Reproductive plasticity of Hawaiian Montipora corals following thermal stress

Ocean warming, fueled by climate change, is the primary cause of coral bleaching events which are predicted to increase in frequency. Bleaching is generally damaging to coral reproduction, can be exacerbated by concomitant stressors like ultraviolet radiation (UVR), and can have lasting impacts to successful reproduction and potential adaptation. We compared morphological and physiological reproductive metrics (e.g., sperm motility, mitochondrial membrane integrity, egg volume, gametes per bundle, and fertilization and settlement success) of two Hawaiian Montipora corals after consecutive bleaching events in 2014 and 2015. Between the species, sperm motility and mitochondrial membrane potential had the most disparate results. Percent sperm motility in M. capitata, which declined to ~ 40% during bleaching from a normal range of 70–90%, was still less than 50% motile in 2017 and 2018 and had not fully recovered in 2019 (63% motile). By contrast, percent sperm motility in Montipora spp. was 86% and 74% in 2018 and 2019, respectively. This reduction in motility was correlated with damage to mitochondria in M. capitata but not Montipora spp. A major difference between these species is the physiological foundation of their UVR protection, and we hypothesize that UVR protective mechanisms inherent in Montipora spp. mitigate this reproductive damage.


Results
The following comparative experiments explored the differences in some of the reproductive characteristics for two Hawaiian scleractinian corals-M. capitata, which has been substantially studied, and Montipora spp. which has been minimally studied-years after consecutive ocean warming and bleaching events. Reproduction data were collected during the summer spawning season of June to August/September of 2018 and 2019 from colonies of Montipora capitata and Montipora spp. found throughout Kāneʻohe Bay, Hawaiʻi (Fig. 1).
Overall, Montipora spp. appeared to perform better and more consistently than M. capitata in most of the reproductive metrics assessed: sperm motility and mitochondrial membrane potential, fertilization, and larval settlement (Table 1). We found that, comparatively, M. capitata improved in all assessed reproductive metrics from 2018 to 2019, except fertilization where it declined; settlement in 2019 was the lone metric where M. capitata performed marginally better than Montipora spp. Some preliminary spawning data of comparative sperm motility estimates were previously recorded during the summer of 2017 (Table 1). Whereas M. flabellata and M. dilatata have historically been classified as two species, there has not yet been a formal revision of Hawaiian Montipora taxonomy. Therefore, the Montipora spp. colonies for these experiments were originally visually identified and labeled as M. flabellata and M. dilatata, and those distinctions were maintained throughout the duration of the study. We initially compared fertilization success between combinations of crosses of M. flabellata and M. dilatata and found no difference in their ability to fertilize ( Supplementary Fig. S1). Based upon this and previous genetic analysis, the two species were combined to compare to M. capitata.
Eggs and sperm concentration per bundle. Fecundity can be used as a comparative reproductive health metric, and one commonly used method is to determine the number of eggs and concentration of sperm in each bundle that is released. With larger bundles, M. capitata has nearly 30% more eggs per bundle than Montipora spp. ( www.nature.com/scientificreports/ polyps/cm 2 ) is about three times greater than that of M. capitata (~ 25 polyps/cm 2 ), while the diameter of the corallite for Montipora spp. and M. capitata is about 1 mm and 2 mm, respectively. Since M. capitata has fewer but larger polyps per cm 2 , it's perhaps not surprising that it also produces larger egg-sperm bundles, with more eggs and sperm per bundle than Montipora spp.  www.nature.com/scientificreports/ There were other gross morphological differences, as well, related to the integrity or cohesion of the bundles. After the bundles were collected and placed in 15 mL tubes, the break-up time for bundles was also noted. In general, M. capitata seemed to have robust bundles that typically took about 45-60 min to break apart after initial spawning. By contrast, Montipora spp. bundles would frequently take less than 30 min to break apart to release the eggs and sperm. Egg volume. Egg size is another comparative reproductive health metric used as a proxy for energy investment (via lipids) into reproduction by the parent colony. Despite the smaller bundle size of Montipora spp., their individual egg volume was not appreciably different than M. capitata. The mean egg volume of M. capitata is greater than the mean egg volume of Montipora spp., but only with an approximately 4% difference (Table 1, Fig. 3; p = 0.0002, Welch's t = 3.76). This difference may be supported by greater variability in egg volume observed in Montipora spp. and its extended gametogenic and spawning cycle.
