gene loci is commonly observed among natural populations of marine invertebrates. This differentiation reflects restricted gene flow among conspecific ...
BULLETIN OF MARINE SCIENCE, 39(2): 526-535, 1986 LARVAL INVERTEBRATE WORKSHOP
EVOLUTIONARY CONSEQUENCES OF RESTRICTED GENE FLOW AMONG NATURAL POPULATIONS OF THE COPEPOD, TIGRlOPUS CALIFORNICUS Ronald S. Burton ABSTRACT Extensive differentiation at electrophoretically-detected gene loci is commonly observed among natural populations of marine invertebrates. This differentiation reflects restricted gene flow among con specific populations or the action of natural selection. While differentiation indicates the potential for adaptive evolution, few studies have demonstrated the adaptive nature of genetic differentiation among marine invertebrate populations. Two lines of evidence are presented that suggest that the extensive differentiation observed among T. califarnicus populations has resulted in differential adaptation ofIocal populations. (I) Based on the environmental conditions that might favor the GptF allele (predicted by previous physiological genetic investigations) and knowledge of the population structure ofthis species, an adaptive relationship between salinity variation and allele frequency was predicted. Data suggesting that such a relationship may hold among natural populations is presented. (2) The genetic structure of T. califarnicus populations (isolated populations differentiated at many loci) appears to be conducive to the establishment of polygenic interaction systems postulated by Wright's shifting balance theory of evolution. Inter-population hybridization experiments show that the F2 larvae of between-population matings show significantly higher mortality in response to hyperosmotic stress than those of within-population matings. This work suggests that each population possesses a gene pool consisting of an integrated complex of alleles; hybridization between populations breaks up these complexes and results in individuals lacking the physiological capacities present among individuals from each parental population.
Substantial progress has recently been made toward understanding the genetic structure of natural populations of marine invertebrates (reviews by Burton, 1983; Gooch, 1975; Levinton, 1980). Primarily through the use of protein electrophoresis, these efforts have demonstrated that genetic differentiation of conspecific populations is frequently observed among marine species (Berger, 1973; Berglund and Lagercrantz, 1983; Bulnheim and Scholl, 1981; Burton and Feldman, 1981; Janson and Ward, 1984; Schopf and Gooch, 1971; and many others) and is not uncommon even among species with long-lived planktonic larval stages (Johnson and Black, 1982; Koehn et al., 1980; Marcus, 1977; Theisen, 1978; Tracey et al., 1975). Differentiation in species with high dispersal potential could be the result of restriction in effective dispersal imposed by either the environment or behavior of dispersal stages (Burton and Feldman, 1982a), or may result from natural selection favoring resident over immigrant recruits to a population (Koehn et al., 1980). While in the latter case the adaptive significance of population differentiation is evident, relatively few cases of differentiation (especially at enzymecoding gene loci) have been shown to be the direct result of natural selection. Hence, while allozyme studies have shown ubiquitous genetic differentiation among marine invertebrate populations, we know little ofthe evolutionary consequences of the genetically sub-divided population structures found among these species. In previous papers, we have investigated the genetic structure of natural populations of the harpacticoid copepod Tigriopus calif amicus, a common, freeswimming inhabitant of high intertidal and supralittoral rock pools along the California coast. Despite its apparently high dispersal capacity, T. califomicus 526
BURTON: CONSEQUENCES OF RESTRICTED GENE FLOW
527
