478 
Fishery Bulletin 99(3) 
Figure 1 
UPGMA ( unweighted pair-group method with arithmetic averaging) analysis of Kimura 2-param- 
eter genetic distances based on 547 base pairs of 16S mtDNA among seven haplotyopes of Cal- 
linectes maracaiboensis and C. bocourti (see Table 2 for species designation) corresponding to 14 
specimens from Golfete de Cuare (haplotypes 1-4) and El Mojan (haplotypes 5-7) (both Venezu- 
ela) and two populations of C. sapidus. 
294 base pairs obtained for one Callinectes bocourti from 
Colombia did not reveal any difference from ht-1, ht-3, or 
ht-5, suggesting that genetic homogeneity in these species 
can be expected in other Caribbean localities. 
Comparing sequences from C. danae, C. ornatus , C. 
sapidus, and Portunus ordwayi with those of C. bocourti 
and C. maracaiboensis allowed us to postulate phylogenet- 
ic relationships among these selected portunid taxa (Fig. 
2). Marked genetic differences (Table 3) characterized all 
intra- and interspecific separations within the genus Cal- 
linectes, except in pairings of C. bocourti and C. maracai- 
boensis. 
Discussion 
Comparison of 15 sequences of the 16S mtDNA of Cal- 
linectes maracaiboensis and C. bocourti from within and 
outside Lake Maracaibo revealed no consistent differ- 
ences between these two species of swimming crabs. These 
results support observations from previous studies in 
which morphological characters have proven unreliable in 
separating these two species. Rather than constituting a 
separate species, C. maracaiboensis would appear from 
these findings to represent only a phenotypic extreme of C. 
bocouj'ti, a species of known morphological plasticity (Wil- 
liams, 1974). 
We are highly confident that our sequenced samples of 
C. maracaiboensis represent the population originally as- 
signed to that species. In addition to assuring that our 
samples were topotypic, we confirmed individual specimen 
identities by direct comparisons with deposited paratypes 
of Taissoun (IVIC-LEGP 425). However, we cannot make 
similar claims for the analyzed materials of C. bocourti. 
Although published records (Williams, 1984) indicate that 
C. bocourti is widely distributed (from Jamaica and Belize 
to Brazil, occasionally northward to Florida, Mississippi, 
and North Carolina), the type locality is southern Belize. 
To date, topotypic materials of C. bocourti have not been 
available to us for use in either our sequencing or mor- 
phological comparisons with Venezuelan material; such 
comparisons should, however, ultimately be undertaken 
to firmly anchor the synonymy we have proposed. Among 
specimens that we would presently assign to C. bocourti 
on the basis of structural features, one from the Rio Sinu 
of Colombia shared complete identity in almost 300 base 
pairs of mtDNA with the most common haplotypes from 
Venezuela. This finding at the very least suggests a large 
overall range for materials that we refer to as C. bocourti 
and a relative homogeneity among forms that we have as- 
signed to this name on the basis of structural characteris- 
tics, even over large geographic distances. 
The reported phylogenetic relationships and genetic dis- 
tances (Fig. 1) reflect a low level of genetic divergence 
among the seven haplotypes of C. maracaiboensis and C. 
bocourti, especially when compared with sequences of two 
populations of C. sapidus. The tree resulting from the phy- 
logenetic analysis (Fig. 2) also suggests that C. maracai- 
boensis and C. bocourti are closer to C. sapidus than to C. 
ornatus, C. similis, and C. danae, which is in accord with 
preliminary findings of Norse and Fox-Norse (1979). How- 
ever, our results must also be regarded as preliminary 
because other species of Callinectes should be sequenced 
before we can claim full understanding of sister-species re- 
lationships. Published studies to date that have attempted 
to separate questionable brachyuran species on the basis 
of 16S mtDNA have consistently reported at least a few 
nucleotide differences that are diagnostic for the species 
in question (Geller et ah, 1997; Schneider-Broussard et al., 
1998; Schubart et al. 1998; Schubart et al., 2000). Only 
the Mediterranean species Brachynotus sexdentatus and 
B. gemmellari may constitute an exception to this pattern 
because they are identical in the sequenced fragment of 
16S mtDNA (Schubart et al., in press). 
Comparison of our sequences with those from a previ- 
ous study (Geller et al., 1997) revealed important, appar- 
ently intraspecific differences between specimens of C. or- 
natus as well as between specimens of C. sapidus. Genetic 
