William F. PerrinPrisciffa A. AkinJerry V. KashlwadaLa Jolla Laboratory. Southwest Fisheries Science CenterNational Marine Fisheries Service. NOAAP.O. Box 271. La Jolla. California 92038-0271

Geographic Variation inExternal Morphology of theSpinner Dolphin Stenella longirostrisin the Eastern Pacific andImplications for Conservation

Abstract. - Variation in colorpattern, dorsal fin shape. and bodylength exhibit sharp north/south gra­dients centered at about 5-100 N lati­tude and east/west gradients atabout 120-125°W longitude. A con­servation zone with boundaries inthese regions would provide protec­tion for the morphologically uniqueeastern spinner dolphin SteneUa langi­rost1'is orientalis. A radial pattern ofgeographic variation in the easternPacific and a complex pattern of dis­cordant variation outside the corerange ofS. l. mentalis suggest thatthe present separate management of"whitebelly" spinner dolphins (whichcomprise a broad zone of hybridiza­tionlintergradation between S. l. 01'1.­e1'~talis to the east and the pantrop­ical spinner dolphin S. l. longi:rostristo the west and southwest) north andsouth of the Equator may not bejustified on the grounds of conserva­tion of distinct populations.

Manuscript accepted 15 March 1991.Fishery Bulletin. U.S. 89:411-428 (1991).

Spinner dolphins in the eastern trop­ical Pacific vary geographically incolor pattern (Perrin 1972), in exter­nal and skeletal size and shape (Per­rin 1975, Perrin et al. 1979a, Schnellet al, 1982 and 1985, Douglas et al,1986), and in reproductive seasonal­ity (Barlow 1984). Perrin (1975) andPerrin et al. (1979a) described sev­eral forms that have served as stockunits for management of populationsof dolphins killed incidentally in theinternational purse-seine fishery foryellowfin tuna in the region (Perrinet al. 1985). Kills in recent years havebeen on the order of 100,000 annual­ly. The management units have beenknown as the "Costa Rican," "east­ern," "northern whitebelly," and"southern whitebelly" spinner dol­phins. The geographic ranges of theeastern and whitebelly forms overlapbroadly (Perrin et al, 1985).

The model used in studies and, toa lesser degree, management of thespecies in the region has been one ofseveral discrete, more-or-Iess repro­ductively isolated stocks with somegeographic overlap. However, asmore morphological data have ac­cumulated, a different picture hasemerged. For example, while theboundary between the present north­ern and southern whitebelly manage­ment stocks was drawn latitudinally,

the overall pattern of variation in theeastern Pacific now appears radial(Schnell et al. 1982 and 1985, Perrinet al, 1985). Further, the results ofa genetic study indicate substantialexchange of at least mitochondrialDNA between and among the east­ern and whitebelly forms (Dizon et al,1991). Most recently, Perrin (1990)described two subspecies, Ste-nellalongirostris centroamericana andS. l. orientalis, based on the CostaRican and eastern forms, respective­ly, and concluded that the whitebelly"forms" (Fig. 1) collectively comprisea broad zone of intergradation orhybridization between the easternsubspecies and a third subspecies tothe west and southwest, the pan­tropical spinner dolphin S. l. longi­rostris, a form occurring in the cen­tral and western Pacific, Indian andAtlantic Oceans (Perrin et al. 1981;Gilpatrick et al. 1987).

In this context, the purpose of thepresent study was to reexamine vari­ation in external morphology on afiner geographical scale than hadbeen used in earlier studies. Speci­mens previously identified as "CostaRican," "eastern", or "whitebelly"based on the modal appearance ofadult animals in the schools fromwhich they came were pooled in theanalyses. In this way we hoped to

411

412 Fishery Bulletin 8913). 1991

"··l"· . .." ~\>.:......

Figure 1Typical appearance of adult males of the pantropical spinner dolphin Stenella longirostris longirostris (top),the eastern spinner dolphin S. l. orientalis (bottom), and a hybrid/intergrade "whitebelly spinner" (center).

gain a better understanding of the morphological gra­dients and other geographical patterning involved inthe relationships among the Central American, eastern,and pantropical subspecies and to contribute to thedevelopment of approaches to conservation and man­agement of the several populations.

Materials and methods

Source of the data

We examined variation in the ventral field and capecomponents of the color pattern (terminology of Per­rin 1972), the shape of the dorsal fin, and body length.

Perrin et al.: Geographic variation in morphology of Stenella longirostris 413

Figure 2Character states for coding of ventral field, cape, and dorsal fin in spinnerdolphins.

sketch. The sketches themselves already incorporatedsome artificial stratification. For example, in someanimals the darker region yielding the step of Code 4is only faintly evident. Some observers indicated thepresence of this faint feature, leading us to code theanimal as Code 4; others undoubtedly overlooked suchfaint markings or failed to record them, leading us tocode the animals as Code 5. While such errors couldbe expected to blunt the resolving power of the anal­yses, there is no reason to suspect that they inject asystematic spatial bias that would affect the accuracyof the conclusions; the data were collected by a largenumber of observers working over large areas.

The cape (Fig. 2) was coded as "not noted in thesketch" (Code 1) or "noted in the sketch" (Code 2). Ina living or freshly dead spinner dolphin, the cape canbe discerned on close and careful inspection, if onlyvery faintly, in all cases. In the eastern spinner, thedorsal overlay (terminology of Perrin 1972) is verydense, almost obliterating the underlying cape in evenvery fresh specimens. When a carcass of an easternspinner has lain on the deck of a tuna boat for morethan a few minutes, the dorsal overlay darkens to thepoint of completely obscuring the cape. In the live pan­tropical spinner, the dorsal overlay is less dense andthe cape is sharply defined and obvious. It is detectableeven in specimens dead for several hours, albeit morefaintly. Thus the absence of the cape in a sketch meansthat the observer did not note it upon fairly cursory

<=:- ------4.~~

<\~= ~...

~~...

~~:s~

~ 3===:<{Code 2

Code 1

Code 3

FIN

CAPE

~Codel

~Code2

Code 5

Code 2

Code 4

Code 3

Code 1

VENTRALFIELD

The data on color pattern and dorsal finwere collected by biologists during theperiod 1974-88 aboard commercial tunaseiners (see Acknowledgments). Thelength data were collected in 1968-89. Thebiologists completed a field "life-history"form for each dead dolphin examined,recording basic information such as sexand length and (starting in 1974) sketch­ing in the ventral field, cape (if prominent­ly visible; see below), and dorsal fin on apreprinted generalized dolphin outline.When feasible, the observer also traced thedorsal fin on the back of the data form.Body length was measured to the nearestcm with calipers mounted on a 2-m woodenruler. For male specimens, a testis withepididymis attached was preserved andweighed ashore. For females, the numberof corpora in the ovaries was determinedin the laboratory (methods described inPerrin et al. 1976).

The quality of the sketches varied great­ly. Although a few color pattern sketcheswere discarded as totally unusable, thebasic features of the color pattern could bediscerned in relatively inexpert sketches. However, thefin sketches proved to be unreliable. We found that forthe cases where the observer both sketched and tracedthe dorsal fin, the correspondence between fin shapein the sketch and tracing was often poor. Therefore weused only the fin tracings, which reduced the samplesize for dorsal fin shape. We also examined photo­graphs of the specimens if they were available.

