TT , 'I . .,- 1 'TATr --r,rr 1..[la . I. 5 of the FLO IDA STATE MUSEUM Biological Sciences Volume 33 1988 Number 1 THE STATUS AND ECOLOGY OF THE AMERICAN CROCODILE IN HAITI John B. Thorbjarnarson ''i S i I I. 5 .4 . /5.5.3 -bit- 5 ®®e~X=£%5 '. UNIVERSITY OF FLORIDA GAINESVILLE Numbers of the BULLETIN OF THE FLORIDA SrATE MUSEUM, BIOLOGICAL SCIENCES, are published at irregular intervals. Volumes contain about 300 pages and are not necessarily completed in any one calendar year. S . DAVID WEBB, Editor OLIVER L. AUS['IN, JR., Editor Ememus RHODA J. BRYANT, Managing Editor ~ Communications concerning purchase or exchange of the publications and all manuscripts should be addressed to: Managing Editor, Bulletin; Florida State Museum; University of Florida; Gainesville FL 32611; U.S.A. This public document was promulgated at an annual cost of $3150.00 or $3.150 per copy. It makes available to libraries, scholars, and all interested persons the results of researches in the natural sciences, emphasizing the circum-Caribbean region. ISSN: 0071-6154 CODEN: BF 5BAS Publication date: May 17, 1988 Price: $3.15 THE STATUS AND ECOLOGY OF THE AMERICAN CROCODILE IN HAITI John B. Thorbjarnarson* ABSTRACT The American crocodile (Crocodylus acums) is the most widely distributed of the New WorId crocodiles. Due to a combination of hunting for its hide, habitat destruction, and malicious killing, the American crocodile is currently in danger of extinction, with isolated populations existing for the most part only in relatively undisturbed areas. Conservation and management programs are sorely needed to protect this species, but are hampered by a lack of knowledge concerning the current status of many of the extant populations and biological data concerning many aspects of the crocodile's natural history. In Haiti, crocodiles were once widely distributed throughout coastal and lowland areas where suitable habitat was available. Today, the range of the crocodile in this country has been greatly reduced, and the few extant populations have been severely depleted, remaining only in those areas that have a combination of relatively low human population density and sufficient mangrove habitat. Presently, no commercial hide hunting is occurring, and the taking of crocodiles for food or the use of their by-products are restricted to two areas bordering the Dominican Republic. Habitat destruction and incidental killings, primarily by fishermen, represent the greatest threat to cfocodiles in Haiti today. The largest remaining population is found in Etang Saumatre, Haiti's largest lake (113 sq km). Etang Saumatre is a brackish lake, located in a sparsely inhabited region only 30 km from the capital of Port-au-Prince. The total crocodile population in the lake is estimated at 450. Over a 13-month period, various aspects of the demography, reproductive ecology, diet, and habitat selection of these crocodiles were investigated. The results are discussed in the light of other work which has been done on crocodilians. The final section outlines several recommendations for the conservation of crocodiles in Haiti. * Florida State Museum and Department of Wildlife and Range Sciences, University of Florida, Gainesville FL 32611. THORBJARNARSON, J. B. 1988. The Status and Ecology of the American Crocodile in Haiti. Bull. Florida State Mus., Biol. Sci. 33(1):1-86. 2 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) RESUMEN El cocodrilo americano (Crocodylus acutus) tiene la distribuci6n mds amplia entre las 4 especies de los cocodrilos del Nuevo Mundo. Debido a la caza comercial para su piel, destruci6n de su habitat, y a la simple matanza por maldad, el cocodrilo americano esta en peligro de extinci6n, y hoy en dia solo quedan poblaciones aislados en zonas remotas. Se necesitan urgentemente programas de conservaci6n y manejo para la protecci6n de esta especie, pero el desarollo de cmos programas es limitado por la falta de conocimiento sobre el estado actual de la mayoria de las poblaciones, y muchos aspectos de su historia natural. En Haiti, cocodrilos tenian una distribuci6n amplia a 10 Iargo de la costa. Hoy, estadistribuci6n se ha reducido drasticamente, y las pocas poblaciones que quedan son mypequefias, ubicadas en zonas despobladas donde hay sufficiente habitat con manglares. Hoy en dia, no existe una caceria comercial para cocodrilos en Haiti, y la gente los utilizan paracomida o remedios solamente en dos zonas al lado de la frontera con la Republica Dominicana. Los dos peligros mayores para la supervivencia de los cocodrilos son ladestruci6n de habitat, y la muerte por inmer5ion al enredarse en redes de pescadores. La mayor p9blaci6n de cocodrilos en Haiti estd en Etang Saumatre, el lago mas grandedel pais (113 km«'). Etang Saumatre es de agua salobre, y esta ubicado en una zona semi- arida a solamente 30 km de la capital; Puerto Principe. Approximadamente 450 cocodrilos viven en este Iago. Durante un periodo de 13 meses, se estudiaron aspectos de la demografia, reproduci6n, dieta, y uso de habitat de esta poblacidn de cocodrilos. Los resultados estandiscutidos en relaci6n con otras investigaciones sobre cocodrilos. La secci6n final presenta algunas recomendaciones para la conservaci6n del cocodrilo en Haiti. TABLE OF CONTENTS Introdlirtinn 3 Acknowledgemenk 4 Methods f Survey of Status and Distribution..................................................................................................5 Ecological Study: Etang Saumatre Present Status and Distribution in Haiti.................................................................................................9 Introduction 9 Rpoillk 11 Discussinn 17 Ecological Study: Etang Saumpre 74 Introduction 74 Demogrprhy 71 Reproductive Ecology 41 Diet 51 Habitat Selection.............................................................................................................................53 Conservffinn 58 Discussinn 58 Recommendations...............................................................................................................60 THORBJARNARSON: AMERICAN CROCODILE IN HAITI 3 Summary 61 Literature Cited 6? TAbln 68 Appendiref RO I. Distribution of Mangrove in Haiti.........................................................................................82 II. Aquatic and Semi-aquatic Avifauna, native fish fauna,...................................................... and the Dry Forest Vegetative Association of Etang Saumatre...........................................84 III. Survey Correction Procedure.............................................................................................86 INTRODUCTION The American crocodile (Croco*lus acutus) is the most widely distributed of the four species of New World crocodiles. Although it may be found well inland in freshwater habitats, the American crocodile lives primarily in coastal areas, preferring brackish water habitats associated with mangrove forests, coastal lagoons, and the estuarine sections of rivers. As with most species of crocodilians, unregulated hide-hunting, malicious killing, and habitat destruction have resulted in drastic population declines of the American crocodile throughout its range. Today, these crocodiles exist mostly in disjunct populations where past exploitation and habitat destruction have not been overly severe. Although the general trend of local extirpations and reductions in numbers clearly has been evident for some time (Barbour 1923, Moore 1953, Casas and Guzman 1970, Alvarez del Toro 1974), very little is known about the status of existing populations. A recent review (King et al. 1982) collated reports for C acutus on a country by country basis. Detailed information was available only for a few countries, and the review pointed out the clear need for surveys to determine the present status and distribution of the remaining populations. Data of this sort are needed for the implementation of effective conservation and management schemes designed to insure the continued survival of the species. Primarily a coastal species, the range of the American crocodile includes southernmost Florida, the Atlantic coast of Mexico south through Central America and northern South America (east to the Peninsula de Paria in Venezuela), and the Caribbean islands of Cuba, Jamaica, and Hispaniola. The American crocodile is also found in Pacific Ocean drainages from northern Mexico (Sinaloa) south to the Rio Tumbes in northernmost Peru. The northern distribution of the American crocodile along the Atlantic coast of Mexico remains a question. While crocodiles are known from the states of Tamaulipas and Veracruz, all confirmed specimens from these areas are C moreleti. Nevertheless literature accounts of C acutus exist for the Atlantic coast north of the Yucatan. Further survey work is needed to determine the range of 4 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) these two species in Mexico. Very little was known about the status of crocodiles on Hispaniola prior to the mid-1970's. lIn 1975 scientists from the Museo Nacional de Historia Natural began an ecological study of crocodiles in Lago Enriquillo, the largest lake in the Dominican Republic. This lake was found to contain what may be the largest remaining C acums population anywhere (Inchaustegui and Ottenwalder pers. comm.; pers. observ.). Coastal populations in the Dominican Republic, however, appear to have been eliminated completely with the exception of the Rio Massacre along the northwestern border with Haiti. Today, both these populations are protected by law (Decreto de Veda 861, June 1979). Prior to the present study, however, the status of crocodiles in Haiti was completely unknown (Powell 1971, King et al. 1983). The objectives of this investigation were twofold: (1) to determine the status and distribution of crocodiles in Haiti; and (2) to record aspects of the ecology of this poorly known species. The findings of the status survey are somewhat unexpected in that Haiti, a country with twice the population density and a much lower standard of living than the neighboring Dominican Republic, has several coastal crocodile populations as well as a substantial population in Bang Saumatre, Haiti's largest lake. The ecological study was done in Etang Saumatre, which is, at its closest, only 5 km from Lago Enriquillo in the Dominican Republic. The crocodile population in this lake was someWhat unusual, as it was not found in typical coastal wetland habitat. The crocodiles, however, were