2°“ .IKWIWIHUWWHWIUIIHIWHWIW||Hl|WHI N4 _cncooo WWWIIIHWWWII 293 1074-4 8213 '— ~— ." W “5.14.? .31, If; Y Michigan state University This is to certify that the thesis entitled Characteristics of a Crocodile Population in Papua Nenguinea presented by J. Jerome Montague has been accepted towards fulfillment of the requirements for Master's degree in Fisheriefl Wildlife gym Major professor Date fll(-?/ 0-7 639 OVERDUE FINES: 25¢ per do per item RETURNING LIBRARY MATERIALS: , . _‘ Place in book return to move 0:135, . charge from circuhtton records CHARACTERISTICS OF A CROCODILE POPULATION IN PAPUA NEW GUINEA BY J. Jerome Montague A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE . Department of Fisheries and Wildlife 1981 ABSTRACT CHARACTERISTICS OF A CROCODILE POPULATION IN PAPUA NEW GUINEA BY J. Jerome Montague Between October 1978 and July 1980, day and night crocodile surveys were conducted at low, normal and high water levels over 2703 kilometers of river and lakeshore in the Western and Southern Highlands provinces of Papua New Guinea. The 1353 crocodiles seen ranged between 0.18 and 2.23 animals per linear kilometer. Night counts tallied 12.9 times as many crocodiles as day counts over the same area. This may indicate more nocturnal activity than occurs in other crocodilians. It was found that night counts from one bank of a river could be safely doubled to determine a "both bank" count. Unhunted areas had a greater proportion of large crocodiles than did hunted regions. As the water level rose, the "visible juveniles" category decreased as a result of migration into newly- flooded adjacent swamplands. The upstream range of New Guinea crocodiles on the Strickland River drainage extended only to the Burnett River junction. Large crocodiles were more wary, having a greater flight distance than small ones. There was an increase in flight distance with increasing body size and this rate of increase was over twice as great for hunted populations as for unhunted ones. A ban on crocodile hunting was recommended only for those rivers with a density index under 0.5 crocodiles/kilometer. Saltwater cro- odiles were scarce in the Lake Murray district, and it was recommended further that the sale of live animals and skins be prohibited there. Dedicated to my parents Charles and Irene ii ACKNOWLEDGEMENTS This study was made possible by the cooperative efforts of the Food and Agriculture Organization of the United Nations (FAO), the United Nations Volunteers, the Papua New Guinea Wildlife Division, and the Department of Fisheries and Wildlife at Michigan State University. Deepest appreciation goes to my major advisor at Michigan State University, Professor George A. Petrides, who consulted with me in the field and gave much editorial guidance. I am especially grateful, too, to Mr. Eric Balson, Wildlife Manager for the Papuan Region, for his guidance and encouragement. Mr. Sawaku Kawi and his staff at the Baboa Crocodile Station gave generous assistance with the field work.‘ I wish also to express my gratitude to Mr. Melvin Bolton, Dr. Antoon deVos, Mr. Alistair Graham, Mr. Philip Hall, Mr. Miro Laufa, Mr. Kevin McGrath and Mr. Romulus Whitaker of the National Crocodile Project for their frequent advice. Mr. George Craig was helpful in providing a firsthand account of the status of crocodiles in the Lake Murray district in the 1950's and 1960's. I would like to thank Drs. Rollin Baker and Robert Deans for their kind assistance as members of my university graduate committee. I wish, further, to express my appreciation to Maj. Roger Chapman, SSgt. Dave Weaver, Mr. William Neumeister, and Mr. Yogandra Thami of the Operation Drake Strickland Expedition for including me on that expedition and allowing me to conduct surveys at Lake Kopiagu and on the Upper Strickland River. The Royal Australian Air Force kindly provided helic0pters for a portion of the Strickland River survey. iii Special thanks are extended to Dr. Jeffery Lang of North Dakota State University for his helpful assistance in field counts and camp- fire discussions. For her constant support and encouragement during some difficult periods, I am greatly indebted to Miss Rene Lucha. iv TABLE LIST OF TABLES . . . . . . . . . LIST OF FIGURES. . . . . . . . . INTRODUCTION . . . . . . . . . . Study Area. . . . . . . . . . METHODS. . . . . . . . . . . . . RESULTS 0 C O O I O O O O O O 0 Description of the Strickland Distribution and abundance. . OF CONTENTS Size and age composition vs. hunting pressure Water level effects . . . . . Habitat use . . . . . . . . . General preferences . . . Preferences related to body size . . Preferences related to water level. . Wariness. . . . . . . . . . . CONCLUSIONS 0 O O O O O O O O O 0 Distribution and abundance. . Habitat Use . . . . . . . . . Wariness. . . . . . . . . . . RECOMMENDATIONS. . . . . . . . . LITERATURE CITED . . . . . . . . Page vi viii 11 11 ll 16 16 26 26 29 29 37 37 40 42 43 44 Table LIST OF TABLES Page Location and description of the waterways surveyed. Southern Highlands and Western Provinces, Papua New Guinea, 1978 - 1980 . . . . . . . . 3 Numbers of crocodiles seen during low-water night counts on the unhunted gorge, upper and middle portions of the Strickland River. Western and Southern Highlands Provinces, Papua New Guinea, 1979. . 13 Numbers of crocodiles seen during low-water night counts on hunted areas. Western Province, Papua New Guinea, 1978-1980 . . . . . . . . 15 Summary of all low-water night crocodile counts. Western and Southern Highlands Provinces, Papua New Guinea, 1979 - 1980 . . . . . . . . 17 Numbers of crocodiles seen during normal—water level night counts on hunted areas. Western Province, Papua New Guinea, 1978 - 1980 . . . . . . 20 Numbers of crocodiles seen during high-water night counts on hunted areas. Western Province, Papua New Guinea, 1978 - 1980 . . . . . . . 22 Summary of all the crocodiles seen during all night counts. Western Province, Papua New Guinea, 1978 - 1980. . . . . . . . . . 23 Percentages by size-class of crocodiles and belly slides as determined from low-water day counts on seven rivers. Western Province, Papua New Guinea, 1978 - 1980. . . . . . . . . . 25 Habitat locations of different size crocodiles for all (47) night-light surveys. Western Province, Papua New Guinea, 1978 - 1980 . . . . . . 28 vi Table 10. 11. Habitat locations of different size crocodiles from low-water surveys on the unhunted upper and middle Strickland River areas, Western Province, Papua New Guinea, 1978 - 1980 . . . . Habitat locations of different size crocodiles from night counts at each of three water levels on each of seven rivers. western Province, Papua New Guinea, 1978 - 1980 . . . . vii Page 31 32 LIST OF F IGURES Figure Page 1. waterways of the study area in the Southern Highlands and Western Provinces of Papua New Guinea. . . .. . . . . . . . 4 2. Relationship of belly width (Lever, 1975) to total length based on measurements of 500 New Guinea crocodiles captured in Western Province, Papua New Guinea, 1979 - 1980 . . . . . . 18 3. The relative abundance of crocodiles during each of three water—level periods on seven rivers as based upon (A) night counts (N - 484 crocodiles) and (B) day counts of belly slides (N - 255 slides). Western Province, Papua New Guinea, 1978 — 1980. . . . . . . . . . 21 4. Relationship between the number of crocodiles sighted by day and night on seven rivers. Western Province, Papua New Guinea, 1978 - 1980 . . . . 24 '5. Proportion of crocodiles which were sighted on the right banks during 23 night counts (534 crocodile sightings) on seven rivers. Western Province, Papua New Guinea, 1978 - 1980 . . . . . . . 27 6. Percentages of crocodiles sighted in each habitat location in relation to water level. Data are from three surveys on the same portions of seven rivers. Western Province, Papua New Guinea, 1978 - 1980 . . 