Douglas et al : Geographic variation in cranial morphology of Stenella longirostns 



71 



The W. Rostrum (at Base) exhibits covariation with 

 three of the environmental variables based on the 

 Mantel test (Table 8), two of which involve thermocline 

 depth. The matrix correlation (Table 8) for W. Rostrum 

 (at Base) with Thermocline Depth (Winter) is substan- 

 tial (0.598), as it is with environmental component II 

 (0.508). Mean values also showed an association of W. 

 Rostrum (at Base) with the thermocline variables and 

 Water Depth, as well as with three other environmen- 

 tal variables (Table 7). 



Discussion 



Sexual dimorphism 



Our analysis of sexual dimorphism extends the studies 

 of Douglas et al. (1986) in terms of additional specimens 

 and minor adjustments in previously collected data. We 

 found 15 of 30 variables were statistically dimorphic. 

 Douglas et al. (1986) identified 13 of these as showing 

 sexual differences, the increase of two characters 

 (Postorbital W. and Tooth W) being due primarily to 

 increased numbers of specimens available. In five 

 characters, including two rostral and two toothrow 

 lengths, measurements for females were slightly larger 

 than for males, although the differences were not 

 statistically significant. For all other measures, males 

 are larger than females, including all 15 where statis- 

 tical significance was found. 



Geographic variation 



In 1889, when the existence of spinner dolphins in the 

 eastern tropical Pacific was not yet known. True in- 

 dicated that the absence of adequate samples made 

 very difficult the task of taxonomically evaluating 

 species in the genus Stenella. By the 1970s, Perrin 

 (1975b) had available considerably more material for 

 a monographic treatment of S. attenuata and S. longi- 

 rostris from the eastern tropical Pacific and was able 

 to make significant advances with respect to our 

 understanding of morphological variation of Stenella. 

 However, his work on S. longirostris also was hindered 

 by the paucity of skeletal material from parts of the 

 range. For our study, many additional specimens of 

 S. longirostris were available. On the whole, our results 

 are strongly supportive of those obtained by Perrin 

 (1975b) for cranial characteristics, but we also have 

 been able to substantially extend his analyses. 



Perrin (1972) conducted an initial analysis of geo- 

 graphic variation in color patterns of S. longirostris 

 in the eastern Pacific Ocean. He found geographic 

 variation, particularly in the "dorsal field system" of 

 coloration, which overlies a basic general pattern. 



These and other data suggested differentiation into 

 eastern, whitebelly, and Hawaiian forms. The differ- 

 ences were analyzed further by Perrin (1975a,b), who 

 indicated that the whitebelly form was in some ways 

 similar to the Hawaiian form, but had a proportionately 

 smaller beak. Perrin (1975b) described but did not name 

 four races— the three mentioned above plus a Costa 

 Rican form which occurs off the coast of Central 

 America. Perrin et al. (1979b) evaluated possible dif- 

 ferentiation in S. longirostris involving animals found 

 south of the Equator in the eastern Pacific Ocean. They 

 concluded that these S. longirostris are morphological- 

 ly distinct from those to the northeast; characteristics 

 showing such a trend include coloration, size, shape, 

 and skeletal measures. 



Recently, Perrin (1990) named and described three 

 subspecies: S. I. longirostris (Gray's spinner dolphin), 

 S. I. orientalis (eastern spinner dolphin), and S. I. 

 centroamericana (Central American spinner dolphin). 

 In our geographic variation assessment, we purposely 

 included all adult specimens available, without an 

 attempt a priori to differentiate previously described 

 forms. However, the differences among the named 

 forms are reflected in our results for cranial char- 

 acteristics. 



In our analyses, the Central American spinner 

 dolphin of Perrin (1990) is shown to be different from 

 other S. longirostris by the positioning of block 0516 

 on principal component II; it had the lowest value of 

 any of the blocks (see Fig. 3). Three of the four speci- 

 mens in block 0516 exhibit characteristics of Central 

 American spinners (as does one of the nine from 0615). 

 Character associations with the second principal com- 

 ponent—summarized in Table 3— suggest that, after 

 taking into account general size (summarized in and 

 mathematically removed by component I), animals of 

 this subspecies have relatively longer toothrows, 

 greater numbers of teeth, a narrower skull at the 

 parietals, and a shallower braincase than S. longirostris 

 from other areas. Perrin (1990: table 2) provided com- 

 parative measurements and counts for S. I. longiros- 

 tris and S. I. centroamericana, which show the Cen- 

 tral American form to have longer toothrows and 

 greater numbers of teeth. He did not include data on 

 Ht. Braincase, but did characterize the Central Ameri- 

 can form as having a relatively long and narrow skull. 

 While Skull W. (at Parietals) is slightly greater for S. I. 

 centroamericana than S. I. longirostris (Perrin 1990: 

 table 2), the former has a relatively narrow skull given 

 its considerably greater length. Additional S. I. centro- 

 americana specimens are needed, since the diagnosis 

 and understanding of cranial variation in this sub- 

 species still is based on very few animals. 



Perrin (1990) suggested the existence of a zone of 

 hybridization/intergradation between S. I. longirostris 