relatively easy to locate and capture due to the limited amount of vegetative cover. In addition, the lake represented a closed system, containing an easily defined population. In contrast, coastal populations are often widespread over large areas and are not easily delimited. The inaccessible nature of many of these habitats also precludes any easily accomplished quantitative work on population parameters. Based on the findings of the status survey and the ecological investigation, the last section outlines several recommendations for the conservation of crocodiles in Haiti. ACKNOWLEDGEMENTS Many people contributed to this project. Firstly, I am indebted to F. Wayne King, mymaster's thesis chairman, for his continual support during the course of this project, and Charles Woods, who first suggested the feasibility ofwork in Haiti and then provided many ofmy initial contacts in that country. They, along with the other two members of my graduate committee, J. Robinson and M. Collopy, read and made many constructive commentsconcerning this thesis. My stay at Etang Saumatre was made possible through the generosity of Pastor Wallace and Eleanor Turnbull and Pastor Eric and Irene Lange, who provided me with a place to stay and store my equipment at Tete Source. The Langes provided constant support and companionship, as well as many greatly appreciated home-cooked meals. Work in Haiti was sponsored through the Institut de Sauvegarde du Patrimoine National, Albert THORBJARNARSON: AMERICAN CROCODILE IN HAITI 5 Mangones, Director. Paul Paryski especially provided a great deal of support. Permission to conduct the work was obtained from Jean-Baptiste, Jean-Francois, and Edmond Magny of the DEpartement de l'Agriculture des Ressources Naturelles et du D6veloppment Rural (DARNDR). Joseph Felix, also of DARNDR, kindly supplied permits for the export of some specimens. Ekke Lempke provided a great deal of information concerning crocodiles and first introduced me to Etang Saumatre. Jaques Durocher and Jimmy Stecher also supplied much information on the current distribution of crocodiles in Haiti. Jimmy Stecher and Ted Steinhauer provided logistical support for trips to the Laborieux region and La Gonave. Dan Cordier and Brent Mitchell assisted during portions of the coastal surveys. Ragnar Arnesson and Roland Roy, of the Organization of American States in Port-au-Prince, supplied considerable assistance in attending to many of the problems which arise during work in foreign countries. My assistants in Etang Saumatre were Eldee Antoine and Tony Samveiss. Pierre Milfort of DARNDR also helped during the early stages of the project. Others who provided assistance were Peter Blanchard, Drew Kutchenreuter, Jim Keith, David Pulle, Tom Greathouse, Mara McDonald, and Mike Binford. Identification of prey items was done by Robert Woodruff (insects), G.B. Edwards (spiders), Mintor Westfall (dragontly larvae), Eleanor H. Stickney (birds), and George Burgess and Richard Franz (fishes). Plant identifications were performed by David W. Hall, S. Davis, and D. Griffin. Expertise for the step-wise discriminant analysis was provided by C. Abercrombie. Work in the Dominican Republic was sponsored by the Museo Nacional de Historia Natural, in particular Lic. Jose Alberto Ottenwalder, Lic. Sixto Inchaustegui, and David Robinson. I would aIso like to thank Leslie D. Garrick for allowing me to cite some of his unpublished data from Jamaica. Funding for work, both in Haiti and the Dominican Republic, was provided by the Wildlife Conservation International (WCI) (formerly Animal Research and Conservation Center [ARC]) of the New York Zoological Society. Research in Haiti was also funded through a fellowship from the Organization of American States. Additional funding was provided by two grants-in-aid of research from Sigma Xi, The Scientific Research Society. Reynolds Aluminum kindly supplied the use of a pick-up truck over a 6-month period of time in Haiti. METHODS Survey of Status and Distribution Surveys of coastal areas and inland lakes (other than Etang Saumatre) were conducted primarily during May-June 1983. Potential crocodile habitats were initially identified on the basis of information provided by reliable informants familiar with wildlife in Haiti. Next, 1:50,000 topographic maps were used to pinpoint areas of possible crocodile habitat prior to visiting those areas. The bulk of the information on the presence or absence of crocodiles was obtained from conversations with knowledgeable local residents (mostly fishermen). Areas were surveyed on foot or by boat during the day to determine the quantity and the quality of the available habitat. Whenever possible, night surveys were also made using a headlamp to spot crocodiles or their reflected eyeshine. However, due to the limited amount of time available to conduct the surveys and the frequent lack of a suitable boat, night surveys were not conducted at all locations. The past distribution of crocodiles in Haiti was reconstructed based on (1) historical accounts of crocodiles in specific locations, (2) place names referring to "caiman," the 6 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) local word for crocodiles, and (3) the distribution of suitable crocodile habitat in coastal areas. Regions containing extirpated and extant crocodile populations were then compared on the basis of amount of mangrove habitat (determined from 1:50,000 topographic maps) and human population density based on recent population census (IHSI 1983). Ecological Study: Etang Saumatre Demography Crocodiles were censused at night from a boat using a Q-Beam spotlight (200,000 candie power) run off a 12-volt marine cell. Two censuses of the entire lakeshore were performed, one requiring three nights (11, 17, 18 August 1983) and the other two nights (7,8 January 1984). The August survey was not conducted in consecutive nights because of boat trouble. It appears, however, that crocodile movements between survey sections were not sufficiently large to introduce significant error. Surveys were conducted atan average speed of 7 km/hr, approximately 30 m from shore. The light was swept along the shoreline and periodically out into the lake. Crocodiles were spotted by their reflected eyeshine and approached to estimate size (0.3- 0.9 m, 0.9-1.8 m, 1.8-2.7 m, > 2.7 m). Because hatchlings did not form pods and were frequently found in dense vegetation along with yearlings, it was unfeasible to separate these two size classes during counts. If the crocodile submerged before an accurate estimate of size could be made, it was placed into one of the following categories: EO > 1.8 m (eyes only, greater than 1.8 m total length), EO < 1.8 m (eyes only, less than 1.8 m), or EO (eyes only). The location of all crocodiles sighted was plotted on a map of the lake as they were spotted. Surveys produced a base estimate of population size and size- class distribution. These data were then corrected for: (1) reduced sightability in areas of dense shoreline vegetation; (2) known animals that were not seen during the survey; and (3) in the January 1984 survey, line transects were conducted in two dense Conocatpus sWamps to estimate the number of crocodiles that could not be seen from the lake. A full description of the correction procedufes is given in Appendix III. After determining the size-class distribution of the population, the EO, EO < 1.8 m, and EO > 1.8 m sighting classes were divided proportionately between the four known-length size classes. As large animals generally are more wary than small ones, this may have resulted in a slight bias against the larger size classes. Sex ratio and a length-weight relationship were determined from captured individuals. Large crocodiles (> 1.8 m) were caught at night from a boat. Smaller individuals usually were captured when THORBJARNARSON: AMERICAN CROCODILE IN HAITI 7 wading through shallow water habitats. Crocodiles were caught by hand, with pilstrom tongs, or by using breakaway locking cable snares mounted on the end of a pole (Jones 1965). Crocodiles were weighed on Homs spring scales (100 g, 2 kg, or 10 kg capacity) or Hansen spring scales (136 kg capacity). All captured crocodiles were marked in two ways; by placing self-piercing monel tags in the webbing of the hind foot, and cutting numerical sequence of dorsal caudal scutes. Due to uncertainties in the sexing procedure of juveniles (> 1 m) during 1983, sex data from these animals were not used. Because of the difficulty of sexing small animals based on morphological differences in the penis/clitoris (Joanen and McNease 1979), data on the sex of animals less than 40 cm long were not taken. Growth rates were obtained using successive lengths from recaptured animals. Hatchling growth rate also was estimated by assuming a mean hatching date and length, based on hatchlings found in recently opened nests. In this manner the age and growth of first-year animals could be estimated. Biomass was determined from the length-weight relationship of captured animals and the size-class distribution of the population. The mass of a crocodile at the midpoint of each size class was used as an estimate of the average mass of a crocodile in that size class. These values then were multiplied by the total number of crocodiles for each class and summed for all four size classes. For the > 2.7 m size class, mean size was assumed to be 2.9 m. Reproductive Ecology The location of the crocodile nesting beaches and the number of 1983 nests were determined by extensively searching lakeshore habitats during May 1983. Active nests were identified by the presence of an open hole surrounded by eggshells and egg membrane fragments. In 1984, nesting beaches were monitored beginning in early January for signs of activity. Nests were located by following the tracks left by females and probing by hand under the substrate in areas where obvious digging had occurred. Once located, nests were carefully excavated to determine clutch size and egg fertility rate (by egg banding, Ferguson and Joanen 1983) and to measure nest hole dimensions. Measurements of egg dimensions and egg mass were made on a sample (n = 5) of eggs from each clutch. A 100-200 g soil sample was taken for later analysis of water content (by drying over a butanestove) and soil particle size distribution (by passing through a series of sieves). All weights were measured on a 200 g Pesola spring scale (0.5% accuracy). A variety of other parameters