Sperm motility. One of the greatest differences between the species was observed in their sperm motility, with Montipora spp. having consistently more motile sperm, and in order to quantify sperm motility, a Computer Assisted Sperm Analysis (CASA) software system was used in conjunction with computer-aided video microscopy. Overall, the mean percent sperm motility differed for the two species in each year (Table 1, Fig. 4a). Montipora spp. total motility was nearly double that of M. capitata in 2018 (84% and 47%, respectively). Montipora capitata sperm had much higher total motility in 2019 (63%) but still lower than that of Montipora spp. (74%). The mean percent sperm motility for Montipora spp. was greater in 2018 than in 2019, while the reverse was true for M. capitata (Table 1, Fig. 4a). A two-way ANOVA determined that there was a significant interaction between species and year (F = 36.64, p < 0.0001). Species (F = 189.34, p < 0.0001) and year (F = 14.52, p = 0.0002) are both significant factors, but species had a stronger effect on sperm motility (Supplementary Table S2).
Additionally, there were differences in the longevity of sperm motility for M. capitata but not for Montipora spp. In 2018, M. capitata motility declined approximately 10% over one hour (Table 1, Fig. 5a, p = 0.0041, Welch's t = 3.03) but maintained its motility in 2019 ( Mitochondrial membrane potential. The key energy source for sperm motility resides in the mitochondria. Mitochondrial respiration, fueled by a pH gradient, yields a mitochondrial membrane potential that details the energetic state of the cells; low potential indicates less energy and high potential indicates higher energy and is correlated with sperm motility [64][65][66] . Therefore, flow cytometry analysis of the percentage of sperm with high mitochondrial membrane potential was also used as an indicator of healthy sperm cells. Mirroring motility, the patterns of mitochondrial integrity were similar with Montipora spp. (2018: 78%; 2019: 81%) having a greater percentage of sperm with high membrane potential than M. capitata (2018: 36%; 2019: 65%) for both years ( Table 1, Fig. 4b). A two-way ANOVA determined that species (F = 262.73, p < 0.0001), year (F = 61.46, p < 0.0001), and interaction between species and year (F = 25.8, p < 0.0001) were all significant factors (Supplementary Table S3).    Table S2 in Supplementary Fig. 4). The results of the two-way ANOVA for both motility and mitochondrial integrity indicate that the interaction of species and year was a significant factor, and species had a significant main effect for both. While year alone did not have a significant effect on sperm motility, it had a significant effect on mitochondrial integrity; however, given the relatively large F-ratios for species in both analyses, it is likely that species was the primary driver for differences in the analysis (Supplementary Tables S2 and S3). www.nature.com/scientificreports/ Correlation of sperm motility and mitochondrial membrane potential. There was a significant correlation between percent sperm motility and percent of sperm with high membrane potential for both species in each year, but the strength of that correlation varied for each season. For M. capitata in 2018, about 10% of the variance is shared between motility and membrane integrity (  Fig. 6d; R 2 = 0.3655, p < 0.0001). The quality of Montipora spp. sperm (motility and membrane integrity) outperformed that of M. capitata both years. Additionally, correlation of sperm motility with membrane potential for each species and year suggests that a significant positive relationship between the two exists, though the correlation appears to be relatively weak, indicating that other factors could have been influencing sperm motility.
Fertilization. Successful fertilization was examined using bundle-bundle crosses in scintillation vials.    Fig. 7b). A two-way ANOVA for settlement determined that both species (F = 7.947, p = 0.0066) and year (F = 15.79, p = 0.0002) alone had significant effects, and there was also a significant interaction between species and year (F = 17.14, p = 0.0001) on settlement (Supplementary Table S5). The settlement trials were performed in six well plates with a small chip of crustose coralline algae (CCA) as a settlement cue. Because M. capitata is found over a wider range of habitats than Montipora spp. (which seems to be restricted to areas of higher wave energy), settlement cues such as water movement, light intensity, and other physical and chemical parameters are not replicated in a stagnant six well plate and may help explain the low settlement of Montipora spp. Considering all the downstream effects that impact larval coral development and the variable habitat where it is found, it may be that M. capitata is less restrictive in its preference for settlement than Montipora spp., although a different experimental design would be needed to assess that than was utilized here.

