a) , and had a bimodal temporal distribution with peaks in the morning and the afternoon. The smaller number of snaps during the midday period was due to reduced overall activity during this

PAGE 184

175 Number of Snaps Midwaier bnaDs 6:00 8:00 10:00 12:00 Hour 14:00 50 40 30 20 10 Number of Snaps 16:00 Edge Snaps 6:00 8:00 10:00 12:00 Hour Number of Snaps 14:00 16:00 Total Snaps 6:00 8:00 10:00 12:00 Hour 14:00 16:00 Figure 7-2. Surface feeding snaps by hour of day, borrowpit #9. a) midwater snaps b) edge snaps c) total snaps

PAGE 185

176 period (chapter 5) . Edge snaps were observed throughout the day, but were more common during the morning hours (Fig. 72b) . Float-fishing and stationary snaps had a higher success rate near the edge of the lagoon (Fig. 7-3b; mean 10.5%), than in open water away from the edge (Fig. 7-3a; mean 4.15%). The overall mean success rate for surface snaps (stationary and float-fishing) was 7.2%. Trapping was another behavior that concentrated fish in a small area where they could be more easily captured (Fig. 7-4a-c) . Typically, the caiman would begin by slowly swimming towards shore with its body oriented more or less perpendicular to the shoreline. As it approached the shore the caiman would slowly turn its body and orient itself parallel to the shore. In this manner the caiman would angle in towards the shore and, when close enough, sweep its tail and head towards the shore, effectively trapping a volume of water between itself and the shoreline. When trapped in this fashion, fish would begin to swim around frantically and were captured by a quick sideswipe of the caiman's jaws. Frequently fish would escape by jumping over the caiman's back, but on one occasion the caiman was able to grab an escaping fish in mid-air. Trapping was successful in 11.3% of the observed attempts, and was seen throughout the day (Fig. 7-5, 7-6) .

PAGE 186

177 50 40 30 20 10 % Success Midwater 6:00 % Success 10 50 40 30 20 10 % Success Hour 8:00 IttOO 12:00 14:00 16:00 Hour Edge 6:00 8:00 10:00 12:00 14:00 16:00 Total 6:00 8:00 10:00 12:00 14:00 16:00 Hour Figure 7-3. Percent of surface feeding snaps successful, borrow-pit #9. a) midwater snaps b) edge snaps c) total snaps

PAGE 187

Figure 7-4. The trapping technique used by caiman to capture fish.

PAGE 188

Number 20 | 15 10 1 1 1 1 I.I 6:00 8:00 10:00 12:00 14:00 16:00 Hour Figure 7-5. Number of trapping behaviors observed by hour, Borrow Pit #9. % Success 100 80 60 40 20 J | | | | |_ li 6:00 8:00 10:00 12:00 14:00 16:00 Hour Figure 7-6. Percent of trapping attempts successful by hour of day, Borrow Pit # 9.

PAGE 189

180 While trapping, caiman would make use of irregularities along the shoreline, frequently forcing fish into small coves or indentations in the shore. Fishing in Flowing Water In shallow, moving water caiman would use two techniques for catching prey. Both of these feeding behaviors were seen principally at night when it was difficult to make extensive observations without disturbing the animals. The first technique was to orient the body parallel to the flow of water and capture prey by making rapid sideswipes. This behavior was observed several times, including once when seven caiman were oriented in two rows facing into the current (31 December 1984) , with their mouths slightly opened and their heads elevated in the water. Twenty-seven feeding attempts were observed, of which 8 (29.6%) were successful. Caiman were more frequently observed using a variation of this feeding technique that I refer to as weir fishing. Weir-fishing caiman would orient their bodies almost perpendicular to the flow of water, but with the head angled slightly downstream. By positioning its body across the current, the caiman would divert much of the stream flow, forcing it to pass by its head. In this posture the caiman would sit with its mouth partially open and snap at passing fish. Of 9 attempts by weir-fishing caiman to capture prey, four (44.4%) were successful. In one case an extremely

PAGE 190

181 large caiman (ca. 2.7 m) was observed weir fishing in a stream on another ranch. This caiman had selected a natural dam formed by a rock ledge in a location where the stream (averaging approx. 3-4 m across) narrowed to a width of about 2 m. Weir fishing in this location allowed the caiman to divert all the fish moving up or down the stream past his jaws. Ambush Feeding "Ambush" feeding was a nocturnal sit-and-wait technique seen not infrequently among caiman less than 90 cm SVL. It consisted of a caiman's sitting motionless in shallow water, among floating or emergent herbaceous vegetation, with its head and neck protruding out through the vegetation in an "alert" posture (head held parallel to the surface of the water) . No observations of prey capture were made by caiman in this ambush posture, but it was assumed to be prey capture technique. From this position, caiman would have a good view of prey moving through or over the surface of the vegetation. In most cases where caiman were seen in this posture they were in floating mats of Eichhornia . Jumping Jumping behavior was commonly seen in shallow water during the dry season, particularly in the early dry season (Fig. 7-7). Jumps were made by pushing off with the hind limbs, propelling the caiman upward and forward with the

PAGE 191

182 Jumps/mln 1.4 N-182 1.2 1 0.8 0.6 0.4 0.2 n N-175 N«7 N«45 N=5 Feb Mar Apr May Jun Month Figure 7-7. Frequency of feeding jumps observed by month, Guacimos Lagoon. Data for period 1985-1988.

PAGE 192

183 forelimbs folded against the body. This would result in the caiman's only partially emerging from the water, and only rarely was more than one-third of the body visible out of the water. When reentering the water the caiman would lean off to one side or another, avoiding a mid-ventral ("bellyf lop") entry. Jumps were freguently followed by a sharp turn of the body, up to 180°, and a sideways sweeping of the head. Caiman would typically initiate a jump while submerged, or less freguently from a floating posture. Jumps were either single, or repeated. On occasion caiman were seen jumping into schools of small caracoids, many of which would wildly leap out of the water when the caiman jumped. As a fishing technique, jumping had a very low success rate. Of 414 jumps observed at the Guacimos lagoon only four (1%) resulted in the observed capture of a fish. In the borrow pit where 12 hr activity observations were done, 71 jumps were made without one observed capture. Underwater Feeding Observations on underwater feeding behavior were impossible because of the highly turbid water. However, on a number of occasions caiman were observed searching for underwater prey while in shallow water. Caiman would place their head underwater and make tight circling movements of up to 360°. This circling behavior was not unlike the movements made following jumps, and both presumably are attempts to locate and capture prey very close to the

PAGE 193

184 caiman's body. On two occasions caiman followed this circling behavior by extending their bodies parallel to the shore and moving sideways in a trapping behavior. Discussion Three of the prey-capture techniques described here (surface snaps, trapping and weir fishing) are very similar to those reported by Schaller and Crawshaw (1982) for Caiman crocodilus vacare in the Pantanal of Brazil. All the feeding behaviors discussed here (with the exception of the ambush technique) appear to be fish-catching techniques. Adult caiman in the llanos feed largely on fish (chapter 6) , yet lack the morphological specializations (long slender snout) for piscivory present in other crocodilians such as the gharial (Thorbjarnarson 1990) . Caiman have a relatively broad, unspecialized jaw morphology, and rely instead on a suite of behavioral specializations for catching fish. These techniques serve to concentrate fish into small areas where they are more easily captured (trapping, float fishing, weir fishing) , or to capitalize on surprise (jumping) . Although underwater feeding was never observed, the majority of the caiman's prey are captured below the surface of the water. Caiman apparently do not have good underwater vision (Fleishman et al. 1988), and do not need any visual cues for underwater prey capture (Fleishman and Rand 1989) . In the case of the llanos, this is intuitively apparent

PAGE 194

185 because most dry season bodies of water are extremely turbid and have zero visibility. Underwater, caiman may be locating fish by touch, or perhaps by using a mechanoreceptive system involving integumentary sensory organs (ISOs) . ISOs are small pit-like structures present in the integument of all crocodilians (Brazaitis and Garrick, unpublished data) . The ultrastructure of ISOs in caiman are similar to mechanoreceptors seen in other reptiles (von During 1973a, b) . These organs, or other functionally similar structures (Siminoff and Kruger 1968) , may provide caiman with clues to the presence of moving prey underwater. Caiman feeding behaviors were all relatively low-energy foraging techniques. Weir fishing is the epitome of a lowcost, highly effective foraging technique that capitalizes on a strongly developed rheotaxis among llanos fish. Seasonal "lateral migrations" (Lowe-McConnell 1975) are common among fishes living on seasonal floodplains, and are usually associated with feeding or reproduction. Typically, at the end of the wet season fishes tend to move up channels, and down them at the end of the dry season. During the dry season the tendency to move up current and out of drying anoxic ponds was common among several species with accessory air breathing apparatuses. Especially notable in this respect was Hoplias malabaricus . large numbers of which were observed moving up small currents that would only cover a small fraction of their body. Weir

PAGE 195

186 fishing behavior has been also reported by Ayarzagiiena (1983) . Quick lateral strikes were the principal technique used to grab prey. However, only a small percentage of these strikes were successful. Strikes near the edge of a water body (10.5% success) tended to have a higher success rate that strikes in midwater (4.2%), but the overall success rate of 7.2% was somewhat lower that the 15.9% reported by Schaller and Crawshaw (1982) . Quick lateral strikes are presumably how caiman capture other prey as well. At night caiman were frequently seen in shallow water under mats of vegetation with their head and neck protruding above the plant material in an "alert" posture. By remaining motionless in this posture caiman may be able to ambush prey (e.g. amphibians, reptiles) that are nocturnally active. A similar technique was described by Ayarzagiiena (1983) as a method by which large caiman captured small capybara ( Hvdrochoerus hvdrochaeris ) . Quick lateral strikes can also be used to capture birds as well. On 21 February 1986 a caiman grabbed a flying black skimmer (A. Schmitz, pers. comm.). Jumping is an unusual foraging technique. This behavior was first described by Staton and Dixon (1977) as a courtship display. The authors did note, however, that five of 150-200 jumps resulted in the capture of fish. Ayarzagiiena (1984) implied that in most cases jumping behavior was related to fishing behavior. Fitzgerald (1989)

PAGE 196

187 classified jumping as a foraging technique, noting that it was the only energetically expensive foraging behavior observed. During this study jumping was observed among all size-classes, including hatchlings who would jump in a nearly identical fashion in attempts to capture aquatic insects in shallow water. Jumping was most commonly noted, however, among adult caiman during the dry season. Nevertheless, the use of this foraging technique remains somewhat enigmatic, as it is the most energetically costly of all the fishing techniques observed, and the least effective. Caiman also will feed on carrion when it is available (Staton and Dixon 1975) . One individual was found with a maggot-infested opossum (Didelphis marsupialis ) in its stomach, indicating that the prey had been eaten as carrion. Another caiman had consumed a whistling duck ( Dendrocvana sp.) riddled with shotgun pellets. Caiman were also observed feeding on dead birds, as well as dead caiman. Larger individuals would also frequently try to chase smaller caiman and steal their food. This study has shown that in lieu of morphological specializations for the capture of fish, caiman relied on a number of behavioral specializations. These fishing techniques facilitated prey capture, principally by bringing the prey into close proximity to the jaws, or by confusing the prey. Nevertheless, the majority of prey that were targeted by these techniques were surface, or mid-water

PAGE 197

188 column dwelling species, which together comprised a small proportion of the caiman's diet (chapter 6). Crabs, snails and bottom-dwelling fish were the prey most frequentlyconsumed by caiman, yet very little is known about foraging techniques for these animals. Also, little is known about ontogenetic shifts in foraging that may coincide with dietary shifts (chapter 6) , and the investigation of these topics should be the thrust of future work on caiman feeding behavior.

PAGE 198

CHAPTER 8 BEHAVIORAL AND SOCIAL ASPECTS OF REPRODUCTION Introduction The study of crocodilian behavior is still a relatively new topic. Little was known about social behavior prior to the 1960 's, and suggestions of behavioral complexity were often denounced as fable (e.g., Neill 1971). Some work on the behavior of wild Nile crocodiles was done in the 1960's (Cott 1961, Modha 1967), but the study of crocodilian social behavior did not come into its own until the 1970' s with the landmark investigations of Garrick and Lang (1977) and Garrick et al. (1978) on captive populations of American alligators and American crocodiles. To this day much of what we know regarding crocodilian social behavior is based on studies of captive populations (e.g., Vliet 1986, 1989, Lang 1987a) . Pough (1980) pointed out the interlocking physiological and morphological features of ectothermal tetrapods which are manifested in a low-energy lifestyle guite different from that of mammals and birds. This low-energy approach to life has important conseguences for the behavior and ecology of all ectotherms. The social behavior of reptiles, while 189

PAGE 199

190 generally less complex than in endotherms, is difficult to compare with that of birds and mammals because they occur on different time scales (Lang 1987) . Prolonged bouts of activity are rarely seen among amphibians and reptiles who depend, to a large extent, on anaerobic metabolism during peak activity levels (Bennett and Dawson 1972, Bennett 1982) . Because of their reduced aerobic scope, reptiles cannot sustain high levels of activity, for doing so results in rapid exhaustion due to the depletion of metabolic energy stores and the accumulation of lactic acid. This low-energy lifestyle dictates that most reptilian behavioral acts do not follow one another in rapid seguence as they may among mammals and birds, but may be separated by minutes or even hours (Pough 1980, Watanabe 1980, Lang 1987). This lengthened time span often makes reptilian behaviors difficult to comprehend for human observers, who may misinterpret inactivity for a lack of social interaction. Crocodilians are among the most behaviorally sophisticated of the reptiles. Recent work has demonstrated that crocodilians possess social behaviors that are on a level of complexity similar to birds and mammals. Communication involves visual, vocal, non-vocal acoustic, tactile and olfactory signals (Garrick and Lang 1977, Garrick et al. 1978, Vliet 1989). Although social displays among crocodilians are complex, there are many interspecific similarities, suggesting that these behaviors, like

PAGE 200

crocodilian morphology, are generally conservative (Vliet 1989) . A good understanding of the social behavior of the caiman is essential for the interpretation of other aspects of its reproductive ecology. Studies of the life-history and population ecology of reptiles have for the most part ignored the role of social behavior, which given the behavioral complexity of these animals cannot be justified. Many important ecological parameters such as reproductive cycles, mortality schedules, and patterns of movement and habitat utilization reguire information on the species' social behavior for adeguate assessment. Investigations of social behavior are also of importance for conservation and management purposes (Lang 1987) . For instance the caiman management program in Venezuela is based on a harvest of adult males, on the assumption that the species has a polygynous mating system and an excess of reproductive males. In-depth treatments of crocodilian social behavior to date have invariably been restricted to captive populations (Garrick and Lang 1977, Garrick et al. 1978, Vliet 1989), where artificial conditions may result in aberrant behavior. However, the study of social behavior in the wild is extremely difficult because crocodilians are usually found at low densities, are wary of human presence, and occupy areas where behavioral observation is difficult (e.g., due to the presence of extensive aguatic vegetation) . For these

PAGE 201

192 reasons information on the social behavior of wild crocodilians is limited (Lang 1987) , and previous studies of wild spectacled caiman populations contain only rudimentary information on social interactions (Alvarez del Toro 1974, Staton and Dixon 1977, Gorzula 1978, Ayarzagiiena 1983) . This study is the first attempt to conduct a relatively detailed examination of the social behavior of a species of crocodilian in the wild. Despite the limitations of having a large, open population of largely unmarked individuals, behavioral observations were greatly facilitated by the annual concentration of caiman into readily observed groups inhabiting shallow-water lagoons almost completely free of aquatic vegetation. The objectives of this study were to describe the behavioral and social aspects of reproduction in a wild population of spectacled caiman, to compare these observations with what is known about the social behavior of other crocodilians, and to relate these to other aspects of the reproductive ecology of this species. Methods Observations on caiman reproductive behavior were made seasonally from May 1985 to June 1989, principally in two areas on Hato Masaguaral: the Guacimos Lagoon, and in borrow pits and natural marshes surrounding the ranch living quarters (Casa area) . The Guacimos Lagoon is a shallow, 18 ha body of water situated in a mosaic of open savanna and thorn woodland (see chapter 2). Observations at the

PAGE 202

193 Guacimos Lagoon were made from a blind constructed of scaffolding material set up along the western edge of the lagoon, or from a stationary vehicle. Most observations at the Guacimos Lagoon were made during the early morning hours when the caiman were most active. Observations were initiated before sunrise and continued until the rising wind created significant wave action in the lagoon, causing most of the caiman to submerge. Observations were restricted to the western one-fifth of the lagoon (ca. 3 ha) and were made by continually scanning the lagoon and recording noteworthy behaviors or interactions. Comparable observations were made less frequently during the midday or afternoon hours. Most observations were made during the late dry season and very early wet season (March-June) . Certain caiman could be identified from numbered plastic tail tags, or from the presence of a radio transmitter on the tail. However, in most cases the caiman were not individually identifiable. Among unknown individuals large males were easily recognized by their size, but smaller males were similar in size to adult females. In the latter case assumptions could frequently be made concerning the sex of the individuals based on their behavior. During observations individuals were placed into the following size categories: size-class IV (>90 cm snoutvent length, adult males), size-class III (60-90 cm SVL, adult females and small adult or subadult males) , and sizeclass II (20-60 cm SVL, juvenile caiman) .

PAGE 203

In the Casa area, observations were made opportunistically from in or around the house. Behaviors (bellows, chases, fights, patrolling, adult distress calls) were noted whenever seen or heard. When possible, the number of sub-audible vibrations associated with each bellow was recorded. The majority of the observed behaviors in the Casa area took place in a small lagoon immediately adjacent to the house, or in one of several borrow pits within 50 m. Bellows from caiman in these bodies of water were easily detected, and would usually wake me up when I was asleep, allowing me to keep 24 hour records. The same was true to a lesser degree for other behaviors that produced loud splashing noises (e.g. chases or fights) . Behavioral interactions of marked and unmarked caiman were followed either continuously over short periods of time (to one hour or more) or periodically over longer periods of time. The contents of the submandubular and clocal glands of scarificied caiman were examined on a monthly basis from June 1986-July 1987. Gland contents were classified as either no musk present, waxy secretions present, or liquid secretions present. Following Garrick et al. (1978), I divided signals into three categories: vocal, non-vocal acoustic, and visual signals. Many of the behaviors involved several sensory modalities and were placed in one category or another according to the apparently dominant sensory input.

PAGE 204

195 Results Catalog of Behavioral Signals Visual signals Visual communication among caiman involved a complex series of body postures and movements. Postures comprised the presentation of the head, back or tail above the water's surface in varying amounts or in different orientations. The principal corporal posture components for caiman in the water are briefly described below. Each posture component is identified by an abbreviation (e.g., H) , which is used together with others to denote overall body posture (e.g., HBT=head, body and tail visible above the water's surface). Head postures . a. Head Low (HL)Head low in the water. Eyes, nostrils and part of cranial table visible. b. Head (H)Normal head posture in water. Eyes, nostrils, cranial table and most of the mid-dorsal snout region visible (Fig. 8-1) . c Head Elevated (HE)Head elevated in the water. Eyes, nostrils, cranial table and most of lateral maxillary region visible. d. Head High (HH)Head further elevated in water. Entire maxillary region, teeth and part of mandibular region visible (Fig. 8-2) . e. Head Alert (HA) Head elevated completely out of the water. Entire lateral mandibular surface visible.

PAGE 205

Figure 8-1. Caiman in typical "head" (H) posture. Figure 8-2. Caiman in a "head high" (HH) posture.

PAGE 206

197 Mandible more or less parallel with the water's surface. f. Head Oblique (HO)Head inclined at 30-40 degree angle out of water (Fig. 8-3) . Body postures . a. Back (B)Back slightly elevated in the water. Only mid-dorsal region visible b. Back Elevated (BE)Body inflated. Mid-dorsal and upper dorso-lateral region visible. Tail postures . a. Tail (T)Only dorsal-most region of tail visible. b. Tail Elevated (TE)Tail slightly elevated in water. Less than one-third of the mid-lateral tail surface visible. c. Tail Arched (TA) Middle region of tail obviously elevated in the water. More than one-third of the midlateral tail surface visible (Fig. 8-4) . d. Tail Oblioue (TO)Basal region of tail lifted at a 45-60 degree angle out of the water. Virtually every combination of these head, back and tail postures was seen among caiman on Hato Masaguaral. However, only certain combinations (described below) were regularly used in behavioral interactions. Male caiman demonstrated the most complex series of body postures in

PAGE 207

Figure 8-3. Caiman in a head inclined, tail elevated posture . Figure 8-4. Caiman in a head elevated tail arched posture (left) being approached by a caiman in a head elevated back elevated tail elevated posture (right) .

PAGE 208

199 association with thermoregulation, patrolling, agonistic encounters and courtship. Complex posturing among females was less freguently observed. A total of 10 visual signals are described. Yawning, included as a visual signal for alligators by Garrick et al. (1978) , was observed infreguently in this study but was not included here because yawns played no obvious social role. Staton and Dixon (1977) included "jumping" in a list of courtship displays for Caiman crocodilus . However, in my opinion jumping is a feeding behavior and is not related to courtship. During this study most jumps were seen in the early to mid dry season, well before the onset of courtship activity. Jumps were also most commonly observed during the afternoon hours, a time when few courtship encounters were noted. Furthermore, jumping did result in the capture of prey (see chapter 7) . Jumping was used as a technigue for prey capture by caiman of all size-classes, including hatchlings. Submergence. In a social context, subordinate caiman would submerge when they were challenged by a patrolling male, and unreceptive females would submerge when approached by a courting male. In both cases submergence was freguently accompanied by underwater movement away from the approaching individual. Submergence also plays a role in courtship activity. Apparently receptive females were observed to submerge and remain in place in front of an

PAGE 209

200 approaching male who would then proceed to mount and attempt copulation. Surface fleeing . Another antisocial response to an approaching male was to swim away without submerging. This behavior served a similar role as submergence, but did not remove the individuals from visual contact. Surface fleeing was occasionally accompanied by grunting vocalizations. In some cases the fleeing animals would be chased by the approaching dominant male. Patrol . A patrol was a slow, leisurely swim, with the dorsal region of the head, torso and tail visible, usually in a HBT, HEBTE, or HEBETE posture. Patrolling behavior was only observed in adult males and was most prevalent in June, at the beginning of the rainy season. Patrolling caiman occasionally stopped and raised their tail and head slightly, entering a HEBTA or a HETA signalling posture. Patrolling males approached other caiman, who generally submerged or swam off. Approaches of equally large males frequently led to agonistic face-offs. During the mating season approaches of receptive females initiated courtship. Inflated posture. This posture was an elevated, obviously aggressive stance with the body inflated, partially uplifted by the limbs, and head in an alert posture, usually with the occipital region raised higher than the rest of the skull (Fig. 8-5) . Inflated postures were frequently accompanied by tail wagging. Inflated postures were regularly seen following aggressive encounters

PAGE 210

201

PAGE 211

202 involving large males. Females defending pods of hatchlings from other caiman would also occasionally adopt the inflated posture following an encounter. Tail wagging . The tail wag consisted of a vigorous side to side motion of the tail given from an elevated, or inflated posture. Tail wags were frequently given by adult males either before, or following a chase. Chase . Chases were vigorous pursuits where the caiman rapidly propels itself forward, the anterior part of body out of water, with a forceful thrust of its tail, and occasionally with a push from the hind legs. During the chase the caiman would swim with its body inflated, legs folded at its side, mouth slightly open, and head held high out of water. Following a chase, the dominant individual would frequently adopt a stationary inflated posture which lasted up to several minutes, which was infrequently accompanied by submergence of the snout and vigorous bubbling through the nostrils. Chases were occasionally preceded by, or followed with, a tail wag by the dominant animal . The dominant caiman usually began the encounter by slowly swimming toward the subordinate, then suddenly speeded up, lunging with a rapid thrust of the tail. The subordinate animal would submerge and rapidly swim off underwater, rarely escaping by running out onto land. The vast majority of observed chases involved large males, but

PAGE 212

203 adult females defending hatchlings were also observed chasing caiman away from their pods. Head elevated tail arched posture . This posture was identical to the head-emergent tail-arched posture described by Garrick et al. (1978) for alligators (Fig. 8-4). The HETA posture was most commonly seen among adult males in the late dry or early wet season. The HETA was a common display just before, during, or after bellowing choruses, but was also seen unassociated with bellowing. Adult males would normally adopt this posture before bellowing, but a HETA display did not mean that the individual would bellow. Adult females would occasionally adopt the HETA posture prior to chasing caiman away from pods of hatchlings. Adult male caiman in the HETA posture in the Guacimos Lagoon were always well spaced with a minimum nearest neighbor distance of approximately 5 m. Head alert tail arched posture . This posture was used by adult males when approaching females for courtship. Typically the male began by approaching a female from a HETE posture, gradually assuming a HATA posture as he moved forward. The HATA posture has not been reported in the courtship of any other crocodilian. Staton and Dixon (1977) reported a social signal they termed the "tail display," which apparently includes both the HETA and the HATA, and noted that their observations of copulation always were preceded by this display.

PAGE 213

204 Head oblique tail oblique posture . The HOTO display was the characteristic posture from which male caiman bellowed. Rarely, males would enter the HOTO as if to bellow, but would not, and occasionally males were seen in a tail oblique posture without elevating the head. Snout lifting . During courtship, females would raise their heads into a head-inclined posture for rubbing the head or neck of the male. Snout-lifting as a submissive posture (Garrick and Lang 1977) was never observed among caiman. Non-vocal acoustic signals Sub-audible vibrations . The sub-audible vibration (SAV) is an infrasonic signal produced by caiman during bellowing and courtship. The mechanism for the production of these vibrations is unclear but presumably involves the rapid contraction of muscles in the dorso-lateral region of the chest and abdomen (K. Vliet, pers. comm.). These vibrations cause a physical agitation of the water over the caiman's back, adding a characteristic visual component to this signal. Headslapping. In American alligators headslaps consist of a vigorous slapping of the ventral surface of the head against the water, accompanied by a rapid opening and shutting of the mouth (Garrick et al. 1978, Vliet 1989). On Hato Masaguaral, no direct observations of caiman headslapping were ever made. Nevertheless, on five

PAGE 214

205 occasions a sound very similar to the headslap of an alligator was heard. In four of the five instances, the sound was followed by 1-4 SAVs. In three of the five instances a large adult male was seen in the area from where the sound originated. Bubbling . Adult male caiman frequently bubbled vigorously through their submerged nostrils. Bubbling was most commonly observed following a bellow, but caiman would also bubble after chases. Vocal signals Bellowing . Bellows were principally comprised of an infrasonic signal (SAV) , but at the end of each bellowing bout caiman would typically produce a weak audible bellow consisting of 2-3 vocalizations (see Bellowing behavior , below) . Hisses . Prolonged hisses are produced by the slow release of air from the lungs and serve a threatening function. Hisses were only heard by females defending their nest or pods of young, or by adults and juveniles when cornered in shallow water or on land during capture procedures. Grunts . Caiman grunts were given by individuals of all size classes. Caiman hatchlings would grunt from within the egg as a signal for the female to open the nest. Groups of hatchlings and juveniles were especially vocal, and captive caiman would vocalize much more frequently than captive

PAGE 215

206 Orinoco or American crocodiles kept under identical conditions. Hatchling grunts were short vocalizations between 0.1 KHz and 2 KHz. Occasionally, hatchling grunts were accompanied by longer growling vocalization. With increasing size the frequency of the grunts dropped. Grunts by hatchling and juvenile caiman were given in a variety of circumstances: in response to human approach, to the approach of an adult caiman, following the bellow of an adult male, or in response to the movement of other hatchlings . Adult caiman emitted a series of grunts when they felt threatened. These grunts were produced by individuals escaping from an approaching male, or as a response to the approach of a human. Grunts were also used as a form of communication between adult females and pods of hatchlings. Females approaching hatchlings frequently grunted several times and the hatchlings responded with similar, but higher pitched, grunts. I observed females leading hatchling caiman on overland treks more than one kilometer from the nearest water. On two occasions females leading young were heard to vocalize in an apparent attempt to lead the hatchlings and maintain group cohesion. In these situations the hatchlings responded similarly by emitting short choruses of grunts. Juvenile distress calls . Juvenile distress calls were initially similar to grunts but were louder and contained more high freguency elements (up to 4.5 KHz in alligators;

PAGE 216

207 Herzog and Burghardt 1977). However, with continued stimulation these distress calls would grade into a longer, two component vocalization containing a high freguency "screech" and a downward freguency "moan" (Herzog and Burghardt 1977) . This pattern was noted both for alligators and spectacled caiman. On Masaguaral, distress calls were occasionally given by hatchlings when captured or when grabbed by predators. Distress calls would occasionally provoke a defensive reaction from nearby adults, especially females, but there was much variability. Adult distress calls . Adult distress calls were similar to the ones produced by small caiman, but were much lower in pitch. Adult distress calls were occasionally given by animals that had been captured, or in free ranging animals in association with fighting between males, or between females contesting the "ownership" of pods of hatchlings (Thorbjarnarson, unpublished observation) . Olfactory signals The examination of 72 sacrificed caiman adult collected at monthly intervals over a 14 month period showed that both males and females had a reservoir of a viscous musk in the cloacal glands throughout the year. Only one female, collected in December, did not have musk. The gular glands were never found to contain large guantities of this viscous fluid, but contained either a waxy or milky secretion. In the dry season a majority of both sexes had the dry, waxy

PAGE 217

208 secretion. In the early wet season animals would have either this waxy secretion, or a less viscous milky secretion. In the later wet season all animals examined had this liquid material in the gular glands (Fig. 8-6a,b) . In November and December a few individuals were found with dry gular glands. Agonistic Behavior and Territoriality Toward the end of the dry season and into the early part of the rainy season (April-June) caiman in the llanos were found in dense concentrations in the few permanent water lagoons or streams. Agonistic behavior (e.g. chasing) was observed throughout the dry season, but intensified as the start of the rainy season approached (Fig. 8-7) . Water level controlled caiman density and played an important role in regulating the severity and timing of agonistic behavior. For instance, in April 1986 the Guacimos Lagoon water level dropped dramatically due to a pump failure, causing the caiman to concentrate in a small part of the lagoon. The number of chases increased dramatically from March to April 1986, and was higher than for 1985 (Fig. 8-7). However, within any one year the general pattern was one of increasing agonistic encounters during the late dry season. Under the high density conditions of most water bodies on Hato Masaguaral during the late dry season, males could not establish well-defined territories. Patrolling males typically swam around a loosely defined area, which varied

PAGE 218

Percent Contents Males JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Month Dry B§3 Liquid [ZD Waxy Figure 8-6. Contents of male and female submandibular glands of caiman in 1986.