were recorded at each nest site. Height of vegetation was estimated to the nearest 0.5 m, height above the 8 BULLEIIN FLORIDA STATE MUSEUM VOL. 33(1) lake to the nearest 0.3 m. Percent of shrub, grass, and leaf litter cover was estimated to the nearest 10%. In six nests copper-constantan thermocouples were implanted at the top and the bottom of the clutch. Nest temperatures were recorded over a 30- hour period using an Omega 871 digital thermometer. To determine the environmental parameters important in the selection of nest sites, 15 null sites were randomly chosen along the major nesting shore. The same environmental parameters measured at the nest sites were recorded at each of the null sites (except distance to lake which was standardized at 25 m) and a stepwise discriminant analysis performed on the data set. Dietary Analysis Stomach contents from crocodiles under 1 m total length were obtained using the stomach flushing technique described in Taylor et al. (1978). For crocodiles longer than lma modified scooping technique was used. Crocodiles were strapped to a wood plank with their taped jaws immobilized in an open position around a 7.6 cm diameter section of PVC pipe. Stomach contents were extracted using natural latex (Paramold, Imperial Adhesives and Chemicals, Inc.) scoops moulded from small funnels (opening diameter 6.5 cm, 7.0 cm long). The narrow end of the funnel-shaped scoop was pushed down the esophagus using a 2.5 cm diameter rod until the scoop was felt to reach the end of the stomach. The scoop and rod were generously lubricated with vegetable oil to reduce the chance of trauma to the esophagus during this procedure. Once in the stomach, the rod was extracted and the scoop was slowly pulled out using two strings attached to opposite sides of the funnel rim. The narrow, open end of the scoop was covered with cheesecloth to allow fluids to pass through as it was drawn through the stomach. This procedure was repeated at least three times and was used on animals up to 2.88 m total length. All stomach contents were preserved in alcohol for later identification and analysis. Stomach contents were categorized into three groups: fresh, partly digested, and fragments. Items from the first two groups were individually weighed on an Ainsworth 21N analytical balance to the nearest 0.1 g. The presence of gastroliths, vegetation, and nematodes was noted. For purposes of comparison, invertebrate prey items generally were classified to the ordinal level. Representative invertebrate prey items were identified to family or genus to allow analysis of prey ecology and, by inference, crocodile foraging modes. Vertebrate prey was identified to genus or species in all cases. THORBJARNARSON: AMERICAN CROCODILEIN HAITI 9 Habitat Selection The various lakeshore habitat types were categorized by physiognomy of the vegetation or shoreline features. The extent of the habitat types was mapped on 1:25,000 topographic maps of the lake during day surveys by boat. The locations of all crocodiles seen during the population surveys were marked on a map of the lake and later assigned to one of the habitat types. Shorelines or the lake-vegetation interface also were assigned an exposure index (protected, moderate, or exposed) based on the amount of wave action they received from the predominant easternly winds. Spatial distribution of crocodiles around the lake was examined by dividing the lake into eight segments of varying length (4.3-16.0 km long) and comparing the crocodile population in each of these segments. PRESENT STATUS AND DISTRIBUTION IN HAITI Introduction The Republic of Haiti (Fig. 1) occupies 27,700 sq km of the western third of the island of Hispaniola, the second largest of the Greater Antilles (Woodring et al. 1924). A French colony until 1804, Haiti became the world's first black republic following a bloody revolution that lasted nearly 15 years. The name Haiti is derived from an Arawak Indian word meaning "mountainous land" and provides a very apt description of this country, which has peaks up to 2680 m and more than 65% of its surface area sloped greater than 20 degrees (AID 1982). Within its diverse topography, Haiti supports a wide variety of ecological life zones and associated plant communities ranging from dry thorn scrub to mountain pine forests. In fact, the Holdridge life zone classification system was first developed during work on the mountain vegetation of the La Selle ridge of southeastern Haiti (Holdridge 1947). Today, however, very little of the natural vegetation remains because of extensive deforestation. With an estimated population of more than 5 million, Haiti has one of the world's highest population densities. This, in combination with a paucity of arable land, has resulted in a degree of environmental degradation that is perhaps without equal in the world today. The chief problem is one of rampant erosion, resulting from the nearly complete deforestaton of many hillsides 10 BULLETIN FLORIDA STATEMUSEUM VOL. 33(1) ILEDE LA TORIUE PAIX CAP, HAI"EN rA.L« GONAIVES 41*/ 6 - RIVIERE ARTIBONITE ILE DE LA - GONAv E TRou CAIMAN 0 JEREMIE PORT-AU-PI~INCE E TAN 0. SAUMATRE AQUIN CAYES VACHE N ~ 20 49 km <2, FIGURE 1. Map of Haiti including some major landmarks. Inset is a map of the West Indies with the relative location of Haiti shown as a solid black area. without the implementation of any soil conservation techniques. Today, the effects of such past practices are being sorely felt by the Haitian people who have the lowest per capita income in the western hemisphere. Recognition of the problem has been slow, but currently the Haitian government, in cooperation with several foreign aid agencies and volunteer organizations, is beginning to implement reforestation projects. Centuries of human depredation and the virtually complete loss of lowland forested regions have had drastic effects on the local fauna. The hardest hit have been the endemic non-volant mammals, as out of a pre- Colombian total of 25, today only 2 species survive, and both of these are quite rare, persisting in only a few relatively undisturbed areas of suitable habitat (C. Woods pers. comm.). Similarly with reptiles, the giant rock iguanas (Cyclum comum and C dcordi) have become Very rare and today are found only in a few dry, rocky areas inhospitable to man. THORBJARNARSON: AMERICAN CROCODILE IN HAITI 11 The crocodile has managed to survive in Haiti principally because man and American crocodiles are essentially allopatric in their distribution, as man finds the coastal wetlands the crocodiles prefer marginal for agriculture or habitation. In contrast to the near total destruction of the terrestrial forests, the coastal forests, especially mangrove, have fared considerably better. During the period between 1956 and 1977 only 7% of the existing mangrove disappeared. The corresponding figure for loss of pine forest was 40-70% (FAO 1978). While the mangroves are not cleared for agricultural use as are the terrestrial forests, mangrove is used for firewood in bakeries, home cooking, distilleries, and dry cleaners. Mangrove wood also is used as fuel for burning coral rock to produce lime for cement, making charcoal, and for construction purposes. Mangrove has been spared to a large degree in the past, but as the human population continues to grow, more and more pressure will come to bear on these forests. Already the pace of mangrove destruction appears to have been accelerated (pers. observ.). Although less apparent than the cutting of mangrove, diversion of freshwater for agriculture has probably had more lasting, although as of yet largely undetermined, effects upon some of these ecosystems. This is most notable in the l'Ester region, which contains Haiti's largest mangrove swamp. Mangrove forest still remains in many parts of Haiti, however, and these areas serve as nuclei for the present coastal distribution of crocodiles (see Appendix I). This section will present the results of a country-wide survey to determine the present status of crocodiles in Haiti. The findings will then be examined in relation to the past distribution of crocodiles, as determined from historical records, to document the retreat of the crocodiles into isolated populations and to provide some useful insights into the ability of crocodiles to survive in man-dominated ecosystems. Results Past Distribution of Crocodiles in Haiti Fourteen historical accounts and eight place names were found that made reference to crocodiles (Fig. 2). These sources, plus information on the distribution of suitable crocodile habitat, were used to reconstruct the probable former distribution of crocodiles in Haiti (Fig. 3). 12 BULLETIN FLORIDA STATE MUSEUM VOL 33(1) Present Distribution of Crocodiles in Haiti Four regions containing extant coastal crocodile populations were identified: (1) the southern coast of the Tiburon Peninsula from Cote-de-Fer west to the RiviBre l'Acul, including Ile-8-Vache, (2) Ile de La Gonave, (3) the l'Ester-Artibonite mangrove swamp, and (4) the Rivi6re Massacre-Lagon aux Boefs region bordering on the Dominican Republic. In addition, the largest remaining crocodile population is found in Etang Saumatre, an inland lake not far from the capital of Port-au-Prince. The present range of crocodiles in Haiti is summarized in Figure 4. A comparison of the present and past distributions of crocodiles reveals that crocodiles have been 0 20 40i km <-J 3,7,10,13,A Dll ~ B 4,5,8 0,j 2,4,5,6 4 H 0 If 3,4,14 4 I2 FIGURE 2. Locations of historical records of crocodiles and place names referring to "caiman" in Haiti. (Historical Accounts: 1 - Las Casas 1561, 2 - Lescallierer 1764, 3 - Moreau de St. Mery 1796, 4 - Moreau de St. Mery 1797-8, 5 - Descourtilz 1809, 6 - Hearne 1834, 7 - Ritter 1836, 8 - Gosse 1851, 9 - Fortunat 1889, 10 - Hazard 1873, 11 - Rodriguez 1915, 12 - Wetmore Perrygo 1931, 13 - Loederer 1935, 14 - Steedman 1939. Place Names: A - Isleta de los Caimanes, B - Bassin Caiman, C - Calman, D - Bassin Calman, E - Carman, F - Trou Carman, G - Rividre Carman, H - Trou Carman.) THORBJARNARSON: AMERICAN CROCODILE IN HAITI 13 0 20 40 11, 1-1 N FIGURE 3. Reconstructed historical range of crocodiles in Haiti, based on Figure 2, the present crocodile distribution, and the availability of suitable habitat. extirpated from approximately 70% of their former coastal range, and two of the three inland lake systems where they were once found (Etang Laborde, Etang Miragoane). Southern Coast of the