2 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 124 



Weber 1795, Callinectes Stimpson 1860, and Arenaeus Dana 1851 by 

 other than the classical techniques of descriptive systematics. 



The work began by detailed comparisons between pairs of species 

 in each of the genera. Rathbun (1930) has listed eight pairs of analo- 

 gous species (or twin species, or geminate species) in which one of 

 each pair is a western American species and one an eastern. Most of 

 these are not clearly detectable by a classical approach (Garth and 

 Stephenson, 1966), which instead has suggested "confused" relation- 

 ships between groups of western and groups of eastern species. Nine 

 of the ten western "species" (including one subspecies; authors and 

 dates of species are given in table 1) of Portunus appear closely related. 

 Several different dichotomous keys can be devised for their separation, 

 but none has obvious precedence for convenience or indications of 

 relationships. (The key to the western species that eventually was 

 adopted employed an initial pentachotomy.) It seemed that all 

 characters had equivalent hierarchial significance. The problems of 

 establishing a hierarchy are emphasised by the fact that Rathbun 

 (1930) had used two subgenera, Portunus and Achelous de Haan 

 1833, that are linked by continuous variation within one species, 

 P. xantusii. 



To some extent the present work was a trial of numerical techniques 

 and initially involved a small number of species, the nine western 

 species of Portunus. When the eastern species of the genus were added, 

 the increased complexity of information gave added convenience to 

 the numerical methods. 



As the work progressed, it was widened to include the relationships 

 among Callinectes, Arenaeus (which is very close to it; see Garth and 

 Stephenson, 1966, p. 52), and Portunus. Callinectes is a particularly 

 interesting case. Stephenson and Campbell (1959, p. 88) questioned 

 whether Callinectes differs sufficiently from the general span of the 

 genus Portunus for it "to have more than the status of a subgenus if 

 such are to be recognised." 



When Callinectes species are compared with western American 

 Portunus species, there are numerous differences. Garth and Stephen- 

 son (1966), partly influenced by preliminary results of the present 

 work, retained Callinectes as a genus. It is diagnosed by three features 

 only (Garth and Stephenson, 1966, p. 42): (1) male abdomen J_- 

 shaped, which is shared with certain Indo-West Pacific species of 

 Portunus and with Arenaeus; (2) anteroexternal angle of merus of 

 third maxilliped strongly produced outward, which is shared with 

 many species of Portunus; (3) wrist of cheliped without an inner 

 spine, which is the only unique feature. 



When eastern American Portunus species are considered, the 

 morphological gap to Callinectes becomes partly bridged. If the 



