Reproductive natural history and successful juvenile propagation of the threatened Caribbean Pillar Coral Dendrogyra cylindrus

Predictable spawning times across months and years

For the threatened Caribbean pillar coral Dendrogyra cylindrus, we report consistent spawning times across three consecutive years and across five
total lunar cycles (Figure 2). As in other Caribbean corals, spawning was closely synchronized with the lunar
cycle and daily sunset times (e.g., 17],18]). Only two prior observations of D. cylindrus spawning are published, both from Florida: a single male colony was seen spawning
at 112 minutes after sunset, three nights after the early August full moon in 2006
9] and multiple males and females were seen spawning 95 to 110 minutes after sunset,
three and four nights after the early August full moon in 2012 10]. Our observations are consistent with these reports, though it appears that spawning
in Curaçao occurs slightly later relative to sunset. We did not conduct monitoring
in other months, nor did we monitor outside of nights 2 to 5 after the full moon.
This leaves the possibility that additional D. cylindrus spawning occurs outside of the window documented here.

Asynchronous spawning times of males and females

By tracking individual colony spawning times, we found that males consistently and
predictably spawned earlier than females (Figure 2), an observation reported by Neely and colleagues 10], but which was not yet confirmed with individual colony data and statistical support.
On some nights of our study, all observable male spawning ceased before any female
spawning was seen.

After males began spawning, we observed some females with open mouths (Figure 3D), which is a rare behavior in corals. We also observed females exposing their eggs
to ambient seawater though radial slits around the mouth (Figure 2D-G). As D. cylindrus colonies do not release eggs and sperm in synchrony, it is possible that some or
all eggs are fertilized in situ, prior to release. The release of recently-fertilized
zygotes or embryos is one variant of so-called “spermcasting,” a term that encompasses
any fertilization strategy in marine invertebrates in which free-spawned sperm are
used for internal fertilization 19]. For example, in some dioecious coral reef gorgonian species, free-spawned sperm
are used by females to fertilize oocytes internally; females either promptly release
these newly-fertilized embryos into the seawater or brood them externally on their
surfaces for a number of days 20]-23].

In scleractinian corals, researchers previously described males spawning before females
in the dioecious Caribbean species Stephanocoenia intersepta (Blushing Star Coral) and Montastraea cavernosa (Great Star Coral; 18],24]-26]). We previously presented photographic evidence supporting the possibility of internal
fertilization in S. intersepta26]. Further, when Hagman and colleagues collected eggs from females of both M. cavernosa and S. intersepta, they found surprisingly high rates of fertilization without adding any sperm 27]. This led the authors to suggest that these two so-called “broadcast spawners” were
fertilized internally. Based on our observations, D. cylindrus appears to have both morphological and behavioral traits that would enable internal
fertilization.

Regardless of the precise location of fertilization, the asynchronous release of gametes
by males and females has the potential to increase individual fitness. One of Thorson’s
rules is that, for dioecious marine invertebrates, males generally spawn before females
28]. By delaying reproductive investment until fertilization is likely, individual females
may improve their overall rates of fertilization. Individual males may also benefit
from spawning early and therefore dominating the available gamete pool under conditions
of sperm competition (e.g., 29]). The apparent benefits of asynchronous spawning are reflected in the wide diversity
of dioecious marine animal taxa, and even four genera of green algae, in which male
spawning is known to occur before female spawning (e.g., 28],30],31]).

In sum, D. cylindrus is one of many dioecious marine broadcasters that have adopted a fertilization strategy
other than synchronous male and female spawning. For now, the precise timing of fertilization
in D. cylindrus and the window of gamete viability in the water column remain to be determined. Because
there are currently no population genetic data from D. cylindrus, we cannot yet predict whether individuals of this species generally fertilize only
their very close neighbors or whether gametes have the potential to survive dispersal
and achieve fertilization over relatively long distances.

Successful propagation of a threatened coral species

In the lab, we achieved the successful propagation of D. cylindrus larvae to the primary polyp settler stage. In the field, we documented physical characteristics
for identifying gravid males and females on spawning nights. We also recorded the
first cases of September spawning anywhere in this species’ range, thereby demonstrating
that populations of this species can distribute reproductive effort across two consecutive
months (so-called “split spawning”). The known, region-wide reproductive season for
D. cylindrus now extends across three lunar cycles, from early August in Florida to late August
and late September in the southern Caribbean.

In our propagation efforts, we only achieved successful larval development after we
injected sperm underneath a watertight egg collection tent underwater. However, we
could not determine the definitive timing of fertilization for these embryos because,
once on shore, we combined approximately 30 eggs collected in this manner with additional
sperm and with approximately 30 additional eggs collected underwater without this
step. This was done to maximize gamete density and diversity in hopes that any egg would be fertilized. Ultimately, 28 of these 60 eggs in total underwent cell
division and developed into larvae. It therefore remains possible that we had nearly
100% fertilization from the tented colony underwater, and 0% fertilization from gametes
mixed on shore. Alternately, it is possible that lesser amounts of fertilization occurred
both underwater and onshore. For researchers attempting to rear D. cylindrus larvae, we recommend collecting sperm underwater in syringes and transferring this
directly to tented female colonies that either exhibit pre-spawning characteristics
(Figure 2D-F) or that have been observed spawning on previous occasions. It also remains possible
that researchers will succeed with the traditional method of mixing gametes onshore.