Tiburon Peninsula; A diffuse crocodile population exists in this region from immediately west of Cotes-de-Fer, west to the vicinity of the Rivitre l'Acul (Fig. 5). Crocodiles are regularly seen in a number of the more isolated coastal wetlands, specifically: the vicinity of Laborieux-L'Osiendieu, the Aquin-RiviBre Capolo region, the Rivitre Cavaillon, and the Rivitre Bondonne. Crocodiles are sporadically seen in pockets of suitable habitat between these areas, primarily associated with mangrove swamps or the dense vegetative cover at river mouths. Nesting reportedly occurs in vegetated beach strand habitats along the lower reaches of the RiviBres Cavaillon and Bondonne. During surveys, crocodiles were only seen in the Laborieux region (5 subadults seen, 17 April 1983), but according to local residents the largest population remains in a 14 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) 92,0 4,0 9-J N 0 00 FIGURE 4. Present range of crocodiles in Haiti. N 0. Alo AQUIN 3 4 2 6 CAYES BAIE D'AQUIN 7 COTES·DE·FER 8 0 20 40 KM i LE A VACHE CARIBBEAN SEA FIGURE 5. South-central Tiburon Peninsula; marked locations indicate the postion of extant or former crocodile populations (1 - Laborieux; 2 - Osiendieu, 3 - Rividre Capolo, 4 - Rividre Millionaire, 5 - Rividre Cavaillon, 6 - Rivitre Torbeck, 7 - Rivitre Bondonne, 8 - Rividre l'AcuI, 9 - l'Etang, 10 - Etang Laborde). THORBJARNARSON: AMERICAN CROCODILE IN HAITI 15 mangrove swamp at the mouth of the Rivitre Cavaillon. A recently killed specimen (1.2 m total length) was found in the Riviare Capolo (25 June 1983, UF 54208). Crocodiles were also reported from a small freshwater lake (called L'Bang) on the northwestern end of Ile a Vache, located approximately 10 km off the southern coast of Haiti. Ile de La Gonave., La Gonave is the largest of Hispaniola's satellite islands (658 sq km), or about 2.3% of Haiti's land surface. Much of the coast is protected from wave action by a barrier reef, permitting the growth of a mangrove fringe even though freshwater runoff is almost non-existent. Crocodiles were reported from several of the coastal lagoons on the western end of the island. During a night survey of Lagon Blanch (11-12 June 1983), a shallow water lagoon along the north coast near the town of Richard, a total of five subadult crocodiles (to 1.5 m total length) were observed. l'Ester-Artibonite., Hispaniola's largest mangrove swamp (8490 ha) is found in the l'Ester region just south of the town of Gonarves. Immediately to the south of the l'Ester region is the mouth of the RiviBre Artibonite, Hispaniola's longest river (240 km). Crocodiles were well documented historically from this area by Descourtilz (1809) in his treatise on the "crocodile du St. Domingue." Today, crocodiles are well known to local fishermen, although they are not seen with any frequency. During a daytime survey by sailboat (4 June 1983) and a night survey of mangroves near Gonarves (5 June 1983), no crocodiles were seen. Crocodiles were also sporadically observed by local residents of Grande Saline, at the mouth of the Rivitre Artibonite, which lacks any protective mangrove forests. Rividre Massacre-Lagon aux Boefs., Lagon aux Boefs is a 4-sq-km, mangrove-lined, freshwater lagoon connected at its northern end to the Rivitre Massacre, which forms the northeastern boundary between Haiti and the Dominican Republic. The estuarine section of the RiviBre Massacre has a mixed riverine-fringe type mangrove swamp (after Lugo and Snedaker 1974) of approximately 1030 ha, most of Which is in the neighboring Dominican Republic. Local Haitians were very familiar with crocodiles, which they would catch and eat whenever possible. This is in stark contrast to the rest of Haiti where crocodiles are not eaten. A daytime survey of Lagon aux Boefs (16 June 1983) revealed no crocodiles. During a previous night survey of the lower Rivi6re Massacre in December 1981 one 3 m crocodile was seen approximately 2 km upstream from the river's mouth. Etang Saumatre and Trou Caiman., Haiti's largest remaining crocodile population, approximately 450 individuals of all sizes, is found in Etang Saumatre, located in the Cul-de-Sac valley 30 km northeast of Port-au-Prince (see Ecological Study: Etang Saumatre). A large lake (113 sq km), Etang Saumatre is surrounded by a relatively sparsely inhabited region of the 16 BULLETIN FLORIDA SrATE MUSEUM VOL. 33(1) country. The uninhabited eastern lakeshore, bordering the Dominican Republic, contains a significant amount of juvenile crocodile habitat and virtually all the nesting sites. Trou CaYman is a marshy freshwater lake (6.9 sq km) located 6 km west of Etang Saumatre. The two are connected by a small canal. The sporadic accounts of crocodiles in this lake suggest that it does not contain a breeding population but probably serves as a dispersal area for crocodiles from Bang Saumatre. Extirpated Populations North Coast West of Ft. Libert6., Crocodiles were known historically as far west as the mouth of the Rivitre Limb6 (Bassin Cayman). It is not known exactly when the last crocodiles in this region were extirpated, but none has been seen for many years. It is likely that at one time crocodiles ranged as far west as Port-de-Paix, as pockets of suitable habitat exist at regular intervals along the coast. West of Port-de-Paix, a rocky, high-energy coast predominates, providing little in the way of crocodile habitat. Crocodiles also were reported historically in the vicinity of Cap Haitien (Rodriguez 1915) where a moderate-sized (760 ha) mangrove forest still exists along the lower reaches of the RiviBre Haut-du-Cap. Crocodiles were also found in the Caracol region between Cap Haitien and Fort Libert6. The only published account is that of Ritter (1836), who mentions seeing crocodiles at Ft. Real, which may have been the old site of Puerto Real, a Spanish colonial settlement in the Caracol region. The presence of a small lagoon (the old river bed of the Grand Rivitre du Nord) called Bassin Cayman, and the finding of crocodile mandibles in Indian middens from the area (W. Hodges pers. comm.), however, attest to the fact that crocodiles were indeed in the area at one time. North and West Coasts of the Tiburon Peninsula., Historically crocodiles were found in the vicinity of Petit Goave (Riviere Caiman), the Rivitre Grande Anse near J6r6mie (Moreau de St. Mery 1797), and the Tiburon region at the tip of the peninsula (Las Casas 1552). They also were formerly found in several inland lakes in the area (see following section). Currently there are no verified populations anywhere in this part of Haiti. Crocodiles may still be found in small numbers in the BaradBres region, although this is doubtful. The small individuals reported from Petit Goave and JEremie (all in the 1.2-1.5 m range) probably represent vagrant individuals. Inland Lakes: Etang Laborde., Etang Laborde is the largest of four lakes located on the coastal plain 12 km north of Cayes (1978 size estimate THORBJARNARSON: AMERICAN CROCODILE IN HAITI 17 0.9 sq km). The region is heavily populated and extensive agriculture occurs around the lake. The lake itself is very shallow and reportedly dried up in 1975 after an extended drought. Moreau de St. Mery (1797) stated that crocodiles were found in this lake (then called Bang Vert) but had long since disappeared, making this the earliest known extirpation of crocodiles in Haiti. No crocodiles are currently found in Etang Laborde or any of the other lakes in the region. Inland Lakes: Etang Miragoane., Etang Miragoane presently consists of two lakes (8.3 and 1.3 sq km) near the north coast of the Tiburon peninsula, adjacent to the town of Miragoane. This region also is densely populated, and there is currently much fishing activity in the lake. The lake is fresh water with extensive grass fringes. Other vegetation includes Nuphar, Nelumbo, Typha, and Potamogeton. Moreau de St. Mery (1797) mentioned that the lake had many crocodiles 2.5-3.5 m long which nested in sandy areas surrounding the lake during the summer. Bang Miragoane has no crocodiles, and no one interviewed could remember there ever being crocodiles in the area. Discussion Analysis of the Present Distribution of Crocodiles Coastal Crocodile Movements., Croco*lus acutus is one of two primarily coastal dwelling crocodiles, the other being C porosus from northern Australia, the Indo-Malayan Archipelago, and Southeast Asia. The wide coastal distribution and probable recent evolutionary derivation of both these species (Densmore 1981) suggests that they are adept at moving along coasts and possibly even making transoceanic journeys. This is best documented in the case of C porosus (Bustard and Choudhury 1980) where specimens have been spotted at sea nearly 500 km north of New Zealand (Robb 1980), on Cocos-Keeling Island in the Indian Ocean nearly 1000 km from the closest known population in Indonesia, and on Ponape in the Western Caroline Islands some 1360 km from the nearest population (Allen 1974). Messel et al . ( 1982) developed a model of C porosus population dynamics for tidal rivers in northern Australia which proposes that a large fraction of the subadult crocodiles leave the productive nesting rivers when they reach a size (0.9-1.8 m) that brings them into conflict with larger territorial adults. Such individuals, if not killed outright, are forced to leave the river and find other, usually marginal, habitats. Some perhaps move along the coast and manage to find another river where territorial adults are less common (rivers less suitable for nesting) and take up residence there. 