The analyzed data consisted of body length in cm andcoded values for ventral field, cape, and dorsal finshape. These were accompanied by date and locationof capture and reproductive data (weight of testis +epididymis for males; number of ovarian corpora forfemales).

The ventral field received a code from 1 to 5 (Fig.2). Code 1 is the state typical of the eastern subspecies;Code 5 is that typical of the pantropical subspecies (Fig.1). Codes 2-4 are intermediate. In Code 1, the animalis gray laterally and ventrally, with white patches ingenital and axillary regions. In Code 2, the separategenital and axillary white areas of Code 1 are joinedby a speckled zone. In Code 3, the two areas are con­fluent, forming a white ventrum. In Code 4, the ven­tral field extends dorsally above the umbilical-genitalregion, yielding a stepped pattern in lateral view. InCode 5, the anterior portion of the ventral field extendsdorsally to behind the eye, eliminating the step of Code4. This variation is continuous, of course; a code wasassigned for the state closest to that evident in the

414 Fishery Bulletin 89(3). 199 J

30·'r---------<:"0:'----_=--~-___,

A

"WHITEBELLY"

TYPE SCHOOLS

Figure 3Localities for purse-Beine sets on schools of spinnerdolphins for samples included in analyses, by schooltype (see text for school-identification criteria).

"COSTA RICAN" (..)

..

"MIXED EASTERN

& WHITEBELLY"

TYPE SCHOOLS

"EASTERN" TYPE SCHOOLS

& "UNIDENTIFIED" (e)

TYPE SCHOOLS

o

c

10·

o·

10·

70·

o·

10·

20·

o·

10·

10·

10·

20·

20

Perrin et al.: Geographic variation in morphology of Stenella longirostris 415

inspection, not that it was not actually absent. Cape"presence" is a relative measure of the density of thedorsal overlay, with some precision lost due to post­mortem darkening. Again, under the assumption thattime before examination of the carcasses and diligenceof the observers did not vary with area, there shouldbe no systematic bias.

The dorsal fin was coded to one of three states (Fig.2): canted-closest to a right triangle with hypotenuseposterior (Code 1; most "eastern-like"), erect-isos­celes triangle (Code 2), or falcate-right triangle withhypotenuse anterior (Code 3; most "pantropical-like").Sources of error in this code include partial tracingsof fins (part of the base of the fin not included) andinjuries distorting the fin. The first was difficult tocope with and could not be resolved completely. Sometracings were discarded because of sharply anomalousshape obviously due to the fin being incomplete, butsome small portion of the variation recorded is un-

doubtedly due to undetected errors of this sort. Whena distal portion of the fin was missing due to injury,this was indicated in the observer's sketch, and thetracing was not used. The tracing was also not usedif a very large notch or gap was indicated. Smallnotches and holes do not affect the overall shape ofthe fin.

The total of the code scores for the three charactersranged. from 3 (typical of the adult male eastern spinner)to 10 (typical of the adult male pantropical spinner).

The sample

The sample for color pattern and fin shape included8350 specimens from 971 net sets on "eastern"-typeschools, 1200 sets on "whitebelly"-type schools, 27 setson "mixed eastern and whitebelly" schools, 2 sets on"Costa Rican"-type schools, and 54 sets on spinnerdolphins of unidentified stock (Figs. 3 and 4). Not all

WHITEBELLY WHITEBELLY

..10' rtl' -so' ••0" 111)" ':l'C" 110'

I I ':-"~.\ ! ,- (18)TI - '-"L '---"--.' -\

_L~!!i!_~~~"-,1974-d-=~~~ Wr-:

I Q" '.1)", " II I ...

...1979 H-++++-Hf--r"'I-+.......

..... ".

!=RI I ~~~ r (36)r~ \_~ '-j' I •I . ,1"1.

-~ • T;1

1+1: I I I I ,-'" l'1980 I

·rJ150' ,. ".. ".. .... ... ...

...1977 f-+++--H-+-H-++.!.~ 1981 H-++++-H--F'H-.......1977 H-+-+-++++-+-+-I--l,

.~"".-,=..."-"'!:.,..~,,=-.. '--:.-!,.-... --'--±,....,-l--,!...,.-L,...!;--L-,!rclf II!..-'·""'""...---;t;.".,.-J-;j.,-L-,~~-.!.-'-;!;.-J'-;!1981 H-+-+++-+-f-+-+-+-'~

~=-.,-!;J~L.",,=-.. .L.:!:....--'-;;!;...~..............rJ~ ISO" 1110" 130" 13D" 110' 100" ICI" Ill" rJ'"

Figure 4Distribution, by year, of purse-seine sets on "eastern" and "whitebelly" schools of spinner dolphins for samples included in analyses.

416 Fishery Bulletin 89(3). 1991

...'\' I (0)

~'~S '-~,(' ,

1/\I

I '. .1988 I r,... .... ".. ".. " .. ... ... ...

ALL YEARSt-+++++-+-,>--

1',. . 10"

'.1988 I-+++--H--I-H-t+........ ,..

EASTERN WHITEBIiLLYrr-rr-.--r--.-;;o:-~7"""'~-"""--"'-'-""T"'T~.,....--~.,......,..~---,...r (44) I-+-1f--H-t-++'~·iI-'.. r (91)-::::·'L H-++++++-+-"._ \--..:'L ,..

~ i " I...."'.r 10" •

-+-H-~

'.I-+t+--H~I-+t+-+-~+Ih

150" ,to" 130" 120' II

,..

WHITEBIiLLY

1984 l-+-i-I--++-H--fU'l-+-',

rr-rr-.--r..,....,,..----,..,T"""""~-,,.,.

f-+--iI-H--H-+'~' '. r (97)1-+-++++-+-1--+-,,_....JL.,

I~'":.,.,...; 10"

I

,w

I (52).. I-+-+++-+-+--I-H. _ \.JL

- 1. r" '..

I 'i

.. f-L-.J......L...++-+--I-+-H

'.. f-L--'---'-+-+-+--1'-f-t-+++~,

...

EASTERN

... ,

1985 H-++-+-t-+-H-+-+.......

.,~ ,,

''t I (21)" ',--JL ,

I ....'71\

1/

I~

1984 I r·,.. " .. ,... ... ... ...

1985 I-+-+-+-+++-+-H-+-',~IISO' l.acr 130" 120" ncr 100" IG" lit' rJII' IWb---"=...----±,,.,.r'--::,,..,,.L-,,,,!.-.. '-:,b.....L.,!,,...L--,!......-'--ol.,J,f1 1H"k--=,.-:=,,.,.,.......,,,...:o-;-;';;-'-==,...,...L-;!;.................~

~:'i ., ( _(20)~ '-.-J L '

i ,,\'-'i12'j:" I

- --- V"-\

1983...

I r·,'w '... ' .. ".. ,,.. ... ... ,..

"~~ ( (35),

,~_rL

~ I .,';p,"" ,

,

f"1982 ~ ,'w ,... . ".. ,... .., ,..

