Variation in Spinner Dolphins (Stenella longirostris) from the Eastern Tropical Pacific Ocean: Sexual Dimorphism in Cranial Morphology

VARIATION IN SPINNER DOLPHINS (STENELLA LONGIROSTRIS) FROM THE EASTERN TROPICAL PACIFIC OCEAN: SEXUAL DIMORPHISM IN CRANIAL MORPHOLOGY MICHAEL E. DOUGLAS, GARY D. SCHNELL, AND DANIEL

J. HOUGH

Oklahoma Biological Survey and Department of Zoology, University of Oklahoma, Norman, OK 73019 Present address of MED: Department of Zoology, Arizona State University, Tempe, AZ 85287 ABSTRACT.-Four meristic and 32 morphometric cranial characteristicswere analyzed for sexual dimorphism in 200 adult spinner dolphins (Stenella longirostris) from the eastern tropical Pacific Ocean (ETP). Specimens were first separated into "whitebelly" and "eastern"forms based on external criteria and compared in a two-way ANOVA of color versus sex. Differences between the two color forms were minor, thus allowing the specimens to be pooled for subsequent analyses of sexual dimorphism. We assessed animals from nine 50 latitude-longitude blocks containing four or more individuals with a two-way ANOVA involving sex and geographic block. Sexual dimorphism was significant in 13 of the 36 charactersand ranged from 0.04 to 11.32%difference. We could correctly identify 67% to 69%of the specimens to sex based on a discriminant function involving six characters (three rostralwidths, a tooth width, the length of the temporal fossa, and one braincase measure). Classification functions are included to classify unknown specimens to sex. In addition, correction terms are included that will enable investigators to adjust measurements and factor out sexual differences in studies where it is advantageous to treat the sexes simultaneously.Trends in sexual dimorphism were very similar to those found in spotted dolphins (S. attenuata) from the same geographic region, although the differences were less pronounced in the spinners.

Sexual dimorphism in morphology is of long-standing biological interest; yet the evolutionary significance of this phenomenon is still actively being debated. Several hypotheses have been advanced concerning the causes and significance of sexual dimorphism. These include sexual selection (Darwin, 1871), differential niche utilization (Selander, 1966), bioenergetic pressures (Downhower, 1976; Ralls, 1976), predation pressures (Bergmann, 1965), type of mating system (Trivers, 1972), and various combinations of these factors (see Sigurjonsdottir, 1981). In addition to theoretical considerations, there also are practical reasons elucidating the degree of sexual differences within species. For example, sexual dimorphism and other forms of nongeographic variation should be evaluated (and possibly corrected for) before geographic variation can be properly assessed. In order to facilitate later geographic variation analysis, Schnell et al. (1985) analyzed and provided correction terms for sexual dimorphism in spotted dolphins (Stenella attenuata), an economically important cetacean found in the eastern tropical Pacific Ocean (ETP). Here we have examined sexual differences in the spinner dolphin (S. longirostris) and compare findings with those for the spotted dolphin. Both species of Stenella are currently affected by tuna fishing activities in the ETP (see Perrin, 1969; Smith, 1983). As part of our overall evaluation of S. longirostris, we include the correction terms needed to treat males and females simultaneously in subsequent analyses of variation. METHODSAND MATERIALS We measured 32 morphometric and 4 meristic characters on skulls and mandibles of 200 museum specimens (see Fig. 1 for sample sizes and grid-block localities). The characters (listed in Table 1) are described and figured in Schnell et al. (1985). Most measures were recorded to the nearest 0.1 mm with dial calipers (characters 4-11, 14-28, 33-35), while others were taken to the nearest 0.5 mm with an anthropometer (12-13) or a large vernier caliper (1-3, 36). Characters29-32 were tooth counts. Most of the characterswere selected from those used by Perrin (1975b), while others either represent slight modifications of previously used measures (5, 17, 18) or new characteristicssuggested by knowledgeable investigators(23, 25-27, 35). J. Mamm., 67(3):537-544, 1986