18 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) The majority, however, probably never survive. As adults it is possible the crocodiles will move back into more suitable rivers for reproduction. The model proposes, then, that movement and mortality are quite high for these intermediate-sized crocodiles. Logically, a similar pattern may hold for C acums, which is ecologically similar but poorer known from a biological standpoint. Both Alvarez del Toro (1974) and Medem (1981) reported C acutus moving from river to river using overland routes. Alvarez del Toro (1974) stated that such movements are in response to territorial fighting and the drying up of temporary lagoons. Mazzotti (1983) found C acutus moved considerable distances in the coastal regions of southern Florida. The presence of C acums on several mid- oceanic islands (Cuba, Jamaica, Hispaniola, and formerly the Cayman Islands) as well as a number of near-shore islands (Isla Margarita, Venezuela; Islas del Rosario, Islas de Sn. Bernardo, Isla Fuerte, Isla Tortuguilla, Colombia; and the Archipelago de los Canarreos, Cuba; as well as Ile a Vache and Ile de La Gonave off Haiti) strongly suggests the species is adept at moving long distances along the coast or across open water. Fishermen in Haiti reported that on occasion crocodiles could be seen in the ocean. Crocodile movement along the coast would explain many of the unusual reports of local residents along the southern coast of the Tiburon peninsula. In this region, crocodiles were said to be found in several areas with little or no suitable habitat, and their presence in many of these areas was reported to be of irregular occurrence. Similarly, reports of crocodiles on the northern coast of the Tiburon peninsula are most likely transient individuals, as are the ones feported from Anse-a-Galets on eastern La Gonave. A pertinent point concerning these last reports was that all the animals were 1.2-1.5 m individuals, the size class that would be dispersing the most according to the Messel model. Adult crocodiles will also move along the coast, resulting in large individuals being reported in small coastal wetlands. Most of the areas where crocodiles have been reported along the Tiburon peninsula contain more than 60 ha of habitat. Only a fraction of this area, however, is available to the crocodiles because of extensive fishing, rice cultivation, and other human activities. The habitat at the mouth of the Rivitre Capolo is much smaller, however; so small it is impossible to make a size estimate from 1:50,000 topographic maps. A similar situation is found on the southern coast of Jamaica where crocodiles may be found in almost any coastal wetland one hectare or larger in size (Plotkin and Faibairn MS, referenced in Groombridge 1982). This frequent usage of small coastal habitats suggests considerable movement by crocodiles along the coast. Human-related Mortality., Five direct causes of human-related crocodile mortality can be identified in Haiti (other»than habitat destruction): (1) incidental capture in fishing nets or traps, (2) malicious killing, (3) THORBJARNARSON: AMERICAN CROCODILE IN HAm 19 hunting for sport, (4) hunting for food, and (5) nest raiding. By far the most important of these is being trapped in fishing nets and traps. When caught, crocodiles either drown or are killed by the fishermen when pulled to the surface. Inmost cases the body is simply discarded in the water. Occasionally crocodiles may be eaten (see below) or the body may be disposed of by burial or by dumping·it at sea (see Folk Beliefs section below). In Bang Saumatre, where gill netting is not commonplace, one 1.5 m crocodile was drowned in a net during the 10-month period of time I spent there. As most bodies of shallow water in Haiti are heavily fished, this source of mortality probably is considerable. Malicious killings often occur in response to crocodiles taking livestock or, in one reported case, killing humans. A large crocodile, approximately 3 m long, was killed at Tete Source in Bang Saumatre after it had taken livestock on several occasions (goats and sheep). This crocodile was also considered a nuisance because it would take fish from gill nets, leaving gaping holes. The crocodile was caught using a baited hook and then beaten to death with a long stick. Large crocodiles also are killed occasionally when they accidentally wander into populated areas. Likewise, hunting for sport claims adult crocodiles in certain areas, although the incidence of this appears to have declined in the recent past. Throughout Haiti, the people are so poor that the number of firearms is very limited, being restricted in rural areas to certain local leaders and the military. Hunting excursions from Port-au- Prince were popular in the past, especially during the 1917-1934 United States Marine occupation (Steedman 1939, Cave 1952). Crocodile hunting was also a popular pasttime of the colonial French prior to the independence of Haiti (Descourtilz 1809). Use of crocodiles for food is limited to only two areas in Haiti, the Riviare Massacre and, to a much smaller extent, in Bang Saumatre. Both these regions border on the Dominican Republic, and the usage of crocodiles is undoubtedly due to a Dominican cultural influence (where crocodiles are eaten and their by-products used). Although the better educated people in Haiti will occasionally eat crocodiles, the vast majority of Haitians will not. Active hunting of crocodiles only occurs near the Rivitre Massacre, where they are caught with harpoons, set nets, cast nets, and baited hooks. Most of the hunting in this region centers on Lagon aux Boefs, with a smaller number caught in the river itself. The meat is eaten and the fat is used to render an oil to treat pneumonia and rheumatism. The only other account of crocodile by-products being used comes from the Cayes region. Although unconfirmed, three separate sources claimed that crocodile teeth are sometimes used for false teeth in people, the work being done in a hospital in Cayes. Nest robbing probably is not a major source of mortality, although nests may be dug up and the eggs left to die. In most areas, residents reported 20 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) nests as being very difficult to find and stated that like crocodile meat, they did not eat the eggs. The only area where eggs were reported eaten (other than the Rivitre Massacre region) was Etang Saumatre. Residents of the town of Ponds Parisien, on the southern shore of the lake, will eat them on occasion (F. Conway pers. comm.). One nest near Tete Source was excavated in 1983 and the eggs sold in the local market of Thomazeau (I. Lange pers. comm.). There is no present market for crocodile skins in Haiti, and hide hunting is non-existent. Interest was shown by an Italian firm during the 1950's when it approached Mr. E. Lempke of Port-au-Prince about obtaining crocodile hides. Nothing came of this, however, and little interest in commercial hide production has surfaced since. There is a tannery in Port-au-Prince that deals with reptile leathers (mostly lizard and snake skins from South America [J. Wilson pers. comm.]). The skin of the crocodile that was killed for taking livestock in Etang Saumatre was reportedly taken to this tannery. Until recently there also was a small export trade in live juvenile crocodiles for pets. This apparently has stopped in the last few years. Folk Beliefs., The fact that crocodile meat generally is considered inedible is undoubtedly one of the major reasons why there are still crocodiles in Haiti today. Such folk beliefs evolved with the culture of the Haitian people and are deeply tied to their religious beliefs. The indigenous Indians ate crocodiles, as is evidenced by the presence of crocodile bones in middens near the Caracol mangrove swamp (W. Hodges pers. comm.). Personal observation, however, has shown that today there is a widespread cultural taboo against eating crocodiles or their eggs. In most areas crocodiles are considered inedible and simply thrown away when killed; however, there is local variation in the method of disposal. In some areas (e.g. Gonaives), dead crocodiles are buried in a grave, often in coffin-like boxes with a grave marker. Around Etang Saumatre crocodiles are beheaded, with only the head being buried, usually away from the rest of the body. This represents an attempt to keep the spirit of the dead crocodile (which resides in the head) separate from the body so that it can create no "mischief' after death (pers. obs.). On the island of La Gonave crocodiles actually are considered poisonous. When crocodiles are killed in this region, they are disposed of by being weighed down with rocks and towed out to sea. Crocodiles are eaten only in areas where there is cross-cultural exchange with Dominicans. Around the RiviBre Massacre crocodiles are accepted as being edible. Near Bang Saumatre some Haitians also will eat crocodile, but it is considered poor repast, equivalent to the Cyclura iguanas in the area. An interesting example of cultural differences occurred at Las Lajas, a border military post in the Dominican Republic. A Dominican guard had shot a 2.7 m crocodile in Etang Saumatre, dragged it ashore and cut out the tail meat to eat and the penis (which is widely believed to be an aphrodisiac in the THORBJARNARSON: AMERICAN CROCODILE IN HAITI 21 Dominican Republic). Afterward, Haitians chopped off the head and buried it several meters from the carcass to prevent the crocodile's spirit (called its loa) from doing any harm. Crocodile Distribution in Relation to Habitat., Crocodiles are found only where there is suitable habitat. The habitat provides two important functions: an environment that meets the biological needs of the animal and, if it is to sustain a crocodile population, sufficient cover to protect the crocodiles from man. Because of the extremely dense human population in Haiti, the only significant coastal habitat type that satisfies both these criteria is mangrove swamp. In areas where human densities are considerably less, the second criterion obviously becomes less important. In historical times, crocodiles probably were commonly associated with virtually all coastal wetland habitats. The pattern of disappearance from these habitats is inversely correlated with the degree of human activity in the area. Mangrove swamps provide good habitat because they are relatively impenetrable to humans and offer a wealth of hiding places. Mangroves also are a very common form of tropical coastal wetlands, growing under a variety of physiognomic conditions and, as such, have been well documented as a preferred habitat of the American crocodile throughout its range (Alvarez del Toro 1974; Ogden 1978, Medem 1981). In Haiti, there is a correlation between the amount of mangrove habitat and crocodile presence in each of the coastal D6partements (a political subdivision) and the major satellite islands (Table 1). Coastal regions with crocodiles contain, on average, significantly more mangrove (0.390 sq km/km shoreline, as measured from 1:50,000 topographic maps) than areas without (0.05 sq km/km shoreline) (t-test, p < 0.05). The limited occurrence of crocodiles in coastal non-mangrove habitats emphasizes the need for sufficient cover in which crocodiles can hide. Although much of Haiti's coastal wetlands are dominated by