populations exhibit extensive differentiation (Burton et aI., 1979; Burton and Feldman, 1981); populations inhabiting neighboring rock outcrops often possess unique (i.e., found in only one or a few local populations) alleles in high frequency at one or more loci. Field transplantation experiments (Burton and Swisher, 1984) have further refined our understanding of population structure in this species by demonstrating that gene flow is extensive among T. californicus populations inhabiting pools located on single rock outcrops. Gene flow between outcrops separated by sandy beach, however, is highly restricted (Burton and Feldman, 1981). In another line of investigation, Burton and Feldman (1983) presented evidence for physiological differences among genotypes at one of the loci that had been used as a genetic marker: Gpt. This locus codes for the enzyme glutamate-pyruvate transaminase, which catalyzes the final step of alanine biosynthesis. Since rapid accumulation of alanine appears to play an important role in response to hyperosmotic stress in this species (Burton and Feldman, 1982b), polymorphism at the Gpt locus could result in differences in physiological response among genotypes. Data presented in Burton and Feldman (1983) suggest that such differences are observed not only at the physiological level (i.e., rates of alanine accumulation), but also in differential larval survivorship among Gpt genotypes subjected to hyperosmotic stress. Here I will attempt to integrate our population structure analyses with our physiological genetic investigations in order to address the evolutionary consequences of restricted gene flow among T. californicus populations. While our previous work on the genetic structure of natural populations of this species has focused on documenting levels of population genetic differentiation and gene flow, the investigations presented here provide our first indications that restriction of gene flow has resulted in adaptive differentiation of T. californicus populations. MATERIALS
AND METHODS
Sampling of natural populations of T. californicus in the vicinity of Santa Cruz, California, was carried out on an irregular basis from January 1980, to August 1984; some of the data used in the analysis here were previously reported in Burton and Feldman (1981). Pool salinities for each population sampled were measured to the nearest part per thousand using a hand-held refractometer. In March 1984, 10 coastal populations from Moss Beach (San Mateo County) to La Jolla '(San Diego County) were sampled and returned alive to Philadelphia where laboratory populations were established and maintained in continuous culture using artificial seawater (Instant Ocean). Field-collected animals were used for electrophoretic analysis, while animals used for the inter-population crosses reported below were FI and F, individuals from the laboratory populations. Electrophoretic analyses were carried out as described in Burton and Feldman (1981; 1982). Intra- and inter-population matings were obtained by isolating a group of adult males (approximately 20 individuals) from one population in a 100 x 15-mm petri dish and adding approximately 20 virgin females from the second population, Virgin females were obtained by collecting clasped pairs of T. californicus from the required population and dissecting the adult male from his immature "mate"; immature individuals obtained in this way are invariably virgin females (Burton et aI., 1981; Burton, 1985). Inter-population hybrids (FI individuals) were then reared en masse, or as clasped pairs were formed, single-pair crosses were isolated. The F, larvae used for the salinity-shock treatments were the progeny ofFI x FI matings. Salinity-shock treatments consisted of taking Stage I-II nauplii hatched and reared for at least 24 h in 50% seawater (salinity = 170/00) and pipetting known numbers (25-35 per trial) into 100% seawater (340/00).Survivorship was counted after 6 days. It should be noted that this stress simulates that which occurs when a low salinity tidepool (common habitat for T. californicus) is inundated by seawater during a period of high wave action; adult T. californicus show no mortality in response to this shock. RESULTS
Salinity Versus Gpt Allele Frequencies. -Over the period of 1980-1984, allele frequencies and salinities were determined for 53 samples representing pools
528
BULLETIN OF MARINE SCIENCE, YOLo 39, NO.2.
.30
•
. 25
>u z UJ ::>
.20
0
•
•• •
•
UJ
•
0
Cl..
Cl
0
. 10
0
.05
o
•
SCN
o PLA
*• WCAP NB
o•
0
• •
*; • *" * **
0
u.. I-
•
0
0
•
.15
•
•• • • • • • •• •
Cf
c:r: u..
1986
• • ** •
o CAP
0
o 00.
