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



Ti 



zone of intergradation/hybridization of S. I. orientalis 

 with S. l. longirostris is consistent with limited data 

 on movements from tag returns (Perrin et al. 1979a), 

 which indicate "a home range of a diameter of hun- 

 dreds rather than thousands of kilometers." 



Our findings strongly support these conclusions. All 

 30 characters studied showed geographic variation, 

 with two-thirds having demonstrable regional pattern- 

 ing, and 25 of the 30 showing local patterning. These 

 patterns emerge even though data are based on 

 specimens pooled over season and for a number of 

 years; consistent geographic patterns largely would be 

 obscured if animals typically moved long distances 

 within or between years. Clearly, as found by Schnell 

 et al. (1986) for S. attenuata, in S. longirostris "there 

 are notable patterns of geographic variation ... in- 

 dicating that geographic subdivision exists among 

 populations." 



We found concordance of geographic patterns in 

 S. longirostris for a number of cranial characters as 

 noted by Perrin et al. (1991) for external characters. 

 Yet, some patterns are not concordant; in fact, there 

 is a mosaic of patterns involving different characters 

 and/or character suites. For example, many of the tooth 

 counts, toothrow measurements, and rostrum and 

 ramus lengths show very similar patterns of variation 

 (as indicated in Fig. 2), while other characters like 

 Tooth W. have a pattern among blocks that is not close- 

 ly related to that of any other character. Not surpris- 

 ingly, a number of skull widths covary. Overall, the 

 findings for S. longirostris parallel the situation typical- 

 ly found in other mammals where geographic variation 

 in morphological characters has been studied. Some 

 observed patterns may be the consequence of action 

 by selective forces, while others simply result from and 

 are maintained because of isolation by distance. The 

 findings are consistent with S. longirostris being 

 genetically subdivided, stemming from individual 

 animals or groups of animals having relatively limited 

 home ranges. 



For management stocks, Perrin et al. (1991) pro- 

 posed an alternative management scheme where "an 

 'eastern spinner conservation zone' could be devised 

 that would offer appropriate and unequivocal protec- 

 tion to the unique and coherent gene pool of the eastern 

 subspecies." For instance, a zone bounded on the south 

 by 10°N and and on the west by 125° W would encom- 

 pass 84% of the schools that were identified in the field 

 as being composed of "eastern" spinners, and would 

 include very few "whitebelly" animals (Perrin et al. 

 1991). Based on the cranial measures we employed, 

 spinners from the blocks in this portion of the range 

 are very similar; the blocks typically were closely linked 

 in cluster analyses and ordinations. Blocks from most 

 other parts of the range did not show the same degree 



of consistency and concordance. Our data also provide 

 additional biological justification for establishing a 

 geographically defined management zone for S. I. 

 orientalis that, operationally, would be easily under- 

 stood and more effective for management purposes. 

 Perrin et al. (1991) also concluded that data on ex- 

 ternal characters do not support the division of white- 

 belly spinners into northern and southern stocks for 

 management purposes. For cranial features, if one con- 

 siders only eastern blocks, it is possible to achieve a 

 considerable degree of separation between northern 

 and southern whitebelly spinners. However, the situa- 

 tion becomes notably more complex when more wester- 

 ly blocks are added. For virtually all cranial characters, 

 the western blocks group with the more southerly 

 blocks even though they are at the same latitude as 

 blocks to the east containing northern whitebelly spin- 

 ners; the only possible exception is W. Internal Nares, 

 which shows a strong north to south gradient involv- 

 ing all blocks except for one in the vicinity of the 

 Hawaiian Islands (i.e., 0802). The addition of cranial 

 specimens from western locations has provided a more 

 sophisticated picture of geographic variation of 

 S. longirostris in the region under study. 



Morphological-environmental covariation 



Considerable heterogeneity exists in environmental 

 parameters over the range of S. longirostris in the 

 eastern tropical Pacific (see examples of environmen- 

 tal variation in Figs. 11-13). With two circulatory gyres 

 adjacent to the region, one to the north and the other 

 to the south, the eastern tropical Pacific has an 

 easterly-flowing equatorial counter-current from 3° to 

 10°N latitude, and a number of fronts and conver- 

 gences (Wyrtki 1966, 1967). These coupled with 

 latitudinal and other gradients result in substantial 

 spatial differences in environmental characteristics. 



Spotted dolphin/environmental comparisons 



Schnell et al. (1986) evaluated covariation in a similar 

 suite of environmental and cranial morphological 

 features for offshore S. attenuata in the eastern 

 tropical Pacific. The S. attenuata investigation was 

 focused in eastern areas (only 1 of 19 blocks was west 

 of 115°W). Our analysis of S. longirostris covers con- 

 siderably more of the ocean, and includes areas around 

 the Hawaiian Islands, which potentially could have 

 substantially different marine environments. The 

 importance of particular environmental variables, of 

 course, could be quite different when different geo- 

 graphic levels and different-sized areas are considered. 

 Furthermore, environmental influence could well vary 

 between species. Yet it can be instructive to compare 

 results of environmental-morphologic patterns for 