mangroves, there are a number of other small estuarine environments that could support crocodiles. Most of these are on the north coast between Cap Hartien and Port-de-Paix, or on the southern peninsula (e.g. Cayes Plain rivers, Rividre Grande Anse near J6r6mie). Of the non-mangrove habitats, only two still have crocodiles, both of which offer sufficient cover for the crocodiles in them: the Rivitre Artibonite (deep, murky water) and the RiviBre Bondonne (extensive herbaceous vegetation). Even with cover, however, the presence of crocodiles in these areas probably is related strongly to the nearby occurrence of mangrove habitat. When considering the suitability of mangrove habitat for crocodiles, however, there are considerations other than total area that need to be addressed. Perhaps the most important are freshwater input and the availability of nesting habitat. As Dunson (1982) demonstrated, hatchling C acums can osmoregulate properly in water up to approximately 20 parts per 22 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) thousand (ppt), but cannot tolerate full strength sea water (35 ppt) for extended periods of time. Periodic access to areas of fresh or brackish water therefore is essential for hatchling recruitment (Mazzotti 1983). The apparent lack of crocodiles in the extensive mangrove swamp at Caracol may, in fact, be due to a lack of freshwater input into the region. One river, the RiviBre Trou du Nord, does empty into the mangrove adjacent to the town of Caracol (population 3982). The heavy human use of the area and a possible lack of suitable nesting habitat provide unsuitable conditions for crocodile reproduction and recruitment. The xeric nature of the surrounding area suggests also that periodic access to freshwater lenses formed by rainfall also is unlikely. In fact, in most of the mangrove areas in Haiti, freshwater influx is very low. The majority of mangrove habitat is located in the subtropical dry forest life zone, areas receiving less than 100 cm of rain annually. The lack of freshwater runoff also is related to the high coastline to total country area ratio. The amount of freshwater runoff is proportional to the amount of land that receives rain, but mangrove habitat is a function of the length of the coastline. Hence, Haiti, with a long irregular coast and a small surface area, provides little freshwater runoff for its coastal mangrove forests. This situation is further aggravated by the fact that much of Haiti is arid, and in several areas (such as the l'Ester region) freshwater runoff is diverted for agriculture. The fringe forest of the l'Ester historically had a much larger freshwater influx than at present, because of irrigation and channelization, and this region once supported a very large crocodile population (Descourtilz 1809). Diversion of freshwater probably had a significant effect on the habitat which, in combination with intense fishing activity and past crocodile hunting, has resulted in a drastic decline in the crocodile population. The situation on the island of La Gonave also is worth mentioning in regard to freshwater availability. No surface streams exist on the island; instead, rainfall percolates down through the limestone bedrock, emerging in springs, some below sea level. In this respect, suitable hatchling habitat may be severely limiting, being restricted to the vicinity of freshwater springs that feed into coastal lagoons. The area visited on the north coast of La Gonave did, in fact, have at least one small spring adjacent to a mangrove swamp in which juvenile recruitment was occurring. Crocodile Distribution in Relation to Human Population Density., Most human-related crocodile mortality in Haiti is accidental, therefore it can be assumed to be directly proportional to the frequency of human-crocodile encounters. This, in turn, is related to the prevalence of activities that bring people into crocodile habitat or to accidentally kill cr6codiles; these activities include fishing, cutting mangrove, or collecting mangrove oysters. Although it is virtually impossible to quantify such activities, they can be assumed to be THORBJARNARSON: AMERICAN CROCODILE IN HAITI 23 more or less directly proportional to human population density in the area. In Haiti, population densities are, in fact, higher in areas where crocodiles have been extirpated than in those regions which have extant crocodile populations (Tables 2,3). Because of the large variances involved, the differences are not significant for the coastal areas (Table 2, t-test p > 0.05). Inland lakes with crocodiles, however, are in areas of significantly lower human population density (Table 3, t-testp < 0.05). A noteworthy relationship exists between the status of crocodile populations and the ecological life zone in which they are found (Table 4). All historical crocodile populations were located in the two life zones that predominate in the lowland regions: the subtropical moist and the subtropical dry life zones (OAS 1972). The majority of the crocodile populations in the moist zone (100-200 cm rain annually) have been extirpated, only remaining today on the south coast of the Tiburon peninsula and Ile A Vache. On the other hand, only two populations in the dry forest zone have disappeared (Caracol region, L'Etang near Gonalves). In the latter cases the crocodile populations were probably vulnerable because of a lack of freshwater habitat, and small population size and ephemeral habitat respectively. As adult crocodiles are often found in freshwater and hatchlings require at least some freshwater, this relationship is not due to the availability of freshwater per se. Apparently the presence or absence of crocodiles in these areas is moderated indirectly through human population density, which is higher in the greater rainfall moist zone, leading to the unexpected result of finding crocodiles primarily in areas surrounded by semi-arid habitat. In the previous discussion, crocodile distribution in Haiti has been examined in relation to habitat availability and human population density. It is reasonable to assume, however, that crocodile distribution is affected simultaneously by both factors. Ranking coastal regions by population density and the amount of mangrove habitat, neither parameter alone is significantly correlated with the presence or absence of crocodiles (Wilcoxon Rank test,p > 0.05). A composite rank combining both parameters (Table 5), however, is significant (p < 0.05), indicating that the combination of mangrove habitat and population density is a better indicator of an area's suitability for crocodiles than either parameter alone. The extirpation of crocodile populations then appears to be synergistically related to human population density and the amount of suitable habitat. As crocodiles are rather long- lived, prolific, and secretive animals, given sufficient habitat offering retreats (mostly mangrove in Haiti), they can survive even in areas of dense human populations (as along the south coast of the Tiburon peninsula). This is due in no small part to their ability to move along coasts between pockets of suitable habitat. Human-related mortality is, however, quite high and can result in extirpation, especially where crocodiles are not afforded sufficient cover (non-mangrove habitats) or where they are vulnerable (e.g. lack of 24 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) freshwater). These generalizations, however, only apply to areas where mortality is mostly incidental, that is where there is no economic motivation for killing crocodiles. Where active hunting occurs for food, or more importantly for hides, crocodilian populations are much more likely to be extirpated. ECOLOGICAL STUDY: ETANG SAUMATRE Introduction General Features of the Region Etang Saumatre is Haiti's largest lake (113 sq km) and is located in the (Jul-de-Sac Valley approximately 30 km northeast of Port-au-Prince (Fig. 6). The Cul-de-Sac graben, which has been referred to as perhaps the most NHOW.r 'r .). etfpea DE NE\P'- f41 BAIE DE ~h PORT·AU-PRINCE / 0 '' TROU R ITANG CAI#UN '~ h~ SAUMATRE PORT-AU PRINCE WPJM- lAGO NRIQUILLO LAGUNA DEL UNCON • BARAHONA CARIBBEAN SEA ¥ 0 10 20 30 km FIGURE 6. Map of the Cul-de-Sac/Valle de Neiba region. THORBJARNARSON: AMERICAN CROCODILE IN HAITI 25 striking surface feature in Haiti (Woodring et al. 1924), is a low-lying valley extending east-southeast from Port-au-Prince Bay completely across the island to the Caribbean coast of the Dominican Republic, where it is referred to as the Valle de Neiba. Mountains border the valley to the north and the south, with Morne La Selle, Haiti's highest point (2680 m) lying directly south of Etang Saumatre. The abrupt changes in elevation provide a great diversity of ecological life zones within a relatively small area (Holdridge 1947). The Cul-de-Sac/Valle de Neiba depression contains a series of lakes, from east to west: Laguna del Rincon, Lago Enriquillo, Etang Saumatre, and Trou Caiman. Lago Enriquillo, in the Dominican Republic, is the largest (180 sq km) and most unusual of the four, as it is one of the lowest lakes in the world (35 m below sea level) and is hypersaline (50 ppt in 1981). Because of hypersaline lake water, the distribution of flora and fauna of the lake are restricted principally to the fringing freshwater habitats. Etang Saumatre (elevation 15 m) is located along the Haitian- Dominican Republic border approximately 10 km west of Lago Enriquillo; the two lakes are separated by a ridge of Pleistocene limestone and alluvial GLORE TETE SOURCE, N LAS DOMINICAN 0 LAJAS REPUBLIC HAITI ETANG SAUMATRE 0 1 2 3 4 km ' GAN THIER FONDS PARISIEN • MALPASSE FIGURE 7. Map of Etang Saumatre. 