Evolution,42(4), 1988, pp. 804-813

REPRODUCTIVE ISOLATION AND DISTINCT POPULATION STRUCTURES IN TWO TYPES OF THE FRESHWATER PA UCIDENS

SEINENCHOW,' YOSHIHISA Fujio, AND TADASHINOMURA Departmentof Fishery Science, Tohoku University, Amamiyamachi, Sendai 980, JAPAN

Abstract.-Twenty local populations of the Japanese freshwater shrimp Palaemon paucidens were electrophoreticallyand morphologicallysurveyed. Based on the diagnosticdistributions of some alleles at Gpi, Mpi, Mdh-J,and Mdh-2, these populationswere largelyclassified into two types(A and B). The A type occurredin lakes, ponds, and rivers,while the B type was observed only in rivers. Average Nei's genetic distance (D) between the two types fell into the subspecies range (D = 0.1 186). The coefficientof gene differentiation,GsT, varied considerablybetween the two types.In 12 populations of the A type,with a GST value of 0.281, nine pond and lake populations showed a higherGsT (0.246) than the threeriver populations (0.151). On the otherhand, GsT was 0.036 forthe eightlocal populations of the B type.The lower rostrumtooth numberhad a mode of two in typeA and threein typeB. Type-A populationslargely varied in the upperrostrum tooth numberand egg size but typeB did not. Under laboratoryconditions, mating frequently occurred withineach type,but not between types.Furthermore, no embryonicdevelopment was observed in the fewcases of intertypemating. These resultsindicate thatthe A and B typeshad experienced cladogenic separationwith pre- or postmatingisolation, whereafterthe A type,under geographic isolation,underwent genetic and phenotypicdifferentiation, while the B type,under extensive gene flow,did not undergodifferentiation.

Received February6, 1987. Accepted January13, 1988

It has been long recognizedthat intraspe- caughtin the same area, we concluded that cific genetic differentiationof sexually re- the typesare reproductivelyisolated in the producingspecies is promotedby geograph- singleriver basin. Differencesin eggsize and ic and reproductiveisolation. Cladogenic rostrumtooth number between the types are separation accompanying intrinsicrepro- apparentlymaintained by the reproductive ductive isolation is now generallyaccepted isolation. Additional analysis of the allelic as speciation. Even so, genetic differentia- distributionssuggested that the geneticdis- tion can occur in geographicallyisolated tance between the types fell into the range populations without such intrinsicrepro- of subspecies, according to the rough cri- ductive isolation. Thus, the origin,nature, teriongiven by Nei (1975) (Chow and Fujio, and functionof isolatingmechanisms play 1985c). Nishino's (1980) discoveryof great importantroles in evolutionarychanges. variation in egg size among local popula- A previous paper provided biochemical tions of this led us to presume a evidence of two typesof common Japanese highlysubdivided and geneticallydifferen- freshwatershrimp Palaemon paucidens in- tiated population structureand convinced habitingthe same watersystem (Chow and us thatan extensivesurvey of the local pop- Fujio, 1985a). The two types,named A and ulationswould be worthwhile.In thispaper, B, were clearly distinguishedby the pres- we present evidence for reproductiveiso- ence or absence of diagnosticalleles at the lation and distinctpopulation structures be- glucosephosphate isomerase (Gpi), man- tweenthe two typesof Palaemon paucidens nosephosphateisomerase (Mpi), and malate and discuss the processes involved in di- dehydrogenase(Mdh-1 and Mdh-2) loci. vergence between and within these two Since both typeshave almost the same re- types. productiveseason and since mature males and femalesof each typeare simultaneously MATERIALS AND METHODS All samples of 20 local populations of P. paucidens were collectedwith dip nets dur- I Present address: Marine Resources Research In- ing the months of March-September in stitute,Charleston, SC 29412. 1983-1986. Sitesand theirnames are shown 804 REPRODUCTIVE ISOLATION 805