30 7. Relationship between crocodile size and flight distance on unhunted (solid line, N = 241) and hunted areas (dashed line, N = 389). Western Province, Papua New Guinea, 1978 - 1980 . . . . . . . 33 8. Average flight distance for different size crocodiles from unhunted (clear, N = 241) and hunted (shaded, N = 389) areas. Western Province, Papua New Guinea, 1978 — 1980. . . . . . . . . . 35 viii Figure 9. Page Relationship between the percentage of crocodiles approached to within 1 meter or less ("sure kills") and the animal's size on hunted (solid line, N = 193 crocodiles) and hunted regions (dashed line, N = 332 crocodiles). Western Province, Papua New Guinea, 1978 - 1980. . . . . . . . . . 36 ix INTRODUCTION Due to habitat destruction and excessive exploitation, many countries with crocodile populations have restricted or banned hunting and/or trade in crocodiles. In Papua New Guinea (PNG), however, the harvest of crocodile skins is not only a major but somewhat-expanding industry. In PNG, the eastern half of the large island of New Guinea, vast areas of prime crocodile habitat are considered to be too remote to be subject to serious harvest pressure (Medem, 1976). This could be assumed to be especially true of the 96,400 km2 Western Province where there are only 65,000 people and 90 km of permanent roads (Ford, 1973). One can fly over hundreds of kilometers of rivers, swamps and lagoons there without seeing even a dwelling. The PNG pOpulation is judged to be between 100,000 and 200,000 crocodiles (Pooley, 1976) and the annual harvest of skins is estimated to be between 25,000 and 50,000 (IUCN, 1978). In many villages, crocodile hunting is the only source of cash income. Of the crocodiles harvested, 75-95% usually are the New Guinea or freshwater crocodile (Crocodylus novaeguineae) with the remainder being the saltwater crocodile (E;_porosus) (Lever, 1975; Whitaker, 1980). Despite its economic significance, the New Guinea crocodile was discovered only 50 years ago (Schmidt, 1928) and is one of the least known of the world's 20 or so crocodilians. It is believed to be restricted to the main island of New Guinea (Lever, 1975; Whitaker, 1980). 1 Neill (1946) has stressed the need for research into the species' population dynamics and habitat preferences. In addition, Pooley (1976) has emphasized that it is essential to test the assumption that the current harvest is from an adequate population. A crocodile hunting and skin industry is likely to be a part of Papua New Guinea's commerce for years to come. A Food and Agri- culture Organization of the United Nations (FAO) project PNG/74/029 "Assistance to the Crocodile Skin Industry" was established to guide exploitation over the period 1977-1981. Although it has been stated (Gore, 1978) that the New Guinea crocodile is the only crocodilian not in danger of extinction, this opinion has not been verified nor has basic population data been determined. If a proper management policy can be established for crocodiles in New Guinea, it is felt that this will be a contribution to croc- odilian conservation programs everywhere. The present study was conducted from 22 October 1978 to 18 July 1980 in the Southern Highlands and the western Provinces of Papua New Guinea. It was designed to determine: 1) the pOpulation distributions and density indices of both species of crocodiles, 2) their age and size structure, 3) the extent of habitat utilization at various water levels, and 4) the effects of hunting pressure on flight distance with guidelines for crocodile management. StudygArea The areas surveyed (Table 1) were located between latitude' o o . o n o n 5 l9"S and 7 58"S and between longitudes 140 55 E and 142 34 E on Lake Murray, the middle Fly River and its major tributaries including the Strickland River (Figure 1). Lake Kopiagu, the highest point 3 Table 1. Location and description of waterways surveyed, Southern Highlands and Western Provinces, Papau New Guinea, 1978-1980. Waterwa Location Bank Water Levelsz y vegetationl low normal high Agu River 141°7'E, 7°5'S* to savanna x x x 74 km upstream swamp forest Aiema River 142°E, 7°6'S* to rainforest x 30 km upstream Boi River 141°25'E, 6°50'S* to rainforest x x Damami River Fly River, middle Fly River, lower Herbert River June River Kaim River, upper Kaim River, lower Lake Kopiagu Lake Murray 22 km upstream 142°4'E, 6913'S* to rainforest x 6 km upstream 141°7'E, 7°5'S to savanna x x 141°23'E, 7°35'S l4f’23'E,7°35'S to savanna x 14f’52'E,7058'S rainforest l4I°34'E, 7°20'S* for savanna x x entire length 141021'E, 6°20'S* to savanna x x 51 km upstream rainforest 62 to 104 km upstream rainforest x 141°32'E, 6°54'S* to savanna x x 62 km upstream rainforest 142°32'E, 5023' S (center) (4 km around lake) croplands x 141°30'E, 70$ (center) (87 km N.W. shore) swamp forest x Leva River Mamboi River Nomad River ** Rentoul River Strickland River, Gorge Strickland, upper Strickland, middle Strickland, lower Tomu River 141°36'E, 705 to 20 km upstream 14192’E, 7°10'S* to 44 km upstream 6 km upstream swamp forest x swamp forest x 142°8'E, 6°19'S* to rainforest x 142014 'E, 6018's 142°4'E, 6°21'S* to rainforest x 142°8'E, 6°19'S 142°19'E, 5019's to rainforest x 142°10'E, 5047's ‘ 142°10'E, 5047's to rainforest x 142°5'E, 6038's 142°5'E, 6°38'S to rainforest x 141°34'E, 7°20'S .141034'E, 7020's to gallery x x 141°23'E, 7035's rainforest 142°7'E, 6°37'S* to rainforest x * Coordinates of the river mouths. **Day counts only. After Paijmans et al., 1971. 2 at the time(s) of (mostly nocturnal) surveys. .mmCAzo 302 momma cuonuoom on» :a mono >©sum may no m>m3umu~3 .H ouomwm MO WTUCH>OHAH Chmummha UCM wvcmflnmflc . i... 2: J l sol: " r o I .0 0’ . . ... x U 3 N 1 V n d C m a c . I .11. d a g . .I O I o N a . u v 9 Iooau . M. y o u 32.30 Bu: .5... . oaoa studied, was situated 1349 m above sea level. The Strickland River Gorge, at the highest point surveyed on the main river was 370 m above sea level. Ranging in width from 1 m in the Kagwezi Chasm, the Strickland was over 800 km wide in its lower reaches. Characterized by rapids and falls at higher elevations, its average depth was 9 m and flow rate 6 knots (Paijmans et al., 1971). Both the Strickland and the Fly carried considerable sediment loads, but the former was the more heavily burdened. Lake Murray was 750 km2 in area and irregularly shaped (Figure 1). Lying roughly at the center of the study site, it averaged 7 m deep. This depth fluctuated up to 3 m between most wet and dry seasons and up to 5 m during 7-year drought years (1958, 1965, 1972, 1979) (Roberts, 1978). The area of Lake Murray in the wet season may be enlarged 5 times (Wheeler, 1979). Surprisingly, among world streams only the Amazon and the Congo are greater in volume than the Fly (Roberts, 1978). The Fly's width ranged from 250 to 920 m in the study area and had a 2-3 knot speed. Its bed was only 19 m above sea level at a point 800 km from the mouth. Both the Strickland and the Fly flooded 2 m or so above their banks during the wet season. Their water temperatures ranged from 21.7OC in the Strickland Gorge to 31.00C in Lake Murray. ' The local climate (Paijmans et al., 1971) is mild tropical with a mean annual temperature of 26.700, a diurnal temperature range of 8.3OC, and a mean monthly variation of only 2.20C. Maximum tempera- tures of 31-350C occurred daily in the early afternoon. Rainfall at Lake Murray averages 325 cm/year but could be as high as 500 cm/year on the upper Strickland River. Rainfall is greatest during the somewhat- warmer and less-humid January to June period. Relative humidity is high, with monthly means ranging from 80% in the wet season to 90% during the dry. During 1979, the sun shone 342 days (recorded at Baboa by the crocodile station manager) for an average of 4.5 hrs/day. That average was 3.9 and 5.2 hrs/day for the January - June and July - December periods respectively. The soils of rivers and swamps were