PAGE 219

210 Chases per Minute 0.08 i Figure 8-7. Observed frequency of dry season chases at the Guacimos Lagoon.

PAGE 220

211 in size depending on the density of the population and the presence of other adult males. The shape and size of these individual areas were very dynamic, varying from day to day. Caiman would bellow from within these areas, and attempt to exclude other males. Under some circumstances, well-defined territories were observed even under relatively high population density conditions. On 4 June 1985 at the Guacimos Lagoon five adult males were observed dividing 150 m of shoreline, each defending a segment against its neighbors. However, on the following day this had changed and one of the largest males was not seen, but a smaller male was visible in the same area. On a large body of water like the Guacimos Lagoon most behavioral interactions (agonism, courtship) occurred during the early morning hours (05:00-08:00 h) . After 07:00 h an easterly breeze picked up that would cause significant wave action in the lagoon. As the waves increased in intensity caiman would haul out to bask, or sit submerged on the bottom of the shallow lagoon. With the flooding of the savannas, the vast majority of the caiman dispersed from the dry season permanent water habitats (see chapter 4) . Adult males established territories under much lower density conditions in marshes, and it is there that the majority of courtship and mating took place. Associated with this dispersal and the establishment of male territories were peaks in three behaviors: bellowing, agonistic encounters, and courtship.

PAGE 221

212 Male patrolling Patrolling males would approach other caiman, who if smaller would indicate submission by submerging or swimming away. Retreating caiman were often chased by the larger animal. If the two males were of approximately the same size, these encounters would potentially lead to a face-off where neither male immediately withdrew. Face-offs were observed to last up to 35 minutes with both males remaining in an elevated posture or opening their mouths and further elevating their bodies to assume an inflated posture. During face-offs one male would slowly begin to lower its posture and eventually submerge or swim off, or the encounter would escalate and the participants continued to approach one another slowly. Usually when the caiman were within 1 m of one another, one male would suddenly break off the engagement and flee, and was usually chased vigorously by the dominant male. If neither male fled physical combat would ensue. Observed fights were of very short duration with one male guickly departing after exchanging bites. However, injury data (Thorbjarnarson, unpublished observation) suggest that some fights inflict serious, even fatal injuries. When the seasonal rains began the caiman dispersed from the dry season concentrations to seek out suitable habitats in the flooding savannas. Dispersal on Masaguaral generally occurred during May but varied according to prevailing

PAGE 222

213 hydric conditions (chapter 4) . Agonistic encounters in marsh habitats such as the Casa area, where breeding males established territories, peaked in June (Fig. 8-8) and declined throughout the wet season. Another, smaller peak was noted at the beginning of the dry season (December) . Male patrolling behavior also peaked in June (Fig. 8-9) . Males were seen patrolling territories throughout the day, but agonistic encounters (chases) were most commonly observed between 08:00 and 09:00 h (Fig. 8-10). Bellowing behavior The bellow and its associated body postures are a striking visual and auditory signal. Bellows contain vocal, non-vocal acoustic and visual (and possibly tactile) signals. Unlike the American alligator, for which both sexes bellow (Garrick et al. 1978, Vliet 1989) , bellowing in Caiman crocodilus on Hato Masaguaral was only noted among males. The vast majority of observed bellows were performed by adult males (size-class IV, >90 cm SVL) . On several occasions size-class III animals were observed bellowing, but in the few cases where the sex of these animals could be determined (from tail tags), all were males. There was no evidence of bellowing in female caiman. Bellowing was almost exclusively done in shallow water. Caiman were observed bellowing on land on only two occasions, and these bellows did not follow the sequence of stereotyped body postures as described below. Bellows were

PAGE 223

214 Chases 25 i Month Figure 8-8. Seasonal frequency of observed chases in the Casa area over the period 1985-1988.

PAGE 224

215 Patrols Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 8-9. Monthly frequency of male territorial patrols Casa area data 1985-1988. No. Observed Behaviors 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour I Adult Distress Calls EBB Chases EO Fights M Patrols Figure 8-10. Diel frequency of observed territorial or agonistic behaviors; Casa area data 1985-1988.

PAGE 225

216 typically part of a behavioral act sequence (Fig. 8-11) that began with the male in an elevated posture, either HB, HBT, HEBT, OR HEBTE. If moving, the caiman became stationary, and elevated its head and tail, usually to a HETA posture. From the HETA posture the caiman either continued with the bellowing sequence, or simply remained in this signalling posture. Prior to bellowing the caiman lifted its head into a head alert or a head oblique posture. At the same time the caiman would assume a tail oblique position by raising its tail to a 30-45° angle out of the water. From this posture, with the mouth slightly open, the actual bellow began with a series of SAVs that emanate from the torso of the caiman. These powerful infrasonic signals could be felt through the ground by human observers, and would also produce a characteristic visual display of water "dancing" off the back of the bellowing caiman (Fig. 8-lle) . Bellows were comprised of a series of 2-24 distinct infrasonic bursts. The number of these SAVs in a bout was distributed bimodally with many short (3-5 SAVs) or long (13-16 SAVs) bellows (Fig. 8-12), but this pattern varied seasonally (Fig. 8-13a-c) . During the courtship season, most bouts were short with 4-5 SAVs. From September to December more long bellows were noted (13-15 SAVs) . In the dry season both long and short bellows were observed but the highest frequencies were at intermediate numbers of SAVs (89) . At the end of the bellowing bout the caiman typically produced 1-3 less powerful SAVs which were associated with

PAGE 226

Figure 8-11. Body postures comprising the bellow behavioral act sequence. Note the production of infrasonic signals in the fifth photograph as evidenced by the disturbed water over the caiman's back. The last photo demonstrates bubbling.

PAGE 227

218

PAGE 228

219

PAGE 229

220 Bellows p — 1 i \ 01 f iff T — T — T — T — T — T — T — T — T — T — T — T T X ™ 1 1 1 1 1 1 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Subaudible Vibrations Figure 8-12. Frequency distribution of the number of infrasonic bursts (subaudible vibrations) produced by male caiman during a bellowing bout. All data combined.

PAGE 230

20 Bellows 15 10 Dry Season (Jan-Apr) 0 *-! r — i 1 — P-^-^P — i r nil I t 1 — r-n — l — i — i — I — i — r— l — i — I — I 1 1 1 — l r2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 20 15 10 Bellows Early Wet Season (May-Aug) 1 1 • 1 1 • 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 20 15 Bellows Late Wet Season (Sept-Dec) 10 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Subaudible Vibrations Figure 8-13 . Frequency distribution of the number of infrasonic bursts (subaudible vibrations) produced by male caiman during a bellowing bout; by season.

PAGE 231

222 rather weak audible signals. On rare occasions no audible bellow was produced. The production of these audible bellows was normally accompanied by a small upward swinging motion of the snout, and a slight opening of the mouth. The audible bellows were not very powerful, often sounding like a faint "croak", and usually were not heard from more than 50 m away. In a typical long bellow sequence the caiman started producing SAVs at a rate of approximately one per second. During the middle of the bellowing bout the cadence dropped to nearly half that rate, but again increased near the end of the bout until it was difficult to distinguish consecutive SAVs. The last SAV prior to the audible bellows was extended and occasionally lasted over two seconds. Behavior immediately following a bellow was variable. In all cases the caiman lowered its posture in the water. Frequently the caiman would submerge the tip of its snout and bubble vigorously. Bubbling was noted in 21 of 41 bellows (51.2%) at the Guacimos Lagoon. Under high density conditions at the end of the dry season caiman would frequently submerge or remain in a low profile float following a bellow. In 34.3% of the cases (N=35) at the Guacimos Lagoon the male assumed a "head" posture (head only visible) and in 5.7% the male submerged following the bellow. In the remainder of the observations the caiman adopted a semi-elevated (head-back, 25.7%) or elevated posture (34.3%) .

PAGE 232

223 Over a four year period (June 1985-May 1989) the occurrence of 618 bellows was noted in the Casa area (Fig. 8-14) . Bellowing was heard throughout the year, but peaked in the early rainy season (May-July) , with the greatest number in June (134 bellows) , coinciding with the peak in courtship and mating. A secondary peak in bellowing was found in January, associated with the return of caiman to permanent water habitats and the subseguent re-concentration of adults animals. A secondary peak in chases was also noted in the early dry season (Fig. 8-8) . These data indicate that there are two peaks in male agonistic behaviors during the year: at the start of the rainy season with the initiation of courtship and territorial behaviors, and at the beginning of the dry season when the drying savannas force the dispersed caimans to concentrate and reestablish their social hierarchy. Although bellows were heard during all hours of the day, bellowing was principally an early morning activity (Fig. 8-15) . Bellowing began to increase after midnight, and peaked between 06:00 h and 07:00 h. On a seasonal basis, during the late wet season bellows were strongly concentrated in the early morning hours with few bellows heard during the day or at night (Fig. 8-16c) . In the dry season there was an increase in bellows heard at night, especially in the hours before dawn (Fig. 8-16a) . Bellowing during the early wet season, when courtship and mating occur, also peaked during the early morning hours,

PAGE 233

224 Bellows 140 i Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 8-14. Monthly frequency of bellows observed in the Casa area 1985-1988. Bellows 160 1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2C 21 2223 Hour Figure 8-15. Diel frequency of bellows observed in the Casa area 1985-1988. All data combined.

PAGE 234

No. Bellows 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour No. Bellows 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour 60 50 40 30 20 10 Late Wet Season: Sep-Dec J , wMm 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour Figure 8-16. Diel frequency of bellows observed in the Casa area 1985-1988; by season.

PAGE 235

226 but bellows could be heard more frequently during the day and night (Fig. 8-16b) . Many of the bellows heard during the day in the early wet season occurred during brief to moderate rainshowers and were accompanied by an increase in male patrolling behavior. Responses to bellowing One caiman's bellowing elicited bellowing by other individuals. Bellowing choruses were frequently heard at the Guacimos Lagoon during the early morning hours. In the Casa area bellowing choruses (defined as more than one bellow within a three minute period) accounted for 226 of the 618 bellows (36.6%) noted over a four-year period. Bellowing choruses comprised the largest percentage of bellows during the early dry season (60.3% of observed January bellows) . The largest bellowing chorus noted in the Casa area was 8 bellows, with at least one individual bellowing twice. Larger bellowing choruses were observed at the Guacimos Lagoon with its larger caiman population. On 24 April 1986, 13 males were observed bellowing during a two minute interval. The total caiman population in the observation area at the time was approximately 400. Similarly, on 5 June 1988, 12 caiman bellows were heard within five minutes at the Guacimos Lagoon (est. population 90 caiman) . Caiman bellowing was also elicited by the sounds of automobile engines, thunder, loud explosions, and the bellows of Orinoco crocodiles.

PAGE 236

227 Following bellowing choruses the number of adult male caiman in high profile positions in the water increased. For instance, prior to the 24 April 1986 bellowing chorus (06:00 h) , only five males were seen in high profile postures. The chorus lasted from 06:04-06:06 h, and at 06:07 h 16 males were seen in elevated postures with their tails arched. By 06:15 h this number had decreased to 9, only three of which had the tail arched. On 16 May 1986 a bellowing chorus of 9 bellows was observed from 06:44-06:52 h at the Guacimos Lagoon. Prior to the chorus (06:30 h) only four males were visible in elevated postures. Following the chorus (06:52 h) , 31 individuals were elevated. Many, but not all, of the males in the elevated postures were individuals that had bellowed. In certain cases the bellowing elicited an elevated posture but not a bellow. Bellowing rarely elicited other obvious behavioral responses. Courtship and mating Most courtship and mating activity among caiman on Hato Masaguaral took place at the beginning of the rainy season in flooded marshes or other habitats following the dispersal from the dry season habitats. However, when late rains delayed the onset of the wet season caiman began courtship and mating while still in the dry season lagoons. This was the case in the Guacimos Lagoon in 1988 when late rains delayed caiman dispersal until mid to late June. The

PAGE 237

majority of the following observations on caiman courtship and mating were made in 1988. Most observed courtship attempts were elicited by the male's approaching the female. On only three occasions (5.1% of observed encounters in the Guacimos lagoon) was the female observed to approach the male and initiate courtship. The initial approach was not unlike that for aggressive encounters between males. The male approached the female by swimming slowly toward her in an elevated posture, but then would gradually assume a HATA posture. During 5 of 58 (8.6%) observed approaches males produced 2-3 SAVs. In the majority of observed encounters the female was unreceptive and submerged and swam away. In only 18 of 58 (31.0%) encounters was a mounting attempted. Frequently, the male slowly followed an unreceptive female, approaching her in a HATA or a HHBTE posture when she returned to the surface. Receptive females did not move away from the approaching male, but remained in a low profile position with only the head visible. As the male drew alongside, a receptive female assumed a head-inclined posture and rubbed the tip of her snout along the lateral surfaces of the male's mandibles, or on the male's gular region. On one occasion the female was observed in a head inclined posture, oriented perpendicular to the male and rubbing his left nuchal region with her snout. Following this tactile encounter the female would then swim slowly posteriorly

PAGE 238

229 alongside the male's body, allowing the male to turn and mount her. Upon mounting, males would almost immediately begin tail searching to oppose cloaca for intromission. When the male succeeded in arching his tail under the female's, he would lift her tail slightly, and slide off to the side. Males that were presumed to be copulating were seen in this position with the tail arched under the female's. The tip of the male's tail would frequently emerge from the water on the opposite side of the female. At the Guacimos Lagoon the mounted pair would frequently split up, apparently before intromission could occur, either because the female moved off underwater, or because the water was too shallow to permit copulation. In only 8 of 18 (44.4%) observed mounts did the pair remain together long enough to allow tail searching and the apparent opposition of cloacae. Dominant males interrupted the courtship of other caiman in their surrounding area. On six occasions large males were observed to approach courting or mounted pairs of caimans. In all cases the male broke up the pair and chased them off. In one case the male was subsequently observed courting the female from the interrupted courtship. Discussion As with all crocodilians studied to date, caiman appear to have a polygynous mating system. Males establish territories during the early wet season, patrol these

PAGE 239

230 territories to exclude other males, and attract females with bellowing advertisement displays. Communication is achieved visually using a series of graded body postures, with vocal and non-vocal acoustic signals and, presumably, by olfactory signals produced by the two sets of paired musk glands. Many of the visual and acoustic signals used by the spectacled caiman were very similar to ones reported for American alligators, (Garrick et al. 1978), and Crocodvlus acutus and C. niloticus (Garrick and Lang 1977) , but there were some striking differences. Headslapping, a commonly observed assertion display among crocodilians (Lang 1987) , appears to have a minor role, if any, in the caiman's behavioral repertoire. Headslapping is a freguently observed non-vocal acoustic signal in both Alligator mississippiensis and Crocodvlus acutus and C. niloticus (Garrick and Lang 1977) . The headslap produces an airborne auditory signal as well as an underwater percussive one. Waves produced by the headslap and the accompanying tail wag may also be part of the signal (Vliet 1989) . Among alligators headslaps are given by both sexes, but it is predominantly a male activity (Vliet 1989). Although headslaps were never observed during this study, sounds similar to headslaps were heard on occasion. Ayarzagiiena (1983) also reported hearing headslaps, but never seeing them. It seems probable that caiman do headslap, but that it is an infreguently used signal.

PAGE 240

231 Bellowing behavior has been well described for American alligators (Garrick and Lang 1977, Garrick et al. 1978, Vliet 1989) , and similar behaviors have been reported for other crocodilians including the Nile crocodile (Cott 1961, Modha 1967) , estuarine crocodile (Deraniyagala 1939) , Chinese alligator (Garrick 1975) , American crocodile (Thorbjarnarson, unpublished data) , and Orinoco crocodile (Thorbjarnarson, unpublished data) . However, in all of these species bellows (termed "roars" for some species) contain a major audible sound component. Conversely, bellowing in caiman consisted almost entirely of subaudible frequencies. Furthermore, bellowing in caiman appeared to be exclusively performed by males, and while this also appears to be true for some crocodiles ( Crocodylus niloticus . Garrick and Lang 1977; C. acutus . C. intermedius . Thorbjarnarson, unpublished data) , bellows are performed by both males and females in the alligators (A. mississippiensis and A. sinensis ; Garrick et al. 1978, Garrick 1975) . The SAV, described in American alligators by Garrick and Lang (1977) and Garrick et al. (1978), was shown by Vliet (1989) to be an important male-specific component of the bellowing display. Vliet (1989) also reported the use of SAVs in six other species and postulated that this signal may be universal among crocodilians. Preliminary work by Vliet (1987) suggests that alligator SAVs are in the 10 Hz range. The actual sensory modality used in perceiving these

PAGE 241

232 signals is uncertain. Crocodilians are acknowledged to have an acute sense of hearing (Bellairs 1969) and may be sensitive to very low frequencies. However, the SAV can also be felt by human observers and tactile reception through the water may be important for crocodilians as well. The heads of all the Alligatoridae, and the entire bodies of the Crocodylidae and Gavialidae, contain integumentary sense organs (ISOs) (Brazaitis and Garrick, unpublished data) which are similar in structure to rapidly-adapting mechanoreceptors (von During 1973a, b) . These ISOs may serve a role in tactile behaviors in courtship, or in the sensing of waterborne low-frequency vibrations or other non-vocal acoustic signals involving splashing (e.g. tail splashing, headslapping) . Two principal functions have been ascribed to crocodilian bellowing: an agonistic signal for the demarcation and defense of territories (Kellog 1929, Vliet 1989) , and an advertisement display for the attraction of mates (Joanen and McNease 1971, 1975, Garrick and Lang 1977, Garrick et al. 1978, Vliet 1989). A bellow conveys information regarding the sex, location, size and possibly individual identity of the animal (Garrick et al. 1978). Garrick and Lang (1977) and Garrick et al. (1978) suggest that a vocal signal is better than visual displays for the attraction of mates over long distances with reduced visibility. They go on to suggest that visual signals may be more important for short to moderate distance

PAGE 242

233 communication in species that live in more open habitats (e.g. Crocodylus niloticus ) . The authors suggest the prevalence of vocalizations in American alligators is an adaptation for living in marsh habitats. Caiman live in a wide variety of habitats, so few generalizations can be made in terms of specific adaptations to particular habitat types. However, caiman bellowing is significant in that it is comprised almost entirely of a low frequency, subaudible signal of non-vocal origin. SAVs are incorporated into the bellows of American alligators (Vliet 1989) , and are part of the courtship display of all crocodilians that have been examined to date (A. mississippiensis . C. acutus, C. niloticus (Garrick and Lang 1977) , C. intermedius (Thorbjarnarson, unpublished observation)). However, the almost exclusive use of an infrasonic signal during bellowing sets Caiman crocodilus apart from any other crocodilian that has been studied to date. The low frequency SAV would be propagated efficiently over long distances in water (Hawkins and Myrberg 1983), and appear to be well suited for communication in densely vegetated aquatic environments. Apart from the auditory signal, bellowing in caiman (and other crocodilians) involves a stereotyped series of body postures that typically involves an elevation of both the head and the tail. Similar postures are also associated with other social displays (e.g., headslapping) and may facilitate the location of the signalling animal by its

PAGE 243

234 neighbors, or even permit its size and identity to be determined (Vliet 1989) . However, the postural display of bellowing caiman is unusual in the degree of elevation of the tail. This exaggerated tail posture may increase the effectiveness of the visual signal by permitting it to be observed from further away. Staton and Dixon (1977) also reported this posture as a "tail vertical" display, noting that it was apparently only seen in males. A similar posture was described among caiman at Hato El Frio in Apure state (Ayarzaguena 1983), but was described as being rarely seen among adults (although common among hatchlings) , and was suggested to be a signal indicating danger. However, in no previous study has the relationship between the HOTO posture and bellowing been noted. Courtship in caiman involved much tactile stimuli, and in this respect was not very different from the reported courtship behavior in American alligators (Garrick and Lang 1977, Vliet 1987). However, the lengthy courtship bouts seen in captive crocodilians were never observed among wild caiman. Caiman courtship typically consisted of a male approach, and guick mounting after a few preliminary tactile behaviors. Lengthy courtship among captive animals may result from individuals being familiar with one another. Among captive Orinoco crocodiles on Hato Masaguaral a pair that had been housed together for 20+ years would engage in lengthy, leisurely courtship bouts. Conversely, other individuals of the same species that had been together less

PAGE 244

235 than one year were never observed in drawn-out courtship activities. Alternatively, the abbreviated courtship among caiman in this study may simply be a result of most observations being made under high-density dry season conditions when many courtship attempts were interrupted by neighboring males. The observations of caiman courtship and mating noted by Staton and Dixon (1977) and Alvarez del Toro (1974) differed in some respects from those made in this study. Alvarez del Toro (1974) noted it was the male that moved around the female during courtship activities, whereas my observations and those of Staton and Dixon (1974) suggest the opposite. The latter authors, however, suggested that jumping behavior was a courtship display, whereas my interpretation is that jumping is solely a feeding activity (chapter 7). Staton and Dixon (1977) also reported an instance where a female climbed onto the back of a male during courtship, and suggested that this was a copulatory posture utilized under low water conditions. On 13 June 1988 I observed a similar female riding behavior in a pair of courting caiman. The female was situated partially over the anterior one-third of the male's dorsum, with her head directly over the male's occipital region. The male moved vigorously and dislodged the female, who then submerged. However, I believe this behavior to be a courtship act where the female attempts to rub her snout over the male's head or

PAGE 245

236 nuchal region, and may be analogous to the pressing behavior described for alligators (Vliet 1986) . The role of olfaction in social behavior is poorly understood. Caiman, like all crocodilians, possess paired cloacal and gular glands (Reese 1921) . The function of these glands has not been clearly demonstrated but they presumably play a role in chemical communication. The "musk" from these glands may serve an intraspecif ic role as a pheromone, or to repel potential predators (Wright and Moffat 1985) . Male and female caiman had musk present in the cloacal glands throughout the year. However, the presence of, and the composition of, secretory material in the submandibular glands varied on a seasonal basis. The increase in the amount of liguid material in the submandibular glands during the early wet season suggests that it may play a role in reproduction. The lack of seasonal variation in cloacal gland contents suggests that these glands may have a different function. Alligators have been observed to evert their mandibular glands during bellowing (Schmidt 1922, Mcllhenny 1934), although this is said to be uncommon (Vliet 1989) . A female Orinoco crocodile was observed everting the gular glands while opening her nest (Thorbjarnarson, unpublished observation). Juvenile alligators have been shown to react to the airborne scent of the musk (Johnsen and Wellington 1982), but the principal mode of transmission may be along the water's surface (Vliet 1989) .

PAGE 246

Intraspecif ic aggression among caiman appears to play an important role regulating many aspects of their behavior and ecology. The outcome of agonistic encounters between males appears to be settled based on size, with larger animals being dominant. Following a face-off involving body posturing, encounters where one animal was clearly larger than the other were characteristically ended with the fleeing of the subordinate animal. When the two caiman were more evenly matched in size, these encounters could escalate into actual combat, and these battle could result in severe injury and even death. The results of the radio-telemetry study indicated that after dispersal, large males preferentially occupied the areas near the dry season lagoons, where the density of females was the greatest. Smaller males dispersed longer distances and occupied more marginal areas, presumably due to agonistic encounters with dominant males. Males that were recently physiologically mature remained near the dry season lagoons, but tended to wander more than larger males, indicating that they were not establishing breeding territories (chapter 4). These results suggest that the reproductive success of males is strongly linked to body size, and that ontogenetic changes in movement patterns and habitat utilization must be interpreted in light of reproductive considerations. Small physiologically mature males may either not attempt to establish reproductive territories, or do so and be excluded

PAGE 247

238 from suitable areas by larger males. As they grow larger and move up the social hierarchy, and at the same time become more familiar with the nearby savanna habitats, males may find themselves capable of establishing territories in marginal areas. In the case of Hato Masaguaral, this appears to be principally marsh habitats located more than 3-4 km from the dry-season lagoons. With continued growth, males may be able to establish territories in some of the preferred habitats closer to the lagoons. Hence, as the caiman grows, his ability to compete against other males increases and his social rank increases, permitting him to establish territories in preferred areas and presumably increase his reproductive success. A somewhat similar scenario has been put forth for two species of crocodiles (C. porosus . Messel et al. 1981, Messel and Vorlicek 1987; C. niloticus . Hutton 1989). These studies have emphasized the importance of intraspecif ic aggression, particularly agonistic encounters by larger crocodiles against ones just entering sexually maturity (Messel et al. 1981). However, these investigations have emphasized more the role of intraspecif ic aggression on species size-class segregation and mortality schedules. While this undoubtedly occurs in caiman as well, the situation in the llanos has an added dimension, that of time. Intraspecif ic aggression and cannibalism undoubtedly play an important role in dry season caiman mortality. These effects are to a large extent density-dependent, with

PAGE 248

239 hatchling and juvenile survivorship being much higher in low-density lagoons (chapter 3) . Under these high-density dry season conditions, aggression of adults towards subadults (Hunt and Watanabe 1982) may also be exacerbated. These kinds of interactions among size-classes appear to be a major structuring component for the population dynamics of all crocodilians (Hutton 1989). In the llanos, however, following dispersal into the flooded savannas caiman have a enormous amount of potential habitat available, and sizeclass segregation can probably take place without much mortality. In this case the intraspecif ic aggression during this time of year appears to be more important in regulating reproductive success, especially among males. It is clear that spectacled caiman in the Venezuelan llanos exhibit a highly developed signalling system involving all the sensory modalities. These signals were utilized for the establishment of dominance hierarchies among males during the latter dry season, and the creation of breeding territories and courtship and mating during the early wet season. Many of the observed behaviors were very similar to ones reported for American alligators and other species of crocodilians, especially those involving body postures. However, of the two principal crocodilian advertisement displays, headslapping was never observed, and caiman bellowing is unigue in its virtual exclusive use of subaudible vibrations. Caiman courtship activities also appeared to be abbreviated and simple relative to other

PAGE 249

240 crocodilians that have been studied in captivity. However, much remains unknown regarding the social behavior of spectacled caiman. More extensive work is needed, both with the wild and captive populations before we can claim to have a good understanding of this species' reproductive behavior. Of particular importance, both for the understanding of the caiman's mating system as well as for management purposes, will be to investigate the dynamics of caiman breeding groups and to determine the degree of polygyny in this species.

