This study was based on 15 nesting pairs of horseshoe
crabs and their associated satellite males that spawned
along the beach at Seahorse Key on 31 March and
14–18 April 1995 on both night and day (17–18 April)
tides. We observed, marked and measured a total of 26
mating groups on these dates, but seven were excluded
because no eggs were laid or the eggs were mixed with
those of a female that nested earlier in the season. An
additional four were excluded because paternity was
ambiguous due to allele sharing in two or more of the
putative sires. We do not know of any specific bias that
was introduced by the exclusion of these four mating
groups. These 15 pairs and their 106 associated satellites
completed 40 clutches (two to five clutches from each
pair) with two to four satellites each. (One clutch in the
sample was completed with one satellite when a male left
after completing two clutches with that nesting pair).
Because satellites were free to move among nests, and
because there were no differences attributable to clutch
number, we treated these 40 clutches as independent data
points for analyses of satellites. However, to avoid problems
of pseudoreplication, when the analyses involved
the female or the attached male (such as whether the size
of the female affected satellite male paternity), then we
used data from the first clutch of each nest (N=15).
Descriptive data are reported as meansSE for continuous
data and as proportions for categorical data. We
used a repeated measures analysis of variance (ANOVA)
on ranked data within the general linear model procedure
(GLM, SAS Institute 1990) to evaluate the combined
effects of variables hypothesized to affect paternity
(Agresti 1996).