537

-

538

JOURNAL OF MAMMALOGY

0"?7j -

I

1

2

8

-4 1 13

2---

04

06

02

jI

MALE A1 FEI4ALf

01

03

5

Vol. 67, No. 3

o1

4

3

1

-'k4

0

EASTER 0

05

3 32 /4TEBEL

W

07

01

09

8

2

2

1 22

-

2

05

-

I

X 10

03

9

-

8

07

09

FIG.1.-(A) Known range of S. longirostris in the eastern tropical Pacific Ocean (after Scott, 1981), with numbers of males (above) and females (below) available for each 5? latitude-longitude block (total of 200 specimens). Asterisksindicate nine blocks included in analysis of sexual dimorphism. (B) Numbers of eastern (above) and whitebelly (below) spinner dolphins from each block (total of 193 specimens). Only adult specimens were measured, these being recognized on the basis of the fusion of the premaxillaries with the maxillaries at the distal end of the rostrum (described in Dailey and Perrin, 1973). Sex was assigned as designated on the specimen label; typically, gonadal data were available for the animals used. If there was some question concerning the sex of a specimen, it was excluded from the analysis. Incomplete or damaged specimens also were excluded. Thus, for the variables and specimens assessed, only 0.53% of the total measurements (38 of 7,200) were missing; these were estimated by linear regression onto the character that explained the greatest proportion of the variance for that character over all specimens ("missing data estimator" computer program developed by D. M. Power). Each specimen was assigned to a particular 5? latitude-longitude block of the ETP according to the coordinates of the capture vessel. Blocks are identified by numerical values printed in the left and bottom margins of Fig. 1. The analysis of sexual dimorphism was based on specimens from approximatelythe same geographic area, which enabled us to partition out variance due to geographic differentiation. Specimens were taken over a period of years and in different seasons. There were no evident trends of morphological change in skulls over years. Differences between specimens collected in the winter and summer were analyzed using a two-way analysis of variance (ANOVA; Proc. GLM, SAS Institute Inc., 1982) of geographic block and season. Only two variables (characters29 and 30; number of upper teeth on the left and right sides) exhibited statistically significant interactions (P < 0.01 and 0.05, respectively). None of the characters demonstrated significant seasonal differences. Since no evidence was found in our data for seasonal shifts that would influence the analysis of sexual dimorphism, we did not consider seasonal differences further. Whitebelly-eastern differentiation.-Perrin (1972, 1975a, 1975b) and Perrin et al. (1985) described two color patterns in spinner dolphins of the ETP. One, designated as the "eastern"form, is found to the north and northeast;the other is a "whitebelly" form typically inhabiting areas to the south and southwest. The two forms are considered the same species, and their distributionsoverlap. Figure lB shows the geographic distribution of the 103 eastern and 90 whitebelly spinner dolphins; 3 other specimens were not identified as eastern or whitebelly, and 4 specimens were of a separate "Costa Rican" form that was not included in this part of our analysis. As an initial assessment,we conducted a two-way ANOVA on specimens from the geographic block with the largest number of specimens (0504; see Fig. 3). Of the 8 eastern spinners from this block, 2 were males and 6 were females, whereas the 17 whitebelly spinners included 9 males and 8 females.

TABLE 1.-Geographic variation and sexual dimorphism in Stenella longirostris evaluated for 36 characters and comp with S. attenuata. F-value' Character

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Condylobasal L. L. Rostrum (frm. Base) L. Rostrum (frm. Pterygoid) W. Rostrum (at Base) W. Rostrum (at ? L.) W. Rostrum (at ?2 L.) W. Premax. (at Ih L.) W. Rostrum (at % L.) Preorbital W. Postorbital W. Skull W. (at Zygomatic P.) Skull W. (at Parietals) Ht. Braincase L. Braincase Max. W. Premax. W. External Nares W. Lf. Premax. (md. Nares) W. Rt. Premax. (md. Nares) L. Temporal Fossa W. Temporal Fossa Orbital L. L. Antorbital P. Sep. Pterygoids W. Internal Nares L. Lf. Tympanic Cavity L. Rt. Tympanic Cavity W. at Pterygo. Sutures L. Up. Toothrow No. Teeth (Up. Lf.) No. Teeth (Up. Rt.) No. Teeth (Low. Lf.) No. Teeth (Low. Rt.) L. Low. Toothrow Ht. Ramus Tooth W. L. Ramus

