University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Anthony Zera Publications Papers in the Biological Sciences

1981

Genetic Structure of Two Species of Waterstriders (: ) with Differing Degrees of Winglessness

Anthony J. Zera University of Nebraska - Lincoln, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/bioscizera

Part of the Microbiology Commons

Zera, Anthony J., "Genetic Structure of Two Species of Waterstriders (Gerridae: Hemiptera) with Differing Degrees of Winglessness" (1981). Anthony Zera Publications. 33. https://digitalcommons.unl.edu/bioscizera/33

This Article is brought to you for free and open access by the Papers in the Biological Sciences at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Anthony Zera Publications by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. Published in Evolution, 35(2), 1981, pp. 218–225. Copyright © 1981 Society for the Study of Evolution; published by Black- well Publishing. Used by permission.

Genetic Structure of Two Species of Waterstriders (Gerridae: Hemiptera) with Differing Degrees of Winglessness

Anthony J. Zera Biological Science Group, University of Connecticut, Storrs, Connecticut 06268 (Present address: Department of Ecology and Evolution, State University of New York at Stony Brook, Stony Brook, New York 11794) Submitted November 1979; revised July 1980. This paper is part of a dissertation submitted to the Graduate School of the University of Connecticut in partial fulfillment of the requirements for the M.S. degree.

The Gerridae (Hemiptera: Insecta) is a 1976; Selander, 1976). However, additional fac- worldwide family whose constituent species tors may counteract the effects of reduced dis- exhibit dramatic inter- and intra-specific varia- persal by flight. Gene flow among populations tion in the degree of winglessness (Brinkhurst, may occur via alternate modes of dispersal, in- 1960; Vepsäläinen, 1978; Calabrese, 1979). At cluding passive stream drift and overland dis- one extreme, the family contains species which persal (Riley, 1920). Furthermore, marked ge- are fully winged in all populations and during netic differentiation among populations is not a all seasons, while several species consist almost necessary consequence of severely reduced dis- exclusively of wingless morphs over large geo- persal if locality-independent balancing selec- graphical ranges and during all seasons. Many tion is operating (McKechnie et al., 1975). species exhibit the intermediate case of wing- In this study I compare patterns of spatial polymorphism: the occurrence of various com- variation of polymorphic enzyme-loci and levels binations of fully winged, partially winged and/ of variability in two species of waterstriders (Ger- or wingless morphs in the same population at ridae: Hemiptera) with differing degrees of wing- the same time. Various wing-polymorphic spe- lessness: the nearly wingless remigis and cies show differing patterns of spatial and/or the wing-polymorphic canaliculatus. temporal changes in morph ratios and patterns may vary both inter- and intraspecifically. Species Studied The dramatic differences in frequency of winged morphs pose intriguing questions re- Gerris remigis is the most widely distrib- garding the evolutionary forces responsible for uted and the most abundant North American ger- degree of winglessness and the relationship be- rid species (Drake and Harris, 1934), occurring tween degree of winglessness and genetic struc- throughout the continental United States, Mex- ture of water-strider species. One might expect ico, south to Guatemala. In the eastern U.S., G. genetic structure to be strongly influenced by remigis occurs principally on flowing water. Ca- degree of winglessness via reduction of flight labrese (1979) recorded greater than 99% wing- dispersal ability and consequent reduced gene less morphs in museum and field collections of flow. Thus, species composed almost exclu- G. remigis in Connecticut. Limnoporus canalic- sively of wingless individuals should exhibit ulatus, which until recently was known as Ger- patterns of marked genetic differentiation and ris canaliculatus (Anderson, 1975), ranges from reduced levels of within-population variability Florida to Maine and west to Illinois and Mich- typically found in organisms with reduced dis- igan (Calabrese, 1974). The species occurs on persal (Avise and Selander, 1972; Laing et al., ponds, swamps and sluggish streams. Genetic structure of 2 waterstriders 219