alluvial clays while those of extensive floodplains were high in organic matter, with sloppy alluvial clay. ~All of the study area below the Tomu River on the Strickland River was without visible rocks or stones. The substrate on the upper Strickland, near the beginning of the gorge, was mainly granite or limestone. The vegetation in the upstream Fly River areas, north of the bulge toward Irian Jaya and upstream from the Herbert/Strickland junc- tion on the Strickland River, was open rainforest-woodland. The swamp regions there near the river were characterized by dense sago palm (Sage sage) stands. South of those points, the vegetation was primarily marsh-grass (Gracilius indicus) and wild sugar cane (Saccharum robustum) with thin strips of swamp forest and water-logged Open rainforest. The back swamps (flooded areas away from the river) were mostly Melaleuca savannas (Brandes, 1929; Paijmans et al., 1971). Exploratory expeditions in the past reported that nocturnal sightings and/or signs of crocodiles were plentiful on the middle Fly River drainage but few crocodiles were seen during the day (Everill, 1885; Bauerlen, 1886; Hurley, 1924; Hides, 1939; Archbold and Rand, 1940). Commercial hunting for crocodile skins began in 1948 and by the mid 1950's was a firmly-established industry in the area. Most hunting was by the indigenous pe0p1es who traded skins with expatriate skin-buyers for salt and cartridges. Until 1966, the area supplied 5,000 - 7,000 skins per year, primarily adults of both species (Bustard, 1968; Whitaker, 1980). After that year, crocodiles became scarce both on the middle Fly River and in Lake Murray (Neill, 1971; Lever, 1975; Craig, personal communication). METHODS All surveys except for Lake Kopiagu and the gorge, upper and middle sections of the Strickland River, were made from one of two square-ended flat-bottomed boats, either a 5.6 m aluminum craft with a 35 hp outboard motor or a 6.6 m wooden barge with two 25 hp outboards. Lake Kopiagu was investigated from a hand-paddled 5 m dugout canoe. The Strickland gorge was explored using two 5 m inflatable white-water boats with steerage oars. The upper and middle Strickland counts were made from a 4 m inflatable runabout with a 25 hp motor. In addition to studies by boat, a portion of the upper Strickland River was surveyed from Iroquois helic0pters. Studies were carried out at high, normal and low water levels which roughly corresponded to the months of February - April, June - August, and September - December, respectively. The night surveys began a half-hour after complete darkness. Since wave size was found (Woudward and Marion, 1979) to be negatively correlated with night sightings for the American alligator (Alligator mississippiensis), 8 surveys were not conducted on windy nights. At the beginning of each survey the name of the observer(s), operator, boat and motor, starting location and time, water level, and weather conditions (moon brightness and/or cloud cover, rainfall, air and surface water temperature) were recorded. Upon completing a night count the stopping location, time, and distance covered were recorded. The tapetum of the crocodilian eye reflects bright red when shined with a light at night (Chabrek, 1966; Whitaker and Whitaker, 1978; Woodward and Marion, 1979) . When their eyes are exposed, this permits crocodiles to be tallied. The light used for all but one night count was a 12-volt, 100 watt, hand-held, sealed-beam light powered either ' by a 12-volt wet-cell battery or a Honda EM 300 portable generator. Lake Kopiagu was surveyed with a 6-volt headlamp. The lZ-volt light was effective to 300 m, while the 6-volt's range was limited to 75 m. The spotlight illuminated the watercourse for navigational as well as counting purposes. 0n rivers 100 m or less in width, the survey craft was operated in midstream and the light was oscillated in front of the boat from bank to bank in an arc of approximately 180°. 0n the rivers over 100 m wide and on lakes, the boat was operated 30-50 m off one shore and the light was oscillated in a 120° arc from mid-river to the near bank perpendicular to the boat. A rather constant cruising speed of 18-22 km/hr was maintained. The observer sat in the bow of the boat behind a safety rope, while the boat-handler sat in the stern near the motor(s). The light was held ahead of the boat's bow in the hope that this would prevent the craft from becoming visible to the crocodile. The majority of all surveys were made with the same boat-handler/observer combination, as 9 was suggested by Chabrek (1966) and Goodman and Marion (1979). It was necessary for the observer to wear eye protection due to the numbers of flying insects. When a crocodile was spotted, the observer vibrated the beam at the crocodile's location as a signal to the operator. Without stopping or reversing the boat, the animal was then approached as closely as possible, following an are that went first next to the shore and then extended toward mid—stream. This pattern was followed in order to prevent having to reverse back out into the mainstream after each sighting. When a crocodile was spotted, the observed dictated on a small tape recorder the time of the sighting, the side of the river, the flight distance, the species of crocodile if possible, the animal's size and the habitat situation. The flight distance has been described (Heathwole, 1968; Bustard, 1968; Webb and Messel, 1979) as the distance a potential predator (here, the observer and boat was substituted) could approach a prey before the latter flees. Salt-water crocodiles were distinguished from New Guinea crocodiles by their sleeker appearance, lighter color, sharper saute-crests and lack of ossified scutes on the dorsal portion of the neck between the nuchal rosette and the skull plate (Neill, 1971; Lever, 1975; Whitaker, 1979 and 1980). Field estimates of a crocodile's size are based on the distance between the eyes and the snout tip (Chabrek, 1966; Graham, 1968; Messel et al., 1977). Since it was the standard and familiar reference used in PNG, the size was recorded in classes of inches belly—width (b.w.) (Lever, 1975). The belly-width size categories used were: 4" and under (hatchling), 5-6", 7—9", 10-12", 13-15", 16-19", 20-24", 10 and greater-than-24". The ranges of the belly-width categories were progressively increased for each larger size class because the possi- bility of mistaking the size of a crocodile is greater with bigger animals (Messel, 1977). Those above 20" b.w. were considered to be adults (Tago, 1977) and their maximum known belly width was 34" in the wild. An "eyes only" (e.o.) category was given to crocodiles that submerged before being classified or which were in shallow water (designated "5") or among obstructions ("0") which prevented close approach. The habitat location upon initial observation was classed according to a system modified from Messel et al. (1978): (l) M§_(mid-stream), where crocodiles were well out from shore in water so deep that their limbs could not contact the substrate, (2) SW9§_(shallow water on edge), where the animals were in open water near shore and presumably could touch bottom, (3) 9§_(on bank), where they were on bare soil between the water's edge and the line of vegetation, (4) I! (in vege- tation on shore), (5) $33!.