Unlike broadcast spawning corals, brooding corals reproduce by internally fertilising their gametes and brooding their larvae until release. The vast majority of scleractinian corals broadcast spawn, and consequently spawning corals have been the primary focus for reproductive studies. However, brooding corals are important reef builders and can be some of the most abundant corals within their communities. Here we investigated the timing of reproduction for two common brooding coral species at Scott Reef; Isopora brueggemanni and Seriatopora hystrix. We found that both I. brueggemanni and S. hystrix had protracted reproductive periods, but with slightly higher output around spawning times. We also noted asynchrony in the development of testes and egg maturity, with limited mature testes recorded at times when mature eggs were present, particularly in October. Our results demonstrate that these brooding species at Scott Reef, have mature gametes and planulae during many months of the year. Year round gametogenesis is a likely consequence of the year round favourable environmental conditions at these lower-latitude reefs, but with peak reproductive output around the times of broadcast spawning reflecting a common response by corals with different reproductive modes to the cues for reproductive timing.

The 2 modes of reproduction employed by scleractinian corals are broadcast spawning, where corals release their gametes for external fertilisation, and brooding, where eggs are fertilised and embryos are maintained internally until they reach the planulae (larval) stage of development. Broadcast spawning species typically have a single, longer gametogenic cycle per year (5–7 months), whereas brooding species have multiple short cycles per year (<3 months)^[1][2][3]. There are far fewer brooding coral species than broadcast spawning species, with approximately 83% employing broadcast spawning, 14% employing brooding and 3% using both modes (of the 428 species for which reproductive mode is known)^[3]. The reproductive cycles of the broadcast spawning species at the Oceanic Kimberley reefs of Scott Reef and the Rowley Shoals have been described, with mass spawning occurring biannually in spring (October) and autumn (March)^[4][5]. Broadcast spawning is thought to occur around these times to fit into optimal environmental windows for fertilisation and larval survival and settlement, during which there are warmer water temperatures and lower wind speeds^[6][7][8]. Although spawning corals are typically more abundant on most tropical reefs, brooding corals can also make a significant contribution to the community. For example, at the Oceanic Kimberley reefs, groups of corals that include brooders, such as Isopora and Pocilloporidae contribute to approximately 25% of the coral community^[4], yet little is known of their cycles of gametogenesis and times of planulae release.

Globally, there is not a clear trend in breeding season and duration across brooding coral species^[9][10]. However, periods of planulae release seem to be shorter at higher latitudes, and occur over long periods or year round in the tropics. For example, Acropora (Isopora) palifera releases once per year in spring at Heron Island (23°S), but releases all year round at Lizard Island (14°S)^[11]. Similarly, Pocillopora damicornis from Rottnest Island (32°S) releases planulae in the summer months, but releases over protracted periods or year round on lower latitude reefs of the Great Barrier Reef^[12][13] and Seriatopora releases planulae seasonally (summer months) at the higher latitude Red Sea ^[9] but releases year round in tropical Palau^[14]. It should be noted however that with Pocillopora damicornis, patterns in reproduction are further complicated by both brooding and spawning behaviour^[15] and recent subdivision of species^[16].

There is currently little information available on the timing of gametogenesis and planulation for brooding corals on most Western Australian reefs, including those in the Oceanic Kimberley, despite their being ubiquitous and important reef builders. Managing local impacts on reefs requires knowledge of the times of reproduction and recruitment, and it is also important to understand the role that corals with different reproductive strategies may play in reef renewal, particularly in a changing climate. For example, because brooding corals release planulae over several months of the year, as opposed to in a single event as with spawning corals, reproduction and recruitment in brooding corals may be less impacted by pulse disturbance events (such as coral bleaching) than broadcast spawning corals, which has implications for future changes in community structure. In order to assess how brooding corals may impact overall coral recruitment at Scott Reef, we describe the timing of gametogenesis and planulation in two common brooding corals at Scott Reef.

Of the colonies with gametes present, 57% of Isopora brueggemanni and 68% of Seriatopora hystrix colonies had multiple developmental stages of gametes at any one time, indicating that both species had multiple, overlapping gametogenic cycles. Consequently, the proportions of colonies with different stages summed to more than 100% (Figure 1B-E). Other studies on brooding corals have also reported multiple overlapping gametogenic cycles or continuous planulae release (including four Madracis species from the Caribbean, Acropora palifera from the Great Barrier Reef and Papua New Guinea and Stylophora pistillata, Seriatopora hystrix and Alveopora daedalea from the Red Sea)^{[11][10][2][17][18]}. Overlapping gametogenic cycles and asynchronous planulae release in brooders has been linked to polyp size and water temperature. Brooders with larger polyps are able to hold more planulae and consequently have longer brooding periods and more synchronous planulae release^[18], while yearly temperature cycles have also been shown to regulate planulation in Caribbean brooding corals^[17].