Populations of L. canaliculatus consist of There are no published data for dispersal in various proportions of fully-winged, wing- either species. However, it is likely that pop- less, and (very rarely) short-winged individu- ulation mixing occurs in L. canaliculatus as als. The frequency of fully-winged individuals a result of dispersal of winged individuals to varies seasonally, the highest frequency occur- and from overwintering sites, as has been re- ring from September to November in Connect- ported for several European gerrid species (see icut and Virginia populations (Bobb, 1951; Ca- Discussion). labrese, 1979). Calabrese (1979) reported that Connecticut populations contained greater than Sampling and Electrophoresis 25% winged individuals during all seasons. However, I have found Connecticut and Mas- Samples were taken from eight populations sachusetts populations to consist almost ex- of Gerris remigis ranging from Maine to Vir- clusively of wingless individuals from July ginia and seven populations of Limnoporus can- through early August (Zera, unpubl.). This ob- aliculatus ranging from Maine to Georgia during servation is similar to the data of Bobb (1951) July–October of 1976 (Table 1). No sampling for Virginia populations. sites were directly connected by water. All sites Both G. remigis and L. canaliculatus are at contained water throughout the year. least bivoltine in Connecticut (Zera and Saks, The number of winged and wingless unpubl.), while in Florida, L. canaliculatus morphs was recorded in each fall (Sept.) pop- breeds throughout the year (Penn and Gold- ulation sample of L. canaliculatus. The fre- smith, 1950). In Connecticut, both species quency of morphs was recorded in each pop- overwinter as adults. Individuals emerge the ulation of G. remigis in both summer and fall following April-May and mate throughout this since there is no evidence for seasonal change time. Adults of the subsequent summer gener- in the near unitary frequency of wingless ation begin to emerge in late June, and adults morphs (Calabrese, 1979). of both species can be seen continuously from The following enzymes were studied in the this time through late October. Mating pairs are two species: α-Glycerophosphate dehydroge- commonly seen from July through August. nase (α-Gpdh, E.C. 1.1.1.8); Malate dehydro- 220 A. J. Zera in Evolution 35 (1981) genase (Mdh, E.C. 1.1.1.37), anodal Mdh-1 ing buffer “B” of Ayala (1972). Peptidase, NAP, and cathodal Mdh-2; Isocitrate dehydrogenase and Esterases were resolved using the LiOH (Idh-1, E.C. 1.1.1.42); Malic enzyme (Me, E.C. buffer of Selander et al. (1971). 1.1.1.40); and Esterase (Est, E.C. 3.1.1.1), Est- Gels were stained for the appropriate en- 3 and Est-4. zymes as in Shaw and Prasad (1970). For In order to obtain a larger sample of loci ACPH, gels were soaked in 0.25 M Na-acetate for heterozygosity estimates, three populations buffer, pH 4.5, for 30 min prior to staining. of L. canaliculatus (Petersham, Pink Ravine Observed genotypic numbers were com- and Niantic) and two populations of G. remi- pared with their Hardy-Weinberg expectations gis (Pink Ravine and Niantic) were resampled by the χ2-test (Sokal and Rohlf, 1969). Geo- in 1977 for variation in the following additional graphic variation of polymorphic loci was an- enzymes: Glutamate-oxaloacetate transaminase alyzed using R × C tests of homogeneity of (GOT, E.C. 2.6.1. 1), GOT-1 and GOT- 2; 6- genotypic numbers (Workman and Niswander, Phosphogluconate dehydrogenase (6- PGDH, 1970). Polymorphic loci surveyed during 1976 E. C. 1.1.1.43); Phosphoglucose isomerase as well as additional polymorphic loci discov- (PGI, E. C. 5.3.1.9); Glucose-6- phosphate de- ered during the 1977 sampling were analyzed. hydrogenase (G-6-PDH, E.C. 1.1.1.49); Leu- Rare genotypic classes were pooled with their cine napthylamidase (NAP, E.C. 3.4.—.—); most electrophoretically similar class until not Acid phosphatase (ACPH, E.C. 3.1.3.2); Fuma- more than 20% of the cells of the R × C ta- rase (FUM, E.C. 4.2. 1.2); Xanthine dehydro- ble contained expected values less than five genase (XDH, E.C. 1.2.3.2); Phosphoglucomu- and no cell contained an expected value less tase (PGM, E.C. 2.7.5.1); Peptidase (PEP, E.C. than one (Conover, 1971). These criteria could 3.4.1.—); and an additional Isocitrate dehy- not be met for the weakly polymorphic Idh-1 drogenase locus (IDH-2, E.C. 1.1.1.42). In to- and Me loci of L. canaliculatus. R × C tests tal, 18 loci were examined in G. remigis and 16 of homogeneity were still performed on these loci in L. canaliculatus. loci, since they conform to less stringent cri- In addition to the populations listed above, a teria (i.e., no cell of an R × C table with an new population (Swift) of G. remigis was sam- expected value less than one, Lewontin and pled during 1977 for variation at the 18 loci sur- Felsenstein, 1965). A posteriori STP tests (So- veyed in the other populations of this species. kal and Rohlf, 1969) were; done on all R × C Gerrids were homogenized in 0.1 ml (G. tests of homogeneity. The Mdh-2 and Est-4 remigis) or 0.05 ml (L. canaliculatus) of 0.05 loci of G. remigis were not analyzed statisti- M Tris-HCl buffer, pH 7.0, containing 0.01 M cally for geographical variation, since it is ob- mercaptoethanol. Crude homogenates were ab- vious by inspection that marked geographical sorbed onto paper wicks and applied to horizon- variation is present. tal starch gels prepared with Electrostarch (lots Two values of average heterozygosity per