o
o
10
20
30
40
SALINITY
50
60
70
80
90
(Ppt)
Figure 1. Frequency of the Gpt allele in 53 T. californicus population samples plotted against the salinity of the pool from which the sample was taken. Different symbols represent samples taken from pools on different rock outcrops. F
located on 5 outcrops located between Natural Bridges State Park and Capitola in Santa Cruz County, California. Sample sizes for allele frequency determinations ranged from 30 to 160 individuals (average = 76). The relationship between pool salinity and GptF allele frequency is plotted in Figure 1; no significant relationship can be discerned (linear regression of arc-sine transformed GptF allele frequency on pool salinity was not significant at 0.05 level). This plot did, however, suggest that the five outcrops differed in average allele frequency and the range of salinity variation observed. These outcrop parameters are tabulated in Table 1. Allele frequencies are highly heterogeneous among outcrops (Kruskal-Wallace test, H = 31.0, P < 0.001). While mean salinities do not differ significantly (ANOV A, F = 0.99, P > 0.25), they are heteroscedastic (Bartlett's test, P < 0.01) and clearly differ in range and coefficient of variation. Without continuous monitoring of numerous pools on each outcrop it is difficult to assess the actual salinity regimes each presents to its resident T. califomicus population. However, characterization of each site (based on observations of fresh water input, range of pool elevations, and wave exposure) suggests that the differences in levels of salinity variation among outcrops tabulated in Table 1 accurately reflect both short- and long-term differences in the salinity regimes of the outcrops. For example, the SCN and PLA outcrops show the greatest range in salinity among pools at anyone time and are also more variable through time than the other three outcrops. With the exception of the typically low salinity CAP site, mean salinities on each of the other outcrops are near 100% seawater. There is a suggestive tendency for outcrops with extensive salinity variation to have higher GptF frequencies; all samples with GptF frequency above the median (0.167) were taken from the two outcrops with
BURTON: CONSEQUENCES
OF RESTRICTED
529
GENE FLOW
Table I. Salinity variation and Gpt'" allele frequencies on five Santa Cruz outcrops sampled on 5 to II dates in 1980-1984 Outcrop
Dates
SeN PLA NB WCAP CAP
II
6 9 5 7
No. samples
24 8 9 5 7
Salinity:
Mean,
range (0/00)
34, 3-96 28, 4-80 36,34-52 34,31-40 20, 5-34
Salinity CV(%o)
65.2 85.7 15.4 11.1 42.8
GptF x (±SEM)
0.219 0.150 0.134 0.086 0.056
(0.008) (0.013) (0.009) (0.023) (0.018)
the greatest salinity variation (SCN and PLA). Also notable is the perfect rankorder correlation between maximum outcrop salinity and allele frequency. Geographic Distribution of Allozyme Variants at Four Enzyme-encoding Gene Loci.-Allele frequencies for Got-I, Got-2. Gpt, and Me at 10 sites along the central to southern California coast are presented in Table 2. As we have previously observed with other loci and populations, many alleles have highly restricted, disjunct distributions and reach high frequency where they occur (Burton and Feldman, 1981; Burton, 1983). Inter-population Crosses and Larval Survivorship During Hyperosmotic Stress.Inter-population crosses were carried out using four source populations: two from central California, Santa Cruz (SCN) and Pescadero (PES) and two from southern California, Palos Verdes (AB, Abalone Cove Park) and La Jolla (LJ). Three classes of crosses were obtained: (1) within-population (control) crosses, where both male and female were taken from the same population; (2) within-region crosses, where the male was taken from one and the female taken from the other population within either the central or the southern California region; (3) between-region crosses, where one parent was taken from a central California population and the other from a southern California population. Survivorships ofF2 larvae subjected to hyperosmotic stress are presented in Table 3. Most of the potential cross types, including reciprocals, are represented in the data. Progeny of within-population crosses show significantly higher survivorship following hyperosmotic stress than those produced by interpopulation hybridization. Furthermore, there is a significant effect of region such that the apparent break-down of adaptations to hyperosmotic stress was more pronounced for between-region hybridizations than for within-region hybridizations. DISCUSSION
Inference of Gene Flow from Genetic Data. - The level of gene flow among geographically separated con specific populations may have important implications for the evolutionary response of a species to differences in selective forces among local environments (Crisp, 1978, for pertinent discussion focused on marine invertebrates). The extent to which populations can adapt to environmental heterogeneity depends both on environmental "grain" and gene flow. At the extremes, panmixia over broad geographic ranges will prevent evolutionary response to a spatially patchy environment, while total restriction of gene flow can result in the evolution of extensive local adaptation and, in the limit, speciation. The inference of levels of gene flow from either morphological or biochemical genetic data is often difficult. In the former case, the genetic basis ofthe trait must first be established and then the relative roles of natural selection and gene flow
530
BULLETIN OF MARINE SCIENCE, VOL. 39, NO.2,
1986
Table 2. Frequencies of common alleles (rounded to 0.1) for Gpt. Got-I, Got-2, and Me at study sites along the California coast (All frequencies are based on electrophoretic analysis of at least 50 individuals collected in March, 1984) Gpr
Site
F
M
GOI-I
F
S
Central California study sites Moss Beach Pescadero Santa Cruz 0.3 Capitola 0.1 Monterey Carmel
1.0 1.0 0.7 0.9 1.0 1.0
0.1
0.9 1.0 1.0 1.0 0.5 0.9
0.1
Southern California study sites Abalone Cove Laguna Beach Aliso Beach La Jolla 0.1
1.0 1.0 1.0 0.9
0.7
0.1
0.5
ES
VF
F
S
0.1
0.9 1.0 1.0 1.0 1.0 1.0
1.0 0.2 1.0 0.3
0.7
Me
Gor·2
VS
S
VF
F
0.2 0.1 0.4
1.0 0.3
0.7 1.0 1.0
M
S
1.0 0.8 0.9 0.6 1.0 1.0 1.0
1.0 0.9
0.1 1.0
must be somehow disentangled. Struhsaker's (1968) study of shell sculpture polymorphism in Littorina pieta suggests that this trait is under some level of genetic control and that selection influences the frequencies of morph types at different sites (although the mechanism by which selection acts is unclear since, as Struhsaker points out, the distribution of morphs in L. pieta is exactly opposite that in another polymorphic species, L. saxatilis rudis). More pertinent to this discussion, however, is the fact that the role of gene flow in this system remains unclear; while morph distributions of new recruits differ from those of older individuals at the same site (strong evidence for selection), the higher diversity ofmorph types among juveniles could be due to Mendelian segregation of resident multilocus genotypes affecting sculpture as well as gene flow from neighboring differentiated populations. Because this species has planktonic larvae, the latter explanation has been rather uncritically excepted (Gooch, 1975). However, Struhsaker (1968) reported that plankton tows suggest that most larvae are found close to shorelines where large adult populations occur, indicating that levels of gene flow might be low. An obvious advantage of protein polymorphisms as genetic markers for studies of population genetic structure is their single-locus co-dominant mode of inheritance, which eliminates the potential problem posed by complex segregation of morphological characters. A second and equally important advantage is the (frequent) availability of numerous polymorph isms independently segregating in the same natural populations, Still, inference of gene flow patterns among natural populations based on protein polymorphisms is not always straightforward and is usually based on ad hoc arguments specific to individual study systems (Christiansen and Frydenberg, 1974). Slatkin (1981) has proposed a semi-quantitative approach for inferring levels of gene flow in natural populations based on the frequencies of alleles present in varying proportions of the study populations. While this method is simple, objective, and gaining in popularity (Buroker, 1984), the extent to which it is an improvement over ad hoc methods is unclear for at least two important reasons: (1) it does not take the actual geographic distribution of alleles into account (an allele's presence in 2 of 10 populations can mean quite different things depending on whether the two populations are adjacent or dis-
BURTON: CONSEQUENCES
Table 3.
531
OF RESTRICTED GENE FLOW
Survivorship ofF2 larvae of within- and between-population
crosses to hyperosmotic stress Survivorship
Cross type
Within populations (PES x PES, AB x AB, etc.) Within regions (PES x SC, AB x U) Between regions (PES x U, U x SC)
N
(%)
626
77.3
716
68.6
406
53.2
R x C Test of Independence: G = 65.1. df = 2. P « 0.005.
junct), and (2) in averaging the frequencies of alleles with the same occupancy number (proportion of populations where the allele occurs), much valuable information may be lost. For example, a unique allele fixed in one population tells much about restriction of gene flow; yet when its frequency (1.0) is averaged with many rare unique alleles, its information content can be effectively lost. More recently, Slatkin (1985) has presented simulation results that suggest that the joint parameter Nm (effective population size multiplied by the proportion of immigrant individuals per generation) can be estimated from the average frequency of "private" (occurring in only one population) alleles. While overcoming problem (1) above, the estimate shows strong sample size dependence (Slatkin provides a sort of correction factor). More importantly, since we lack any information on effective population sizes among marine invertebrate species, it remains difficult to resolve actual migration (gene flow) rates. Given these reservations, Slatkin's methods should be combined with simple analyses of the raw data in order to provide a reasonable first approach for the analysis of gene flow based on allozyme data. Consequences of Restricted Gene Flow among T. californicus Populations: Microadaptation Among Neighboring Populations. - The widespread occurrence of unique alleles in high frequency among Tigriopus califarnicus populations indicates that gene flow in this species is highly restricted (Burton and Feldman, 1981; Burton, 1983). Both high predation pressure in the lower intertidal zone and the behavior of T. californicus probably contribute to this restriction of gene flow (Burton and Feldman, 1981). Consequently, the population structure of this species is conducive to the evolution of adaptations to local environmental conditions. Such adaptations could potentially involve the specific gene loci at which we have documented population differentiation or they may involve a diversity of structural and regulatory loci which we have not yet studied. The data presented above attempt to address both possibilities. In the former case, we know of only one locus for which an appropriate gene/environment adaptation argument may be constructed, i.e., Gpt (Burton and Feldman, 1983). Any number of adaptations might be used to address the latter; here we have focused on larval survivorship following hyperosmotic stress. The Gpt locus might participate in local adaptation in the following way: Burton and Feldman (1982b) demonstrated the rapid accumulation of intracellular free amino acids (FAA) in T. californicus subjected to hyperosmotic stress, with alanine playing an important role in the early stages of the response. Burton and Feldman (1983) showed that Gpt genotypes differ significantly in rates of alanine accumulation during stress response (Gpt catalyzes the final step of alanine synthesis) and larval survivorship following stress. While we lack information concerning
532
BULLETIN OF MARINE SCIENCE, VOL. 39, NO.2,
1986
other forces acting on the Gpt polymorphism, I propose that the higher activity GptF allele will be favored in environments where T. calijomicus experience more episodes of hyper osmotic stress. Under laboratory conditions of constant salinity (54, 34, or 170/00), I have observed no change in Gpt allele frequencies for periods of over 6 months (unpubl. data), suggesting that in the absence of salinity stress, directional selection on this polymorphism may be weak or nonexistent. Unfortunately, I have not yet been able to maintain large population sizes in a laboratory fluctuating salinity experiment to test the above hypothesis. As demonstrated in Figure I, there is no direct relationship between Gpt allele frequencies and pool salinities. Given that I now propose that the action of this locus is as described above, the salinity history of the pool and not its current salinity would clearly be a more appropriate environmental parameter. The relationship is further complicated by the population structure of this species because there is extensive gene flow among pools on a given rock outcrop (Burton and Swisher, 1984). Restriction of gene flow among these outcrops, stretching over approximately II km of coastline, is evident from the differentiation observed at the pgm locus (Burton and Feldman, 1981). Hence, the relevant parameters are
those quantifying salinity variation on an outcrop (in time and space) and the average allele frequency on that outcrop. While we do not yet have an extensive set of such data, Table 1 suggests that Gpt allele frequencies respond to the differing levels of salinity variation experienced by local populations of T. califomicus. Consequences of Restricted Gene Flow Among T. californicus Populations: Differentiation of the Genetic Basis of Physiological Response. -On a broad geographic scale, T. calijomicus populations are genetically isolated to the point where they are currently undergoing independent evolution. A predicted consequence of such independence is that the genetic basis of common physiological responses could differentiate over time since each population will experience different mutational input, genetic drift, and selective regimes. Hence, while all T, califomicus populations along the California coast will experience some degree of osmotic stress and have the ability to respond to that stress, the biochemical genetic processes underlying the physiological response may not be the same in each population. Such predictions arise from the work of King (1955) on integration of the gene pool in lines of Drosophila melanogaster independently selected for DDT resistance and other related investigations (Wallace, 1981). These studies indicate that physiological response (in this case DDT resistance) was built up "by the consolidation of polygenic systems which are not identical in independently developed lines and in which the constituent factors are not simply additive (King, 1955)." Such consolidation of polygenic systems would correspond to the establishment of new "harmonious interaction systems," the adaptive peaks of Wright's shifting balance theory of evolution (reviewed in Wright, 1977). When populations with different interaction systems are hybridized, adaptations will break down in the F2 generation as segregation breaks up the pleiotropic systems found intact in each parental population. This breakdown of adaptations is also thought to underlie the phenomenon of "optimal" outcrossing distances in plants (refs. in Willson, 1984); a reduction in female reproductive success is observed when pollen donors are from a relatively distant population ("outbreeding depression"). Extensive differentiation among T. calijomicus populations is apparent from the data presented in Table 2. Based on the four loci studied here as genetic markers, we can safely conclude that central and southern California populations are differentiated at many loci and that the population structure of this species is
BURTON: CONSEQUENCES
OF RESTRICTED GENE FLOW
533
highly subdivided. These conditions are clearly conducive for the establishment of different polygenic interaction systems (Wright, 1977). The data presented in Table 3 indicate an apparent "outbreeding depression" exists both within and among regional T. califomicus populations; I conclude that the reduced larval survivorship among F2 interpopulation hybrids is the result of breakdown of the successful interaction systems characterizing each of the parental lines resulting from meiotic segregation. It should be noted that this conclusion makes no assumptions concerning the involvement of specific loci in osmotic response or whether the differentiation observed at the four loci studied here is the result of selective forces or random drift. Evolutionary Consequences of Restricted Gene Flow. -My laboratory, like many others, has put considerable effort into studying the spatial distribution of allelic variants at enzyme coding gene loci among natural populations of marine organisms. The goal of this work is to understand the evolution of adaptations to local environmental conditions. There are two general ways in which electrophoretic data might contribute to this goal: (1) By indicating the extent of population differentiation, allozyme data can aid in determining the extent of gene flow among populations. Combined with other genetic data, such as heritable morphological variation (Struhsaker, 1968) and mitochondrial DNA analyses (Avise and Lansman, 1983), allozyme studies can suggest the appropriate spatial scale at which one should focus studies of adaptation. With respect to T. californicus electrophoretic surveys of natural populations and field transplantation experiments indicate that while gene flow is too extensive to expect adaptive differentiation among pools on a single outcrop, restriction of gene flow among neighboring outcrops appears to be sufficient to allow such differentiation. (2) Since some enzymes are known to function in specific, physiologically-relevant biochemical pathways, allelic frequencies in those enzyme systems might directly reflect population adaptations. Unfortunately, while suggestive patterns of differentiation at single gene loci have frequently been noted (i.e., allele frequencies are correlated with some environmental parameter), a causal relationship has seldom been established (Koehn et aI., 1983). In conclusion, surveys of allozyme variation have made invaluable contributions to our understanding of gene flow among marine invertebrate populations. In fact, the detailed studies needed to resolve the genetic structure of natural populations of marine invertebrates are still relatively few in number. However, such studies alone can only aid in demarcating the boundaries of natural populations. While this may be a valuable goal from a fisheries management point of view, it is rarely the ultimate goal of evolutionary ecology; here I have suggested some ways in which this work can serve as a first step in elucidating the evolution of physiological adaptations. ACKNOWLEDGMENTS This work was supported by a grant from the National Science Foundation (DEB-8207000). I thank P. Noga for laboratory assistance. LITERATURE CITED Avise, J. C. and R. A. Landsman. 1983. Polymorphism in mitochondrial DNA in populations of higher animals. Pages 147-164 in M. Nei and R. K. Koehn, eds. Evolution of genes and proteins. Sinauer, Sunderland, Massachusetts. Berger, E. 1973. Gene-enzyme variation in three sympatric species of Littorina. BioI. Bull. 145: 83-90.
534
BULLETIN OF MARINE SCIENCE. VOL. 39. NO.2.
1986
Berglund, A. and U. Lagercrantz. 1983. Genetic differentiation in populations of two Palaemon prawn species at the Atlantic east coast: does gene flow prevent local adaptation? Mar. BioI. 77: 49-57. Bulnhein, H.-P. and A. Scholl. 1981. Genetic variation between geographic populations of the amphipods Gammarus zaddaehi and G. sa/inus. Mar. BioI. 64: 105-115. Buroker, N. E. 1984. Gene flow in mainland and insular populations of Crassostrea (Mollusca). BioI. Bull. 166: 550-557. Burton, R. S. 1983. Protein polymorphisms and genetic differentiation of marine invertebrate populations. Mar. BioI. Letters 4: 193-206. --. 1985. Mating system of the intertidal copepod Tigriopus ealifornieus. Mar. BioI. 86: 247252. --and M. W. Feldman. 1981. Population genetics of Tigriopus ealifornieus: II. Differentiation among neighboring populations. Evolution 35: 1192-1205. --and ---. 1982a. Population genetics of coastal and estuarine invertebrates: does larval behavior influence population structure? Pages 537-551 in Y. S. Kennedy, ed. Estuarine comparisons. Academic Press, New York. --and ---. 1982b. Changes in free amino acid concentrations during osmotic response in the intertidal copepod Tigriopus ealifornieus. Compo Biochem. Physiol. 73A: 441-445. -and --. 1983. Physiological effects of an allozyme polymorphism: glutamate-pyruvate transaminase and the response to hyperosmotic stress in the copepod Tigriopus ealifornieus. Biochem. Genet. 21: 239-251. --and S. G. Swisher. 1984. Population structure of the intertidal copepod Tigriopus ealifornieus as revealed by field manipulation of allele frequencies. Oecologia 65: 108-111. ---, M. W. Feldman and J. W. Curtsinger. 1979. Population genetics of Tigriopus ealifornieus (Copepoda: Harpacticoida): I. Population structure along the central California coast. Mar. Ecol. Prog. Ser. I: 29-39. ---, --and S. G. Swisher. 1981. Linkage relationships among five enzyme-coding gene loci in the copepod Tigriopus ealifornieus: a genetic confirmation of achiasmatic meiosis. Biochem. Genet. 19: 1237-1245. Christiansen, F. B. and O. Frydenberg. 1974. Geographic patterns offour pOlymorphisms in Zoarees viviparous as evidence for selection. Genetics 77: 765-770. Crisp, D. J. 1978. Genetic consequences of different reproductive strategies in marine invertebrates. Pages 257-273 in B. Battaglia and J. A. Beardmore, eds. Marine organisms: genetics, ecology, and evolution. Plenum Press, New York. Gooch, J. L. 1975. Mechanisms ofevo]ution and population genetics. Pages 349-409 in O. Kinne, ed. Marine ecology, Yol. II, Part 1. Wiley, London. Janson, K. and R. D. Ward. 1984. Microgeographic variation in allozyme and shell characters in Littorina saxatilis Olivi (Prosobranchia: Littorinidae). BioI. J. Linn. Soc. 22: 289-307. Johnson, M. S. and R. Black. 1982. Chaotic genetic patchiness in an intertidal limpet, Siphonaria sp. Mar. BioI. 70: 157-164. King, J. C. 1955. Evidence for the integration of the gene pool from studies of DDT resistance in Drosophila. Cold Spring Harbor Symp. Quant. BioI. 20: 311-317. Koehn, R. K., R. I. E. Newell and F. Immermann. 1980. Maintenance of an aminopeptidase allele frequency cline by natural selection. Proc. Natl. Acad. Sci. U.S.A. 77: 5385-5389. --, A. J. Zera and J. G. Hall. 1983. Enzyme polymorphism and natural selection. Pages 115136 in M. Nei and R. K. Koehn, eds. Evolution of genes and proteins. Sinauer, Sunderland, Massachusetts. Levinton, J. S. 1980. Genetic divergence in estuaries. Pages 509-520 in Y. S. Kennedy, ed. Estuarine perspectives. Academic Press, New York. Marcus, N. H. 1977. Genetic variation within and between geographically separated populations of the sea urchin, Arbacia punetulata. BioI. Bull. 153: 560-576. Schopf, T. 1. M. and J. L. Gooch. 1971. Gene frequencies in a marine ectoproct: a cline in natura] populations related to sea temperature. Evolution 25: 286-289. Slatkin, M. ] 981. Estimating levels of gene flow in natural populations. Genetics 99:323-335. --. ]985. Rare alleles as indicators of gene flow. Evolution 39: 53-65. Struhsaker, J. W. 1968. Selection mechanisms associated with intraspecific shell variation in Littorina pieta (prosobranchia: Mesogastropoda). Evolution 22: 459-480. Theisen, B. F. 1978. Allozyme clines and evidence of strong selection at three loci in Mytilus edu/is (Bivalvia) from Danish waters. Ophelia 17: 135-142. Tracey, M. L., K. Nelson, D. Hedgecock, R. A. Shleser and M. L. Pressick. 1975. Biochemical genetics oflobsters: genetic variation and the structure of American lobster (Homarus amerieanus) populations. 1. Fish. Res. Board Can. 32: 2091-2101.
BURTON:CONSEQUENCES OF RESTRICTED GENEFLOW
535
Wallace, B. 1981. Basic population genetics. Columbia University Press, New York. 688 pp. Willson, M. F. 1984. Mating patterns in plants. Pages 261-276 in R. Dirzo and J. Sarukhan, eds. Perspectives on plant population ecology. Sinauer, Sunderland, Massachusetts. Wright, S. 1977. Evolution and the genetics of populations, Yol. 3: Experimental results and evolutionary deductions. University of Chicago Press, Chicago. 611 pp. DATEACCEPTED: February II, 1986. ADDRESS: Department of Biology, University of Pennsylvania. Philadelphia, Pennsylvania 19104.
