Printer Friendly

The effect of sperm competition on male gain curves in a colonial marine invertebrate.

INTRODUCTION

Hermaphroditic plants and animals can potentially maximize reproductive success through a wide variety of different strategies. Since reproduction occurs via both genders, the total reproductive success of an individual must be evaluated as the sum of the reproductive success obtained through male and female function (Charnov et al. 1976, Maynard Smith 1978). Understanding how resources should be allocated to male vs. female gamete production in order to maximize total reproductive success in different ecological situations is one of the fundamental problems of sex allocation theory (Charnov 1982).

Most models of sex allocation are based on the concept of male and female gain curves (the relationship between relative investment in either male or female gamete production and resulting reproductive success; Charnov 1979, 1982). Evaluating how total reproductive success will be affected by gender-specific allocation patterns requires determining the shape of these two relationships. Female gain curves are thought to be linear in most hermaphrodite taxa and so have received relatively little empirical attention (Brunet 1992). However, this relationship may be nonlinear if physiological limitations constrain the number of offspring that can be brooded (Heath 1979, Charnov 1982, Strathmann et al. 1984) or if the philopatric dispersal of offspring leads to sibling competition (Maynard Smith 1978).

In contrast, male gain curves are expected to vary as a function of a wide variety of ecological conditions, including intensity of sperm competition (Fischer 1981, Lloyd 1984, Petersen 1991), size of local mating group (Charnov 1982, Raimondi and Martin 1991), rate of self fertilization (Charlesworth and Charlesworth 1981), and pollinator behavior (Harder and Thomson 1989). In particular, male gain curves are expected to be linear when males compete for access to ova (sperm competition), but saturate (plateau) in the absence of sperm competition. In the absence of alternative sperm sources, a male may be able to fertilize most of the ova in his vicinity with relatively few sperm. The production of additional male gametes would then result in diminishing gains in fertilization success via male function. However, when males compete for access to female gametes, additional male gamete production should result in continued increases in male fertilization success, especially if a male's fertilization success is proportional to the concentration of his gametes relative to that of other males (Yund and McCartney 1994).

Predictions about variation in the shape of male gain curves have stimulated much recent empirical work. Several studies have employed genetic markers to directly assay paternity and estimate male gain curves in a variety of terrestrial plants (Schoen and Stewart 1986, Broyles and Wyatt 1990, Devlin et al. 1992) and a marine invertebrate (McCartney 1997). Others have evaluated male gain curves via indirect measures of male fertilization success (e.g., pollen removal or deposition patterns [Thomson and Thomson 1989, Young and Stanton 1990] but see Brunet 1992, Snow and Lewis 1993 for drawbacks to this approach). In lieu of assaying actual gain curves, other authors have instead tested for expected comparative allocation patterns under linear vs. saturating male gain curves (fish: Fischer 1981, Petersen 1991; plants: Burd and Allen 1988; invertebrates: Sella 1990, Raimondi and Martin 1991).

Most of these studies have implicitly assumed that male gain curves and allocation patterns are relatively invariant within a species and largely determined by reproductive systems (but see Raimondi and Martin 1991). Some studies have considered the possibility of temporal variation in male gain curves and the consequences of different assay methods (Young and Stanton 1990, Devlin et al. 1992). However, there have been no published attempts to directly test hypotheses about ecological factors that may alter the shape of male gain curves by manipulating conditions in experimental populations. Such experiments might be difficult in terrestrial plant or marine fish systems (which have received the most empirical attention) due to the problems inherent in manipulating wind, insect pollinators, and fish behavior. However, sessile, colonial marine invertebrates that brood eggs and release sperm into the water column possess several attributes that facilitate an experimental approach. First, sperm of these and other free-spawning marine invertebrates are generally rapidly diluted in the water column (Pennington 1985, Yund 1990, Levitan 1991, Brazeau and Lasker 1992, Oliver and Babcock 1992, Benzie et al. 1994; but see Thomas 1994a, b), so fertilization is only possible if males and females are in relatively close proximity. This facilitates the assembly of experimental populations that are largely isolated from natural populations. Second, closely spaced male-phase colonies of some taxa are known to compete for fertilizations, and the intensity of competition can be manipulated by altering the number and sperm production levels of males present (Yund and McCartney 1994, Yund 1995, Atkinson and Yund 1996). Thirdly, embryos brooded by maternal colonies can easily be extracted for paternity analysis (Grosberg 1991, Yund and McCartney 1994). In small experimental populations, allozymes provide sufficiently diverse genetic markers to permit paternity analysis. Finally, extensive genetically based variation in sperm production occurs in some taxa (Hughes and Hughes 1986, Hughes 1989, McCartney 1997, Yund et al. 1997).

In this paper I explore the effect of the intensity of sperm competition on male gain curves in the colonial ascidian Botryllus schlosseri, a cyclical hermaphrodite. I present results from in situ fertilization studies with experimental populations in which the intensity of sperm competition was manipulated via three treatments. Allozyme markers were utilized for paternity assignment. The male gain curve in this species saturated at low sperm production levels in the absence of nearby competitors for fertilizations, but became progressively more linear with increasing intensity of sperm competition. Increased sperm competition also reduced male fertilization success at all sperm production levels. These results have implications for the evolution of hermaphroditism, the evolution of allocation patterns within hermaphrodite taxa, and the evolution of high levels of sperm production in free-spawning marine invertebrates in general.

MATERIALS AND METHODS

Study organism

Botryllus schlosseri is a colonial ascidian with a cosmopolitan distribution that is common on firm substrata in New England coastal waters (Gosner 1971). Colonies are composed of numerous asexually produced zooids organized into systems (typically 7-10 zooids/system), with up to 100 systems comprising a colony. Each zooid possesses its own inhalant (oral) siphon and all zooids in a system share a common exhalant (atrial) siphon. Colonies are hermaphrodites that exhibit an unusual reproductive system in which male and female phases alternate in a cycle. This sexual cycle is linked to an asexual zooid replacement cycle that occurs synchronously throughout a colony. In the asexual cycle, a new generation of zooids (buds) grows between the existing generation of zooids. After [approximately]7-10 d of growth, the buds are fully formed and take over the function of the old zooids, which are quickly resorbed (Milkman 1967). Eggs brooded by the new asexual generation are fertilized at the time of takeover, when the siphons of the new zooids open and water enters the atrial chamber. Zooids do not commence releasing sperm until [approximately]1-2 d after fertilization, effectively preventing self fertilization (Yund and McCartney 1994). Sperm are released over a period of several days (Stewart-Savage and Yund 1997), so colonies are functionally male through most of the reproductive cycle. Fertilized eggs develop into larvae that are brooded until release just before the end of the asexual cycle. The stages of this asexual/sexual cycle are easily identified in laboratory-reared colonies, and colonies can be selected for deployment in field experiments when they are in either male or female phase. Henceforth, I will use dioecious gender assignments to indicate the functional gender of a colony at the time of an experiment or fertilization event.

Male-phase B. schlosseri colonies of approximately equal sperm production are known to compete for fertilizations when they are closely spaced and near a female, but not when they are separated by greater distances (Yund and McCartney 1994). When males are farther from a female, sperm are in limiting supply and the addition of extra males simply results in the fertilization of more eggs. The intensity of sperm competition in this species can thus be controlled by altering the number and sperm production of local males.

Variation in sperm production

Botryllus schlosseri colonies exhibit substantial variation in life history strategy. Although most work to date has focused on variation in egg production and colony growth rates (Grosberg 1988, Chadwick-Furman and Weissman 1995), colonies also exhibit substantial variation in sperm production. Total sperm production varies among colonies for two reasons. First, sexually mature colonies vary in sperm production due to variation in the number of sperm produced within each testis (each zooid in a sexually mature colony contains two testes). Variation in testis-level sperm production is readily apparent as variation in testis cross-sectional area (visible from the underside if colonies are reared on glass substrata in the laboratory), which shows a linear correlation with actual sperm counts (r = 0.71; Yund et al. 1997). Testes within a colony do exhibit some variation in size. One testis in each zooid is typically larger than the other (Sabbadin 1958, Mukai and Watanabe 1976), and zooids with larger testes are often flanked by neighbors that have smaller testes (P. O. Yund, personal observation). Nevertheless, average testis cross-sectional area exhibits almost four-fold variation among genetically distinct colonies (Yund et al. 1997).

Colonies also vary in total sperm production because they vary in size (number of zooids), which in turn varies during colony ontogeny. However, unlike most colonial invertebrates, B. schlosseri does not exhibit indeterminate growth. Instead, colonies reduce growth as they approach a terminal size, which is then maintained over subsequent asexual zooid replacement cycles (Boyd et al. 1986, Grosberg 1988). Clonal replicates grown in a common-garden experiment exhibited more than four-fold variation in terminal size among genotypes (Yund et al. 1997). Combined variation in terminal size and testis size results in [approximately]1 order of magnitude variation in total sperm production among terminal sized colonies. Both terminal size and testis size have fairly high clonal repeatabilities (a form of broad sense heritability [Falconer 1981]) of 0.42 and 0.57, respectively (Yund et al. 1997), suggesting that a substantial proportion of the variation in sperm production among colonies is genetically based. However, breeding experiments to establish the transmission genetics of testis size have not yet been conducted. Natural populations composed of a mix of juvenile and terminal-size colonies are likely to exhibit variation in total sperm production that spans [greater than]2 orders of magnitude due to added effects of ontogenetic variation in both colony and testis size.

Colony sperm production in this experiment was assayed as the total cross-sectional area of testes within a colony. An ocular micrometer was employed to measure the length and width of a subsample of 16 of the mulberry-shaped testes in each colony. Testes were selected for measurement in groups of four, with each group consisting of the two pairs of testes in neighboring zooids (ensuring that subsamples incorporated variation within testes pairs and between testes in neighboring zooids). The four groups of testes measured were located in four different systems of zooids, which were distributed throughout the colony. Testis cross-sectional area (length x width) varied greatly among experimental male colonies, with colony means ranging from 0.031 to 0.189 [mm.sup.2] and averaging 0.111 [mm.sup.2]. Total testis production was then calculated for each colony as the average size of the testes sampled multiplied by two times the number of zooids in the colony. In order to incorporate as large a range of variation in total testis production as possible, some experimental colonies had not yet achieved their maximum testis or colony size.

Paternity assignment

The paternity of embryos brooded by experimental female colonies was assigned on the basis of electromorphs at the GPI (glucose-6-phosphate isomerase) locus. GPI electromorphs in B. schlosseri are known to be inherited as Mendelian alleles (Sabbadin 1982). Three alleles are moderately common in B. schlosseri populations in the Damariscotta River (numerical designations reflect relative mobility on a gel). Frequencies of the two most common alleles appear to vary slightly in space and time (GPI-100, 0.68 to 0.79; GPI-115, 0.21 to 0.30; Yund and McCartney 1994, Yund 1995) while the rarer GPI-105 allele has a frequency of 0.01 (Yund and McCartney 1994). Electrophoresis was performed on cellulose acetate gels (Titan III, Helena Laboratories, Beaumont, Texas) under previously described, standard conditions (Yund and McCartney 1994, modified from Grosberg 1987, 1991) with the exception that duration of electrophoresis runs was increased from 30 min to 38 min to improve separation of the GPI-100 and GPI-105 alleles.

Colonies used in mating experiments were collected from the field, cultured on glass microscope slides in a flowing seawater system, and screened for genotype at the GPI locus. Colonies homozygous for the GPI-100 and GPI-115 alleles were utilized as experimental females and as competing males. The experimental treatments assayed the fertilization success of focal males that varied in testis production under different intensities of sperm competition. Consequently, accurate paternity assignment for focal males was essential. Past in situ mating experiments utilizing homozygotes of the two more common alleles (Yund and McCartney 1994, Yund 1995, Atkinson and Yund 1996) received a low level of contamination from exogenous sperm, which could potentially confound paternity estimates. To minimize contamination errors in paternity assignment, all colonies selected for use as focal males in this experiment carried the rare GPI-105 allele. Focal males were either GPI-105/GPI-115 heterozygotes collected from nature or GPI-105/GPI-105 homozygotes bred in the laboratory. Colonies homozygous for the GPI-105 allele were bred by isolating pairs of field-collected GPI-105 heterozygotes in different gender phases (after the method of Grosberg 1987, 1991). Five crosses were conducted with different parental colonies that carried the GPI-105 allele. Offspring from these crosses were allowed to settle and metamorphose on microscope slides, reared in the flowing seawater system, and screened for GPI genotype. The 25% of the offspring that were homozygous for the GPI allele were reared to sexual maturity in the flowing seawater system prior to deployment in field experiments

Experimental design

All work was conducted at the University of Maine's Darling Marine Center and a field site in the Damariscotta River, in a channel between Carlisle Island and the mainland, 1.8 km downriver of the laboratory. This site has a fairly smooth, uniform bottom of small rock cobbles embedded in soft sediment. Naturally occurring B. schlosseri colonies are absent from this site due to the lack of appropriate substrata. A previous characterization of the flow regime at this site indicated that the tidally driven current is almost perfectly bi-directional and nonzero [greater than]98% of the time (Yund and McCartney 1994).

Laboratory cultured colonies were assembled into experimental populations in three treatments designed to manipulate the intensity of sperm competition. In all three treatments, a focal male of measured testis production (male F, [ILLUSTRATION FOR FIGURE 1 OMITTED]; either a GPI-105/GPI-105 homozygote or a GPI-105/GPI-115 heterozygote) was mounted in the center of a rectangular acrylic plate (15 x 91 x 0.6 cm), and two females (homozygous for either GPI-100 or GPI-115) were mounted 5 cm up- and downriver of the male (approximately north and south of the focal male; [ILLUSTRATION FOR FIGURE 1 OMITTED]). The competitor-free treatment contained no other males [ILLUSTRATION FOR FIGURE 1A OMITTED]. The intermediate intensity sperm competition treatment contained two additional males mounted between the focal male and the females ([ILLUSTRATION FOR FIGURE 1B OMITTED], males L). Competing males were homozygous for GPI-100 if the focal male was a GPI-105/GPI-115 heterozygote, and homozygous for either GPI-100 or GPI-115 if the focal male was a GPI-105/GPI-105 homozygote. The two competing males in this treatment had a combined testis cross-sectional area of 34.3 [+ or -] 2.6 [mm.sup.2] (mean [+ or -] 1 SE). The high intensity sperm competition treatment contained two competing males of the same GPI genotype in the same positions ([ILLUSTRATION FOR FIGURE 1C OMITTED], males H), but in this treatment the competing males combined had approximately double the testis production of the intermediate intensity treatment (68.4 [+ or -] 6.4 [mm.sup.2] of testes).

Replicate trials of these three treatments were deployed in the field during July, August, and September of 1994 and 1995. The testis production of the focal male varied in each trial. Fifteen trials of the competitor-free and 13 trials of the intermediate intensity sperm competition treatments were performed during 1994. In 1995, I conducted 7 competitor-free trials, 5 intermediate intensity sperm competition trials, and 11 high intensity sperm competition trials. In each field trial, an experimental population assembled in one of the three treatments was mounted 4 cm above a partially cement-filled cinder block and placed at [approximately]7 m depth (MLW) with the long axis of the plate oriented parallel to tidal flow. Trials were generally conducted in pairs, with each pair of experimental populations separated by at least 10 m cross-current to prevent fertilization between arrays (typical dispersal distances of fertilizing sperm in B. schlosseri are [less than]50 cm [Grosberg 1991, Yund 1995]). The area around the mating arrays ([approximately]300 [m.sup.2]) was patrolled twice weekly by divers who removed the few B. schlosseri recruits that became established. Divers also cleared drift kelp (a small fraction of which carried B. schlosseri colonies) from around the cinder blocks to minimize potential contamination from drifting sperm sources.

Each trial was deployed in the field for 7 d. Colonies were placed in the field [approximately]48 h prior to the anticipated time of sperm release and the onset of egg viability, and remained in place for [approximately]5 d after fertilization to permit embryos to complete development. The number of eggs brooded by each female colony was counted prior to field deployment and controlled within a limited range (mean [+ or -] 1 SE = 110.0 [+ or -] 7.3 eggs per female). At the end of each trial, female colonies were returned to the laboratory where all brooded embryos were counted and 30-33 embryos were surgically extracted for paternity analysis. Embryos were prepared for electrophoresis and gels run as previously described (Yund and McCartney 1994). Offspring that were heterozygous for either the GPI-105 allele (all trials) or the GPI-115 allele (trials with GPI-105/GPI-115 heterozygous focal males only) were nominally designated progeny of the focal male, though possible errors in paternity assignment via the more common GPI-115 allele are considered below.

Data analysis

The success of the focal male colonies in fertilizing eggs brooded by each female colony was assayed as the percentage of the female's initially available eggs that were fertilized by the focal male, which was calculated according to the following formula:

percentage of available eggs fertilized

= [(no. embryos brooded)

x (percentage of embryos fathered)]

[divided by] no. eggs initially produced.

This measure of male fertilization success incorporates variation in both the percentage of embryos lathered by a male and the proportion of eggs fertilized by all males. Since the proportion of eggs fertilized by all males may increase with local sperm concentration (in turn a function of the number of local males and their testis production), this assay provides a useful estimate of absolute male fertilization success that is independent of the number of eggs available to be fertilized (Yund and McCartney 1994, Yund 1995). In addition, I calculated the female fertilization success of each maternal colony as the percentage of her initially produced eggs (as counted prior to deployment) that developed into embryos (counted after retrieval from the field). All male and female fertilization success values were arcsine transformed prior to statistical analysis to normalize distributions. For ease of comprehension, the raw percentage values are reported in the text and figures.

Before considering the central issue of the shape of the male gain curve in each treatment, I conducted analyses to address five subsidiary issues. First, each focal male had the opportunity to fertilize eggs brooded by females positioned both up and down river. Since sperm swim very slowly relative to naturally occurring currents, they are advected exclusively downstream in moving water (Denny 1988, Denny and Shibata 1989). Consequently, in the bidirectional, tidally driven flow at my field site, sperm released by the focal male (at any point in time) could only fertilize eggs brooded by one of the two females. The fertilization of eggs brooded by the two females occurred at different points in time (when current direction reversed approximately every 6 h) and were likely the result of different episodes of sperm release by the focal male. Thus, the two values of male fertilization success obtained for each trial are in a sense independent estimates of that focal male's fertilization ability. Using both male fertilization success estimates in the gain curve analysis doubles the sample size, which increases the power to distinguish between a linear and saturating relationship. However, this approach would be invalid if male fertilization success systematically varied between the two females, as might be expected since current velocity at this field site is consistently higher on flood then ebb tides (Yund and McCartney 1994). An increase in current velocity should decrease male fertilization success due to increased sperm dilution (Pennington 1985, Denny 1988). I tested for this effect by comparing male fertilization success on eggs brooded by up- and downriver females (data combined across all treatments) with a Student's t test. Other possible sources of lack of independence for male fertilization success estimates (e.g., temporal dependence of sperm release, genotype-specific fertilization success) could not be explicitly tested with this data set.

Secondly, focal males in some trials were GPI-105/GPI-115 heterozygotes, while other focal males were GPI-105 homozygotes. Since the GPI-105 allele is quite rare in natural populations (allele frequency = 0.01), fertilizations obtained by exogenous sperm sources are not likely to bias the paternity estimates for GPI-105 homozygotes. However, the GPI-115 allele is more common, so paternity estimates for GPI-105/GPI-115 heterozygotes might be inflated if contamination by exogenous sperm was a frequent occurrence. To test for this possible effect, I compared the paternal success of GPI-105/GPI-115 focal males via the GPI-105 and the GPI-115 alleles (data combined across all treatments) with a Student's t test.

The previous comparison tests for the possibility of a contamination bias between the two alleles, but does not explicitly assay contamination levels from exogenous sperm. To estimate contamination rates directly, I also examined the incidence of contamination by exogenous sperm as a function of local male testis production in the competitor-free treatment. Since no local competing males were present in this treatment, the presence of GPI-100 alleles (in all trials) or GPI-115 alleles (in trials with GPI-105 homozygous focal males) in brooded embryos had to be the result of fertilization by sperm from exogenous males. Local males apparently acquire fertilizations at the expense of more distant males (Yund and McCartney 1994, Yund 1995), so I expected contamination rates to decline with increased local sperm competition. Consequently, I calculated the percentage of available eggs fertilized by known, exogenous males and regressed these values on testis production values for local males.

Fourthly, in the two treatments with competing males present, competitors varied slightly in testis production among trials. Although competing male testis production was narrowly controlled within each treatment, the remaining minor variation in testis production by competing males could nevertheless affect the shape of the focal male gain curve if competing male testis production was nonrandomly associated with focal male testis production. To test for this effect, I examined correlations between focal male testis production and competing male testis production within the two treatments in which local competitors were present (intermediate and high intensity sperm competition treatments).

Fifth, female fertilization success (the percentage of eggs fertilized by all males) was expected to increase among treatments with increased sperm availability (as previously demonstrated; Levitan 1991, Levitan et al. 1992, Yund and McCartney 1994, Yund 1995). I tested the one-tailed hypothesis that mean female fertilization success for the three treatments would exhibit the expected ranking (competitor-free [less than] intermediate intensity sperm competition [less than] high intensity sperm competition) via isotonic regression (Gaines and Rice 1990).

Finally, to evaluate the shape of the male gain curve in each of the three treatments, I fit power functions (of the form y = a[x.sup.b]) to the data on focal male fertilization success (y) as a function of testis production (x). To avoid using iterative techniques, I performed the power function fit as a linear regression on the log-transformed x and y values (equivalent to fitting a linear regression to a log-transformed power function of the form log y = log a + (b x log x)). Power functions have traditionally been used to model gain curves and the value of the exponent (b) in the function permits the interpretation of results in the context of existing sex allocation models (Charnov 1979, 1982). For each treatment I performed a t test to evaluate whether the exponent differed significantly from one (i.e., the value that the exponent would assume if the function were linear). Exponent values significantly [less than]1 indicate a saturating function.

RESULTS

Directional effect

For 12 the 102 female colonies present in the 51 field mating trials performed in this study, I was not able to collect data on progeny due to either the death of or damage to the female during the course of a trial, or due to the premature release of larvae (five females in competitor-free trials, four females in intermediate intensity sperm competition trials, and three females in high intensity sperm competition trials). For the remaining 39 trials for which I had data for both female colonies, I first tested for variation in the ability of focal males to fertilize eggs brooded by female colonies up- vs. downriver. Although there was a trend toward higher male fertilization success on eggs brooded by downriver females (the direction with lower average current velocities [Yund and McCartney 1994]), this difference was not significant (mean of 45.0 vs. 40.1% of the available eggs fertilized; Student's t test on arcsine-transformed values: t = 0.68, P [greater than] 0.45). In the absence of an up- /downriver effect, fertilizations of eggs brooded by the two females in each field trial are treated as independent estimates of focal male fertilization success in the remaining analyses. However, dropping this assumption and substituting the mean of the two male fertilization success estimates in all subsequent analyses changes none of the qualitative conclusions of this study (P.O. Yund, unpublished analysis). For the 39 trials with data for two females I use two male fertilization success estimates for each focal male testis allocation measurement. In the remaining 12 trials (with data for only single females) I have one male fertilization success estimate per testis allocation measurement.

Contamination bias

Focal males in all trials were either GPI-105/GPI-105 homozygotes or GPI-105/GPI-115 heterozygotes. Since the GPI-105 allele is quite rare in natural populations, this allele will provide a more accurate estimate of male fertilization success than the GPI-115 allele if many sperm from outside of the mating arrays fertilized the eggs of experimental females. Contaminating sperm contain GPI alleles in proportion to allele frequencies in natural populations (Yund and McCartney 1994), potentially resulting in the erroneous assignment of paternity to focal males via the more common GPI-115 allele. For the 23 trials (combined among treatments) that utilized GPI-105/GPI-115 heterozygotes as focal males, I tested for a possible contamination bias by comparing the paternity of the focal male via the GPI-105 and GPI-115 allele. Although paternal success estimated via the GPI-115 allele (across all experiments) tended to be slightly higher than estimates via the GPI-105 allele, this difference was not significant (mean of 28.6 vs. 24.0 % of the available eggs fertilized; Student's t test on arcsine-transformed values: t = 1.59, P [greater than] 0.10). Since paternity assigned due to the presence of the GPI-115 allele in embryos was not significantly higher than paternity assigned via the GPI-105 allele, any contamination from exogenous sperm did not significantly elevate the assigned fertilization success of GPI-105/GPI-115 focal males relative to GPI-105/GPI-105 homozygotes.

Relationship between fertilization by exogenous sperm and local testis production

While the previous analysis tests for contamination bias between the two alleles possessed by focal males, data from the competitor-free trials permit an explicit assessment of how many fertilizations were obtained by males from outside of the experimental arrays. Since no local competing males were present in this treatment, and only GPI-105 and GPI-115 (in 11 trials) alleles were present in focal males, embryos carrying the GPI-100 (in all trials) and GPI-115 alleles (in 11 of 22 trials) must have been fathered by males from outside the experimental populations. A regression of the percentage of initially available eggs fathered by known exogenous males on local testis production (production of focal males only) indicates a strongly negative relationship [ILLUSTRATION FOR FIGURE 2 OMITTED]. A power function (significant at P [less than] 0.005) provides the best fit to the data, with [r.sup.2] = 0.23. Above a local testis production of [approximately]25 [mm.sup.2] of testes, exogenous sperm accounted for only [approximately]8-12% of the fertilizations, a value comparable to past estimates from other methods (Yund and McCartney 1994, Atkinson and Yund 1996, Yund 1995). The relatively high rate of fertilization by exogenous sperm experienced in some trials with low testis production, focal males apparently did not bias paternity (per previous analysis) because all of the lowest allocation focal males happened to be GPI-105/GPI-105 homozygotes [ILLUSTRATION FOR FIGURES 2 AND 3 OMITTED].

Relationship between focal and competing male testis production

Although testis production of competing males was controlled within relatively narrow limits, the remaining minor variation could potentially alter the shape of the focal male gain curve if the testis production of focal and competing males was nonrandomly associated. To test for this effect, I examined correlations between testis production of focal and competing males in the intermediate and high intensity sperm competition treatments. There were minor trends toward a positive association in the intermediate intensity treatment and a negative association in the high intensity treatment, but in neither treatment was the relationship significant (r = 0.28 and 0.30, respectively, P [much greater than] 0.05 in both cases). Thus, minor variation in competing male testis production is not likely to have affected the shape of the focal male gain curve.

Female fertilization success

As expected, female fertilization success (the percentage of eggs fertilized by all males) increased among treatments with increased sperm availability from local males (isotonic regression on arcsine-transformed values; hypothesis, high intensity sperm competition [greater than] intermediate intensity sperm competition [greater than] competitor-free, [[E.sub.3].sup.-2] = 0.09, P [less than] 0.005). With just the variable focal male colony present (competitor-free treatment), 77.92 [+ or -] 2.51% (mean [+ or -] 1 SE) of the females' eggs were fertilized. With the addition of other males and increased testis production, female fertilization success increased to 80.52 [+ or -] 2.44 and 87.88 [+ or -] 3.43% in the intermediate and high sperm competition treatments, respectively.

[TABULAR DATA FOR TABLE 1 OMITTED]

The effect of the intensity of sperm competition on male gain curves

In all three treatments there was a highly significant, positive relationship between testis production and male fertilization success ([ILLUSTRATION FOR FIGURE 3 OMITTED], Table 1). However, the shape of the relationship varied among treatments according to the intensity of sperm competition. In the absence of local competitors for fertilizations [ILLUSTRATION FOR FIGURE 3A OMITTED], male fertilization success started to level off above a fairly low testis production level ([approximately]25 [mm.sup.2] of testes). The exponent in the power function was significantly [less than]1 (Table 1), indicating a nonlinear, saturating relationship. Except for males at the lowest end of the testis production spectrum, increased testis production would have resulted in only a marginal increase in fertilization success.

In the intermediate intensity sperm competition treatment [ILLUSTRATION FOR FIGURE 3B OMITTED], male fertilization success started to level off at a higher testis production level ([approximately]50 [mm.sup.2] of testes) than in the competitor-free treatment [ILLUSTRATION FOR FIGURE 3A OMITTED]. The exponent in the power function was still significantly [less than] 1 (Table 1), indicating a nonlinear relationship, but the exponent value was higher than in the competitor-free treatment. Male fertilization success in this treatment tended to be lower at all testis production levels than in the competitor-free treatment (compare the vertical height of the regression lines in [ILLUSTRATION FOR FIGURE 3A AND B OMITTED].

Finally, in the high intensity sperm competition treatment, male fertilization success continued to increase linearly throughout the range of testis production examined [ILLUSTRATION FOR FIGURE 3C OMITTED]. The exponent in the power function did not differ significantly from one (Table 1), indicating an almost strictly linear relationship. Male fertilization success in this treatment was apparently lower at all testis production levels than in the first two treatments (compare the vertical height of the regression lines in [ILLUSTRATION FOR FIGURE 3A-C OMITTED]), as expected under conditions of increased sperm competition (Yund and McCartney 1994).

DISCUSSION

Sperm competition and male gain curves

As previous theoretical work has predicted (Fischer 1981, Lloyd 1984, Petersen 1991), the male gain curves assayed for focal males in my experimental populations became progressively more linear as the intensity of sperm competition increased among treatments [ILLUSTRATION FOR FIGURE 3 OMITTED]. In the absence of nearby competitors for fertilizations, even males with low testis production were quite capable of fertilizing nearby brooded eggs, and higher testis production levels resulted in only minor increases in fertilization success [ILLUSTRATION FOR FIGURE 3A OMITTED]. In the presence of two competing males of low testis production, the fertilization success of focal males continued to increase with total testis cross-sectional area, but this relationship still decelerated at higher testis production levels [ILLUSTRATION FOR FIGURE 3B OMITTED]. Only in the presence of two competitors with higher testis production did focal male fertilization success increase linearly with testis cross-sectional area throughout the range examined [ILLUSTRATION FOR FIGURE 3C OMITTED]. The increased intensity of sperm competition across treatments reduced overall male fertilization success (compare heights of the curves in [ILLUSTRATION FOR FIGURE 3 OMITTED]) in spite of minor (but significant) increases in the percentage of eggs fertilized by all males combined.

My results support past studies that have interpreted male allocation patterns (Fischer 1981, Petersen 1991) and gain curves (McCartney 1997) under expectations of a linear relationship in the presence of sperm competition and diminished gains in its absence or reduction. A comparison with McCartney's (1997) result is particularly illuminating. He demonstrated a linear male gain curve in a bryozoan that is known to compete for fertilizations (Yund and McCartney 1994), under experimental conditions in which sperm competition would have been expected to be intense (McCartney 1997). The close similarity between the outcome of his study and the results from my high intensity sperm competition treatment [ILLUSTRATION FOR FIGURE 3C OMITTED] suggests that the effect of sperm competition on male gain curves may be similar for very distantly related taxa that have similar reproductive strategies (i.e., internal fertilization and feeding structures that may be able to collect and concentrate sperm).

Exogenous sperm and contamination

Exogenous sperm fathered a substantial number of embryos when very little local sperm was available and an average of [approximately]8-12% of the embryos in trials in which lone focal males had relatively high testis production ([greater than]25 [mm.sup.2] cross-sectional area; [ILLUSTRATION FOR FIGURE 2 OMITTED]). Although the presence of some exogenous fertilizing sperm did not result in a significant contamination bias for paternity estimates via the GPI-115 allele, the nonsignificant trend toward higher GPI-115 vs. GPI-105 estimates ([approximately]4% of available eggs fertilized), combined with the inverse relationship between exogenous fertilizations and local sperm production [ILLUSTRATION FOR FIGURE 2 OMITTED], suggest that a contamination bias would have occurred if GPI-105/GPI-115 heterozygotes had been used for the focal males with lowest testis production. The [approximately]4% contamination bias estimate via GPI-115 alleles is lower than the 8-12% exogenous fertilization level for lone focal males with higher testis production [ILLUSTRATION FOR FIGURE 2 OMITTED] because the latter estimate includes fertilizations via the more common GPI-100 allele. The ratio of the contamination bias to the exogenous sperm fertilization level is well within the range expected on the basis of relative allele frequencies. If the exogenous fertilization level continued to decrease with increasing local sperm availability beyond the range examined [ILLUSTRATION FOR FIGURE 2 OMITTED], exogenous sperm may actually have accounted for [less than]8-12% of the fertilizations in the two sperm competition treatments. However, the presence of all three alleles in experimental colonies precludes a quantitative assessment of exogenous fertilization levels in these treatments.

Male fertilization success in nature

The results from these experimental populations will be valid for males in natural populations to the extent that the experimental conditions were representative of conditions in the field. In nature, Botryllus schlosseri populations exhibit substantial spatial and temporal variation in density, ranging from [approximately]1 to 1000 colonies/[m.sup.2] (Grave 1933, Grosberg 1982; P.O. Yund, unpublished data). Since colonies release sperm throughout most of their reproductive cycle (Stewart-Savage and Yund 1997), synchronous release of sperm by neighbors is probably common and likely to lead to sperm competition. Natural variation in density and the intensity of sperm competition is likely to encompass and exceed the range of conditions considered in this study.

Female colonies were located only 5 cm from the focal males, thus possibly elevating male fertilization success and leading to a saturating gain curve in the absence of competitors. However, these colony spacing distances are typical of high density natural populations, where colonies routinely grow in contact with each other (Grosberg 1987, Rinkevich and Weissman 1987a, b). Moreover, a previous study conducted in a dense field population demonstrated that male fertilization success declines rapidly with distance, with males fathering only [approximately]15% of the embryos at 10-cm distance (Grosberg 1991). These data suggest that fertilization in dense populations is unlikely to occur over spatial scales much greater than those examined here.

However, in the absence of competing males, isolated males do have the ability to fertilize eggs at greater distances (at least tens of centimeters; Yund 1995). Additional evidence of long distance dispersal by some fertilizing sperm is provided by the contamination results in this study [ILLUSTRATION FOR FIGURE 2 OMITTED]. Given the spatial isolation of my experimental populations, exogenous sperm had to come from either neighboring arrays (10 m), distant natural populations ([greater than]40 m), or colonies moving through the study site on drift kelp (Yund and McCartney 1994). An inverse relationship between male density and effective fertilization distance suggests that mating area may vary with population density (Yund 1995). If so, the relevant spatial scale for assessing gain curves for males living at low population densities may be greater than for males in dense populations.

This study manipulated the intensity of sperm competition by increasing the number ([ILLUSTRATION FOR FIGURE 1B OMITTED] vs. a) and testis production ([ILLUSTRATION FOR FIGURE 1C OMITTED] vs. b) of competing males, while controlling the number of female colonies present. This approach was adopted to discern the effects of increased sperm competition on the male gain curve for fixed position females. In natural populations, the intensity of sperm competition will also vary as a function of total population density (Atkinson and Yund 1996). As total population density increases, an increase in the number of female-phase colonies with eggs available to be fertilized may offset the loss of fertilizations of eggs brooded by each female (Atkinson and Yund 1996). However, a male's population-wide fitness gain should be the product of his gain (for the level of sperm competition experienced) on eggs brooded by each female and the total number of females present. Consequently, variation in total population density may alter the height of the gain curve, but should not change the basic shape of the relationship.

Evolutionary implications

Although the impact of variation in the intensity of sperm competition on the fertilization success that a male achieves for a given level of sperm production is an ecological problem, the implications of these changing relationships are largely evolutionary. The shape of the male gain curve under different ecological conditions dictates the selective pressures imposed on gamete allocation patterns. Three evolutionary consequences merit consideration.

First, my results have implications for the evolution of sperm production levels in marine invertebrates that release sperm into the water column (free-spawners). Males in these taxa typically produce vast quantities of sperm and gonads often comprise a large fraction of a male's body mass (Ansell and Lander 1967, Loosanoff 1969, Giese and Kanatani 1987). The traditional explanation for this phenomenon, fueled by recent observations that egg fertilization rates in the field are often [less than]100% (reviewed by Levitan 1995, Levitan and Petersen 1995), is that fertilization may be a markedly inefficient process in the marine environment. If released sperm are rapidly diluted in the water column, especially in turbulent flow (Denny 1988), high levels of sperm production may be the result of selection to maximize egg fertilization rates.

While sperm dilution and limitation undoubtedly occur in many marine taxa (and especially in certain habitats), the results presented here suggest an additional consideration. In some free-spawning taxa, high levels of sperm production may be at least in part the result of selective pressures to maximize male fertilization success in competitive situations. When male gain curves saturate at fairly low sperm production levels in the absence of competitors, higher levels of sperm production may only be adaptive in competitive situations. Sperm competition is probably most prevalent in taxa that brood eggs, inhabit relatively benign flow regimes, and possess feeding structures that might be used to collect and concentrate sperm (Yund and McCartney 1994).

Alternatively, high levels of sperm production might instead be adaptive if they enhance the ability of a male to fertilize distant eggs. Thus, even if the male gain curve saturates for fertilization of nearby eggs, it may remain linear through higher allocation levels if distant fertilizations are considered. Theoretical predictions suggest that the rapid dilution of sperm in the water column will minimize a male's fertilization success on distant eggs (but see Babcock and Mundy 1992) and render it relatively insensitive to variation in sperm production (Denny 1988, Denny and Shibata 1989). Nevertheless, this hypothesis merits further empirical attention because the area covered by a male's expanding sperm cloud (and hence presumably the number of potential mates) will scale with the square of the radius of that cloud. Thus, even a minimal change in a male's fertilization success with distance might dramatically increase total male fertilization success (Grosberg 1991).

Secondly, my results generally support the proposition that hermaphroditism in some colonial invertebrates may have evolved in order to maximize total individual reproductive success under conditions of decelerating gains via male function (Charnov 1979, 1982). However, an important component of allocation theory is that hermaphroditism will be favored by a saturating relationship between relative male allocation and fertilization success, with an assumed trade-off between male and female allocation (Charnov 1979, Brunet 1992). Due to the dual source of variation in sperm production in B. schlosseri (testis size and number of zooids), I had to measure absolute (as opposed to relative) male allocation in order to consider the full range of variation in sperm production present. Although allocation to testis size exhibits a three-way trade-off with allocation to egg production and asexual growth (i.e., zooids in future asexual generations; Yund et al. 1997), colony size (the second source of variation in colony sperm production) varies throughout colony ontogeny and is unlikely to exhibit a trade-off with female allocation.

Finally, in hermaphroditic, brooding marine invertebrates like B. schlosseri, the optimal allocation to male vs. female function may vary with population density. An increase in either population or male density should result in an increase in the intensity of sperm competition (Yund and McCartney 1994, Atkinson and Yund 1996). Under these conditions, colonies should be subject to a linear male gain curve [ILLUSTRATION FOR FIGURE 3C OMITTED] and male fertilization success should continue to increase with increased allocation to sperm production. In contrast, at lower male or population densities, colonies are likely to experience a saturating male gain curve [ILLUSTRATION FOR FIGURE 3A OMITTED]. Under these conditions, increased allocation to sperm production will yield few gains in male fertilization success. Consequently, assuming a linear female gain curve, total reproductive success in low density conditions should be maximized by increased allocation to female reproduction. Since B. schlosseri is subject to very limited gene flow via the dispersal of fertilizing sperm (Grosberg 1991, Yund 1995) and larvae (Grosberg 1987), the selective pressures that local density conditions exert via fertilization processes may result in locally adapted gamete allocation patterns

ACKNOWLEDGMENTS

R. Collin, R. Grosberg, S. Johnson, P. O'Neil, and M. McCarthey provided valuable comments on earlier drafts of this manuscript. M. Cashner assisted immeasurably with electrophoresis, and Y. Greenawalt, G. River, and B. Wagstaff provided additional valuable assistance in the field and lab. T. Miller and the other staff members of the University of Maine's Darling Marine Center also helped make this work possible. Financial support was provided by the National Science Foundation (OCE-94-16548).

LITERATURE CITED

Ansell, A. D, and K. F. Lander. 1967. Studies on the hard-shell clam Venus mercenaria, in British waters. Journal of Applied Ecology 4:425-435.

Atkinson, O. S., and P. O. Yund. 1996. The effect of variation in population density on male fertilization success in a colonial ascidian. Journal of Experimental Marine Biology and Ecology 195:111-123.

Babcock, R. C., and C. N. Mundy. 1992. Reproductive biology, spawning and field fertilization rates of Acanthaster planci. Australian Journal of Marine and Freshwater Research 43:525-534.

Benzie, J. A. H., K. P. Black, P. J. Moran, and P. Dixon. 1994. Small-scale dispersion of eggs and sperm of the crown-of-thorns starfish (Acanthaster planci) in a shallow coral reef habitat. Biological Bulletin (Woods Hole) 186:153-167.

Boyd, H. C., S. K. Brown, J. A. Harp, and I. L. Weissman. 1986. Growth and sexual maturation of laboratory-cultured Monterey Botryllus schlosseri. Biological Bulletin (Woods Hole) 170:91-109.

Brazeau, D. A., and H. R. Lasker. 1992. Reproductive success in a marine benthic invertebrate, the Caribbean octocoral Briareum asbestinum. Marine Biology 114:157-163.

Broyles, S. H., and R. Wyatt. 1990. Paternity analysis in a natural population of Asclepias exalta: multiple paternity, functional gender, and the "pollen-donation hypothesis." Evolution 44:1454-1468.

Brunet, J. 1992. Sex allocation in hermaphroditic plants. Trends in Ecology and Evolution 7:79-84.

Burd, M., and T. F. H. Allen. 1988. Sexual allocation strategy in wind-pollinated plants. Evolution 42:403-407.

Chadwick-Furman, N. E., and I. L. Weissman. 1995. Life histories and senescence of Botryllus schlosseri (Chordata, Ascidiacea) in Monterey Bay. Biological Bulletin (Woods Hole) 189:36-41.

Charlesworth, D., and B. Charlesworth. 1981. Allocation of resources to male and female functions in hermaphrodites. Biological Society of the Linnaean Society 14:57-74.

Charnov, E. L. 1979. Simultaneous hermaphroditism and sexual selection. Proceedings of the National Academy of Sciences (USA) 76:2480-2484.

-----. 1982. The theory of sex allocation. Princeton University Press, Princeton, New Jersey, USA.

Charnov, E. L., J. Maynard Smith, and J. Bull. 1976. Why be an hermaphrodite? Nature 263:125-126.

Denny, M. W. 1988. Biology and the mechanics of the wave-swept environment. Princeton University Press, Princeton. New Jersey, USA.

Denny, M. W., and M. F. Shibata. 1989. Consequences of surf-zone turbulence for settlement and external fertilization. American Naturalist 134:859-889.

Devlin, B., J. Clegg, and N. C. Ellstrand. 1992. The effect of flower production on male reproductive success in wild radish populations. Evolution 46:1030-1042.

Falconer, D. S. 1981. Introduction to quantitative genetics. Second edition. Longman Scientific and Technical, Essex, England, UK.

Fischer, E. A. 1981. Sexual allocation in a simultaneously hermaphroditic coral reef fish. American Naturalist 117:64-82.

Gaines, S. D., and W. R. Rice. 1990. Analysis of biological data when there are ordered expectations. American Naturalist 135:310-317.

Giese, A. C., and H. Kanatani. 1987. Maturation and spawning. Pages 251-329 in A. C. Giese, J. S. Pearse, and V. B. Pearse, editors. Reproduction of marine invertebrates. Volume IX. General aspects: seeking unity in diversity. Published jointly by Blackwell Scientific, Palo Alto, California, and Boxwood, Pacific Grove, California, USA.

Gosner, K. L. 1971. Guide to the identification of marine and estuarine invertebrates: Cape Hatteras to the Bay of Fundy. John Wiley and Sons, New York, New York, USA.

Grave, B. H. 1933. Rate of growth, age at sexual maturity, and duration of life of certain sessile organisms, at Woods Hole, Massachusetts. Biological Bulletin (Woods Hole) 65: 375-386.

Grosberg, R. K. 1982. Ecological, genetical, and developmental factors regulating life history variation within a population of the colonial ascidian Botryllus schlosseri (Pallas) Savigny. Thesis. Yale University, New Haven, Connecticut, USA.

-----. 1987. Limited dispersal and proximity-dependent mating success in the colonial ascidian Botryllus schlosseri. Evolution 41:372-384.

-----. 1988. Life-history variation within a population of the colonial ascidian Botryllus schlosseri. I. The genetic and environmental control of seasonal variation. Evolution 42:900-920.

-----. 1991. Sperm-mediated gene flow and the genetic structure of a population of the colonial ascidian Botryllus schlosseri. Evolution 45:130-142.

Harder, L. D., and J. D. Thomson. 1989. Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. American Naturalist 133:323-344.

Heath, D. J. 1979. Brooding and the evolution of hermaphroditism. Journal of Theoretical Biology 81:151-155.

Hughes, D. J. 1989. Variation in reproductive strategy among clones of the bryozoan Celleporella hyalina. Ecological Monographs 59:387-403.

Hughes, D. J., and R. N. Hughes. 1986. Life history variation in Celleporella hyalina (Bryozoa). Proceedings of the Royal Society (London), Series B 228:127-132.

Levitan, D. R. 1991. Influence of body size and population density on fertilization success and reproductive output in a free-spawning invertebrate. Biological Bulletin (Woods Hole) 181:261-268.

-----. 1995. The ecology of fertilization in free-spawning invertebrates. Pages 123-156 in L. McEdward, editor. Ecology of marine invertebrate larvae. CRC Press, Boca Raton, Florida, USA.

Levitan, D. R., and C. Petersen. 1995. Sperm limitation in the sea. Trends in Ecology and Evolution 10:228-231.

Levitan, D. R., M. A. Sewell, and F. Chia. 1992. How distribution and abundance influences fertilization success in the sea urchin Strongylocentrotus franciscanus. Ecology 73:248-254.

Lloyd, D. G. 1984. Gender allocations in outcrossing co-sexual plants. Pages 277-300 in R. Dirzo and J. Sarukhan, editors. Perspectives on Plant Population Ecology. Sinauer, Sunderland. Massachusetts, USA.

Loosanoff, V. L. 1969. Maturation of the gonads of oysters, Crassostrea virginica, of different geographic areas subjected to relatively low temperature. Veliger 11:153-163.

Maynard Smith, J. 1978. The evolution of sex. Cambridge University Press, Cambridge, England, UK.

McCartney, M. A. 1997. Sex allocation and male fitness gain in a colonial, hermaphroditic marine invertebrate. Evolution 51:127-140.

Milkman, R. 1967. Genetic and developmental studies on Botryllus schlosseri. Biological Bulletin (Woods Hole) 132:229-243.

Mukai, H., and H. Watanabe. 1976. Relation between sexual and asexual reproduction in the compound ascidian Botryllus primigenus. Science Reports of the Faculty of Education, Gunma University 25:61-79.

Oliver, J., and R. Babcock. 1992. Aspects of the fertilization ecology of broadcast spawning corals: sperm dilution effects and in situ measurements of fertilization. Biological Bulletin (Woods Hole) 183:409-417.

Pennington, J. T. 1985. The ecology of fertilization of echinoid eggs: the consequences of sperm dilution, adult aggregation, and synchronous spawning. Biological Bulletin (Woods Hole) 169:417-430.

Petersen, C. W. 1991. Sex allocation in hermaphroditic sea basses. American Naturalist 138:650-667.

Raimondi, P. T., and J. E. Martin. 1991. Evidence that mating group size affects allocation of reproductive resources in a simultaneous hermaphrodite. American Naturalist 138: 1206-1217.

Rinkevich, B., and I. L Weissman. 1987a. A long-term study on fused subclones in the Ascidian Botryllus schlosseri: the resorption phenomenon (Protochordata: Tunicata). Journal of Zoology (London) 213:717-733.

Rinkevich, B., and I. L Weissman. 1987b. The fate of Botryllus (Ascidiacea) larvae cosettled with parental colonies: beneficial or deleterious consequences? Biological Bulletin (Woods Hole) 173:474-488.

Sabbadin, A. 1958. Analisi sperimentale dello sviluppo delle colonie di Botryllus schlosseri (Pallas). Archivio Italianio di Anatomia e di Embriologia 63:178-221.

-----. 1982. Formal genetics of ascidians. American Zoologist 22:765-773.

Schoen, D. J., and S. C. Stewart. 1986. Variation in male reproductive investment and male reproductive success in white spruce. Evolution 40:1109-1120.

Sella, G. 1990. Sex allocation in the simultaneously hermaphroditic polychaete worm Ophryotrocha diadema. Ecology 71:27-32.

Snow, A. A., and P. O. Lewis. 1993. Reproductive traits and male fertility in plants. Annual Review of Ecology and Systematics 24:331-351.

Stewart-Savage, J., and P. O. Yund. 1997. Temporal pattern of sperm release in a colonial ascidian, Botryllus schlosseri. Journal of Experimental Zoology, in press.

Strathmann, R. R., M. Strathmann, and R. H. Emson. 1984. Does limited brood capacity link adult size and brooding, and simultaneous hermaphroditism? a test with the starfish, Asterina phylatica. American Naturalist 123:796-818.

Thomas, F. I. M. 1994a. Physical properties of gametes in three sea urchin species. Journal of Experimental Biology 194:263-284.

-----. 1994b. Transport and mixing of gametes in three free-spawning polychaete annelids, Phragmatopoma californica (Fewkes), Sabellaria cementarium (Moore), and Schizobranchia insignis (Bush). Journal of Experimental Marine Biology and Ecology 179:11-27.

Thomson, J. D., and B. A. Thomson. 1989. Dispersal of Erythronium grandiflorum pollen by bumblebees: implications for gene flow and reproductive success. Evolution 43:657-661.

Young, H. J., and M. L. Stanton. 1990. Influences of floral variation on pollen removal and seed production in wild radish. Ecology 71:536-547.

Yund, P. O. 1990. An in situ measurement of sperm dispersal in a colonial marine hydroid. Journal of Experimental Zoology 253:102-106.

-----. 1995. Gene flow via the dispersal of fertilizing sperm in a colonial ascidian (Botryllus schlosseri): the effect of male density. Marine Biology 122:649-654.

Yund, P. O., and M. A. McCartney. 1994. Male reproductive success in colonial invertebrates: competition for fertilizations. Ecology 75:2151-2167.

Yund, P. O., Y. Marcum, and J. Stewart-Savage. 1997. Life history variation in a colonial ascidian: broad-sense heritabilities and tradeoffs between growth and allocation to male and female reproduction. Biological Bulletin (Woods Hole) 192:290-299.
COPYRIGHT 1998 Ecological Society of America
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Yund, Philip O.
Publication:Ecology
Date:Jan 1, 1998
Words:9024
Previous Article:Effects of maternal and larval nutrition on growth and form of planktotrophic larvae.
Next Article:Stem tilting and pseudocephalium orientation in Cephalocereus columna-trajani (cactaceae): a functional interpretation.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |