Circatrigintan instead of lunar periodicity of larval release in a brooding coral species

Larval release by brooding corals is often assumed to display lunar periodicity. Here, we show that larval release of individual Stylophora pistillata colonies does not comply with the assumed tight entrainment by the lunar cycle, and can better be classified as a circatrigintan pattern. The colonies exhibited three distinct reproductive patterns, characterized by short intervals, long intervals and no periodicity between reproductive peaks, respectively. Cross correlation between the lunar cycle and larval release of the periodic colonies revealed an approximately 30-day periodicity with a variable lag of 5 to 10 days after full moon. The observed variability indicates that the lunar cycle does not provide a strict zeitgeber. Other factors such as water temperature and solar radiation did not correlate significantly with the larval release. The circatrigintan patterns displayed by S. pistillata supports the plasticity of corals and sheds new light on discussions on the fecundity of brooding coral species.

could not find conclusive evidence for lunar periodicity in planulation, timing and shed of larval numbers from gravid S. pistillata colonies, whereas other studies 30,38 claimed to reveal such lunar periodicity in larval release. This study investigates this deliberation of circalunar periodicity in S. pistillata reproduction as displayed by shallow water colonies from Eilat.

Results
Larvae were collected from eight colonies and counted per colony per day (during 48-59 nights per colony, over a total period of 84 days). In total, 16,586 planulae were collected during 427 sampling sessions ( Table 1). The numbers of planulae released per colony per day varied greatly among the eight coral colonies (Fig. 1). Planula numbers varied between maximum peaks of 903 (colony #6) to 12 (colony #7) planulae caught on a single night from a specific colony (Fig. 2). The most gravid colony (#6) released on average 156.1 planulae per day (ranging from 5 to 903 on a collection night), whereas the least productive colony (#7) released on average 3.8 planulae per day (ranging from 0 to 12 on a collection night; Table 1). The average number of planulae collected per colony per day increased with colony volume (Pearson Correlation: r = 0.81; n = 8, p(1-tailed) = 0.007).
Auto-correlation analysis revealed three distinct patterns of peak larval release (Fig. 3). The first pattern was exhibited by colonies (#1) and (#2), which had a short peak-to-peak period of approximately 27 and 25 days, respectively. The second pattern was exhibited by colonies (#5), (#7) and (#8), which had a long peak-to-peak period of 34, 33 and 35 days respectively. The third pattern was exhibited by colonies (#3), (#4) and (#6), which did not display any repetition of planulae releasing peaks during the April-June reproductive period (Fig. 3).
Coral fecundity was high during the studied reproductive season. In April, 98.2% of the 109 collected samples contained planulae larvae, a trend that was repeated in the following months (97.1% of the 175 samples and 96.5% of the 143 samples for May and June, respectively). Coral colonies #1 and #2 released greater numbers of planulae as the season progressed, whereas the other six S. pistillata colonies released lower numbers of planulae as the season progressed (Table 2).  Table 1. Reproductive effort of eight S. pistillata colonies (#1-8; see Fig. 1) during the reproductive season (April-June). Ecological volume as described by Shafir et al. 46 . Cross-correlation analysis revealed no tangible link between S. pistillata larval releases and daily maximum solar irradiance or UV irradiance (Fig. 4). Cross-correlation analysis of the 5 periodic colonies showed coherent oscillations of S. pistillata larval releases with the lunar period and tidal range, characterized by a variable periodicity of 27 to 33 days and a variable time lag of 5 to 10 days after full moon depending on the colony (Fig. 5).
In total, peaks in larval release were not related to water temperature or solar irradiance. They were loosely associated with the lunar cycle, occurred during both neap and spring tide, and both before and after full moon (Fig. 6).

