Introduction

Marine organisms are characterized by a remarkably diverse array of reproductive strategies (Mercier and Hamel 2009; Desjardins and Fernald 2009; Harrison 2011, Giese 2012). Sexual reproduction plays a vital role in producing genetically diverse offspring, generating adaptations to the environment, and enabling the persistence of species (Kondrashov 1988; Barton and Charlesworth 1998; Pianka 2011). Consequently, the reproductive characteristics of numerous benthic marine animals have been extensively studied (Olive 1995; Levitan 1996; Ramirez Llodra 2002; Mercier and Hamel 2009; Baird et al. 2009; Giese 2012). Most studies of reproduction in tropical coral-reef ecosystems have focused on scleractinian corals (Cnidaria: Anthozoa), commonly known as ‘hard’ or ‘stony’ corals (Richmond 1997; Harrison 2011; Levitan et al. 2011; Shlesinger and Loya 2019a; Randall et al. 2020; Baird et al. 2021; Lin and Nozawa 2023), largely because they are the main organisms contributing to the construction of reefs. However, in some places, a large share of the reef’s structure is built by species of the calcareous hydrozoan genus, Millepora (Cnidaria: Hydrozoa; Linnaeus, 1758), commonly called hydrocorals or ‘fire corals’ due to the strong stinging sensation they generate when touched. The ecological contribution of Millepora species as key habitat-forming organisms has been widely recognized and, accordingly, they are gaining increased attention (López et al. 2015; de Souza et al. 2017; Dubé et al. 2017, 2020a, 2021; Arrigoni et al. 2018; Boissin et al. 2020; Ortiz González et al. 2021; Garrido et al. 2022). However, fundamental knowledge of the ecology and biology of fire corals has remained limited, particularly regarding their population dynamics and reproduction.

The ongoing rapid decline in scleractinian coral cover worldwide (Knowlton 2001; Hughes et al. 2017; Sully et al. 2019) and the restructuring of many reef species assemblages (Graham et al. 2014; Stuart-Smith et al. 2018) continues to occur due to local and global stressors (Hoegh-Guldberg et al. 2007; Hughes et al. 2017; Donovan et al. 2021). Consequently, the ratio between scleractinian corals and fire corals may shift towards a higher relative abundance of fire corals in some reefs (e.g., Cramer et al. 2021). Despite several reports on fire coral susceptibility to bleaching (van Woesik et al. 2011; Fitt 2012; Smith et al. 2014), several authors have suggested that as the abundance of scleractinian corals declines, the role of Millepora species in the overall coral-reef ecology may be increasing (Brown and Edmunds 2013; Kayal and Kayal 2017; Dubé et al. 2020b).

Like most other reef-building corals, Millepora species are colonial cnidarians that dwell in a symbiotic relationship with unicellular, photosynthetic endosymbionts of the family Symbiodiniaceae (Lewis 2006; LaJeunesse et al. 2018; Dubé et al. 2020b; Garrido et al. 2022). Millepora species occur worldwide throughout the tropical seas and have been usually documented in shallow reefs (Lewis 2006; Dubé et al. 2020b). As hydrozoans, fire corals exhibit the most complex life cycle of all corals, including the benthic stage of a zooid, the basic repetitive unit of hydrozoan colonies, and two pelagic stages—a medusoid and a planula-larva (Lewis 2006; Dubé et al. 2020b). Each colony is gonochoric, with individual colonies releasing either male or female medusoids into the water column, where they shed their gametes and external fertilization occurs, emphasizing the importance of intra-species breeding synchronization (Soong and Cho 1998; Bourmaud et al. 2013; Shlesinger and Loya 2019b).