Block

18.51*** 12.04*** 16.52*** 10.47*** 6.88*** 6.55*** 4.14*** 2.99** 25.88*** 36.51*** 35.48*** 5.64*** 11.25*** 13.68*** 4.42*** 2.61* 6.29*** 8.66*** 3.93*** 7.03*** 2.82** 8.26*** 0.86 22.04*** 2.17* 1.57 23.16*** 13.00*** 1.49 1.37 1.31 1.67 10.66*** 14.69*** 4.99*** 15.39***

Meanb Sex

Male

Female

0.28 1.19 0.50 1.49 4.52* 4.62* 4.17* 17.10*** 5.17* 3.19 6.57* 7.78** 6.90** 4.46* 0.05 0.02 0.17 0.05 7.23** 15.82*** 0.26 9.92** 1.76 2.42 0.46 0.07 0.86 2.49 0.31 0.00 0.06 0.50 1.40 5.63* 2.67 0.10

402.82 256.60 298.12 74.06 52.72 44.71 20.72 33.31 139.26 154.73 153.39 130.07 90.40 104.51 62.40 40.21 23.85 37.46 48.51 36.90 39.95 42.86 1.68 43.05 46.09 46.12 43.24 220.60 53.35 53.06 51.28 50.93 215.37 55.07 2.77 344.09

403.96 258.57 299.49 73.43 51.73 43.73 20.15 31.42 137.53 153.43 151.48 128.16 88.94 103.28 62.52 40.16 23.75 37.38 46.99 35.07 40.10 41.54 1.50 42.52 45.80 46.01 42.84 223.09 53.10 53.04 51.38 51.22 217.25 54.23 2.71 344.73

Corr

-0 -0 -0

-0

-0

-1

-0 -0 -0

-0

F-value from main effects two-way analysis of variance (sex vs. 5? block) involving 9 blocks (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Total of 139 individuals for between sexes. b Unweighted mean for the 9 blocks. c Added to all individual female measurements and subtracted from all individual male measurements to correct for sexual differences. d The difference between sexes (males minus females) multiplied by 100, with the resulting value divided by the average of the male and female means. Data on S a

TABLE

2.-Statistics for stepwise discriminant analysis of male and female Stenella longirostris. Character

6 7 8 14 19 28

Vol. 67, No. 3

JOURNAL OF MAMMALOGY

540

W. Rostrum (at l/2 L.) W. Premax. (at 1/2 L.) W. Rostrum (at % L.) L. Braincase L. Temporal Fossa L. Up. Toothrow Constant

F-value to enter

Order of entry

4.03 5.40 21.24 7.45 5.50 7.56

5 6 1 4 2 3

Coefficients' Unstandardized

-0.2419 0.2640 0.3415 0.1307 0.0871 -0.0500

Standardized

-4.1634 4.5427 5.8768 2.2484 1.4980 -0.8607

-12.3775

Classificationfunctionsb Male

Female

-1.525 0.767 3.617 4.424 1.551 0.620

-1.266 0.485 3.253 4.284 1.458 0.674

-370.245

-357.042

a For canonical variable, which in the two-group case is equivalent to the discriminant function. bUsed with original measurements. Add products of measurements and function values to constant; classify as male or female depending on which results in the higher value for its classificationfunction.