26 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) deposits. Although the vast majority of the lake lies in Haiti, two small sections (at Malpasse and south of Las Lajas) extend into the Dominican Republic (Fig. 7). Like Lago Enriquillo, Etang Saum9tre has no surface outlets, although the possibility of subsurface drainage into Lago Enriquillo cannot be ruled out. Etang Saumatre is only slightly saline, and because of this its ecology is quite different from that of Lago Enriquillo, in general supporting a much more diverse flora and fauna. Geology and Shoreline Features Within the last several million years the Cul-de-Sac valley was a shallow water marine strait that separated the Sierra de Neiba/Montagnes de Trou d'Eau to the north from the Sierra de Bahoruco/Morne La Selle ridge to the south (Woodring et al. 1924). This prior separation of Hispaniola into "north" and "south" islands still is reflected in the biogeographical relations of many taxa, as has been perhaps best documented for reptiles and amphibians (Schwartz 1980). Miocene 13 Qual'inory N 3·:'::,;: 01,gocene ~~~1 Quaion,ory *In,1 1 1 Alluviola Etang -SIS* Saumatre 01234k- Nflin'11'Mnmi~mp1I,-~ FIGURE 8. Geology of the Etang Saumatre region. THORBJARNARSON: AMERICAN CROCODILE IN HAITI 27 The abundance of Quaternary limestones composed of extant coral species is evidence of the recent marine inundation of the valley floor (Woodring et al. 1924). In many areas throughout the region coraliferous deposits are still recognizable, and in a few areas, such as Isla Cabritos in Lago Enriquillo, virtually intact exposed coral reefs can be observed. The placement of the lakes within the valley has been attributed by Woodring et al. (1924) to uneven alluvial deposition from sfreams draining the mountain watersheds to the north and the south. Areas that did not accumulate sediments are now depressions that have filled with water and remain as a series of lakes. The maximum known depth of the Etang Saumatre depression is 30 m below the lake's surface, or about 15 m below sea level. The shoreline features of the lake reflect the geology of the area (Fig. 8). The western lake margin, from Tete Source south to the vicinity of Ganthier, is composed of Quaternary alluvial deposits, creating a shallow lakeshore gradient. Continuing east along the south shore, Quaternary alluvials are mixed with similar-aged limestones creating a mosaic of shallow gradient and rocky, medium gradient shorelines. Just east of Fonds Parisien, older limestones (Miocene and Oligocene) form the northern flanks of the Morne La Selle ridge, and descend abruptly to the lake creating a rocky, steep gradient shoreline. A similar rocky coast is found along the north shore from Glore east to the vicinity of Las Lajas. However, along the north coast a number of prominent valleys, filled with recent alluvial deposits, extend down to the lake and create shallow gradient, non-rocky shoreline (coves). The eastern shore, northwest of Malpasse, is a medium gradient shoreline dominated by Quaternary limestone, containing several areas of shallow gradient mudflats at the openings of arroyos. Limnology Bang Saumatre is a brackish lake, with a salinity range of 8-10 ppt. In the shallow northwestern lake region where much of the freshwater input occurs, salinity is at the lower end of this range. The lake level iluctuates periodically, apparently as a function of rainfall. During the period 1979- 1983, following a series of hurricanes, the lake rose approximately 2 m. At the turn of the century, the lake level apparently was even higher than at present. Tippenhauer (1901) reported its elevation as 20 m above sea level (5 m above present), and Wells (1893) stated that the water was potable, being only slightly brackish. Similar fluctuations in lake level and salinity are known from adjacent Lago Enriquillo (La Fuente 1976). 28 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) A number of freshwater springs (conductivity 500-600 mohmos/cm) empty into the lake, primarily in the northwestern section of the lake, the largest being at Tete Source. Several small springs are found in the southeastern section of the lake near Malpasse, located directly along a fault, giving them a high hydrogen sulfide content. Although much of the freshwater input into the lake comes from these solution channel springs, a significant amount of water enters the lake in seepage areas and, to a lesser extent, through freshwater irrigation canals. Various chemical analyses have been done on Bang Saumatre lakewater since 1921 (Table 6). The fluctuations in salinity probably reflect changes in lake volume associated with variable rainfall. The high phosphate content of the water indicates the lake is eutrophic (M. Binford and M. Brenner pers. comm.). Bond (1935) termed Etang Saumatre eutrophic and "thalassohaline," meaning its ion concentrations are in the same relative proportion as seawater. Bond (1935) argued that this was evidence of the lake's marine origin. Climate The area surrounding the lake receives an annual average of 70-90 cm of rain. With a mean monthly temperature of 26 ' C, the region is classified in the subtropical dry zone, as defined by Holdridge (1967). Rainfall follows a predictable annual pattern (Fig. 9), peaking in May and October with a long winter (November-March) and a short summer dry season (June-July). Mean monthly temperature varies from 27.9 'C in August to a low of 23.7 ' C in February. Winds are predominantly from the east, and steady 18-36 kph winds were not unusual. Under these conditions whitecaps cover the lake, with wave amplitudes in excess of 1 m. The less common westerly winds were usually associated with rainfall. Vegetation Terrestrial Vegetation., The riparian vegetation is a virtual monoculture of buttonwood mangrove (Conocatpus erecms) growing in fringe 5-20 m wide around the lake where there is sufficient soil. Because of the recent rise in lake level, the Conoca,pus frequently extended out into shallow water to a depth of approximately 1 m. The most prolific Conocarpus stands are found along the shallow gradient alluvial shorelines on the western shore between THORBJARNARSON: AMERICAN CROCODILE IN HAITI 29 -300GANTHIER MEAN MEAN50 -100 MONTHLY MONTHLY TEMPERATURE RAINFALL ('C) 30 60 (MM) 10- 11 JAN MAR MAY JUL SEP NOV MONTH FIGURE 9. Walter diagram of rainfall and temperature variation in Thomazeau, adjacent Etang Saumatre. Shaded areas represent the dry season (after Walter 1973). Tete Source and Ganthier, and on the western shore between Las Lajas and Malpasse. Most of the Conocatpus is of low stature, rarely exceeding 4 m in height. The upland areas immediately surrounding the lake are vegetated with a xeric Acacia, Prosopis, and cactus association (Appendix II). Aquatic Vegetation., The freshwater flora is restricted to a few springs, canals, and freshwater seeps around the lake. In these areas the dominant vegetation consisted of grasses (Echinochloa cnisgalli, Paspalidium geminatum), sedges (Eleochan's cellulosa, Scitpus pungens), floating vines (Ludwigia leptocarpa, L. repens, Enhydra sessilis, Commelina geniculata), submerged plants (Najas manna, NiteUa sp., Saggitaria sp.), and cattails (7*ha domingensis). The flora of the lake itself is rather limited, being restricted to several halophytic algae and a few vascular plants. Shallow water sublittoral areas usually supported dense bottom mats of algae (Chara homemannii and Batophom oersteadi). These were the most productive areas of the lake and supported large numbers of fish (mostly Tilapia). Another alga (Cladophora 30 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1) sp.) frequently was found growing attached to Conocarpus roots. Ruppia mantima forms dense mats in shallow water in several parts of the lake, usually adjacent to freshwater seeps. The grass Halodule beaudettei was often encountered along shores or in shallow water areas on mud or mari, usually under Conocarpus. In the southeastern section of the lake near Malpasse, Salicomia perennis grows along the lakeshore and out into shallow water. Vertebrate Fauna Fish., West Indian freshwater fish communities generally are lacking in diversity (Myers 1937). Etang Saumatre only supports eight native species (Appendix II), three of which (Strongylura nomta, Gobionellus sp., and Dormitator maculatus) typically are marine-coastal dwellers, reflecting the marine origin of the lake. In terms of biomass the dominant native species are Cichlasoma hatiensis (in the lake and spfings) and C>prinodon bondi Cake only). The other native fishes all are poeciliids, the two genera (Limia and Gambusia) being characteristic members of Hispaniolan freshwater fish communities. Two species of introduced fish also are found in Etang Saumatre: Tilapia mossambica and carp ((*tinus calpio). Both were first introduced into the lake during the early 1950s as part of an FAO fisheries pfoject. The Tilapia have become quite abundant and probably are the dominant fish overall in the lake in terms of biomass. The carp, on the other hand, are rarely seen and almost never caught by fishermen. As carp are long-lived fish and generally require cool water conditions for spawning, it is possible that no reproduction has occurred in the lake, and the few seen are remnants of the introduced stock. Reptiles and Amphibians., Aside from crocodiles, the only aquatic reptile found in Etang Saumatre is Trachemys decorata, a little-known freshwater turtle restricted to the Cul-de-Sac/Valle de Neiba region. Very secretive in their habits, the turtles were rarely seen, with most observations being restricted to shallow water areas in or around Conocatpus. No amphibians were f6und in the lake itself although several anurans (Bufo gunthed, B. marinus, and Osteopilus dominicensis) bred in freshwater areas adjacent to the lake. Avifauna., Aquatic and semi-aquatic birds are quite common in the lake, especially during the winter when there is an influx of migrants from North America. The aquatic and semi-aquatic avifauna of the lake is listed in Appendix II. Mammals., No native mammals are found in the Etang Saumatre area. The only local mammals (besides domestic stock) were rats (Ratms ratms, R. THORBJARNARSON: AMERICAN CROCODILE IN HAITI 31 norvegicus), mice (Mus musculus), feral cats (Felis domesticus), and mongoose (Herpestes auropunctatus). Demography Population Size, Sex Ratio, and Size-class Distribution Based on the corrected survey data the total population size in Etang Saumatre was estimated to be approximately 450, including crocodiles of all size classes (Fig. 10). Because sexing juvenile crocodiles is difficult, much of the juvenile sex data was not used, so sample sizes are small (Table 7). The predominance of males in the two smaller size classes is noticeable; however, it is not significantly different from 50% (chi square, 0.25
2.7 m). This may be
a slight overestimate as sexual maturity is not attained until 2.2-2.3 m total
length (see Reproductive Ecology). Nevertheless, the great majority of the
animals seen in the 1.8-2.7 m size-class were 2.3-2.5 m long, and the
overestimate of reproductive animals is probably quite small. Only 10% of
the population was in the subadult (0.9-1.8 m) size-class.