Our study sought to establish a foundation for understanding the essential life-history traits of fire corals and fill some fundamental knowledge gaps regarding the reproductive characteristics of these key reef-building organisms. In this process, we also endeavored to examine their ecology (e.g., abundance and size-frequency distribution) in light of their unique reproductive biology. We examined all three Red Sea Millepora species in the Gulf of Aqaba/Eilat (GoA/E), northern Red Sea: Millepora dichotoma, M. exaesa, and M. platyphylla (Arrigoni et al. 2018). Here, we present and discuss the following: (i) field surveys quantifying the size-frequency distribution and abundance of fire corals across depth and time; (ii) extensive in-situ monitoring of the breeding events of Millepora species from June to August over six consecutive years; (iii) histological examination of the two species over a period of 13 months, characterizing their annual reproductive cycle and sex ratio; and (iv) fertilization and settlement experiments of M. dichotoma to explore whether selective settlement under different light regimes may explain the differential distribution of this species at different depths. Our goal was to address key knowledge gaps pertaining to the life history and reproductive biology of these important, yet relatively understudied reef-building fire corals.

Materials and methods

Study area

All sampling and surveys for this study were performed ca. 3 km north and south of the Eilat Coral Beach Nature Reserve (29°30′30.0"N, 34°55′20.6"E) in the northern Red Sea (Online Resource 1, Figure S1). Daily sea-surface temperature (SST) measurements for 2019—2020 were provided by the Israel National Monitoring Program in the GoA/E, accessible at http://www.meteo-tech.co.il/eilat-yam/eilat_download_en.asp.

Abundance and population structure

We used three survey methods to quantify the different ecological aspects of Millepora populations: (i) size-frequency distribution, (ii) abundance over time, and (iii) distribution across depth.

To quantify species size-frequency distributions, in August 2021 we surveyed 49 belt transects (10 × 1 m) parallel to the shoreline, where nine transects were haphazardly deployed at a depth of 0.5 m, nine transects at 0.5–4 m, 18 transects at 4–10 m, and 13 transects at 10–20 m. The orthogonal longest axes of all colonies (n = 210) of M. dichotoma (length and height) and (n = 379) M. exaesa (length and width) were measured to the nearest cm up until the few colonies that reached 100 cm or more. Colonies larger than 100 cm in diameter were grouped together.

To examine the possible long-term persistence in Millepora abundance, we compared surveys conducted in the same reef in 1969 and 2021. In 1969, seminal quantitative underwater surveys were conducted using the line-transect method (Loya and Slobodkin 1971; Loya 1972). In August 2021 we used a similar methodology and counted every colony of M. dichotoma and M. exaesa along 16 line transects at 0.5 m, 15 transects at 0.5–3 m, and 15 transects at 3–6 m. During the 1969 surveys the two species were considered as different morphs of the same species (i.e., M. dichotoma), we therefore also pooled them in the current study.

To assess Millepora distribution along depth (0–40 m), we used 50 randomly chosen photo-quadrats per six depth stations (i.e., 2, 5, 10, 20, 30, and 40 m depth) taken during our earlier study in 2017 (see Shlesinger et al. 2018 for more details).

Reproductive cycles

To quantify and assess the reproductive patterns of Millepora species in the GoA/E, extensive in-situ observations were conducted every night for six consecutive breeding seasons (June–August 2016–2021). Observations included examinations of hundreds of fire coral colonies each night, which began at or before dusk and lasted at least 0.5–2 h, during which the timing and number of colonies releasing medusae were documented. Breeding events of 15 or more colonies releasing medusae were considered large-scale breeding events within a population, based on preliminary observations in which when we reached a count of ca. 15 colonies releasing medusae, practically every colony we observed on the reef released medusae. All medusae release events were categorized as either major or minor events for further examination of the intensity and distribution of breeding events. To differentiate between "major" and "minor" events, an analysis of the distribution of the number of breeding colonies for each species was conducted, and the thresholds for distinguishing between the two were determined to be the median values. For M. dichotoma major events, the threshold value was set at 13 or more breeding colonies per night, whereas in M. exaesa and M. platyphylla, this threshold was set at six or more colonies per night (Online Resource 1, Figure S2). Minor events were those in which at least one colony released medusae, but the overall number of these colonies reached values below the respective threshold.