Figure 4 (contlnuedl

Table 1Sample sizes for analyses of color pattern and fin shape ofspinner dolphin in the eastern tropical Pacific Ocean.

sex or age or both (Perrin 1972). We endeavored totake into account these factors in the geographicalanalyses by considering males and females separatelywhere necessary and by applying a minimum lengthor sexual-maturity criterion for specimens to be in­cluded in analyses of each of the characters separately.

of the characters were recorded for all of the specimens(Table 1). Body-length data were available for 3304adult dolphins (1461 males and 1843 females). (Femaleswith at least one ovarian corpus and males with testislepididymis weight of 100g or more were consideredadult, following Perrin et al. 1977.)

Analytical methods

Areal stratification For purposes of the analyses, wedivided the study area into 60 5-degree blocks withboundaries at the Equator, 5 degrees, 10 degrees, etc.,and 80 degrees, 85 degrees, etc. We experimented with4-degree blocks and with different placement of theboundaries but obtained less consistent results (moreembedded and peripheral anomalous blocks) than withthe scheme adopted.

Effects of sexual dimorphism and developmentColor pattern and fin shape are known to vary with

Character

Total individualsVentral fieldCapeFin shape

Adults with all3 characters

Males

4121337733762053

610

Females

4229346334622421

683

Total

8350684068384474

1239

Perrin et al.: Geographic variation in morphology of Stenella longirostris 417

o L...-...L..._--.L__...L..._----I__....I-_----I__-J

~in in '" ~

~ r::IL iiJ iiJ ~ '" Ii r:: gj :$ l8 :: on

on '" l8 ~-; .,;

~:::;; on N 0 g .,; 0 & ~~ ~ !:l. e !!!. !!!. ::!. !!!.

25

Figure 6Change in percentage with noted cape in relation to bodylength in males and females. All specimens over 195cm in­cluded in last interval. Sample sizes in parentheses (total males3376, total females 3462).

200180160

CAPE

140120

BODY LENGTH (em)

10080

;;- r:: ;;;- r:: r:: M" ~ S ;;; in~ !i! ~ ~ ~ ~ ~

...!!!. ~ ~ e ~::!.

80

"70

VENTRAL FIELD ""60

" •wc:l 50 • •j:5 l>. • "• •Z 40 " •w0 • • " " "II:w 30 " eo "D.. • •

20 •10 •

80 100 120 140 160 180 200

BODY LENGTH (em)

Figure 5Change in ventral field in relation to body length in spinnerdolphins. Circles are Code 1; triangles are Code 5. Allspecimens over 195cm included in last interval. Sample sizesin parentheses (total 6840).

spermatogenesis occurs (Perrin and Henderson 1984).Thus, the canted fin in the eastern-type animals wasstrongly associated with sexual maturity, althoughsome mature males had erect fins.

Because they relate primarily to the methodology ofthe study, the results of analyses to establish thelength/maturity criteria are presented here rather thanbelow with the results of the analyses of geographicalvariation.

We found no difference between the sexes in extentof the ventral field but a statistically significant associa­tion between ventral field state and body length (con­tingency test of association with Peterson's chi-square,a 0.05; computer program BMDP4F in Dixon. et al.1990). As described by Perrin (1972), all calves tendtoward the "whitebelly" color pattern (codes 4 or 5),with reduction in the ventral field in "eastern" spin­ners with age (to Codes 1, 2, or 3). The proportion inthe overall present sample with Code-5 ventral field(most "whitebelly-like") declined with body length untilabout 120cm, when it stabilized at about 30-35% (Fig.5), and the proportion of individuals with Code 1 ven­tral field (most "eastern-like") behaved similarly butconversely. We therefore limited the analyses of ven­tral field to specimens of 120cm or larger. (After about170cm, the proportion of Code-5 animals again climbedand Code-l animals decreased; this is because mostspecimens above that length were "whitebelly" spin­ners-see analyses below of geographical variation inbody length.)

Noted presence of the cape exhibited a similar pat­tern of change with growth (Fig. 6). Calves in all re­gions tend to exhibit a strongly defined cape, which isprogressively obscured with development of the "east­ern" pattern (Perrin 1972). Again, relative stability inthe proportion of all animals for which the cape wasnoted was reached at about 120cm in both sexes; thesample for the geographic analyses was limited tospecimens of this length or greater.

We also found a slight association of expression ofthe cape with gender; the cape was noted for 10.1%of all females but only 7.8% of males (Fig. 6). In thegeographical analyses, we used the average of male andfemale means in order to avoid effects of possible arealvariation in sex ratio in the samples.

The degree of sexual dimorphism in the shape of thedorsal fin varies geographically, ontogenetically, andindividually (Perrin 1972). In the present sample, inschools classified as "eastern" (Fig. 7), about 60% ofsmall calves of both sexes had erect dorsal fins (Code2); the balance had falce.te fins (Code 3). Beginning atabout 135cm, erect fins increasingly predominated, tomore than 80% in large juveniles of both sexes andabout 90% in adult females. In males, the first cantedfins (Code 1) appeared at about 160cm. They increasedin frequency to about 70% in large adults of 180cm ormore. This corresponds roughly to the length at whichabout 80% (the asymptotic level) have achieved at leastminimum testis-epididymis weight (94g) at which

418 Fishery Bulletin 89(3). J991

• canted

180 190140 150 160 170

BODY LENGTH (em)

~ erect

70 120 130o

10

100

90

ffl so..J

~ 70

~ 60

W 50

~Z 40wU 30a:~ 20

10

o

90

ffl so..J<::i: 70W

~ 60oW 50

~ 40ZW 30Uffi 20a-

140 150 160 170 180 190(521 152) (6ll1 11881(1281 (186)1194)(1041 (52)

100

90(J)W 80..J<::i: 70WU-

60U-0W 50Cl

~ 40ZW 30Ua: 20Wa-

10

0

100

90

(J) 80W..J< 70::i:U- 600W 50Cl~ 40ZWU 30a:W 20a-

10

070 120 130

155) 1521

............. '.' / ....:;.. ....... ':, "~ '.. "../ "/"~"'/' /.(~'

:-:. /,.".-:., <-. -::,- <..;:,... /j' ...... ../"/ .~/,

Vi'L>iY: »;. ~>;/~;.../:,~ / .. '.. ,".. .." ,/' ..

70 120 130 140 150 160 170 180 190

BODY LENGTH (em)

L2J falcate 0 erect • canted

Figure 7Change in percentage of males (top) and females (bottom)from "eastern"-type schools with canted (Code 1), erect (Code2), and falcate (Code 3) dorsal fins in relation to body length.Sample sizes in parentheses.

Figure 8Change in percentage of males (top) and females (bottom) from"whitebelly"-type schools with canted (Code I), erect (Code2), and falcate (Code 3) dorsal fins in relation to body length.Sample sizes in parentheses.

In the "whitebelly" schools (Fig. 8), roughly 90% ofcalves, juveniles, and adult females had falcate dorsalfins and about 10% erect fins. One adult female hada canted fin. In adult males, the frequency of erect finsincreased to about 40% in individuals of 175cm or more(2.3% had canted fins). Again, this is about the lengthat attainment of sexual maturity (loc. cit.). We limitedthe analyses of dorsal fin shape in the pooled samplesto sexually mature dolphins and considered males andfemales separately.