Since the number of specimens from block 0504 is relatively small (albeit the largest from any given block), a similar analysis was conducted with specimens combined from blocks where eastern and whitebelly spinners are in sympatry (0501, 0504, 0505, 0506, 0507, 0508, and 0409 of Fig. IB; although we had only eastern specimens from blocks 0507 and 0508, both color types probablyoccur there). This combined sample of 59 specimens included 8 male and 18 female eastern spinners, and 18 male and 15 female whitebelly spinners. Although an analysis of this type is more likely to confound geographic factors with those of color and sex, the statistical power of our analysis is increased with the larger sample sizes. If a significant interaction were found between sex and color for a particular character in our analysis, it would indicate that whitebelly and eastern spinners differ in their levels of sexual dimorphism; lack of interaction would suggest that in subsequent analyses the two forms could be treated together with respect to their sexual differences. Sexual dimorphism.-Since differences between the two color forms were not substantial (see Results), subsequent analyses were conducted on combined samples. A two-way ANOVA was performed on each character with the classes being geographic block and sex. Nine blocks containing four or more specimens of each sex (marked with asterisksin Fig. 1A) were evaluated (a total of 139 specimens). Utilizing all available specimens, we employed stepwise discriminant analysis (Program P7M of BMDP79; Dixon and Brown, 1979) to determine which combination of the 36 characters provides maximum discrimination between males and females given the variability within each sex. Variableswere considered one at a time, with an ANOVA F-statistic (i.e., the F-to-enter) calculated to determine which variable should be added to the discriminant function at each step. These values are conditioned on the variables already added to the function. At each step in the process, the function is recomputed so as to maximize the separation between the two sexes. A variable was not included in the function unless the F-to-enter was at least 4.0. Specimens were then projected onto the resulting discriminantaxis. Also, we obtained classification functions that can be used to assign a specimen of unknown sex to one or the other gender based on the likelihood of membership in that group. Correction for sexual dimorphism.-For the spotted dolphin, Schnell et al. (1985) suggested a procedure for adjusting measurements so that specimens of both sexes can be analyzed simultaneouslyfor geographic variation. We produced similar correctionsbased on morphometric and meristic cranial data for the spinner dolphin. RESULTS the two-way ANOVAs for spinner dolphins from Whitebelly-eastern differentiation.-In block 0504, only 2 of the 36 characters (5.6%) showed a significant interaction term (P < 0.05) between sex and color. These were W. External Nares and W. Internal Nares, neither of which show significant sexual dimorphism (see below). Similarly, when the specimens were assessed from the area of sympatry, only 2 of the 36 cranial characters-Skull W. (at Parietals) and Tooth W.-had significant interactions between color and sex. Only the former character is sexually dimorphic (see below). The proportions of characters exhibiting statistically significant interaction in both of these analyses are about what one would expect by chance alone. Sexual dimorphism.-The two-way ANOVAs of block and sex indicated that three characters

August 1986

>

.

DOUGLAS ET AL.-VARIATION IN STENELLA LONGIROSTRIS

541

EStenella longirostris

z io10 -

-4.2

n

Female

-3.5

-2.8

-2,1

-1,4

-0.7

0.0

0.7

1.4

2.1

2.8

DISCRIMINANT FUNCTION

FIG.2.-Projectionsof spinnerdolphinspecimensontothe discriminant functionaxis,whichprovidesthe maximumseparation of malesandfemalesbasedon thesetof 36 characters. Arrowsindicatemeanprojection valuesfor each sex. had significant statistical interactions (P < 0.05)-W. Temporal Fossa, W. at Pterygo. Sutures, and No. Teeth (Up. Rt.). In addition, the interaction term for No. Teeth (Up. Lf.) was highly significant (P < 0.01). We found statistically significant sexual dimorphism in 13 of the 36 morphological characters(Table 1). Females were slightly larger for several charactersinvolving the length of the rostrum (Table 1), although not significantly so. Males were larger in most other characters, with a number of widths exhibiting statistically significant differences. Geographic variation, which is not addressed in detail in this paper, was found for 30 of the 36 characters (Table 1). Nonsignificant differences were recorded for the four tooth counts, Sep. Pterygoids, and L. Rt. Tympanic Cavity. The classification functions in the discriminant analysis were based on the equal probability of a specimen being male or female. The stepwise discriminant function includes six characters (Table 2), with W. Rostrum (at % L.) being the best single character for differentiating between the sexes (i.e., the first character entered into the discriminant function). The 200 specimens are projected onto the discriminant function in Fig. 2. Although considerable overlap exists, we were able to provide some differentiation between the sexes. The standardized coefficients in Table 2 give a measure of the relative importance of the individual characters in placing specimens on this axis. Unstandardized discriminant coefficients in the table can be used to project additional specimens onto the discriminant axis. When assessing new specimens it is important that measurements be taken as done in our analysis, since differences between sexes are relatively small; measurement bias of individual investigators could reduce the effectiveness of the discriminant function in correctly identifying the sex of additional specimens.