(in emergent vegetation in the water), (6) ILIW_(in logs in the water), where crocodiles lay among dead wood. Crocodiles seen feeding or exhibiting unusual behavior were so noted. Day counts which were conducted the day before the night counts, were similar to nocturnal surveys except that animals were not approached. If a crocodile was spotted, the time, species and belly width were recorded. Where crocodile "slides" were seen on the bank, the belly- width was estimated from marks in the mud or soil. RESULTS Description of the Strickland River The gorge, upper and middle regions of the Strickland River (Figure 1) had not been fully explored prior to this survey. In the Strickland Gorge, the river drops from 388 m to 100 m over a 100 km course. This steep fall within the river's narrow width produced almost continuous rapids, some perhaps the largest anywhere. In addition, water temperatures there ranged 3-4OC lower than on the lower Strickland. Within the gorge and 436 km from the mouth, a 6 m waterfall may have been a physical barrier to upstream crocodile dispersal. The upper Strickland extends downstream from the gorge. It drOps from an elevation of 110 m to 50 m and has no adjacent swamp- lands. Its shoreline is soft gravel or sandstone. Its tributaries, the Burnett, Murray, Carrington and Cecilia Rivers, are clear, fast, boulder-strewn, mountain streams. The Damami is a typical lowland river. On the next section downstream, the middle Strickland, sand and clay substrate replaced the stone and gravel of the upper streambed. Shore areas were generally shallow and muddy, with protected coves containing many downed trees. In this section the current was slow and the course twisted. Distribution and Abundance Very few crocodiles could be identified to species. Data for New Guinea and saltwater crocodiles necessarily were grouped together. 11 12 The night count conducted on Lake Kopiagu revealed no crocodiles (Table 2) and this was the report also of the local Duna people. The Strickland Gorge was surveyed for 30 km downstream from the Tumbudu River Junction from inflatable white-water boats and revealed no crocodiles. After the white-water expedition ended, helicopter flights over the rest of the gorge still disclosed no crocodiles above the Burnett River (Figure 1). Two adult New Guinea crocodiles were seen at the Burnett Junction. Mr. Tom Hoey, an Asia Pacific Christian missionary reported (in conver- sation) that on an extensive foot patrol along the Burnett River in 1970 he also had found crocodiles. Crocodile-skin trader Keith Tetley (in conversation) had made the same observation in 1968. Though no evident barriers to crocodile range expansion occurred there, the upstream limit of New Guinea crocodiles on the Strickland River evidently was in this region at a 110 m elevation. As we surveyed downstream on the upper Strickland, the Damami River was the first tributary in which we recorded crocodiles (density index 1.33/km, Table 2). The crocodile density index on the upper Strickland was 1.22/km. Only a day count could be conducted on the Nomad River (Figure 1, Table l) but three adult crocodile slides were noted, indicating that a breeding population occurred there. Crocodile numbers of 2.23/km on the middle Strickland (Table 2) were the highest found during the study. Except for the Damami, tribu- taries of the upper and middle Strickland (Table 2) had lower density indices (.42, .16 and .07) than the main river (1.22 and 2.24). All crocodiles positively identified during night counts on the upper and middle Strickland were New Guinea crocodiles. But two belly l3 .Huxmu mom .mHmEHcm mmHmH mo mocmummcommum m mmsHocH ou memosn pony mHnHmmom muHm mo mmumEHumm o: =>Hco mmhm: H .mcoHumHomom mHHmooouo mchHmucoo monoumuum How cho xmmCH huHmcmm mmmum>¢ « m.m~ v.~ «.4 m.m m.m m.HH e.mH m.mm o.m 00H mommucmoumm mvH mH mm on am we omH meH mm .m.H vom Hem mHmuoe H H ho. Om N Hm>Hm msmHm H mH. m H um>Hm sacs H H H m we. NH m Hm>Hm Hsoucmm m H H H m mm.H m m um>Hm Henson HHH m 0H mH «H mm Hm GMH mm m~.~ mHm ope mHmmHs . um>Hm vameoHuum mm A m m m an em mm a -.H mmH mmH “mod: n um>Hm ocmeoHHum mm o mmuom . um>Hm ocmeoHupm u o :mmHmox wxmq .1Hom am emuom mHumH mHumH ~H-OH one mum aw .EH “mm H.2xv mumn «mum nuon NHHmn mmnosH :H mmNHm meHmooouo mmocmumHo 16:2 .mhmH .mmcHso 3oz momma .mmo:H>oum mmcmHanm mumaunom mam cumummz .Hm>Hm mcmHHOHHum man no mcoHuuom mHmmHE mom momma .mmuom commons: on» :o mucooo uan: Hmumzuon mcHHsp cmmm mmHHpooono mo mumoEdz .N mHQmB 14 slides of 40" and 50" width were seen at km 195 and no wild New Guinea crocodiles that large have been recorded. It may be, therefore, that saltwater crocodiles range that far upstream. No crocodile hunting was known to occur in the gorge, or on the upper and middle Strickland River and its tributaries (Table 2). Hunting was practiced, though, during low water on the lower Strickland. That area, with a density index of 1.66 crocodiles/km, was similar to the remaining portion of the study area in that it had vast areas of adjacent swamplands. Several large saltwater crocodiles were seen there during day counts, though none were identified at night. Lake Murray, with a density index of only .09 crocodiles/km of shoreline, had the lowest (and probably most-depleted) crocodile stocks (Table 3, 4). Many of that lake's tributaries, including the Boi, Lower Kaim and Leva Rivers (Figure l) with crocodile density indices of .32, .29 and .30/km, respectively, were judged to have been overhunted. Other streams entering that lake, such as the upper Kaim, June and Mamboi Rivers, nevertheless held considerably higher crocodile density indices (.76, 1.51 and 1.0 crocodiles/km, respectively). These low densities corresponded to areas of known high hunting pressure. Crocodile numbers were much lower (.35 crocodiles/km) on the middle Fly River and its tributary the Agu River (.43 crocodiles/km) than on the otherwise-similar habitat of the lower Fly River (1.8 crocodiles/km) where hunting pressure was much lower (Table 3). Except for Lake Murray, the middle Fly River was the most overhunted. waterway on the study area (Table 3). 15 so ommmo HmmHucss mason Hmmw mum>mmv m mam HHHc I .mmum mnu mo mommHsocx Hmumcmm momupv o cmmzumn mHH on ommmofl mm3 mummmmwm mcHucsm m .m oHnme «mm H .xcmnfimmuum mco mo mucooo Eoum UmHQDOp mmOHUcH qumch « o.mH 5.0 ~.~ m.H H.m m.m m.mH v.4m o.h~ o0H mammucmoumm em m 0H m 4H mH on mmH NNH Hm. was Hme Hmuoe H H H m mm mH .mm.H He mm uoon n um>Hm mcmeoHuum m OH H m N H OH RH oo.H av we um>am Hogans m m m H on. om m um>Hm m>wH m m m m «mH. em m saunas ome H m H m m m a on. me mm “mam: u Hm>Hm EHmM m m H H m m 3 mm. mm mH HosoH u um>Hm EHmM H OH H v m m m HH vH om Hm.H Hm we um>Hm «can m e H H m m m Hm. mm mm um>Hm uumnumm m e H a e 0 «mm. omH mm mHmmHs u um>flm aHm H mH H m m m mH mm on .m.H mmH mHH umon I Hm>Hm hHm m H H m H 3 mm. mm m uo>Hm Hom m e H H H e HH A me. 4» mm um>Hm 5mm «musmmmum How am emuom mHumH mHumH NH-0H mus mum 4w” .sH umml, ..sxv mumn «mum mmHHp mmocmumHo Ifidz @GHUCSI nupH3I>HHmn mmcomH :H mNHm Iooouo .ommHnmemH .mmcHso zwz madam .mo:H>oum cumummz .mmmum omucon co muosoo uzch umum3|3oH mcHuom comm mmHHpooouo mo mumnsoz .m mHomB 16 Size and Age Composition vs. Hunting Pressure Size (age) composition of the crocodile population was markedly different on unhunted and hunted areas (Table 2, 3 and 4). Only 5% of the population in the unhunted areas were :4" b.w. (about 40 cm body length, Figure 2) while in hunted areas these small crocodiles averaged 27%. Sub-adults (10"-l9" b.w.) comprised 19.3% of the population on the upper and middle Strickland but only 8.2% on the hunted areas. The pr0portions of animals in the 3?0" b.w. and E0 classes were 29.3% vs. 14.9% for unhunted and hunted regions, respectively. The upper and middle Strickland populations were primarily of adults and sub-adults and exhibited a low rate of reproduction. Indicating a stable condition, , these two areas were judged to be at historic pOpulation levels. This was in marked contrast to the hunted regions (Table 3) where almost 77% of the crocodile population were <10" b.w., indicating a population far below habitat carrying capacity. Reduced numbers, seemingly due to hunting by the local people, were evident in all size classes <10" b.w. Adults of known breeding size (320" b.w.) comprised only 2.9% of the population, although much of the 12% E0 class would likely add to the total of mature animals. These proportions, nevertheless, were much lower than the 6.8% known adults and 22.5% E0 for the unhunted upper and middle Strickland areas (Table 2). Water-level Effects Seasonal changes in water level had a marked effect on the numbers of crocodiles seen and possibly on crocodile densities. High-water number83mnrkilometer of shoreline were 56% below and low-water counts l7 .m mHnma 0mm H m.mH u.H m.m m.m v.m m.m o.mH m.mm m.MH oOH mmmmucmoumm mom mH mm mm mm mm oom Hmm mmH mo.H NOMH NHHH Hmuoe Hom em vmnom mHan mHan mHnoH one mum vw_ .sx mum H.2xv mama mmHHm mocmumwa I352 numenHHHon mmnocH cH mNHm .mmo:H>0Hm mUGmHsmHm :Hmnusom mam cumummz nooonu .ommHImhmH .mmcHso 3oz mammm .mucsoo mHHpooouo uan: Hmum333oH HHm mo humEEom .v mHQmB 18 BELLY WIDTH 50 100 150 200 250 300 350 TOTAL LENGTH (cm) Figure 2. Relationship of belly-width (Lever, 1975) to total length based on measurements of 500 New Guinea crocodiles captured in Western Province, Papua New Guinea, 1979-1980. 