*~HNP 30 km IHPO KMP

2 km

FIG. 1. Map showingcollection localities ofPalaemon paucidens. Miyagi Prefecture:IHP = Ishihoga Pond, HNP = Hinata Pond, KMP = Katamase Pond, HKP = Hanukizawa Pond, TTP = Tomita Pond, ATP = Aburata Pond, KKR = Kitakami River, NRR = Naruse River, HRR = Hirose River, NTR = Natori River, ABR = Abukuma River, SIR = ShiroishiRiver; Ibaragi Prefecture:NKR = Naka River; Tokyo: SPP = Sanpoji Pond, TMR = Tama River; Chiba Prefecture:KYR = Kiyosumi River, OBR = Ohbiso River; Ishikawa Prefecture: LK = Lake Kahokugata; Shiga Prefecture:LB = Lake Biwa; Hiroshima Prefecture:ASR = Ashida River. Symbols show the A (0) and B ( H)types. in Figure 1. Live samples were transferred exceed 0.95. Observed heterozygosity(h.) to the laboratory,except forthose of Lake was obtained fromdirect counts of hetero- Kahokugata which were frozen on dry ice zygotes,while expected heterozygosity(he) fortransport. was calculated as he = 1 - : xi2,where xi Rostrum tooth number and egg volume is the frequencyof the ith allele at each lo- were adopted as the morphological char- cus. Mean heterozygosities and He) for acters. Ten eggs at the pre-eye stage of (H. each local population were obtained from development were measured from each mother shrimp. The egg volume (V) was the formulas,H. = : ho/rand He = : he! calculated from the formula V==rh2/6, r, where r is the number of loci. Allele-fre- where I and h are lengthand width of an quency data were used to calculate genetic egg in mm, respectively.(Data forrostrum distance(D) (Nei, 1972) and a coefficientof toothnumber and eggvolume in the Natori gene differentiation(GST) (Nei, 1973). and Hirose riverswere derived fromChow Mating experiments were designed to and Fujio [1985a].) Afterthese measure- compare the matingpropensities of mature ments,some individuals were rearedin 40- males and femaleswhen they are pairedwith liter aquaria at 250C for mating experi- potential mates fromthe same versus dif- ments,and otherswere wrapped and stored ferentlocal populations and of the same at -800C forelectrophoresis. versus differenttypes. In decapod crustacea Horizontal starch-gelelectrophoresis and (especially ), mature females molt nomenclatureof alleles followedChow and priorto matingand spawning.In captivity, Fujio(1985a, 1985b).Allele-frequencydata the mature females of P. paucidens were for the Natori, Kitakami, Naruse, Hirose, distinguishedby fully developed ovaries, Abukuma, Naka, Ohbiso, and Kiyosumi whichcould be observed throughthe trans- rivers are from Chow and Fujio (1985a, lucent exoskeleton. Mature females were 1985b). A locus was assumed to be poly- takenout ofcommunal holdingaquaria and morphicin each local population if the fre- maintained individually; these usually quency of the most common allele did not underwent a prespawning molt within a 00 O

TABLE 1. Summarized allele distributionsat fivediagnostic loci of 20 local populations.Localities and abbreviationsare as given in Figure 1. Symbols: #= most common allele (P _ 0.50); * = otheralleles (0.05 _ P < 0.50). Rare alleles (P < 0.05) are not shown here.Detailed allele-frequencydata is available upon request fromthe authors.