307 or 371; Otto Hiller, Madison, Wisconsin) at individual (H) were calculated. One value, Ht, a concentration of 12% (w/v). was calculated using all loci surveyed in each

The 0.135 M Tris-citrate buffer of Shaw and species, while a second value, Hh, was calcu- Prasad (1970) was adjusted to different pH’s by lated using only the 15 presumed homologous varying the concentration of citric acid. MDH, loci surveyed in both species. The calculation ME, IDH, GOT, a-GPDH, FUM, ACPH in L. of average heterozygosity using only homol- canaliculatus only, and PGI were resolved at ogous loci was done since this eliminates the pH 7.3. G-6-PDH and 6-PGDH were resolved confounding effects of interlocus variation in at pH 8.0 and ACPH in G. remigis was resolved interspecific comparisons of average heterozy- at pH 6.8. Phosphoglucomutase was resolved gosity (Selander, 1976). For this reason, only using the Tris-maleate-EDTA buffer of Shaw Hh, was used in the comparison of heterozygos- and Prasad (1970) and XDH was resolved us- ity between the two species. Genetic structure of 2 waterstriders 221

RESULTS Frequency of Wingless Morphs

Eleven of 18 loci surveyed in G. remigis were All sampled G. remigis were wingless, except monomorphic (a-Gpdh, Me, Got-1, Got-2, 6- for six individuals from the Maplewood pop- Pgdh, Fum, Pgi, G-6-pdh, Nap, Pep, and Xdh), ulation (Table 1). In contrast, the frequency while 8 of 16 loci in L. canaliculatus were of wingless individuals varied considerably monomorphic (Got-1, Got-2, Fum, Nap, G-6- among populations of L. canaliculatus in the pdh, Mdh-1, Idh-2, and Xdh). A monomorphic Beddington and Bass River populations, re- locus is defined as a locus containing one and spectively (Table 1). the same allele with a frequency greater than 0.95 in all sampled populations. Geographical Variation Allele frequencies of polymorphic loci are given in Tables 2 and 3. No significant devia- Five of six polymorphic loci in G. remigis tions from Hardy-Weinberg expectations were exhibited significant among-population hetero- observed, except at the 6-Pgdh locus in the Ni- geneity (Mdh-1, G(8) = 243.72, P < .005; Mdh- antic population of L. canaliculatus. No bio- 2, heterogeneous by inspection; Est-3, G(3) = logical significance is attributed to this single 14.35, P < .005; Est-4, heterogeneous by in- deviation. spection; Pgm, G(2) = 13.81, P < .005). Acid 222 A. J. Zera in Evolution 35 (1981)

