Marine Turtle Newsletter Number 49 April 1990 Editors: Editorial Board: Karen L. Eckert & Scott A. Eckert Nat B. Frazer Physiological Research Lab (A-004) Nicholas Mrosovsky Scripps Institution of Oceanography David W. Owens University of California-San Diego Peter C. H. Pritchard La Jolla, California 92093 USA James I. Richardson KEMP'S RIDLEYS ARE RARER THAN WE THOUGHT In 1989, 835 nests of the Kemp's ridley (Lepidochelys kempi) were recorded by the bi-national beach monitoring crew at Rancho Nuevo, Tamaulipas, Mexico (Márquez, personal communication). Despite intensive patrols, it was not possible to encounter all of the nesting turtles; the turtles spent a short time on land (about 45 minutes), showed unusually broad dispersal north of the camp headquarters at Barra Coma, and also a new tendency toward very early morning nesting during the 1989 season. Nevertheless, 201 turtles were tagged with Monel metal tags in 1989, and 74 turtles tagged in previous seasons were encountered. Of the 201, 116 were recorded nesting once, 72 twice, and 13 three times. Of the 74, 47 were seen once, 23 twice, and 4 three times. These data allow the calculation of an estimate of the average number of nests per female per season as follows. Out of the 835 total nesting events, the turtle was seen (and tagged, or the tag number noted) on 404 occasions. Thus, based on the assumption that beach coverage was consistent throughout the season, there was 404/835 = 0.484 chance of witnessing a given nesting event and consequently a (0.484)3 probability of witnessing a three-time nester on all three occasions. So, if three-time nesters were observed on 13 + 4 = 17 occasions, the actual season's total of three-time nesters can be estimated at 17/(0.484)3 = 150. Similarly, to estimate the actual total of two-time nesters, I note that the observed total of 72 + 23 = 95 includes a subset of three-time nesters that were actually observed only twice. The chance of seeing a three-time nester on exactly two of its three nestings (i.e., on nestings 1 and 2, 1 and 3, or 2 and 3) may be estimated as 3x(0.484)2(1-0.484) = 0.363. Thus, 150 x 0.363 = 54.5 of the three-time nesters would have been seen just twice, leaving 95 - 54.4 = 40.6 actual double nesters observed both times. This corresponds to a true total (observed + unobserved) of 40.6/(0.484)2= 173.3 double-nesters. The triple and double nesters together thus produced (150 x 3) + (173.3 x 2) = 796.6 nests for the season, leaving just 38.4 nests (835-796.6) made by single nesters. So 835 nests were made by (150 + 173.3 + 38.4) = 361.7 turtles, giving an average of 2.31 nestings per turtle. This figure is much higher than accepted literature values; for example, Márquez et al. (1982) calculated a value of 1.326. Later this figure was revised upwards to 1.47 (1.45 for neophytes, 1.55 for remigrants); but it is clear in the latter calculation (Márquez et al., 1989) that no correction was made for the diminishing probability of observing a multiple nester on 1 all of its nesting emergences. The important implication of this revised estimate is that the estimated total population of nesting female Kemp's ridleys becomes substantially lower than thought, by a factor of 1.326/2.31 to 1.47/2.31. This calculation assumes that there were no four-time nesters, and that there was no intraseasonal tag loss. These assumptions may be unrealistic; if more than 50% of the nests were made by triple-nesters, there may have been a few four-time nesters, any one of which would only have a (0.484)4 chance of being recorded all four times. Some of the unrecorded one-time nesters and recorded three-time nesters could have been composite records for single animals. Indeed, Márquez et al. (1989) noted that four nests were sometimes made in a season. In addition, in 1989, 70 turtles (34.8% of those encountered) had at least one tag scar, and while many of these may have been tagged in previous years, it definitely raises the possibility of intraseasonal tag shedding. The uncertainties in our two assumptions would tend to over-estimate the season's nesting population, which may thus be significantly less than the 361.7 animals calculated. Moreover, since Lepidochelys is known often to nest in successive years (Márquez et al., 1982), one can apply only a modest (and still uncertain) multiplier to estimate the overall (multi-season) nesting population. The population does appear to be a dynamic one, in that each season only a minority of nesting turtles bear tads from previous seasons, and many of the untagged ones show no trace of tag scars and subjectively appear to be "young", although clearly this conclusion could be erroneous. The scenario that seems likely is that the population does receive