A-typelocalities B-typelocalities Locus Allele IHP HNP KMP HKP TTP ATP NTR SPP TMR LK LB ASR KKR NRR HRR ABR SIR NKR OBR KYR Gpi A # # # # # # # * # # # # B * * * * * * * # * * # # # # # # # # Q MPI A' * * A # * * # # # # # # # # # C B * # # * * * * * C * * * * * * * * * # * # # # # # # D * * # * * * * * * > Mdh-J B # # # # # # # # # # # # # # # # # # # C' * C * * * * * * * * * Mdh-2 A # # # * # * * B * * * # * # # # # # # # # # # # # # # # 6Pgd A * * * * * * * # # * * B # # # # # # # # # # # # # # # # # # REPRODUCTIVE ISOLATION 807 week. Each molted female was then trans- rangedfrom 0.0602 to 0.1945 witha mean ferredto an aquarium containing 10-20 of 0.1186. The dendrogramdrawn by the males, obtained fromeither the same or a UPGMA methods(Sneath and Sokal, 1973) differentlocality. The aquarium was then and based on the genetic-distancevalues is observedfor five minutes for signs of sexual shown in Figure 2. There are two major behavior (chasing, mounting,and mating clustersat a discriminantlevel of0.1, clearly by males). Females were transferredfrom separatingA and B types.The diversitylevel one aquarium to another until mating oc- among the A-type populations was large, curred.Females thatdid notmate with males and one subcluster(LK and LB) was quite fromany otherlocalities were finallypaired differentfrom the other localities. The D with males fromthe same localityin order values betweenthe subclusterof LK and LB to evaluate whetherthey were still sexually and otherA-type populations ranged from attractiveor not. If a femalewas not mated, 0.0548 to 0.1057, with a mean of 0.0770; the matingexperiment was abandoned for D values betweenthe subclusterand B-type thatfemale. After mating, females were iso- populations range from 0.1645 to 0.1879 lated and maintained individuallyto con- with a mean of 0.1770. Formation of this firmfertilization. distinctsubcluster is due to the diagnostic allele distributionat the 6Pgd locus. The RESULTS diversitylevel among the B-type popula- Genetic Variation and Differentiation.- tions was extremelysmall. The difference Out of the total of 15 loci encoding 1 1 en- in diversitylevels betweeneach typeis fur- zymes, eight(Aat-2, aGpd-1, aGpd-2, Est, thershown by the coefficientof gene differ- Idh, Ldh, Me-i, and Me-2) were mono- entiation, GST (see Table 2). For the 12 morphicin all local populations,and seven A-type populations (overall GST =0.281), (Aat-1, Gpi,Mdh-1, Mdh-2, Mpi, 6Pgd, and the pond and lake populations displayed Pgm) showed polymorphismor variation higherdifferentiation (0.246) than the river in at least one locality.The 20 local popu- populations(0.151). On the otherhand, the lationswere classified into the A and B types eightB-type populations had a considerably accordingto Chow and Fujio (1985a), and smallerGST (0.036). Even when the data for the diagnosticallele distributionsare sum- thethree distant localities (LK, LB and ASR) marized in Table 1. At all loci examined, were deleted (to avoid an artifactdue to theobserved and Hardy-Weinbergexpected sampling-areadifference between types), the numbersof genotypesagreed well (G test), nine A-type populations still had a higher except in the case of a homozygoteexcess GSTvalue (0.163). for the Mpi locus in the Katamase Pond. Meristic Characters.-The upper- and Overall averages (?SD) of H0, He, and P lower-rostrumtooth numbers counted in (proportionof loci that are polymorphic) each local population are shown in Table 3. are 0.097 ? 0.030, 0.100 + 0.032, and The upper tooth number varied from4 to 0.261 ? 0.083, respectively,in the A type, 8 in the A type and from 3 to 8 in the B and 0.078 ? 0.009, 0.081 + 0.008, and type.Average upper-rostrumtooth number 0.218 ? 0.031 in the B type.No significant in the A type (x = 5.77) was slightlylower differencesin theseaverages were foundbe- thanin theB type(x = 6.10) (t = 2.46, d.f = tween types (t tests). However, significant 16, P < 0.05). The mode varied among the differencesbetween variances of these in- A-typepopulations (5 and 6) but not among dices (two-tailedF test) indicate that there theB-type populations (always 6), and there was greatergenetic variability among lo- was a significantdifference in tne among- calities in the A type. locality variances (F = 42.25, P < 0.001). Nei's geneticdistance (D) was calculated The nine A-typepopulations (withoutdis- for each pair of local populations. Among tantpopulations LK, LB and ASR) stillhad the A types,D values varied from 0.0026 significantlylarger among-locality variance to 0.1066 witha mean of 0.0462; forthe B than was found forthe B-typepopulations types,D values rangedfrom 0.000 to 0.0125 (F= 17.36, P < 0.005). The lower-rostrum with a mean of 0.0039. The D values be- toothnumber provided a moreuseful means tween the A- and B-typelocal populations to distinguishthe two types.Type A had a 808 S. CHOW ET AL.