PAGE 250

CHAPTER 9 SEXUAL MATURITY, REPRODUCTIVE CYCLE, AND EGG AND CLUTCH CHARACTERISTICS Introduction Previous studies on the reproduction of Caiman crocodilus have dealt principally with nesting ecology (Alvarez del Toro 1974, Rivero-Blanco 1974, Crawshaw and Schaller 1980, Medem 1981, 1983, Crawshaw 1987, Ouboter and Nanhoe 1987, Cintra 1989), although some information has been published concerning other aspects of the species' reproductive ecology (Staton and Dixon 1977, Ayarzagiiena 1983). However, most investigations of caiman reproductive cycles have been descriptive in nature, with few quantitative data. Given the fact that caiman in the llanos occupy one of the most seasonal habitats of any crocodilian, it is of interest to examine how the timing of the reproductive cycle coincides with seasonal changes in the environment. Another important aspect of reproduction is the relationship between size and fecundity. Among species, clutch size has been shown to increase with female size in crocodilians (Greer 1975) , but work on intraspecif ic trends 241

PAGE 251

242 has lagged considerably behind analyses of other groups such as turtles (Wilbur and Morin 1988) . Size-specific fecundity is an important life-history parameter of all animals, but is particularly so for long-lived, iteroparous species with indeterminate growth (Congdon and Gibbons 1990) . Fecundity is a complex variable determined by the interaction of a number of life-history parameters such as size at maturity, the relationship between clutch size and female size, tradeoffs between clutch size and egg size, and the frequency with which females breed. This study represents the first time that all these parameters are evaluated for a population of crocodilians. The goals of this study of caiman reproductive ecology were: 1) to quantify the minimum reproductive size for both sexes, 2) to explain the seasonal reproductive cycle in relationship to environmental variability, 3) to examine seasonal variation in stored lipids and how this is related to reproduction, 4) to determine egg and clutch characteristics and how these are related to female size, 5) to estimate the size-class distribution of the breeding female population, 6) to evaluate egg viability, and 7) to determine breeding effort and its variation among years and female size-classes. Methods Reproductive tracts were obtained from dead caiman collected in one of three ways. Fresh road-killed specimens

PAGE 252

243 (N=233) were picked up from a highway located adjacent to Hato Masaguaral, between September 1984 and May 1989. A sample of 72 adult caiman from Hato Masaguaral and adjacent Hato Flores Moradas was captured and sacrificed under government permit between June 1986 and June 1987 (Licencia de Caza con Fines Cientificas No. 001744, 9 June 1986, Ministerio de Ambiente y de Recursos Naturales Renovables) . Thirdly, 14 caiman found dead from natural causes were necropsied. The sex, snout-vent length, total length and weight of specimens were noted. In some cases weight and total length of road-killed individuals could not be accurately measured. The mass of the lateral fat body, testes, and/or ovaries was weighed on an O-Haus Triple Beam balance to the nearest 0.1 g, although gonad or fat body mass from road-killed specimens was not always measurable. The diameters of the six largest ovarian follicles in each ovary were measured with calipers to the nearest 0.1 mm. Vitellogenesis was determined by macroscopic examination of the ovarian follicles. Ovaries were examined for the presence of different sized follicles or copora lutea. Oviducal eggs were weighed (0.1 g) and measured (length and width, 0.1 cm) . Egg and clutch data were collected during the examination of nests (chapter 10) . Egg dimensions were measured with calipers to the nearest 0.01 cm and egg weight was weighed with a 100 g Horns spring scale to the nearest 1 g-

PAGE 253

244 To examine the changes in reproductive parameters with increasing clutch size I followed the statistical analysis conducted by Ford and Seigel (1989) , performing a partial correlation analysis of the reproductive parameters (egg width, egg length, egg mass, and clutch mass) while controlling for female size (SVL) . Histological analyses of testes were performed for a sample of adult road-killed and sacrificed individuals. Testes were collected and preserved in 10% formalin, then prepared using standard histological techniques. Tissues were hematoxalin-eosin stained, mounted on glass slides and observed at lOOx magif ication. Results Sexual Maturity and Adult Size Females The smallest vitellogenic caiman examined was a 60.0 cm SVL road-killed individual found on 13 May 1986 (mean follicle diameter=2 . 15 cm). The examination of the physical state of oviducts also suggested that sexual maturity is reached at a minimum size of approximately 60 cm SVL. The oviducts of immature female caiman are narrow, translucent structures less than 0.5 cm wide posteriorly, and 0.3 cm wide anteriorly. After the hypertrophy of the oviducts associated with ovulation, the oviducts remain distended, are more heavily muscled and present noticeable longitudinal convolutions in the posterior segment. The examination of

PAGE 254

the oviducts (classified as distended, partially distended or not distended) of 104 females over 50 cm SVL indicated that very few individuals under 60 cm SVL had previously nested (1 of 17, 5.9%: (Fig. 9-1). The majority of females reached reproductive maturity between 60 and 70 cm SVL, corresponding to an age of 6-10 years based on growth rates calculated in chapter 3. Average female size at first reproduction (50% of females with distended oviducts) was calculated to be 64.0 cm SVL, which corresponds to an average age of 7 years. Only one female greater than 70 cm SVL had not yet reproduced. No females under 50 cm SVL showed any signs of oviducal hypertrophy. Three of the four females under 65 cm SVL with partially distended oviducts were examined during the breeding season (July and September) . The mean snout-vent length of live-captured females over 60 cm SVL was 72.6 cm (SD=6.6 cm, N=134) (Fig. 9-2). Among road-killed caiman, mean SVL of females over 60 cm SVL was 70.7 cm (SD=6.8 cm). The largest female examined on Masaguaral measured 86.1 cm SVL. However, in other parts of the Venezuelan llanos females attain larger sizes. Two females captured at nests on Hato La Guanota (Apure State) in 1987 measured 88 cm and 92 cm SVL. Two females in a sample of 34 commercially harvested caiman from Hato Merecure (Apure State) in April 1989 measured 90 cm and 99 cm SVL. A larger sample of 144 harvested caiman from several ranches in Apure State contained two females

PAGE 255

% Distended Oviducts ——Distended -4" Partially Distended Undeveloped 50-54,9 55-59.9 60-64.9 65-69.9 70-74.9 75-79.9 80-84.9 85-89.9 Size-Class (cm SVL) Figure 9-1. Percent of females with distended, partially distended, and undeveloped oviducts by sizeclass. Number 60 65 70 75 80 85 90 95 100 105 110 115 120 125 Snout-Vent Length Interval (cm) Figure 9-2. Frequency distribution of captured male and female caiman over 60 cm SVL.

PAGE 256

247 measuring 90.5 cm and 104 cm SVL (C. Molina, pers comm.). The latter figure represents the largest reported female Caiman crocodilus and may be approaching the maximum size for females of this species. Males Based on the presence of spermatocytes in the testes, the smallest physiologically mature male measured 90.0 cm SVL (N=16) . However, mature spermatocytes were noted in testes as small as 13.3 g during May, near the peak of the annual testes mass cycle. Using 15 g as a conservative estimate of the minimum mature testes mass, the smallest male that was assumed to be exhibiting spermatogenesis measured 77.4 cm SVL (Table 9-1). This individual had testes weighing 29.0 g, with enlarged fluid-filled epididymides. Another individual (DOR 39), measuring 69.0 cm SVL, had enlarged testes weighing 14.6 g and may have been reproductively mature. However, two other individuals between 70 and 75 cm SVL had small, undeveloped testes, indicative of immaturity. Assuming a minimum reproductive size of 75 cm SVL, males become sexually mature at an average age of seven years (see growth curves, chapter 3). The mean size of live-captured males greater than 75 cm SVL was 96.4 cm SVL (SD=13.6 cm, N=113) (Fig 9-2) and for road-killed males was 94.9 cm SVL (SD=13.6, N=47) . The largest male captured on Hato Masaguaral measured 127.0 cm SVL.

PAGE 257

248 Table 9-1. Paired testes mass of caiman over 60 cm SVL collected during May-July. Reproductive status based on a Caiman SVL Testes Mass Month Reproductive (cm) (g) State DOR 45 60.5 1.7 June Immature ND 6/5/85 64.5 2.7 June Immature DOR 179 65.5 1.9 July Immature DOR 39 69.0 14.8 May Immature? DOR 93 71.5 2 . 3 June Immature DOR 171 72.0 9.3 June Immature DOR 29 77.4 29.0 May Mature DOR 92 82.7 22.4 June Mature DOR 98 83.0 30.6 June Mature S 10 84.5 17.7 14.6 June Mature DOR 88 88.0 June Immature? S-70 90.0 18.5 June Mature Female Gonad Cycle No evidence of significant year-to-year differences in the timing of reproductive events was noted, so all the data collected over the study period (October 1984-May 1988) were analysed as one data set. Vitellogenesis began at the end of the dry season, with the mean ovarian follicle diameter of sacrificed and road-killed females (Table 9-2) showing an initial increase in May, although in some individuals vitellogenesis was apparent in late April. The mean follicle size of vitellogenic females increased throughout the early wet season (May-July) and reached a peak in July and August. Paired ovary mass and relative ovary/ova mass (ovary + ova (or oviducal eggs) mass/female mass) showed similar seasonal patterns (Fig. 9-3 and 9-4) .

PAGE 258

249 CP tn V) o o > O "O CD i_ '5 Q_ 1800^ 16001400-1200-1000-800-600-400-200 -0-O — O Non-vitellogenic • — • Vitellogenic 0^0-^0 — #—0—0 — ® — 0 — ® — 0 — Q— 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 9-3. Mean paired ovary mass by month for breeding and non-breeding females. Data collected 19851988. C/) U) O CD 20 15->^ 10O > o 50 O — O Non-vitellogenic • • Vitellogenic I '1 /I J* Q — © — 0— % — 0 — 0 0 © © 0 — O — 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 9-4. Month Mean paired relative ovary mass (ovary mass/body mass x 100) by month for breeding and non-breeding females. Data collected 19851988.

PAGE 259

250 Table 9-2. Mean monthly ovarian follicle diameter (cm) and standard deviation for vitellogenic and non-vitellogenic females. Month Vitellocrenic Non-Vitelloaenic Total Mean SD N Mean SD N Mean SD N Jan 0.29 0.06 6 0.29 0.06 6 Feb 0.33 0.08 5 0.33 0.06 5 Mar 0. 33 0. 07 5 0.33 0.07 5 Apr 0.44 0.01 2 0.29 0.04 3 0.35 0.08 5 1.73 0.58 10 0.30 0.11 6 1 1 Q i . iy U.BJ 16 Jun 3.05 0.51 7 0.25 0.04 7 1.65 1.44 14 Jul 3.60 0.74 8 0.23 0.02 3 2.68 1.63 11 Aug 3 .86 0.16 2 0.23 0.07 6 1.14 1.57 8 Sep 0.28 0.10 5 0.28 0.10 5 Oct 0.28 0.08 2 0.28 0.08 2 Nov 0.34 0.07 7 0.34 0.07 7 Dec 0.22 0.03 7 0.22 0.03 7 Grossly, the oviducts of reproductive females are divided into two distinct segments: a translucent, heavily vascularized upper (anterior) section, and a white, heavily convoluted lower region proximal to the cloaca (uterus) . Seasonal hypertrophy of the oviducts is noticable during the late dry and early wet season. Non-breeding adult females typically had two sizes of follicles in their ovaries, minute granular follicles less than 1 mm in diameter, and 2-3 mm diameter follicles. Breeding females had a third set of vitellogenic follicles. The number of vitellogenic follicles declined as the average size of the follicles increased. The overall mean number of ova in breeding females decreased from 35.17 in May, to 24.20 in July and 25.67 in August (Table 9-3). However, this decreasing trend was only noted among large females (>75 cm SVL) . Large females (>75 cm SVL) produced

PAGE 260

251 more follicles than small females (F lf2 l= 9 ' 15 / p<0.01). The reduction in the number of vitellogenic follicles was presumably caused by atresia. An interesting result is that a loss of follicles could only be shown for the right ovary (Fig. 9-5) . No significant difference in the production of ova was noted between the right and left ovaries (Fig. 9-5) . Table 9-3. Mean number of enlarged ovarian follicles by month for large '>75 cm snout-vent length) and small (<75 cm SVL) caiman. Month Small Females Larae Females All Females Mean N Mean N Mean N May 25.00 2 40.25 4 35.17 6 June 25.75 4 32.00 2 27.83 6 July 18.67 3 32.50 2 24.20 5 August 25.50 2 26.00 1 25.67 3 An examination of the mean monthly percent of the adult female population that was vitellogenic (Table 9-4) indicated there was no significant drop in the proportion of vitellogenic females during the reproductive season. A declining trend would indicate that some initially vitellogenic females do not reproduce. These data, however, suggest that once a female became vitellogenic, she ovulated and nested, thereby permitting the inclusion of females examined early in the vitellogenic cycle in the estimation of breeding effort (see below) .

PAGE 261

252 30 25-O Left Ovary • Right Ovary 20d) 15 _Q 10O i o o May June July Month August Figure 9-5. Mean number of vitellogenin follicles in the right and left ovaries by month. Data collected 1985-8.

PAGE 262

253 Table 9-4. The number and monthly percentage of females in breeding condition over the period 1985-1988. Month Non-breedina Breeding April 3 60.0% 2 40.0% May 7 41.2% 10 58.8% June 8 50.0% 8 50.0% July 3 25.0% 9 75.0% August 6 33.3% 12 66.7% September 13 72.2% 5 27.8% Total 40 46.5% 46 54.5% Ovarian follicles were shed into the oviducts when they reached a diameter of approximately 3.7-3.9 cm. Examination of females with oviducal eggs suggested that transovarian migration (ova entering the opposite oviduct) is not uncommon (Table 9-5) . In two of the five females with oviducal eggs, three ova had crossed from the right ovary to the left oviduct. In both cases where displacement had occurred, one ovary had produced a larger number of ova than the other. Transovarian migration is not uncommon among reptiles and has been noted in turtles (Legler 1958) , lizards (Mayhew 1966a, 1966b, Cuellar 1970), and snakes (Tinkle 1957) . A comparison of the number of corpora lutea with oviducal eggs also suggested that some ova apparently did not enter the ostium of the oviduct (Table 9-5) . The ectopic loss of ova has been previously reported for American alligators (Lance 1989), with the ova enterring the body cavity where they may be found encrusted with calciumlike deposits (Lance, pers. comm.).

PAGE 263

254 It was not unusual to find a variable number of intermediate sized follicles remaining in the ovary following ovulation. The largest unshed follicle encountered in a female with oviducal eggs was 2.2 cm diameter. Some evidence indicated that these intermediate sized follicles did not become atretic but entered an arrested growth phase, or atrophied very slowly. Three adult females were found with intermediate sized follicles in the late wet season (two females each with two intermediate sized follicles 0.9-2.8 cm diameter in November) or early dry season (one female with two follicles 2.25 cm and 2.27 cm diameter in January). One of the females from November had regressed corpora lutea visible in the ovary. The presence of these intermediate sized follicles may be indicative of females that ovulated the previous reproductive season. By late July and August, most reproductive females had shelled eggs present in their oviducts (three of four reproductive caiman examined) . Oviducal eggs were found from mid-July until early September (Table 9-6) . The mean date of encountering females (N=7) with oviducal eggs over the 1986-1988 period was 5 August. Over the same period of time the mean date of oviposition varied from 12-17 August, suggesting that females retained eggs in their oviducts for at least 7-12 days. Oviducal eggs were never found in the anterior section of the oviducts, indicating that they pass to the posterior oviduct (uterus) fairly rapidly. Eggs

PAGE 264

255 found in the uterus were always either completely, or nearly completely, shelled. Recent work on gopher tortoises suggests that albumin is secreted from endometrial glands in the tube portion of the anterior oviduct, and the eggshell fibers originate in the endometrial glands of the posterior or uterine section of the oviducts (Palmer and Guillette 1988) . Table 9-5. Comparison of the number of corpora lutea (CL) and oviducal eggs (OE) in five adult female caiman from Hato Masaguaral. Female snout-vent lengths are given in parentheses . Caiman Date Left Side Riaht Side Notes (SVL) CL OE CL OE DOR 103 8/3/86 11 10 15 15 One ovum (80. 1cm) missing S14 7/16/86 7 10 11 8 Three ova (72.9cm) 15 crossovers S15 7/16/86 16 15 14 Two ova (82.8cm) missing S16 7/25/86 9 9 8 8 (70.2cm) 17 S21 8/22/86 11 14 14 Three ova (70.2cm) crossovers Table 9-6. Dates when female caiman had oviducal eggs. Over the same period of time mean dates of oviposition ranged from 12 August to 17 August. 1986 1987 1988 16 July 16 July 25 July 3 August 22 August 13 August 1 September

PAGE 265

256 Male Gonad Cycle The size of the testes exhibited seasonal variation associated with the annual breeding cycle. Testicular recrudesence began in April, at the end of the dry season (Fig. 9-6) , coinciding with the initiation of vitellogenesis in females. Testes mass peaked during the early rainy season (MayJune) , concurrent with the annual peak in bellowing and male agonistic enounters (see chapter 8) . In July-August, testes mass dropped sharply, and testes remained in a regressed state throughout the remainder of the wet season and the early dry season. Testicular mass was correlated with body size, with large (120 cm SVL) males having maximum testes weights up to 60 g. Larger males (to 145 cm SVL) examined on Hato Merecure in Apure state had testicular masses up to 80 g in late April, indicating that maximum testes mass in very large males may be well over 100 g. Testicular mass expressed as a percentage of total body mass (relative testes mass) exhibited a seasonal pattern identical to that of uncorrected testicular mass (Fig. 9-7) . These data also suggest that testicular recrudesence in large males began earlier in the season (April) than for small males (Fig. 9-6 and 9-7; >100 cm size-class). Large males exhibit a greater annual change in testicular mass (398% increase) than do males under 100 cm SVL (280% increase; Table 9-7, Fig. 9-6). During the reproductive season (May-August) , there was no significant difference between the mass of the right (mean

PAGE 266

co a> +-> co 60 50 CO 40 CO 30 20 10 0 O 60-80 cm SVL • 80-100 cm SVL •A > 100 cm SVL Figure Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month 9-6. Mean paired testicular mass by month for three size-classes of caiman. Data collected 1985-8 0.300 CO CO o CO CD CO CD (Y. 0.200-0.100-0.000 O 60-80 cm SVL • 80-100 cm SVL A > 100 cm SVL Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 9-7 Month Mean paired relative testicular mass (testes mass/body mass x 100) by month for three sizeclasses of caiman. Data collected 1985-8.

PAGE 267

258 15.3 g, SD 10.1, N=27) and the left testes (mean 13.1 g, SD 9.3, N=27) . Table 9-7. Mean paired testes mass (g) for small and large male Caiman during the early dry season (January-March) , late dry-early wet season (April-August) , and late wet SVL Class Jan-Mar Apr-Aua Sep•Dec % Chancre (cm) Mass N Mass N Mass N 80-100 7.3 4 15.4 13 5.5 8 280% 100-120 11.4 5 41.0 14 10.3 5 398% Histological examination of testes from sacrificed adult males revealed a similar annual pattern in the abundance of mature spermatozoa (Fig. 9-8) , but spermatozoa remained abundant into August. Moderate numbers of spermatozoa were found in the seminiferous tubules of large males from mid March, and there is some indication that larger males began spermatogenesis earlier than smaller individuals. After August, there was a reduction in size of the seminiferous tubules, and a proliferation of interstitial tissue. During this time few or no spermatozoa were seen in the seminiferous tubules, although some were still found in the efferent ductules of the testes. No spermatozoa were visible from November to February (N=4) . Seasonal Fat Cycle The wet mass of the lateral abdominal fat body was used as an index of stored lipids. To correct for differences in

PAGE 268

259 Abundant 90.0 101.5 97.0 Moderate 102.8 11 2.0 109.0 1 10.0 Few 96.8 93.0 None 98.4 90.7 i 79.5 i l l l 106.0 94.2 i i 91.5 89.1 I Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Figure 9-8. Abundance of mature spermatozoa in the seminiferous tubules by month. Numbers refer to the snout -vent length of the caiman.

PAGE 269

260 body size, a relative fat body (RFB) index was calculated ((wet fat body mass/wet body mass) x 100). Both male and female caiman demonstrated seasonal cycles in fat body mass, but the pattern differed between the sexes. The mean RFB value for females was 0.48% of total body mass; that of males was 0.14%. However, an analysis of covariance, using snout -vent length as the covariate, found no significant differences in fat body mass between males and females. Relative fat body mass in females (>50 cm SVL) peaked in the late wet and early dry season (Fig. 9-9) . Peak RFB values were noted in January (mean 1.51%, N=4) . RFB values declined during the dry season and on into the early wet season (February-August) . Although the mean values of RFB during this period were consistently higher among breeding females than in non-breeders (Fig. 9-9) , the differences were not significant (Kruskal Wallis; p=0.45). The decline in RFB from the late dry to the early wet season was significant (Kruskal Wallis; p=0.05) and apparent in both breeding and non-breeding females. In the late wet season (September-November) , the RFB increased in nonbreeding animals. The one value for a breeding female (captured adjacent to a pod of hatchlings) , was comparable to non-breeding females and suggests that RFB values remained low following nesting. The seasonal cycle of male (>80 cm SVL) RFB differs from the female cycle (Fig. 9-10) . Although there is much variablility, males tend to have high RFB levels in the dry

PAGE 270

261 CO CO o o CO o Ll. CD 2.000 1 .800 1.600 1 .400 1.200 1.000 0.800 0.600 0.400 0.200 0.000 O O Non-vitellogenic # — # Vitellogenic o-o -8rt4 V .o .O' O' o\6 -A o Figure 9-9. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Relative lateral fat body mass (fat body mass/body mass x 100) among vitellogenic and non-vitellogenic females by month. Data collected 1985-1988. CO CO o o CD o Li_ 0) Figure 9-10, Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Relative lateral fat body mass (fat body mass/body mass x 100) among male caiman. Data collected 1985-1988.

PAGE 271

262 and early wet seasons (JanuaryJune) , and lower values in the mid to late wet season (July-December) . The highest RFB values were in June, at the height of the courtship and mating season. Egg and Clutch Characteristics Caiman eggs are roughly elliptical, with a distinctly rugose outer eggshell that is very distinguishable from the smooth exterior of crocodile and alligator eggs. Scanning electron micrographs of similar egg shells from Caiman latirostris and Melanosuchus niger (Schleich and Kastle 1988) reveal that caiman egg shells possess a reticulate series of deep meandering grooves termed lacunae by Erben (1970) . The function of these unusual structures is unknown . Mean caiman egg mass on Hato Masaguaral was 62.5 g (SD=7.13; 38-85 g; N=1015) (Table 9-8). Mean egg mass did not vary significantly among the five habitats (F4 t 144=0 . 474 ; p>0.05), but did vary among years ( F 4,148 =3 « 27 ' P<0.05) with eggs from 1985 (58.7 g, SD=6.73) being significantly smaller than eggs from either 1984 (64.6 g, SD=9.02; LSD p<0.01) or 1986 (64.1 g, SD=6.81; LSD p<0.05). Mean egg dimensions were: length-6.45 cm (SD=0.30), width-3.96 cm (SD=0.14). The smallest mean clutch egg mass recorded was 4 0.7 g for a nest in 1988. The largest was 80.7 g, also in 1988. Most eggs within a clutch were of relatively uniform size, but on one occasion an

PAGE 272

263 obviously dwarf egg was found (mean clutch egg mass=58.9 g; dwarf egg mass=40 g) . No double or triple yolk eggs were found. Both egg length and egg width were significantly correlated with egg mass (Table 9-9) . Egg mass (EM) was predicted from both egg length (EL) and egg width (EW) by the eguation: EM=12 . 08 (EL) +6 . 85 (EW) -121 . 80 (N=153 ; r 2 =0.88). During five nesting seasons on Masaguaral, clutch size ranged from 4 to 36, with a mean of 22.2 (SD=5.57; Table 910) , and a mode of 23 (Fig. 9-11) . There were no significant differences in clutch size (number of eggs per clutch) among years (F4 f 155=0. 91, p=0.46), or among habitats (F4,152=l73 / P=0.14) . Estimated clutch mass (mean egg mass x clutch size) ranged from 201 g to 2336 g, with a mean of 1392 g. There were no significant differences in clutch mass among years (F4 f 143=0. 27 ; p=0 . 90) (Table 9-10), or among habitats (F4 f 14 o = l • 18 ; p=0.32). There was a slight tendency for heavier clutches to be laid later in the nesting season (Fig. 9-12; r=0.19, p=0.03). Egg mass did not vary significantly among small (6069.9 cm SVL) , medium (70-79.9 cm SVL) , or large (>80 cm SVL) females (F2 / 22 =0 66 » P=0.53). No significant correlations were found between female size and egg mass, egg length or egg width (Table 9-9) . Because there is no significant increase in egg mass with female size, relative egg mass

PAGE 273

264 Table 9—8 Mean egg mass (g) values by year. Values represent means of average clutch egg mass (N=6 eggs measured per clutch) N is the number of clutches measured. Year Mean Minimum Maximum Std. Dev. N 1984 64.6 45. 6 77 . 0 9.02 12 1985 58 . 7 48.2 73.0 6.73 17 1986 62.4 50.3 71.5 7.10 12 1987 60.4 49.2 71.7 5.82 35 1988 64.1 40.7 80.7 6.81 70 All 62.5 40.7 80.7 7.13 146 Table 9-9. Linear correlations between egg, clutch and female size parameters. Clutch means of egg length (cm) , width (cm) and mass (g) measurments were used for calculating correlations. Eqcr Mass (EM) Egg Length (EL) EM=19.50(EL)-63.18 N=153 r=0. 83*** Egg Width (EW) EM=41 . 90 (EW) -103 . 68 N=153 r=0. 84*** Clutch Size (CS) Egg Mass CS=0.241(EM)+7.38 N=152 r=0. 28*** Clutch Mass CS=0.013(CM)+4.54 N=152 r=0. 94*** Clutch Mass (CM) Egg Mass CM=36.66(EM)-875.07 N=152 r=0. 57*** Clutch Size CM=70. 02 (CS) -157. 16 N=152 r=0. 94*** Female Snout-Vent Lenath (cm) Egg Length (cm) EL=0. 008 (SVL) +5.93 N=25 r=0. 20ns Egg Width (cm) EW=0.006(SVL)+3.53 N=25 r=0. 33ns Egg Mass (g) EM=0 . 182 (SVL) +48 . 57 N=25 r=0. 22ns Clutch Size CS=0.592(SVL)-22.02 N=27 r=0. 66*** Clutch Mass (g) CM=42 . 39 (SVL) -1783 .96 N=25 r=0. 69*** ***=p<0.001, ns=p>.05.

PAGE 274

265 y=6.094x + 1 1 64.2 21 July 10 Aug 30 Aug 1 9 Sept Date of Oviposition Figure 9-12. Relationship between clutch mass and date of oviposition for the 1985-8 nesting seasons.

PAGE 275

266 ({egg mass/female mass} x 100) decreases significantly with increasing female size (p<0 . 001) (Fig. 9-13). Table 9-10. Mean clutch size and clutch mass over the period 1984-1988 on Hato Masaquaral. Year Clutch Size Clutch Mass (g) Mean SD N Mean SD N 1984 23.1 5.64 17 1512 527.5 11 1985 24.9 5.21 18 1472 365.0 17 1986 22. 1 6.32 15 1354 479.4 12 1987 21.2 4.75 35 1291 364.6 35 1988 21.9 5.61 69 1411 430.3 69 Total 22.2 5.56 154 1392 426.4 144 Both clutch size (F2,24 =4.53, p=0 .02) and clutch mass ( F 2,22= 4 92, p=0.02) varied significantly among small , medium and large females (Table 9-11) . Significant differences in clutch size and mass were only noted between large and small females (clutch size; LSD, p=0.03: clutch mass; LSD, p=0.03). Also, a least squares linear regression model found both clutch size and clutch mass to be significantly correlated with female SVL (Table 9-9) . A partial correlation analysis of egg and clutch attributes demonstrated that, independent of female size, as clutch size increased there was no significant increase in egg mass or egg width, but there was a decrease in egg length (Table 9-12) . Relative clutch mass (RCM) , the ratio of clutch mass to female body mass, was determined for females captured at

PAGE 276

267 nest sites, sacrificed or road-killed females with oviducal eggs, and nesting females with radio transmitters. RCM values were typically in the 10% to 16% range with a mean of 13.9% (N=26, SD=2.41) (Fig. 914) • Table 9-11. Clutch size and clutch mass for three female size classes • Size-Class Clutch Size Clutch Mass (q) (cm) Mean SD N Mean SD N 60.0-69.9 19.5 2.12 2 1185 141.3 2 70.0-79.9 21.1 3.82 14 1305 267.6 14 80.0-89.9 26.9 6.49 11 1732 448.5 9 Total 23.4 5.71 27 1449 393.0 25 Table 9-12. Partial correlation coefficients (r a b) between the indicated reproductive parameters and clutch size, female SVL held constant. Parameter r a b N Clutch Mass 0.91** 25 Egg Length -0.48* 25 Egg Width -0.01 25 Egg Mass -0.12 25 * p<0.05, ** p<0.001 Egg Viability Egg viability was determined by the presence of a visible opague ring or spot indicating the attachment of the embryo to the inner surface of the eggshell membrane (Ferguson 1985, Webb et al 1987). The absence of a band indicated the egg was infertile, or that very early embryonic mortality occurred, the combination of which is here referred to as inviable eggs.