Discussion
Montipora capitata is one of Hawaiʻi's main reef-building corals and is vital to reef habitat and growth 51 . Both Montipora flabellata and Montipora dilatata-Montipora spp.-are Hawaiian endemics that inhabit a much narrower ecological range than M. capitata 51,52 . See Fig. 8 for details. While we do not have bleaching histories for the colonies used here, we do have long-term sperm motility data for the M. capitata population in Kāneʻohe Bay that encompasses the years prior to bleaching 49 . With its relative scarcity in the bay, Montipora spp. is rarely used for reproductive studies, individual colony history is not typically tracked, and no similar long-term reproductive data set exists for Montipora spp. Therefore, it is possible that the colonies used here have differences in bleaching and/or recovery from the consecutive bleaching events of 2014 and 2015. www.nature.com/scientificreports/ Following bleaching, coral species differ widely-ranging from months to years-in their ability to recover from thermal stress and invest in sexual reproduction 27,33,39,67 . By increasing heterotrophy, M. capitata has previously been demonstrated to quickly recover lipid reserves after bleaching, likely allowing it to maintain fecundity, at least in terms of egg production 36,62 . Given the results of our experiments, certain aspects of M. capitata reproduction may be more sensitive to environmental perturbations and might take longer to recover than previously thought and, additionally, longer than Montipora spp. For example, it appears that Montipora spp. maintained higher quality sperm production more consistently and with less variation than M. capitata. It might be, then, that sperm production in M. capitata is more heavily influenced by sustained temperature extremes, UVR exposure, and/or other stressors than Montipora spp.
Many reproductive studies are done by means of histological analysis which can provide a snapshot of reproductive condition at the time of sampling but cannot clearly delineate the dynamics of reproductive physiology. Because oocytes are orders of magnitude larger than sperm and generally more energetically expensive 68,69 , studies following reproductive investment in corals after stress events have primarily focused on maternal colony effects and impacts to oocytes [25][26][27][28]36,37,70 . Analysis of sperm-if even examined-is typically limited to presence/ absence of mature testes, sperm per bundle, and/or histological measurements 26,28,37 . Using histological processes, Johnston, et al. 71 suggested that thermal stress in 2014 and 2015 in Kāneʻohe Bay had a greater and longer-term impact on oocyte development than on the testes in Pocillopora meandrina, but without active physiology measurements at spawning, it is difficult to fully assess the state of recovery.
One of the most susceptible coral reproductive metrics impacted by ocean warming is sperm sensitivity, exhibited by reduced motility 49,72 . Prior to 2014, M. capitata was commonly observed to have sperm motility greater than 70% to 90% 49 . When preliminary data were collected in 2017-two years after the sequential bleaching events of 2014 and 2015-sperm motility in Montipora spp. and M. capitata was 58%, and 42%, respectively (Table 1). In 2018, three years later, M. capitata sperm motility was still comparable to levels observed at the height of bleaching in 2015 49 , suggesting that its motility had not markedly recovered until 2019, four years after the consecutive bleaching events, yet still had not returned to levels observed prior to 2014 (Table 1, Fig. 4a).
Unfortunately, no similar long-term data for Montipora spp. reproduction exist for comparison, as this species has not been the target for this type of reproductive research, but after a relatively high preliminary motility assessment in 2017, Montipora spp. sperm motility outperformed M. capitata the following two years (Table 1, Figs. 4a, 5). Additionally, mitochondrial potential mirrored motility assessments in both 2018 and 2019, with M. capitata improving and Montipora spp. maintaining an already high level by comparison (Table 1, Fig. 4b). www.nature.com/scientificreports/ This suggests that if motility and mitochondrial membrane potential were impacted from bleaching in 2014 and 2015, Montipora spp. recovered more quickly than M. capitata, possibly in half the time, or perhaps it was never impacted at all by those events. Understanding sperm characteristics and physiology is important, especially in light of climate change's impact on and the response to reproduction. But in a larger context, are marine populations sperm-limited? This is largely going to depend on the environment, life history traits, and current population status of the taxon in question. As reviewed by Levitan and Petersen 73 , percent fertilization in marine organisms increases due to proximity, population density and size, and number of gametes released and declines with gamete dilution via current velocity. In his review, Yund 74 indicates that mobile and pair-bonding organisms, like some fish species, have very high fertilization success and are likely not sperm limited while benthic, sessile, and sedentary broadcasting invertebrates have some of the lowest rates of successful fertilization; however, there are documented cases of broadcasting invertebrates with high fertilization rates. Additionally, there are adaptations to enhance fertilization-synchronous spawning, sperm storage, larger egg size, sperm chemotaxis, and increased motile longevity-and mechanisms to prevent polyspermy 74 . Field measurements combined with in vitro fertilization assessments in corals, urchins, and sea stars have suggested that sperm limitation is a factor in some populations and could be due to limited sperm energy reserves, gamete dilution, proximity of conspecifics, and changes in sperm:egg ratio, but there are also life-history traits to reduce these effects [75][76][77][78][79][80] .