The paradox of the missing juveniles

We found that rearing D. cylindrus larvae in the laboratory was relatively easy due to fast development and a short
time to settlement competence. Further, settlers were surprisingly robust in the laboratory
setting relative to Orbicella and Acropora spp., two other genera of spawning Caribbean corals that are listed as threatened.
This presents a new paradox for the early life history of D. cylindrus: if settled primary polyps survive so well under relatively stressful conditions,
why are sexually-produced recruits absent in all large-scale Caribbean reef surveys
published over the past three decades? Given the fact that D. cylindrus extends its tentacles fully during the day, even small recruits (1 cm diameter) should
be easily distinguished from other species in the Meandrinidae family, including those
whose juveniles are similar to one another in appearance such as Eusmilia fastigiata and Meandrina meandrites.

If D. cylindrus settlers are not likely to be misidentified by researchers, what explains their absence
in surveys? Are colonies experiencing low or failed fertilization (i.e., suffering
from Allee effects due to mate limitation, because populations densities are below
a critical threshold)? Are embryos or larvae highly sensitive to eutrophication or
microbial attack in the water column? Are competent larvae missing a critical cue
for settlement? Do settlers face a pathogen, predator, or competitor that causes extensive
post-settlement mortality? Locating this population bottleneck is an important next
step for conservation. Encouragingly, the advances we report here should help to make
D. cylindrus a viable subject for research on coral early life history, and perhaps restoration,
provided that gametes can be collected in sufficient numbers and that good fertilization
rates are achieved. This is the first dioecious, spawning coral species in the Caribbean
for which larval propagation methods have been described.

Population biology of a threatened coral

Caribbean coral species suffer together through habitat destruction, overfishing,
eutrophication, sewage, pollution, disease, and global climate change, yet D. cylindrus garners heightened concern because its particular life history characteristics, limited
habitat preferences, and disease susceptibility together pose an additional threat
to its viability 2],3]. Its listing as a threatened species is not due to its historical rarity per se,
but rather due to these species-specific factors that disproportionately threaten
the continued persistence of individuals. The peculiar life history characteristics
of D. cylindrus do partly explain its low historical abundance, but more worryingly, these traits
then further magnify its conservation plight on modern reefs by limiting recruitment
and population growth. As far back as 1986, Szmant described the risk of local extinction
for D. cylindrus due to its small population size, the rarity of small colonies, a limited geographic
range, and the occasional occurrence of a lone colony in a vast area 8]. In addition, reproductive success is limited by dioecy 8],32], which reduces the number of potential mates relative to hermaphroditic species.
Fragmentation 33], slow growth 15],33],34], and a long lifespan can create populations with many genetically identical individuals,
among whom mating is impossible because fragments originating from one colony are
all the same sex. Over the long term, extremely low sexual recruitment rates 11]-14] also limit the introduction of new genetic diversity into a population.

We identified additional traits in D. cylindrus with potential consequences for population viability. Rapid development and fast
settlement competence may affect average dispersal distances 35], thereby affecting population connectivity and local extinction risk. Importantly,
our observations of fast development were not due to unusually high temperatures.
We conducted our experiments at approximately ambient August/September seawater temperature
in Curaçao. Over the past eight years in Curaçao, we have found that other spawning
species such as Acropora palmata and Orbicella faveolata do develop slightly faster at warmer temperatures, but neither of these species have
ever developed nearly as quickly as D. cylindrus did at the same temperature.

We observed some potential for long-distance dispersal in larvae that remained swimming
for over four days. However, given the species’ virtually undetectable recruitment
rates, this perhaps does more to explain the occurrence of lone colonies in vast reef
swaths (e.g., Puerto Rico; 8], Barbuda; [K. Marhaver, unpublished data]) than it provides evidence that populations
can rebound from local extinction or near-extinction on timescales relevant to ecology
and conservation.

The occurrence of split-spawning likely affords male colonies an additional lunar
cycle to produce gametes. However, the gametogenesis cycle for D. cylindrus females is reported to be three months long 8], meaning total annual population fecundity may not be increased by splitting reproduction
over two consecutive months. Rather, split spawning could potentially reduce population
viability if this results in gamete concentrations below the density required for
fertilization on a given spawning night 36]-38].

With D. cylindrus now officially listed as a threatened species by the U.S. Government, a history of
scientific neglect burdens the conservation planning process. We still have no data
on its size at sexual maturity or the relative contribution of sexual versus asexual
reproduction to population dynamics. It remains possible that many dense stands of
D. cylindrus are in fact made up of very few, or even single genotypes (e.g., 39]). Our knowledge gaps reduce the accuracy of population viability assessment, which
is difficult even in well-studied corals due to their clonality, coloniality, fragmentation,
and partial death 39]-43]. Given these life history characteristics, demographic surveys of D. cylindrus and the assumption of long-distance dispersal may easily contribute to overestimates
of population viability. This warrants a precautionary approach to conservation.