Discussion
This is the first time, to our knowledge, that the planulae release of individual coral colonies has been monitored, in situ, for a prolonged period (84 days). A key advantage of this time series approach is that it enables assessment of possible periodicities in planulae releases for individual brooding corals. All other studies on S. pistillata thus Figure 2. Daily collections of planulae for each of the eight coral colonies (a-h) of S. pistillata during April-July. Peaks are indicated with light grey highlights and are defined by number of planulae collected being greater than or equal to 50% of the maximum number of planulae collected on a single day during the study of that particular colony (see Table 1).     37 , which is comparable to the seasonal decline from April to June 2011 observed in our study (Table 2). Hence, our results are consistent with previous studies.
Of the eight coral colonies monitored in our study, three had no discernible periodicity and five had consecutive peak larval releases during the sampling period; two of these five coral colonies displayed a period of 25 and 27 days, while the other three coral colonies had periods of 33-35 days. This variation in the timing of planulae release does not support population-wide synchronization of planulae release to a zeitgeber. Despite earlier suggestions that water temperature and UV or solar radiation might be potentially important environmental drivers of larval release periodicity 2,17,31,40,41 , our results did not reveal such a relationship with these variables. It is possible that due to the limitations of field work and the high variability of planulation a local peak in larval release might have been missed during days when no traps were placed in the field.
Cross-correlation analysis showed a relation between the reproductive patterns of the periodic coral colonies and the lunar cycle (~29.5 days) as well as between reproductive patterns of the periodic coral colonies and the spring-neap tidal cycle (as captured by the variation in tidal range). The periodicities of these cross-correlations varied from 27 to 33 days, with a time lag of 5 to 10 days depending on the colony, and none of the coral colonies actually showed a ~29.5-day periodicity in larval release. Consequently, peaks in larval release gradually shifted phase with respect to the lunar cycle, and occurred during both spring tide and neap tide, and both before and after full moon (Fig. 6).
It could be argued that S. pistillata in the Gulf of Eilat (Aqaba) is loosely circalunar, having approximately 29.5-day peak larval release cycles when averaged over all periodic colonies (Fig. 5), but the large lunar phase variation (from well before full moon to new moon and spring to neap tide; Fig. 6) implies that the entrainment by an external zeitgeber, which is required for the presence of circalunar periodicity 25 , is lacking. At the moment this ca. 30-day cycle would therefore be more aptly described with the term circatrigintan 24 and within that definition the reproductive cycle still shows large variation among the colonies (25-to 35-day periodicity).
It is possible that the timing of sperm release and internal fertilization in this brooding coral are more tightly controlled by the lunar cycle. Few studies have touched upon the timing of sperm production and release in this species 13,39 and, to our knowledge, no study has been done on the possible periodicity of sperm release in this species. Hence, this possibility would require further study. If fertilization rates vary in sync with the lunar cycle but subsequent rates of larval development vary among the brooded planula larvae 13,19,42 , this developmental variation could also cause larval release at different times during the lunar cycle, which would offer an interesting explanation for the observed circatrigintan pattern of larval release.
The presence of some sort of free running endogenous clock, controlling the planulae output of shallow water S. pistillata in the northern Red Sea, might be a vestige from earlier times when the lunar cycle did act as zeitgeber. The intense light pollution at night in Eilat 43 , potential impacts of global changes causing shifts in reproductive seasonality 10 or perhaps a low evolutionary need for entrainment of larval release with the lunar cycle in the Gulf of Eilat (Aqaba) could be possible reasons why the reproductive cycle of this brooding species has been decoupled from the lunar cycle. This would also explain why some of the colonies do not have a pronounced periodicity in their larval reproductive output. The lack of a strict zeitgeber for larval release and the variation in circatrigintan patterns displayed by S. pistillata colonies sheds new light on the discussion on coral plasticity and fecundity, and the possible implications for other brooding coral species that have previously been thought to display a tight lunar periodicity.

Methods
S. pistillata planulae were collected from a coral reef in the Gulf of Eilat (Aqaba), located in front of the Interuniversity Institute for Marine Sciences (IUI) in Eilat, at a depth of 3-5 metres from 1st of April to 24th of June 2011. Planulae traps were placed at nightfall covering the coral colonies and collected the following morning with their contents, as described 30,37 . Eight large S. pistillata colonies (Fig. 1) were chosen haphazardly, situated no more than 20 metres apart. Initially only colonies #1-#4 were sampled in the period 1-13 April 2011. From the 14 th of April onwards all eight colonies were sampled with small breaks in between (4 × 1-day breaks, 1 × 2-day break, 1 × 3-day break and 2 × 6-day breaks). The traps for colonies #5-#8 experienced technical failures from 19 to 22 April, and therefore the data collected during that time were not taken into analysis.
The collected nets were retrieved from the coral colonies and immediately transferred on land to a cool box on wheels with seawater. The traps were re-suspended to minimize the time exposed to air and desiccation, and transported to a wet lab situated nearby (less than 1 minute away). Each trap was rinsed separately with seawater, and the collected planulae were flushed from each trap to a separate glass container. The planulae were counted using a stereoscope and a pipette.
Auto-correlation 44 was applied to establish the presence of periodic patterns in larval production. IUI data loggers collected data on site every 10 minutes, which are readily available via a website 45 . The data included water level (cm), water temperature (°C), UV irradiance (mmol m −2 sec −1 ) and solar irradiance (W m −2 ). Tidal range was calculated as the difference between high tide and low tide, maximum daily solar and UV irradiance were obtained from the irradiance data, and lunar period was calculated as: Cross-correlation was applied to investigate relationships between fluctuations in larval production and fluctuations in these environmental variables. Furthermore, we calculated daily solar insolation (kWh m −2 day −1 ), which was compared to peak reproduction timing. Images of the coral colonies were analysed using ImageJ software, and an ecological volume index was established for each colony, by approximating colonial structures to the shape of a half sphere 46 . Graphs were created in Sigmaplot 12.5 and statistical analysis were performed with SPSS 21. Figure 6. Compilation of reproduction peaks and concomitant changes of tidal range (green line, solid), average water temperature (black line, solid), daily solar insolation (red line, solid) and full moon phases (black line, dashed). The light grey areas comprise the reproductive peaks of all eight S. pistillata colonies (see Fig. 2). Full moon occurred on the 18 th of April, 17 th of May, and 15 th of June.