HNP KMP TTPI ATP

SPP

ASR LKB LB KKR NRR HRR ABR h B type SIR OBR l KYR NKR 0 0.05 0.10 Genetic Distance (Nei) FIG. 2. UPGMA clusteringby Nei's geneticdistances among the local populations of Palaemonpaucidens. Abbreviationfor localities are as in Figure 1.

mode of 2 while type B had a mode of 3, River and threefrom the Naruse River were witha significantdifference between means typeB. Sixty-eightcombinations within and (t = 7.438, P < 0.005). Berriedfemales were between types were performed (Fig. 3). captured from 10 localities of the A type For all cases in which the female was ini- and fiveof the B typefrom April to August. tiallytransferred to an aquarium containing The mean egg volumes in each mother males fromthe same locality(trials 1-8 for shrimpwere averaged, and the population type-Afemales and trials 24-29 fortype-B means are shown in Table 4. No significant females), several males quickly responded correlationwas observed between egg vol- by chasing and mountingthe female, and ume and female body size. The mean egg mated almost withina minute. Even after volumes in the A type varied greatlyfrom a series of handling,females were still "at- the smallest (0.142 mm3 in Lake Biwa) to tractive"to males (as indicatedby trials12- the largest (0.998 mm3 in Tomita Pond) 18 for type-Afemales and by trials 34-41 while those of the B type were all of inter- for type B). These results show that a fe- mediate size (0.497-0.623 mm3). A two- male's thresholdfor acceptance of a poten- tailed F test indicated that the difference tial mate is not compromised by repeated between the variances was statisticallysig- handling,thereby demonstrating the inde- nificant(P < 0.001). pendence of the experimental combina- Mating Experiments. -Among 41 fe- tions. males used forthe matingexperiments, 17 Table 5 summarizes the male responses individualsfrom the Natori River,five from and gives the resultsof 2 x 2 G tests.There theAshida River, and one fromthe Tomita were 12 instancesin which typeA females Pond weretype A, while 15 fromthe Hirose were transferredto aquaria containingtype

TABLE 2. Coefficientsof gene differentiationwithin each type. HT = heterozygosityin the total population; HS = average heterozygositywithin populations.

Number of Type localities HT HS GST Mean D ? SD A 12 0.139 0.100 0.281 0.0462 ? 0.0295 Pond and lake 9 0.142 0.107 0.246 0.0475 ? 0.0316 River 3 0.093 0.079 0.151 0.0221 ? 0.0063 B 8 0.084 0.081 0.036 0.0039 ? 0.0039 TABLE 3. Number of individuals having a given number of upper- and lower-rostrumteeth in each local population. Localities and abbreviationsare as given in Figure 1. Mean numbersof teethdiffered significantly between typesA and B (upper rostrum:t = 2.46, d.f = 16, P < 0.05; lower rostrum:t = 7.439, d.f = 16, P < 0.001).

Number of upper-rostrumteeth Number of lower-rostrumteeth Type Locality 3 4 5 6 7 8 Mean SD 0 1 2 3 4 Mean SD A IHP 6 8 5.57 ? 0.51 1 10 3 2.14 ? 0.53 HNP 3 38 25 4 5.43 ? 0.67 1 52 17 2.23 ? 0.46 KMP 1 26 23 4 5.55 ? 0.66 39 14 1 2.30 ? 0.50 HKP 1 6 21 2 5.80 ? 0.61 1 26 3 2.07 ? 0.37 8 TTP 17 29 2 5.69 ? 0.55 29 18 1 2.42 ? 0.54 ATP 10 30 5 5.89 ? 0.57 1 30 14 2.28 ? 0.51 NTR 2 46 68 3 1 5.63 ? 0.61 5 96 19 2.12 ? 0.43 SPP 10 17 2 5.72 ? 0.59 26 3 2.10 ? 0.31 TMR 2 14 3 5.05 ? 0.52 4 13 2 1.89 ? 0.57 LK 4 31 24 6.34 ? 0.60 1 35 23 2.37 ? 0.52 0 LB 1 7 6 1 6.47 ? 0.74 15 2.00 ? 0.00 D ASR 1 5 26 9 6.05 ? 0.67 5 35 1 1.90 ? 0.17 Mean 5.77 ? 0.39 2.15 ? 0.17 0 B NRR 3 10 6 6.16 ? 0.69 3 15 1 2.89 ? 0.46 HRR 2 21 123 46 2 6.13 ? 0.64 2 1 30 145 16 2.89 ? 0.58 ABR 1 1 14 29 20 6.06 ? 0.77 1 12 48 4 2.85 ? 0.54 SIR 6 19 7 6.03 ? 0.65 1 8 19 4 2.81 ? 0.69 OBR 1 3 44 15 6.16 ? 0.57 28 35 2.56 ? 0.50 KYR 4 15 5 6.04 ? 0.62 10 14 2.58 ? 0.50 Mean 6.10 ? 0.06 2.76 ? 0.15