PAGE 277

1.500 CO CO o 1 .000-cn Ld CD 0.500 -t— < _o CD 25 CO co 20 o 1510Y=-0.1 17x4-22.727 r=0.30 o o 0 o >8 o o o 0 — I o° 0 1 o 1 60 70 80 90 100 Figure 9-14 Female SVL (cm) Relative clutch mass (clutch mass/body mass : 100) as a function of female snout-vent length.

PAGE 278

Over the five nesting seasons mean egg viability rates ranged from 82.2% to 99.7%, with an overall mean of 94.4% (Table 9-13) . Mean egg viability was significantly lower in 1984 (82.2%) than in all other years (F 4 ( 127=3 . 11, p<0.05). Table 9-13. Mean annual egg viability rates over the period 1984-1 9 8 8 . Year Mean Eaa Viabilitv (%) SD N 1984 1985 1986 1987 1988 82.3 98.0 99.7 98.3 93.7 25.9 2.9 0.9 3.4 19.9 16 10 13 32 61 Mean 94.4 17.0 132 Breedina Effort The percentage of adult females that nest in a given year was determined for a sample of 81 females (>59.9 cm SVL) over the period 1985-1988. Reproductive state was ascertained for road-killed (DOR; N=28) , sacrificed (N=30) , natural mortality (ND; N=2), and radio-tracked (N=21) females (Table 9-14) . Among DOR, sacrificed, or ND specimens, only caiman examined during the reproductive season (May-September) were used to determine breeding status . During the period 1985-1989, the mean annual breeding effort of females over 59.9 cm SVL in the Hato Masaguaral study population was 54.3%, indicating that on average an adult female nested approximately every other year.

PAGE 279

Breeding effort varied from a low of 44.8% in 1985 to a peak of 68.8% in 1988, but no significant differences were noted among years (F3 / 77=0. 68 ; p=0.57). Breeding effort was strongly positively correlated with female size (Table 915). Only 16.7% of the females in the 60-65 cm SVL range were reproductive, whereas the corresponding figure for females over 80 cm SVL was 77.8%. Size-class Distribution of Breeding Female Population Because there was a significant effect of female size on reproductive parameters, the size of the nesting female can be predicted from the egg and clutch data collected at nests. With a large sample size of nests this, in turn, can be used to make a prediction about the size-class structure of the breeding female population. To this end a multiple linear regression equation was constructed to predict female snout-vent length based on egg and clutch characteristics. The model predicted female size (SVL) based on clutch mass (CM), egg length (EL) and clutch size (CS) : SVL=15.522EL +1.653CS-0.012CM-45.725 (r 2 =0.61), and predicted the sizeclass distribution of females nesting on Hato Masaguaral from 1984-1988 (N=144 nests; Fig. 9-15). A separate means of estimating the size structure of the breeding female population is to multiply the known size structure of the female population (based on captures) by the size-class specific breeding frequencies. The predictions from these two methods were in general

PAGE 280

271 Table 9-14. Percentage of females over 59.9 cm SVL breeding as derived from four sample populations. All samples taken during the months May-September over the period 1985-1988. DOR Sacrificed ND Radio Total Sample Size 28 30 2 21 81 Minimum SVL (cm) 60.0 60.0 77.4 61.0 60.0 Maximum SVL (cm) 86.1 82.8 80.5 82.5 86.1 Mean SVL (cm) 70.1 73.4 79.0 76.5 73.2 Std Dev. SVL (cm) 7.46 5.54 1.55 5.15 6.66 No. Breeding 14 16 1 13 44 % Breeding 50.0 53.3 50.0 61.9 54.3 Table 9-15. Breeding effort of female caiman greater than 59.9 cm snout-vent length by size class and year. Data are presented as number breeders/sample size and percent Size-class 1985 1986 1987 1988 Total (cm SVL) 60-65 0/1 1/3 1/3 0/5 2/12 0.0% 33.3% 33.3% 0.0% 16.7% 65-70 1/2 0/5 3/4 0/0 4/11 50.0% 0.0% 75.0% 0.0% 36.4% 70-75 1/2 5/8 5/9 2/2 13/21 50. 0% 62.5% 55.6% 100. 0% 61.9% 75-80 0/0 2/7 6/9 3/3 11/19 0.0% 28.6% 66.7% 100.0% 57.9% 80-85 1/1 5/5 1/5 6/6 13/17 100.0% 100.0% 20.0% 100.0% 76.5% 85-90 0/0 1/1 0/0 0/0 1/1 0.0% 100.0% 0.0% 0.0% 100.0% Total 3/6 13/29 16/30 11/16 44/81 50.0% 44.8% 53.3% 68.8% 54.3%

PAGE 281

272 agreement, with the majority of the breeding caiman population composed of females between 70 and 80 cm SVL (Fig. 9-15) . Because of the higher breeding effort of the large females, they were overrepresented in the nesting freguency sample when compared to the actual size-class distribution of captured females (Fig. 9-15) . Discussion Seasonal Reproductive Cycle The reproductive cycles of reptiles, including crocodilians, are usually tightly correlated with environmental factors (Duval et al. 1982) . However, the environmental factors controlling the onset of gonadal recrudescence and the timing of reproductive events in caiman are unknown. Among the more temperate species of crocodilians (e.g., Alligator mississippiensis ) . the reproductive season appears to be constrained by temperature limitation. Good correlations have been found between spring temperature regimes and mean nesting date for alligators (Joanen and McNease 1979, 1989, Kushlan and Jacobsen 1990) . Environmental factors other than temperature that control seasonal reproductive cycles among reptiles include photoperiod, rainfall, moisture, humidity, and food supply (Duval et al. 1982). Caiman in the Venezuelan llanos exhibited a highly seasonal reproductive cycle that was stongly correlated to

PAGE 282

273 50-55 55-60 60-65 65-70 70-75 75-80 80-85 85-90 Size-Class (cm) Figure 9-15. Estimated size-class distribution of the breeding female population based on 1) predictions from clutch and egg parameters, and 2) the female size-class distribution and size-class specific breeding effort. Stippled bars represent the size-class distribution of captured females.

PAGE 283

274 the annual cycle of rains and flooding. Vitellogensis and testicular recrudescence both began in the late dry season (April) , two to three months prior to the peak period of courtship and mating, and approximately four months before the peak nesting period. Ovulation and peak sperm motility occurred during the early wet season, coincident with the dispersal of caiman from dry season lagoons, and the establishment of breeding territories by males. An earlier seasonal onset of gonadal recrudescence would result in courtship and mating beginning during the late dry season under high density conditions. Under these circumstances the establishment of breeding territories by males is very difficult. Courtship and mating under these high density conditions is also very inefficient, with many attempts being interrupted by neighboring males (chapter 8) . Studies on American alligators reveal that higher densities are associated with higher stress levels (as measured by plasma corticosterone) and lower nesting success (Elsey et al. 1990) . A later onset of gonadal recrudescence would delay nesting and result in hatching occurring at the very end of the wet season. Studies on nesting ecology indicate that this would lead to increased nest depredation (chapter 10) , and probably higher neonatal mortality as well. The resulting timing of gonadal recrudescence allows the caiman to begin courtship and mating immediately following dispersal from the dry season lagoons, and ensures that the

PAGE 284

275 peak of oviposition will take place during the height of the annual savanna water levels. Lance (1989) suggested that ovulation may be a response to copulation or courtship. If this is true it suggests another disadvantage of courtship and mating prior to dispersal. During extremely dry years, suitable conditions for nesting may not be available. Although no evidence was found during this study indicating that females would enter a phase of general atresia, it is a physiologically feasible means of recuperating the energy invested in ovarian follicles. Following ovulation there is no known means by which a female can resorb the energy invested in the shelled eggs. This suggests that once a female ovulates, she is committed to ovipositing. By courting and mating (and ovulating) after dispersal from the dry season lagoons (when the savannnas are flooded) , females may be able to avoid nesting during extremely dry years when nesting success would presumably be very low. If courtship and mating occurred in the dry season lagoons, and the females ovulated prior to dispersal, females run the risk of having to nest during a year when extensive flooding did not occur. The rains at the study site were delayed in 1988, but extensive flooding still occurred by late June-early July and nesting levels were very high. However, in the year following the end of this study (1989) the rains were delayed even longer, extensive flooding did not occur until mid-late July, and the observed nesting levels were very low (M. Munoz, pers.

PAGE 285

276 comm.)These observations indicate that as long as the caiman are able to disperse by early July, normal nesting can take place. If caiman dispersal is delayed much beyond this, nesting levels are probably severely diminished. One conseguence of the timing of the reproductive events is that caiman nests hatch during the latter part of the wet season. This provides the young with only a few months of wet season environment prior to the onset of the stressful dry season. This results in a very high mortality of hatchling caiman during their first months of life. Nesting during the early wet season would probably increase neonatal survivorship to some degree, but is unfeasible in the context of the environmental factors (low water levels, high population density and unavailability of nesting material) that govern the reproductive cycle. Seasonal changes in the mass of the lateral fat body were evident both among male and female caiman, however the pattern was sexually dimorphic. Derickson (1976) defined four types of reptilian lipid cycling patterns: 1) no cycling, 2) storing and utilization for periods of dormancy, 3) storing and utilization for reproduction, and 4) storing and utilization for both dormancy and reproduction. However, the seasonal cycle of relative fat body mass on Hato Masaguaral does not clearly support one or the other of these patterns. It is evident that females are utilizing stored lipids throughout the dry season, and the decrease in relative fat body mass among vitellogenic females is

PAGE 286

277 consistent with mobilization of energy for reproduction. However, this trend was also seen in non-vitellogenic females, suggesting that lipids are being used for maintenance during this period of reduced food availability. Nevertheless, no such tendency was found among males, who exhibited what Derickson (1976) suggested was the typical pattern for tropical reptiles living in a seasonal environment, with the largest relative fat body masses found during the late dry and early wet season (March-June) . The fat body mass of non-reproductive females increased throughout the dry season, but decreased among males. These seasonal lipid patterns did not parallel any seasonal changes in the guantity of food consumed (chapter 6) , but suggest that the dry season was a more energetically stressful time for females than for males. The low stored lipid levels among males in the wet season may reflect increased energy demands associated with the establishment and maintenance of territories. Sexual Maturity and Size-Specific Fecundity The definition of sexual maturity is less clear cut and more difficult to demonstrate in male caiman than it is in females. Males may be physiologically mature (i.e., producing viable sperm) but be excluded from mating by larger, more dominant males (Modha 1967, Joanen and McNease 1975, Lang 1987, Lance 1989). Because of this, the size at which males start to engage in courtship and mating may also

PAGE 287

278 depend on social factors such as density and the population size-class distribution. The discussion presented here deals with the attainment of physiological maturity in male caiman, as opposed to the size/age at which males may actually reproduce. Testes mass and spermatogenic activity are closely related in crocodilians (Joanen and McNease 1980, Lance 1989) . Based on histological examination of testes, the minimum size of mature testes (during the seasonal peak in testes mass) was ca. 15 g, suggesting that caiman on Masaguaral are capable of producing sperm at approximately 75 cm SVL, and an average age of seven years. It is likely that some individuals may become sexually mature at smaller sizes (approximately 70 cm SVL) , and in fact Ayarzagiiena (1983) reported obtaining sperm from two males measuring 69.5 cm SVL and 73.5 cm SVL on Hato El Frio in Apure state. Graham (1968) examined testes from the much larger species Crocodylus niloticus and noted that the percentage of animals producing sperm increased with increasing testicular mass, but found sperm in some animals with a testicular mass of 6-10 g (corresponding to a paired testes mass of approximately 12-20 g) . Although the largest male measured on Hato Masaguaral was 127.0 cm SVL, males do grow significantly larger in other parts of the llanos. In a sample of 144 caiman harvested from various ranches in Apure State, five (3.5%) were over 130 cm SVL and the largest measured 139.5 cm SVL

PAGE 288

279 (C. Molina pers. comm.)* From a sample of 34 caiman harvested in 1989 on Hato Merecure, Apure State, eight (23.5%) were over 130 cm SVL, and the largest was a 145.5 cm SVL (Thorbjarnarson, personal observation) . Maximum size for male caiman in the llanos is probably over 150 cm SVL. Small adult male caiman (approximately 70-80 cm SVL) bellow, but it is unknown if they actually reproduce. Evidence from the study of caiman movement patterns (chapter 4) suggested that males less than 90 cm SVL did not disperse very far, but tended to have large home ranges, indicating they may not seek out breeding territories. Males in the 90-100 cm size class, however, tended to disperse greater distances than large males, suggesting they may be relegated to marginal areas for establishing breeding territories. Large males apparently monopolize the preferred breeding habitats located close to the dry season lagoons. Small male caiman do reproduce in captivity. Alvarez del Toro (1969) noted that a 1.5 m TL male Caiman crocodilus ( chiapasius ) (approximately 79 cm SVL assuming its tail was complete) mated with a 1.2 m TL female, who produced fertile eggs. Hunt (1969) reported a male Caiman crocodilus 1.65 m TL (approximately 87 cm SVL) that successfully mated with a female in the Atlanta Zoo. Stribrny (1978) reported a 1.13 m TL male copulating with a female, but the female laid the eggs in the water from 1974-1976. In 1977, the eggs were laid in a nest and were found to be fertile when the male was 1.6m TL.

PAGE 289

280 Female Caiman crocodilus in the llanos began reaching sexual maturity at a size of 60 cm SVL, and an average age of six years. The minimum size of sexual maturity of caiman on Masaguaral agrees well with other estimates made for the Venezuelan llanos (Ayarzaguena 1983), but is somewhat smaller than the minimum size (67.7 cm SVL) noted by Staton and Dixon (1977) . Because caiman mature at a small size for crocodilians (Table 9-16) and natural growth rates for most species are not grossly dissimilar over the first few years of life, caiman achieve sexual maturity at a young age for a crocodilian. Females on Hato Masaguaral may become sexually mature in as little as 4-5 years, although in most cases 6-8 years are reguired (chapter 3) . This has important implications for the population dynamics of the species and is one of the reasons caiman populations have been more resilient to human exploitation than larger species of crocodilians (Magnusson 1982) . Reproductive Parameters and Female Body Size The relationship between egg size and female size varies considerably among species of reptiles (Ford and Seigel 1989) . A number of studies have demonstrated an increase in egg size with female size, but others have found no such relationship. Interspecific comparisons among crocodilians reveal that larger species do lay larger eggs (Table 9-16) , and this also appears to hold for many

PAGE 290

rH O cO > 0) -P <4H (0 I E-" n •H 0) o w d) co a e 0) i s u oj 05 a) a (0 U o x: c 0 a) p rH CQ u 1 o E u n cr 0 0) > 0 •« to a> (c •h e u <0 0) e rH +j e (0 nj 3 Eh woo 05 u <*> 05 — 5 rH •» CO" a) a) CO +J N c U O XI •H 1 u -P CO O O U X! 3 Eh 73 O 10 E se u to o fa X a re ^ 0) E fa c tn • to 0) 1 to •J "e ess H W E BO a) •H o 0 a w vo cm cn cn cn oo cm h in in h h o O O O O O H rH rl P> ffl O N CO O 10 H t >* Mtt VO H H H H rH 10 •H 10 c o •H a o -rl to to Vh (0 •H to •H 0 CO to 3 rH & •H c rH p H 10 0 H to c o (0 c TS o (0 •H •H 0 u to E to 0 •H •rl 0 -p x; U M u (0 u 0 o 0 u rH 3 p -p -p to (0 CO CD c c 0 0 0 0 CO (0 c •H •H •H E E CO rH rH rH --H •H rH rH rH rH (0 (0 0 < < < 0 u s ID U Qj-P ininHVOfflNniMCM^OtfJNO HnHflNClrlNNHNINrllO oooooooooooooo coinnvooir^Hin^rsovocoN ^rtn^ i *?cncnmvo<4 | i0<*tncMco CO ID rH vo vo in i VO o VO rH in n in in cn O in H 1 VO CO rcn VO H O 1 cn co tf CO CM VO in rH p» rH cn cn 1 cn IT) rH ID CM CO cn O 1 H CO cn co CM cn in in VO in cn CM vo I o CM rH rH CM in H 1 CM rH CM CM CM in CM CM cn VO o cn O VO I> ^jCO 1 VO O rH in CM CM in in rcn vo in r~ i n in cn cn n cn i cn cn CM CM VO O CO cn o cn CM 1 • rM CM CM CM n rH rH | cn rH * rH rH cn CM CM m CM CM CM rH I cn rH CM VO CM vo VO VO 1 CO O in I"» CM in CM in in o cn cn cn O 1 H cn CO CM VO cn CO CM 1 CM VO in cn cn cn CO cn o CO vo cn in i • ># p» tf vo r» vo ri H H vo rr*> o CO cn rH o o cn in I VO H H rH rH rH rH rH rH H H r» VO cn VO O CO CM 1 in CO in o m in cn rrcn in o cn I O o rH H rH rH rH rH n H o co H 1 VO CM m co co vo rH CO CM n CO CM in cn ID vo co o CN 1 VO m cn co VO co CO o n VO CM cn o cn i CM CM H H H CM H rH I CN CM CM rH H rH CM CM CM CM CM CM cn H 1 cn > O o o u o to 3 -P 3 O CO to 3 rH > O o o u u to 3 -P u CO u x; Q CC -p u to 3 rH > T5 O D O U > > > > > o vw •H X! E O x: Sh to 3 rH > O U O u to •H D 10 CO U o -p to 3 E o CD rH o o -p 10 O

PAGE 291

282 intraspecific comparisons as well (Table 9-17) . Theories of optimal egg or offspring size predict an inverse relationship between clutch size and egg size (Smith and Fretwell 1974, Brockelman 1975) and some evidence has been found to support these theories among lizards, but not among other reptiles (Congdon and Gibbons 1985) . Among the caiman in this study a simple regression of egg size on female Table 9-17. Summary of relationships between female size and reproductive parameters for crocodilians at the intrapopulation level. + indicates a significant positive relationship, 0 indicates no relationship. Species Egg Clutch Clutch Breeding Source Size Size Mass Effort C. niloticus A. mississippiensis + 0 + + + + + 0 + 0 + + + 0 A B C D E F H A-Cott 1961 B-Graham 1968 C-Hutton 1984 D-Ferguson 1985 E-Deitz and Hines 1980 F-Wilkinson 1984 G-Joanen 1969 H-Joanen and McNease 1980 length found no clear relationship between these variables. Even by controlling for female size no significant change in egg mass was found as clutch size increased. Hence this study provides no real support for the idea of a trade off between egg size and clutch size. No evidence was found either for morphological constraints on egg mass, which have been noted in some species of small freshwater turtles (Congdon and Gibbons

PAGE 292

283 1987) . Although larger females had larger pelvic canals (Thorbjarnarson, unpublished data) , they produced eggs that were not significantly larger than those produced by small females. However, caiman egg length decreased with increasing clutch size, and this may be related to morphological factors. The number of eggs that can be shelled at any one time should be limited by the length of the uterus, so for a female of given size, an increase in clutch size should result in less space being available per egg. This tighter packing of eggs in the oviduct during shelling may be responsible for the observed inverse relationship between egg length clutch size. Fecundity in female caiman is manifested in two parameters: the freguency of breeding, and clutch size. Both these parameters were shown to increase with body size, demonstrating a clear size-specific trend in fecundity. Reported values of the mean annual percentage of females nesting vary considerably among crocodilians. Studies on the genus Crocodylus have generally reported high annual nesting values: 90% (C. iohnsoni : Webb et. al 1983d), 87.6% (C. niloticus : Graham 1968), 72% (C. acutus; Mazzotti 1983) and 63.8% (C. acutus ; Thorbjarnarson 1988). Work on American alligators has suggested somewhat lower values: 68.1% (Louisiana; Chabreck 1966), 63% (Louisiana; Joanen and McNease 1980), 29.8% (Louisiana, Kinler et al. 1987), 29% (Florida Everglades; Kushlan and Jacobsen 1990), 28% (Louisiana; Taylor 1984), 25% (South Carolina, Wilkinson

PAGE 293

284 1984), and <10% (North Carolina, Lance 1989). Other than the present study on Caiman crocodilus . the only investigations that have examined annual variation in breeding effort were with American alligators. In the Florida Everglades, it was found that over a seven-year period the percentage of nesting females ranged from 16% to 58% (Kushlan and Jacobsen 1990) . A radio telemetry study of alligators in northwestern Louisiana found nesting levels of 33.3% (N=9 females), 12.5% (N=8) and 37.5% (N=8) over a three-year period (Taylor 1984) . Over a four-year period nesting effort on Hato Masaguaral was relatively constant, varying from 48.3% in 1986 to 68.8% in 1988. The overall mean was 54.3%. Although some females began nesting at a size of 60 cm SVL or slightly below, most individuals did not nest until they reach a slightly larger size. Average size at first reproduction was estimated to be 64.0 cm SVL, but some females did not nest until they were over 70 cm SVL. Small females reproduced at more infrequent intervals than larger females. A similar relationship has been shown among female Nile crocodiles (Cott 1961) and American alligators (Wilkinson 1984). However, Joanen and McNease (1980) found no size-related difference in breeding effort among 2 5 alligators in Louisiana. Larger females also tended to have larger clutch masses. No significant trend was found between female size and egg size, so clutch mass was increased by laying more

PAGE 294

285 1984), and <10% (North Carolina, Lance 1989). Other than the present study on Caiman crocodilus . the only investigations that have examined annual variation in breeding effort were with American alligators. In the Florida Everglades, it was found that over a seven-year period the percentage of nesting females ranged from 16% to 58% (Kushlan and Jacobsen 1990) . A radio telemetry study of alligators in northwestern Louisiana found nesting levels of 33.3% (N=9 females), 12.5% (N=8) and 37.5% (N=8) over a three-year period (Taylor 1984) . Over a four-year period nesting effort on Hato Masaguaral was relatively constant, varying from 48.3% in 1986 to 68.8% in 1988. The overall mean was 54.3%. Although some females began nesting at a size of 60 cm SVL or slightly below, most individuals did not nest until they reach a slightly larger size. Average size at first reproduction was estimated to be 64.0 cm SVL, but some females did not nest until they were over 70 cm SVL. Small females reproduced at more infreguent intervals than larger females. A similar relationship has been shown among female Nile crocodiles (Cott 1961) and American alligators (Wilkinson 1984). However, Joanen and McNease (1980) found no size-related difference in breeding effort among 25 alligators in Louisiana. Larger females also tended to have larger clutch masses. No significant trend was found between female size and egg size, so clutch mass was increased by laying more

PAGE 295

286 3500 3000 RE (kcal) RRE (%) 60 65 70 75 80 Female Length (cm SVL) Figure 9-16. Estimated mean annual reproductive expenditure expressed in kilocalories (RE) , and as a percent of total body caloric value (RRE) , as a function of female size.

PAGE 296

287 of its effects on female survivorship and foraging mode. But RCM has also been used in some studies of reptiles as a measure of reproductive effort (e.g., Vitt 1974, Tinkle and Hadley 1975, Jackson 1988). Problems with the use of this ratio for lizard populations have been reviewed by Cuellar (1984), but the use of RCM as a measure of reproductive effort for turtles was considered to be valid by Jackson (1988) . However, it is clear that because RCM does not account for differences in breeding freguency, it is unsuitable for intrapopulation comparisons of reproductive effort. This same statement can also be applied to interpopulation or interspecific comparisons as well. A simple measure of the relationship between female mass and clutch mass is not sufficient to estimate reproductive effort because the female may not reproduce every year (as with most crocodilians) , or she may lay multiple clutches in one year (as with many species of lizards and turtles) . The RE value calculated here represents a minimum value based on the energetic cost of egg production. No attempt has been made to guantify other energetic costs associated with reproduction, such as nest construction, nest defense, defense of the neonates, or the opportunity costs of nest attendance (e.g. loss of foraging opportunities) . These costs are most certainly higher for crocodilians than for other reptiles who do not have a well developed system of parental care.

PAGE 297

CHAPTER 10 NESTING ECOLOGY Introduction Nesting has long been realized as an important link in the life history of reptiles. However, it is only recently that we have begun to understand just how critical the period of embryonic development is for crocodilians. The initial work on the importance of the nest environment on crocodilian ecology focused on temperature-dependent sex determination (Ferguson and Joanen 1982, Hutton 1987, Webb et al. 1987) . More recently it has become clear that conditions experienced by the developing embryos also affect a host of other non-sexual factors such as embryo survivorship, body size, and the frequency of abnormalities and pigmentation of hatchlings, post-hatching growth, survivorship and thermoregulation patterns (see review in Webb and Cooper-Preston 1989) . Hence the period the crocodile spends in a egg to a large degree determines its future prospects, and is the most critical 2-3 months of its life. Because of these recent findings, much of the focus of current work on nesting biology has centered on investigations on the effects of incubation conditions on embryonic sex determination, growth, survival, and post288

PAGE 298

289 hatching fitness (Ferguson and Joanen 1982, Joanen et al. 1987, Webb et al. 1987, Deeming and Ferguson 1988). Despite the obvious importance of these topics, investigations such as these need to be complemented by studies of nesting ecology under field conditions if adequate interpretations of laboratory findings are to be sought. Studies of crocodilian nesting habits have been conducted on a large number of species (Ferguson 1985) , including the spectacled caiman (Alvarez del Toro 1974, Rivero Blanco 1974, Staton and Dixon 1977, Crawshaw and Schaller 1980, Medem 1981, Ayarzagviena 1983, Crawshaw 1987, Ouboter and Nanhoe 1987, and Cintra 1988). Nevertheless, our understanding of many aspects of crocodilian nesting ecology remains relatively poor, and additional work is needed to provide useful insights into the recent findings regarding embryonic development and how these may affect the species population ecology. Of particular interest in this regard are the factors that influence nest site selection, nest hatching rate, and the seasonal timing of nesting. The objectives of this study were to quantify aspects of spectacled caiman nesting as part of an overall study of the species' reproductive ecology. Among the topics investigated were physical nest characteristics, nest thermal relations, nest fate, and nest timing, and how these parameters varied with respect to relevant environmental parameters. The results of this study are compared with other studies conducted in the llanos, as well as

PAGE 299

290 investigations of caiman nesting in the ecologically similar Brazilian Pantanal. Methods Studies of the nesting ecology of the spectacled caiman were conducted principally on Hato Masaguaral. A small sample of nests was located on the neighboring ranch, Flores Moradas, and in 1986 and 1987 brief visits were made to Hato La Guanota in Apure state (85 km S of Masaguaral) to locate nests. Nests were found by slowly searching through potential habitat on foot or on horseback. In certain areas (e.g., along roadsides) searching could also be done from atop a vehicle. All nests were marked with plastic flagging tape (all years) as well as numbered metal markers (1984-5) or plastic gardening tags (1987) to allow the identification of nest sites used in previous years. Nest microhabitat and predominant vegetation were briefly described for each nest. Nest mound dimensions were recorded to the nearest centimeter. Nest height was measured from the base of the nest with the aid of a string level. The volume of the nest mound was calculated using the formula for a hemi-ellipsoid (Crawshaw 1987) . A list was made of material used in the construction of the nest mound, in order of predominance, and the position of the nest in relation to the nearest tree was recorded (distance and compass orientation) . The amount of surface flooding 5 m around each nest was visually estimated to the nearest

PAGE 300

291 10%. The presence or absence of an attending adult caiman was noted, although the water around nests was not extensively searched (in order to reduce the likelihood that the disturbed adult would abandon the nest) . In order to test for the effects of nest opening on subseguent nest predation, only a sample of nests was opened in 1984, 1985 and 1986. In 1987 and 1988 virtually all nests found early during the nesting season were opened. Eggs removed from the clutch cavity were carefully marked with a wax pencil or ink marker to preserve proper orientation at all times. Among nests that were opened, depth to the egg clutch (top and bottom) was noted, along with clutch size, the presence or absence of ants and termites, and mean egg dimensions and weight for a sample of six eggs. When visible, the egg band width (minimum and maximum width) was measured. For clutches where the egg bands had not reached the egg poles, the rate of egg inviability was determined by the percentage of unbanded eggs. Egg dimensions were measured with calipers to the nearest 0.01 cm and egg weight was determined with a 100 g Horns spring scale to the nearest 1 g. To monitor interior nest temperature a sample of nests had chromel-alumel thermocouples implanted in the center of the egg clutch in 1984, 1987 and 1988. Nest temperatures were recorded at regular intervals on specific days, and in 1987 and 1988 were also checked at regular weekly intervals. Temperature was measured using a hand-held Omega 871A

PAGE 301

292 digital thermometer (0.25% accuracy) to the nearest 0.1'C. The distance and direction of nests from the four principal permanent bodies of water in the study area (Guacimos, Piscina, and Alta Venega lagoons and a borrow pit adjacent the ranch entrance) were measured from aerial photographs (scale 1: 11, 300) . Nesting studies in 1984 began in October, too late to permit the estimation of oviposition date. In subsequent years when nests were opened, one egg was collected and opened to estimate embryo age and date of oviposition. Eggs less than 10 days old were aged principally by band width. Eggs greater than 10 days old were aged by embryo head length as determined from calibration curves for nests artificially incubated under natural conditions in 1985, 1986 and 1987. Comparison of egg band width or embryo head length with the calibration curves resulted in a range of potential oviposition dates. The best estimate of oviposition date was taken to be the midpoint of the range. During the 1985-1988 nesting seasons a sample of nests was visited at weekly intervals to note depredation, nest flooding, or hatching. The specific identity of the nest predator was assigned based on the manner in which the nest was opened. Nests opened by adult caiman at times appeared to superficially resemble predated nests, but usually could be distinguished by the manner with which the eggs had been opened and by their position relative to the nest cavity. Some of the nests were visited at more frequent intervals

PAGE 302

293 close to the estimated time of hatching in order to ascertain more precisely the date of hatching, as well as to mark the hatchlings. Nests found within one or two days of hatching would usually exhibit signs of recent hatching (e.g., moist eggshell membranes) and the date of hatching could be closely estimated. Where this was not possible, the date of hatching was estimated to be the midpoint between the date the nest was found hatched, and the previous visit. Date of nest depredation was estimated in a similar fashion. The presence of dead or unhatched eggs was noted at all opened nests. The use of the terminology for nesting and hatching success follows Crawshaw (1987) , where nesting success is the percentage of nests that hatched at least one egg successfully, and hatching success is the proportion of eggs that hatched in any one nest. Results Over five nesting seasons (1984-1988) a total of 271 nests were located (1984-28 nests; 1985-45 nests; 1986-60 nests; 1987-59 nests; 1988-87 nests). The fate of 217 (80.1%) of these nests was determined. Of the total nest sample, 262 were located on Masaguaral or in the neighboring ranches Hato Flores Moradas or Hato Matadero. A total of 9 nests was found on Hato La Guanota, four in 1986 and five in 1987.