Classification coefficients in Table 1 can be used to evaluate specimens of unknown sex. Measurementsfrom such a specimen are first multiplied by the coefficients from a classification function and the resulting products are added to the constant. The animal is then assigned to the sex for which the resulting classification value is the greatest. For the 104 males in our analysis, 71 (68.3%)were identified correctly using the classification functions. For females, 67 of 96 (69.8%) were correctly allocated. Overall, 69.0% were appropriately identified. The percentages of correctly identified males and females were slightly lower when a pseudo-jackknifed classification procedure was employed (65.4% of males, 68.8% of females, and 67.0% of total sample correctly identified). The jackknifed procedure effectively leaves out the individual specimen being considered, recomputes the coefficientsof the functions, and then evaluates the specimen (see Neff and Marcus, 1980); thus, this approach is likely to give a better indication of the behavior of classificationfunctions when they are applied to new specimens.

Correction for sexual dimorphism.-It is often of interest to evaluate variation in each sex separately. However, when the number of specimens is not large-such as for the spinner dolphin in the ETP-one may receive a better overall picture of variation if the sexes are combined into a single analysis. In such instances, it may be helpful to eliminate the influence of variation due to sexual dimorphism. In Table 1, we have included correction terms that will remove most if not all of the influence

542

JOURNALOF MAMMALOGY

Vol. 67, No. 3

of sex on skull measures. These terms are based on weighted differences between the sexes. We obtained the correction terms by computing the difference between males and females for each character within each of the nine blocks; a weighted mean of the within-block differences was then calculated. The weighting factor for each block was the harmonic mean of the number of males and number of females for that block. A harmonic mean reflects the fact that an estimate of between-sex difference is more reliable when the numbers of males and females are equal than when they are unequal (given the same total number), just as larger samples provide a better estimate than smaller-sized samples. The correction term is one-half of the resulting weighted difference. Measurementsfor a particularcharacter are adjustedupward in specimens of the smaller sex and downward in the larger. Schnell et al. (1985) called such sex-adjusted measures "zwitter" (or intersex) values, reflecting the intermediacy of such measures. DISCUSSION

In his early work on the common dolphin (Delphinus delphis), Fischer (1881) was the first to suggest that there might be intersexual differences in a small delphine. Kleinenberg (1956) included tabular material showing sexual dimorphism for D. delphis in absolute size of the cranium, and relative lengths of facial and cerebral sections. Males were larger than females in absolute size, including the width and depth of the rostrum,although their overall rostrallength was smaller than that of females. Kleinenberg's (1956) work also showed that male harbor porpoises (Phocoena phocoena) had smaller crania than did females. Perrin (1975b) evaluated sexual dimorphism in both spinner and spotted dolphins, while we (Schnell et al., 1985) studied the latter species in greater detail. Using external measurements, Perrin (1975b) showed that female spinner dolphins had longer snouts than males. He had only a relatively limited sample of adult spinner skulls available from the ETP (25 males and 17 females), and analyzed both absolute measures as well as a series of ratios. Although several characters were statistically different, none of these were absolute skull measurements. Perrin (1975b:164) suggested "it is likely that all or most of the 'differences' result from chance alone, that is, from the large number of simultaneous tests," and concluded that sexual dimorphism was in general negligible. In our studies, which were based on larger numbers and geographically partitioned samples, 13 of 36 skull charactersexhibited significant dimorphism. We believe that sexual differentiation is characteristicof spinner dolphins in the ETP. It is therefore likely that correcting for sexual differences would be helpful in the evaluation of geographic and other types of variation. Schnell et al. (1985) indicated that spotted dolphins had a greater number of sexually dimorphic characters (23 of 36; P < 0.05) than we report here for spinners. Sample sizes for the former were substantially larger and, thus, the disparity might at least in part be explained by the relation of sample size to levels of statistical significance (but see below). In both species, males were larger than females for 25 characters,and smaller in 7 others. For two characters, spotted dolphins showed no sexual differences. For two other measures, there was a difference in the sign of deviation, with females being bigger in spinners and smaller in spotted dolphins; however, differences in both cases were very small. In evaluating the extent of sexual dimorphism in the two species, we tabulated the number of cases where the absolute sex differences were greater in one of the species. For the 25 characters where (for both species) the males were larger, spotted dolphins exhibited more pronounced sexual dimorphism in 19 cases, while in only 6 cases were the deviations greater in spinner dolphins. In six of the seven characters where females were larger, the disparity was greater in spotted than in spinner dolphins. Since the spotted dolphin is somewhat larger for some characters than the spinner dolphin, the actual intersexual differences in the former might be expected to be greater. Thus, we also compared percentage differences between the species, which are summarized in Table 1. These findings were identical to those based on absolute differences (see above). Thus, it is clear that the degree of sexual dimorphism in spotted dolphins is greater than in spinner dolphins. How-