We followed the same procedure for analyses of bodylength, limiting the sample to sexually mature animalsand analyzing males and females separately.

Effects of seasonal variation The apparent overlapof collection localities for the two kinds of schools isnot an artifact of seasonal migrations. The region ofsubstantive overlap of "eastern" and "whitebelly"

schools is between 8° and 12°N and 95° and 125°W(Fig. 3). We examined the data by quarter of the yearand found overlap in this region in all four quarters.

Pooling to obtain adequate sample sizes Samplesizes in many cases were very small for some of theperipheral 5-degree blocks. In these cases we pooledthe samples over two or more blocks, to achieve aminimum sample size of 25. In the pooling we at­tempted to approximate the apparent radial patternof variation, i.e., pooling was mainly longitudinal in thesouth and mainly latitudinal in the west. To the extentpossible, we maintained the same pooling scheme fromanalysis to analysis, although in some cases differen­tial pooling was required for males and females. Thesample sizes and pooling schemes for the variousanalyses are given in the Appendix.

Perrin et al.: Geographic variation in morphology of Stenella longirostris 419

o'J------L---,--------t----'-,

30'r---..,-----,----.-......-

10'

20'

10'

20'L-_-'-__--'-_--'-_----L__-'-_-'-_-'--_--I

150' 140' 130' 120' 110' 100' 90' 80' 70'

20'

O'

eo'100'120'\40'

~,.A~ •. '~-rl:~..A 90 .."..r·°'lo,t n

~.,~~

~15 23 52

62 84 86 -.....,,'. .,. .

~-(78 83 77 81 86 89, ",--,,_ ri:10 I 0 9 2 11 35 15 25 33 37 54 25 'D~>" '.- .

0 o 0 0 0 149 ;.''--- 3 0 •• "l: i:,.:

o 1 8 119 ".". "0( ~: ~::t <t •

Most "eastern"-Iike o ~~~~ ..i· ~ •

.1 '<:>I I .....l..- ! I I

o

20

20'

30"50' 140' 130' 120"

20'

19 13 832 20

10'27 23 29 11 17

120'

Figure 10Contours ofventral-field codes for spinner dolphins ~ 120cm.Based on 5-degree block averages (see text).

Figure IIPercentage of spinner dolphins with observed cape, by area.

10· CAPE

20'

O'

20'

80'

eo'

80'

100'

100'120'

120'

120' 100'

140'

IJO'

140'

B ,'-''''-'~19 16 14~"':":"" <'

~16 19 23

17 18 13 ~,<13 14

124 128 13 20 19 19 28 14 11 21 19 47~'

8 17 8 24 32! 28 V'--- 4 13

0 3 111 2d.Intermediate

3 1fventral field 5 1--\

0

~:~~:'~ >(C

~ .. ..0'"' .. /'

""~L• >

19 0 ci"-....~ .. ·• ..

M8869 59 25

:RIl:. ~ •••

9 3 6 1 1 ; '::,1 .~176 174 78 78 64 46 59 61 50 42 27 27 "'\v~.."",

92 83 94 76 681 26 ,...,,~

'--- 93 87 .... :t:r oAl:

100 96 7 1 154 :.HoY..,1" :.:. ).

97 N:*"Most "whitebelly"-Iike95 -*"If""'$

100 "ir·'~--;

20

O'

20'

O'

20'

20' '---_-'-_--L._--'-_----l._~'---_ _'___ _'__ ___l 20'

Figure 9Percentage of spinner dolphins exhibiting (A) Code 1 (most"eastern-like"), (B) Codes 2, 3, and 4 (intermediate), and (C)Code 5 (most "whitebelly-like") ventral fields. by area.

Contouring We used the ZSMTH, ZGRID, and ZCSEGsubroutines of the PLOT88 graphics software package(Young and Van Woert 1987) to generate contours ofventral coloration scores for unweighted 5-degreeblocks. The center of a block was used as the positionfor the mean ventral coloration score of specimens col­lected within the block. We used ZGRID with the "cay"parameter set at 5 and "nrng" set at 4 to generate adata grid for contouring. ZSMTH with "nsm" set at1 was used to smooth the grid. Contour lines weregenerated using ZCSEG with "ndiv" set at 4,

Results

Ventral field

The pattern of disjunct axillary and genital white areas(Code 1) occurred at high frequencies (~750/0) in arelatively small part of the study area, in 12 5-degreeblocks north of lOON and east of 1200 W (Fig. 9A). Itwas totally absent south of 50 S and absent from mostof the peripheral blocks between 15°N and 50 S. Itoccurred at intermediate frequencies near CentralAmerica and northern South America, between lOONand 50 S. The steepest gradient would appear to havebeen at about lOON, extending from the coast out topossibly beyond 125°W. This pattern was mirrored inthe relative distribution of the Code-5 ventral field(most "whitebelly-like"), which was absent or nearlyabsent from most of the northeastern blocks where theCode 1 pattern predominated and occurred at frequen-

420 Fishery Bulletin 89(3). 1991

80'

\'-----''----------' \"";

100'120"140'

A

O'

20'

20' 1--_'---_'--_'--_'--_'--__'-------',_

20' I--_'--_'--_'--_'--_'--_'--_'-------l 20'

Figure 12Percentages of adult male spinner dolphins with CA) canted,(B) erect, and (C) falcate dorsal fins, by area.

cies above 75% in a broad band encompassing the studyarea south of 15°N (Fig. 9C). Again, per force, thesteepest gradient was at about lOoN.

Dolphins with intermediate ventral fields (Codes 2-4)were least frequent in the peripheral blocks to the southand southwest and most frequent in a region off Cen­tral America below lOON (Fig. 9B).

Contouring of the ventral-field data based on 5­degree block averages (Fig. 10) yields a clear patternof extensive, largely Code-1 and Code-5 regions dividedby a steep gradient.

Dorsal fin

Because of sexual dimorphism and the need to restrictthe analyses to sexually mature animals, sample sizeswere very much smaller for this character than for thecape and ventral field, necessitating more extensivepooling. Parts of the areas off northern South Americaand in the southwestern portion of the study area thatwere included in the coverage for the color patterncharacters are not covered in the fin samples.

Despite the reduced resolution in the analyses of dor­sal fin variation, some patterns are evident (Figs. 12,13). In the males the sharp gradient at about lOON(Figs. 12A, C) is concordant with the pattern for theventral field (Fig. 9); the canted fin, like the disjunctventral field, occurred at high frequency (~75%) onlyin a region bounded approximately by lOON and120oW. The same is true of the erect fin in the female(Fig. 13). The falcate fin in males was more commonsouth of the Equator; 90 of 127 males (71%) in the threesouthernmost cells that contained no males with cantedfins had falcate fins, but only 25 of 66 (44%) in the

Figure 13Percentage of adult female spinner dolphins with erectdorsal fins, by area. Of remaining, one had canted fin,others had falcate fins.

Cape

The percentage of dolphins with observed cape washighest in the far south and far west (Fig. 11). Thispattern differs from that for the ventral field. For theventral field (Fig. 9C), the zone of high frequency ofCode-5 fields was continuous along a southeast-north­west axis, whereas for the cape the two regions of rela­tively high frequency in the south and north were sep­arated by a broad zone of relatively low frequency. Inaddition, the north-south gradient in expression of thecape was not as steep (at lOON) as for the ventral field.