To provide further insight into the gametogenesis and medusogenesis of fire corals, we collected ca. 5 cm samples monthly from both M. dichotoma and M. exaesa colonies (n = 146 and 100, respectively) over a 13-month period, between August 2019 and October 2020, and performed histological sections. Since these species exhibit multiple breeding events throughout one season, between June and August 2020, we sampled both species every 10 days. To increase the temporal resolution of the medusogenesis examination, five colonies of M. dichotoma were tagged immediately after the first observation of a breeding event in June 2021, and samples were repeatedly collected from these colonies every two or three days until the subsequent breeding event (ca. 14 days later). Immediately after collection, all samples were fixed in a 4% formaldehyde solution in seawater for 24–48 h, washed in running tap water, and preserved in 70% ethanol. The samples were then decalcified in a sodium citrate-buffered 25% formic acid solution for 12–24 h. The decalcified tissues were dehydrated, embedded in paraffin, cross-sectioned at 7 µm, mounted on slides, and stained with Mayer’s hematoxylin and Putt’s eosin (H&E stain) to highlight reproductive structures. Histological sections were examined under a Nikon Eclipse 90i microscope mounted with a DS-5 M camera, and NIS Elements D 3.2 software was used to measure and describe medusogenesis. To assess the reproductive state of each sample the following two parameters were measured: (i) mean medusa size, obtained by measuring the longest diameter of all medusae, and (ii) the number of medusae. Medusae identified in the fine temporal resolution samples (i.e., June-July 2021) were classified into two developmental stages with a size threshold of 300 µm, indicating developed medusae with distinguishable oocytes and sperm. To provide a standardized estimate of the reproductive potential of each sample (i.e., fecundity), all histological slides were scanned using an Epson V750 PRO scanner, and the area of each mounted sample was measured using Adobe Photoshop CS5 software. The number of medusae was standardized according to the tissue surface area of the histological section.

Identifying the sex of the medusae using histology was only possible when they were at an advanced stage and had developing gametes (i.e., medusae > ca. 300 µm). Therefore, the sex ratio was calculated based on a subset of samples. To increase the sample size for the sex ratio estimate, we used additional samples of M. dichotoma and M. exaesa collected during the 2017 reproductive season and processed as described above. A total of 28 samples were considered in the sex-ratio analysis of each species.

Fertilization and settlement experiment

To obtain planulae for the settlement experiment, we collected medusoids from 7–10 colonies of M. dichotoma in-situ in July 2021. Immediately thereafter, the medusoids were transferred to a 20 L aquarium filled with freshly filtered seawater (0.22 µm). Fertilized oocytes were present ca. 12 h (hrs) post-medusae release, and competent, swimming planulae appeared ca. 48 h post-fertilization. Thirteen beakers (2 L each) filled with 1500 ml of filtered seawater were placed in shaded aquaria with running seawater and maintained at a temperature similar to that of the natural environment. Once swimming planulae appeared, 125 planulae and one pre-conditioned terracotta settlement plate (3 × 3 cm) were placed in each beaker. Six beakers were further covered with a filter simulating the light regime at ca. 40 m (“Lagoon blue”, Lee Filters, see Eyal et al. 2016 for more details) and seven beakers without the filter served as a control. Every two days, freshly filtered seawater was added to the beakers, and salinity and temperature were monitored daily. The experiment began ca. 60 h after the initial medusae released into the sea, and 19 days later, all settled and metamorphosed planulae in all the beakers were counted.