August 1986

DOUGLAS ET AL.-VARIATION IN STENELLA LONGIROSTRIS

543

ever, the trends are basically the same for both species, suggesting that common behavioral and/or ecological factors are influencing sexual dimorphism in these dolphins. ACKNOWLEDGMENTS The following individuals provided access to museum specimens: G. L. Worthen (Utah State Univ., Logan); W. F. Perrin and L. J. Hansen (NOAA, Southwest Fisheries Center, La Jolla, California); J. G. Mead and C. W. Potter (National Museum of Natural History, Washington, D.C.); P. J. H. van Bree (Instituut voor Taxonomische Zoologie, Univ. Amsterdam, Amsterdam, Netherlands); W. Z. Lidicker, Jr. (Museum of Vertebrate Zoology, Univ. California, Berkeley); M. L. Johnson and E. Kritzman (Univ. Puget Sound, Tacoma, Washington); D. W. Rice and A. A. Wolman (NOAA, Northwest and Alaska Fisheries Center, Seattle, Washington); D. R. Patten and J. Heyning (Los Angeles County Museum of Natural History, Los Angeles, California); C. P. Lyman and M. Rutzmoser (Museum of Comparative Zoology, Harvard Univ., Cambridge, Massachusetts); L. C. Binford and J. Schoenwald (California Academy of Science, San Francisco); A. Rey (San Diego Museum of Natural History, San Diego, California); G. F. Mees (Rijksmuseum van Natuurlijke Historie, Leiden, Netherlands). Useful suggestions on earlier drafts of this paper were provided by W. F. Perrin, J. Barlow, D. K. W. Au, D. M. Armstrong,and an anonymous reviewer. D. E. Sergeant provided several references, and D. L. Fields (Oklahoma Biological Survey) assisted in various ways during the project. Support was received through Contract 79-ABC-00167 from the U.S. Department of Commerce, National Oceanic and Atmospheric Administration,and Purchase Order 84-ABA-02177 from the National Marine Fisheries Service, Southwest Fisheries Center, La Jolla, California, to the University of Oklahoma.

