74 



Fishery Bulletin 90(1), 1992 



these two dolphin species with broadly overlapping 

 geographic ranges in the tropical Pacific Ocean. 



Sea surface temperatures (variables 6 and 7; July 

 values depicted in Fig. 12) have negative correlations 

 with a large number of morphological features in both 

 studies (Table 7 and Schnell et al. 1986: table 6), in- 

 dicating a general trend of larger animals in warmer 

 waters. Surface Salinity (Fig. 13B) exhibits relatively 

 strong morphologic correlations in both studies, reflect- 

 ing a pattern that has both east-west and north-south 

 components. Also, Thermocline Depth (Summer), 

 which has relatively low values in northern localities 

 and higher numbers in blocks as one proceeds to the 

 west and south, is positively associated with a number 

 of morphological measures in S. longirostris (Table 7), 

 and covaries with S. attenuata cranial features as well. 



Solar Insolation (Jan.) registers a north-south gra- 

 dient. Our S. longirostris study produced virtually no 

 significant correlations with this measure, while there 

 were numerous positive correlations in the S. attenu- 

 ata investigation. In the eastern portion of the S. longi- 

 rostris range, a number of cranial features have north- 

 south gradients, but the overall statistical association 

 is negated with the addition of the western blocks, 

 where animals often (irrespective of latitude) exhibit 

 characteristics similar to those found in southern areas. 

 The same findings were obtained for Solar Insolation 

 (Ann.). 



Three environmental variables— Water Depth (Fig. 

 IIB), Thermocline Depth (Winter), and Surface Dis- 

 solved Oxygen— are positively correlated with cranial 

 measures in S. longirostris, but show few of these 

 associations in S. attenuata. Again, the differences in 

 findings simply may reflect the inclusion of a wider 

 geographic range of blocks in the S. longirostris study. 



Schnell et al. (1986) indicated that for S. attenuata 

 it would be helpful to have additional samples, par- 

 ticularly from western locations. They suggested that 

 "Such a geographic broadening of representation may 

 enable investigators to separate, at least in part, 

 environmental-morphological correspondences that 

 reflect causal relationships from trends [in morphology] 

 maintained primarily as a result of isolation by dis- 

 tance." For S. longirostris, where additional western 

 blocks are now represented (albeit in some cases with 

 very limited samples), it is clear that the gradients in 

 a relatively large number of cranial characteristics are 

 not simply north-south trends, but rather what Perrin 

 et al. (1991) described as a radial pattern. From north- 

 eastern blocks, these characters in S. longirostris 

 change more-or-less gradually as one moves to the 

 south, the southwest, or the west. There are several 

 environmental variables exhibiting this type of pattern 

 (e.g.. Surface Salinity; Fig. 13B). At the same time, the 

 January and July sea surface temperatures (for July 



values, see Fig. 12) have a predominantly north-south 

 orientation (with the Hawaiian Island blocks being 

 lower than expected, given their latitude) and are cor- 

 related with the largest number of cranial characters 

 (see Table 7). When additional specimens of S. at- 

 tenuata become available from westerly blocks not 

 represented in samples available to Schnell et al. (1986), 

 it will be of interest to determine whether patterns of 

 cranial variation (and covariation with environmental 

 measures) in this species will mirror those we have 

 found for S. longirostris. 



Significance of covariation with environmental 

 measures While previous literature has little infor- 

 mation on the relation of environmental and cranial 

 variation in S. longirostris, other investigators (Au and 

 Perryman 1985, Reilly 1990) have evaluated physical 

 environmental parameters with respect to distributions 

 of S. longirostris and several other species in the 

 eastern tropical Pacific. They pointed out that the 

 highest school densities for S. longirostris are in the 

 area off the Mexican coast, which also is the most 

 tropical and least seasonally variable portion of the 

 range. We have demonstrated notable associations of 

 cranial variation with physical environmental char- 

 acteristics. Analyses involving environmental-cranial 

 correlations, by their very nature, are descriptive and 

 do not provide direct information on causal factors per 

 se. Nevertheless, they clearly indicate that between 

 areas where animals are different cranially, there often 

 are marked habitat differences involving the physical 

 environment. 



The first two environmental principal components 

 (Fig. 10) describe independent, orthogonal environmen- 

 tal patterns: component I has a general configuration 

 of high values between 5° and 15°N, slightly lower 

 values further to the north, and low values to the south 

 (Fig. lOA), which is overlain by basically an east-to-west 

 trend summarized in environmental principal compo- 

 nent II. The important individual environmental covari- 

 ates with morphological characters include surface 

 temperatures, salinity, and measures of water depth. 

 The physical environmental differences reflected by the 

 principal components, as well as by individual environ- 

 mental measures, describe basic habitat differences and 

 likely reflect, indirectly, geographic differences in 

 available prey species and their abundances. Given the 

 marked environmental differences exhibited in the 

 range of S. longirostris, the most surprising result 

 would have been if this species had been relatively 

 uniform geographically in cranial features— clearly, this 

 is not the case. Our initial assessment of morphologic- 

 environmental covariation further underscores the 

 appropriateness of treating different parts of the range 

 of S. longirostris in the eastern tropical Pacific 