some recruitment from the 25 years of release of hatchlings at Rancho Nuevo, but that the loss of both juveniles and adults to trawlers is so severe as to negate any significant overall increase in the nesting population. It is not immediately clear if the 2.31 figure applies to seasons other than 1989. This figure may vary, depending upon the ratio of neophytes to remigrants, and possibly other factors also. Indeed, a similar manipulation of the 1985 data does yield a slightly different result. In that year, 706 nestings were recorded (Burchfield and Foley, 1985), and 212 neophyte and remigrant turtles encountered, of which 57 were recorded as nesting twice and 11 three times. This implies that the chance of seeing a given nesting was 0.41, leading to a calculation of 159.6 actual triple nesters, 56.5 actual double nesters, and 114.2 singles, yielding an average of 2.14 nests per turtle. The modest number of double nesters as compared with the numbers of triple or single nesters is odd, and may be an artifact of a bias of some kind. For example, if some tags were poorly applied and fell off quickly, while those that stayed on for the subsequent nesting were likely to remain attached for the rest of the season, we would get a result of this kind. Correction for this bias would increase the 2.14 figure. The revised value for the average number of nests per season does not affect the overall form of the population trend line; but by reducing the estimate of the 1989 nesting population from 545 or 472 (Ross et al., 1989) to 361.7 turtles, the consequences of the loss of even a single adult ridley become correspondingly more serious to the population as a whole. The arguments presented in this paper are in line with the observations of Tucker (1989) that on localized, intensively-patrolled nesting beaches a program of "saturation tagging" can significantly increase the estimates of both average and maximum number of nests-per-season for leatherback, loggerhead, hawksbill, and green turtles and correspondingly reduce estimates of population size. Burchfield, P. M. and Foley, F.J. 1985. Report of Republic of Mexico/USA conservation effort on behalf of Kemp's ridley sea turtles at Playa de Rancho Nuevo, Tamaulipas, Mexico, 1985. U.S. Fish Wildl. Serv., Albuquerque, New Mexico. 59p. 2 Márquez, R., Villanueva, A. and Sánchez, M. 1982. The population of the Kemp's ridley sea turtle in the Gulf of Mexico -- Lepidochelys kempii, p.159-164. In: Biology and Conservation of Sea Turtles, K. A. Bjorndal (Editor). Smithsonian Inst. Press. Márquez, R., Sánchez, M., Diaz, J. and Carrasco, M. 1989. Notes on the reproduction of the Kemp's ridley at Rancho Nuevo, p.107-110. In: Proc. Ninth Annual Workshop on Sea Turtle Conservation and Biology, S. A. Eckert, K. L. Eckert and T. H. Richardson (Compilers). NOAA Tech. Memo. NMFS-SEFC-232. Ross, J. P., Beavers, S., Mundell, D. and Airth-Kindree, M. 1989. The Status of Kemp's Ridley. Center for Marine Conservation, Washington D.C. 51p. Tucker, A. D. 1989. So many turtles, so little time: underestimating fecundity and overestimating populations?, p.181-183. In: Proc. Ninth Annual Workshop on Sea Turtle Conservation and Biology, S. A. Eckert, K. L. Eckert and T. H. Richardson (Compilers). NOAA Tech. Memo. NMFS-SEFC-232. PETER C. H. PRITCHARD, Florida Audubon Society, 1101 Audubon Way, Maitland, Florida 32751 USA. GUEST EDITORIAL : SEA TURTLE CONSERVATION AND MANAGEMENT, THE NEED FOR A UNIT STOCK APPROACH Two major problems in sea turtle conservation and management need attention. The first stems from the Convention on International Trade in Endangered Species' (CITES) definition of a "population" of Appendix I species (New Delhi, India, 1981, Conf. 3.15) as that which occurs within the jurisdiction of [CITES] parties. Rather than using this definition for populations of wide-ranging, migratory species such as sea turtles, the concept of unit stock or management unit should be applied. Well established definitions of unit stock or management unit have been used in the management of many migratory marine resources (e.g., tuna, salmon, whales). Whether one uses ideal or working definitions of unit stock, it is fundamental that unit stock be defined, and this is no less true for sea turtles than for other living marine resources. If the best available data are inadequate to characterize unit stocks of marine turtles, geographic regions (e.g., northern half of the Atlantic Ocean and the Gulf of Mexico for Kemp's ridley, Lepidochelys kempi) could be used instead as starting points for international conservation and management until more data were collected by the usual techniques (tagging, morphometric-meristic, physiological, biochemical, etc.) to delineate the stocks.