The results from Scott Reef align with other studies on tropical brooders that reported planulae release all year round^{[19][20][21][11]}. The seasonality of planulation in brooding corals varies with location and species, but generally there appears to be a shorter reproductive season at higher latitudes^[11]. At high latitude reefs such as Heron Island (23°S) and Rottnest Island (32°S), Pocilloporidae brooders only released planulae in the summer months^[13][22], with some colonies observed to have up to three cycles of maturing gametes and planulae per season ^[13]. The same species (Seriatopora hystrix, Pocillopora damicornis and Stylophora pistillata) reproduced all year round near the equator (Palau, 7.5°N)^[23][24][14]. This pattern is likely related to latitudinal and seasonal variations in temperature, light and colony energetics. For example, at Heron Island and Rottnest Island, water temperatures fall well below 20°C in winter, suggesting that winter water temperatures are too cold for reproduction^[22]. However, not all brooding corals from high latitude reefs have short reproductive seasons. For example, P. damicornis releases larvae only during the summer months at Heron Island (23°S), but at a similar latitude in Hawaii (24°N), the same species releases all year round^[20][22]. Biological drivers such as competition, may also influence the length and seasonality of reproduction. In the Red Sea, benthic algal cover increases significantly in the winter months, reducing settlement space for coral larvae and thus providing more favourable conditions for reproduction in the summer months when algal cover declines^[9]. Globally, there does not appear to be a consistent trend in breeding season and duration of release for brooding corals, and both appear to vary substantially between species^[9][10]. At Scott Reef, S. hystrix and I. brueggemanni had the protracted reproduction generally seen in tropical brooders, but also appeared to have higher output around broadcast spawning times.

S. hystrix and I. brueggemanni had increased proportions of mature gametes and planulation in the mass spawning months, in autumn (March) and spring (October) (Figure 1). Both species had mature eggs, mature testes and planulae in most of the months studied, with the exception of no mature testes in S. hystrix and no planulae in I. brueggemanni in February. Less maturity in February compared to March, as well as both species having fewer colonies with eggs and testes during April compared to other months, indicated that more colonies were reproducing around the March spawning event. There are advantages to timing the release of planulae to the months of mass spawning, including ideal environmental conditions for gamete maturation, larval settlement and survival^[6][7][8]. More synchronous release of larvae during spawning times may also improve survival due to predator satiation^[25][17]. There are several records of brooders releasing planulae on the night of mass spawning events, however these records come from other locations and refer to different species to those studied here. For example, mass planulae release by the brooding coral Madracis senaria, has previously been reported to coincide with mass broadcast spawning events in the Caribbean^[17]. Additionally, in some colonies of Goniastrea aspera in Okinawa, both brooding and spawning behaviour was observed within the same colony around mass spawning times^[26].

The apparent peak in gamete maturation and planulae release around the times of mass spawning at Scott Reef may reflect ancestral responses to the same proximate and ultimate cues influencing mass spawning in corals^[27][28][29]. An earlier study suggested that broadcast spawning was the ancestral mode of reproduction in corals, and brooding the derived trait^[2], however more recent phylogenetic studies have described a more complex evolutionary pathway, with the most probable path being gonochoric spawning to gonchoric brooding, then to hermaphroditic brooding and finally hermaphroditic spawning^[29]. It is also not uncommon for species to exhibit both brooding and spawning modes of reproduction^[15][26][3] and there are species that exhibit ‘pseudo brooding’ behaviour for a few hours to a few days^[17][18][30].

We also recorded some asynchrony in the timing and proportions of colonies with mature testes and colonies with mature eggs. In October, both S. hystrix and I. brueggemanni had fewer colonies with testes compared to eggs, with ~45% of colonies in both species having mature eggs, but only ~20% of I. brueggemanni and ~8% S. hystrix colonies had mature testes. However, in March, higher proportions of eggs and testes appeared to correspond with one another (Figure 1B-E). The onset of oogenesis precedes spermatogenesis by a few months in most corals, but both sexes usually reach maturity simultaneously^[1]. This greatly increases the chances of seeing only eggs in samples and is the most likely reason for many early studies recorded only eggs in their samples^[1]. Asynchrony in the initial stages of oogenesis and spermatogenesis partially explains the lower proportions of testes in our samples. However, it is interesting that we recorded high proportions of mature eggs in October for S. hystrix and I. brueggemanni with corresponding low proportions of colonies with testes. One possibility is that reproduction may be limited by sperm production. Similar asynchrony was observed in brooding corals in Panama, and the authors suggested that there was little advantage in synchronising sperm release since unfertilised eggs were not wasted (eggs were either reabsorbed or remained viable for later fertilisation)^[31].