phosphatase, the only polymorphic locus which .005), significant heterogeneity at the Me locus did not exhibit significant among-population was due to the Bass River population. The sub- heterogeneity, approached statistical signif- set of samples excluding the Bass River popu- icance (G(2) = 5.88, .l > P > .05). In contrast, lation was statistically homogeneous. Similarly, five of seven polymorphic loci surveyed in L. the majority (but not all) of the heterogeneity at canaliculatus exhibited no significant among- the Est-3 locus was due to the Bass River popu- population heterogeneity (α-Gpdh, G(6) = 7.54, lation. This population was the only population P > .l, Idh-1, G(6) = 10.99, .1 > P > .05; Pgm, to form a unique subset in the STP analysis. G(4) = 1.57, P > .l; Pgi, G(4) = 7.48, P > .1; The pattern of variation was very different Acph, G(6) = 4.34, P > .l). Of the two loci which for polymorphic loci of the two species. Four were significantly heterogeneous (Est-3, G(20) of six polymorphic loci of G. remigis (Mdh-1, = 133.65, P < .005 and Me, G(6) = 53.93, P < Mdh-2, Est-3, Est-4) exhibited sharp disconti- Genetic structure of 2 waterstriders 223 nuities in allele frequencies with different al- L. canaliculatus exhibits the pattern of geo- leles fixed in some populations at the Mdh-2 graphical variation and levels of variability typ- locus and nearly fixed at the Mdh-1 locus (Ta- ically found in continental species (Ay- ble 2). At the Est-4 locus, an allele which was ala, 1972; Lewontin, 1974) and other winged nearly fixed in the Shenandoah population was gerrid species (Varvio-Aho et al., 1978; Varvio- absent from the Maplewood population. How- Aho and Pamilo, 1979). In these species (as in ever, at each locus in L. canaliculatus, the same L. canaliculatus) allele frequencies of polymor- allele (or pair of alleles at the Est-3 locus) pre- phic loci are often homogeneous over large dis- dominated in all populations, with the excep- tances. Average heterozygosity for L. canalic- tion of the Est-3 locus in the Bass River popu- ulatus (0.234 ± 0.072) is within the range for lation (Table 3). “typical” or invertebrates (Powell, 1975; Local differentiation among populations of Selander, 1976). G. remigis was marked. For example, the STP In contrast to L. canaliculatus, polymor- analysis of the Est-3 locus showed significant phic loci of G. remigis typically exhibit marked among-population heterogeneity for the Con- spatial variation of allele frequencies, some- necticut samples. Different alleles predomi- times over short distances. Average heterozy- nated at the Mdh-2 locus in the Swift and Pink gosity in G. remigis (H = 0.058 ± 0.026) is well Ravine populations. These were also signifi- below the value for “typical” insects or inver- cantly different at the Pgm locus. Populations tebrates (Powell, 1975; Selander, 1976). Al- separated by five miles or less were -signifi lozyme data for G. remigis are similar to data cantly different at the Mdh-2 locus (Pink Ra- obtained for species inhabiting geographical or vine and Schoolhouse Brook, G(1) = 4.77, P ecological islands (Avise and Selander, 1972; < .05; Swift and Petersham, G(1) = 15.17, P < Saura et al., 1973; Laing et al., 1976; Selander, .005). No local differentiation was observed at 1976). For example, in six cave populations of polymorphic loci in L. canaliculatus. Esterase- troglobite beetle, Ptomaphagus hirtus, amounts 3, although exhibiting significant heterogeneity of genetic variability were about one-third that for the total sample of all populations, exhib- of “typical” invertebrates. Of six polymorphic ited no significant among-population hetero- loci studied, different alleles were fixed at three geneity for the subset of samples north of Bass loci, while large spatial differences in allele fre- River. quencies were observed at the other three poly- morphic loci (Laing et al., 1976). Average Heterozygosity The similarity of the allozyme data between G. remigis and other “island” species, where