PAGE 303

294 Nest Characteristics Caiman nests were composed of live or dead vegetation, decomposing organic material, or soil scraped together into a small, roughly hemispherical mound. Females would begin collecting nest material one to two weeks prior to oviposition. Nest material was initially scraped together into a small mound to which new material was added periodically. Material from within a 3-4 m radius around the nest was utilized, and in some cases females had to utilize submerged decaying organic matter for nest construction. The mound did not take on its final appearance until after oviposition, when the female would shape it into a well formed mound. If nest mounds were disturbed prior to oviposition, females would freguently abandon the site and move to a new nest site. All nest construction took place at night. In certain cases females periodically added new nest material to the mound during the course of incubation. Eggs were deposited in a roughly spherical mass in a hole excavated near the top of the nest mound. Mean depth to the top of the egg clutch was 16.9 cm (SD=5.39, range=533, N=172) and to the bottom of the egg clutch was 27.9 cm (SD=5.31, range=15-45, N=125) . No significant differences were found in clutch depth either among years or habitat types . Nest material varied greatly depending on the habitat and the nature of the available organic material. The

PAGE 304

295 materials used most commonly in nest construction were grass, small diameter woody vegetation (twigs and sticks), deciduous leaves from trees and shrubs, fallen palm petioles, or other herbaceous vegetation such as the stems and leaves of herbs. The use of these nest materials varied among habitats as a function of their local abundance (Table 10-1). In palm savanna habitats, grass was the principal nest material with 72 of 83 nests (86.7%) containing grass, and 69 (83.1%) having grass as the principal nest component. Caiman nests in sandhill habitats were also composed principally of grass, with 61 of 67 nests (91.0%) containing grass and 64.2% having grass as the principal nest material. The nest material from 36 nests in forests was more varied than in sandhill or palm savanna nests, and was composed primarily of dead tree and shrub leaves (present in 58.3% of nests; principal component in 33.3%), and twigs and sticks (77.8%; 19.4%). Spoil bank nests (N=22) were built mostly of tree and shrub leaves (present in 72.73% of the nests; principal component in 22.7%), grass (59.1%; 36.4%), and other herbaceous vegetation (50.0%; 27.3%). Marsh nests (N=20) were made mostly from grass (75.0%; 65.0%). Mean nest dimensions were: length-108.0 cm (SD=2 2.08, range=60-170, N=233) ; width-88.0 cm (SD=17.37, range=40-135 , N=209); height-39.6 cm (SD=9.77, range=20-80, N=218) . Average nest volume was 0.197 m 3 (SD=0.098, range=0.0430.534, N=205; Table 10-2). Nest volume did not vary significantly among years (F 4 f 2 00 =0 • 227 ? P>0.05). However,

PAGE 305

296 significant variation in nest volume was noted among habitat types (F4 /195 =3.69; p<0.01). Nests in forest habitats (mean volume 0.238 m 3 ) were significantly larger than nests from all other habitats except palm savanna (mean 0.215 m 3 ; LSD p<0.05). The smallest nests were found along raised dikes or roadsides (spoil banks) and were significantly smaller (mean 0.161 m 3 ; LSD p<0.05) than nests in both forest and palm savanna habitats. A large percentage of the palm savanna and marsh nests was constructed using soil, usually a fine grained clay. When baked in the heat of the sun, this material would form a hardened outer surface that was very difficult to penetrate. Significant differences in mean nest height ( F 4,209 =6 « 77 ' P<0.001) and egg height (=nest height-depth to clutch top) (F4 f 163=3 . 76 ; p<0.01) were found among habitat types. Egg clutches were elevated higher in palm savanna and marsh habitats (Table 10-3). Clutches in these habitats were significantly higher than those in sandhill and spoil bank habitats (LSD; p<0.05) and marginally higher than in forest habitats (LSD; p=0.06). These differences reflect the probability of nest flooding among the different habitats which, based on topographic and hydrographic factors, can be ranked in decreasing order as: marsh, palm savanna, sandhill, forest, and spoil bank. These rankings do not reflect the actual rate of nest flooding among habitats (see Nest Fate, below) because this has changed in

PAGE 306

297 Table 10-1. Nest material used in different habitats. Material categories are: grass (GR) , soil (SO) , deciduous leaves (DL) , palm petioles (PP) , twigs and sticks (TS) , other herbaceous vegetation (HV) and sand (SN) . Classifications are: 1) material present in nest (present) , 2) material is principal component of nest mound (principal) . Values are percent of all nests of that Habitat GR SO DL PP TS HV SN Palm Savanna a. present 86.7 69.9 57.8 54.2 51.8 18.1 0.0 b. principal 83.1 7.2 4.8 10.8 4.8 1.2 0.0 Sandhill a. present 91.0 20.8 55.2 0.0 74.6 32.8 40.3 b. principal 64.2 3.0 7.5 0.0 10.5 7.5 3.0 Forest a. present 27.8 30.6 58.3 47.2 77.8 22.2 0.0 b. principal 16.7 5.6 33.3 11.1 19.4 2.8 0.0 Spoil Bank a. present 59.1 30.6 72.8 0.0 68.2 50.0 2.8 b. principal 36.4 2.8 22.7 0.0 9.1 27.3 0.0 Marsh a. present 75.0 60.0 35.0 5.0 70.0 45.0 5.0 b. principal 65.0 0.0 15.0 5.0 10.0 5.0 0.0 Table 10-2. Mean nest volume (m 3 ) by habitat. Nest volume was calculated from length, width and height data using the formula for a hemi-ellipsoid. Habitat Tvpe Mean Volume SD N Sandhill 0.173 0.107 64 Palm Savanna 0.215 0.101 73 Spoil Bank 0.161 0.065 18 Marsh 0.179 0. 051 18 Forest 0.238 0.090 27 Overall 0.197 0.098 200

PAGE 307

298 recent years due to human intervention. Damming for cattle management purposes has had a disproportionately greater effect on sandhill nests, and 8 of 12 (66.7%) nests flooded on Hato Masaguaral were in this habitat type. Significant differences in nest height were also noted among years (F4 213= 4 « 20 '* p<0.01)). Mean nest height was the greatest in 1984 (43.2 cm) and 1987 (43.3 cm), significantly higher than in 1986 (36.9 cm) and 1988 (37.3 cm; LSD; p<0.05) and marginally higher than in 1985 (39.1 cm; LSD; p=0.07). Similar annual differences in egg height were also evident with values for 1984 and 1987 being higher than all other years (LSD; p<0.05). The reasons for these annual differences in egg height are unknown but presumably are related to the likelihood of nest flooding (Kushlan and Jacobsen 1990) , with egg clutches being elevated higher in wetter years. However, no correlations between nest or egg height and rainfall or July/August water levels were found. Table 10-3. Mean nest height (cm) and egg height (cm) for nests from five different habitats on Hato Masaguaral. Habitat Type Nest Heiaht Eaa Heiaht Mean SD N Mean SD N Sandhill 35.3 8.18 66 19.9 6.78 55 Palm Savanna 43.2 10.31 79 24.7 8.44 58 Spoil Bank 36.6 7.78 20 20.0 7.48 17 Marsh 39.8 7.78 18 24.8 10.11 13 Forest 39.9 10.17 31 20.7 5.49 25 Overall 39.5 9.74 214 22.1 7.84 168

PAGE 308

Caiman nesting takes place when much of the available savanna habitat is flooded and the raised nest mounds serve as convenient sites for ants and termites to occupy. A significant number of ants (i.e., greater than a few isolated individuals) was found in 86 of 179 nests (48.0%). Termites were noted in 34 nests (19.0%). Only one nest was found adjacent to a termite mound. No obvious nest habitat differences in nest occupancy by ants or termites were noted. On three occasions, the presence of large, black, biting ants ( Pacvcondyla sp.) prevented the opening of nests. In many other nests, colonies of fire ants ( Solenopsis sp.) made nest opening a difficult process. The presence of these noxious ants may play a role in predator deterrence. False Nests Recently constructed nest mounds that did not contain any eggs were classified as false nests. Of the overall sample of 261 nests, 18 (6.9%) were false nests. False nests probably resulted from a variety of causes. If nest sites were disturbed prior to oviposition (e.g. by potential predators or by flooding) females would abandon the site and begin nest construction at another site. However, in these cases the nest could usually be distinguished as a nest under construction. If nests were depredated early in the incubation period, females would typically rebuild the nest, even if all the eggs had been consumed by a predator. Most

PAGE 309

300 "false" nests appear to have been rebuilt following predation. On one occasion I removed all the eggs from a nest for artificial incubation, and collected all the nest material as well. The nest site was revisited 14 days later and the mound had been completely rebuilt, without eggs. The entire nest mound was collected twice more, and each time the female reconstructed the nest mound. By the third time the rebuilt mound was rather small, apparently from the lack of nesting material. Nest Habitat Five distinct macrohabitat types were classified for the purpose of nesting habitat analysis: sandhills, palm savanna, spoil banks (dikes or roadsides) , marshes, and forests (see habitat descriptions; chapter 2) . Of 2 62 nests from the Hato Masaguaral area, the largest numbers of nests were found in the palm savanna (35.5%) and sandhill (30.2%) habitats (Fig. 10-1) . Using the nests of radio-tagged females (N=13) as an unbiased sample of nest macrohabitat utilization, suggests the utilization of sandhill habitats is less (15.4%), and nesting in forests more common (23.1%) than the overall nest sample would indicate. The two distributions are significantly different (chi-square; p<0.001). The sandhill nests may be more easily found because they are located along the easily traversed linear sandhill-marsh ecotone. In other habitats, nests were

PAGE 310

Figure 10-1. Percent of nests in each of five habitat types for all nests found (N=262) and nests produced by radio tracked females (N=13) .

PAGE 311

302 spread out in a three-dimensional space and were more difficult to sample. Within habitat types, the actual location of nests was closely related to microtopographical features of the environment. In all habitats, nests were situated on elevated sites above the water level. Nests under construction that were flooded prior to oviposition were abandoned by females. Important factors for choosing nest sites were elevated microtopography , the presence of sufficient material for nest mound construction, and frequently the presence of a nearby aquatic refuge (e.g., pool, stream) for the female. However, the last factor was not a rigid requirement. Females preferentially selected small, elevated, isolated nesting sites, presumably to reduce the likelihood of nest predation. Many potential nesting areas such as sandhills and linear dikes were heavily used by feral pigs (Sus scrofa) and were not preferred nesting sites. Most nests were situated at the base of, or adjacent to, trees or large shrubs (>2 m high). Overall, 51.1% of the nests were located immediately adjacent to the base of a tree or large shrub and another 9 . 1% were found within 1 m of one. Only 28.1% of the nests were located more than 4 m from a tree, and 21.7% of these were under recumbent Mimosa shrubs, principally along the edge of sandhills. An intriguing finding is that for a sample of 121 nests located adjacent to a tree or a large shrub, the majority of the

PAGE 312

303 nests were located on the west side of the tree (Fig. 10-2), but the significance of this is unclear. Table 10-4. Percent of nests in each of seven distance categories relative to the nearest tree or large shrub (>2 m high) by habitat type. Nests classified as >4m Herbs were more than 4 m from the nearest tree/ shrub but in a dense herbaceous growth; >4 m Mimosa were nests in dense growths of the recumbent shrub Mimosa piara. Habitat M Distance Cateaorv (%) At 4m >4m >4m Base Open Mimosa Herbs Sandhill 67 22.4 5.6 6.0 7.5 6.0 52.2 0.0 Palm Savanna 83 84.3 6.0 1.2 2.4 1.2 4.8 0.0 Spoil Bank 21 19.0 9.5 23.8 4.8 0.0 28.6 14.3 Marsh 20 40.0 15.0 5.0 0.0 20.0 15.0 5.0 Forest 30 53.3 20.0 10.0 13.3 0.0 0.0 3.3 Overall 221 51.1 9.1 6.3 5.4 4.1 21.7 2.3 Palm savanna nests The principal nesting habitat for caiman on Hato Masaguaral was palm savanna. Nests in this habitat were typically located on small, raised microtopographical features, the vast majority being at the base of palm (or associated) trees on sites I refer to as palm "islands" (core-tree matas of Troth 1979) . Palms ( Copernicia tectorum ) growing in the open savanna would freguently form the nucleus of a small group of trees or shrubs with the most common associates being stranger figs Ficus sp. , and various thorny shrubs or small trees, especially Zanthoxvlum culantrillo. Randia venezuelensis f and Annona sp.. Around the base of the palm was a small, elevated site where the

PAGE 314

305 associated trees and shrubs rooted, and where caiman nested. Of a sample of 93 palm savanna nests, 70 (75.3%) were located at the base of a tree and only 10% were found more than 4 m from a tree or shrub (Table 10-4) . Of the nests found at the base of a tree, 48.6% were at the base of a palm, 17.2% adjacent to a palm and strangler fig, and 14.3% near a palm and other tree (Table 10-5) . Within the palm savanna habitat, some preference was shown for nesting near localized depressions that would form shallow pools (< 1 m deep) and serve as a refuge for females attending the nest. As the pools gradually dried up during the late wet season some females would make shallow depressions or wallows within the drying pools. Nevertheless, it was not unusual for the surrounding savanna to dry completely prior to the end of egg incubation, forcing the female to remain far from the nest (Thorbjarnarson, personal observation) , or even abandon it altogether. Sandhill nests Sandhill nests were located along the ecotone between the raised sandhill dune habitat and flooded marshes; or along the edges of isolated ponds surrounded by sandhill. Nests were preferentially situated in areas where there was a relatively steep physical gradient between the sandhill and the wetland habitats. These areas were often dominated by dense growths of the spiny recumbent Mimosa shrub, or by small isolated tree patches. The latter areas were

PAGE 315

306 especially favored by nesting caiman. However, owing to the scarcity of tree patches along the ecotone, only 15 of 79 nests (19.0%) were situated at the base of a tree or large shrub. A wide variety of tree species was utilized in this respect, most commonly Cocoloba caracasana (26.7%) and Guazuma tomentosa (13.3%; Table 10-5). Nevertheless, sandhill nests were most commonly found greater than 4 m from a tree and in the Mimosa sandhill fringe (44.3% of sandhill nests, Table 10-4). Forest nests Forest nests, like those in palm savanna, were frequently adjacent to a tree on an elevated microtopographical feature. Seventeen of 40 nests (42.5%) were found at the base of a tree and an additional six (15.0%) were located within one meter of one (Table 10-4). Palms (23.5%), palms in association with other trees (11.8%), and Ficus sp. (17.7%) were the trees most frequently associated with nests (Table 10-5) . Spoil bank nests These nests were similar to sandhill nests in that they occurred in a narrow fringe along a linear elevated feature, in this case dikes or roadsides. As in sandhill habitats, spoil bank nests were less closely associated with tree bases than in palm savanna, forest or marsh habitats (Table

PAGE 316

307 « Table 10-5. Relative frequencies (%) of tree and large shrub species associations with caiman nests for three different habitat types. Species Relative Species Relative Frequency Frequency Palm Savanna (N=70 nests) Copernicia tectorum C. tectorum/Ficus sp. C. tectorum/ other sp. Zanthoxvlum culantrillo Randia venezuelensis 48. 6 Cocoloba caracasana 2. 9 17. 1 Ficus so. 1. 4 14. 3 Vernonia brasiliana 1. 4 8. 6 Pithecellobium saman 1. 4 2. 9 Annona sp. 1. 4 Sandhill fN=l5) Cocoloba caracasana 26.7 Zanthoxvlum culantrillo 7 .7 Guazuma tomentosa 13.3 Randia venezuelensis 7 .7 Cordia collococca 7.7 Dead Tree 7 .7 Caesaria mollis 7.7 Unknown sp. 13 .3 Annona iahnii 7.7 Forest (N=17) Copernicia tectorum 23.5 Ficus sp. 17.7 C. tectorum and other sp. 11.8 Pithecellobium quachpele 11.8 Dead Tree n.8 Randia venezuelensis 11.8 Zantoxvlum culantrillo 11.8 Unknown sp. 17.7

PAGE 317

308 10-4). Only 4 of 24 (16.7%) spoil bank nests were found at the base of a tree, and this may reflect in part the availability of raised sites not directly associated with tree bases. Of the nests located more than 4 m from a tree or shrub more than 2 m high, all were located either in Mimosa (25.0% of the nests) or in dense herbaceous growth (12.5%). Because spoil banks were frequently found adjacent to marshes these nesting sites also provided good retreat sites for females attending nests. Marsh nests Marsh nests were the least commonly encountered and this may reflect the limited number of nesting sites in the deeply flooded marsh habitats. Typically nests in these areas were located on small raised "islands" within the marshes, but one site used in three consecutive years (19868) was on the base of a reclining tree ( Ruprectia hamanii) . Of 21 marsh nests, eight (38.1%) were situated at the base of a tree, and another three (14.3%) were within one meter of a tree or shrub. Trees utilized included Ruprectia hamanii (three nests) ; Cordia collococca (two nests) ; dead trees (two nests) and a palm/fig (one nest) . Distanc e From Permanent Water The distance from a nest to permanent water reflects the minimum distance a female has to travel from a dry season lagoon to a nest site. Conversely, it also indicates

PAGE 318

309 the minimum distance a female must return at the end of the wet season, and if the nest has been successful, over what distance she must eventually lead her pod of young. From a sample of 168 nests, 75 (44.6%) were found within 0.5 km of permanent water (i.e. a water body that does not dry up in the dry season), and another 56 (33.3%) were found within 1 km (Table 10-6) . However, few were located at the very edge of the larger permanent water bodies such as the Guacimos lagoon. Data from radio-tracked females (principally from Guacimos and the Piscina) present a similar picture (Table 10-7) with 10 of 13 nests (76.9%) located within 2 km of permanent water. Nevertheless some females do nest considerably further away, with 5.65 km being the greatest recorded distance. Most nests encountered in the Guacimos region were found to the south and west of that lagoon. That region appeared to contain the best nesting habitat, as well as more marshes, and was the preferred direction of dispersal of all radio-tracked caiman from the Guacimos lagoon (see chapter 4). Large areas to the north and east of the lagoon either did not flood very deeply, or flooded deeply but were open savannas without trees. Conversely there was little directionality of nests in the Piscina region. Nest Site Reuse As was expected, the frequency of nest site reuse was an increasing function of the number of previous nesting

PAGE 319

310 Table 10-6. Distance of nests (percent, in 0.5 km sections) from permanent bodies of water. Distance measured in a straight line from the nest site to the center of the closest permanent body of water (Guacimos, Piscina, Alta Nest Distance Guacimos rlSCilla nil iicaLs (km) (N=73) / XT — C.C. \ \ N — Do ) \Lt — lUU f 0.0-0.5 16.44 57.58 44.64 0.5-1.0 39.73 36.36 33.33 1.0-1.5 17.81 4.55 9.52 1.5-2.0 13.70 0.00 6.55 2.0-2.5 0.00 0.00 0.00 2.5-3.0 4.11 0.00 1.79 3.0-3.5 1.37 1.52 0.60 3.5-4.0 0.00 0.00 0.60 4.0-4.5 1.37 0.00 0.60 4.5-5.0 2.74 0.00 1.19 5.0-5.5 1.37 0. 00 0.60 5.5-6.0 1.37 0.00 0.60 Table 10-7. Distance of nests (percent, in 0.5 km sections) from permanent bodies of water. Distance measured as in Table 10-6. Data from radio tagged females only (1987Nest Distance (km) Guacimos (N=8) Piscina (N=3) All Nests (N=13) 0.0-0.5 12.50 0.00 7.69 0.5-1.0 25.00 66.67 38.46 1.0-1.5 25.00 0.00 15.38 1.5-2.0 12.50 0.00 15.38 2.0-2.5 0.00 0.00 0.00 2.5-3.0 0.00 0.00 0.00 3.0-3.5 0.00 0.00 0.00 3.5-4.0 0.00 33.33 7.69 4.0-4.0 0.00 0.00 0.00 4.5-5.0 0.00 0.00 0.00 5.0-5.5 12.50 0.00 7.69 5.5-6.0 12.50 0.00 7.69

PAGE 320

311 seasons studied (1984-0% reuse; 1985-8.9%; 1986-5.4%; 198713.3%; 1988-20.5%), so an absolute estimate of nest site reuse cannot be made. However, the comparison of relative values of nest site reuse among habitat types is revealing. A high nest reuse figure suggests a relatively small number of suitable or preferred nest sites, whereas a low figure indicates a larger number of suitable nesting sites. The highest overall nest reuse figure was obtained for marsh habitats (22.7% of nest sites having been used in previous years; Table 10-8) . This is consistent with the idea of a limited availability of nest sites in marsh habitats and the preferred use of marsh sites by females because of the low nest predation rates (see Nest Fate below) . Nest site reuse in sandhill habitats was also quite high (20.3%). Females nesting along the sandhill-marsh edge preferred areas of steep physical incline along the ecotone, especially where there were small groups of trees and the high reuse figure here reflects the tendency of nests to be placed in the same few sites year after year. However, the sandhill habitats offered a great deal of apparently suitable nesting habitat that was not used and so the high reuse value indicates a limited number of highly "desirable" nest sites rather than a small number of suitable sites. The low nest reuse figures for palm savanna and forest habitats reflected the large number of potential nesting sites available to females in those habitats. The intermediate value for spoil bank habitats was due to the

PAGE 321

312 small number of nests in the sample (22) and the reuse of one particular nest site. Table 10-8. Nest site reuse in 5 different habitat types; 1985-1988. Number and percentage of nests on sites used in previous years. . Habitat Tvoe No. Nests Site Reuse % Reused Sandhill 79 16 20.3 Palm Savanna 93 5 5.4 Spoil Bank 24 3 12.5 Marsh 22 5 22.7 Forest 37 2 5.1 Overall 262 31 11.8 Few data were obtained on nest site reuse by individual females. Only one radio-tracked female nested in consecutive years, and she nested in two sites approximately 250 m apart. Conversely, on rare occasions two females may oviposit in the same nest, as happened in one spoil bank nest in 1988. However, the information from the radiotelemetry study (chapter 4) indicated that adult caiman returned to the same areas during consecutive wet seasons, so the potential exists for individual female to reuse nest sites. Some idea of nest site reuse by individual females can be obtained by comparing egg and clutch characteristics at reused nest sites between years. Clutch size and clutch mass are all correlated with female size (see chapter 9). Similar egg and clutch size parameters at a reused nest site do not necessarily indicate that the same female has reused the site. However, if egg and clutch size are very

PAGE 322

313 different, it implies that different females were involved. Table 10-9 presents comparative clutch and egg data for reused nest sites. It is likely that in comparison A and F, and the 1987 nests in comparisons C and H, different females had nested. So in at least 4 of 9 (44.4%) reused nest sites it appears that different females have nested. This is a very conservative figure. Table 10-9. Comparison of clutch and egg parameters of nests on the same nest site in different years. ^ X U. LU11 Size Mass (g) A. 10/07/84-1 30 2160 08/03/88-1 17 1130 B. 08/06/87-1 23 1445 09/04/88-1 29 1798 C. 08/31/86-1 28 2002 09/06/87-1 14 789 09/11/88-2 25 1858 D. 09/09/86-5 21 1252 08/04/87-1 23 1357 E. 08/24/87-1 30 1855 09/18/88-1 27 1804 F. 07/31/87-3 14 810 08/17/88-3 21 1195 G. 07/31/87-4 29 2011 08/21/88-1 28 1652 H. 08/27/85-1 27 1408 08/10/87-1 16 896 08/25/88-3 30 1710 I. 08/21/87-1 28 1829 09/11/88-3 28 2063

PAGE 323

314 Nest Fate Nest fate bv year During the 1984-1988 nesting seasons the fate of 247 nests was determined. Excluding false nests, and nests that were collected for artificial incubation, a sample of 199 nests remained for analysis (Table 10-10) . Nest fate categories were defined as: depredated (all eggs depredated) , partially depredated (nest depredated but some eggs hatched) , flooded (all eggs drowned) , and hatched. Hatching success averaged 52.8%, and did not vary significantly from year-to-year (Kruskal-Wallis ; p>0.05). A total of 54.8% of the nests produced some live hatchlings (hatched + partially depredated) . Most nests that failed were depredated (mean 39.2%). Flooding only accounted for 6.0% of the failed nests. All partially depredated nests were opened by Tupinambis lizards (see Nest Predators) . However, these estimates of nest depredation rates may be conservative. Many of the false nests (nest mounds without eggs) may be nests rebuilt following early depredation (see False Nests , above) . If all the false nests were depredated, then the overall depredation rate would be 44.2%, and the hatching rate would be lowered to 48.4%. Nest fate by habitat type Nest fate varied significantly among habitat types (Kruskal-Wallis, p<0.05; Table 10-11). Nest predation rates

PAGE 324

315 Table 10-10. Nest fate by year. Year Depredated Partially Depredated Flooded Hatched N 1984 6 40.0% 0 0.0% 1 6.7% 8 53.3% 15 1985 9 28.1% 0 0.0% 3 9.4% 20 62.5% 32 1986 18 40.9% 0 0.0% 1 2.3% 25 56.8% 44 1987 10 31.3% 2 6.3% 4 12.5% 16 50.0% 32 1988 35 46.1% 2 2.6% 3 4.0% 36 47.4% 76 Total 78 39.2% 4 2.0% 12 6.0% 105 52.8% 199 Table 10-11. Nest fate by habitat type. Habitat Depredated Partially Depredated Flooded Hatched N Sandhill 22 34.4% 4 6.3% 8 12.5% 30 46.9 64 Palm Savanna29 42.7% 0 0.0% 1 1.5% 38 55.9 68 Spoil Bank 7 50.0% 0 0.0% 1 7.1% 6 42.9 14 Marsh 1 5.3% 0 0.0% 2 10.5% 16 84.2 19 Forest 19 55.9% 0 0.0% 0 0.0% 15 44.1 34 Total 78 39.2% 4 2.0% 12 6.0% 105 52.8 199 were high in both forest and spoil bank habitats (55.9% and 50.0%, respectively). Nest depredation was somewhat lower in the palm savanna (42.7%) and sandhill (34.4%, and 6.3% partially depredated) habitats. Nests in marsh habitats experienced by far the lowest rates of predation (5.3%). Although they had a relatively high incidence of flooding (10.5%), the hatching rate of marsh nests was much higher than for any other habitat type. Sandhill nests also had a high rate of flooding (12.5%). However, in 6 of the 8 cases, sandhill nest flooding was directly related to artificially high water

PAGE 325

316 levels created by damming water drainages for cattle management. Excluding these nests, loss due to flooding would have only accounted for 3.1% of all sandhill nests. Nests in all habitats were typically placed on raised microtopographical features, usually surrounded to some degree by water. Locating nests on these small "islands" presumably makes them more difficult for predators to locate. From a sample of 64 nests where the amount of surface flooding 5 m around the nest was recorded (when the nest was first located) , increased flooding was associated with decreased depredation (Fig. 10-3) . Nest Predators The three principal predators on caiman eggs were crabeating foxes ( Cerdocyon thous) , tegu lizards ( Tupinambis texeguin ) and feral pigs (Sus scrofa) . Each predator had a distinctive style of opening a caiman nest. Tegus would tunnel into the nest, leaving a circular entrance hole approximately 10 cm in diameter. Eggs were pulled out individually and eaten. Frequently, tegus would not consume all the eggs during the first visit, but returned periodically over a period of several weeks before completely depredating the nest. In some cases the female caiman would remake the nest mound, filling the tegu's entrance tunnel. If this happened, and the tegu failed to return, the nest was considered to be partially depredated (see previous section) . Tegus were found throughout the

PAGE 326

317 Percent Nest Depredation 60 ( 0-20 20-40 40-60 60-80 80-100 Percent 5 m Radius Flooded Figure 10-3 . The rate of nest depredation as a function of the degree of surface flooding in a 5 m radius around the nest.

