FISHERY BULLETIN: VOL. 81. NO. 4 



Now, three independent investigations which con- 

 cern elements of this complex have produced results 

 which bear on the status of these forms. The 

 emphasis and method of each study are different,, but 

 each employs reference specimens from the crusta- 

 cean collection of the U.S. National Museum of 

 Natural History. The results of the studies are 

 arranged here as a trilogy. To simplify nomenclature, 

 the first (this paper) presents a taxonomic revision of 

 the complex but rests in part on evidence in the 

 following parts; the second (Sullivan et al. 1983) 

 employs electrophoretic analyses of hemocyanins in 

 the four forms (species) from the Carolinas, the gulf 

 coast of Louisiana and Florida, and the Florida Keys; 

 the third (Reames and Williams 1983) deals with 

 morphological and ecological features of two forms 

 (species) in local populations from southern 

 Alabama. 



This paper contains specific diagnoses of these 

 populations partly conceived by Rathbun (1930) 

 through designation as "forms" and includes 

 diagnoses of two additional species from the 

 southern part of the range; it also gives full syn- 

 onymies of published descriptions with variant 

 spellings of names and critical distributional records, 

 outlines geographic distributions, and provides a key 

 for determination. Some of the older records already 

 included in published synonymies can be associated 

 with nominal species by place of origin. 



MATERIALS AND METHODS 



Conclusions are based on results presented in Sul- 

 livan et al. (1983) and Reames and Williams (1983), 

 on study of specimens in the crustacean collection of 

 the U.S. National Museum of Natural History, 

 Smithsonian Institution (USNM), on type materials 

 in the Yale Peabody Museum (YPM), and photo- 

 graphs of type material in the Museum National 

 d'Histoire Naturelle, Paris (PM). 



Form and ornamentation of the carapace have been 

 the main bases for distinguishing "varieties" in this 

 complex in the past, but imprecision in setting limits 

 for the variations led earlier workers to arrive at iden- 

 tities which form improbable species distributions 

 over a broad range of latitude between the northern 

 and southern temperate regions. Factors that con- 

 tributed to these interpretations include ill- defined 

 "varietal" differences as well as allometric changes, 

 sexual dimorphism, and wear of exoskeletal parts. To 

 increase the number of definitive characters, other 

 features such as shape and ornamentation of the 

 chelipeds and color pattern were employed, although 

 the ranges of variation were increased by these 



additions. To avoid the consequent confusion, dis- 

 tributional patterns were outlined by noting the 

 occurrence of "typical" specimens in a large se- 

 ries; once that framework was perceived, the vari- 

 ants could be recognized as such and placed 

 properly. 



Mensuration data are treated in two ways. Species 

 diagnoses include proportional measurements, i.e., 

 one measure expressed as a percent of another. 

 Minimum values, maximum values, and arithmetic 

 means are presented in a table to supplement the 

 accounts of each species. In order to test for differ- 

 ences among species, the analysis of covariance 

 (ANCOVA) procedure is used. This procedure tests 

 null hypotheses concerning 1) overall coincidence, 2) 

 equality of slopes, and3) equality of intercepts of two 

 or more regression equations. If hypothesis 1 is re- 

 jected at a prespecified level (P= 0.05 in this study), 

 hypotheses 2 and 3 are then tested If the null 

 hypotheses concerning the slopes or intercepts of 

 three or more groups are rejected, the Newman- 

 Keuls multiple range test (MRT) is used in order to 

 determine which regressions are different from 

 others. It is not uncommon for the ANCOVA pro- 

 cedure to reject a null hypothesis in which the MRT 

 procedure is unable to detect differences. It is 

 generally accepted that the ANCOVA procedure is a 

 more powerful test than the MRT. 



Statistical tests were performed on the DEC Sys- 

 tem 10 Computer' of the National Marine Fisheries 

 Service, using computer programs written and main- 

 tained by Joseph L. Russo for the NMFS Sys- 

 tematics Laboratory. 



All statistical procedures and notation used in this 

 investigation follow those presented by Zar (1974), 

 with the exception that the probability associated 

 with the calculated value of the F statistic of the 

 ANCOVA procedure is generated by the computer 

 programs which for the purpose of simplification 

 and clarity is used instead of the calculated value of 

 the F statistic with its associated numerator and 

 denominator degrees of freedom. The indepen- 

 dent variable used for all regressions is carapace 

 width. 



First pleopods of male crabs were studied with the 

 aid of a light microscope and scanning electron 

 microscope. 



Citations in synonymies are limited to descriptive 

 or distributional accounts that can be associated with 

 species. Many references that cannot be assigned to 

 species with certainty are not included. 



'Reference to trade names does not imply endorsement by the National 

 Marine Fisheries Service, NOAA. 



864 