Data analysis

Analyses were performed using R v4 (R Core Team 2021). All data were tested for normality, homogeneity of variances, variable dependencies, and model residuals and dispersion using visual inspections, the Shapiro–Wilk normality test, and Levene’s test for homoscedasticity. When the assumptions of parametric tests were not met, a permutation ANOVA was performed using the R package ‘predictmeans’ with 4999 permutations (Luo et al. 2014). Count data used in this study (i.e., the abundance data in 1969 and 2021 and the number of medusae per cm2 of tissue) were log-transformed for normalization. To examine the colony size-frequency distributions, we used the R package ‘moments’ (Lukasz and Novomestky 2015) and calculated skewness (g1), indicating the relative abundance of small and large colonies, and kurtosis (g2), indicating extreme values of the tail of the distribution.

Results

Abundance and population structure

The colonies of M. dichotoma were significantly larger than those of M. exaesa (permutation ANOVA, p = 0.0002; Fig. 1 and Online Resource 1, Figure S3). The mean colony size and abundance of both species decreased significantly with depth (permutation ANOVA, colony size: p = 0.0182 and p = 0.01 for M. dichotoma and M. exaesa, respectively; abundance: p = 0.0002 for both species; Online Resource 1, Figure S3). Both species exhibited positive skewness at all depths, i.e., a right-tailed distribution of size structure, indicating a larger number of small-sized colonies (Online Resource 1, Table S1 and Figure S3). The comparison of the historical abundance of Millepora with contemporary data at the study site did not show a significant difference at depth stations of 0.5 m and 3–6 m. In contrast, at the 0.5–3 m station, there was a significant decrease of 61% in the mean number of colonies per transect between 1969 and 2021 (ANOVA, p = 0.015; Fig. 1c).

Fig. 1
figure 1

Size-frequency distribution of (a) Millepora dichotoma and (b) M. exaesa colonies measured between 0.5–20 m depth (n = 210 and 379 for M. dichotoma and M. exaesa, respectively). Dashed lines indicate the median colony size. (c) Comparison of Millepora abundance (per 10 m line transect: n = 44 and 46 in 1969 and 2021, respectively) at the study site over five decades. Note that the data for M. dichotoma and M. exaesa are pooled, as in the 1969 surveys they were not distinguished as different species. Asterisk indicates a significant decrease. Lower and upper boxplot limits indicate the 25th and 75th percentiles, respectively, and center lines indicate the medians

Full size image

Reproductive cycles

The development of medusae and gametes occurred over 19 days, and the analysis of histological sections sampled monthly in 2019–2020 revealed that they were present in the tissues of both M. dichotoma and M. exaesa only between June and August. Accordingly, in all the other months, reproductive structures were completely absent (Fig. 2a). The medusae were randomly distributed within the tissue and two medusae cohorts developed next to each another (Fig. 3a and 3c). M. dichotoma and M. exaesa medusae exhibited maximum diameters of 667 µm and 402 µm, respectively. These medusae developed inside empty ampullae cavities found after a breeding event. Additionally, distinguishable oocytes and sperm were only noted in medusae > ca. 300 µm in diameter in both M. dichotoma and M. exaesa, with no advanced development of one sex before the other.

Fig. 2
figure 2

Temporal patterns of medusogenesis cycles. (a) Annual pattern of medusae development in Millepora dichotoma and M. exaesa indicated by blue and pink boxplots, respectively. N = number of sampled colonies and n = number of measured medusae. Lower and upper boxplot limits indicate the 25th and 75th percentiles, respectively, and center lines indicate the medians. (b) Fine temporal resolution of medusae development between two consecutive breeding events (i.e., within 13 days) of M. dichotoma. Violin plots indicate the distribution of the data (as probability density), highlighting the relative quantities of medusae of different sizes. Red points indicate the mean medusae size (µm) and n = number of measured medusae

Full size image
Fig. 3
figure 3

Key phases in the development of Millepora dichotoma medusae, gametes, and early life-history stages as seen in (a–e) histological sections, (f) confocal microscopy, and (g and h) stereoscope images. Panels (a and b) depict male medusae and (c—f) depict female medusae; (g) zooxanthellate planula 24 h post fertilization; (h) primary zooid (following planula settlement and metamorphosis) ca. 48–96 h post-fertilization. The medusae depicted in (a-d) are incorporated into the hydrocoral tissue while in (e and f) they are in a free-living state. Letters and arrows in the figure indicate different structures as follows: A ampulla; DM developed medusa; EA empty ampulla; EM: early medusa; GS gastrodermis; N nucleus; O oocyte; S spermaries; SC supporting cells; TB tentacle bulb; Z zooxanthellae