PAGE 327

318 ranch, but were especially common in the sandhill -marsh ecotone, where most sandhill nests were located. Tegu predation on nests occurred throughout the nesting season but peaked early in the incubation period (Fig. 10-4) . Foxes opened nests by digging with their forelimbs, usually leaving a distinctively v-shaped excavation. Foxes were the most common nest predators in the palm savanna habitat (Table 10-12) and typically preyed on nests late in the incubation period (Fig. 10-4) when savanna water levels were dropping. Feral pigs preyed on nests throughout the incubation period (Fig. 10-4) . Pigs would usually completely destroy the nest mound, in some cases making it difficult to locate the former nest site. Human depredation at Masaguaral was relatively minor and occurred principally along the highway (two spoil bank nests) . Table 10-12. Principal nest predators on caiman nests on Hato Masaguaral divided by habitat type. Habitat Tvoe Predator Fox Pig Tegu Human Unknown N % N % N % N % N % Sandhill 6 27.3 4 18.2 11 50.0 0 0.0 1 4.6 Palm Sav. 11 37.9 8 27.6 7 24.1 0 0.0 3 10.3 Spoil Bank 1 14.3 1 14.3 3 42.9 2 28.6 0 0.0 Marsh 1 100.0 0 0.0 0 0.0 0 0.0 0 0.0 Forest 8 42.1 5 26.3 6 34.6 0 0.0 0 0.0 Total 27 34.6 18 23.1 27 34.6 2 2.6 4 5.1

PAGE 328

Nests Depredated 61 1-10 11-20 21-30 31-40 41-50 51-60 61-70 71-80 81-90 Days After Oviposition Crab Eating Fox K§3 Feral Pig CU Tegu Figure 10-4. The timing of nest depredation (days after oviposition) for the three principal nest predators on Hato Masaguaral.

PAGE 329

320 Nest Attendance Based on captures and size estimates of caiman observed at nest sites, no male caiman were ever found near a nest and nest attendance was assumed to be an exclusively female behavior. In order not to disturb nesting females, extensive efforts to determine if an adult was present at the nest site were not made. As a result, females were noted as being present at only 16.8% of the nests on the initial visit (N=202) . Females would frequently be hidden, either submerged in a nearby pool or in vegetation, so this figure undoubtedly underestimates the true value. One radio-tagged female (#60; 80.1 cm SVL) was located every two hours over a 24 hour period on two occasions during the 1988 nesting season (3-4 October, 9-10 November) . This female had nested on a spoil bank located at the north end of the Piscina Lagoon, and followed a similar pattern of behavior on both occasions. Most of the day was spent in the Piscina Lagoon, 70-150 m south of the nest, in relatively deep water (ca. 1 m) . Shortly after nightfall, the female moved north through shallow water habitat, and passed through a culvert adjacent to the nest. In October, the female spent most of the night in a shallow canal located 20 m from the nest. It is uncertain if she actually visited the nest itself. The following morning between 04:00 h and 06:00 h, she returned to the main pool of the Piscina Lagoon. In November, the female repeated the same pattern of movements, but instead of moving into the canal,

PAGE 330

321 she remained out of the water within 2 m of the nest from 00:00 h to between 04:00 h and 06:00 h. The attendance at nests by female caiman presumably deters egg predators. However, no significant difference in predation rate was found between nests where the female was seen on the initial nest visit (37.0% depredated), and nests where no female was visible (45.8% depredated). Timing of Nesting Caiman oviposition on Hato Masaguaral peaked during mid-August (Table 10-13; Fig. 10-5). During the period 1985-1988, the mean date of oviposition varied only nine days (12-21 August) . Mean nesting date was earliest in 1987 and latest in 1985, but no significant difference was found among years (F3 t i20 = 1-92; p>0.05). Hatching peaked during the first half of November (Table 10-13, Fig. 10-5), although in 1987 the date of mean hatching was 24 October, significantly earlier than in any other year (F4 f 95=8 . 64 ; p<0.001; LSD; p<0.01). This was in part due to an earlier mean oviposition date, and a slightly shorter mean incubation period (see below) . No significant differences in mean date of oviposition or mean date of hatching were found among habitat types. Caiman nesting takes place during the peak of the rainy season, when water levels and savanna flooding were at their highest (Fig. 10-6) . Although no significant differences were found in mean date of oviposition during the period

PAGE 331

c c • fO CO > CO (0 cr> W H I •o m c CO (0 CT\ H c o •H «H A U u o) -p (0 X! Q) TJ -P c o o -P T) •H C W H O 0) > P -P 03 ftj a u in i o H 2 •H

PAGE 332

323

PAGE 333

200 150 100 50 Water Depth (cm) PPT (mm) Water Depth Precipitation CourUhlp/matlng Ovlpoiltlon Hitching Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month Figure 10-6. Timing of major reproductive events with respect to rainfall and water levels.

PAGE 334

325 1985-8, mean oviposition dates were somewhat earlier in years when flooding began early (1987, 1986; Fig. 10-6). Table 10-13. Mean, standard deviation and range of dates of caiman oviposition and hatching on Hato Masaguaral; 19851988. Year N Mean Date SD Rancre (days) Ovioosition 1985 10 21 August 14 . 9 1 August-25 September 1986 15 15 August 13.5 28 July-10 September 1987 38 12 August 12.4 18 July-1 September 1988 61 17 August 13.7 17 July-24 September Hatchina 1984 5 7 November 13.6 18 October-24 November 1985 14 10 November 8.2 22 October-2 3 November 1986 16 5 November 11.0 22 October-5 December 1987 17 24 October 10.7 2 October13 November 1988 34 15 November 11.0 24 October-8 December Nest Temperature Regime Measurements of nest temperature were made on a total of 32 nests. Nineteen of these nests were monitored for periods of up to 35 hours to examine diurnal variation in egg cavity temperature. Nests in 1984 were monitored over a 35-hour period (N=7) , whereas nests in 1987 and 1988 were monitored over a 24-hour period (N=12) . Diurnal nest temperature variation The egg cavity temperature of 12 nests was monitored in 1987 (N=5) and 1988 (N=7) over a 24-hour period both early (September) and late (October) in the incubation period.

PAGE 335

Temperature in the egg chamber followed a diurnal cycle which, due to the thermal inertia of the nest mound, was out of synchrony with diurnal air temperature variation and resulted in peak nest temperatures being reached late at night (Fig. 10-7a) or in the evening (Fig. 10-7b) . However, the daily air temperature regime had an important effect on nest temperature, and two basic patterns were noted. When air temperature exceeded nest temperature for a significant portion of the day, nest temperature continued to rise and peaked at night (20:00-22:00 h; Fig. 10-8a,b) . However, in 1988 nests were monitored on days when air temperature exceeded nest temperature for only short periods (Fig. 109a, b) and under these conditions egg cavity temperature peaked earlier (16:00-20:00 h) . Minimum nest temperatures on all days were encountered during the morning hours (10:00-12:00 h) . Rainfall caused nests to cool faster. Little or no rainfall was recorded during the nest temperature monitoring period on three of four days during the 1987/8 seasons. However, rain at 20:00 h on 28 September 1988 (Fig. 10-7b, lower curve) caused an increase in the rate of heat loss, and nest temperature declined more rapidly during the night. Temperature was monitored in seven nests in October 1984 (Fig. 10-10) during a period of heavy rainfall (61.5 mm rain over a two-day period) . Mean October 1984 nest temperature was significantly lower than the mean 1988 October temperature readings (LSD; p<0.01: see following section for

PAGE 336

Nest Temperature 6:00 10:00 14:00 18:00 Hour 22:00 2:00 6:00 Nest Temperature Figure 10-7. Mean 24 hour nest temperature variation for two days in 1987 and 1988.

PAGE 337

Temperature (C) 16 October 1987 6:00 10:00 14:00 18:00 22:00 2:00 6:00 Hour Temperature (C) 6:00 10:00 14:00 18:00 22:00 2:00 6:00 Hour Egg Cavity — I— Air Figure 10-8. Mean 24 hour nest and air temperature variation in 1987.

PAGE 338

329 Temperature (C) 6:00 10:00 14:00 18:00 22:00 2:00 6:00 Hour Temperature (C) 6:00 10:00 14:00 18:00 22:00 2:00 6:00 Hour Egg Cavity — ' — Air Figure 10-9. Mean 24 hour nest and air temperature variation in 1988.

PAGE 339

Temperature 34.0 1 0 1 1 i i i i i i 8:00 12:00 16:00 20:00 2:00 6:00 10:00 14:00 18:00 Hour Figure 10-10. Mean 34 hour nest and air temperature variation in 1984.

PAGE 340

331 account of high October 1988 nest temperatures) . However, neither mean October 1987 nest temperature nor 1987-8 24hour temperature range was significantly different from October 1984 readings. Although there were few quantitative differences in mean nest temperature or temperature range during the rainy period of 1984, there was a difference in the timing of the diurnal temperature cycle. Egg cavity temperature in 1984 was more tightly coupled with air temperature, peaking during the daylight hours and reaching a minimum at night (Fig. 10-10) . Mean nest temperatures for individual nests (>24 hour period) ranged from 29.0°C to 35.5°C (Table 10-14). The overall mean temperature recorded was 32.2°C (17 nests, N=350) . Minimum and maximum recorded egg cavity temperatures were 21. I'd and 36.5°C. Temperature variability over a 2 4 -hour period ranged from a maximum of 4.9°C to a minimum of 0.4°C, with a mean value of 1.69'C. Season al nest temperature variation Egg cavity temperature was monitored at weekly intervals for a sample of 12 nests in both 1987 and 1988. Because data were collected at various times throughout the day, all egg cavity temperatures were corrected to a mean value based on the average 24-hour temperature variation for all 1987/1988 nests. This was done by calculating an hourly conversion factor from the combined temperature data set. Nest temperature increased throughout the incubation period

PAGE 341

332 Table 10-14. Mean egg cavity temperatures and temperature (°C) ranges for individual nests over a 24 hour period. Temperature data collected at two hour intervals. Nest Date Mean Minimum Maximum N 1984 Nests 10/6/84-1 10/16/84 31.80 31.1 32.9 9 10/7/84-1 10/16/84 32.46 32.0 33.3 9 10/11/84-1 10/16/84 32.50 32.0 32.9 9 10/6/84-2 10/16/84 29.58 29.0 30.0 9 10/11/84-5 10/16/84 32.43 31.8 33.1 9 10/10/84-7 10/16/84 30.83 30.0 31.7 9 10/10/84-8 10/16/84 29.02 27.7 30.8 9 10/11/84-7 10/16/84 29.73 28.9 31.3 9 1987 Nests 8/21/87-3 9/8/87 32.52 31.7 33.7 13 10/16/87 29 . 70 29 . 0 30.4 13 8/21/87-4 9/8/87 32.37 32.0 32.9 13 10/16/87 32.42 33.1 34.0 13 8/21/87-5 9/8/87 31.42 31.2 31.8 13 10/16/87 32.55 32.3 33.0 13 8/24/87-1 9/8/87 30.12 29.1 31.8 13 10/16/87 32.72 32.1 33.5 13 8/30/87-1 9/8/87 31.89 30.9 32.8 13 10/16/87 30.35 29.1 31.2 13 1988 Nests 8/25/88-1 9/28/88 32.31 30.4 35.3 13 10/26/88 34.05 33.0 34.9 12 8/25/88-2 9/28/88 32.50 31.5 33.8 13 10/26/88 34.30 32.8 8/25/88-3 9/28/88 32.79 32.2 33.3 13 8/30/88-1 10/26/88 34.88 34.1 35.7 12 9/28/88 30.11 28.7 31.9 13 10/26/88 30.97 29.5 32.5 12 9/11/88-1 9/28/88 32.65 32.4 32.8 13 10/26/88 34.56 34.2 34.9 12 9/11/88-2 9/28/88 33.68 33.2 34.2 13 10/26/88 35.50 34.6 36.5 12 9/11/88-3 9/28/88 33.23 32.9 33.5 13 10/26/88 34.28 33.8 34.8 12

PAGE 342

333 in both years (Fig. 10-11) . This increase appeared to be related to three factors: a slight increase in ambient temperature from August through October (Table 10-15) , a decrease in rainfall, and metabolic heat generated by the developing embryos. The midpoint of monthly mean maximum and minimum air temperatures increases approximately 1°C from August through November (1987-1.34°; 1988-0.47°; combined-0.91° ; Table 10-15), which would account for some 75% of the observed increase in mean egg cavity temperature over the same period: 1987-1.4°; 1988-1.0° (Table 10-16). Factors influencing nest temperature Mean egg cavity temperature varied significantly among habitats (F 4fl9 =9.44; p<0 . 001) (Table 10-17). Nests in sandhill, marsh and forest habitats had significantly higher mean temperatures than did palm savanna or spoil bank nests. Data on mean egg cavity temperature were highly correlated with differences in mean incubation time among habitats (Fig. 10-12). The "hot" nests from marsh, sandhill and forest areas had relatively short incubation times (mean 7880 days) , whereas the "cool" palm savanna and spoil bank nests had longer incubation periods (85-86 days) . No significant relationship between egg depth or nest mound volume and mean clutch temperature was found. However, the range of twenty-four hour clutch temperatures tended to decrease both with increasing egg depth (Fig. 1013), and with increasing nest mound volume (Fig. 10-14),

PAGE 343

334 34.0 33.0 32.0 31.0 30.0 29.0 Mean Nest Temperature Rainfall (mm) 200 150 100 23412341234 August September October 12 3 4 November 34.0 33.0 32.0 31.0 30.0 29.0 Mean Nest Temperature Rainfall (mm) 200 150 100 23412341234 1234 August September October November ~~ — Mean Nest Temp. Rainfall Figure 10-11. Seasonal trends in nest temperature and rainfall for 1987 and 1988.

PAGE 344

335 Table 10-15. Mean minimum and maximum shade air temperatures (°C) and rainfall on Hato Masaguaral during the 1987 and 1988 nesting seasons. Year Aug Sept Oct Nov 1987 Min. 21.11 22.02 22.04 21.80 Max. 32.47 33.12 33.73 34.45 Rain (mm) 260.9 222.7 381.9 80.9 1988 Min. 21.43 22.09 21.76 21.41 Max. 32.04 32.38 33.66 33.00 Rain (mm) 419.5 196.5 102.0 94.0 Table 10-16. Mean egg cavity temperatures by month. Sample size in parentheses. Year August September October November 1987 1988 31.4 (23) 32.3 (51) 31.5 (33) 32.9 (18) 32.9 (42) 32.8 (2) 32.5 (14) Table 10-17. Mean egg cavity temperature by habitat type. Early temperatures are for nests prior to the 6th week postoviposition. Total represents the mean of all nests. Habitat Mean Temperature f'CJ Early Total Marsh Sandhill Forest Palm Savanna Spoil Bank 33.1 (7) 32.4 (37) 32.2 (9) 30.7 (16) 30.3 (7) 33.1 (15) 32.8 (99) 32.7 (18) 31.4 (40) 30.3 (21)

PAGE 345

Incubation (days) Spoil Bank + Palm Savanna + Sandhill + Marsh Forest + + i i i i 30 30.5 31 31.5 32 32.5 33 33.5 Mean Nest Temperature (C) Figure 10-12. Incubation time as a function of mean nest temperature for the five nesting habitat types.

PAGE 346

6. 0 5. 5 5. 0 4. 5 4. 0 3. 5 3. 0 2. 5 2. 0 i. 5 l. 0 0. 5 0. 0 Temperature Variation 0 ^4 cases , r.hdu, h . u^.t o lllllllllll s 0 " o o " — o O 0 o 8 8 o 10. Figure 10-13 15.81 21.38 27.19 Depth to Eggs (cm) 33. Twentyfour hour nest temperature variation as a function of depth to the top of the clutch. Figure 10-14. Twenty-four hour nest temperature variation as a function of nest volume.

PAGE 347

suggesting a thermal buffering effect for deeply buried eggs. Incubation The length of the egg incubation period (days between oviposition and nesting) was determined for 58 nests during this study. The overall mean was 81.8 days (SD=6.14, range=68-98) . The longest recorded incubation period (98 days) was for a nest that was not opened by the adult, and incubation time was prolonged by the fact that the young had to dig themselves out of the nest. Mean incubation period ranged from 79.2 days in 1987 to 83.4 days in 1988; however no significant differences among years were noted (F3 f 54=1. 72 ; p>0.05). Caiman nests in different habitat types did show a significant difference in incubation time (F4 t 53=3 . 39 ; p<0.05). The shortest mean incubation time was found in both marsh (78.7 days; SD=2.06; N=4) and forest (78.7 days; SD=7.06; N=7) habitats. Nests with the longest mean incubation intervals were in spoil bank (86.5 days; SD=2.88; N=6) and palm savanna habitats (85.2 days; SD=7.83; N=13) , and both were significantly longer (LSD; p<0.05) than for nests in forests and marshes. Sandhill nests (mean 80.4 days; SD=4.80; N=28) were intermediate in incubation length. In 16 nests, mean clutch temperature and incubation period were negatively correlated (r=-0.621; p=0.01; Fig. 10-15) . Differences in mean incubation period among

PAGE 348

Incubation (days) 16 cases, r=-.621, p=.Q10 I 1 1 I I I I I I I I I I M I I I I I I I I I I I I I 1 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 29.60 30.64 31.63 32.66 33.70 Mean Nest Temperature Figure 10-15. Incubation time as a function of mean nest temperature for individual nests.

PAGE 349

340 habitats were correlated with differences in measured nest temperatures (Fig. 10-12) . However, the reasons for these differences are unclear. No significant correlations were found between incubation length and any of the following parameters: nest volume (r=0.08), clutch mass (r=0.02), egg mass (r=0.13), egg depth (r=0.04), date of oviposit ion (r=0.15) or date of hatching (r=0.28). The degree of nest insolation may be an important factor, but was not quantified here. Another potentially important parameter is nest moisture. Very wet nests may delay incubation or even kill the embryos (nest flooding) . Conversely, very dry nests may be less thermally buffered and lack sufficient moisture for compost heat generation via bacterial decomposition of nest material. Discussion The mean hatching rate reported here was much higher than for other studies conducted in Venezuela, but comparable to results of studies in the Pantanal of Brazil. Staton and Dixon (1977) reported only 7.5% of the nests in Hato La Guanota (Apure State) hatched successfully. Also in Apure State, Ayarzaguena (1983) found a hatching rate of 20%. In both studies nest predation was the principal cause of egg mortality, with Tupinambis being the most important nest predator. Human nest depredation was also an significant factor in the study of Staton and Dixon (1977).

PAGE 350

341 In the Pantanal of Brazil, Crawshaw (1987) reported a mean nesting success rate of 36%, although this value varied significantly between two study sites. In the same area Cintra (1988) found a 55.3% hatching rate. In both studies depredation was the most important cause of egg mortality, and the principal nest predator was the coati (Nasua nasua) . This study is the only one to report significant nest depredation by crab-eating foxes ( Cerdocyon thous) and feral pigs (Sus scrofa) . Foxes preyed largely on palm savanna nests late in the incubation period. At this time savanna water levels were falling and fox utilization of the savanna habitat increased (Brady 1979) . There was some evidence that foxes were attracted to nests by the vocalizations of fully developed hatchlings waiting to be released from the nest (Thorbjarnarson, personal observation) . Unlike most areas in the llanos, feral pigs were extremely abundant in the study site, and would forage in groups in many of the same areas where caiman nested. However, pig predation on nests freguently appeared to be almost a chance event, for numerous unpredated nests were encountered with pig sign around them. Another significant finding was that the degree of surface flooding plays an important role in determining the potential of nest depredation, presumably because nests in extensively flooded areas are more difficult to locate. In this study, the highest hatching rate was found in marsh habitats, where nests were typically located on small

PAGE 351

342 elevations surrounded by a deeply flooded zone. These sites also had the highest nest reuse values, indicating that they are highly preferred by females. The use of small, raised, isolated features for nesting has also been reported by Ouboter and Nanhoe (1987). Crawshaw and Schaller (1980) found most nests in the cerrado habitat of their study site to be located on "tree-covered hillocks protruding above flooded terrain." Cintra (1988) also reported that most nests in his study were situated at the base of trees but does not indicate if the sites were elevated. The high freguency of nests located adjacent to tree trunks in certain habitats on Hato Masaguaral was clearly related to microtopographical features. This was particularly true in the palm savanna where the bases of palm trees provided the only raised features in an otherwise uniformly low-lying landscape. The area immediately surrounding the base of the palms typically consisted of a raised hummock upon which palm leaf debris accumulated. A wide variety of other trees and shrubs also germinated and grew on these sites, leading to the formation of distinctive clumps of woody vegetation surrounding the palm. These "palm islands" provided ideal conditions for nesting, not only because they were raised above the surrounding terrain, but also because they furnished large amounts of organic material for the construction of the nest. A somewhat similar situation was noted in the forest habitat where over half of all caiman nests were situated within 1 m of a tree.

PAGE 352

343 Egg laying is timed to coincide with peak annual water levels, thereby reducing the probability of nest flooding. Nevertheless, nest flooding was an important secondary cause of egg mortality in this (6.0%), and other studies (Staton and Dixon 1977-12.5%, Crawshaw 1987-13.7%, Cintra 198816.6%). Nest flooding in Masaguaral was caused by the inundation of the nest and egg cavity, asphyxiating the eggs. However, in some cases mortality appeared to have resulted from partial inundation of the nest, and may have been caused by low temperature stress, or a combination of temperature stress and reduced oxygen availability. Apart from causing outright mortality, sub-lethal levels of flooding may have significant effects on the embryonic development of caiman, and result in the lowered fitness of hatchling caiman (Joanen et al. 1987, Webb et al. 1987). Caiman nest temperatures were within the normal range for crocodilians (Magnusson et al. 1985), but increased throughout the incubation period. Seasonal variation in incubation temperature has been well documented among crocodilians (Chabreck 1973, Webb et al. 1983b, Lutz and Dunbar-Cooper 1984, Webb and Preston-Cooper 1989), but there are various interpretations regarding the underlying cause of this phenomenon. All studies to date, except that of Chabreck (1973), have found an increasing trend in nest temperature. Heat production from the decomposition of nest material has been hypothesized to play an important role in elevating nest temperatures (Chabreck 1973, Staton and Dixon

PAGE 353

344 1977, Magnusson 1979), but would presumably result in a decrease in nest temperature with time as microbial activity declines (Ferguson 1985) . Caiman eggs are incubated at the end of the wet season, during a period of generally diminishing rainfall, and increasing ambient temperature. Much of the increase in incubation temperature can be explained simply by the increase in mean air temperature, however rainfall patterns may also be important. Periods of heavy rainfall, which were more common early in incubation, may have an important cooling effect by reducing insolation, and inundating the bases of nests or partially flooding the egg cavity. The drying environmental conditions would also result in reduced nest cooling from evaporative water loss late in the incubation period. The role of embryonic metabolism in raising nest cavity temperature is somewhat controversial (Magnusson et al. 1985, Webb and Cooper-Preston 1989), but may be more important among species that lay large clutch masses (e.g., Crocodvlus porosus ) . If this is true the relatively small clutch mass of caiman would argue for a reduced role of embryo metabolism in raising egg cavity temperature. The reasons for different nest temperature regimes among habitat types remain unclear. The warmest nests were found in marsh habitats, and these were also most often found in exposed areas (20% in the open away from trees) . The cool nests in palm savannas and spoil bank habitats were frequently well shaded by trees and tall herbaceous plants,

PAGE 354

respectively. These observations suggest that insolation may play an important role in the difference among habitats. Another potential factor influencing egg cavity temperature is the material the nest is made from. Although there was no effect of nest volume on mean egg cavity temperature, differences in the thermal properties of nest materials and their susceptibility to decomposition may be important. For instance, palm petioles constituted a large fraction of the volume of palm savanna nest mounds, and were comprised of a very light, porous material very resistant to decomposition. Palm petioles and other woody vegetation incorporated into the nest would not provide a suitable substrate for composting and this may result in cooler nest temperatures. Differences in nest moisture may also influence the rate of nest material decomposition and nest temperature. Crocodilian nests are of two basic types: hole nests and mound nests (Greer 1970, Campbell 1972). Hole nests consist of an excavation, usually into a sandy soil, in which the eggs are deposited and then covered up. Mound nests are made principally from an accumulated mass of decomposing organic matter, shaped into a roughly hemispherical mound. Both types of nests provide a thermal buffering effect, as well as a humid environment for the developing clutch of eggs (Staton and Dixon 1975, Ferguson 1985). Greer (1970), following Schmidt (1924), suggested that the type of nest mound used by a species may be indicative of phylogenetic relationships. Campbell (1972)

PAGE 355

countered that nesting habits are ecologically malleable, and that some species (such as Crocodvlus acutus) are typically hole-nesters, but will construct elevated mound nests in low-lying areas prone to flooding. Nevertheless, the two types of nesting modes do indicate different nesting strategies among the Crocodylia. Not only is there a difference in the form of the nest, but there is also a good correlation between nesting mode and the seasonal timing of nesting (Ouboter and Nanhoe 1987) . Hole-nesting species typically oviposit during periods of dropping water levels, with hatching usually occurring near the beginning of the annual rainy period (Cott 1961, Webb et al. 1983). Mound nesters, on the other hand, characteristically nest during the midst of the annual rainy period (Ouboter and Nanhoe 1987) . The few exceptions to this rule are almost invariably mound nesting species that nest during the annual dry season (Cox 1985, Magnusson et al. 1985) , or species where temperature constrains the seasonal nesting period. Mound nests can also be found in a much wider variety of habitats than hole nests, which are typically made in seasonally exposed sandbars. Some typical hole nesting species (e.g. Crocodvlus acutus P C. rhomb ifer . C. moreletii ) have a tendency to produce mound nests in low-lying areas where seasonal exposure of sandbars does not occur (Campbell 1972, P. Moler, pers. comm.). Furthermore, in the absence of temperature constraints, the seasonal timing of hole nesting species appears to be more limited by environmental

PAGE 356

347 conditions than for mound nesters. Among populations of mound nesters, it is not unusual for nesting to take place over a 2-3 month period (Webb 1983, Cox 1984), whereas among hole nesters the breeding populations usually oviposit within a period of one month (Cott 1961, Webb et al. 1983e, Thorbjarnarson 1988) . Taken together, these observations suggest that mound nesting is a more adaptable strategy, allowing greater flexibility in where and when a female nests. The advantages of nesting during the wet season, for mound nesting species, have been discussed by various authors (Staton and Dixon 1977, Crawshaw and Schaller 1980, Ouboter and Nanhoe 1987) and include: 1) nesting material is more available during the wet season, 2) nests can be more dispersed, thus reducing nest depredation or intraspecif ic competition among neonates, 3) females can remain in flooded areas adjacent to the nest and deter egg predators, 4) nesting during the period of peak water levels reduces the probability of nest flooding, 5) the wet season environmental conditions favor embryo development, and 6) mound nests are more susceptible to desiccation and overheating, which are less likely to occur in the wet season. Despite the fact that they are typically wet season nesters, on a regional basis caiman do exhibit a certain amount of variability in the timing of their nesting. A review of available information on the timing of caiman

PAGE 357

348 nesting with respect to the annual precipitation cycle suggests nest timing can be divided into four categories: nesting during the middle (peak) of the wet season (N=14) , nesting during the late dry season or early wet season (N=9) , nesting during the late wet season (N=3), and nesting during the dry season (N=l) (Fig. 10-16, Table 10-18) . Throughout much of the range of the spectacled caiman, hydric fluctuations are not as extreme as in the llanos and suitable nesting conditions can be found throughout the year (Ouboter and Nanhoe 1987) . Under these more equitable environmental conditions caiman should have more latitude in the timing of nesting. Nevertheless, under these circumstances nesting is still seasonal, with most nests found in the late dry and early wet season (Fig. 10-16) . Nesting just prior to, or during the early wet season may serve to maximize the amount of time the neonates have in the wet season prior to the seasonal drop in water levels. However, nesting at the beginning of the wet season would also be expected to increase egg mortality due to flooding. This might suggest that hatchling survivorship, and not egg survivorship, is the overriding factor determining when nesting occurs in these areas. The relative importance of this seasonal nesting strategy for Caiman crocodilus may be underestimated in the sample due to the fact that the majority of caiman reproductive investigations have been done in seasonal savanna habitats.