LITERATURE CITED

in Proc. Sixth Ann. Conf. Biol. Sonar Diving Mammals (T. C. Poulter, ed.). Stanford Res. Inst., phismus der Anatiden als Anpassung an das HohMenlo Park, California, 113 pp. lenbruten. Commentat. Biol., 28:1-10. W. F. 1972. Color patterns of spinner porM. D., ANDW. F. PERRIN.1973. Helminth PERRIN, DAILEY, poises (Stenella cf. S. longirostris) of the eastern parasitesof porpoises of the genus Stenella in the Pacific and Hawaii, with comments on delphinid eastern tropical Pacific, with descriptions of two new species: Mastigonema stenellae gen. et sp. n. pigmentation. Fish. Bull. (U.S.), 70:983-1003. W. F. 1975a. Distribution and differentia(Nematoda: Spiruroidea) and Zalophotrema pa- PERRIN, tion of populations of dolphins of the genus Stecificum sp. n. (Trematoda: Digenea). Fish. Bull. nella in the eastern tropical Pacific. J. Fish. Res. (U.S.), 71:455-471. Board Canada, 32:1059-1067. C. 1871. The descent of man and selecDARWIN, . 1975b. Variation of spotted and spinner tion in relation to sex, 2 vol. Appleton, New York. DIXON,W. D., AND M. B. BROWN(EDS.). 1979. porpoise (genus Stenella) in the eastern tropical Pacific and Hawaii. Bull. Scripps Inst. Oceanogr., Biomedical computer programs. P-series. Univ. 21:1-206. California Press, Berkeley, 880 pp. PERRIN, W. F., M. D. SCOTT, G. J. WALKER, AND V. DOWNHOWER,J. F. 1976. Darwin's finches and the L. CASS. 1985. Review of geographical stocks of evolution of sexual dimorphism in body size. Nature, 263:558-563. tropical dolphins (Stenella spp. and Delphinus FISCHER,P. 1881. C6tac6s du Sud-Ouest de la delphis) in the eastern Pacific. Natl. Marine Fisheries Service, U.S. Dept. Commerce, NOAA Tech. France. Actes Soc. Linn. Bordeaux, 35(4th ser., 5): 5-219 + PI. 1-8. Report NMFS 28, 28 pp. S. E. 1956. Mammals of the Black RALLS, K. 1976. Mammals in which females are KLEINENBERG, Sea and the Sea of Azov. Results of joint biologilarger than males. Quart. Rev. Biol., 51:245-276. cal-commercial dolphin whaling studies. U.S.S.R. SAS INSTITUTE,INC. 1982. SAS user's guide: statistics. SAS Institute, Inc., Cary, North Carolina, 584 Academy of Sciences Press, Moscow, 287 pp. (In pp. Russian; Trans. Serv. No. 4319 [1978] by Translation Bureau [Multilingual Serv. Div., Dept. of SCHNELL, G. D., M. E. DOUGLAS,AND D. J. HOUGH. 1985. Sexual dimorphism in spotted dolphins Secretaryof State of Canada] for the Fisheries and Marine Service [Arctic Biol. Stat., Ste. Anne de (Stenella attenuata) in the eastern tropical Pacific Ocean. Mar. Mamm. Sci., 1:1-14. Bellevue, Quebec].) SCOTT, M. D. 1981. Dolphin stocks in the eastern NEFF, N. A., ANDL. F. MARCUS. 1980. A survey of tropical Pacific. Pp. 97-107, in Report on the multivariate methods for systematics. Amer. Mus. workshop on tuna-dolphin interactions (P. S. Nat. Hist., New York, 143 pp. Hammond, ed.). Inter-American Tropical Tuna PERRIN,W. F. 1969. The problem of porpoise morCommission, La Jolla, California, 259 pp. tality in the U.S. tropical tuna fishery. Pp. 45-48,

BERGMANN,G. 1965. Der sexuelle Grossendimor-

544

JOURNAL OF MAMMALOGY

R. K. 1966. Sexual dimorphismand difSELANDER, ferential niche utilizationin birds.Condor,68:113151. H. 1981. The evolution of sexual SIGURJ6NSDOTTIR, size dimorphism in gamebirds, waterfowl and raptors. Ornis Scand., 12:249-260. T. D. 1983. Changes in size of three dolphin SMITH,

Vol. 67, No. 3

populations (Stenella spp.) in the eastern tropical Pacific. Fish. Bull. (U.S.), 81:1-14. TRIVERS,R. L. 1972. Parental investment and sexual selection. Pp. 136-179, in Sexual selection and the descent of man, 1871-1971 (B. Campbell, ed.). Aldine-Atherton,Chicago, 378 pp.

Submitted 23 July 1985. Accepted 11 November 1985.