19 were 46% above those recorded during periods of normal water levels (Table 3, 4, 5 and Figure 3). While counts of crocodile belly slides (Figure 3) were a poorer indicator of abundance than night counts and were useless during high water, they do indicate that crocodiles moved out of adjacent swamplands and were more concentrated in rivers during low water. Changes in water level also affected the distribution of crocodile by size class. In high water, 38.6% of the visible population (including E0) were <10" b.w. (Table 6) but during low water were only 23.1% (Table 3). Larger crocodiles seemed to stay in the rivers regardless of water level while smaller ones tended to disperse into adjacent flooded swamps. Night Counts vs. Daytime Tallies Since 12.9 times as many crocodiles were seen at night than during the day (Figure 4), day counts were weak representations of crocodile abundance. Day counts (Table 8) favored the sighting of large crocodile slides. Seventy three percent of the slides seen were from animals :10" b.w. while in the corresponding low-water night counts over the same area, only 23.1% of the population (including EO's) were :10" b.w. Big crocodiles were simply easier to see during the day than were small ones. But slides showed a more realistic distribution of the larger size classes than did either of the live crocodile survey types. The number of crocodiles seen on one bank of a river during night counts were not significantly different (two tailed "t" test, =.05, Steel and Torie, 1980) than the number seen on the other bank. When 20 .m mHnme mam N .N mHomB mom H .xcmnsmmwum moo mo mucooo Scum mmHodom_mmoncH wuHmcmo .. ¢.H~ m.v m.v m.H m.q m.HH H.mm mm OOH mammucmoumm mm m m m m 0H mv mm mm. one mmH Hmuoe H HH v N m 0H 0 me. He mv umon I mcmHHOHHum N N N N m n 0N. No mH HmsoH I Hm>Hm EHmM H m H H m m o mH mm. Hm Nv wm>Hm mash m H H H v m H Hm. mm HH Hm>Hm pumnumm m N H H OH N «0N. omH mH mHmmHE I Hm>flm mHh N m H H H N m m mu. NN 0H Hm>Hm Hom N m H m H H m H vN. vb mH wm>flm 9mm amusmmmum How an vmnom mHImH mHImH NHIOH a-» mum aw .ex “mm H.5HV mama «and mmHHm mocmumHo Ieoz mcHucsm numH3I>HHmn monosH CH mNHm Iooowo .ommHImhmH .mmcst 3oz momma .mocH>oum cumummz .mmmum means: so mucooo uan: Hm>mHIumum3 Hmsuo: mcHHsm comm mmHHmooono mo mumofioz .m mHome 21 .ommHImhmH .mmcHow 3mz mommm .HOHuumHo mmwuoz mme .HmmcHHm mmNuzv mmmHHm mHHmo mo muoooo amp Amy mom HmmHHoooouo vmvuzv muoooo uano Amy somo momma mm mum>Hu om>mm co mmoHumm Hm>mHIumum3 momma mo some mcHuom mmHHmooono mo mommmoonm m>HumHmH one .m muomHm ..m>m._ amh<>> Boo 25.02 :9: 30. 25.02 :9: o o me mad 0.. and 9. who o.« 8.. Ru 3.. c...” on; ma 2.. o6 < oo.~ m... mud O.m OmN NOIlV'lndOd 13A3'1 dBlVM 1VWHON :IO NOIIUOdOUd .m mHnma mum m .N mHnme mum H .omnEmmuum mco mo mucooo Eowm omHnooo mmoHQCHthmomo « 22 m.hN m.o v.H m.o o.m m.m H.mH m.hN 00H mommuomoumm 0N m m m N m mH 0N mH. 05v Nb Hmuoa H m N H m HH cH om. Hm ov um30H I mcmeoHHum N H N mmo. No m um30H I Hm>Hm EHmM H m H H H m NN. Hm HH Hm>Hm moon m H N H H H NH. mm o H0>Hm uwmnumm m H H H ammo. 00H m mHmoHa I Hm>Hm mHm N m H H H N Nm. NN h Hm>Hm Hom N N hNo. on N Hm>Hm omm Nmuommmmm, Hom vN «NION mHImH mHImH NHIOH th mIm vw .53 wmmI H.Exv moon mmum mmHHo moomumHo IEoz moHucom numH3I>HHmo monocH :H muHm Iooouu .ommHImhmH .mmoHou Bmz mommm .mo:H>oum oumummz .mmmum omuoon co mucooo noch umumsIsmHn mcHHom comm mmHHoooouo mo mumoaoz .m mHome 23 mHnt mom H o.mH v.H m.m N.m v.m m.h ©.©H H.0N h.mH OOH unmoumm mmm OH mm me me mOH «mm 4mm «Hm Om. HORN mmmH Hmuoe How am vNIOm mHIOH mHImH NHIOH mus mum 4w .ax\ H.2HO meme mmHHo mocmumHQ IEoz sumHsumHHmn masocH :H muHm Iooouu .OmmHImhmH .mmcHsO 3mz mommm .mooH>oum cumummz .mucooo uan: HHm msHHoo comm mmHHoooouo on» HHm mo mamasom .e mHnme 24 500 480 444) 400 360 320 280 240 200 160 120 NUMBER OF CROCODILES SIGHTED 80 4O DIKY lfllCSliT TIME OF SIGHTING Figure 4. Relationship between the number of crocodiles sighted by day and night on seven rivers. Western Province, Papua New Guinea, 1978-1980. 25 m.m ©.mH N.©N m.NH m.mH w.HH o.m 0.5 mmUHHm mHHUOOOHO Hom N.ON m.OH m.mH III m.mH m.mH N.N v.m mmHHOooouo em «N ONION NHIOH mHImH NHIOH mIN Im ng leuOHzINHHmn manuaHO mmuHm .ommHIman .mmoHow 3oz mommm .wum>Hu om>mm co muoooo amp Hmum3I3oH scum mmoHaumumo mm mmmHHm NHHmo Jam mmHHoooouo mo mmmHOImuHm an mmmmuomoumm .m mHQmB 26 counts show fewer than 20 crocodiles (Figure 5), however, statistical variation may be too great to insure that doubling is a reliable procedure. Habitat Use General Preferences. The habitat locations in which crocodiles were most commonly sighted were, in decreasing order of use (Table 9), SWOE, IVIW, OB,MS and least frequently, IV category. The three that were used most,where 76.2% of the crocodile population were seen,were the SWOE, IVIW, and ILIW. While no surveys were conducted to indicate precise habitat availability, they were judged to be in the decreasing order: MS, IV, SWOE, IVIW, OB, and ILIW. Considering that it was the least available, ILIW was the most preferred habitat location. Almost every time that a clump of deadwood was seen during a night count, it had crocodiles in it. In contrast, only 0.9% of the population were seen in the IV location which was the second most common in the environment. During any one night, an average of 16.9% (16.0% OB and 0.9% IV) of the population were seen on land. On the upper and middle Strickland areas (Table 10), 17.4% of the crocodiles were on land. Only 13.3% were seen on land, though, during low-water surveys in hunted areas (Table 9). Preferences Related to Belly Size. When the size-classes of crocodiles in each habitat situation are compared with the average in those habitats (Table 9), it becomes apparent that: 27 .uanu m.um>ummoo on» on xcmn one « .ommHImhmH .mmsHoO 3oz mommm .mocH>0Hm cumumm3 .mum>HH cm>mm co AmmoHuanm mHHoooouu vmmv muoooo uano MN ocHHom «Homo usmHH mnu co mouanm mum3 :oHnB mmHHpooouo mo COHuHomoum .m muomHm «v.24; FIG; 20 mm.:oOUO~_U “.0 «mmioz mm 00 mm on me 0? mm on mN ON mp or m o .J‘IIIIII1IF1I q I 1. dI a HI. H II! H H 1 ,..4 H O «.0 .. NO 0 I 0.0 o O . . I to 0 all: IIII IIII III. III. .III. III. III .IJ1 .III .Ild .III. Idl. III. IIdIo 0.0 o o .. O.O O C so ad ad 0.? )INVS 1H9”! NO ("1“98 LING) 1V101 :10 NOIlHOdOBd 28 .N mHomB mom «« .Humum3 :H mmOH oHV SHAH “HmcmH co QOHumummm> CHO >H “meum3 GH cOHumummm> :HV 3H>H “Axomn mwmo coy mo «Amoco co umum3 30HHmnw. mozm “Hemmuum oHEV m: « OOH O.ON O.O m.NN O.OH O.OH O.O mmO Hmuoe OOH N.mN H.H N.NN O.N O.mm H.HH ONH II on OOH N.N O.O N.N N.N m.HO O.OH mH ON OOH .H.mH O.O m.NN H.O N.ON m.NN mm ONION OOH O.OH O.O m.NN O.OH O.ON O.OH NN OHIOH OOH O.OH O.O O.OH O.OH O.OH O.OH Om OHIHH OOH O.ON H.H N.O O.OH O.OH O.O Om NHIOH OOH O.ON O.O H.mN m.NH H.Om O.O OOH O I N OOH N.HN O.O O.ON H.NN N.Om H.H NNN O I m OOH m.NH m.H N.ON O.OH N.Om m.H NON Ow Heeoe O zHHH >H 3H>H mo Noam m2 Ommm manocH OH mumnssz AgumHz NHHmnO RooHumooH umanm: Hmm mmmHo mNHm some :H mmHHoooouo mo mommucmoumm muHm .OOOHIONOH .mmcHoo 3mz momma .mo:H>oum cumumm3.m>m>uom uanHqumHo Anvv HHm Mom mmHHoOUOHo mNHm ucmHOMMHm mo mGOHumooH umanmm .m OHQmB l) 2) 3) 4) 5) 6) 29 Larger crocodiles (>12" b.w.) were more likely than smaller ones to be found in mid-stream. Juvenile crocodiles (<12" b.w.) were twice as likely than the rest of the population (not including E.O.) to be in the ILIW situation. Juveniles were more evident on bare banks than adults and sub-adults. The SWOE, IVIW and IV locations were utilized about equally by all size segments of the population. The "eyes only" class most closely resembled the 'over 24"' class, as also found by Messel et a1. (1977). But the large percentage of E.O. in the IVIW situation probably indicates that some juveniles were also seen "eyes only." The "zero sightings" of four size-classes in the IV situation may not be significant but there is evidence (see beyond) that they were absent, not merely hidden, by foliage. Preferences Related to Water-Level. When the percentages of crocodiles in each habitat situation are compared with changes in water level (Figure 6, Table 10 and 11), it can be seen that: 1) 2) 3) 4) Wariness Use of the IVIW and IV location increased as water levels rose. Use of the SWOE location was negatively correlated with water level. The same proportion of the crocodile population was seen in mid-stream regardless of water levels. Crocodiles used bare banks (OB) and fallen trees (ILIW) locations more at normal water level than at any other time. The low and high water usages were about the same for OB but use of ILIW was greater during low water. There is an increase in flight distance with increasing crocodile size (implying age and experience) and with hunting pressure (Figure 7). 9‘ OF TOTAL CROCODILES SIGHTED 555 5!) 445 4C) AAS Figure 6. 3O VVATWER LE\/EL ...... ha. I a... ..... I... ..... a... lllll I... ..... ccccc I... ..... I... ..... I... mmmmm O... ..... O... IIIII O..- OOOOO O... IIIII I... ooooo O... ooooo III. ..... Cl.- ..... I... ..... III. ccccc I... ccccc I... ccccc I... ..... O... ..... ooooooooo I... ..... c... ccccc O... ..... a... ..... Q... ..... cul- ..... I... ......... ..... Oll- ooooo IOU. ..... I... 00000 lo. I... ..... O... ..... I... ......... ..... O. O ..... Oil. ..... I... 000000 I... ..... I... ..... O... ..... O... ..... I... ..... U... ..... I... ..... I. U ..... I... IIIII 0... IIIII I... ..... C... ..... a... ..... I... ..... I... ..... I... ..... a... O ..... I .00- ..... ..... ooooo ..... IIIII ..... ..... IIIII ..... .......... nnnnn ..... ooooo ..... nnnnn ..... ...... ooooo 00000 .......... ..... IIIII ooooo 00000 ..... ..... .......... ..... ..... ..... ..... ..... ..... ..... .......... ooooo ..... :.:.:.;.;. '''''''''' ..... ooooo 0000000000 ..... ..... swoe on IVIW Iv ILIW HABITAT TYPE * Percentages of crocodiles sighted in each habitat location in relation to water level. Data are from three surveys on the same portions of seven rivers. Lake Murray district, Papua New Guinea, 1978-1980. * see Table 9. 31 .N mHnme mmm II .O mHnme mam I N.NN 0.0 O.NH 0.0H 0.0m 0.0 mommucmoumm hNH m hm No ©ON Hm H4909 HN N mH m mo mH ««.O.m H H H m :ON Hm>o v N H h H :ON I =0N m H v m N =mH I :0H m N m m :mH I :mH NH H N m 0H N :NH I :OH ON mH 0H on H :m I :5 mm 0N mm ow H :m I :m mH N Nm OH Om m mmmH Ho :O SHAH >H 3H>H m0 mOSm «ooHumUOH umanmm HmmImmmHO mNHm 30mm :H mmHHoooouo mo mumnaoz Oz zumHs NHHwn mmnocH . mNHm .ommHImbmH .mmcHoO 3oz mommm .mocH>oum cumummg .mmmum Hm>Hm mcmeoHHum meoHE mom Hmmmo mmuoosco on» so m>m>uom umum3I30H moHHoo comm mmHHmooouo muHm pomHmMMHm mo mcoHmeOH umanmm .OH mHnma 32 IN OHQME Omm «O .m OHQMB Omm « O O H m H O z O O z x O H Hm>mH noumz N.O O.N O.OO O.NH O.OH O.O O.N O.H O.ON N.OH O.OH N.O N.OH O.O O.ON O.OH O.OO O.O OOOO Iucmo lumm O O ON O NH O OO N NO ON ON NH NO OO ON ON OH Hmuoe O N O O H N O OH N O N HH O O H II.O.m N VN HO>0 N H N H O O H H O ONION N H H N H N H N OHIOH O H H H H H O OHIOH N O H O N N N NHIOH H H O H N O O N N N HH N HH N O I N HH N NH OH O OH H O OH NH HO H O I O H H NH N O OH H O O O O O O ON Ow SHAH >H 3H>H mo mozm m2 BHHH >H 3H>H mo mo3m m: SHAH >H 3H>H mo m03m m2 I£H©H3 , NHHon «ZOHBOHQ Chmflmwz .mH0>HH C0>Om MO SUMO CO mHm>mH Hmums woman «0 nomm um mucooo uanc Eoum mmHHoooouo mNHm HomHOMMHo mo mooHumOOH umanmm .HH mHnt 33 .ommHImhmH .mmoHoO 3oz mommm .mocH>oum cumummz .Hmmmuz .mcHH Umsmmpv mmmHm vmucos pom HHvNuz .mCHH UHHOOO pmuoonoo :0 moomumHo uanHm mam mNHm mHHpooouo ommBumn mHnm:0HumHmm .5 muomHm 2:23 3.: .22.: ON; E.Oonvomu OO - ON ON O. O. O O O . O \ \. O. \ \ \ ON xOO.O+ON.OI no \ \ OO \ \ \ OO \ xO..N+ O.O.Iuo (w) BDNVlSIG lH‘DI'IJ 34 The average flight distance increases from 2.5 m and 3.5 m for thefi4" b.w. class, to 25.0 m and 52.4 m in the 20-24" b.w. size class, respec- tively, for unhunted and hunted populations. When a least squares regression (Neter and Wasserman, 1974) is plotted of crocodile size against flight distance, the rate of increase in flight distance for the hunted population is 2.2 times that of unhunted populations (Figure 7). Evidently, hunting pressure was over twice as important as age in determining flight distance. The average approach distance of the several size classes :16" b.w. was similar to that of the "eyes only" class (Figure 8), which presumably indicated that these E.O. sightings mainly were larger crocodiles, which further supported Messel et a1. (1977 and 1981). The percentage of crocodiles in each size class approached to within 1 m or less (Figure 9), the distance at which a village croc- odile hunter would have nearly 100% success with spears or harpoons, ranged from 35.0% and 27.2% for the :4" b.w. size class to 0.0% and 9.5% in the 20-24" b.w. size class for hunted and unhunted populations respectively. It can be seen (Figure 9) that hunting pressure had a greater effect than size in determining flight distances of 1 m or less. The slope of the hunted population's regression line was over 4.5 times as steep as the one from the unhunted areas. Comparing this pr0portional difference of 4.5 to the 2.2 in the regressions plotting flight distance against size (Figure 7) indicates that hunting pressure was %f%-= 2.05 times as important in determining the percentage of approaches :1 m in each size class than it was in determining flight distances in general. 3S .OOOHIONOH .OOOHOO 3mz Osmmm .mmmum ANONIz .Omcmnmv ewuqss OOO HOOHIz .NOOHOO meanness so .mo:H>oum cumummz momma H w.ou omsomoummm mum3 noH£3 mmmHo mNHm Na mmHHoooouo mo mmmmuomonom .m muomHm 2.23 :_.O .22.: ON; O.__O0uo._u vNION 97: 2.5. «To. OIH. oIn CW CD . .9. :2.- 2 (w) aamnsm mom Ci «5 36 .OOOHIONOH .OOOHOO 3mz mommm .mooH>oum cumummz .mooHomu HmmHHpooouo Nmmuz .moHH pmnmmmv omucon com HmmHHoooouo mmHuz .mcHH oHHOOV mmuconco mo mNHm m.Hmchm may pom H:mHHHx mwom=v mmmH Ho HmumE H cHnuHS ou omnomoummm mmHHmooouo mo mmmuomoumm mop :mm3umo mHnmooHumHmm 2:23 .23 2:12.: ON; 33953 on mN ON 0.. or m xOO..IO OOuo / x Dad-Ioduflo .m mnomHm or m_. ON mm on 8313W ll ; Ol OHHDVOUddV‘XI CONCLUS IONS Distribution and Abundance The absence of crocodiles in Lake Kopiagu and the Strickland Gorge seems likely to be due to a scarcity of forage fishes rather than to physical obstacles or high altitudes. Fish are the largest part of a crocodile's diet (Cott, 1954; Corbett, 1960) and local Duna tribesmen reported (in conversation) that Lake Kopiagu declines to a mere mud hole during drought years. Roberts (1978) found that fish also were scarce in the mountainous regions of the upper Fly River, an area similar to the Strickland Gorge. Since it has been found that fish diversity and numbers increase as one travels down the Fly, a river similar to the Strickland (Roberts, 1978), this may be correlated with the continuous increase in crocodile density observed on the latter. The crocodile density index on the upper Strickland was 1.22/km. The middle Strickland, the next section downstream, had