Since Hh and Ht (see Materials and Meth- reduced gene flow is either known or strongly ods) were essentially identical in each species, inferred, suggests that gene flow is similarly only Hh values will be reported. Average het- reduced among populations of G. remigis. Re- erozygosity for G. remigis (H = 0.058 ± 0.026) duced gene flow would result in a population was significantly lower than for L. canalicula- structure consisting of isolated demes which tus (H = 0.234 ± 0.072), as tested by Wilcox- could differentiate due to selection and/or drift, on’s signed ranks test (Ts = 13, P < ,025; two accounting for both the degree and pattern of tailed test). spatial differentiation of allele frequencies. Re- duced levels of variability could result from Discussion either small present population sizes or the founder effect during colonization. Gerris remigis and Limnoporus canalicu- High mortality during the overwintering latus differ markedly, both in degree of popu- period was recorded in British populations of lation differentiation at polymorphic loci and the wingless riverine waterstrider, Gerris na- in amounts of variability within populations. jas (Brinkhurst, 1966). In a six-year study, win- 224 A. J. Zera in Evolution 35 (1981) ter mortality was always greater than 55%, and canaliculatus which was composed almost ex- sometimes exceeded 90%, in each of four pop- clusively of wingless individuals (94% wing- ulations. Populations of G. remigis were much less individuals, Table 1). The almost total ab- smaller in the spring than in the summer or fall sence of winged individuals indicates that gene (Zera and Saks, unpubl.), indicating that high flow via flight into Bass River must be severely winter mortality may also occur in this species. reduced. Both the Est-3 and the Me loci exhibit Population bottlenecks resulting from high highly divergent allele frequencies in this pop- winter mortality can severely reduce variability ulation, compared with other populations of L. (Nei, 1975), and may be an additional factor re- canaliculatus (Table 3). Thus, as in G. remigis, sponsible for the low variability found in popu- when dispersal by flight is severely reduced, lations of G. remigis. genetic differentiation may result. The gener- Spatially homogeneous allele frequencies ality of this hypothesis is being tested further among populations of L. canaliculatus may be by investigations of differences among popu- the result of extensive dispersal and gene flow. lations of L. canaliculatus composed predom- However, homogeneous allele frequencies do inantly of wingless morphs. not necessarily imply extensive gene flow since only a small amount of gene flow is needed to SUMMARY swamp out local differentiation due to random drift (Lewontin, 1974). Similar allele frequen- The relationship between degree of wing- cies in different populations may also be due lessness and genetic structure was investigated to locality-independent balancing selection in in two waterstrider species (Gerridae: Hemip- the absence of gene flow. Therefore, allozyme tera) with differing degrees of winglessness: data for L. canaliculatus cannot decisively sep- the nearly wingless Gerris remigis and the arate the relative contributions of dispersal, se- wing-polymorphic Limnoporus canaliculatus. lection and drift. However, several additional Five of six polymorphic loci in G. remigis ex- points should be made. Dispersal is reported to hibited significant spatial variation of allele fre- commonly occur in many winged gerrid spe- quencies. There was fixation or near fixation cies (Brinkhurst, 1960; Vepsäläinen, 1971; Cal- of different alleles in different populations at lahan, 1974; Landin and Vepsäläinen, 1977). three loci. In contrast, five of seven polymor- Winged adults of G. lacustrus have been found phic loci surveyed in L. canaliculatus exhibited hibernating far from water (Douglas, 1882; geographically homogeneous allele frequen- Brinkhurst, 1960) and winged forms of the con- cies. Average heterozygosity in G. remigis (H tinental European species, G. rufoscutellatus, = 0.058 ± 0.026) was one-quarter the value of have been found in Britain (Brinkhurst, 1960). L. canaliculatus (H = 0.234 ± 0.072) and was Winged gerrids have been found in small pools one-third to one-quarter the value for “typical” on top of three story buildings (Calabrese, insects or invertebrates. 1974) and in the laboratory winged morphs of These data indicate that differing degrees of L. canaliculatus commonly fly from con- winglessness have resulted in very different ge- tainers. Hence, it seems unlikely that dispersal netic structures in the two species. is restricted in L. canaliculatus. The importance of flight dispersal as a fac- Acknowledgments tor responsible for the homogeneous allele fre- quencies in nearly all populations of L. cana- Special thanks are extended to my advisor liculatus may also be inferred by examining Alari H. Brush and members of my committee, allele frequencies in the one population where Carl Schaefer and Fred A. Streams, for much flight dispersal seems to be severely reduced, encouragement and constructive criticism. Mr. the Bass River population. The Bass River pop- Cecil Smith collected and sent L. canalicula- ulation was the only surveyed population of L. tus from Athens, Georgia. Diane Calabrese pro- Genetic structure of 2 waterstriders 225 vided information on collecting localities near 29:156–160. Storrs, Connecticut, and provided valuable as- Lewontin, R. C. 1974. The Genetic Basis of Evolutionary sistance in the initial stages of this project. Mr. Change. Columbia University Press, N.Y. and Mrs. Raymond E. Saks generously pro- Lewontin, R. C., & J. Felsenstein. 