PAGE 358

Figure 10-16. The timing of nesting with respect to seasonal peaks in rainfall for caiman populations.

PAGE 359

350 Rio Parnalba, Brazil (Barra do Corda) Putumayo/Caquata, Colombia (Iqultoa) Ba|o Caquata/Apaporla. Colombia (Iqultoa) Rio Napo/Paataza/Aguarlco, Eouador (Iquiloa) Frtnch Guiana (Cayanna) Guyana (Oaorgatown) Rio Araguala, Brazil (Conoaieayo do Araguaya) Iqultoa, Paru (Iqultoa) Manu National Park, Paru (Manu) C.C. crocodllua 1 1 L Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Wet Season Nesting

PAGE 360

El Mantaco. vtnizuili (El Mantaco) Ida Maraio, Brazil (8alara) Rupununl, Ouyana (Dadawana) Coaaawlina Rlvar. Surinam (Coaaawlina Rlvar) Trinidad (Port of Spain) Llanoa. Vanazuala (San Farnando da Apura) Colombia (Barlnaa) Orooua, Alia Mata. Colombia (Barlnaa) Puarto Carrano, Colombia (Puarto Ayacucho) C.c. crocodllus J I l Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec HI Wet Season Hi Nesting Figure 10-16 — continued.

PAGE 361

Chlapae. Mexico (Retalhuleu) C.c. apaporienala La Moaqultla, Honduraa (Puarts Cabezaa) Cano Negro, Coata Rica (La Cruz) \w//^//Mmmmmmi\ C.c. chlapaaiua C.c. fuscua Kit 8alaraanca, Colombia (Crlatobal) Lower Bani/Mamora. Bolivia (Santa Cruz) 18— Pantanal. Brazil (Corumba) mmmm H Paraguay (Aaoanalon) wmmmmsmsmtmmw*. Upper Rio Maraora, Bolivia (Sanla Cruz) C.c. yacare i i i i i i i Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Wet Season Wm Nesting Figure 10-16 — continued.

PAGE 362

353 Table 10-18. Four categories of nest timing with respect to the annual pattern of rainfall. Data from Figure 10-16. Location Source Middle Wet Season Llanos, Venezuela Staton and Dixon 1977 Ayarzagiiena 1983 This study Arauca/Casanare, Colombia Medem 1981 Alta Meta, Colombia Medem 1981 Puerto Carreno, Colombia Medem 1981 Chiapas, Mexico Alvarez del Toro 1974 La Mosauitia. Honduras Klein 1979 Caho Negro, Costa Rica Allsteadt 1988 Isla Salamanca, Colombia Medem 1981 Pantanal, Brazil Crawshaw and Schaller 1980 Putamayo/Cagueta, Colombia Medem 1981 Bajo Cagueta/Apaporis Medem 1981 Trinidad Medem 1981 Coeswijne River, Suriname Ouboter and Nanhoe 1987 El Manteco, Venezuela Gorzula 1978 Late Drv-Earlv Wet Season Rio Parnaiba, Brazil Medem 1983 Rio Araguaia, Brazil Medem 1983 Isla Marajo M<=>H cam 1 Qpi neuciu 1)70 J Iguitos region, Peru Dixon and Soini 1977 Rio Napo region, Ecuador Medem 1983 Manu , Peru Herron 1985 Lower Beni region, Bolivia ncuciii x z? o o Upper Beni/Mamore, Bolivia King and Videz-Roca 1989 Medem 1983 King and Videz-Roca 1989 Paraguay Medem 1983 Late Wet Season Rio Apaporis, Colombia Medem 1981 French Guiana Medem 1983 Lowland Guyana Medem 1983 Dry Season Rupununi region, Guyana Medem 1983

PAGE 363

354 The advantages of late-wet season and dry season nesting are unclear. In coastal Guyana and French Guiana, caiman appear to be nesting in the late wet season. It is possible that in these areas neonatal survival is not compromised by nesting late in the wet season, or that the potential for nest flooding is high throughout most of the wet season. The one data point for dry season caiman nesting comes from the Rupununi savanna region of Guyana (from a letter by Lee cited in Medem 1983) and is somewhat enigmatic. Obviously, further work on caiman reproductive cycles in a variety of these habitats needs to be done before we can speculate on the adaptive significance of these nesting strategies. Nevertheless, the advantages of nesting during the middle of the wet season for caiman in the llanos, and other extreme seasonal environments such as the Brazilian Pantanal (Crawshaw and Schaller 1980, Crawshaw 1987), are clear. In these habitats the timing of the reproductive cycle is tightly constrained by the annual pattern of rainfall and flooding (chapter 9). Acting through the survivorship of eggs and hatchlings, these rigid environmental conditions provide strong selection on nest timing. Nesting earlier in the wet season would reguire courtship and mating to take place in the crowded dry season lagoons, which is unfeasible based on behavioral considerations (chapter 8), and would not permit females to adjust their reproductive output in accordance with environmental conditions (e.g., reduced

PAGE 364

nesting in very dry years; chapter 9) . Earlier nesting would also result in the reduced availability of nesting material, and the increased probability of egg mortality due to flooding. Late wet season nests would be more vulnerable to overheating and depredation. Nesting later in the wet season would also result in a reduction in the habitat available for neonates, and little opportunity for feeding and growth prior to the onset of the stressful dry season.

PAGE 365

CHAPTER 11 SUMMARY AND CONCLUSIONS Many aspects of the behavior and ecology of the spectacled caiman in the Venezuelan llanos are shaped by the seasonal variation in water availability. On a yearly basis the llanos are transformed from a virtual inland sea into a near desert habitat. The annual reproductive cycle, patterns of movement and habitat use, and feeding ecology all vary in concert with the availability of wetlands habitats. Despite the rigorous environmental conditions, populations of caiman have managed to thrive in most parts of the llanos, and much of this success has been due to recent human activities. Accounts by early naturalists who visited the llanos would almost invariably comment on the presence of large numbers of Orinoco crocodiles ( Crocodvlus intermedius ) in the riverine habitats, but very little mention was made of spectacled caiman (Gumilla 1741, Humboldt 1860, Paez 1868, Appun 1871). Doubtless, this was in part related to the much more impressive nature of the crocodile, as well as its proclivity for eating people. However, the general impression regarding caiman populations was that, while present in savanna ponds, they were not very numerous. With the virtual extirpation of Orinoco 356

PAGE 366

357 crocodiles (Thorbjarnarson and Hernandez, 1989), and the creation of many new permanent water habitats associated with cattle management and road construction, caiman populations have greatly expanded. During the annual dry season caiman densities reach prodigious levels. In some borrow-pits on the ranch densities exceeded 1500 caiman/ha, and biomass surpassed 20,000 kg/ha. Other borrow-pits and windmill ponds, generally isolated and far from the natural llanos drainage systems, had much lower caiman densities and it was in these bodies of water that most of the hatchlings survived. During the study period the ranch population was declining. Over the five year period 1985-1989, total population size was reduced by 20%. Most of this decline occurred during the period 1985-1986, and consisted of a large loss of juvenile caiman. The reasons for this loss are not clear but were assumed to represent mortality. The size-class distribution of the caiman population at Hato Masaguaral was typical for llanos caiman populations. Approximately 60% of the population at the end of the dry season was composed of size-class III individuals (60-89.9 cm SVL) , containing adult females and subadult males. The low percentage of juveniles (size-class II; 20-59.9 cm SVL) was attributed to high mortality from predators and cannibalism during the dry season. Adult males (>90 cm SVL) comprised some 16% of the total population. The overall sex

PAGE 367

358 ratio of the caiman population did not differ significantly from a 1:1 ratio. Caiman growth came to a virtual halt during the dry season. Wet season growth, even among juveniles, was slow. Sexual dimorphism in growth rates became apparent when caiman reached a size of 50-60 cm SVL, and was caused by female growth slowing faster than male growth. The reproductive behavior of the spectacled caiman involved a complex communication system that uses body postures, vocal and non-vocal acoustic signals, tactile, and olfactory signals. Many of the observed behaviors were similar to ones reported for other species of crocodilians. Long distance communication was done by bellowing, an exclusively male behavior, which relied on the non-vocal production of infrasonic sounds. Caiman bellowed throughout the year, but bellowing exhibited a definite peak during the early wet season courtship period. A secondary peak early in the dry season was presumably related to the establishment of social hierarchies as caiman were forced to occupy the few permanent water habitats. In the dry season lagoons, male caiman maintained a dominance hierarchy based on largely on size. Signalling was done by bellowing, and by visual means using a graded series of body postures. Aggressive postures involved the display of increasing amounts of the caiman's body above the water's surface. Agonistic behaviors among males tended to increase throughout the dry season. Male caiman established

PAGE 368

359 territories in newly flooded habitats shortly after dispersal from the dry season lagoons (June-July) . When the annual rains were delayed, courtship activity was seen in the dry season lagoons as well. Courtship was only observed under the latter conditions, and was comprised of tactile behaviors, principally snout rubbing, followed by the male mounting the female. Caiman gonadal cycles mirrored the seasonal changes seen in reproductive behaviors, with gonadal recrudescence on both males and females beginning in the late dry season. Testicular mass peaks in May and June, just prior to courtship. Peak numbers of spermatozoa were observed in the testes from June through August. Among vitellogenic females, ovarian mass and follicular size increased throughout the early wet season, and peak at the beginning of the nesting period. Nesting was a wet season event with peak levels coinciding with maximum savanna water levels in mid-August. Nests consisted of a small mound (average 108 cm long x 88 cm wide x 40 cm high) constructed by females on elevated sites. Nest mounds were made from locally available organic matter, often incorporating soil. Nest temperatures varied daily and on a long term basis. Overall mean temperature was 32.2 °C, and the average 24 hour variability was 1.69*C. Nest temperatures also increased during the incubation period, and was related to increasing mean air temperature, a decrease in rainfall, and possibly the metabolic heat of

PAGE 369

360 the embryos. The mean length of incubation was 81.8 days, but varied significantly among habitat types and was inversely correlated with mean nest temperature. From a sample of 199 nests, the overall hatching rate was 52.8%, with an additional 2.0% of the nests being partially depredated but hatching at least some young. Nest losses due to flooding were only 6.0%, but 39.2% of the monitored nests were totally destroyed by predators, principally the lizard Tupinambis texeguin . crab-eating foxes, and feral pigs. Nest fate varied among habitats with marsh nests having the highest hatching rates (84.2%), which was primarily related to a very low rate of nest depredation (5.3%) . Females began reproducing at a size of approximately 60 cm SVL and an age of seven years, although some females did not nest until they are closer to 70 cm SVL. Mean size at first reproduction was 64.0 cm SVL. Mean clutch size was 22.2 (SD=5.57), and ranged from four to 36. Very small clutches may have been the result of partial nest depredation. Female fecundity increases with body size. Clutch size and clutch mass were significantly correlated with female size, but egg size was not. Breeding effort, defined as the percent of females over 60 cm SVL that nest in any one year, was highly correlated with body size but overall averaged 52.8%. Male caiman reached physiological sexual maturity at a size of approximately 75 cm SVL, although some individuals

PAGE 370

361 may mature at smaller sizes. However, because the social hierarchy was based largely on size, breeding may not occur until males are significantly larger. During the dry season, caiman typically remained in one body of water, and movements were restricted. Mean dry season home range was 11.3 ha, but for most individuals was less than 10 ha. During the course of the dry season some caiman moved overland between bodies of water, principally as a response to low water levels. Dispersal from the dry season bodies of water coincided with the flooding of the savannas in the rainy season. Due to late rains, dispersal in 1988 (mean date 30 June) was more than one month later than in 1987 (23 May) . No significant difference in the date of dispersal was found between males and females. The patterns of dispersal and movement varied widely among caiman. Three types of wet season home ranges were defined: 1) wet season and dry season home range identical, 2) wet season home range overlaps dry season home range, and 3) wet and dry season home range totally separate. The great majority of caiman dispersed to new areas during the wet season (home range types 2 and 3) . Caiman moved away from the dry season lagoons during the early wet season, and reached a peak dispersal distance from July to September. The longest straight-line dispersal distance measured was 6.5 km. Nevertheless, even during the peak period of dispersal, the majority of caiman did not move more than 1 km from their dry season habitat. During most wet season

PAGE 371

months 40-50% of the caiman were within 0.5 km of a permanent water lagoon. Dispersal distance was greater in 1987 than in 1988, and this was attributed to the late flooding in 1988. No difference was found in dispersal distance between males and females. Among adult males the greatest dispersal distance was among intermediate size individuals (90-100 cm SVL) . An index of daily movements indicated that the peak periods of movement occurred during the early and late wet season, and were associated with dispersal and return to the dry season lagoons. Of 23 caiman that were followed through an entire dry/ wet season cycle, 21 (91.3%) returned to the same lagoon. Wet season home ranges averaged 39.8 ha, and no significant differences in home range size were noted between the sexes. However, ontogenetic changes in home range size were found with smaller individuals of both sexes having larger home ranges. Adult caiman used the same wet season home ranges in consecutive years, with the exception of subadult males, whose home ranges appear to be less fixed on a year-to-year basis. Although non-nesting females tended to disperse farther than nesting females, no significant differences in home range size or daily movement index were noted. These results suggest that nesting does not alter female movement patterns to any great extent. Evidence from this study suggests that male caiman undergo a change in movement patterns with size, and that

PAGE 372

363 this shift is related to social factors. Small adult or large subadult males tend not to disperse long distances, have relatively small daily movement indexes, and have large type 2 home ranges. Intermediate sized males disperse long distances, move more freguently, and have large type 3 home ranges. Large males do not disperse long distances, move relatively little, and have small type 2 or 3 home ranges. This ontogenetic shift in movement patterns appears to be in part related to the fact that large males preferentially occupy the best breeding territories closest to the dry season lagoons where the density of females is the greatest. Wet season macrohabitat use differed significantly from the relative availability of habitat types. Caiman showed a strong preference for lagoon and borrow-pit habitats, whereas the extensive palm savanna habitat was under utilized. There was a great deal of variability in wet season macrohabitat selection, both among individuals and on a seasonal basis. No significant differences between males and females were noted. The late flooding of the savannas in 1988 shifted the pattern of habitat usage towards an increased use of permanent water sites. Males consistently selected deeper water habitats than did females, and this was not simply related to differences in body size. Microhabitat selection was highly influenced by macrohabitat type, but suggested that caiman were mostly found in open water areas or among light growth of herbaceous vegetation. Use of microhabitat types varied

PAGE 373

364 throughout the course of the wet season, to a large extent reflecting the changing savanna environment. Use of open water habitat was greatest in the early wet season, prior to the growth of extensive herbaceous vegetation. Caiman found among woody vegetation peaked in the middle of the wet season when flooding was most extensive. In the late wet season, caiman would frequently use dense colonial patches of Thalia, or be located under mats of floating vegetation. There was a tendency for males to be found in more open habitats than females. Caiman feeding ecology also demonstrated significant seasonal and ontogenetic shifts. Pomacea snails, fish, and freshwater crabs were the most important dietary items on Hato Masaguaral. Prey items consumed principally during the dry season were fish, reptiles and mammals. Pomacea . crabs, amphibians, and birds were most frequently taken in the wet season. Juvenile caiman fed largely on Pomacea . insects (principally Coleoptera) , and crabs. With increasing size, vertebrate prey became increasingly important and fish, crabs, mammals and reptiles were the most important prey of caiman over 90 cm SVL. Caiman lack the morphological specializations of some other crocodilians (i.e. a long, narrow snout) for catching fast moving aquatic prey, and instead relied on behavioral specializations. For catching fish, caiman utilized a number of behaviors including float fishing, trapping, and weir fishing, all of which increased the probability of prey

PAGE 374

365 capture. Nevertheless, the success rate of feeding attempts was very low, averaging 7.2%. One enigmatic feeding behavior consisted of jumping partially out of the water, which presumably served to disorient prey. However, the feeding success rate of this energetically costly behavior is extremely low. The results of this study, and others conducted on the spectacled caiman in the Venezuelan llanos, have detailed many aspects of this species' life history. We now have relatively good information on basic aspects of the reproductive ecology, feeding ecology, and seasonal patterns of movements for this species in the hyperseasonal llanos habitat. Nevertheless, there still is a great deal of work that remains before we can profess to have a good understanding of the caiman's population ecology. Very little information is available on such important lifehistory traits as age-specific mortality rates, or reproductive lifespan, and no data whatsoever address the question of density dependent factors and how these influence fecundity and mortality schedules. Currently, work is underway to produce a life table for the caiman population at Hato Masaguaral, which will permit us to employ population simulations as a tool for investigation of population dynamics. Nevertheless, this model will only be for one population at one point in time, and extrapolations to other areas will be tenuous at best. The amount of work needed for an adequate understanding of a species population

PAGE 375

ecology is truly daunting, particularly for large, longlived species such as crocodilians. However, I believe that study of the spectacled caiman in the Venezuelan llanos provides us with one of the best opportunities for undertaking this task. Caiman are relatively small, easily captured and censused, and occur in a wide variety of densities from steady state (such as on Hato Masaguaral) , to areas where they have been depleted from hunting. Apart from this, the presence of a large caiman harvest program provides biologists with the opportunity to collect large amounts of data, and carry out manipulative experiments that cannot be done with endangered species. Unfortunately, there has been little effort made to take advantage of this situation. The Venezuelan government has not established a research program to accompany the caiman management program, and even the monitoring of exploited populations is inadequately done. Also, universities have been slow to capitalize on the opportunities provided by this harvest. Research programs need to be established in a number of areas throughout the llanos, not only for the study of caiman population ecology, but also to monitor the effects of the harvest program on wild caiman populations. Nevertheless, the design of the current harvest program, based on the harvest of caiman over 90 cm SVL, is a good one. The offtake consists almost entirely of males, and very few reproductive females are taken, providing the

PAGE 376

lower legal size limits are adequately enforced. The results of this study indicate that males reach sexual maturity at sizes below 90 cm SVL, but are prevented from breeding by larger males. Additionally, the presumed polygynous breeding system would suggest that there exists a surplus of males in the population. The removal of a portion of the male population over 90 cm SVL should stimulate population growth by reducing population density, and do so without affecting the important reproductive segment of the population, the adult females. Recently, the management program has been expanded to permit caiman ranching, or the collection of eggs from wild nests for artificial rearing. This program will provide biologists with many new opportunities for research into a number of areas of caiman husbandry, as well as investigations of reproductive ecology. One of the main requisites for land owners participating in the caiman harvest program is that they improve caiman habitat on their land. This usually takes the form of creating more permanent water habitat such as borrow-pits or tapas (dammed streams) . The results of this study suggest that the best way to do this would be to create a large number of small ponds, rather than a few larger ones. This would increase hatchling and juvenile survivorship, the factors that most limit llanos caiman populations.

PAGE 377

368 Venezuela has taken the lead among Latin American countries in developing a model management program based on sustained yield harvesting. The design of the program is sound, and has much promise for application in other developing countries. However, the link between research and management has been left out. The Venezuelan government needs to sponsor a strong research program on caiman biology, either by employing government biologists, or by funding university or other non-governmental organizations who are willing to carry out such work. The caiman program has been generating a great deal of revenue, part of which should be turned back into research as it is in other countries with good crocodilian programs. With a strong commitment to investigation, the Venezuelan caiman management program can become a true model program for sustained yield conservation programs in developing countries.

PAGE 378

LITERATURE CITED Allsteadt, J. 198. Nesting ecology of the Central American caiman, Caiman crocodilus fuscus, in northern Costa Rica. Unpubl. Abstract. 1 p. Alvarez del Toro, M. 1969. Breeding the spectacled caiman Caiman crocodylus at Tuxtla Gutierrez zoo. Int. Zoo Yb. 9:35-36. Alvarez del Toro, M. 1974. Los Crocodylia de Mexico. Instituto Mexicano de Recursos Naturales Renovables, Mexico D.F. 70 p. Andrews, H. 1989. An unusual record of Crocodylus moreletii nesting. Hamadryad 14:11-13. Appun, K.F. 1871. Unter den Tropen. Wanderungen durch Venezuela, am Orinoco, durch Britisch Guiana und am Amazonenstrome in den jahren 1849-1868. Ed. Hermann Costenoble, Jena. 519 p. Ayarzagviena, J.S. 1983. Ecologia del caiman de anteojos o baba ( Caiman crocodilus L.) en los Llanos de Apure (Venezuela). Donana 10:7-136. Bartholomew, G.A. 1982. Physiological control of body temperature, pp. 407-424 in: C. Gans and F.H. Pough (eds.) Biology of the Reptilia. Vol. 12. Academic Press, New York. Bayliss, P.W. , g.J.W. Webb, P.J. Whitehead, K. Dempsey and A.M. A. Smith. 1986. Estimating the abundance of saltwater crocodiles, Crocodylus porosus Schneider, in tidal wetlands of the Northern Territory: a mark recapture experiment to correct spotlight counts to absolute numbers, and a calibration of helicopter and spotlight counts. Aust. Wildl. Res. 13:309-320. 369

PAGE 379

370 Bellairs, A. 1969. The Life of Reptiles. 2 Vols. Weidenfield and Nicholson, London. 590 p. Bennett, A.F., and W.R. Dawson. 1972. Aerobic and anaerobic metabolism during activity in the lizard Dipsosaurus dorsalis . J. Comp. Physiol. 81:289-299. Bennett, A.F. 1982. The energetics of reptilian activity, pp. 155-200 in: C. Gans and F.H. Pough (eds.). The Biology of the Reptilia. Vol.13. Academic Press, New York. Berroteran, J.L. 1985. Geomorfologia de un area de llanos bajos centrales Venezolanos. Bol. Soc. Venez. Cienc. Nat. 143:31-77. Brady, C.A. 1979. Observations on the behavior and ecology of the crab-eating fox ( Cerdocyon thous ) . pp. 161-171 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics . Smithsonian Institution, Washington D.C. Brattstrom, B.H. 1965. Body temperatures of reptiles. Amer. Midi. Nat. 73:376-422. Brockelman, W.Y. 1975. Competition, the fitness of offspring, and optimal clutch size. Am. Nat. 109:677-699. Campbell, H.W. 1972. Ecological or phylogenetic interpretations of crocodilian nesting habits. Nature 238:404-405. Campbell, H.W. 1973. Observations on the acoustic behavior of crocodilians. Zoologica 58:1-11. Carbonneau, D.A., and R.H. Chabreck. In press. Population size, composition and recruitment of American alligators in freshwater marsh. Proc. 10th Working Meeting of the Crocodile Specialist Group-IUCN-The World Conservation Union. Gland, Switzerland. Chabreck, R.H. 1966. Methods of determining the size and composition of alligator populations in Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 20:105-112.

PAGE 380

371 Chabreck, R.H. 1971. The foods and feeding habits of alligators from fresh and saline environments in Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 25:117-124. Chabreck, R.H. 1973. Temperature variations in nests of the American alligator. Herpetologica 29:48-51. Chabreck, R.H. , and T. Joanen. 1979. Growth rates of American alligators in Louisiana. Herpetologica 35:5157. Cintra, R. 1988. Nesting ecology of the Paraguayan caiman ( Caiman crocodilus yacare ) in the Barazilian Pantanal. J. Herpetol. 22:219-222. Congdon, J.D. and J.W. Gibbons. 1985. Egg components and reproductive characteristics of turtles: relationships to body size. Herptologica 41:194-205. Congdon, J.D. and J.W. Gibbons. 1987. Morphological constraint on egg size: a challenge to optimal egg size theory? Proc. Natl. Acad. Sci. USA. 84:4145-4147. Congdon, J.D. and J.W. Gibbons. 1990. The evolution of turtle life histories, pp. 45-56 in: J.W. Gibbons (ed.). Life History and Ecology of the Slider Turtle. Smithsonian Institution. Washington D.C. Cott, H.B. 1961. Scientific results of an inquiry into the ecology and economic status of the Nile crocodile (Crocodvlus niloticus l in Uganda and Northern Rhodesia. Trans. Zool. Soc. London 29:211-358. Coulson, R.A. and T. Hernandez. 1983. Alligator metabolism: studies on chemical reactions in vivo. Comp. Biochem. Physiol. B. 74:1-182. Cowles, R.B. and CM. Bogart. 1944. A preliminary study of the thermal requirements of desert reptiles. Bull. Am. Mus. Nat. Hist. 83:261-296.

PAGE 381

372 Cox, J. 1985. Crocodile nesting ecology in Papua New Guinea. Field Document No. 5. Assistance to the crocodile skin industry. PNG Dept of Primary Industry/FAO Publication. 149 p. Crawshaw, P.G. , Jr. 1987. Nesting ecology of the Paraguayan caiman ( Caiman vacare ) in the Pantanal of Mato Grosso, Brazil. Unpublished MS Thesis. University of Florida, Gainesville, FL. 69 p. Crawshaw, P.G. , Jr., and G. Schaller. 1980. Nesting of the Paraguayan caiman ( Caiman vacare ) in Brazil. Pap. Avulsos de Zoologia 33:283-292. Cuellar, 0. 1970. Egg transport in lizards. J. Morph. 119:7-20. Cuellar, 0. 1984. Reproduction in a parthenogenetic lizard: with a discussion of optimal clutch size and a critique of the clutch weight/body weight ratio. Amer. Midi. Nat. 111:242-258. Davenport, J.D., J. Grove, J. Cannon, T.R. Ellis and R. Stables. 1990. Food capture, appetite, digestion rate and efficiency in hatchling and juvenile Crocodvlus porosus . J. Zool., Lond. 220:569-592. Deeming, D.C., and M.W.J. Ferguson. 1988. Environmental regulation of sex determination in reptiles. Phil. Trans. Roy. Soc. London (B) 322:19-39. Deitz, D.C. and T.C. Hines. 1980. Alligator nesting in north-central Florida. Copeia 1980:249-258. Delany, M.F., and C.L. Abercrombie. 1986. American alligator food habits in northcentral Florida. J. Wildl. Manage. 50:348-353. Deraniyagala, P.E.P. 1939. The tetrapod reptiles of Ceylon Vol. I. Testudinates and crocodilians. Colombo Museum of Natural History. Colombo. 412 p. Derickson, W.K. 1976. Lipid storage and utilization in reptiles. Amer. Zool. 16:711-723.

PAGE 382

373 Dixon, J.R., and P. Soini. 1977. The reptiles of the upper Amazon basin, Iquitos region, Peru. Part II. Crocodilians, turtles, snakes. Milwaukee Pub. Mus. Contrib. Biol. Geol. 12:1-91. Duval, D.L., L.J. Guillette Jr., and R.E. Jones. 1982. Environmental control of reptilian reproductive cycles, pp. 201-231 in: C. Gans and F.H. Pough (eds.). Biology of the Reptilia. Vol. 13. Academic Press, New York. Eddy, W. F. 1977. A New Convex Hull Algorithm for Planer Sets from ACM. Transactions on Mathematical Software 3:398-403. Eisenberg, J.F., M.A. O'Connel and P.V. August. 1979. Density, productivity, and distribution of mammals in two Venezuelan habitats, pp. 187-210 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution, Washington D.C. Ellis, T.M. 1980. Caiman crocodilus : an established exotic in south Florida. Copeia 1980:152-154. Elsey, R.M. , T. Joanen, L. McNease and V. Lance. 1990. Stress and plasma corticosterone levels in the American alligator-relationships with stocking densities and nesting success. Comp. Biochem. Physiol. 95A: 55-63. Erben, H.K. 1970. Ultrastrukturen und mineralisation rezenter und fossiler Eischalen bei Vogeln un Reptilien. Biomineralisation 1:1-66. Ewel, J.J., a. Madriz and J. A. Tossi, Jr.. 1976. Zonas de Vida de Venezuela. Ministerio de Agricultura y Cria, Caracas. Ferguson, M.W.J. 1985. The reproductive biology and embryology of the crocodilians. pp. 329-491 in: C. Gans, F.S. Billet and P.F.A. Maderson (eds.). Biology of the Reptilia. Vol. 14. John Wiley and Sons, New York. Ferguson, M.W.J. , and T. Joanen. 1982. Temperature of egg incubation determines sex in Alligator mississippien g-i ^ Nature 296:850-853.