an index of 2.23/km. This change in the density index from 1.22 to 2.23 is an 83% increase. If this rate of increase can be extrapolated so as to apply to the lower Strickland, an index of about 4/km would occur rather than the 1.66 found in the survey. If that extrapolation is justified, a considerable reduction in the croc- odile population, probably due to hunting, may be indicated. At present in PNG, streams with 2 or more crcodiles/km must be considered to be maintaining healthy populations. The density index of 2.23 animals/km however, is not impressive when compared to some other crocodilian 37 38 populations. Neill (1971) reported an American alligator density index of 9.3/km. Nile crocodiles (g;_niloticus) numbered 20/km on Ethiopia's Awash River (CRDtt and Pooley, 1972). Graham (1968) reported 56 crocodiles/km on portions of Lake Turkana in Kenya. Aside from this last location, though, many of thelruflidensities reported in the litera- ture were for shorter distances than surveyed on the middle Strickland. Isolated exceptional index levels could have yielded inflated results due to encounters with one or more atypical groups of crocodiles. The lack of adjacent swamplands on the upper and middle Strickland may be a reason that crocodile densities were much lower on the trib- utaries than on adjacent portions of the main river (Table 2). Low water in the upper regions presumably would affect the tributaries first as opposed to the swamplands in the lower areas. Crocodiles probably move out of the tributaries of the upper reaches during periods of low water much as the crocodiles lower down move out of the swamplands. Crocodiles were plentiful in Lake Murray during the 19405 and 19505 (Bustard, 1968 ; Neill, 1971) and supported densities (Craig and Kune', tribesmen, personal communication) that were higher than any of the lake's tributaries. If Lake Murray did have a greater density index than the 1.51/km (Table 3) noted for its June river tributary, then the crocodile population in the lake has been reduced by about 88% (l- Ll§-) from its original level. 1.51 Crocodiles seem to be making a comeback on the lower Fly River and, to a much lesser degree, on the middle Fly and Lake Murray. Bustard (1968) and Neill (1971) reported sighting only 4 crocodiles/night on the lower and middle Fly River and l per night on Lake Murray in 1967, yet 39_ the present study revealed density indices of 1.8, 0.35, and 0.18 crocodiles/km on these respective areas. This improvement probably results from the National Crocodile Project's efforts toward better legislation, public education and management. While some of the rivers and Lake Murray had low crocodile stocks, there was some evidence that additional crocodiles occurred in the lagoons and backwaters. The low water surveys of over 1302 km of river and lakeshore on this study area resulted in a count of 1112 crocodiles (Table 4). Messel et a1. (1981) suggested that 63% of crocodiles actually present are seen on night counts. If this percentage is accepted, it could be concluded that actually 1112 .63 there were ( ) 1765 crocodiles present. In this same area in 1980 alone, however, 2,002 small live crocodiles plus approximately 1,100 skins were taken out. Undoubtedly, the permanent swamplands adjacent to the main navigable areas yielded the additional animals. ' Furthermore, the saltwater crocodile's primary habitat is riverine and not flooded backwaters (Lever, 1975). The low percentage of this species in the harvest would be a further indication that much hunting likely takes place in backwater areas. Saltwater crocodiles were undoubtedly scarce in the Lake Murray District. Of 2,002 small live crocodiles purchased on the study area in 1980, only 1.1% were saltwater crocodiles (Balson, 1980). If captive saltwater crocodiles were available, especially as mature specimens, then even a small-scale restocking program might increase this species population above its critically-low level. -40 Habitat Use That juvenile crocodiles dispersed into flooded areas during high water (Tables 3 and 6) also was determined by Chabrek (1965) who observed that juvenile American alligators tend to disperse with increas- ing water levels. Juvenile crocodiles may suffer harassment (Messel et al., 1981) and even cannabalism (Nichols et al., 1976) from larger members of their species. Hence, small crocodiles may tend to seek sanctuary in areas away from large crocodiles when suitable habitats are made available by flooding. The fact that 12.9 times (Figure 4) as many crocodiles were seen at night as by day may explain why early explorers into the middle Fly drainage reported seeing few live crocodiles during the day despite the many belly-slides present. Chabrek (1966) reported that between 5 and 12% of American alligator populations were seen on land during night counts in Louisiana. His figure closely approximates the 16.9% observed in.thisstudy. Lang (1979) also found that American crocodiles (g;_acutus) spend most of their on-land time at night. Crocodiles in the study region were a reclusive group that were seldom out of the water except at night. The low values (0.9%) in the vegetated (IV) location may have been affected by poor visibility due to foliage. Yet since the pro- portions of on-land (OB and IV) sightings were not below those on-land sightings for other crocodilian species may indicate that crocodiles IV were not being overlooked. Juvenile crocodiles were more likely to be found OB than were adults and sub-adults, because young crocodiles tend to seek warm 41 places (Lang, 1981) to a greater degree than larger crocodiles. Also, large crocodiles have more difficulty moving on land than do small ones. The use of the mid-stream (MS) habitat type by mainly larger— sized crocodiles was also observed by Messel et a1. (1977, 1981). The smaller crocodiles may be reluctant to leave the safety of near-shore cover and to venture out into the open water with its dangers of inter- and intra-specific predation (Nichols et al., 1976; Valentine et al., 1972). Large crocodiles, on the other hand, have no predatory enemies in New Guinea other than man. Juvenile crocodiles ($12" b.w.) utilized ILIw habitat more than larger crocodiles probably because the smaller size-class feeds heavily on the schools of small fish (Corbett, 1960) that frequent brushy areas. Evidently, the differential utilization of habitat types with changing water level was more the result of changes in availability and accessability of habitats rather than changes in habitat prefer- ence. The MS category was the only habitat location whose availability was not altered by fluctuations in water level. And, there was no significant difference in the use of MS across water levels (Figure 6, Table 11). As water levels rose, SWOE locations became IVIW. It seemed likely that as the water level beyond the line of emergent aquatic plants became too deep for crocodiles to rest their feet or tails on the substrate, they moved back into the vegetation near shore where they could touch bottom. The decrease in the proportion of crocodiles found SWOE was balanced by a corresponding increase in the propor- tion using IVIW (Figure 6). 