Considering coral reproduction and increasing ocean temperatures, Omori, et al. 72 reported a decline in sperm motility and fertilization success in Acropora corals following the 1998 worldwide bleaching event; sperm concentrations one and two orders of magnitude greater than prior to bleaching were required to achieve 80% and 95% fertilization, respectively, suggesting that decline in sperm performance rather than competency of eggs was the more significant change. By contrast, Armoza-Zvuloni, et al. 37 suggested that acclimatization and/or increased heterotrophy in the coral Oculina patagonica might explain why there was no difference in developed oocytes and testes between those that had repeatedly bleached and those that had not. However, development of gametes was only investigated histologically; no other metrics of reproduction after spawning were assessed in that study.
Even if populations across taxa were not historically sperm limited, anthropogenic disturbances through overharvesting, pollution, disease, climate change, etc. could potentially reduce populations to levels lower than their evolutionary history, making sperm limitation more likely 74,81 . A reduction in sperm motility accompanying a population decline could, conceivably, further limit successful fertilization and complicate recovery. If certain aspects of Montipora spp. reproduction are more robust than M. capitata, what might be contributing to that? The combination of high sea surface temperatures and solar irradiance likely exacerbate coral bleaching and/or sublethal stress 15,82,83 . Increased temperatures have a long history of detrimental effects on many aspects of reproduction, but UVR has also been implicated in negative impacts as well. For example, damage to coral gonads 84 , fish sperm chromatin 85 , and urchin sperm mitochondrial membrane damage, DNA and chromatin damage, and morphological abnormalities 86 all correlate with high UVR.
Shallow water corals inhabit areas of the reef with supersaturating irradiance and high exposure to UVR 87 . Corals in these waters are known to produce UVR photoprotective pigments in their tissues, chromoproteins and fluorescent proteins, and/or concentrate symbiont-produced UVR photoprotective mycosporine-like amino acids (MAAs) in their surface mucus [88][89][90][91] . UV-protective chromoproteins in Montipora spp. are found colony-wide in a thick band in the upper layer of epidermis, superficial to the nuclei and concentrated at the surface 48 . These chromoproteins are not found in other Hawaiian coral species such as Porites compressa, P. evermanni, Pocillopora meandrina, and M. capitata; interestingly, they are found in the polyps but not in the remaining colony tissue of M. patula 48 . Chromoproteins and fluorescent proteins are not very energetically costly and would need to be in abundance if serving as photoprotection 89 . The correlation of having higher concentrations of UVR-protective chromoproteins produced by the coral colony has, however, been linked with higher susceptibility to bleaching and mortality 92 , and Montipora spp. appears to be extremely sensitive to thermal stress, as observed during the bleaching events in 2014 48,93 . Additionally, Richards Donà 48 found Montipora spp. to be inefficient at removing sediments and histologically is significantly lacking in mucocytes compared to other Montipora corals, which could explain why it is restricted to areas of higher water flow. Montipora spp. symbionts were also enrobed with UVR-absorbing melanin which was not present in its congener, M. patula 48 .
Taken together, the lack of mucocytes producing very little mucous, likely relying less on the translocation of on MAAs for photoprotection, and instead incorporating endosymbionts enrobed with UVR-absorbing melanin and a thick band of superficial chromoproteins in the epithelium suggests a different strategy to mediate the supersaturated light conditions on shallow reefs 48 . Perhaps this alternative photoprotective strategy also is better at shielding gametes or their associated stem cells during and/or prior to gametogenesis. See Fig. 9 for details.
