Genetic effects of germination timing and environment: an experimental investigation.
For species with seed dormancy, there is often considerable variation, both genetic and environmental, in the expression of dormancy. That is, within a population, some seeds may germinate, whereas other viable seeds experiencing the same environment remain dormant in the soil (e.g., Cavers 1974; Harper 1977; Westoby 1981; Silvertown 1984). As a result, there is considerable variation among seeds in the timing of germination as well as the environment in which germination and establishment occur. Understanding the causes of this variation is critical, because germination timing and response to the environment in which seeds emerge are of overriding importance for subsequent plant performance and success (Harper 1977, Weiner 1988). However, the proximate regulation of intraspecific variation in dormancy and germination is complicated by the mix of tissues contained within seeds, and by the often complex interactions among seed genotypes, maternal seed maturation conditions, and the abiotic germination environment (see reviews and discussion in Silvertown 1984; Roach and Wulff 1987; Baskin and Baskin 1989; Cabin 1995; Evans and Cabin 1995; Vleeshouwers et al. 1995).
Despite these complications, there is considerable evidence that seed dormancy and germination are under some degree of heritable genetic control (Jann and Amen 1977; Burass and Skinnes 1984; Bewley and Black 1985; Garbutt and Witcombe 1986; Rice 1989; Levin 1990; Meyer et al. 1995). Genetic variation in the timing and success of seed germination and establishment could have several important implications for the demographic and genetic structure of plant communities. First, seed dormancy could function as a type of sieve, screening when and where particular seed genotypes germinate and establish from the soil (Cabin 1996). This could in turn have profound consequences for the evolution of both germination and postgermination characteristics, if the germination environment is correlated with the adult environment and natural selection on adults differs among environments (Templeton and Levin 1979; Ritland 1983; Brown and Venable 1986; Klinkhamer et al. 1987; Evans and Cabin 1995). Second, differential germination and establishment of seed genotypes in response to spatial and temporal environmental variation could help explain the creation and maintenance of nonrandom patterns of genetic variation found in many mature plant populations (e.g., Antonovics 1971; Snaydon and Davies 1972, 1976; Turkington and Harper 1979; Schmitt and Antonovics 1986; Nevo 1988; Hamrick 1987, 1989). Despite these intriguing possibilities, few empirical studies have explored the extent to which seed genotypes may differentially respond to different germination conditions. This paucity of empirical investigations may stem from practical considerations, as it often is extremely difficult to distinguish differential germination from differential survival (see Discussion for further examination of this issue).
In this study, we experimentally investigate genetic effects of germination timing and environment in the desert mustard Lesquerella fendleri. We used a two-way factorial design to compare genotypes that germinated and established at different times and under initially different soil water levels in a greenhouse. Specifically, we asked whether there are genetic differences between plants derived from seeds that germinate and establish (1) relatively early versus relatively late, and (2) under relatively high versus relatively low initial soil water conditions.
Lesquerella fendleri (Brassicaceae) is a self-incompatible, short-lived perennial native to southwestern North America (Rollins and Shaw 1973). This species is well-suited for this study because previous work has shown that (1) there are significant genetic differences in the field between Lesquerella seeds that germinate and other viable seeds remaining dormant in the soil (Evans and Cabin 1995; Cabin 1996); (2) Lesquerella maintains a persistent seed bank, and the germination and establishment of seeds from the seed bank may be significantly affected by spatial and temporal variation in soil water availability (Evans and Cabin 1995); and (3) there are significant differences in the survival and morphology of plants derived from relatively early and relatively late germinating Lesquerella seeds (Cabin et al. in press). Lesquerella seeds used in this study were obtained from a bulk collection made in the summer of 1994 from over 50 individuals within a single population growing at the Sevilleta National Wildlife Refuge (NWR), 80 km south of Albuquerque in central New Mexico. Following collection, these seeds were stored in opaque envelopes at room temperature within the lab. In the field, Lesquerella germination occurs primarily between January and March; for this experiment, we germinated seeds in the greenhouse in early February 1995. Tetrazolium viability tests indicated that 98% of these seeds were viable.
The Experimental Setup
To assess the genetic effects of germination time and environment on Lesquerella plants, we used a two-way factorial design consisting of two germination times (early vs. late) and two water treatments (initially low water followed by high water vs. constantly high water levels). Previous trials revealed that germination is more rapid and more extensive under high versus low water conditions. Therefore, to insure adequate and similar sample sizes of Lesquerella plants in each of the four experimental combinations (high water-early germination, high water-late germination, initially low water-early germination, initially low water-late germination), we sowed seeds in an unbalanced design, with fewer seeds (400) in 40 high water-early pots and many more seeds (8000) in 160 high water-late pots (Table 1). However, since Lesquerella seeds germinate in roughly equivalent proportions in the early and late periods when the initial water level is low, we planted an equal number of seeds (1000) in 50 initially low water-early pots and in 50 initially low water-late pots. This design resulted in different sowing densities across treatments (Table 1). However, previous work demonstrated that Lesquerella does not experience density-dependent germination (analysis of variance [ANOVA] with four values of the main effect of experimental densities [5, 10, 25, and 50 seeds per pot, with nine pots per density], P [less than] 0.86), and survival rates in this experiment were not related to sowing densities (Table 1).
TABLE 1. Germination and survival, along with sample sizes at various stages of the experiment for the four treatments. Germination data include only seeds that emerged during the specified time period (early or late); see text for details. Germination data for all seeds are included in Figure 1. Data for average numbers of germinants and survivors per pot are calculated; we did not keep track of individuals in pots. Germination time Early Early Late Late Initial water treatment Low High Low High Number of pots 50 40 50 160 Number of seeds per pot 20 10 20 50 Total number of seeds sown 1000 400 1000 8000 Number of seedlings emerging 290 168 230 108 Percent germination 29 42 23 2 Average number of germinants per pot 5.8 4.2 4.6 0.7 Number of seedlings surviving 175 148 130 62 Percent survival 60 88 57 57 Average number of survivors per pot 3.5 3.7 2.6 0.4
To effect water treatments, pots were placed into troughs that received different amounts of water. First, we carefully filled 300 8-cm x 8-cm x 7-cm deep plastic pots with equivalent amounts of sand, bottom-watered until the sand was completely saturated, and allowed the pots to drain. We placed ten pots into each of 30 48-cm x 24-cm x 10-cm deep plastic troughs in a greenhouse receiving ambient light (ca. 10.5 hr/day) and 21 [degrees] day and 10 [degrees] C night temperatures. We randomly designated 20 troughs (with 200 pots) as high water troughs and filled them with 1500 mL of water, and designated the remaining 10 troughs (with 100 pots) as initially low water troughs and filled them with 500 mL of water. Troughs were randomly positioned in the greenhouse. These water treatments were designed to span the range of soil water availability observed at the Sevilleta NWR; the high water troughs resulted in moist but not saturated conditions on the soil surface, and the low water troughs simulated the dry desert soils that Lesquerella seeds frequently encounter during germination and establishment (mean [+ or -] SE% soil water = 13.1 [+ or -] 0.13 for high water, 3.7 [+ or -] 0.35 for low water, n = 6 soil samples for each water level). For the first two weeks of the experiment (the "early" germination period), we maintained these contrasting water levels by adding three times as much water to the high water troughs relative to the low water troughs at roughly three-day intervals. Although it does not mimic rainfall, bottom-watering facilitates the maintenance of distinct water regimes without disturbing seeds on the soil surface.
To separate early versus late germination groups, we further manipulated water availability. Previous trials had shown that very few additional seeds germinated under extended dry soil conditions. Therefore, to enhance germination of the initially low water pots during the "late" germination period, after two weeks we added water to the low water troughs so that soil moisture was equivalent to that in the high water troughs. An alternative treatment would be to apply the germination hormone gibberellic acid (GA), but we avoided this because susceptibility to GA varies among families in Lesquerella (Evans et al. 1996). Throughout the remainder of the experiment, water was added equally to all troughs as needed to maintain this relatively high level of moisture on the soil surface. Thus the "initially low water-late" plants are actually derived from seeds that did not germinate in the first 14 days under low water conditions, but eventually germinated under the high water conditions experienced 14 days after planting.
We defined early germinating seeds a priori as those that initiated germination on or before 14 days from planting, and late germinating seeds as those that initiated germination after 18 days from planting. To clearly separate these periods, all seeds that germinated between 14 and 18 days after planting were removed and discarded [ILLUSTRATION FOR FIGURE 1 OMITTED]. To effect these treatments, all newly germinating Lesquerella seedlings that emerged after 14 days from planting within the early pots were carefully removed and discarded as they emerged. Conversely, all seedlings that germinated on or before 18 days after planting were removed from the late pots. To quantify germination for each of the four experimental treatments, five additional pots per treatment were sown with the exact number of assigned seeds (10, 20, or 50) and censused daily throughout the experiment.
Data Collection and Analysis
Genotypes of Lesquerella plants that established under the four experimental conditions were characterized by employing starch gel electrophoresis, using five enzymes and seven polymorphic loci. All electrophoresis was performed using methods and loci described in Cabin (1996). In the high water troughs, all surviving plants in the late pots (62) were collected and electrophoresed, while 100 of the 148 surviving plants in the early pots were randomly selected for electrophoresis. In the initially low water troughs, 81 plants were randomly selected from both the early and late pots (175 and 130 plants survived in these early and late pots, respectively; see Table 1).
Loglinear analyses (Fienberg 1980; Caswell 1989; Weir 1990; Horvitz and Schemske 1994) were used to compare allele frequencies and heterozygosity among surviving plants in the four experimental treatments. To compare allele frequencies, we arranged the electrophoretic data from each locus into a three-way contingency table consisting of water level (W), time of germination (T), and the number of copies of each allele (A) in the sample for the locus in question. Rare alleles were combined when necessary to meet the assumptions of the log-likelihood chi-square statistic (Fienberg 1980). We follow the conventional notation of hierarchical models as illustrated in Caswell (1989) and Horvitz and Schemske (1994). In this notation, the presence of an interaction implies that all terms containing that interaction or lower order interactions are included in the model (Fienberg 1980). Thus, the model WT, A is equivalent to the model W, T, A, WT. In our experiment, the initial water level and timing of germination are fixed explanatory factors, and thus the appropriate null model is WT, A (Caswell 1989; Horvitz and Schemske 1994). This model implies that allele frequencies of Lesquerella plants are independent of water, germination time, and their interaction. The model is statistically evaluated as outlined below.
In loglinear analysis, the goodness of fit of the expected cell frequencies predicted by each model is compared with the observed cell frequencies by means of a marginal and/or conditional log-likelihood chi-square ([G.sup.2]) statistic. Since both our marginal and conditional [G.sup.2] analyses yielded very similar results, here we report only the results of the slightly more conservative marginal tests. To assess the significance of any particular effect, the [G.sup.2] of a model containing this effect is compared against another model that is identical except that it does not contain this effect (Caswell 1989). In these analyses, a significant null model does not negate the potential importance of the experimental variables, since what is of interest is whether inclusion of these additional effects significantly improves the model's goodness of fit.
In addition to the allele frequency analyses, we also performed two parallel analyses that examined the effect of water and germination timing on the heterozygosity of experimental Lesquerella plants. These analyses were identical to the allele frequency analyses except that the allele term was replaced by single-locus heterozygosity for the first analysis (the number of individuals heterozygous and homozygous at each locus), and multi-locus heterozygosity for the second analysis (the number of loci heterozygous per individual plant). For the multi-locus analysis, we only used individuals with scores at all seven loci assayed. The log-likelihood statistics for all three analyses were computed using the LOGLIN option of the CATMOD procedure of SAS (1989).
[TABULAR DATA FOR TABLE 2 OMITTED]
Germination and Survival
As expected, Lesquerella seeds in the high water troughs germinated much more rapidly, and in greater numbers, than did seeds in the low water troughs ([ILLUSTRATION FOR FIGURE 1 OMITTED], Table 1). By the last day of the early germination period, 14 days after planting, over 50% of the seeds in the high water troughs had germinated, compared with 21% of the low water seeds, a difference that was highly significant ([[Chi].sup.2] = 21.94, df = 1, P [less than] 0.001). Beyond 14 days from planting, both sets of troughs were maintained at high water levels. Under these conditions, only 16% of the remaining ungerminated seeds in the high water troughs germinated, compared with 33% of the remaining seeds in the initially low water troughs; this difference was also highly significant ([[Chi].sup.2] = 32.62, df = 1, P [less than] 0.001). In addition, seeds sown in the initially low water troughs continued to emerge as late as 33 days from planting, while germination in the high water troughs ceased after 25 days from planting. Survival to the end of the experiment for the high water-early plants was nearly 90%, while the survival of plants in the other three experimental categories ranged from 57% to 60% (Table 1).
Allele frequencies (Table 2) of surviving Lesquerella plants differed significantly among water treatments and germination times (Table 3). The overall [G.sup.2] statistics (summed over all seven loci assayed) were significant for both these main effects, although the effect of water was somewhat stronger. However, the overall joint effect of water and germination time on allele frequencies was not significant, indicating that the effect of each factor on allele frequencies was largely independent of the other. At the individual locus level, both water and germination time significantly affected allele frequencies at two of the seven loci, and there was also a significant interaction effect at one locus (LAP).
There was a highly significant overall effect of water on single locus heterozygosity, but the effects of germination time and the water by germination time interaction were not statistically significant (Table 4). At all of the electrophoretic loci except PGM1, the observed heterozygosity was greater among the initially low versus high water plants and among the early versus late germinating plants. For water, these single locus differences were statistically significant at two loci, whereas the effect of both germination time and the water by germination time interaction were each significant at only one locus (Table 4). The observed heterozygosity of plants produced by seeds in the initially low water treatment also tended to be slightly (but not significantly) higher than that expected under Hardy-Weinberg equilibrium conditions (Table 2).
In the multilocus examination of heterozygosity, there were [TABULAR DATA FOR TABLE 3 OMITTED] no statistically significant effects of water, germination time, or their interaction, indicating that the number of loci heterozygous within individual plants was independent of these variables (Table 4). However, the failure to detect statistical differences in these multilocus analyses may be partially attributable to the relatively low sample sizes of surviving plants for which we obtained scores at all seven loci (n = 78, 22, 66, and 30 plants for the high water-early, high waterlate, initially low water-early, and initially low water-late categories, respectively). Inspection of the general patterns of heterozygosity suggests a weak but consistent response to the water treatments. Almost 80% of plants in the initially low water treatment had at least three of seven loci heterozygous, [TABULAR DATA FOR TABLE 4 OMITTED] compared with only about 70% of plants in the constantly high water treatment.
We found significant genetic differences among Lesquerella plants that germinated and survived under the four experimental combinations of germination time and initial soil water availability. These results suggest that (1) the genetic constitution of seeds may play a significant role in regulating the timing and success of germination and establishment, and (2) seed genotypes may differentially respond to environmental variation experienced during the germination and establishment periods.
Before discussing the implications of genetic effects of germination timing and environment, it is important to recognize that differential survival as well as germination may have affected the results. Most of the seeds exposed to high water conditions in the first two weeks of the experiment germinated during this "early" period, while very few of the seeds that failed to emerge under these conditions germinated over the following four weeks in the "late" germination period. In contrast, seeds experiencing low water conditions for the first two weeks after planting but high water conditions for the remainder of the experiment germinated in roughly equivalent proportions in the early and late periods. While the survival of emerging seedlings in the high water-early category was nearly 90%, the survival of seedlings in the other three categories was only around 60%. Since we obtained no genetic data from the seedlings that did not survive to the end of the experiment, it is possible that some portion of the observed genetic differences among the surviving plants was caused by nonrandom seedling survival. Evaluating this possibility is to some extent a technical issue, since seedlings must grow large enough to electrophorese, and Lesquerella, like many plants, experiences damping off and/or early mortality. based on the significant results found in this study, one of us (ASE) is now working on distinguishing germination versus establishment effects.
We examined two aspects of genetic variation among treatment groups: allele frequencies and heterozygosity. Overall allele frequency differences were based on data summed over all seven loci. Both initial water availability and time of germination significantly affected overall allele frequencies of surviving Lesquerella plants, although the interaction of these two effects was not statistically significant. While some studies suggest that different germination environments may favor particular alleles or combinations of alleles (e.g., Zangerl and Bazzaz 1984a, b; Mitton et al. 1989), our data show no strong or readily interpretable allele-specific patterns.
Only soil water availability significantly affected the heterozygosity of surviving Lesquerella plants. Seeds that germinated under initially low soil water conditions produced plants that had significantly greater single-locus heterozygosity than did seeds that germinated under constantly high water environments. Many studies have shown that overall allozyme heterozygosity is correlated with various ecological and environmental variables, with more heterozygous individuals often showing higher levels of relative vigor and fitness (reviewed by Mitton 1989, 1994). Indeed, in our study, the seeds that germinated quickly under the presumably more stressful condition (initially low water-early germination) also contained on average the most heterozygous genotypes. Other studies have also found that more heterozygous seeds may differentially persist, germinate, and produce more vigorous plants (e.g., Kalisz 1989; Cabin 1996; see also data and discussion in Hamrick 1989). Thus, it is conceivable that the observed heterozygosity differences among the Lesquerella plants in this study at least partially contributed to the significant allelic differentiation among the four experimental groups.
A final statistical issue worth mentioning is that loglinear analyses assume that loci are independent, and thus significant linkage disequilibrium could potentionally over-estimate the importance of treatment effects. We tested this assumption using the permutation algorithm of P. O. Lewis and D. Zaykin (Zaykin et al. 1995). After correction for departure from Hardy-Weinberg equilibrium, only three of 21 two-locus combinations showed significant linkage disequilibrium, and none of these combinations were significant after Bonferroni correction, indicating that these data conform fairly well to the assumption of independence among loci.
What are the evolutionary consequences of variation in germination response for the genetic constitution of plant populations? Lesquerella plants in this study showed pronounced genetic differentiation based on the time and environment of germination, which suggests that variation in germination response may be an important determinant of the genetic structure of aboveground populations. In other words, variable germination behavior that is related to the genetic constitution of seeds may function as a type of genetic sieve. However, in this and other experiments, it remains unclear to what extent observed genetic differences in seedling success reflect differences within versus among maternal sibships. This issue is important because in the former case, natural selection would be strongest among siblings, while in the latter case selection would primarily be among maternal genotypes (Schmitt and Antonovics 1986).
Evidence is beginning to accumulate that variable germination behavior may allow seed banks to function as sieves. For example, Kalisz (1991) found that the probability of Collinsia verna seeds persisting and emerging from the soil seed bank was significantly affected by spatial and temporal environmental variation, and that seedlings emerged significantly later out of experimental seed banks containing older seeds than adjacent unmanipulated seed banks. Cabin et al. (in press) found that Lesquerella plants that differed in germination behavior (one group of seeds naturally germinated, the other remained dormant under the same environmental conditions, but was forced to germinate by GA application) responded differently to microenvironmental variation in the field. These studies suggest that germination behavior may be influenced by a complex suite of interactions among seed characteristics and environmental variables. Our results raise the possibility that one important seed characteristic may be genetic constitution.
One possible consequence of a sieve phenomenon is that the existence of dormancy and differences among genotypes in germination timing and environment may set the stage for subsequent evolution. If the seeds that actually germinate in any one time and place are a nonrandom subset of all available seed genotypes, then such germination behavior could limit the range of genotypes available for selection to act on. This in turn could set the stage for the joint evolution of germination requirements and adult characteristics (Evans and Cabin 1995). Some support for this idea comes from the work of Bennington et al. (1991), who found that clonally replicated genotypes (to minimize maternal/environmental effects) originating from young and old buried seed populations of Luzula parviplora in Alaska showed persistent phenotypic differences in both common and multiple-environment experiments (see also McGraw 1993). The myriad ways in which seeds have evolved to track and predict the abiotic environment (reviewed in Mayer and Poljakoff-Mayber 1975; Baskin and Baskin 1989) may have resulted in seeds germinating in particular points in time and space that maximize the chance of successful establishment and reproduction (Schupp 1995). Unfortunately, few other studies have empirically examined how environmental variation may affect the germination timing and success of different seed genotypes contained within soil seed banks; it also remains unclear if and to what extent such responses may be adaptive.
Understanding the nonrandom and often intriguing patterns of genetic variation in nature remains a primary goal of evolutionary biology. Although the processes of seed germination and establishment are of obvious demographic importance to plant communities, few studies have examined the genetic consequences of selection acting at this critical life-history stage. Two empirical studies (Tonsor et al. 1993; Cabin 1996) have shown that natural seed bank and surface plant populations may differ genetically. To our knowledge, however, this is the first study to experimentally demonstrate that the germination environment and the timing of germination can significantly affect the genetic structure of emerging plant populations. In nature, microsite differences and spatial and temporal environmental variation interact to produce a mosaic of germination and growth environments much more complex than the simple variation generated in this greenhouse study. We believe that differential response of seed genotypes to this kind of natural variation likely plays a major yet largely overlooked role in generating and maintaining the genetic structure of many plant populations.
We thank J. Heywood, T. Lowrey, D. Marshall, T. Markow, J. Schmitt, S. Tonsor, and M. Price for helpful comments and valuable suggestions. P. Lewis kindly offered expert genetic statistical advice and software, and J. Avritt provided considerable help in the greenhouse and lab. This research was completed in partial fulfillment of a Ph.D. degree to RJC at the University of New Mexico, and was supported by National Science Foundation Grant DEB-9318433 to ASE.
ANTONOVICS, J. 1971. The effects of a heterogenous environment on the genetics of natural populations. Am. Sci. 59:593-599.
BAKER, H. G. 1989. Some aspects of the natural history of seed banks. Pp. 9-21 in M. A. Leck, V. T. Parker, and R. L. Simpson, eds. Ecology of soil seed banks. Academic Press, San Diego, CA.
BASKIN, J. M. AND C. C. BASKIN. 1989. Physiology of dormancy and germination in relation to seed bank ecology. Pp. 53-66 in M. A. Leck, V. T. Parker, and R. L. Simpson, eds. Ecology of soil seed banks. Academic Press, San Diego, CA.
BENNINGTON, C. C., J. B. McGRAW, AND M. C. VAVREK. 1991. Ecological genetic variation in seed banks. II. Phenotypic and genetic differences between young and old subpopulations of Luzula parviiflora. J. Ecol. 79:627-643.
BEWLEY, J. D., AND M. BLACK. 1985. Seeds: physiology of development and germination. Plenum, New York.
BROWN, J. S., AND D. L. VENABLE. 1986. Evolutionary ecology of seed-bank annuals in temporally varying environments. Am. Nat. 127:31-47.
BURASS, T., AND H. SKINNES. 1984. Genetic investigations on seed dormancy in barley. Hereditas 101:235-244.
CABIN, R. J. 1995. An examination of the ecological and evolutionary relationship between the seed bank and the surface plant population of the desert mustard, Lesquerella fendleri. Ph.D. diss. Univ. of New Mexico, Albuquerque.
-----. 1996. Genetic comparisons of seed bank and seedling populations of the desert mustard Lesquerella fendleri. Evolution 50:1830-1841.
CABIN, R. J., A. S. EVANS, AND R. J. MITCHELL. 1997. Do plants derived from seeds that readily germinate differ from plants derived from seeds that require forcing to germinate? A case study of the desert mustard Lesquerella Fendleri. Am. Midl. Nat. 138:121-133.
CASWELL, H. 1989. Matrix population models. Sinauer, Sunderland, MA.
CAVERS, P. B. 1974. Germination polymorphism in Rumex crispus. The effects of different storage conditions on germination responses of seeds collected from individual plants. Can. J. Bot. 52:575-583.
EVANS, A. S., AND R. J. CABIN. 1995. Can dormancy affect the evolution of post-germination traits? An assessment of the desert mustard Lesquerella fendleri. Ecology 76:344-356.
EVANS, A. S., R. J. MITCHELL, AND R. J. CABIN. 1996. Morphological side effects of using gibberellic acid to induce germination: consequences for the study of seed dormancy. Am. J. Bot. 83:543-549.
FIENBERG, S. E. 1980. The analysis of cross-classified categorical data. Massachusetts Institute of Technology, Cambridge.
GARBUTT, K., AND J. R. WITCOMBE. 1986. The inheritance of seed dormancy in Sinapis avensis L. Heredity 56:25-31.
HAMRICK, J. L. 1987. Gene flow and distribution of genetic variation in plant populations. Pp. 53-68 in K. M. Urbanska, ed. Differentiation patterns in higher plants. Academic Press, London.
-----. 1989. Isozymes and the analysis of genetic structure in plant populations. Pp. 87-105 in D. Soltis and P. Soltis, eds. Isozymes in plant biology. Dioscorides Press, Portland, OR.
HARPER, J. L. 1977. Population biology of plants. Academic Press, New York.
HORVITZ, C. C., AND D. W. SCHEMSKE. 1994. Effects of dispersers, gaps, and predators on dormancy and seedling emergence in a tropical herb. Ecology 75:1949-1958.
JANN, R. C., AND R. D. AMEN. 1977. What is germination? Pp. 7-28 in A. A. Kahn, ed. The physiology and biochemistry of seed dormancy and germination. North Holland Publishing, Amsterdam, The Netherlands.
KALISZ, S. 1989. Fitness consequences of mating system, seed weight and emergence date in awinter annual, Collinsia verna. Evolution 43:1263-1272.
-----. 1991. Experimental determination of seed bank age structure in the winter annual Collinsia verna. Ecology 72:575-585.
KLINKHAMER, P. G. L., T. J. DEJONG, J. A. J. METZ, AND J. VAL. 1987. Life history tactics of annual organisms: the joint effect of dispersal and delayed germination. Theor. Popul. Biol. 32: 127-156.
LECK, M. A., V. T. PARKER, AND R. C. SIMPSON, EDS. 1989. Ecology of soil seed banks. Academic Press, San Diego, CA.
LEVIN, D. A. 1990. The seed bank as a source of genetic novelty in plants. Am. Nat. 135:563-572.
MAYER, A. M., AND A. POLJAKOFF-MAYBER. 1975. The germination of seeds. 3d ed. Pergamon, London.
McGRAW, J. B. 1993. Ecological genetic variation in seed banks. IV. Differentiation of extant and seed bank-derived populations of Eriophorum vaginatum. Arct. Alp. Res. 25:45-49.
MEYER, S. E., S. G. KITCHEN, AND S. L. CARSON. 1995. Seed germination timing patterns in intermountain Penstemon (Scrophulariaceae). Am. J. Bot. 82:377-389.
MITTON, J. B. 1989. Physiological and demographic variation associated with allozyme variation. Pp. 127-145 in D. Soltis and P. Soltis, eds. Isozymes in plant biology. Dioscorides Press, Portland, OR.
-----. 1994. Molecular approaches to population biology. Annu. Rev. Ecol. Syst. 25:45-69.
MITTON, J. B., H. P. STUTZ, W. S. SCHUSTER, AND K. L. SHEA. 1989. Genotypic differentiation at PGM in Engelmann spruce from wet and dry sites. Silvae Genet. 38:217-221.
MURDOCH, A. J., AND R. H. ELLIS. 1992. Longevity, viability and dormancy. Pp. 193-229 in M. Fenner, ed. The ecology of regeneration in plant communities. CAB International, Wallingford, U.K.
NEVO, E. A. 1988. Genetic diversity in nature: patterns and theory. Evol. Biol. 23:217-246.
PRIESTLY, D. A. 1986. Seed aging: implications for seed storage and persistence in the soil. Cornell Univ. Press, Ithaca, NY.
RICE, K. J. 1989. Impacts of seed banks on grassland community structure and population dynamics. Pp. 212-230 in M. A. Leck, V. T Parker, and R. L. Simpson, eds. Ecology of soil seed banks. Academic Press, San Diego, CA.
RITLAND, K. 1983. The joint evolution of seed dormancy and flowering time in annual plants living in variable environments. Theor. Popul. Biol. 24:213-243.
ROACH, D. A., AND R. D. WULFF. 1987. Maternal effects in plants. Annu. Rev. Ecol. Syst. 18:209-235.
ROLLINS, R. C., AND E. A. SHAW. 1973. The genus Lesquerella (Cruciferae) in North America. Harvard Univ. Press, Cambridge, MA.
SAS INSTITUTE. 1989. SAS/Stat user's guide. Vers. 6. 4th ed. Statistical Analysis Systems Institute, Cary, NC.
SCHMITT, J., AND J. ANTONOVICS. 1986. Experimental studies on the evolutionary significance of sexual reproduction. III. Maternal and paternal effects during seedling establishment. Evolution 40:817-829.
SCHUPP, E. W. 1995. Seed-seedling conflicts, habitat choice, and patterns of plant recruitment. Am. J. Bot. 82:399-409.
SILVERTOWN, J. W. 1984. Phenotypic variety in seed germination behavior: The ontogeny and evolution of somatic polymorphism in seeds. Am. Nat. 124:1-16.
SNAYDON, R. W., AND M. S. DAVIES. 1972. Rapid population differentiation in a mosaic environment. II. Morphological variation in Anthoxanthum odoratum. Evolution 26:390-405.
-----. 1976. Rapid population differentiation in a mosaic environment. IV. Populations of Anthoxanthum odoratum at sharp boundaries. Heredity 37:9-25.
TEMPLETON, A. R., AND D. A. LEVIN. 1979. Evolutionary consequences of seed pools. Am. Nat. 114:232-249.
THOMPSON, K. 1992. The functional ecology of seed banks. Pp. 231-258 in M. Fenner, ed. Seeds: the ecology of regeneration in plant communities. C.A.B International, Wallingford, U.K.
TONSOR, S. J., S. KALISZ, AND J. FISHER. 1993. A life-history based study of population genetic structure: seed bank to adults in Plantago lanceolata. Evolution 47:833-843.
TURKINGTON, R., AND J. L HARPER. 1979. The growth, distribution, and neighbor relationships of Trifolium repens in a permanent pasture. IV. Fine-scale biotic differentiation. J. Ecol. 67:245-254.
VLEESHOUWERS, L. M., H. J. BOUWMEESTER, AND C. M. KARSSEN. 1995. Redefining seed dormancy: an attempt to integrate physiology and ecology. J. Ecol. 83:1031-1037.
WEINER, J. 1988. Variation in the performance of individuals in plant populations. Pp. 59-81 in A. J. Davy, M. J. Hutchings, and A. R. Watkinson, eds. Plant population ecology. Blackwell, Oxford.
WEIR, B. S. 1990. Genetic data analysis. Sinauer, Sunderland, MA.
WESTOBY, M. 1981. How diversified seed germination behavior is selected. Am. Nat. 118:882-885.
ZANGERL, A. R., AND F. A. BAZZAZ. 1984a. Effects of short-term selection along environmental gradients on variation in populations of Amaranthus retroflexus and Abutilon theophrasti. Ecology 65:207-217.
-----. 1984b. Niche partitioning between two phosphoglucoisomerase genotypes in Amaranthus retroflexus. Ecology 65:218-222.
ZAYKIN, D., L. ZHIVOTOVSKY, AND B. S. WEIR. 1995. Exact tests for association between alleles at arbitrary numbers of loci. Genetica 96:169-178.