00 C) 810 S. CHOW ET AL.

No. r trial rB type-1 A type 1 NTR -NTR* 2 NTR NTR* 3 NTR NTR* 4 NTR -NTR* 5 NTR NTR* 6 NTR - NTR* 7 ASR ASR* 8 ASR -ASR* 9 NTR HKP* 10 NTR ASR* 11 ASR -NTR * A type 12 NTR - - HRR-- - HNPA-- - -NTR* 13 NTR 0-*HRRR@- - HNPO NoNTR* 14 NTR -0-H HRR-- - HKP -m HNP@ - NTR* 15 NTR- -RR HR-- -- HKP@ HNP- *NTR* 16 NTR--HR HRRR---- HKP-- - HNP) egg lost* 17 NTR f HRR@- - ATPO TMR* 18 NTR - HRRA----b- TTP* 19 NTR HRR, egg lost* 20 NTR - HRRE egg lost* 21 ASR -HRR* egg lost* 22 ASR HRRO egg lost* 23 TTP HRR) egg lost*

A type - rB type1,

24 HRR - HRR* 25 HRR - HRR* 26 HRR - HRR* 27 HRR -HRR* 28 HRR - HRR* 29 HRR HRR* 30 HRR -- NRR* 31 NRR - HRR* 32 NRR HRR* B typ. ~33 NRR HRR* 34 HRR - -NTR@ HRR* 35 HRR--- *NTR@ -0HRR* 36 HRR - *NTR- - -- HRR* 37 HRR - NTRR-- *ATPO-- *TMR@-- HRR* 38 HRR -ATATPP NTRO - -HRR* 39 HRR ATPO- *AMR@- NTR) egg lost* 40 HRR ASR* - HRR* 41 HRR - ASR- -- HRR* REPRODUCTIVE ISOLATION 811

B males, and the males responded in seven TABLE 4. Mean eggvolumes in each local population. cases. Althoughmatings took place in five Abbreviationsof localities are as given in Figure 1.

of these seven cases, spawned eggs fell off Number withina week withoutany cleavage. For the of indi- Mean egg volume Type Locality viduals (mm-)+ SD reciprocaltype of pairing,type-A males re- sponded to type-B females in only two of A IHP 4 0.956 ? 0.075 HNP 4 0.993 ? 0.053 13 combinations;although mating was ap- HKP 1 0.860 ? 0.061 parentlysuccessful in one of these, no fer- TTP 14 0.998?0.111 tilized eggs were obtained. Thus, male re- ATP 3 0.991 ? 0.010 sponse was observed only in nine cases NTR 24 0.975 ? 0.186 x B B x A) TMR 1 0.946 ? 0.066 among 25 intertype(A and LK 5 0.659 ? 0.072 combinations;in contrast,males responded LB 5 0.142 ? 0.093 in 37 cases of 43 intratype(A x A and B x ASR 3 0.426 ? 0.047 B) combinations (G = 18.19, P < 0.001). B KKR 5 0.520 ? 0.038 Therewere 14 combinationsbetween A-type NRR 3 0.619 ? 0.050 individuals from differentlocalities, and HRR 25 0.497 ? 0.068 ABR 6 0.538 ? 0.034 males respondedin eightcases. Matingtook OBRa 3 0.623 ? 0.042 place in six ofthese cases, and fertilizedeggs a Sample measured at the eyed-stageof development. were produced by five females fromthese matings. Mating and fertilizedeggs were confirmedin all fourcombinations between crustaceanspecies (Tracey et al., 1975; Ne- B-typeindividuals fromdifferent localities meth and Tracey, 1979; Berglundand La- (trials 30-33). Type-A male response to- gercrantz,1983; Boulton and Knott, 1984; ward type-Afemales fromdifferent locali- Huber, 1985; Chow and Fujio, 1985b) but ties did not differfrom male response for very similar to that between the American intertypecombinations (G = 1.63, P > 0.1), and European lobsterspecies (Hedgecock et and the frequencyof matingswas likewise al., 1977). In the last case, since the two similar (G = 1.47, P > 0.1). However, the homaridlobsters can easily mate in the lab- frequencyof matingsfor intratype pairings oratoryand F, hybridscan be obtained(Tal- involvingB-type shrimp from different lo- bot et al., 1984), it appears thatgeographic calities was significantlygreater than those isolation has provided the effectivemech- forintertype pairings and forintratype pair- anism for genetic differentiationbut that ingsinvolving A-type shrimp from different matingisolation is not yetaccomplished. In localities. In all cases, if spawned eggswere contrast,the two types of P. paucidens in- confirmedto be fertilized,normal devel- habit overlappingzones in the same water opment and hatchingwere observed. systemand yet no gene exchange has been observed betweenthe types(Chow and Fu- DIsCUSSION jio, 1985a). Althoughhybrid zones are ob- The distinctallelic distributionsthat al- served in many species (Barton and lowed classificationof local populations of Hewitt, 1981), we found no intermediate Palaemon paucidens into two typesindicate types in the present study, suggestingthe that thereis littleor no gene flowbetween presence of an effectivemechanism to iso- types. Average Nei's genetic distance be- late these two gene pools. tweenthem (D = 0.1 186) is verylarge com- Rapid and vigorous courtshipbehavior pared with a number of cases among con- of males towardsmature and parturientfe- specific local populations of decapod males has been observed in many decapod