20°

20'

80'

80'

-, if

\" ~.

'-----''----------' ~. ~~

1(10'

100'

120'

120'"

erect

140"

140'

140'

B

c

O'

O'

20'

20~

Perrin et al.: Geographic variation in morphology of Stenella longirostris 421

O·

00'100'140'

O·

20' ,1---,..1-40-,---'----121-

0.----'---..1.

100-.--'---80-'---.---l 20'

20'

O'

20'

O·

20' 1---1...J.40-'--'---'2-'-0-'--'-------'IO~O.----'----'-OO-.---I 20'

0O' O' O' O'

~.·'::.~:i:.s(~;;b";:::,::' ':":':". ,120' 20' 20' 20·

140' 120' 100' 60' 140· 120' 100' 00'

B Bl ..~20' 20' 20' 20'

9 .. ' ",a;37 32 :~"

~~

72

100 96O' O' O' O'

Intermediate Intermediate

20' 20' 20'20'140' 120' 100' 60' 140' 120' 100' 80'

C20' 20' 20· 20'

Figure 14Percentages of adult male spinner dolphins having most"eastern" (A), most pantropical/"whitebeIly" (C), and in­termediate combinations (B) of ventral field, cape, and dor­sal fin, by area.

Figure 15Percentages of adult female spinner dolphins having most"eastern"-type (A), most pantropical/"whitebeIly"-type (C),and intermediate combinations (B) of ventral field, cape, anddorsal fin, by area.

two westernmost cells had falcate fins. This differenceis statistically significant (P<O.OOOl: Fisher's exacttest, two-tailed). The frequency of falcate fins in thetwo southwestern cells between lOoN and 50 S was in­termediate (Fig. 12B). In females, the falcate fin wasalso most common in the south, although in this casethe frequency there was not statistically different fromthat in the two westernmost cells at P = 0.05.

Combined characters

Data on all three characters (ventral field, cape anddorsal fin) were available for 610 sexually adult malesand 683 females. The distributions of the extreme ex­pression of eastern character combinations (a totalscore of 3 for males and 4 for females) and pantropicalcombinations (10 for both males and females) weresimilar for males and females (Figs. 14, 15). Only theextreme eastern combination predominated (more than50%) in any area, in a region bounded by lOON and

422 Fishery Bulletin 89(3). 1991

2~5l-,,0",...----.1'4D--" --,,!,-,O.-----::I:12D--,.---,,,!,-,O.----::!:'OO::-.--:.~O.-----:':.D::-. --::!.

D'

'D'

'D' cf

20'

9.0 9.0 8.9 9.0

8.2 8.6 8.8 8.9 8.2 7.3 8.1 7.9 8.3 8.8 8.8 7.6 9..

&S&7&S&S&8&9&SaO&0O'

'0"

'D'

20'

,00' '40- 130" 120"30"

20'

'D'

D'

'0' d20",..,. 14"- 130- '20' 110· '00' 90" SO'

.,,,.'50' '40" 130" I~O· 90' SO' 70"30"

2";..,.!=,----:-!:'40":::-----:I3~D.,-----::!'2D=.---,'-'::'D.,-----::!'OD=.---;.=D.;--~SO::-. --::!.

Figure 16Average combined score for spinner dolphins with inter­mediate combinations (see Figure 19) of ventral field, cape,and dorsal fin, by 5-degree block (no pooling). Heavy lineis between blocks with average score ~7.0 and those withscore <6.

Figure 17Average body length of adult male and female spinnerdolphins, by area. For males, cells with length ~ 180cm arehatched; for females, cells with length ~ 176cm are hatched(adult males are about 4cm longer than adult females onaverage; Perrin et al. 1985 and present data).

125°W for males (Fig. 14A) and lOON and 1200 W forfemales (Fig. 15A).1t was absent in the cells south of5°N and east of 1200 W and in the cells west of 135°W.The extreme pantropica1 combination reached only34% in males, in the southernmost region (Fig. 14C),and 44% in females (Fig. 15C). It was absent (exceptfor one female) from the cells in which the eastern com­bination predominated (north of lOON and east of120-125oW). It was present only at low levels in thesouthwestern region, about the same as just to thesouth and west of the core eastern region. Except inthe core eastern region, the majority of dolphins wereintermediate (combined scores of 4-9 for males and 5-9for females (Figs. 14B, 15B); in some cells between 0°and 15°N, over 95% of the adults were intermediate.Even in the core eastern region, about one third wereintermediate. The degree of intermediacy was notuniform, however. For both sexes, there was a verysteep latitudinal gradient in average scores of inter­mediate animals at about lOON (Fig. 16). There was

also a steep longitudinal gradient, at about 125oW formales and 115oW for females. Outside this sharplybounded core "eastern" region, variation in interme­diate scores was very much less, with only slight N/Sand EIW gradients and most animals relatively closeto the pantropical extreme.

Proportionately more adult females than adult malesin the westernmost cells (west of 1400 W) were of theextreme pantropical combination (Fisher's exact test,P<O.Ol), whereas this was not as true in the southern­most cells (P>0.10).

Body length

The major shift in average adult length of both malesand females (Fig. 17) was at about 5°N, rather thanat about lOON as for the other characters (e.g., com­pare with Figs. 14-16). The region of smaller animalsextended south along the coast of northern SouthAmerica to about 50 S. Animals from the far west were

Perrin et al.: Geographic variation in morphology of Stenella longirostris 423

78

177.2

7.84

4.43

Pantropical

47

181.9

7.46

4.10

195

172.7

5.90

3.42

Eastern

Males Females Males Females

136

178.9

6.30

3.52

Sample size

Average length (cm)

SD (cm)

CV

Table 2Variability of body length in sexually adult spinner dolphinsof eastern (combined scores of 3 for males or 4 for females)and pantropical (combined score of 10 for both males andfemales) morphology (ventral field. cape, and dorsal fin) fromthe eastern Pacific.

20'

o = 5.4-6.4cmd IZI = 6.5-7.4cm

11II = 7.5-6.3cm

0'

,0·

'0'

20'

2o;5!-::0'~--:'-.!::0';----'+'''':-'--::!:120::-·-7.;"0::-·--;;tIOO:;;""'-....;;.0;;.·-'80~·-~

Figure 18Standard deviations (in cm) of body length for adult male andfemale spinner dolphins. by area.

more closely associated in average length with thosein the east than those in the far south (below 50 S).

Body length is more variable in the regions of greaterbody length, in both males and females (compareFigures 17 and 18; Table 2).

Discussion

Year-to-year variation and home range

The patterns of distribution of the eastern spinnerand the "whitebelly" intergradeslhybrids appear to berelatively stable year to year, although analysis of thepresent data is complicated by the fact that they comefrom incidental kills in a very mobile pelagic tunafishery. The cumulative maps of specimen localities(Figs. 3, 4) show the locations where tuna boats madepurse-seine sets that killed spinner dolphins. It reflectsonly co-occurrence of tuna boats, spinner dolphins, andyellowfin tuna. In addition, the spinner dolphin is notthe main target dolphin species in the fishery; the pan­tropical spotted dolphin Stenella attenuata is relative-

ly more often found together with tuna. The sets onspinner dolphins are largely sets on mixed schools ofspotted and spinner dolphins; in some areas, especial­ly in the southern and southwestern regions of thestudy area, spinner dolphins are not often associatedwith spotted dolphins and seldom are with tuna (Au andPerryman 1985).