Full size image

The fine-temporal resolution histological analysis of tagged M. dichotoma colonies from July 2021 revealed that between two consecutive breeding events occurring within the same breeding season, two medusae developmental stages occurred: (i) a small medusae cohort (hereafter referred to as stage I), which continued to develop throughout the fine-temporal sampling (Fig. 2b and 3a-d); and (ii) a large, presumably older, medusae cohort (hereafter referred to as stage II). Stage I medusae were already present on day one of the fine-temporal sampling, and therefore they had likely commenced their development at least a few days prior to the previous breeding event. A conspicuous difference in medusae size, dividing the medusae population into two groups, significantly intensified at day five following the breeding event (Fig. 2b) (permutation ANOVA, p < 0.001 for days 5, 8, 10, 11, 12, and 13 and p = 0.057 for day 3; day 1 had no stage II medusae and therefore is excluded from the analysis). At day five post-breeding, the sizes of stage I and stage II medusae were 123 ± 26.4 µm and 340 ± 72.5 µm (mean ± SD), respectively. Moreover, the two timeframes (before and after day 5) differed significantly in the overall number of medusae per cm2 of tissue, presenting 29 ± 3.5 and 137.2 ± 40.5 medusae (mean ± SD), respectively (ANOVA, p < 0.0001). Overall, these findings suggest that a full development cycle of one medusae cohort spans a period of up to three weeks, from their initial development until their release.

The Millepora species of the GoA/E showed a distinct pattern of temporal reproductive isolation, with each species breeding on different nights during the breeding season (Fig. 4). M. dichotoma bred most frequently, with four to six breeding events during a single breeding season, while M. exaesa and M. platyphylla bred only once or twice per season. Each breeding event for all three species usually spanned two to five nights, with the initial and final nights featuring minor occurrences, in which only a few colonies released some medusae. During the course of two consecutive nights in the mid-period of each event, numerous colonies of the same species simultaneously released large quantities of medusae, leading to highly synchronized major breeding events.

Fig. 4
figure 4

Reproductive phenology of fire corals. (a) Breeding patterns of the three Red Sea Millepora species during 2016–2021. Species are indicated by different colors and their gradients correspond to the number of colonies observed to release medusae. Gray bars indicate nights when no observations were made. Solid and dashed red lines indicate full and new moon nights, respectively. (b) The distribution of Millepora species breeding events across lunar days. Major events were considered when more than 13 colonies were observed to release medusae in M. dichotoma and more than six colonies were observed in the other two species (see Methods and Online Resource 1, Figure S2 for further details)

Full size image

The breeding patterns of M. dichotoma largely coincided with two opposite lunar phases, i.e., at both the new and full lunar phases. Most major breeding events of M. dichotoma were observed two days before to two days after the new or full lunar phases, but a few were observed around the third quarter (Fig. 4b). In contrast, M. platyphylla breeding events largely occurred around the first quarter, whereas most M. exaesa events occurred a few days prior to the new and full lunar phases. Additionally, the onset of medusae release appears to be strongly linked to sunset, and as the time of sunset shifts throughout the breeding season, the breeding events were also scheduled accordingly, with the duration of the medusae release also being shorter (Online Resource 2).

When examining the sex ratios, M. dichotoma exhibited a 1.0:1.3 female-to-male ratio and a Chi-square test did not reveal a significant deviation from the expected 1:1 ratio. In contrast, M. exaesa exhibited a 1.0:3.0 female-to-male sex ratio, which was significantly different from the expected 1:1 ratio (Chi-square test, n = 28, \(df=1\), \({\chi }^{2}= 7\); p = 0.008).