These data provide insights into the seasons and months of gametogenesis and planulation. The various stages of development within single colonies indicated that multiple gametogenic cycles were occurring in these brooding species. However, further sampling at finer temporal scales would provide the additional information required to determine how many gametogenic cycles are occurring each year, when exactly sperm and planulae are released, and how closely aligned these times are with mass spawning events (e.g. the same night, week or season). These finer scale details have local management implications. For example, anthropogenic increases in sediment concentrations at the times of sperm or planula release can inhibit fertilisation and larval settlement^{[32][33][34][35]}. In coming decades, patterns of coral reproduction will increasingly be influenced by global climate change. It will be critical to determine whether the times for planulation and gametogenesis of brooding corals are shifting with increasing temperature and more frequent and severe disturbance events, and whether there are subsequent implications for coral community structure. Although the cycles of reproduction reported here were not influenced by acute disturbances, the frequency of disturbances is increasing at Scott Reef, as with most reefs around the world. Scott Reef experienced severe bleaching in 1998 and 2016 and moderate bleaching in 2010^[36][37]. Exposure to potentially damaging cyclones also occurred in 2004, 2005, 2007 and 2012^[38]. Broadcast spawning corals synchronise the release of their gametes to one or a few nights of the year. Although a highly successful reproductive mode, synchronised spawning also leaves corals more susceptible to acute disturbances around spawning time. At Scott Reef, autumn spawning corals are particularly vulnerable, releasing their gametes in months that have potential exposure to damaging cyclones and the warmest water temperatures for the year. Consequently, a severe disturbance around a spawning event could eliminate a significant proportion of the year’s reproductive output. In contrast, brooding corals release planulae over many months of the year, making them less susceptible to acute disturbances. Provided that brooding corals do not have higher sensitivity to these stressors, they may maintain a high reproductive output and be better placed to re-colonise a reef following a disturbance event. This, in turn, could facilitate an increase in abundance of brooding corals and a shift in community structure.

This study provides the first information on the reproductive timing of 2 common species of brooding coral from the Scott Reef atoll system off the northwest coast of Western Australia. Although our findings have mostly local relevance, they highlight that these species prescribe to the low latitude, protracted reproduction pattern seen in brooding corals in other tropical locations, but also appear to concentrate their reproduction around mass coral spawning events. This baseline information on brooding coral reproduction is essential to understanding the ecology of these systems, and may become even more important if reproductive patterns and community structure shift through the Anthropocene.

The isolation of the Scott Reef atoll system means the frequency of our sampling regime was limited, and we were unable to sample corals during each month of the year. Although we can see evidence of multiple gametogenic cycles (multiple stages of development within a single colony) we are unable to determine how many cycles and whether there were other peaks in reproduction during the months not covered here. Finer temporal scale sampling would also help to determine the times of sperm release, period of larval development internally, and the times of planula release relative to the lunar cycle and times of mass spawning.

In Western Australia, future research into the reproduction of brooding corals would benefit from sampling at a higher temporal resolution in order to determine the number of gametogenic cycles per year and planulae release dates relative to mass spawning. Globally, it is going to become increasingly important to determine who the winners and losers will be as disturbance regimes on coral reefs increase in frequency and severity. In particular, it will be critical to know whether corals with more protracted reproduction (including brooding corals), will be better positioned to re-seed reefs with new coral recruits than corals with more synchronised gamete release.

Study site and spawning observations

Scott Reef is an isolated system of reefs, located near the edge of the continental shelf, approximately 270 km off the northwest Australian coastline (Figure 1A). Two brooding species, Isopora brueggemanni and Seriatopora hystrix, were sampled across six sites at Scott Reef. Colonies were sampled in February, March, April and October of 2008 and 2009, with between 13 and 39 (median = 26) replicate colonies sampled per species, per month. Mass spawning occurred during March (autumn) and October (spring) in 2008 and 2009, and was not split over consecutive months within each season^[8]. Only sexually mature (>20 cm diameter) colonies were sampled and three branches were collected from each colony. Samples were collected from the centre of the colony to avoid sterile colony margins.

Reproduction was examined by staging the eggs using histological techniques. Decalcified tissues were dehydrated using graded ethanol, then cleared in chloroform and embedded in paraffin wax. Samples were sectioned into 6 micron slices, mounted on slides, and then stained using Harris’ Haematoxylin and Young’s Eosin (see Figure 1F and G for examples of histological sections). Stages of egg and teste development were ranked according to^[39] and^[40].

Woodside Energy Ltd (operator of permit WA33P) funded the study on behalf of the Browse Joint Venture and the Australian Institute of Marine Science (AIMS).

We thank E. Gates for processing laboratory samples, K. Cook, M. Case and D. Ceccarelli for assisting with data management, Z. Richards for taxonomic assistance and K. Cook, K. Fitzgerald, D. McKinney, M. Travers, A. Heyward and the crew of the RV Solander for providing assistance in the field. Woodside Energy Ltd (operator of permit WA33P) funded the study on behalf of the Browse Joint Venture and the Australian Institute of Marine Science (AIMS). Woodside Energy Ltd had no role in the data analysis, data interpretation, the decision to publish or in the preparation of the manuscript. T. Foster's position at AIMS is supported by the Woodside Energy Ltd Coral Reef Research Fellowship.

The authors declare no conflicts of interest.

Not applicable.

No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.