The small number of subadult crocodiles appears to be a general
characteristic of most crocodilian populations. Cott (1961) commented on
this mysterious "disappearance" of small- and intermediate-sized crocodiles
in Africa, quoting Pitman as it being "a conspicuous feature of African inland
waters." Messel et al . ( 1981) noted a similar lack of subadult C porosus in
northern Australia, as did Mazzotti (1983) for C acums in Everglades
National Park. This apparent lack of subadults may be a result of: (1) rapid
growth of young crocodiles, (2) extremely secretive behavior or occupancy of
marginal habitats, or (3) high juvenile mortality. In most populations these
factors are not mutually exclusive and may all play a role. Indeed, the last two
32 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1)
76.9
AUGUST 1983
N=433
100
10.4 76 5.1
79.2
N
U
M
B
E
R
O
F
C
R
O
C
O
D
IL
E
S 300
JANUARY 1984
N=413
100
11.67.0 2.2
74.3
300
COMBINED
N =44 7
100
5.010.0 1 .
over0.3-0.9 ' 0.9-1.8 1.8-2.7 2.7
TOTAL LENGTH (m)
FIGURE 10. Size-class distribution and the total population size estimated from the
August 1983, January 1984, and combined survey data.
THeRBJARNARSON: AMERICAN CROCODILE IN HAm 33
are major components of the population dynamics model proposed by Messel
et al. (1981) for C porosus.
In Etang Saumatre, rapid growth does not appear to be a factor (see
following section). Subadults, however, do have a tendency to be found in
more "marginal" habitats (see Habitat Selection), and this may have resulted
in a slight underrepresentation of these crocodiles in the population surveys.
The possibility of movement of subadults out of Etang Saumatre also cannot
be ruled out. Local residents have reported crocodiles moving between Trou
CaTman and Bang Saumatre, especially during periods of heavy rain.
Crocodiles also have been reported crossing the arid strip between Lago
Enriquillo and Bang Saumatre U. Ottenwalder pers. comm.). In both cases,
however, the animals seen were most often adults; the incidence of migration
of subadults remains unknown.
If we assume that the population in Bang Saumatre has a stable age
distribution, we then would be left with high juvenile mortality as the only
explanation for the small number of subadults. While this mortality probably
plays an important role, recent past events in the lake also must be
considered.
Prior to 1979 the lake was approximately 2 m below its present level. In
many areas one can still see Conocatpus snags standing in water of this depth.
Subsequently, three hurricanes hit Haiti in 1979 and 1980, causing the lake to
rise to its present level and flood out into Conocatpus habitat over 41% of the
lakeshore (Conocatpus fringe and Conocatpus flats habitats; see Habitat
Selection). Unless exposed to considerable wave action, these areas were
ideal habitat for young crocodiles, hiding by day in the Conocarpus root mats
and emerging at night to feed on the abundant fiddler crabs and other
invertebrates (see Diet). Although the amount of suitable juvenile habitat
before the hurricanes is unknown, it almost certainly was considerably less
than it is today. Aerial photographs of the lake when it was at a similarly low
level (Bond 1935) show considerable stretches of barren shoreline.
Furthermore, prior to 5-6 years ago, the major nesting areas on the
eastern shore were the sites of temporary human settlements. People lived
on the eastern lakeshore, cutting Conocatpus, making charcoal, and fishing,
bringing freshwater across the lake from Fonds Parisien by boat. Remains of
old thatch "ajoupas" were found in the middle of several of the nesting
beaches. Nests were robbed and the eggs eaten (F. Conway pers. comm.) by
residents of Fond Parisien. Undoubtedly, juvenile crocodiles also were killed.
Since that time movement across the lake by boat has been outlawed by the
Haitian government because of smuggling across the Dominican border, so
no one lives near the nesting beaches.
Consequently, because of reduced habitat availability and increased
human disturbance for a period of time prior to 1977-1980, hatchling
production and survivorship of juvenile crocodiles may have been
34 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1)
considerably lower than today. Based on growth rates of recaptured
crocodiles (see Growth Rates), the ones in the 0.9-1.8 m size class would have
been born during the period 1977-1980, and recruitment into this size-class
would undoubtedly have been adversely affected by these factors. Thus it
seem5 that both natural mortality and past changes in the level of human-
related mortality and juvenile habitat availability may be factors contributing
to the low observed proportion of subadults.
Density and biomass
Few data exist on any aspect of crocodilian population dynamics. This is
especially true for C acutus where the lack of previous studies is due, at least
in part, to the difficulty of accurately censusing these animals in their typical
habitat (i.e. coastal wetlands). For comparative purposes, however, some
information is available for several other crocodilians species, most notably
Crocodylus niloticus, C. porosus, and Alligator mississippiensis. Density of
crocodiles in Etang Saumatre was calculated on a linear basis (per km
shoreline), as crocodiles are primarily littoral animals and generally do not
move far from shore. Therefore, in [acustrine or riverine habitats, density is
best described in this fashion. Based on the corrected survey data, the mean
density of crocodiles in Bang Saumatre was 6.3/km (crude density, 71.2 km
shore). Eliminating habitats unsuitable for crocodiles (rocky or high wave
energy shores), the ecological density was 9.6/km (46.7 km shore).
Density estimates for various crocodilians are presented in Table 8. The
high degree of variability among these values undoubtedly results from a
variety of factors such as: physical habitat structure, vegetation, water depth,
degree of wave exposure, aquatic productivity and the availability of food,
population structure, and even terrestrial habitat (inasmuch as it determines
the suitability of the area for nesting). Little is known quantitatively about
the roles these various factors play in determining population density.
Crocodile density in Bang Saumatre is variable depending on habitat type
and the degree of exposure to wave action (see Habitat Selection). Wood
and Humphrey (1983) found that density in Alligator mississ*iensis was
correlated to lake productivity (nitrogen concentration). Anecdotal accounts
relate food availability to crocodile density (Montague 1983, Watson et al.
1971) or vice versa (Cott 1961, Fittkau 1970, 1975, Whitaker 1978, Glastra
1983), suggesting that in some instances crocodilians play a beneficial role in
maintaining healthy fish populations.
Although the factors mentioned above determine the attainable density,
or carrying capacity of a particular habitat, in reality actual densities are
usually held well below this value by human predation. Several hunted and
THORBJARNARSON: AMERICAN CROCODILE IN HAm 35
5.0 -
7.3.22*-2.944.0 -
r =0.998
LO
G
M
A
S
S
(g
)
3.0 -
-
2.0-
11I lili
1.4 1.8 2.2 2.6
LOG TOTAL LENGTH (cm)
FIGURE 11. Log-log plot of the length-mass relationship of crocodiles captured in Etang
Saumatre. A maximum of 5 randomly chosen values are plotted for each 20 cm size class
interval.
36 BULLETIN FLORIDA STATE MUSEUM VOL. 33(1)
non-hunted populations are included in Table 8. At the time of censusing,
the non-hunted Crocodylus niloticus populations probably existed in a more
or less undisturbed state, and densities in these areas are quite high (13.1-
21/km). Furthermore, within each of these areas densities were considerably
higher in favorable habitats. In parts of Lake Turkana, for example, densities
reached 55.8/km (Graham 1968). The highest reported value for A.
mississ