Smith et al : Abundance of Limu/us polyphemus in Delaware Bay 



461 



Table 3 



Maximum likelihood estimates of abundance for adult horseshoe crabs iLimultis pnlyphemus) in Delaware Bay during the end of 

 May 2003. Estimates of females and for both sexes combined ("Total") are based on mark-recapture estimates of males and sex 

 ratios among the animals caught and released for this study. Adjusted estimates take into account the possible effect of capture 

 on spawning by reducing releases of males by 0.88, which is an observed relocation rate for radio-tagged males. 



Maximum likelihood estimates 



Adjusted estimates based on relocation 

 rates from radio-tagged males 



Abundance 



90% CI 



Abundance 



90% CI 



Males 



Females 



Total 



16,140,000 



7,350,000 



23,490,000 



9,910,000-22,300,000 



4,520,000-10,160,000 



14,430,000-32,460,000 



13,730,000 8,780,000-19,400,000 



6,250,000 4,000,000-8,840,000 



19,980,000 12,7800,000-28,240,000 



0.001 (909; CI: -0.0007 to 0.003). A model comparison 

 test (likelihood ratio test in Table 2) did not support co- 

 hort-specific recapture probabilities (P=0.77). Although 

 the difference in relative risk was in the direction of a 

 temporary delay in spawning, the difference in recap- 

 ture probabilities tended to be in the opposite direction. 

 On two out of the three recapture occasions, the prepeak 

 release cohort had a higher recapture probability than 

 the preseason cohort. The maximum likelihood estimate 

 of abundance of males with the use of model 2 (Table 

 2) was 20% higher for all releases (16.1 million crabs) 

 than for prepeak releases only (13.4 million). A tempo- 

 rary delay in spawning would tend to cause abundance 

 estimates based on late-releases to be higher than esti- 

 mates including early releases, but in fact the opposite 

 occurred. Thus, based on three lines of evidence, it is 

 unlikely that there was a temporary delay in spawning 

 due to the capture and tagging of males. 



Another way in which tagged animals may behave 

 differently from untagged animals is that tagged ani- 

 mals may forego spawning altogether. In 2004, we radio 

 tagged horseshoe crabs using the same capture process 

 that was used to tag animals in 2003. The radio-tagged 

 horseshoe crabs were detected by radio receivers at 

 high tide when the radio-tagged crabs emerged from 

 the water to spawn. An array of fixed station receivers 

 ensured nearly complete coverage of spawning habi- 

 tat in Delaware Bay. Among radio-tagged males, we 

 observed that age-specific relocation rates were 0.44, 

 1.00, and 0.74 for young, middle, and old-aged males, 

 respectively. The frequencies of these age groups in the 

 end of May 2003 tag releases were 0.07, 0.62, 0.31 for 

 young, middle, and old-aged males. Thus, the average 

 relocation rate predicted for the 2003 releases would be 

 0.88 (i.e., 0.44x0.07-t-lx0.62-h0.74x0. 31=0.88). There are 

 several reasons for a failure to detect radio-tagged ani- 

 mals, namely behavioral response, movement beyond the 

 range of radio receivers, transmitter loss or failure, and 

 animal mortality. It is also possible that some adults 

 migrate but do not spawn in a given year. Nevertheless, 

 to be conservative, we adjusted the tag releases by the 

 observed relocation rate (0.88) and computed estimates 

 using the reduced releases. The 12% reduction in re- 



leases resulted in a 15% reduction in the abundance 

 estimates (Table 3). From currently available estimates 

 of fecundity (88,000 eggs per female; Shuster and Bot- 

 ton, 1985), we estimated that egg production in 2003 

 was 5.5x1011 (90% CI: 3.5x10" to 7.7xl0ii). 



The adjusted estimates of abundance were 13.7 mil- 

 lion (90% CI: 8.8 to 19.4 million) for males and 6.25 

 million (90% CI: 4.0-8.8 million) for females (Table 3). 

 Landings in New Jersey and Delaware during 2003 

 were 2.4% (90% CI: 2-4%) of abundance estimates 

 (Table 4). When landings from Virginia and Maryland 

 are included, landings during 2003 were 4% (90% CI: 

 3-6%) of abundance. Harvest rates were similar for 

 males and females because the sex ratio in the landings 

 was similar to the ratio observed in our fishery-inde- 

 pendent catch. We believe that our fishery-independent 

 catch was a representative sample of mature animals 

 in the bay. 



We caught horseshoe crabs prior to and during the 

 spawning migration. Thus, a comparison of pre- and 

 postmigration catches could indicate the proportion 

 of the population that over-wintered in the bay. We 

 caught, on average, 18 adults per 15-min. tow on a ves- 

 sel with two 2.3-m dredges from 25 March to 3 April 

 2003 prior to the spawning migration, which appeared 

 to begin in mid-April. The catch-per-tow was 39 adults 

 per tow during the period from 13 April to 8 May 2003 

 and 60 adults per tow during 28 to 30 May 2003. Thus, 

 we were catching approximately one third (18/60 = 0.3) 

 of the animals prior to spawning migration; this frac- 

 tion could represent the proportion of the population 

 that over-wintered in the bay and that did not migrate 

 to the ocean between 2002 and 2003. 



Discussion 



There have been few attempts to estimate abundance 

 of adult horseshoe crabs in Delaware Bay during the 

 spawning run when the population is spatially concen- 

 trated. Shuster and Botton (1985) estimated population 

 size from surveys on Delaware Bay beaches. However, a 

 large portion of the bay was not included in the target 