FIG. 3. Experimentalmatings within and between two types of Palaemon paucidens. See Figure 1 for the abbreviationsof locality names. Symbols: 0 = no response; A = searching; * = mating; * = mating and subsequent development of the eggs spawned (indicatingfertilization). * "Egg lost" indicates that all of the spawned eggs were lost withina week withoutcleavage. 812 S. CHOW ET AL.

TABLE 5. Summaryof matingexperiments.

Response Mating

Combination + - G + - G Intratype 37 6 18.19 35 8 22.49 Intertype 9 16 (P < 0.001) 6 19 (P < 0.001)

Interpopulation(A x A') 8 6 1.63 6 8 1.47 Intertype 9 16 (P > 0.1) 6 19 (P > 0.1)

Interpopulation(A x A') 8 6 3.79 6 8 5.61 Interpopulation(B x B') 4 0 (P > 0.05) 4 0 (P < 0.025)

Interpopulation(B x B') 4 0 7.22 4 0 9.81 Intertype 9 16 (P < 0.01) 6 19 (P < 0.005)

species (Burkenroad, 1947; Kamiguchi, sahara and G. D. Gometz of Hiroshima 1972; Ruello et al., 1973; Bauer, 1976; Lee University,and H. Akiyama and M. Mae- and Fielder, 1982). Sexual attraction of hata of the Museum and Aquarium of Lake males is believed to be stimulatedby pher- Biwa fortheir help duringcollection of the omonesthat are releasedby the females upon .We also thank W. D. Anderson, molting(Ryan, 1966; Atema and Engstrum, Jr.of The College of Charlestonand J. M. 1971). Change in these pheromoneswould Macaranas of the Marine Science Institute hinder male recognitionof parturientfe- fortheir kind readingof the manuscript. males and provide effectivepremating re- productive isolation. Poor attraction ob- servedbetween A and B typesofP. paucidens LITERATURE CITED duringmating might reflect a prematingiso- ATEMA, J., AND D. G. ENGSTRUM. 1971. Sex phero- lationmechanism of this type. The fewcases mone in the lobster,Homarus americanus. Nature of intertypemating show that premating 232:261-263. isolation between the types is incomplete. BARTON, N. H., AND G. M. HEWITT. 1981. Evolution However, our matingexperiments between and speciation,pp. 109-145.In W. R. Atchleyand D. S. Woodruff(eds.), Hybrid Zones and Specia- the types indicate that there is also strong tion. CambridgeUniv. Press, N.Y. postmating isolation between types, pre- BAUER, R. T. 1976. Mating behavior and spermato- sumably due to hybridbreakdown. phore transferin the shrimp Heptacarpuspictus Distinctlyhigher genetic and morpholog- (Stimpson) (Decapod: Crustacea: Hippolytidae).J. ical differentiationin the A typethan in the Nat. Hist. 10:415-440. BERGLUND, A., AND U. LAGERCRANTZ. 1983. Genetic B typeis likelyto be attributableto habitat differentiationin populations of two Palaemon differences.Apparently, the A typeexpand- prawn species at the Atlanticeast coast: Does gene ed to ponds and lakes, wheregene flowwas flow prevent local adaptation? Mar. Biol. 77:49- geographicallyrestricted, while the B type 58. BOULTON, A., AND B. KNOTT. 1984. Morphological inhabitedonly rivers,where the larvae can and electrophoreticstudies of the migrate among the estuaries throughthe (Crustaceae) ofthe Perthregion, Western Australia. ocean. The significantlyshorter larval pe- Aust. J. Mar. FreshwaterRes. 35:769-783. riod of the A type(2 weeks) compared with BURKENROAD, M. D. 1947. Reproductive activities theB type(1 month)(Chow and Fujio, 1986) of decapod crustacea. Amer. Natur. 81:392-398. CHOW, S., AND Y. FuJIo. 1985a. Biochemical evi- is a furtherindication of differentpatterns dence of two typesin the freshwatershrimp Palae- ofgene dispersal in riverpopulations of both monpaucidens inhabiting same watersystem. Bull. types. Japan. Soc. Sci. Fish. 51:1451-1460. 1985b. Population geneticsof the palaemo- ACKNOWLEDGMENTS nid shrimps(Decapod; Crustacea). I. Genetic vari- abilityand differentiationof local populations.To- The authors are most gratefulto T. Su- hoku J. Agric. Res. 36:93-108. gawara,a Mayor of Kannari, Miyagi,S. Ka- 1985c. Population geneticsof the palaemo- REPRODUCTIVE ISOLATION 813

nid shrimps(Decapod: Crustacea). II. Geneticvari- . 1975. Molecular Population Genetics and ability and differentiationof species. Tohoku J. Evolution. North Holland, Amsterdam,Neth. Agric. Res. 36:109-116. NEMETH, S. T., AND M. L. TRACEY. 1979. Allozyme 1986. Differentiationof larval and post larval variabilityand relatednessin six crayfishspecies. developmentamong local populations of the fresh- J. Hered. 70:37-43. water shrimp Palaemon paucidens.Fish Genet. NISHINO, M. 1980. Geographicvariation in body size, Breed. Sci. 11:29-33. brood size and egg size of a freshwatershrimp, Pa- HEDGECOCK, D., K. NELSON, J. SIMONS, AND R. SHLE- laemonpaucidens de Haan, withsome discussions SER. 1977. Genetic similarityof American and on brood habit. Jap. J. Limnol. 41:185-202. Europeanspecies of the lobster Homarus. Biol. Bull. RUELLO, N. V., P. F. MOFFITT, AND S. G. PHILLIPS. 152:41-50. 1973. Reproductive behaviour in captive fresh- HUBER, M. E. 1985. Population genetics of eight water shrimp Macrobrachiumaustraliense Hol- species of Trapezia(Brachyura: Xanthidae), sym- thuis.Aust. J. Mar. FreshwaterRes. 24:197-202. bionts of coral. Mar. Biol. 85:23-36. RYAN, E. P. 1966. Pheromone:Evidence in a decapod KAMIGUCHI, Y. 1972. Mating behavior in the fresh- .Science 151:340-341. water prawn,Palaemon paucidens. A studyof the SNEATH, P. H., AND R. R. SOKAL. 1973. Numerical sex pheromoneand its effecton males. J. Fac. Sci. . Freeman, San Francisco, CA. Hokkaido Univ. Ser. VI, Zool. 18:347-355. TALBOT, P., C. THALER, AND P. WILSON. 1984. LEE, C. L., AND D. R. FIELDER. 1982. Maintenance Spawning,egg attachment and eggretention in cap- and reproductivebehavior in the freshwaterprawn tivelobster (Homarus americanus). Aquaculture 37: Macrobrachiumaustralience Holthuis (Crustacea: 239-249. : Palaemonidae). Aust. J. Mar. Fresh- TRACEY, M. L., K. NELSON, D. HEDGECOCK, R. A. waterRes. 33:629-646. SHLESER,AND M. L. PRESSICK. 1975. Biochemical NEI, M. 1972. Geneticdistance between populations. geneticsof lobsters: Genetic variation and thestruc- Amer. Natur. 106:283-292. ture of American lobster (Homarusamericanus) 1973. Analysisof gene diversityin subdivid- populations.J. Fish. Res. Board Can. 32:2091-2101. ed populations.Proc. Nat. Acad. Sci. USA 70:3321- 3323. CorrespondingEditor: T. Mukai