42 The slight increase in the use of the IV habitat with rising water levels likely resulted from the fact that walking distance to on-land vegetation from the water's edge was shorter as water rose. The small proportion of the population using ILIW habitat in high water probably resulted from the fact that as the water came up, dead wood either floated away or became submerged. Likewise, when the water level was down, many woody snags and brush piles were left out of the water. The amount of ILIW habitat available was greatest during normal water levels when falling trees on the bank of a river were not likely to be carried away nor to fall short of the water. Normal water was when the greatest proportions of crocodiles were found in the ILIW type (Table 11, Figure 6). Wariness Flight distance data (Figure 7) indicates that wariness in New Guinea and saltwater crocodiles (webb and Messel, 1979) was a learned response which increases slowly with age under natural conditions. But wariness resulting from an unpleasant experience associated with boats, motors, lights and/or people was in addition to that mentioned above. The hatchling (54" b.w.) class was an exception to this trend because they were more wary in the unhunted area (Figures 7, 8, 9 and Webb and Messel, 1979). This phenomenon may be the result of hunters not attempting to catch small crocodiles and thus imparting no unpleas- ant experience by their activities. Hunting pressure was 2.05 times as important as size in determining approaches :1 m (Figure 9) than it was in determining flight distances in genera1(Figure 7 and 8). Once a flight distance is achieved which allows 43 crocodiles to escape hunters, evidently there is little value in extending the flight distance further. 1) 2) 3) 4) S) 6) RECOMMENDATIONS Crocodile populations were below 0.5 animals/km on Lake Murray, the Leva, Boi, lower Kaim, middle Fly and Agu Rivers. Crocodile harvests there should be halted until there is evidence of further population recovery. All other areas studied may continue to be hunted as long as yields are sustainable by the population. The sale of live saltwater crocodiles or their skins should be banned from the Lake Murray District until the local crocodile population recovers. Restocking may be considered as a management procedure there if hunting can be controlled during the re-establishment period. All saltwater crocodiles and many New Guinea crocodiles at the Baboa Crocodile Station should be secretly released into areas with critically-low populations. Low-water night counts should be conducted on one bank of a stream and be doubled to establish a density index. Such surveys should be considered once every two years in the district to determine if the overhunted areas are indeed recovering and to identify new trouble spots. A test should be made of felling trees and shrubs into adjacent streams. This should be planned so as to determine the effects of this practice on crocodile densities vs. merely encouraging the local crocodiles to become concentrated there (ILIW). LITERATURE CITED Archbold, R. and R. Rand. 1940. New Guinea expedition. R.M. McBride & COO, NoYo: 148-171. Balson, E. 1980. Progress report No. 8 for the Papua Region, to FAO proj. PNG/74/029 Assistance to the crocodile skin industry. Unpub. 14 pp. Bauerlen, W. 1886. The voyage of the Bonito. Gibbs and Shallard & Co. Sydney. 38 pp. Brandes, E.W. 1929. Into primeval Papua by seaplane. Nat'l. Geog. 106(3): 253-332. Bustard, H. 1968. Report on the crocodile skin trade in the territory of Papua New Guinea with recommendations for the future develop- ment of the industry. D.A.S.F., Konedobu, unppb. 22 pp. . 1968a. Rapid learning in wild crocodiles (Crocodylus porosus). Herpetologica. 24(2): 173-175. Chabrek, R. 1965. The movement of alligators in Louisiana. Proc. S.E. Assoc. of Game & Fish Comm. 19: 102-110. . 1966. Methods of determining size and composition of alligator populations in Louisiana. Proc. S.E. Assoc. of Game and Fish Comm. 20: 105-112. Corbett, P. 1960. The food of a sample of crocodiles (Crocodylus niloticus) from Lake Victoria. Proc. 2001. Soc. Lon. 133: 561-572. Cott, H. 1954. The status of the Nile Crocodile in Uganda. Uganda J. 18: 1-12. and A. Pooley. 1972. The status of crocodiles in Africa. I.U.C.N. Pub. New Ser. 33:98 pp. Everill, H. 1886. Exploration of New Guinea - Geographical Society of Australia. Gibbs and Shallard & Co. Sydney. 20 pp. Ford, E. 1973. Papua New Guinea - the land and the people. Jacaranda Press, Melbourne, Australia. 168 pp. 44 45 Gore, R. 1978. Bad time to be a crocodile. Nat'l Geog. 212(1): 90-115. Graham, A. 1968. The Lake Rudolf crocodile (Crocodylus niloticus laurenti) population. Unpub. Masters thesis, Univ. ofIEast Africa, Nairobi, Kenya. 158 pp. Heathwole, H. 1968. Relationship of escape behavior and camouflage in anoline lizards. Copeia 1: 109-113. Hides, J. 1936. Papuan wonderland. Blackie and Sons, Glassgow. 204 pp. Hurley, F. 1924. Pearls and savages. G.P. Putnam & Sons, N.Y. 321-414. I.U.C.N. 1978. Joint report of the Papua New Guinea government/FAD project "Assistance to the crocodile skin industry" at the I.U.C.N./S.S.C. Crocodile Specialist Group 4th working meeting. Feb. 6-12, Madras, India. 1-13. Lang, J. 1979. Crocodilian thermal behaviors: alligator vs. croco- diles. Am. Zool. 19: 975. . 1981. Thermal preferences of hatchling New Guinea crocodiles: effects on feeding and ontogeny. J. Therm. Biol. 1:1-6. Lever, J. 1975. Crocodile industry training manual. D.A.S.F. Wildl. Man. 75/1, Konedobu. 70 pp. Medem, F. 1976. Report on the survey carried out in Papua New Guinea in 1976. P.N.G. Wildl. Div. Pub. 77/26: 48 pp. Messel, H., A. Burbidge, G. Wells, and w. Green. 1977. The status of the saltwater crocodile in some river systems in Northwest Kimberley, Western Australia. Rpt/ 24 Dept. of Fish & Wildl., 50 pp. , G. Wells, and J. Green. 1978. Crocodylus porosus popula- tion studies - survey techniques in tidal river systems of Northern Australia. IUCN/SSC Croc. Spec. Group 4th working meeting, Madras, India. 9 pp. , G. Vorlicek, G. Wells, and J. Green. 1981. The Blyth- Cadell River system study and the status of Crocodylus porosus in the tidal river systems of Northern Australia. Methods for anal- ysis and dynamics of a population of C. porosus. Mon. #1 Pergammon Press, Melbourne (in pressf7_ Nichols, J.D., L. Viehman, R. Chabrek and B. Fernderson. 1976. Simula- tion of a commercially harvested alligator population in Louisiana. La. State Univ. Ctr. for Agr. Sci. Rural Devel. 59 pp. Neill, W. 1946. Notes of Crocodylus novaeguineae. Copeia. 1946(1): 17-20. 46 . 1971. Last of the ruling reptiles. Columbia Univ. Press, N.Y. 486 pp. Neter, J. and W. Wasserman. 1974. Applied linear statistical models. Richard D. Irwin, Inc. Homewood, Ill. 842 pp. Paijmans, K., D. Blake, P. Bleeker and J.R. McAlpine. 1971. Land Resources of the Morehead-Kiunga area, Papua New Guinea. Land. Res. Ser. 29 C.S.I.R.O., Australia. 124 pp. Pooley, A. 1976. Papua New Guinea - a report on crocodile farming. P.N.G. Wildl. Div. pub. 77/27: 14 pp. Roberts, T. 1978. An ichthyological survey of the Fly River in Papua New Guinea with descriptions of new species. Smith. Contr. to Z001. 281: 72 pp. Schmidt, K.P. 1928. A new crocodile from New Guinea. Field Mus. Nat. Hist., 2001. Ser. 12: 175-181. Steel, R. and J. Torie. 1980. Principles and proceedures of statis- tics. McGraw-Hill, N.Y. 697 pp. Tago, S. 1977. The crocodile industry and crocodile laws in Papua New Guinea. P.N.G. Min. of Env. and Cons. pub. 9.1.2: 9 pp. Valentine, J., J. Walther, K. McCartney and L. Ivy. 1972. Alligator diets on the Sabine National Wildlife Refuge, Louisiana. J. Wildl. Manage. 36: 809-815. webb, G., and H. Messel. 1979. Wariness in Crocodylus porosus. Aust. Wildl. Res. 6(2): 227-234. Wheeler, T. 1979. Papua New Guinea: a travel survival kit. Lonely Planet Pub., Victoria, Australia. 224 pp. Whitaker, R. and Z. 1978. A preliminary survey of the saltwater crocodile (Crocodylus porosus) in the Andaman Islands. J. Bombay Nat. Hist. Soc. 75: 43-49. . 1979. A preliminary survey of the crocodile resource in the island provinces of Papua New Guinea. Unpub. rept. to F.A.O. project "Assistance to the crocodile skin industry of Papua New Guinea": 21 pp. . 1980. Status and distribution of crocodiles in Papua New Guinea. FAO project "Assistance to the crocodile skin industry of Papua New Guinea". Field Doc. #1. 59 pp. woOdward, A.R. and w.R. Marion. 1979. An evaluation of factors affecting night light counts of alligators. Contr. #1400 of the J. Ser. Fla. Agr. Exp. Stat., Gainesville, Fla. 21 pp.