PAGE 383

374 Fitzgerald, L.A. 1988. Dietary patterns of Caiman crocodilus in the Venezuelan llanos. Unpublished MS Thesis. The University of New Mexico, Albuquerque. 74 p. Fleishman, L.J., H.C. Howland, M.J. Howland, A.S. Rand and M.L. Davenport. 1988. Crocodiles don't focus underwater. J. Comp. Physiol. A. Sens. Neural Behav. Physiol. 163:441-443. Fleishman, L.J., and A.S. Rand. 1989. Caiman crocodilus does not require vision for underwater prey capture. J. Herpetol. 23:296. Fogarty, M. J. , and J.D. Albury. 1967. Late summer foods of young aligators in Florida. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 21:220-222. Ford, N.B., and R.A. Seigel. 1989. Relationships among body size, clutch size, and egg size in three species of oviparous snakes. Herpetologica 45:75-83. Fuller, M.K. 1981. Characteristics of an American alligator ( Alligator mississippiensis ) population in the vicinity of Lake Ellis Simon, North Carolina. Unpublished MS Thesis. North Carolina State University, Raleigh. 136 p. Garnett, S.T. 1985. The consequences of slow chitin digestion on crocodilian diet analysis. J. Herpetol. 19:303-304. Garrick, L. D. 1975. Structure and pattern of the roars of Chinese alligators ( Alligator sinensis Fauvel) . Herpetologica 31:26-31. Garrick L.D., and J.W. Lang. 1977. Social signals and behaviours of adult alligators and crocodiles. Amer. Zool. 17:225-39. Garrick, L. , J. Lang and H.A. Herzog. 1978. Social signals of adult American alligators. Bull. Amer. Mus. Nat. Hist. 160:155-192.

PAGE 384

375 Goodwin, T. , and W.R. Marion. 1979. Seasonal activity ranges and habitat preferences of adult alligators in a northcentral Florida lake. J. Herpetology 13:157-164. Gorzula, S. 1978. An ecological study of Caiman crocodilus crocodilus inhabiting savanna lagoons in the Venezuelan Guyana. Oecologia 35:21-34. Gorzula, S. 1985. Are caimans always in distress? Biotropica 17:343-44. Gorzula, S. 1987. The management of crocodilians in Venezuela, pp. 91-101 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd, Chipping Norton, Australia. Gorzula, S., and A.E. Seijas. 1989. The common caiman, pp. 44-61 in: Crocodiles: Their Ecology, Management and Conservation. A Special Publication of the Crocodile Specialist Group of the Species Survival Commission of the International Union for Conservation of Nature and Natural Resources. IUCN Publ. N.S. Gland, Switzerland. Graham, A. 1968. The Lake Rudolf crocodile ( Crocodylus niloticus Laurenti) population. Unpubl. report to the Kenya Game Department. 145 p. Greer, A.E. 1970. Evolutionary and systematic significance of crocodilian nesting habits. Nature 227:523-524. Greer, A.E. 1975. Clutch size in crocodilians. J. Herpetol. 9:319-322. Groombridge, B. 1982. The IUCN Amphibia-Reptila Red Data Book. IUCN. Gland, Switzerland. 426 p. Gumilla, J.S.J. 1741. El Orinoco Ilustrado. Historia Natural y Geografia de este Gran Rio y sus Caudalosas Vertientes. Manuel Fernandez, Madrid.

PAGE 385

376 Hawkins, A.D. and A. A. Myrberg Jr. 1983. Hearing and sound communication underwater, pp. 347-405 in: B. Lewis (ed.). Bioacoustics: A comparative approach. Academic Press, New York. Heatwole, H. 1976. Reptile Ecology. University of Queensland Press, St. Lucia, Queensland. 178 p. Herron, J.C. 1985. Population status, spatial relations, growth, and injuries in black and spectacled caimans in Cocha Casu. Unpublished Honors Thesis. Princeton University, Princeton. Herzog, H.A. 1974. The vocal communication system and related behaviors of the American alligator ( Alligator mississippiensis ) and other crocodilians. Unpubl. Masters Thesis. University of Tennessee, Knoxville. Herzog, H.A. and G.M. Burghardt. 1977. Vocalization in juvenile crocodilians. Z. Tierpsychol. 44:294-304. Huey, R.B. 1982. Temperature, physiology, and the ecology of reptiles, pp 25-91 in: C. Gans and F.H. Pough (eds.). Biology of the Reptilia. Vol 12. Academic Press, New York. Humboldt, A.B. 1860. Reise in die Aquinoctial-Gegenden des Neuen Continents. J.C. Cottascher Verlag, Stuttgart. 444 P. Hunt, R.H. 1969. Breeding of spectacled caiman Caiman c. crocodylus at Atlanta Zoo. Int. Zoo Yb. 9:36-37. Hutton, J.M. 1984. Population ecology of the Nile crocodile Crocodvlus niloticus Laurent i, 1768, at Ngezi, Zimbabwe. Unpublished D.Phil Thesis. University of Zimbabwe, Harare. 501 p. Hutton, J.M. 1987. Incubation temperatures, sex ratios and sex determination in a population of Nile crocodiles ( Crocodvlus niloticus l . J. Zool., Lond. 211:143-155.

PAGE 386

377 Hutton, J.M. 1989. Movement, home range, dispersal and separation of size classes in Nile crocodiles. Am. Zool. 29:1033-1050. Hutton, J.M., and M.E.J. Woolhouse. 1989. Mark-recapture to assess factors affecting the proportion of a Nile crocodile population seen during spotlight counts at Ngezi, Zimbabwe, and the use of spotlight counts to monitor crocodile abundance. J. Appl. Ecol. 26:381-395. Iverson, J.B. 1982. Biomass in turtle populations: a neglected subject. Oecologia 55:69-76. Jackson, D. 1988. Reproductive strategies of sympatric freshwater emydid turtles in northern peninsular Florida, Bull. Florida State Museum. Biol. Sci. 33:113-158. Jackson, J.F., H.W. Campbell and K.E. Campbell. 1974. The feeding habits of crocodilians: validity of the evidence from stomach contents. J. Herpteol. 8:378-381. Jacobsen, T. , and J. A. Kushlan. 1989. Growth dynamics in the American alligator ( Alligator mississippiensis l . J. Zool., Lond. 219:309-328. Joanen, T. 1969. Nesting ecology of alligators in Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 23:141151. Joanen, T. , and L. McNease. 1970. A telemetric study of nesting female alligators on Rockefeller Refuge, Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 24:175-193. Joanen, T. , and L. McNease. 1971. Propagation of the American alligator in captivity. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 25:106-116. Joanen, T., and. L. McNease. 1972. A telemetric study of adult male alligators on Rockefeller Refuge, Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 26:252 275.

PAGE 387

378 Joanen, T. , and L. McNease. 1975. Notes on the reproductive biology and captive propagation of the American alligator. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 29:407415. Joanen, T. , and L. McNease. 1979. Time of egg deposition for the American alligator. Proc. Ann. Conf. Southeast. Assoc. Fish. Wildl. Agencies. 33:15-19. Joanen, T. , and L. McNease. 1980. Reproductive biology of the American alligator in southwest Louisiana, pp. 153-160 in: Murphy, J. B. , and J.T. Collins (eds.). Reproductive biology and diseases of captive reptiles. Soc. Study Rept. Amphib. Lawrence, Kansas. Joanen, T. , and L. McNease. 1987. The management of alligators in Louisiana, USA. pp. 33-42 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd. , Chipping Norton, Australia. Joanen, T., and L. McNease. 1989. Ecology and physiology of nesting and early development of the American alligator. Am. Zool. 29:987-998. Joanen, T. , L. McNease and M.W. Ferguson. 1987. The effects of egg incubation temperature on post-hatching growth of American alligators, pp. 533-537 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia. Johnsen, P.B., and J.L. Wellington. 1982. Detection of glandular secretions by yearling alligators. Copeia 1982:705-708. Kellog, R. 1929. The habits and economic importance of alligators. Tech. Bull. USDA No. 147. King, F.W., and P. Brazaitis. 1971. Species identification of commercial crocodilian skins. Zoologica 56:15-70. King, F.W., and R.L. Burke. 1989. Crocodilian, Tuatara, and Turtle Species of the World. A Taxonomic and Geographic Reference. Assoc. Syst. Collections. Washington D.C.

PAGE 388

379 King, F.W., and Videz Roca, D.H. 1989. The caimans of Bolivia: A preliminary report on a CITES and Centro de Desarrollo Forestal sponsored survey of species distribution and status, pp. 128-155 in: Crocodiles. Proceedings of the 8th working meeting IUCN/SSC Crocodile Specialist Group. IUCN Publ. N.S. Gland, Switzerland. Kinler, N. , D. Taylor and G. Linscombe. 1987. 1986 experimental alligator harvest program on Marsh Island Refuge. Report in files of La. Dept. Wildl. Fisheries, New Iberia, La. 27 p. Klein, E.H. 1979. Los crocodylia de Honduras: su biologia y estado actual con recomendaciones para su manejo. Dir. Gen. Recursos Nat. Renov. , Tegucigalpa, Honduras. Krause, S.E. 1983. Reproductive characteristics of the American alligator ( Alligator mississippiensis ) in North Carolina. Unpublished MS Thesis. North Carolina State University, Raleigh. 85 p. Kushlan, J. A., and T. Jacobsen. 1990. Environmental variability and the reproductive success of Everglades alligators. J. Herpetol. 24:176-184. Lance, V.A. 1987. Hormonal control of reproduction in crocodilians. pp. 409-415 in: G.J.W. Webb, S.C. Manolis and P.J. whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia. Lance, V.A. 1989. Reproductive cycle of the American alligator. Amer. Zool. 29:999-1018. Lang, J.W. 1976. Amphibious behaviour of Alligator mississ ippiensis . Roles of circadian rhythm and light. Science 191:575-577. Lang, J. 1987a. Crocodilian behavior: implications for management, pp. 273-294 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia.

PAGE 389

380 Lang, J. 1987b. Crocodilian thermal selection, pp. 301-317 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia. Legler, J.M. 1958. Extra-uternine migration of ova in turtles. Herpetologica 14:49-52. Loveridge, J. P. 1984. Thermoregulation in the Nile crocodile, Crocodvlus niloticus . Symp. Zool. Soc. Lond. 52:443-467. Lowe-McConnell, R.H. 1975. Fish Communities in Tropical Tropical Freshwaters. Longmann, New York. 3 37 p. Lutz, P., and A. Dunbar-Cooper. 1984. The nest environment of the American crocodile ( Crocodvlus acutus) . Copeia. 1984:153-161. Magnusson, W.E., and J. A. Taylor. 1981. Growth of juvenile Crocodylus porosus as affected by season of hatching. J. Herpetol. 15:242-245. Magnusson, W.E. 1982. Biological aspects of the conservation of Amazonian crocodilians. pp. 108-116 in: Crocodiles. Proc. of the 5th working meeting IUCN/SSC Crocodile Specialist Group. IUCN Publ. N.S. Gland, Switzerland. Magnusson, W.E., A. P. Lima and R.A. Sampaio. 1985. Sources of heat for nests from Paleosuchus trigonatus and a review of crocodilian nest temperatures. J. Herpetol. 19:199-207. Magnusson, W.E., E.V. Da Silva and A. P. Lima. 1987. Diets of amazonian crocodilians. J. Herpetol. 21:85-95. Marcellini, D.L. 1979. Activity patterns and densities of Venezuela caiman ( Caiman crocodilus ) and pond turtles (Podocnemis vocrli ) . pp. 263-271 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution, Washington D.C. Mayhew, W.W. 1964. Photperiodic responses in three species of the lizard genus Uma . Herpetologica 20:95-113.

PAGE 390

381 Mayhew, W.W. 1966a. Reproduction in the arenicolous lizard, Uma notata . Ecology 47:9-18. Mayhew, W.W. 1966b. Reproduction in the psamophilous lizard, Uma scoparia . Copeia 1966:114-122. Mazzotti, F. 1983. The ecology of Crocodylus acutus in Florida. Ph.D. Dissertation. Pennsylvania State University. 161 p. Mcllhenny, E.A. 1935. The alligator's life history. The Christopher Publishing House, Boston. 117 p. McNease, L. and T. Joanen. 1974. A study of immature alligators on the Rockefeller Refuge, Louisiana. Proc. Ann. Conf. Southeast. Assoc. Game Fish Comm. 28:495-507. McNease, L. and T. Joanen. 1977. Alligator diets in relation to marsh salinity. Proc. Ann. Conf. Southeast. Assoc. Fish Wild. Agencies 31:36-40. Medem, F. 1981. Los Crocodylia de Sur America. Vol. I. Los Crocodylia de Colombia. Universidad Nacional de Colombia, Bogota, Colombia. 354 p. Medem, F. 1983. Los Crocodylia de Sur America. Vol. II. Universidad Nacional de Colombia, Bogota, Colombia. 270 p. Messel, H. and G.C. Vorlicek 1989. Growth of Crocodvlus porosus in the wild in northern Australia, pp. 110-137 in: Crocodiles. Proc. of the 9th working meeting IUCN/SSC Crocodile Specialist Group. IUCN Publ. N.S. Gland, Switzerland. Messel, H. , G.C. Vorlicek, A.G. Wells and W.J. Green. 1981. Surveys of tidal river systems in Northern Australia and their crocodile populations. Monograph 1. The BlythCadell Rivers system study and the status of Crocodvlus porosus in the tidal waterways of northern Australia. Pergamon Press, New York. 463 p. Modha, M.L. 1967. The ecology of the Nile crocodile (Crocodylus niloticus Laurenti) on Central Island, Lake Rudolf. E. Afr. Wildl. J. 5:74-95.

PAGE 391

382 Neill, W.T. 1971. The Last of the Ruling Reptiles: Alligators, Crocodiles and Their Kin. Columbia University Press, New York. 486 p. Ouboter, P.E., and L.M.R. Nanhoe 1984. Ecological study of Caiman crocodilus in Northern Surinam. Unpubl . Report , Department Animal Ecology, Catholic University of Nij Megen, Netherlands. 65 p. Ouboter, P.E., and L.M.R. Nanhoe. 1987. Notes on nesting and parental care in Caiman crocodilus crocodilus in northern Suriname and an analysis of crocodilian nesting habits. Amphibia-Reptilia 8:331-348. Ouboter, P.E., and L. Nanhoe. 1988. Habitat selection and migration of Caiman crocodilus crocodilus in a swamp and swamp forest-forest habitat in Northern Suriname. J. Herpetol. 22:283-294. Paez, R. 1868. Wild Scenes in South America, or Life in the Llanos of Venezuela. Charles Scribner, New York. 502 p. Palmer, B.D. and L.J. Guillette, Jr. 1988. Histology and functional morphology of the female reproductive tract of the tortoise Gopherus polyphenols . Am. J. Anat. 183:200211. Peaker, M. 1969. Active acquisition of stomach stones in a specimen of Alligator mississippiensis . Brit. J. Herpetol. 4:103-104 Pittier, H. 1942. La Mesa de Guanipa. Ensayo de Fitogeographia. Assoc. para la Protec. de la Naturaleza, Caracas. 46 p. Pooley, A.C. 1982. Discoveries of a Crocodile Man. W. Collins and Sons, London. 213 p. Pough, F.H. 1969. Environmental adaptations in the blood of lizards. Comp. Biochem. Physiol. 31:885-901. Pough, F.H. 1980. The advantages of ectothermy for tetrapods. Amer. Nat. 115:92-112.

PAGE 392

383 Ramos, S. 1975. Evaluacion calorimetrica a nivel indiividual en Caiman crocodilus (Linneaus 1758) de la poblacion localizada en el modulo experimental de Mantecal. Unpubl. Ph.D Dissertation. Universidad Central de Venezuela, Caracas, Venezuela. 88 p. Reese, A.M. 1921. The structure and development of the integumental glands of the crocodilia. J. Morph. 35:581-611. Rivero Blanco, C. 1974. Habitos reproductivos de la baba en los llanos de Venezuela. Natura 52:24-29. Rodda, G.H. 1984. Movements of juvenile American crocodiles in Gatun Lake, Panama. Herpetologica 40:444-451. Rudran, R. 1979. The demography and social mobility of a red howler (Alouatta seniculus ) population in Venezuela, pp. 107-126 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution, Washington D.C. Sarmiento, G. 1983. The savannas of tropical America, pp. 245-288 in: F. Bourliere (ed.). Ecosystems of the World. Vol. 13. Tropical Savannas. Elsevier Scientific Publishing Co., New York. Schaller, G.B. and P.G. Crawshaw. 1982. Fishing behaviour of the Paraguayan caiman. Copeia 1982:66-72. Schleich, H.H. and W. Kastle. 1988. Reptile Egg-Shells SEM Atlas. Gustav Fischer Verlag, New York. 123 p. Schmidt, K. P. 1922. The American alligator. Field Mus. Nat. Hist. Leaflet 3:25-38. Schmidt, K. 1924. Notes on Central American crocodiles. Field Museum Nat. Hist. Zool. Series 12:79-92. Schubert, C. 1988. Climatic changes during the last glacial maximum in northern South America and the Caribbean: a review. Interciencia. 13:128-137.

PAGE 393

384 Seijas, A.E. 1984. Estimaciones poblacionales de babas (Caiman crocodilus ) en los llanos occidentales de Venezuela. Ministerio de Ambiente y de los Recursos Naturales Renovables, (PT) Serie Infonnes Tecnicos DGSIIA/IT/165. 23 p. Seijas, A.E. 1988. Habitat use by the American crocodile and the spectacled caiman coexisting along the Venezuelan coastal region. Unpubl. MS Thesis, University of Florida, Gainesville. 104 p. Seijas, A.E., and S. Ramos. 1980. Caracterizacion de la dieta le la baba, Caiman crocodilus . durante la estacion seca en las sabanas moduladas del Estado Apure, Venezuela. Acta Biol. Venez. 10:373-389. Seymour, R.S. 1982. Physiological adaptations to aquatic life. pp. 1-52 in: F.H. Pough (ed.). Biology of the Reptilia. Vol. 13. Academic Press, New York. Siminoff, R. , and L. Kruger. 1968. Properties of reptilian cutaneous mechanoreceptors . Exp. Neurol. 20:403-414. Smith, C.C., and S.D. Fretwell. 1974. The optimum balance between size and number of offspring. Am. Nat. 108:499506. Spotila, J.R., O.H. Soule and D.M. Gates. 1972. The biophysical ecology of the alligator: Heat energy budgets and climate spaces. Ecology 53:1094-1102. Staton, M.A., and J.R. Dixon. 1975. Studies on the dry season biology of Caiman crocodilus crocodilus from the Venezuelan Llanos. Mem. Soc. Cienc. Nat. 35:237-266. Staton, M.A., and J. A. Dixon. 1977. Breeding biology of the spectacled caiman Caiman crocodilus crocodilus in the Venezuelan Llanos, U.S. Fish Wildl. Ser. Wildl. Res. Rep. No. 5. 21 p. Stebbins, R.C., and D.C. Wilhoft. 1966. Influence of the parietal eye on activity in lizards. Proc. Galapagos Int'l. Science Pro j . California: University of California Press 1966:258-268.

PAGE 394

385 Steel, R. 1973. Crocodylia. in: 0. Kuhn (ed.). Handbuch der Paleoherpetologie. Part 16. Fischer, Portland, Oregon. Stribrny, R. 1978. Nachzucht von Caiman crocodilus fuscus in der Gefangenschaft. DATZ 12:422-424. Sunguist, M.E., F.C. Sunguist and D.E. Daneke. 1989. Ecological separation in a Venezuelan llanos carnivore community, pp. 197-232 in: K.H. Redford and J.F. Eisenberg (eds.). Advances in Neotropical Mammology. Sandhill Crane Press, Gainesville, FL. Taphorn, D.C., and C.G. Lilyestrom. 1984. Los peces del modulo "Fernando Corrales" . Resultados ictiologicos del proyecto de investigacion del CONICIT-PIMA-18 . Rev. UNELLEZ de Cienc. y Teen. 2:55-85. Taylor, D. 1984. Management implications of an adult female alligator telemetry study. Proc. Ann. Conf. Southeast. Assoc. Fish Wildlife Agencies 38:222-227. Taylor, D.J., T. Joanen and L. McNease. 1976. A comparison of native and introduced immature alligators in northeast Louisiana. Proc. Ann. Conf. Southeast. Assoc. Fish Wildlife Agencies. 30:362-370. Taylor, J. A. 1979. The foods and feeding habits of subadult Crocodylus porosus Schneider in Northern Australia. Aust. Wildl. Res. 6:347-359. Thomas, B.T. 1979. The birds on a ranch in the Venezuelan llanos, pp. 213-232 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution. Washington. D.C. Thorbjarnarson, J.B. 1988. The status and ecology of the American crocodile in Haiti. Bull. Florida State Museum Biol. Sci. 33:1-86. Thorbjarnarson, J. 1990. Notes on the feeding behavior of the gharial (Gavialis qangeticus ) under semi-natural conditions. J. Herpetol. 24:99-100.

PAGE 395

386 Thorbjarnarson, J. In press. An Analysis of the spectacled caiman ( Caiman crocodilus ) harvest program in Venezuela, in: J. Robinson and K. Redford (eds.). Neotropical Wildlife Use and Conservation. University of Chicago Press, Chicago. Thorbjarnarson, J., and G. Hernandez. 1990. Recent investigations into the status of Orinoco crocodiles in Venezuela, pp 308-328 in: Crocodiles. Proc. 9th Working Meeting of the IUCN/SSC Crocodile Specialist Group. IUCN Publ. N.S., Gland, Switzerland. Thorbjarnarson, J., and G. Hernandez. In press. Recent investigations of the status and distribution of Orinoco crocodiles ( Crocodylus intermedius ) in Venezuela. Vida Silvestre Neotropical. Thorbjarnarson, J., and T.M. O'Brien. In prep. Ecological and taxonomic correlates of crocodilian reproductive attributes . Tinkle, D.W. 1957. Ecology, maturation and reproduction of Thamnophis sauritis proximus . Ecology 38:69-77. Tinkle, D.W. , and N.F. Hadley. 1975. Lizard reproductive effort: caloric estimates and comments on its evolution. Ecology 56:427-434. Tricart, J. 1985. Evidence of Upper Pleistocene dry climate in northern South America, pp. 197-217 in: I. Douglas and T. Spencer (eds.). Environmental Change and Tropical Geomorphology, I. Allen and Unwin, London. Troth, R.G. 1979. Vegetational types on a ranch in the central llanos of Venezuela, pp. 17-30 in: J.F. Eisenberg (ed.). Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution. Washington. D.C. Vila, P. i960. Geografia de Venezuela. 1. El Territorio Nacional y su Ambiente Fisico. Ministerio de Educacion. Direccion de Cultura y Bellas Artes, Caracas. Vitt, L.J. and H.J. Price. 1982. Ecological and evolutionary determinants of relative clutch mass in lizards. Herpetologica 38:237-255.

PAGE 396

387 Vliet, K. 1986. Social behavior of the American alligator, pp. 203-211 in: Crocodiles. Proc. of the 7th working meeting IUCN/SSC Crocodile Specialist Group. IUCN Publ. N.S. Gland, Switzerland. Vliet, K. 1987. A quantitative analysis of the courtship behavior of the American alligator. Ph.D. Dissertation. University of Florida, Gainesville. 198 p. Vliet, K. 1989. Social displays of the American alligator ( Alligator mississippiensis ) . Amer. Zool. 29:1019-1031. von During, M. 1973a. The ultrastructure of lamellated mechanoreceptors in the skin of reptiles. Z. Anat. Entwickl . -Gesch. 143:81-94. von During, M. 1973b. The ultrastructure of cutaneous receptors in the skin of Caiman crocodilus . RheinischWestfalische Akademie der Wissenschaften, Symposium on Mechanoreception 53:123-134. Watanabe, M.E. 1980. An ethological study of the American alligator ( Alligator mississippiensis Daudin) with emphasis on vocalizations and responses to vocalizations. Unpubl. Ph.D Thesis. New York University. 182 p. Webb, G.J.W., Beal, A.M., Manolis, S.C. and Dempsey, K.E. 1987. The effects of incubation temperature on sex determination and embryonic development rate in Crocodylus iohnstoni and C. porosus . pp. 507-531 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia. Webb, G.J.W. , R. Buckworth, and S.C. Manolis. 1983a. Crocodvlus iohnstoni in the McKinlay river area, N.T. III. Growth, movement and the population age structure. Aust. Wildl. Res. 10:383-401 Webb, G.J.W., R. Buckworth, and S.C. Manolis. 1983b. Crocodvlus iohnstoni in the McKinlay River area, N.T. VI. Nesting Biology. Aust. Wildl. Res. 10:607-637.

PAGE 397

388 Webb, G.J.W., and H. Cooper-Preston. 1989. Effects of incubation temperature on crocodiles and the evolution of reptilian oviparity. Amer. Zool. 29:953-972. Webb, G. J.W. , S.C. Manolis, and R. Buckworth. 1983c. Crocodylus iohnstoni in the McKinlay river area, N.T. I. Variation in the diet, and a new method of assessing the relative importance of prey. Aust. J. Zool. 30:877-899. Webb, G.J.W., S.C. Manolis, and R. Buckworth. 1983d. Crocodylus iohnstoni in the McKinlay River area, N.T. II. Dry season habitat selection and an estimate of the total population size. Aust. Wildl. Res. 10:373-382. Webb, G.J.W., S.C. Manolis, K.E. Dempsey and P.J. Whitehead. 1987. Crocodilian eggs: a functional overview, pp. 417-422 in: G.J.W. Webb, S.C. Manolis and P.J. Whitehead (eds.). Wildlife Management: Crocodiles and Alligators. Surrey Beatty Pty. Ltd., Chipping Norton, Australia. Webb, G.J.W. and H. Messel. 1978. Movement and dispersal patterns of Crocodylus porosus in some rivers of Arnhem Land, Northern Australia. Aust. Wildl. Res. 5:263-283. Webb, G.J.W., G.C. Sack, R. Buckworth and S.C. Manolis. 1983e. An examination of Crocodylus porosus nests in two Northern Australian freshwater swamps, with an analysis of embryo mortality. Aust. Wildl. Res. 10:571-605. Wever, E.G. 1978. The Reptile Ear. Princeton University Press, Princeton. Wilbur, H.M. and P.J. Morin. 1988. Life history evolution in turtles, pp. 387-440 in: C. Gans and R. Huey (eds.). Biology of the Reptilia. Vol. 16B. Alan R. Liss Inc., New York. Wilkinson, P.M. 1984. Nesting Ecology of the American alligator in coastal South Carolina. South Carolina Wildlife and Marine Resources Dept. Div. Wildl. Freshwater Fish. 113 p.

PAGE 398

389 Wright, D.E., and L.A. Moffat. 1985. Morphology and ultrastructure of the chin and cloacal glands of juvenile Crocodylus porosus (Reptilia, Crocodilia) . pp. 411-422 in: G. Grigg, R. Shine and T. Ehrman (eds.). Biology of Australasian Frogs and Reptiles. Roy. Zool. Soc. New South Wales. Sydney.

PAGE 399

BIOGRAPHICAL SKETCH John B. Thorbjarnarson was born on March 23, 1957, in Boston, Massachusetts. After completing his undergraduate work at Cornell University, he enrolled in the graduate program in the School of Forest Resources and Conservation at the University of Florida. From 1980 to 1984 he conducted his masters degree work on the status and ecology of the American crocodile in Haiti and the Dominican Republic. After completing his master's, he visited Venezuela to help run a captive breeding program for the endangered Orinoco crocodile on Hato Masaguaral, a cattle ranch situated in the central llanos region of the country. During his stay in Venezuela he initiated an investigation of the ecology and behavior of the spectacled caiman. This work expanded and eventually became the topic of his dissertation. In October 1990 he will return to Venezuela to resume his work on Orinoco crocodiles and spectacled caiman with a postdoctoral appointment from Wildlife Conservation International . 390

PAGE 400

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree /bk Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Katherine C. Ewel Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. -cfjoTTLp J . €u i Associate Professor of Zoology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of^hilosophy . Kelvin E. Suriquist Associate Research Scientist of Forest Resources and Conservation

PAGE 401

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the deqree of Doctor of Philosophy. Professor of Zoology This dissertation was submitted to the Graduate Faculty of the School of Forest Resources and Conservation in the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. ^ May 1991 Directory Forest Conservation Dean, Graduate School


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EN2XBXOZ2_XERK4E INGEST_TIME 2015-04-14T18:44:31Z PACKAGE AA00029947_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES