Costs of Delayed Mating Opportunities, Overripening, and Mate Choice in Killifish
Females of highly iteroparous, oviparous species must mate multiply and potentially choose mates hundreds to thousands of times throughout their lifespan. Theory indicates that costly female mating preferences may favor low levels of preference in species that have to mate repeatedly. One potential cost of mating preference is overripening, which occurs in some external fertilizers when females ovulate but do not spawn their eggs. Overripening may create a trade-off between the benefits of mate choice (i.e., rejecting males of poor quality) and egg viability. This study tested for overripening as a cost of delayed mating and determined its effects on female mating preferences in the Bluefin Killifish, Lucania goodei. Female L. goodei often mate multiple times each day across multiple days when gravid. Females were divided into three treatments: mate every day, mate every two days, and mate every four days. In the first experiment, females were paired with either a red or a yellow conspecific male on spawning days. In the second experiment, females were paired with either a conspecific male (L. goodei) or a heterospecific male (L. parva) on spawning days. In both experiments, the eggs were retrieved following spawning, and their viability (i.e., survival) was monitored for five days. Females that were forced to ‘hold onto’ their eggs for extended periods had smaller clutch sizes and lower egg viability, demonstrating that costs to rejecting males potentially exist. In the first experiment, there was no evidence of female choice for red versus yellow males. In the second experiment, female preference for conspecifics versus heterospecifics declined when females could only mate once every four days, but the differences among treatments did not reach statistical significance. The results of this study illustrate a potential cost to female mate choice and suggest that the benefits of mate choice may not outweigh the costs due to egg overripening.
ALL sexually reproducing species must mate, and, with the exception of animals engaging in self-fertilization, they must mate with another individual. In many animal systems, mating involves the process of finding and accepting a potential partner (Rosenthal, 2017). The decisions made during mate choice are likely to have significant consequences on the fitness of both the adults and the resulting offspring (Andersson, 1994; Fuller et al., 2005a; Kokko et al., 2006). For the adults engaging in reproduction, the time/energy spent finding mates, the energy exerted while courting (or being courted), and the potential costs due to increased predation risk can affect future reproduction and survival (Rowe, 1994; Houde, 1997; Rosenthal, 2017). However, mate choice can also confer benefits. Even in systems lacking parental care, the individual properties of a potential mate are likely to influence the health/viability of the offspring and may be a target of preference. An extreme example of this phenomenon is the widespread preference to mate with conspecifics (i.e., members of the same species; Coyne and Orr, 2004). There are also potential genetic benefits from mate choice within species. In species with genetic variation in health/viability in prospective mates, choosing mates of high quality may be adaptive as the offspring may also have increased health/viability. There are two primary aspects of mate choice: (i) ‘preference’—the aspect to which an individual ranks prospective mates and (ii) ‘choosiness’—the effort an individual is prepared to invest in mate assessment (Jennions and Petrie, 1997; Rosenthal, 2017). Here, we examine the potential costs of mate choice and ask whether this affects choosiness in females.
An extensive theoretical literature indicates that the costs of expressing mating preference greatly influence subsequent evolutionary dynamics (Pomiankowski, 1987; Houle and Kondrashov, 2002; Kokko et al., 2002). Yet empirically demonstrating costs to preference is a challenge. Proposed costs to mate choice include time/energy expenditure, aggression from conspecifics or heterospecifics, or predation risk (Reynolds and Côté, 1995; Jennions and Petrie, 1997; Hunt et al., 2005). Another potential cost concerns the demands of iteroparity and repeated mating. All else being equal, an individual that needs only to mate once will suffer lower costs of mate choice than an organism that mates many times (Warner, 1998). Another lesser-explored potential cost present in many oviparous species is overripening—the process of decreasing egg viability post-ovulation due to a female choosing not to spawn (Moran et al., 2018). Exerting mating preferences often entails rejecting potential mates, increasing the time required to find a mate, and potentially decreasing egg viability, particularly for external fertilizers. Multiple fish species undergo overripening, including Atlantic salmon Salmo salar (de Gaudemar and Beall, 1998), Atlantic Halibut Hippoglossus hippoglossus (Bromage et al., 1994), Goldfish Carassius auratus (Formacion et al., 1993), and Threespine Stickleback Gasterosteus aculeatus (Roufidou et al., 2016; Roufidou and Borg, 2019). Overripening can decrease the total number of eggs produced and reduce the potential for females to spawn in the future (Foote, 1989; de Gaudemar and Beall, 1998). In Etheostoma caeruleum, the Rainbow Darter, eggs often failed to hatch when the female was isolated for around six hours or more (Moran et al., 2018). Therefore, fish that engage in mate choice must balance these benefits with the costs of overripening. In theory, females that face substantial costs may evolve to be less ‘choosy’ in mating.
Lucania goodei, Bluefin Killifish, is an excellent model for examining the risk–reward tradeoff between the advantage of mate choice and costs to mating due to overripening. First, Bluefin Killifish are highly iteroparous (Breder and Rosen, 1966). The breeding season is long, extending from January through mid-September, but with notable population variation (Foster, 1967; Arndt, 1971). Gravid females can be found year-round in some localities, albeit in lower abundance during winter months (Foster, 1967; Arndt, 1971). In the lab, female Bluefin Killifish will spawn year-round but with reduced egg output from October through January (Fuller, pers. obs.). Females become gravid multiple times over their lifespans (Fuller, pers. obs.). Each time a female becomes gravid, she spawns every day or every other day for approximately 10–14 days. Presumably, females are ovulating daily, similar to Medaka (Leaf et al., 2011; Hilgers and Schwarzer, 2019; Murata et al., 2020). During these times, females spawn ∼10 times each day, depositing a single egg in each spawning event (Arndt, 1971). Hence, females spawn hundreds of times over their lifespans.
Second, female L. goodei show a clear mate choice for conspecific mates over heterospecific mates (Gregorio et al., 2012; Kozak et al., 2015; St. John and Fuller, 2019, 2021). In some populations, L. goodei occur in sympatry with their sister species, Lucania parva (Fuller and Noa, 2008). Hybrids occur at low levels in nature (Hubbs et al., 1943). F1 hybrids have relatively high survival, but the fertility of some male F1 hybrids is 50% of either parental species, and back-crosses into L. goodei have 25% survival relative to the parent species (Fuller, 2008). In areas of sympatry, females of both species and males of L. goodei have very strong preferences to mate with conspecifics (Kozak et al., 2015; St. John and Fuller, 2021).
This study explored the potential tradeoff between females being ‘choosy’ and incurring the costs of overripening in L. goodei. The two overarching goals of this experiment were (i) to establish whether there are costs consistent with overripening due to induced delayed mating and (ii), if so, to determine whether these costs of delayed mating were associated with changes in existing mate choice inclinations in female Bluefin Killifish. We predicted that delayed mating periods would result in decreased egg viability and smaller clutch sizes and that these costs would be associated with decreased ‘choosiness’ in females. To address these predictions, we performed a no-choice study. We assigned female L. goodei to one of three mating schedules and observed their reproductive success when mated with either a conspecific male or a heterospecific male.
MATERIALS AND METHODS
Study site and husbandry.—
This study was conducted at the Shelford Vivarium Greenhouse at the University of Illinois at Urbana-Champaign. Lucania goodei and L. parva were collected using seines and dip nets from the Wakulla River in Wakulla County, Florida, in January 2021. The fish were transported in aerated buckets and coolers to the University of Illinois, where they were housed in stock tanks at ∼2 ppt salinity. The fish were fed frozen Artemia daily. For both experiments 1 and 2, males and females were held separately, except when spawning, in 110-liter glass aquaria. The experiments were conducted in March–June 2021. These aquaria were kept in a greenhouse at the University of Illinois. Air temperatures varied between 16 and 29°C. The photoperiods followed natural seasonal patterns.
Experiment 1: Spawning frequency and conspecific mate choice.—
The first experiment aimed to determine whether delayed mating opportunities affected egg viability, clutch size, and mating preferences. To do this, we manipulated the frequency with which females could mate with males. There were three treatments: (1) a female could interact and mate with a male every day, (2) a female could interact and mate with a male every other day, and (3) a female could interact and mate with a male once every four days. Each female was held separately in a 110-liter tank. On spawning days, individual females were transferred from their tanks and placed with a male with either red or yellow anal fins in his holding tank. Hence, each replicate block had five 110-liter tanks to hold five separate fish (a red male, a yellow male, a female who spawned every day, a female who spawned every other day, and a female who spawned once every four days). We performed six replicate blocks for a total of 30 110-liter tanks.
Experiment 2: Spawning frequency and heterospecific mate choice.—
This experiment was a continuation of the conspecific mate choice study described above. Each replicate block had a single male L. goodei, male L. parva, and three female L. goodei assigned to one of three treatments: spawn daily, spawn every other day, or spawn once every four days. For the male L. goodei, half were of the red color morph, and the other half were yellow. As with experiment 1, individual females were transferred from their holding tanks and placed in a tank containing either a conspecific L. goodei or a heterospecific L. parva. Hence, each replicate block had five 110-liter tanks to hold five separate fish (male L. goodei, male L. parva, a female who spawned every day, a female who spawned every other day, and a female who spawned once every four days). There were six replicate blocks for a total of 30 110-liter tanks. The fish assigned to each replicate were roughly the same size, and the male that each female was assigned to each day was staggered so that each male mated at most once a day. Each tank was also equipped with floating mops and mops that would sink to the bottom. The mops served as spawning substrates and shelter.
On spawning days, females were transferred from their holding tank to a male tank. The fish were then allotted 4–5 hours to interact and spawn before the female was transferred back to her original tank. The spawning mops were then meticulously combed to locate and extract the females’ eggs. These eggs were collected in a labeled container that designated their replicate, fish ID, trial, date, and quantity of eggs. They were then treated with dilute methylene blue, and their survival was monitored over five days post-extraction to determine the survivability of the eggs. Overripening results in inviable eggs. By day five, we could clearly identify dead eggs by a lack of development. Mops were returned to the tanks after egg extraction. Tubs containing the eggs were checked daily, and the number of remaining eggs was recorded for each trial.
Statistical analyses.—
The goals of the experiment were to determine whether delays in mating, which we induced by holding females in isolation for various amounts of time, affected (a) egg survival, (b) the total number of eggs spawned over the experiment, (c) the number of eggs spawned with each male (i.e., clutch size), and (d) the inferred preference (experiment 1: red versus yellow males, experiment 2: conspecifics versus heterospecifics). We summed the total number of eggs laid by each female with each of the males. We measured clutch size as the average number of eggs laid each time a female was paired with a male. We measured egg survival over the first five days of development as the number of eggs alive at day five divided by the total number of eggs obtained. We measured mate choice as the proportion of eggs spawned with and the total number of days spawned with the red versus yellow (experiment 1) or conspecific versus heterospecific males (experiment 2). We performed paired t-tests to ask whether there was an overall preference to spawn with red versus yellow males (experiment 1) and conspecifics versus heterospecifics (experiment 2).
We also performed full mixed models to examine the effects of treatment, male color (or conspecific versus heterospecific), and their interaction on the number of eggs laid, the average clutch sizes, the propensity to spawn when given the opportunity, and the survival of the eggs. The total number of eggs and clutch size were analyzed with a linear mixed model where the replicate block and female ID were random factors. Mixed models used the ‘lme4’ package in R. The fixed effects were analyzed using ‘anova’ from the ‘lmerTest’ package and Sattherthwaite’s degrees of freedom in the error term. We checked that the residuals did not differ from a normal distribution using Shapiro’s test. Egg survival, probability of spawning, and proportion of eggs with either male were analyzed with a generalized linear mixed model with a binomial distribution where replicate block and female ID were random factors. For the binomial models, we analyzed the significance of the model terms using a type 3 model in the ‘car’ package. For experiment 1, we found little effect of male color on any of the measured response variables, so we simplified our models by removing the effect of male color.
All contrasts were set as ‘options(contrasts = c(“contr.sum”,“contr.poly”))’. All analyses were performed in R (version 4.3.2; R Core Team, 2023). Raw data and R scripts can be found at Dryad (see Data Accessibility).
RESULTS
Experiment 1: Red versus yellow males.—
There was little evidence that females had meaningful preferences for red versus yellow males (Table 1A, Supplemental Table 1; see Data Accessibility). There was no effect of male color on any of the variables measured, nor was there an interaction between male color and spawning frequency (Supplemental Table 1; see Data Accessibility). The number of eggs laid and the propensity to spawn did not differ between red and yellow males (paired t-tests, P > 0.22 in all tests).
Delaying spawning reduced egg survival ( = 8.61, P = 0.0135, Fig. 1A). Egg survival was highest when females were allowed to spawn daily (0.96) compared to when they were allowed to spawn every other day (0.88) or once every four days (0.85). Females tended to spawn more total eggs when allowed to spawn daily versus every two or four days (Fig. 2A), but this pattern did not reach statistical significance (F2,10 = 2.85, P = 0.1045). This pattern did reach statistical significance when spawning frequency was treated as a continuous versus a categorical variable (F1,11 = 4.92, P = 0.0488). On average, females that were allowed to spawn daily had 1.6X as many eggs as females that were allowed to spawn once every four days.
Citation: Ichthyology & Herpetology 113, 2; 10.1643/i2024007
Citation: Ichthyology & Herpetology 113, 2; 10.1643/i2024007
Delaying the frequency at which females could spawn also altered clutch size and the likelihood that a female spawned when given the opportunity (clutch size, F2,10 = 4.7, P = 0.036; propensity to spawn, = 6.55, P = 0.0378; Fig. 2B, C). Females that were only allowed to spawn once every four days had large clutches and were more likely to spawn than females that could spawn every other day or every day. Essentially, females were more likely to spawn and lay more eggs when they could only spawn once every four days, but these adjustments in behavior could not compensate for the advantage that females had when allowed to spawn daily.
Experiment 2: Conspecific versus heterospecific males.—
Consistent with previous studies, female L. goodei had a clear preference for conspecific male L. goodei over heterospecific male L. parva (Tables 1B, 2; Figs. 3, 4; paired t-tests P < 0.0011 in all tests). Females were more likely to spawn when paired with conspecific males. Likewise, females spawned more eggs and had larger clutch sizes with L. goodei than L. parva across all three treatments. Females were more likely to forego spawning when paired with L. parva.
Citation: Ichthyology & Herpetology 113, 2; 10.1643/i2024007
Citation: Ichthyology & Herpetology 113, 2; 10.1643/i2024007
There was also good evidence that delaying mating opportunities was costly. Egg survival decreased when females were forced to delay mating (Table 2, Fig. 1B). Egg survival was highest when females were allowed to spawn daily (0.91) and was reduced when females were allowed to spawn once every two (0.62) or four (0.72) days. The larger decreases in egg viability in experiment 2 are likely due to the fact that females were more likely to reject males when paired with L. parva, which increased the realized number of days between spawning. There was also a strong trend for the total number of eggs produced to decrease with spawning frequency (P = 0.055). The total number of eggs spawned was highest when females were allowed to spawn daily (average 97.7 eggs) and was reduced when females were allowed to spawn once every two (average 55.5 eggs) or four days (37.8 eggs). There were no effects of spawning frequency on average clutch size.
Analyses on female preference indicate that preference did not differ among the treatments. Overall, preference for L. goodei was high. However, preference became more variable when females were only allowed to spawn once every four days. Figure 4 shows the proportion of eggs laid with male L. goodei versus male L. parva for each female and the mean and 95% confidence limits. Females who spawned every day or every other day laid a high proportion of their eggs with L. goodei, and this differed from a null expectation of 0.5 (everyday treatment: t5 = 6.24, P = 0.0016; every other day treatment: t5 = 4.42, P = 0.0069). Compared to the null expectation of no preference (0.5), females that spawned once every four days did not exert a significant preference for either male species (t5 = 1.01, P = 0.36).
DISCUSSION
This study showed that female L. goodei have a significant preference for conspecific mates over heterospecific mates, which is in keeping with previous studies. We sought to determine whether there are substantial costs to withholding mating opportunities from females and whether these costs will impact their established mate preference. We initially hypothesized that females would incur greater costs when withheld from spawning for substantial periods. As expected, we found costs to the female’s reproductive success when females were forced to retain eggs over time. These costs included decreased egg viability and reduced total number of eggs spawned. Across the three treatments (mate every day, every two days, and every four days), egg viability and the total number of eggs spawned decreased as the time between opportunities to mate increased.
We also predicted that females withheld from mating opportunities for extended periods would incur greater costs and show reduced mating preferences compared to the other treatments. Since this study was a no-choice study, females could only choose whether or not to spawn with a male with whom she was paired. The design is considered conservative (Houde, 1997; Dougherty and Shuker, 2015; St. John and Fuller, 2019). No-choice studies have been used effectively to show both female and male mating preferences in Lucania, including preference for conspecifics over heterospecifics (Fuller et al., 2007; St. John and Fuller, 2019), increased levels of conspecific preferences in area of sympatry (St. John and Fuller, 2021), and increased levels for native males (i.e., males from the female’s native population) in areas of sympatry (Kozak et al., 2015). In this study, there were statistically significant levels of preference for male L. goodei when females mated once every day or every other day, but this preference was absent when females could only mate once every four days. However, female mating preferences did not differ among the treatments due to high variation in preference when females mated once every four days. Further experimentation is needed to resolve this issue.
This study provided clear evidence for overripening. The costs of overripening align with those of other studies and include lower egg viability and smaller clutch sizes (Foote, 1989; de Gaudemar and Beall, 1998). A remaining question is whether females suffer these costs in nature. We do not know the answer for L. goodei, but overripe females have been found in nature in Threespine Stickleback populations (Roufidou et al., 2016). In Bluefin Killifish (L. goodei), females mate repeatedly across multiple males in a single day and on multiple consecutive days. The availability of males is also potentially limited, as not all males possess spawning territories. Given these constraints—(a) females likely spawn hundreds to thousands of times across their lifespan, (b) not all males possess mating territories, and (c) retaining eggs internally causes reductions in egg viability—natural selection should favor females that reduce choosiness as costs increase.
The existing literature illustrates that highly iteroparous species will be more risk-aversive regarding their mortality and much more likely to abandon current reproduction to preserve future spawning opportunities and egg viability (Stearns, 1992; Warner, 1998; Roff, 2001). Because overall lifetime fitness would not be significantly affected by the fitness gains from a single mating, females of iteroparous species may trade preferential mate choice for future fecundity. In contrast, species that produce few offspring might accept greater costs to preferentially mate with higher-fitness males (Gabor and Halliday, 1997; Bleu et al., 2012). Incorporating these dynamics into models studying sequential mate choice might change the outcome of these models and our understanding of female choosiness in iteroparous species (Crowley et al., 1991).
Finally, multiple aspects of the ovarian physiology of this species are unknown. The spawning cycle resembles Medaka, with an extended breeding season during which females can spawn almost daily (Koger et al., 1999; Leaf et al., 2011; Hilgers and Schwarzer, 2019). In Medaka, ovarian follicles rupture from the ovary in the early morning before dawn and are stored in the ovarian lumen (Murata et al., 2020). Females then promptly spawn with males and release these eggs. The spawning pattern appears similar in Bluefin Killifish, although detailed studies of ovulation have not been conducted. The exact triggers determining when females do and do not spawn are unknown. In many species, females will ‘drop’ their unfertilized eggs and eat them when they lack access to a mate (Myint et al., 2011; darters: Fuller, pers. obs.). This does not appear to be case in Bluefin Killifish. Multiple breeding studies have been conducted using Lucania (McCune et al., 2002; Fuller and Travis, 2004; Fuller et al., 2005b, 2010, 2022; Fuller and Noa, 2010; Berdan et al., 2021). Dead, unfertilized eggs have never been found in female holding tanks (Fuller, pers obs.). Females may possibly resorb the eggs, but further study is needed to resolve this issue.
In conclusion, this study examined whether or not there were costs to inducing delayed mating opportunities in Lucania goodei and, if there were, whether or not the ‘choosiness’ of female killifish would decrease as a result of absorbed costs. We found clear costs for delayed mating. Females incurred lowered egg viability and spawned fewer total eggs as the time between spawning events increased. ‘Choosiness’ measured by the proportion of eggs spawned with male L. goodei tended to decrease as the time between spawning events increased. Overripening is a potential cost to female preference because rejecting a male means increasing the time between ovulation and fertilization of the eggs. This is a novel cost to female mating preferences. The trends observed here provide valuable insight into understanding the life strategies of highly iteroparous, oviparous species such as Bluefin Killifish and the implications of overripening as a cost to mate choice.
DATA ACCESSIBILITY
Supplemental material is available at https://www.ichthyologyandherpetology.org/i2024007. Raw data and R scripts can be found at Dryad (https://doi.org/10.5061/dryad.xsj3tx9s3). Unless an alternative copyright or statement noting that a figure is reprinted from a previous source is noted in a figure caption, the published images and illustrations in this article are licensed by the American Society of Ichthyologists and Herpetologists for use if the use includes a citation to the original source (American Society of Ichthyologists and Herpetologists, the DOI of the Ichthyology & Herpetology article, and any individual image credits listed in the figure caption) in accordance with the Creative Commons Attribution CC BY License.
AI STATEMENT
The authors declare that no AI-assisted technologies were used in the design and generation of this article.
Treatment effects on the proportion of eggs that survived in the (A) first and (B) second experiments. Diamonds and error bars show the mean and 95% confidence intervals.
Treatment effects in the first experiment on (A) the total number of eggs spawned, (B) the average clutch size, and (C) the probability that a female spawned given the opportunity. Diamonds and error bars show the mean and 95% confidence intervals.
The effects of male species and treatment on (A) the total number of eggs spawned, (B) the average clutch size, and (C) the probability that a female spawned given the opportunity in the second experiment.
Treatment effects on the proportion of eggs spawned with L. goodei vs. L. parva. Diamonds and error bars show the mean and 95% confidence intervals. The null expectation of no preference (0.5) is excluded for the ‘every day’ and ‘every other day’ treatments but not for ‘every 4 days.’
Contributor Notes
Associate Editor: M. P. Davis.
These authors contributed equally.