Given that it took M. capitata in Kāneʻohe Bay nearly four years to recover sperm motility levels comparable to those observed prior to bleaching, we speculate that UVR might be causing long-term damage to the reproduction of shallow water corals. Our hypothesis as to why reproductive characteristics in Montipora spp. were superior to M. capitata is that, while bleached, the coral-produced UVR protective chromoproteins observed in Montipora spp. might be guarding its stem and/or gametogenic cells from damage while that UVR protection, MAAs produced by its symbionts, is lost in M. capitata. See Fig. 10 for details.
Perhaps, then, a trade-off for Montipora spp. to bleaching sensitivity and frequent partial colony mortality is highly competitive rapid growth and increased vegetative (asexual) reproduction during periods with normal, lower temperatures and investment in UVR protection of gametes and/or stem cells. While bleached, Montipora spp. would presumably retain that UVR protection, thus protecting against any long-term DNA damage. By contrast, M. capitata appears to be more robust to bleaching events such that colony death might be less prevalent, but the loss of its symbionts possibly yields the loss of most of its UVR protection, potentially leading to higher prevalence of long-term cellular damage. The response to UVR, combined with the interaction www.nature.com/scientificreports/ Figure 9. Physiological mechanism of UVR mitigation and response to thermal stress, M. capitata and Montipora spp. UVR protection for M. capitata is derived from MAAs that are produced by its symbionts whereas Montipora spp. produces its own chromoproteins that serve as UVR protectant. Montipora spp. is also more sensitive to elevated temperatures while M. capitata appears to be more robust, possibly by increasing heterotrophic feeding. This strategy seems unlikely for Montipora spp. www.nature.com/scientificreports/ of temperature and bleaching, and its effect on sperm development, motility, mitochondria, and DNA damage should be investigated.
In order for sperm limitation to not hinder successful reproduction of benthic broadcast spawning marine invertebrates, fertilization is dependent upon, among other factors, population size, density, and spawning synchronization 79,94,95 . As one of the two most common reef-building corals in Kāneʻohe Bay, M. capitata is found in high density and across a variety of depths and reef habitats 51,52 . Furthermore, since M. capitata is highly synchronized in its gametogenesis and spawning behavior 59,60 and has more gametes per bundle than Montipora spp. (Table 1, Fig. 2), then its successful reproduction might not be impeded by a higher percentage of low motile sperm in high density populations.
Prior to anthropogenic disturbance increasing the frequency and severity of thermal stress events, these adaptations and life-history traits were likely well-suited for each species in its habitat. However if, as predicted, large-scale bleaching events continue to occur more frequently 19,24 , Montipora spp. colonies that experience high mortality might not have enough time to regrow to sufficient colony size to sexually reproduce, even though they may have the potential to recover more quickly. Likewise, if M. capitata takes four or more years to fully recover, frequently recurring warming events could feasibly negatively impact its capacity for successful reproduction if mortality pushes some populations to a point that they become limited by colony density whereupon a decline in sperm motile energy exacerbates the negative effects of gamete concentration on successful fertilization. As previously stated, most pre-zygotic studies of bleaching and coral reproduction detail oocyte development, size, and energy investment, with minimal attention to sperm, save histological measurements and counts. Yet arguably the most critical function of sperm physiology and evolution-its ability to swim-is rarely considered. Investigating and tracking both aspects of reproductive physiology and their response to environmental stress will allow for a more complete understanding of the capacity-or inability-of coral populations to adapt to life in the Anthropocene.

Methods
Animal care and gamete collection. In late May and early June 2018 and 2019, a 20-30 cm fragment from colonies of both species was collected from sites across the bay (21°25′57.21″ N, 157°47′16.96″ W). Because Montipora spp. inhabits shallow areas of the reef, colonies for both species were selected from a maximum depth of 3-4 m. Donor colonies were originally selected that were most likely to be of reproductive age, in a size range of approximately 1-3 m 2 . One M. capitata and one Montipora spp. donor colony were each closer to ½ m 2 , and the largest Montipora spp. colony was ~ 10-12 m 2 . However, because large, reproductive-sized Montipora spp. colonies are not common in Kāneʻohe Bay (the species is not found on every patch reef) and the species is not a frequent target of investigation, the study was limited to available colonies. Since finding Montipora spp. was the limiting factor, each time a colony was located, it was tagged along with a nearby M. capitata colony. Most of the Montipora spp. colonies are separated by dozens (and some hundreds) of meters from their nearest conspecific and were often on separate patch reefs in the bay.
Each Small powerheads in the tanks increased water circulation, and a 'surge tank' was added to the Montipora spp. holding tanks to further mimic their high wave energy environment. The colony fragments were housed in these tanks for the summer spawning season and were inspected daily for signs of stress, tissue loss, or lesions. In the rare event that a lesion did appear, an area approximately 2-3 cm deep into apparently healthy tissue and skeleton was removed, and the area was covered with two-part epoxy.
Two hours before their predicted spawning time at 21:00, each colony was isolated to allow independent collection of egg-sperm bundles. Because Montipora spp. spawning dates were uncertain, colonies were monitored nightly from late May through early September in 2018 and nightly in June and July in 2019. If a colony had not spawned by 22:00, it was placed back in the holding tank.
Even though animals were not subject to IACUC certification, colonies were checked every day, and all efforts were made to ensure high quality conditions in the tanks to reduce stress and support reproductive health. After the spawning season, coral fragments were returned to the original locations and secured back onto the reef using two-part epoxy. was measured by capturing images of approximately 30 eggs per colony on the night of spawning (microscope: Olympus SZX12, Center Valley, PA; camera: S01-0801A, software: SSView v. × 64, 1.0.5969, Motic Instruments USA, Schertz, TX). After the bundles had broken apart, the eggs were transferred to a new container with 0.2 µm filtered seawater and allowed to hydrate and expand for at least 30 min. To obtain the highest resolution possible, fewer than 10 eggs were captured in each image. The volume of the eggs was later measured using ImageJ processing software (v. 1.51a, Bethesda, MD) and a micrometer to set the appropriate scale. Eggs that were irregularly shaped or malformed were rarely found but also not used in the analysis.

Sperm motility with computer assisted sperm analysis (CASA). Sperm motility was assessed with
Computer Assisted Sperm Analysis (CASA) software using computer-aided video microscopy (CASA: Hamilton Thorne, Beverley, MA; microscope: Olympus BX41, Center Valley, PA). To prevent the need for multiple serial dilutions, egg-sperm bundle collection targeted the ideal sperm concentration (ca. 5 × 10 6 to 5 × 10 7 cells/ mL) for the specialized fixed depth, 20 µm counting chambers (Leja Products BV, The Netherlands). Transfer pipettes were used to carefully collect the bundles in minimal seawater, and samples were then diluted with filtered seawater to the desired volume. Hawaiian Montipora corals have a toxin (found both in the adult tissue and eggs) that can immediately kill sperm if an egg is ruptured in a small volume of water 96  . If fewer bundles were released (primarily for Montipora spp.), starting seawater volume was adjusted as needed to maintain the target starting sperm concentration. The bundles were allowed to break apart with minimal agitation, and sperm was transferred to a separate 1.5 mL Eppendorf away from the eggs. The CASA slides were loaded with 4 μL of each sperm sample, and per standard CASA software recommendation, at least five videos and a minimum of 200 sperm were recorded per sample to record total motility. If needed and depending on concentration, sperm was diluted in filtered seawater to a target of 1 × 10 7 cells/mL. Both initial and one-hour motility measurements were taken for most samples; a new slide of the existing sample was loaded for the one-hour reassessment. CASA parameters for all samples were modified from Zuchowicz et al. 97 . . Alongside CASA motility analysis, samples were diluted with filtered seawater, if necessary, to a sperm concentration of approximately 1-5 × 10 6 cells/mL for flow cytometry analysis, typically using 50 μL of the sperm sample and 450 μL of filtered seawater; 2.5 μl of JC-1 solution was added to 500 μL of sperm sample, incubated in the dark for 15 min, and then analyzed on a flow cytometer (BD Accuri C6 Plus, BD Biosciences, San Jose, CA). A negative control was made for each species for each night of spawning using 500 μL of sperm sample with 1 μL CCCP (carbonyl cyanide 3-chlorophenylhydrazone, C2759, Millepore-Sigma, St. Louis, MO). After at least 15 min with CCCP, the negative control was then incubated with 2.5 μL JC-1 in the dark for another 15 min and then also analyzed on the flow cytometer. The negative control was later used to assist with sample gating analysis to help identify the population of unhealthy or dead sperm cells. Unstained control samples (500 μL) for each individual coral that spawned each night were also run through the flow cytometer to ensure that there was no other auto-fluorescence interference. The flow cytometer has an argon laser with excitation at 488 nm, and channels measured fluorescence at plots on two wavelengths: green FL1 (533/30 nm) and orange FL2 (585/40 nm). The JC-1 dye will emit green fluorescence at low membrane potential and a higher proportion of orange fluorescence at high membrane potential. BD Accuri C6 Plus software (version 1.0.23.1, BD Biosciences, San Jose, CA) was used for sample gating and analysis. were conducted via bundle-bundle crosses: adding two egg-sperm bundles (one bundle each from two individual colonies) into 5 mL of 0.2 µm filtered seawater in 20 mL scintillation vials, yielding an approximate sperm concentration of 3-6 × 10 5 cells/mL. After the bundles broke apart, the contents were gently mixed and allowed to sit for 60 min at 26 °C, during which the total number of eggs per vial was counted. One hour after the bundles had broken apart, the eggs were rinsed twice in 10 mL filtered sea water to remove sperm and then filled with 15 mL filtered sea water. After three hours, the number of developing, fertilized eggs per vial were counted under a Wild M3 dissecting microscope at 10 × magnification. Selfing controls were performed by adding two bundles from the same colony to 5 mL of filtered seawater in a scintillation vial and analyzing the number of fertilized eggs after three hours.

Settlement.
Developing larvae were kept in glass bowls with ~ 100 mL filtered seawater at 26 °C and allowed to develop for 7-10 days; water was changed daily with 0.2 µm filtered seawater during development. For settlement, larvae were pooled and placed in six-well plates at a density of 10  . Each well was filled with 10 mL of filtered seawater, and a small, 1 cm chip of crustose coralline algae (CCA) was added to provide settlement cues. The CCA species used for settlement, H. reinboldii, can mediate larval settlement 98,99 and was collected near some of the donor colonies at less than 10 m depth. Because M. capitata can have a lengthy pelagic larval duration 98 , larvae were allowed several weeks to settle, and seawater was changed in each well every other www.nature.com/scientificreports/ day. Each well was periodically assessed using a dissecting microscope at 10 × magnification and spat recorded if settled. After one month, final settlement counts were taken.
Statistical methods. Where necessary, data were transformed using the appropriate transformation method in order to meet test assumptions, and normality assumptions were determined visually or with a Shapiro-Wilk's test. Additionally, the possibility that unequal sample size influenced the results in some experiments was eliminated by subsampling to similar sample size at random 10 times to confirm that results did not change. Data for sperm per bundle were transformed with a Box-Cox transformation. Eggs per bundle data were analyzed using a t test with Welch's correction, and sperm concentration per bundle was analyzed with a t test. Egg volume data were square-root transformed and analyzed with a t test with Welch's correction. The data for both sperm percent total motility and percent of sperm with high membrane potential were arcsine transformed for correlation analysis, and a two-way ANOVA was used to investigate the interaction between species and year with a Tukey's post hoc to determine differences among each. The comparison of the initial and one-hour motilities for species/year were each analyzed using a t test with Welch's correction. The M. capitata 2018 initial and one-hour motility motilities were square root transformed to better meet test assumptions.
Percent fertilization data were arcsine transformed, and percent settlement data were square root transformed. For both, a two-way ANOVA was used to investigate the interaction between species and year with a Tukey's post hoc to determine differences among each. Map was generated using Google Earth Pro v. 7.3.3.7786. Graph-Pad Prism 9 software (version 9.0.1; San Diego, CA, USA, www. graph pad. com) was used to analyze gametes per bundle, egg volume, initial and one-hour motilities, correlation analysis of sperm motility with membrane potential, and to create all graphs. Two-way ANOVAs for sperm motility, membrane potential, fertilization, and settlement were analyzed with R version 3.5.3 100 .