Not all regions represented by capture localities inthe cumulative maps (Fig. 3) are represented in themaps for individual years (Fig. 4). Possible reasons forthis are (1) the dolphins moved around from year toyear, (2) the tuna vessels moved around, (3) theyellowfin tuna moved around, and, most likely, (4) somecombination of (1), (2), and (3). For some years (e.g.,1985-87), the samples included no spinner dolphinsfrom south of the Equator, but dolphin research surveycruises in that region in 1986 and 1987 made numeroussightings (Holt and Jackson 1987, 1988). That par­ticular hiatus in the maps of localities for at least someyears therefore represents the absence of tuna vesselsand possibly yellowfin tuna associated with dolphins,not an absence of spinner dolphins. On the other hand,the pattern of paired latitudinal bands of high densityof localities for "whitebelly" schools in the cumulativemap (Fig. 3) was not apparent in annual plots. It re­sulted mainly from relatively large numbers of speci­mens from the area of the upper band in 1975 and1984-87 and from the lower band in 1977 and 1978."Eastern" schools were sampled from the region of thenorthern band in all years. Again, research surveys in1986 and 1987 sighted comparable numbers of "white­belly" spinner schools in the two regions. In this case,it is possible that the pattern or degree of associationof tuna with the two forms varied interannually. How­ever, the numbers of sightings of spinner dolphins dur­ing all the research cruises were relatively small, andit remains possible that some portion of the pattern­ing in the cumulative distribution represents inter-

424

annual variation in distribution; i.e., the northernboundary of the main "whitebelly" distribution mayfluctuate year to year by about 5 degrees of latitude.This would be consistent with some of the results ofthe present study; for example, the intermediate per­centages of the extreme "eastern" and "whitebelly"character states in the band between 5° and lOON(Figs. 9-10, 12-15).

Another possible explanation of the latitudinal pat­tern has to do with distribution of tuna-fishing effort.During the period 1976-79, the Inter-American Trop­ical Tuna Commission set a regulatory demarcation lineat 5°N, between 95° and 1100 W. During this period,when an annual areal quota of yellowfin tuna for theregion north of the line was filled, the internationalfleet was forced to fish south of the line or to the west(peterson and Bayliff 1985). A similar line was in forcein 1973-75 at 3°N. The subsidiary latitudinal band ofdistribution records south of 5°N apparent in the spin­ner dolphin data is also present in data for the pan­tropical spotted dolphin S. attenuata, the principaltarget species of the tuna fishery (Perrin et al. 1983).Illustrations of the effect of the regulatory regime onthe distribution of fishing effort are given in Punsley(1983); the effect was particularly pronounced in 1978.

Some of the morphological results speak against thehypothesis of large-scale latitudinal movements. Varia­tion among sub-areas of the "whitebelly" range isdiscordant. For example, for the character "ventralfield" (Figs. 9, 10), the far-western blocks, far-southernblocks, and southwestern blocks are similar in havingrelatively high frequencies of the Code-5 characterstate (most "whitebelly-like"), while for the "cape"(Fig. 11), the southwestern region differed from thewest and south in having low frequencies of observedcape (they were most like the animals immediately tothe north). For male/female dorsal fin shape (Figs. 12,13), the southern blocks had higher frequencies of thefalcate fin than did the western blocks. While in colorpattern and in cranial characters (Schnell et a1. 1985),the westernmost blocks were most like the southernblocks, in body length and variability of body length(Figs. 17, 18) they were most like the northern and cen­tral blocks. The blocks just north of the Equator wereremarkable in having very high frequencies of femalesintermediate between the most "whitebelly" state andthe most "eastern" state in combined character scores(Fig. 15); the highest frequencies of these in males werein the westernmost blocks. In body length in both malesand females (Fig. 17), the steepest gradient was atabout 5°N, whereas in combined ventral field, cape anddorsal fin (Figs. 14, 15), it was at about lOON. In cape(Fig. 11) and body length (Fig. 17), the dolphins in thePanama Bight and just to the south were most likethose in the core "eastern" region (lacking visible cape

Fishery Bulletin B9(3). 1991

and small in body size), but in ventral field (Fig. 9) theywere intermediate between those to the north andsouth, having approximately equal frequencies of Code1, Code 5, and summed Codes 2-4. This complex patch­work of variation suggests that there is not a largeamount of movement between the various regions andthat the year-to-year variation in the collection local­ities (Figs. 3, 4) reflects mainly movements of the tunavessels and perhaps the tuna.

The complex geographical pattern of variation inthe zone of intergradationlhybridization ("whitebellyspinners") is consistent with the limited informationfrom tag returns (Perrin et al' 1979b), which suggestsa home range of a diameter of hundreds rather thanthousands of kilometers. This is also supported bythe stability of the core eastern region between yearsand patterns of reproductive seasonality that suggestareal structure within the eastern region (Barlow1984).

However, there may be mid-scale seasonal move­ments within an overall pattern of interannual stabil­ity, especially in the "whitebelly" areas outside the corerange of the eastern spinner. Based on census surveysfrom dedicated research vessels, Reilly (1990) reportedincreases in abundance west of 1200 W along lOONcoincident with seasonal shoaling of a thermoclineridge. Because of interannual variation in distributionof tuna fishing effort (discussed above), the coarsegeographic strata used, and the relatively small sam­ple sizes for the outlying strata, the present analysesprobably could not be expected to detect such shifts,although the shifts might be expected to blur theperceived patterns of morphological gradients.

Taxonomic status and management ofthe "whltebelly" spinner dolphin

The "whitebelly" spinner has been managed as two"stocks" divided at the Equator, the northern white­belly spinner and the southern whitebelly spinner, im­plying effective reproductive isolation between themand from the eastern spinner (e.g., Hall and Boyer1990). The complex pattern of discordant geographicvariation and the relatively higher standard deviations(of body length) in the range of the "whitebelly" spin­ner are typical of a hybrid zone resulting from whatMayr (1970) called allopatric hybridization, or "the for­mation of a secondary zone of contact and of partialinterbreeding between two formerly isolated pop­ulations that had failed to acquire complete reproduc­tion isolation during the preceding period of geographicisolation." The discordant pattern speaks against theearlier hypothesis of a locally adapted "whitebelly"population sharing features with populations to the eastand west and defined by morphological clines. An

Perrin et al.: Geographic variation in morphology of Stenella longirostris 425

hypothesis of secondary contact after isolation (Perrinet al. 1985) and consideration of the patterns of varia­tion led to the formal recognition of the eastern formas a subspecies, Stenella longirostris orientalis, and the"whitebelly" form as an hybrid/intergrade betweenthat subspecies and the pantropical spinner dolphin S. l.longirostris (Perrin 1990). The hypothesis of geneticexchange between the eastern and "whitebelly" popu­lations is also supported by results of analysis of mito­chondrial DNA (Dizon et al. 1991); the presence ofunique haplotypes that would be expected in isolatedpopulations was not detected. Given this situation, arelatively higher priority should probably be given toprotection of the eastern subspecies on the groundsthat It is a distinctly and locally adapted form vul­nerable to depletion and to genetic "swamping" fromthe larger and less-exploited population to the west.In any case, the understanding presented here of thecomplex pattern of variation within the "whitebelly"range does not support the present division into north­ern and southern stocks for management on groundsof population distinctness.

Conservation and management ofthe eastern spinner dolphin

There was generally good concordance between thepatterns in the individual morphological characters andthe relative distributions of the two kinds of schoolsas identified by modal appearance (Fig. 3). One excep­tion was the region west of the Bight of Panama; herethe very high frequency of morphologically interme­diate dolphins and virtual absence of animals of the ex­treme "eastern" appearance indicate that these schoolsperhaps would have better been identified as "white­belly" schools. Apart from this discrepancy, the con­cordance between the character patterns and the fieldidentification of schools suggests that the modal­appearance method of identification is largely effective.The substantial number of "mixed schools" and uniden­tified schools, however, indicates that the method canstill offer difficulties. In addition, it is sometimes dif­ficult to identify a school of spinner dolphins to "stock"(eastern or "whitebelly") before the seine is set. Thiscan be due to poor viewing conditions (time of day,angle of sun, amount of wind, height of swell) or to thebehavior of the dolphins (e.g., "laying low," precludingobservation of ventral coloration).

An alternative management scheme based on geo­graphic location would avoid the difficulties of fieldidentification of schools. Although the working hypo­thesis for management of the eastern spinner and the"whitebelly" spinner (easternlpantropical hybrids orintergrades) has been one of broadly overlappingranges (e.g., Perrin et al. 1985), the actual transition

3cf7---------.-.;:~----r""'='"-":"":"""-___,

\J

o·

"EASTERN"

3cf'---------~o:___---r...."..-';"'7""-___,T \)

20·

o·

"WHITEBELLY"

Figure 19Localities of "eastern" and "whitebeIly" schools of spinnerdolphins included in the present analyses in relation to anexample of a possible conservation zone for the easternsubspecies. Stenella longirostris orientalis.

between the two as indicated by gradients in the mor­phological characters and the field identifications (Fig.3) is quite sharp. An "eastern spinner conservationzone" could be devised that would offer appropriateand unequivocal protection to the unique and coherentgene pool of the eastern subspecies. For example, sucha zone bounded by lOON latitude and 125°W longitude,the approximate latitude and longitude of the majorgradients in the morphological characters, would haveincluded 84% of the schools identified in the field as"eastern" (excluding those in the Bight of Panama,discussed above) and only about 5% of those identifiedas "whitebelly" (Fig. 19). Implementation of an inter­national quota or prohibition for such a zone would bestraightforward. The boundaries could be set based ona balance of considerations of conservation and opera­tional practicality and should take into account seasonaland interannual variation in locations of concentrationsof schools of the eastern spinner.

This scheme, of course, is only one of numerous pos­sible alternatives to conservation and management of

426

the subspecies. Others include international applicationof a stock quota, complete bans on this type of fishing,and increased focus on technological efforts to reducefishing mortality rates, to name a few.

The Central American spinner("Costa Rican spinner")

The number of "Costa Rican" schools sampled was toosmall for meaningful separate analysis (2 of 1454; Fig.3) and the effect on the analyses of including them inthe pooled samples can be considered negligible. In anycase, the range of the Central American spinner S. l.centroamericana as presently understood (a coastalstrip about 50nmi wide between about 13° and 7°N)does not overlap those of the eastern subspecies or the"whitebelly" hybrid/intergrade forms (Perrin 1990).This range could serve as a management unit for thesubspecies.

Acknowledgments

The data used in this study were collected by biologicaltechnicians aboard commercial tuna seiners. We wishto thank these people and the captains and crews ofthe seiners; the research would have been impossiblewithout their efforts and support. The sampling pro­grams were supported by the U.S. National MarineFisheries Service and the Inter-American TropicalTuna Commission; both agencies gave us full accessto data and specimens. R.B. Miller and others pro­cessed reproductive samples to determine sexualmaturity. K.E. Wallace and R. Rassmussen provideddata extraction services. R.C. Holland gave invaluableassistance with the contouring. J. Barlow carried outthe Fisher's exact tests. R.M. Allen drafted thefigures. J. Barlow, M. Scott, A.E. Dizon, I. Barrett,and S.B. Reilly reviewed the manuscript and offeredmany useful suggestions.

CitationsAu, D.W.K., and W.L. Perryman

1985 Dolphin habitats in the eastern tropical Pacific. Fish.Bull., U.S. 83:623-643.

Barlow, J.1984 Reproductive seasonality in pelagic dolphins (SteMlla

spp.): Implications for measuring rates. Rep. Int. WhalingComm. Spec. Issue 6:191-198.

Dixon, W.J., M.B. Brown, L. Engelman, and R.I. Jennrich(editors)

1990 BMDP statistical software manual, vol. 1. Univ. Calif.,Berkeley, 629 p.

Fishery Bulletin 89(31. J991

Dizon, A.E., S.O. Southern, and W.F. Perrin1991 Molecular analysis of mtDNA types in exploited popula­

tions of spinner dolphins (Stenella longirostris). In Hoelzel,A.R. (00.) Genetic ecology of whales and dolphins, p. 183-202.Rep. Int. Whaling Comm. Spec. Issue 13.

Douglas, M.E.. G.D. Schnell, and D.J. Hough1986 Variation in spinner dolphins (Sfenella longirostris) from

the eastern tropical Pacific Ocean: Sexual dimorphism incranial morphology. J. Mammal. 67:537-544.

Gilpatrick, J.W. Jr., W.F. Perrin, S. Leatherwood. andL. Shiroma

1987 Summary of distribution records of the spinner dolphin,Stenella longirostris, and the pantropical spotted dolphin,S. attenuata, from the western Pacific Ocean, Indian Oceanand Red Sea. NOAA Tech. Memo. NOAA-TM-NMFS-SWFC­89, NMFS Southwest Fish. Sci. Cent., La Jolla, CA 92038.42 p.

Hall, M.A., and S.D. Boyer1990 Incidental mortality of dolphins in the tuna purse-seine

fishery in the eastern Pacific Ocean during 1988. Rep. Int.Whaling Comm. 40:461-462.

Holt. R.S., and A. Jackson1987 Report of a marine mammal survey of the eastern tropical

Pacific aboard the research vessel McArthur July 29-December6, 1986. NOAA Tech. Memo. NOAA-TM-NMFS-SWFC-77,NMFS Southwest Fish. Sci. Cent., La Jolla, CA 92038, 161 p.

1988 Report ofa marine mammal survey of the eastern tropicalPacific aboard the research vessel McArthur, July 30-Decem­ber 10, 1987. NOAA Tech. Memo. NOAA-TM-NMFS-SWFC­116, NMFS Southwest Fish. Sci. Cent., La Jolla, CA 92038,143 p.

Mayr, E.1970 Population. species and evolution. Harvard Univ. Press,

Cambridge, 453 p.Perrin, W.F.

1972 Color patterns of spinner porpoises (Stenella cf. S. longi­rostris) of the eastern Pacific and Hawaii, with comments ondelphinid pigmentation. Fish. Bull., U.S. 70:983-1003.

1975 Variation of spotted and spinner porpoise (genus Ste­nella) in the eastern Pacific and Hawaii. Bull. Scripps Inst.Oceanogr. 21, 206 p.

1990 Subspecies of Stenella longirostris (Mammalia: Cetacea:Delphinidae). Proc. BioI. Soc. Wash. 103:453-463.

Perrin. W.F., and J.R. Henderson1984 Growth and reproductive rates in two populations of spin­

ner dolphins, Stenella longirostris, with different histories ofexploitation. Rep. Int. Whaling Comm. Spec. Issue 6:417-430.

Perrin, W.F., J.M. Coe, and J.R. Zweifel1976 Growth and reproduction of the spotted porpoise, Stenella

attenuata, in the offshore eastern tropical Pacific. Fish. Bull.,U.S. 74:229-269.

Perrin, W.F., D.B. Holts, and R.B. Miller1977 Growth and reproduction of the eastern spinner dolphin,

a geographical form of Stenella longirostris in the easterntropical Pacific. Fish. Bull., U.S. 75:725-750.

Perrin, W.F., P.A. Sloan [Akin], and J.R. Henderson1979a Taxonomic status of the "southwestern stocks" of spin­

ner dolphin Sumella longirostri and spotted dolphin S. attenu­ata. Rep. Int. Whaling Comm. 29:175-184.

Perrin, W.F., W.E. Evans, and D.B. Holts1979b Movements of pelagic dolphins (Stenella spp.) in the

eastern tropical Pacific as indicated by results of tagging, withsummary of tagging operations, 1969-76. NOAA Tech. Rep.NMFS SSRF-737, 14 p.

Perrin et al.: Geographic variation in morphology of Stenella longirostris 427

'20'

I7 .18

130'

(n=6399)

AppendiX Figure 1Ventral field (see Figure 13).

2a;51-0''''--'4~0'-----:'30L·,-----:",:::0'---''':---'=--o':::---~--=='

O'

10'

10'

2o;5O"~-"""~0·"-------:::'3l::0·-----:::"l::0'--,J~--::!=--~::----=--::!

30.150" 14a' 130' I~O·

20'

6510'

7 : 18O'

10' (n=6406)

zoo

Sample sizes by 5-degree block and pooling for analysesof geographical variation in external morphology of thespinner dolphin (blocks separated by dashed lines werepooled).

AppendixPerrin, W.F., E.D. Mitchell, J.G. Mead, D.K. Caldwell, andP.J.H. van Bree

1981 Stene-lla clymene, a rediscovered tropical dolphin of theAtlantic. J. Mammal. 62:583-598.

Perrin, W.F., M.D. Scott, G.J. Walker, F.M. Ralston, andD.W.K. Au

1983 Distribution of four dolphins (Sten.ella spp. and Delphinusdelphis) in the eastern tropical Pacific, with an annotatedcatalog of data sources. NOAA Tech. Memo. NOAA-TM­NMFS-SWFC-38, NMFS Southwest Fish. Sci. Cent., La Jolla,CA 92038, 65 p.

Perrin, W.F., M.D. Scott, G.J. Walker, and V.L. Cass1985 Review of geographical stocks of tropical dolphins

(Stenella spp. and Delphinus delphis) in the eastern Pacific.NOAA Tech. Rep. NMFS SSRF-28, 28 p.

Peterson, C.L., and W.H. Bayliff1985 Organization, functions, and achievements of the Inter­

American Tropical Tuna Commission. Inter-Am. Trop TunaComm. Spec. Rep. 5, 56 p.

Punsley, R.G.1983 Estimation of the number of purse-seiner sets on tuna

associated with dolphins in the eastern Pacific Ocean during1959-1980. Inter-Am. Trop. Tuna Comm. Bull. 18(3),299 p.

Reilly, S.B.1990 Seasonal changes in distribution and habitat differences

among dolphins in the eastern tropical Pacific. Mar. Ecol.Prog. Ser. 66:1-11.

Schnell, G.D., M.E. Douglas, and D.J. Hough1982 Geographic variation in morphology of spotted and spin­

ner dolphins (Sten.ella attenuata and S. longirostris) from theeastern tropical Pacific. Admin. Rep. LJ-82·15C, NMFSSouthwest Fish. Sci. Cent., La Jolla, CA 92038, 213 p.

1985 Further studies of morphological differentiation inspotted and spinner dolphins (Stene-lla attenuata and S. long;.­rostris) from the eastern tropical Pacific. Admin. Rep.LJ-85-04C, NMFS Southwest Fish. Sci. Cent., La Jolla, CA92038,30 p.

Young, T.L., and M.L. Van Woert1987 PLOT88 software library. Plotworks, Inc., La Jolla, CA.

Appendix Figure 2Cape (see Figure 15).

428 Fishery Bulletin 89(3/. 1991

10'

O'

10" cf (n=639)

20" 20'

10'

O'

10'

130"

cf (n=610)

,.,.._"i90r-'__-'."'O" ':";O"30'

20'

2o;:l1-0"-:---,c'.O""---''':':00"'--"'12"'0':---'1:':'0='".--""'~·---:.':::O·:--~"O:::'-----'70:::'''O" 2o;50~'-----,Ic!..0:-:.---:'3!::0.:---,1:!:20:-:.--.""..,O.:---,,"~":::.--'::."~·:---,BO:':o-· -----,='.

2a;5l-0'-:---,...1.0:-:':---'''~00"' ---::!12l::O':---:I'='0='"'-----,,~OO::.,---:90!;:.:--~BO::.-----,;;!·

20'

,.,.",-...;"":;:.__--=;"0";:-_-=;70"30"120""0"

Q (n=683)

O'

10'

10'

20'

2~5!-:O"::-----::!'.0:::.:---::,30~. --::!,20:::·,---,,,'::0·:---,1Q:!.0"'.---:.'.::0·:---BO:!::::-·-----,::!

20'

.,..,,_.::;90r-·__"';"TO_·__''i0~..

Q (n=1376)

O'

10"

10'

>O"'~'0"~_-'I.~0"!:-_--""",0" ',,20,-·-..=i'i

Appendix Figure 3Dorsal fin (see Figures 15 and 17).

Appendix Figure 4Combined scores for ventral field. cape and dorsal fin (seeFigures 18 and 19).

20'

'0'

O'

10' cf (n=1461)

=..-~·°r-·---'''''T-__':'';0"30.

20' Appendix Figure 5Body length (see Figures 21 and 22).

2o;5~0":---"-!:.0:-:'---::'30~' --1~20:-:·'---::"'::0·:---:'OO:!::::-'--'::.0"'::---BO:!::::-'-----'::!·

.....,,-..:.:;:0"__-.::."'i'-_---'-i7O"30'120"".,."0"

10'

5 116: 20 46O' --i--I--+-+--!-o-/--i

1 : 1 : 1 5

2a;s!-:O":---'I-!:'0::'---::'30='---::!,N'::·:---,":':O·:---,,:!.OO::.---:.!;:O.:--~.O::.-----,;;!.

10' 9 (n=1843)