no. 3645 PORTUNIDAE — STEPHENSON, WILLIAMS, LANCE 7 



identified from type-material or from specimens identified by Rathbun. 

 Most Indo-West Pacific species were known from previous investiga- 

 tions (Stephenson and Campbell, 1959; 1960) or from types; however, 

 one was based upon a recently collected Australian specimen related, 

 but not identical with, P. trituberculatus. (Further study showed this 

 to be an unusual specimen of P. pelagicus, which differs in only a single 

 tabulated feature from P. trituberculatus: it lacks the tubercles on the 

 carapace: feature 13 in table 1). Another Indo-West Pacific species, 

 P. convexus, was based upon the literature, mostly from Crosnier 

 (1962). 



To simplify later treatment, all species and subspecies are referred 

 to in the text below as "species." Throughout the text, ringed numbers 

 that accompany specific names are those listed in table 2. 



Results 



Our investigation considered, first, certain specific questions relevant 

 to the taxonomic problem and, second, the larger scale inter-relation- 

 ships and the evidence they provide for parallel evolution. Conclusions 

 were based, first, on inspection of the NFD and TD values obtained 

 from the tabular data of table 3, and, second, on the 3-dimensional 

 ordination of the numerical models. Since 2-dimensional sections 

 proved somewhat unrevealing, 3-dimensional models were con- 

 structed. Figures 1 and 2 have been drawn from oblique photographs 

 of the model obtained from the basic GOWER ordination. Several 

 of the plotted points in this model may be regarded as spuriously 

 close together since substantial deviations in the fourth, fifth, and 

 sixth axes are suppressed. Each instance of a deviation greater than 

 100 scale units is shown by a black spot in the figures; a considerable 

 number of such discrepancies is evident. Figures 3 and 4 have been 

 derived from similar photographs of the model obtained from the 

 NEWGOWER ordination, which had been reduced by iteration to 

 three factor axes. 



Specific Questions 



Analogous species. — Consideration of both tabular data and the 

 models indicated that none of the following pairs of species of Portunus 

 listed as analogous by Rathbun (1930) can be considered analogous 

 (pairs are listed with western species first) : P. xantusii xantusii (§)/ 

 P. gibbesii ©; P. brevimanus ©/P. sjpinimanm (§); P. angustus 0/ 

 P. ordwayi ©; and P. iridescens ©/ P. spinicarpm (f§). 



While the general conclusions from the two methods are identical, 

 the models distort some of the detailed relationships of the above 

 species to other species; for example, in the models, P. iridescens (£) 