1965. The robust- ness of homogeneity tests in 2 x n tables. Biometrics vided food and lodging during many of my col- 21:19–30. lecting trips. Lawrence Harshman, David Innes, McKechnie, S. W., P. R. Ehrlich, & R. R. White. 1975. Richard K. Koehn, Graham Mark and Peggy Population genetics of Euphydryas butterflies. I. - Ge Saks read earlier drafts of this manuscript. This netic variation and the neutrality hypothesis. Genetics work was supported by NSF grant DEB 76- 81:571–594. Nei, M., T. Maruyama, & R. Chakraborty. 1975. The bot- 20967 to Alan H. Brush and USPHS grant GM tleneck effect and genetic variability in populations. 21133 to Richard K. Koehn. Evolution 29:l–10. Penn, G. H., & R. M. Goldsmith. 1950. The life history of the southern water strider, Gerris canaliculatus (He- Literature Cited miptera, Gerridae). J. Tenn. Acad. Sci. 25:76–81. Powell, J. R. 1975. Protein variation in natural populations Anderson, N. 1976. The Limnogonus and Neo-Gerris of of . Evol. Biol. 8:79–119. the old world with character analysis and reclassifica- Riley, C. F. C. 1920. Migratory responses of water-striders tion of the Gerrinae (Hemiptera: Gerridae). Entomol. during severe droughts. Bull. Brooklyn Entomol. Soc. Scandin. Suppl. 7:7–96. 15:l–10. Avise, J. C., & R. K. Selander. 1972. Evolutionary genetics Saura, A., O. Halkka, & J. Lokki. 1973. Enzyme gene het- of cave-dwelling fishes of the genus Astyanax. Evolu- erozygosity of Philaenus spumarius (L.) (Homoptera). tion 26:1–19. Genetica 44:459–473. Ayala, F. J., J. R. Powell, M. L. Tracey, C. A. Mourao, & Selander, R. K. 1976. Genic variation in natural popula- S. Perez-Salas. 1972. Enzyme variability in the Dro- tions, p. 21–45. In F. J. Ayala (ed.), Molecular Evolu- sophila willistoni group. IV. Genic variation in nat- tion. Sinauer Assoc., Inc., Sunderland. ural populations of Drosophila willistoni. Genetics Selander, R. K., M. H. Smith, S. Y. Yang, W. E. Johnson, 70:113–139. & J. B. Gentry. 1971. Bio-chemical polymorphism Bobb, M. L. 1951. The life history of Gerris canalicula- and systematics in the genus Peromyscus. I. Variation tus Say in Virginia (Hemiptera: Gerridae). Va. J. Sci. in the old-field mouse (Peromyscus polionotus). Univ. 4:102–108. Texas Publ. 7 103:49–91. Brinkhurst, R. O. 1960. Alary polymorphism in the Ger- Shaw, C., & R. Prasad. 1970. Starch gel electrophoresis of roidea (Hemiptera-Heteroptera). J. Anim. Ecol. enzymes—a compilation of recipes. Biochem. Genet. 28:211–230. 4:297–320. ———. 1966. Population dynamics of the large pond- Sokal, R. R., & F. J. Rohlf. 1969. Biometry. W. H. Freeman skater Gerris najas Degeer (Hemiptera-Heteroptera). Co., San Francisco. J. Anim. Ecol. 35:13–25. Varvio-Aho, S., O. Jarvinen, & K. Vepsäläinen. 1978. En- Calabrese, D. M. 1974. The biology of the species of Ger- zyme gene variation in three species of water-strid- ris Fabricius in Connecticut. M.S. Thesis. University ers (Gerris) (Heteroptera, Gerridae). Ann. Ent. Fenn. of Connecticut. 44:87–94. ———. 1979. Pterygomorphism in 10 Neartic species of Varvio-Aho, S., & P. Pamilo. 1979. Genic differentiation of Gem’s. Amer. Midl. Natur. 101:61–68. Gerris lacustris populations. Hereditas 90:237–249. Callahan, J. R. 1974. Observations on Gerris incognitus Vepsäläinen, K. 1971. The role of gradually changing day- and Gerris gillettei (Heteroptera: Gerridae). Proc. En- length in determination of wing length, alary dimor- tomol. Soc. Wash. 76:15–21. phism and diapause in a Gerris odontogaster (Zett.) Conover, W. 1971. Practical Nonparametric Statistics. population (Gerridae:Heteroptera) in South Finland. Wiley, N.Y. Ann. Acad. Sci. Finn. A, IU Biologica 183:l–25. Douglas, J. W. 1882. Gerris lacustris recorded in hiber- ———. 1978. Wing dimorphism and diapause in Gerris: nation far from water. Entomol. Mon. Mag. 19:20. determination and adaptive significance, p. 218–253.In Drake, C. J., & H. M. Harris. 1934. The Gerrinae of the H. Dingle (ed.), Evolution of Insect Migration and Dia- western hemisphere. Ann. Carnegie Mus. 23:179–240. pause. Springer-Verlag, N.Y. Laing, C., G. R. Carmody, & S. B. Peck. 1976. Population ge- Workman, P. L., & J. D. Niswander. 1970. Population stud- netics and evolutionary biology of the cave beetle Ptomo- ies on southwestern Indian tribes. II. Local differentia- phagus hirtus. Evolution 30:484–498. tion in the Papago. Amer. J. Hum. Genet. 22:24–29. Landin, J., & K. Vepsäläinen. 1977. Spring dispersal flights of pond-skaters Gerris spp. (Heteroptera). Oikos Corresponding Editor for Evolution: J. R. Powell