Fertilization and settlement

To examine the potential impact of the light regime on shaping Millepora depth distribution via its effects on the early life stages, we obtained medusae of M. dichotoma. The released gametes were fertilized and we obtained swimming planulae, which were examined under different light conditions as a proxy of depth. All developmental stages, including oocytes, possessed intracellular zooxanthellae (Fig. 3e). At ca. 12 h post medusae collection, fertilized oocytes were observed, and ca. 24 h post-fertilization peanut-shaped planulae were noticeable (Fig. 3g). Finally, at ca. 48 h post-fertilization, fully developed free-swimming planulae were apparent along with numerous first settlers (Fig. 3h). Subsequently, free-swimming planulae were recorded also 10 days post-fertilization. Although no planulae were recorded settling on the terracotta settlement plates inside the glass beakers, settlement and metamorphosis were recorded on all the planes and surfaces of the beakers. Finally, the settlement of M. dichotoma planulae did not significantly differ between light treatments (ANOVA, p = 0.13; Online Resource 1, Figure S4).

Discussion

Given the urgency of the biodiversity and climate crises, which have become increasingly apparent in recent times due to rapid environmental changes (Barlow et al. 2018), it is imperative to investigate the life histories and dynamics of key species that play a critical role in maintaining ecosystem stability. This urgency is particularly pronounced in coral-reef ecosystems, which are undergoing unprecedented degradation (Knowlton 2001; Hughes et al. 2017; Sully et al. 2019). A comprehensive understanding of the reproductive ecology and population structure of key reef-building species is essential to provide a basis for predicting potential shifts in coral-reef communities and their extensive consequences (Riegl et al. 2018; Kayal et al. 2018; Cant et al. 2021; Lachs et al. 2021; Shlesinger and van Woesik 2021). Through our examination, we sought to address current challenges and provide crucial insights into fostering resilience and resistance in coral-reef ecosystems. Although there is extensive knowledge of fire coral distribution and taxonomy (e.g., Boschma 1948; de Weerdt 1984; Meroz-Fine et al. 2003; Razak and Hoeksema 2003; Lewis 2006; Ruiz-Ramos et al. 2014; Arrigoni et al. 2018; Boissin et al. 2020; Dubé et al. 2020b), their life histories are still poorly studied, resulting in major knowledge gaps regarding their ecology and biology. Furthermore, the necessity of establishing a baseline dataset of the Red Sea Millepora species is accentuated when considering that these species might be endemic to the region, due to possible cryptic speciation between the Red Sea and Indian Ocean Millepora populations (Arrigoni et al. 2018). Our study sought to fill in some of these gaps by providing an in-depth examination of the reproductive ecology and population dynamics of the three Red Sea fire corals. Overall, we found that these taxa developed gametes within an extremely brief timeframe of ca. 14 days, which is much shorter than that of any other reef-building coral, which usually spans up to 9–10 months (Harrison and Wallace 1990). These reproductive characteristics, including rapid reproduction cycles and multiple breeding events within a single breeding season, indicate the potential persistence and resilience of Millepora species.

Investigating the population size structure can provide valuable insights into its stability and demographics (Bak and Meesters 1999; Lachs et al. 2021), particularly regarding recruitment, survival, longevity, and responses to environmental impacts (Meesters et al. 2001; Gilmour et al. 2013; Prasetia et al. 2020). Our findings revealed that the size-frequency distributions of both M. dichotoma and M. exaesa were positively or right-skewed (Fig. 1, Online Resource 1, Table S1 and Figure S3), which suggests a prevalence of small colonies. The presence of a right-skewed distribution typically signifies a “healthy” or stable coral population with adequate recruitment rates (Bak and Meesters 1999). This is in contrast to the populations of several scleractinian coral species in the area that were recently found to suffer from within-population spawning asynchrony, leading to low abundance of small colonies and left-skewed distributions (Shlesinger and Loya 2019c). While our findings of a high proportion of small-size fire coral colonies likely imply a large sexually-derived larval influx, it could also be attributed to frequent fragmentation that is common in branching Millepora species (Edmunds 1999; Lewis 2006; Dubé et al. 2017). As small coral colonies have traditionally been considered sexually immature until a certain size threshold is reached (Connell 1973; Harrison and Wallace 1990), a population that includes a large portion of small fragments might have a low overall reproductive output. However, Rapuano et al. (2023) recently demonstrated that large, sexually mature coral colonies that were fragmented into small sizes seemed to “remember” their age and maintain their reproductive viability. Considering the large number of small fire coral colonies we recorded, as well as the wide range of colony sizes and frequent fragmentation, we suggest that the populations of both M. dichotoma and M. exaesa in the GoA/E are robust and thriving. Nevertheless, when comparing the contemporary abundance of fire corals at a depth range of 0.5–3 m, which was historically called ‘the Millepora zone’ due to the clear dominance of fire corals in the area (Loya 1972), to that in 1969, we did find a significant decrease. Although this analysis is temporally constrained and may hinder the ability to capture potential fluctuations, previous studies have indicated a continuous decline over time in the same depth range, as the total number of Millepora colonies decreased from 76 in 1969 to 24 and 12 in 1973 and 2001, respectively (Loya 1975; Wielgus et al. 2003). Moreover, this decline parallels the decrease in scleractinian corals during those years at that site (Loya 2004), followed by an increase and relative stability in the years that followed (Shlesinger and Loya 2016; Shaked and Genin 2021). However, there was no decrease in the abundance of Millepora in the other two depth zones (i.e., on the reef flat and at ca. 3–6 m).

Previous studies have shown a high abundance of Millepora species in shallow reefs (Edmunds 1999; Andréfouët et al. 2014; Smith et al. 2014; Dubé et al. 2020a; Pancrazi et al. 2024); however, their abundance across depth gradients has not been commonly examined. As expected, and known for numerous localities throughout the world, our data shows that the abundance of M. dichotoma and M. exaesa decreased profoundly with depth (Online Resource 1, Figure S3). Additionally, Millepora colony sizes were found to decrease along the depth gradient as well. The settlement of scleractinian coral planulae may be influenced by the depth-dependent light regime (Mundy and Babcock 1998; Strader et al. 2015; Shlesinger and Loya 2021; Prasetia et al. 2022), and we therefore explored whether the same mechanism might explain Millepora distribution along depth. However, the ex-situ manipulation of light conditions did not change the settlement pattern of M. dichotoma planulae (Online Resource 1, Figure S4). Other factors may play a more significant role than light in determining the distribution of Millepora species at different depths. For example, chemical cues from benthic communities may either deter or induce larval settlement and metamorphosis (Ritson-Williams et al. 2009; Doropoulos et al. 2020). Another plausible explanation might be related to the potential absence of depth-compatible photosynthetic symbionts in adult colonies (Frade et al. 2008; Garrido et al. 2022), resulting in energetic constraints (Cooper et al. 2011), which consequently limit certain vital processes that are crucial for colony growth and survival.

Our long-term dataset of Millepora species breeding events revealed a persistent pattern of temporal reproductive isolation of the three Millepora species in the Red Sea, as recorded in other reef-building species in the area (Shlesinger and Loya 1985; Shlesinger et al. 1998), suggesting a convergent adaptation of community-wide reproductive pattern within the region. Consequently, our findings of persistent and within-population synchronized breeding events throughout the course of six years indicate that the different stressors (e.g., rising seawater temperature, local pollution sources, and an unusually severe storm) that characterized the area and time of this study (e.g., Kleinhaus et al. 2020; Rosenberg et al. 2022; Kochman-Gino and Fine 2023; Oron et al. 2023), did not affect the reproductive capability of Millepora species. This highlights the potential resistance of fire corals to some environmental impacts. Additionally, although previous studies did not find a relationship between breeding events of fire corals and lunar phases (Soong and Cho 1998; Shlesinger and Loya 2019b; Dubé et al. 2020b), our long-term dataset indicated that Millepora species tended to exhibit more breeding events during a specific lunar phase (Fig. 4b). Similarly, in other hydroid species, periodicities of oogenesis, ovulation, medusae release, and completion of oocyte meiosis, were correlated with a variety of environmental factors such as dawn, dusk, light intensity, and full moon (Roosen-Runge 1962; Ikegami et al. 1978; Honegger et al. 1980; Boero 1984; Kubota 1996; Genzano and Kubota 2003). Taken together, these results suggest that the temporal reproductive patterns of hydrozoan species follow a consistent, phylogenetically conserved pattern, as has been documented for other marine invertebrate groups (Ramirez Llodra 2002).

As hydrocorals, Millepora species exhibit features similar to those of scleractinian corals, such as the production of three-dimensional structures through skeletal secretion. Consequently, fire corals might be presumed to possess life-history properties similar to those of scleractinian corals. From an ecological standpoint, they are typically regarded as having similar functions. In contrast to the prolonged, month-long gametogenic cycles, most scleractinian species exhibit (Baird et al. 2009; Harrison 2011), we found that medusogenesis (and gametogenesis within the medusae) in both M. dichotoma and M. exaesa spanned only a few weeks (Fig. 2). This duration is comparable to that of other hydrozoan species, such as the well-studied freshwater Hydra (Tardent 1974). Furthermore, histological analysis of M. dichotoma and M. exaesa revealed supporting cell structures surrounding the developing oocytes (Fig. 3d), but not in the free-living medusa stage (Fig. 3e). These supporting structures are believed to be conductive to a large energetic reservoir and fast gametogenic cycles (Mangan 1909; Tardent 1974) through nutritional supply to the definite oocyte or absorption into the final ovum. Sexual reproduction is widely believed to require a substantial amount of energy. Consequently, there may be various trade-offs between different life functions, such as growth, regeneration, and reproduction (Charnov 1982, Kozłowski and Wiegert 1986; but see Rinkevich 1996). Considering this hypothesis, we suggest that the brief reproduction cycles found in Millepora species may offer a selective advantage in terms of energy utilization, thereby accounting for the relatively fast growth rates observed in fire corals as opposed to scleractinian corals (Lewis 1989; Attalla et al. 2011). Furthermore, we suggest that the fast medusogenesis and simultaneous development of multiple medusae cohorts of M. dichotoma enable multiple breeding events within one reproductive season (Fig. 4), resulting in a robust larval supply (e.g., Hock et al. 2019) and a high capacity for population replenishment. Our study demonstrated that while scleractinian corals and hydrocorals exhibit certain functional similarities and ecological roles, they display distinct life-history characteristics, particularly in their reproductive patterns.

Understanding the life histories, especially the reproductive biology, of key species that are ecosystem engineers, is essential for assessing ecosystem functionality (Giangrande et al. 1994) and estimating possible future trends (Kayal et al. 2018; Hock et al. 2019). This study highlights the unusually rapid reproductive cycles of Millepora species, which may be a crucial factor in enabling fire coral populations to achieve high abundance on the reef and to persist through either resistance or resilience to local and global stressors (Brown and Edmunds 2013; Smith et al. 2014; Kayal and Kayal 2017; Dubé et al. 2020b; Pancrazi et al. 2024). Moreover, we suggest that these life history traits might position fire corals as prominent reef-building organisms in future tropical reefs. Our study underscores the importance of extensive, long-term investigations into the reproductive activities of marine organisms to understand their phenology. However, the extent to which Millepora species will dominate or endure in contemporary rapidly changing environments remains uncertain, and further exploration is necessary to uncover their fundamental biology, ecology, and response to environmental shifts.