Demographic dynamics of the smallest marine vertebrates fuel coral reef ecosystem functioning

See allHide authors and affiliations

Science  21 Jun 2019:
Vol. 364, Issue 6446, pp. 1189-1192
DOI: 10.1126/science.aav3384

Little fish make a big contribution

Coral reefs represent one of the most biodiverse and rich ecosystems. Such richness conjures up images of coral heads and large colorful reef fishes. Brandl et al. show, however, that one of the most striking and important parts of the reef ecosystem is almost never seen (see the Perspective by Riginos and Leis). Small cryptobenthic fish, like blennies, make up nearly 40% of reef fish biodiversity. Furthermore, the majority of cryptobenthic fish larvae settle locally, rather than being widely dispersed, and have rapid turnover rates. Such high diversity and densities could thus provide the biomass base for larger, better-known reef fish.

Science, this issue p. 1189; see also p. 1128


How coral reefs survive as oases of life in low-productivity oceans has puzzled scientists for centuries. The answer may lie in internal nutrient cycling and/or input from the pelagic zone. Integrating meta-analysis, field data, and population modeling, we show that the ocean’s smallest vertebrates, cryptobenthic reef fishes, promote internal reef fish biomass production through extensive larval supply from the pelagic environment. Specifically, cryptobenthics account for two-thirds of reef fish larvae in the near-reef pelagic zone despite limited adult reproductive outputs. This overwhelming abundance of cryptobenthic larvae fuels reef trophodynamics via rapid growth and extreme mortality, producing almost 60% of consumed reef fish biomass. Although cryptobenthics are often overlooked, their distinctive demographic dynamics may make them a cornerstone of ecosystem functioning on modern coral reefs.

How coral reefs maintain high diversity and productivity in oligotrophic tropical oceans—often termed “Darwin’s paradox”—remains poorly understood (13). Both pelagic subsidies (2) and internal energy and nutrient cycling (3) have been put forward as drivers of productivity on reefs, but their relative importance has not been resolved.

Fishes form the largest reservoir of consumer biomass on reefs and are involved in most internal energy and nutrient fluxes (4), but they also bridge the pelagic zone–reef interface during their larval development (5) and may therefore subsidize reefs with pelagic productivity (6). Over the past two decades, reconstructions of larval dispersal pathways (7, 8) have revealed the importance of larval dynamics for reef fish ecology, evolution, and conservation (911). However, the role of larval stages in coral reef ecosystem functioning remains virtually unknown.

Integrating published surveys of coral reef fish larvae, new data on adult reef fish communities, and a theoretical population model (12), we demonstrate that cryptobenthic reef fishes [a group of 17 reef-associated fish families characterized by species <50 mm in length (12, 13)] play a previously unrecognized but critically important role in coral reef ecosystem functioning because of their larval dynamics, rapid growth, and extreme mortality.

Cryptobenthic larvae greatly outnumber large reef fish larvae near coral reefs (Fig. 1) despite limited adult reproductive outputs. As more gametes produce more larvae, family-specific larval abundance predictably increases with gamete output for both cryptobenthic and large reef fishes [calculated from adult densities, fecundity, and spawning frequency (12, 14, 15)]. However, the respective slopes differ drastically, with the relationship for cryptobenthics being more than an order of magnitude steeper than that for large reef fishes (Fig. 2 and table S1), and this difference is consistent across the locations studied (Australia, Belize, and French Polynesia). We found no statistical evidence for effects of pelagic larval duration (i.e., the time that larvae spend in the plankton) or broadcast (i.e., eggs released into the water column) versus benthic clutch spawning (i.e., distinct clutches of guarded eggs) on larval supply (table S2) despite the high prevalence of benthic clutch spawning in cryptobenthics. Furthermore, the difference in slopes was not affected by the uncertainty in the input data, including increases in spawning frequency for cryptobenthics (fig. S2 and tables S3 and S4) and uncertainties in the body mass–fecundity relationship, adult densities, and larval composition (figs. S3 to S5). Thus, cryptobenthics are disproportionally more successful at converting adult gametes into larval supply, which likely underpins their extreme abundance in the nearshore ichthyoplankton.

Fig. 1 Global dominance of cryptobenthic reef fishes in the near-reef ichthyoplankton.

(A) Cryptobenthic larvae (blue) account for two-thirds (65.7%) of the larval reef fish pool <10 km from reefs, whereas large reef fish larvae (light gray) dominate >10 km from reefs. Crossbars represent predicted medians (±95% credible intervals) from a Bayesian beta-regression model; circles represent the raw data (12). (B) The high proportion of cryptobenthics (colored pie slices) in the near-reef ichthyoplankton is consistent across major biogeographic coral reef regions. Dots represent separate studies. (C) Average contribution of reef fish taxa to global near-reef ichthyoplankton. The three highest contributing taxa are cryptobenthic, represented by photographs of adult Eviota infulata (Gobiidae) (D), Scartella cristata (Blenniiformes–Blenniidae) (E), and Cheilodipterus quinquelineatus (Apogonidae) (F). Families contributing <0.1% were omitted for clarity.

Fig. 2 Differences in the relationship between larval supply and adult gamete output for cryptobenthic and large reef fishes.

Dashed lines and ribbons represent predicted fits from a Bayesian beta-regression model [±95% credible intervals (CIs)]; circles (broadcast spawners) and diamonds (demersal brooders) represent raw data averaged across three sampling locations (±SE). Both axes represent proportional shares.

The abundance of cryptobenthic larvae, despite limited gamete output, may provide the continuous (16) and copious inflow of larvae that is assumed necessary for adult population maintenance in small, short-lived reef fishes (17). Our results show that this, in turn, forms the basis of a critical energy and nutrient pump that operates across the pelagic zone–reef interface and may help to explain the enigmatic productivity of coral reef ecosystems (Fig. 3). Using a demographic model that simulates the daily arrival, growth, and death of larval recruits from all reef fish families over 1 year (fig. S6), we estimate that juvenile and adult cryptobenthics provide most (57.5 ± 0.1% SE) of the consumed fish biomass on reefs. However, because of extreme mortality rates and small sizes, cryptobenthics appear to make negligible contributions to net productivity and standing fish biomass (Fig. 3C), which mirrors empirical evidence (18). Standing biomass is the most commonly quantified metric of ecosystem functioning on reefs (19); however, it does not capture the rapid turnover in cryptobenthic populations (693.1 ± 2.7% SE annually) that is enabled by their extraordinary demographic dynamics. Thus, cryptobenthics represent the “dark productivity” of coral reefs, which fuels reef fish biomass production but is rarely perceived because it is consumed almost as quickly as it is produced.

Fig. 3 Ecosystem-scale effects of the demographic dynamics of cryptobenthic reef fishes.

(A) Cryptobenthics far outnumber large reef fishes in cohorts of larvae that recruit to reefs. (B) At settlement, large reef fish recruits are, on average, slightly larger than cryptobenthics. (C) Cryptobenthics contribute little (~13%) to net biomass production but produce almost 60% of consumed reef fish biomass because of exceptionally high turnover. (D) Standing stock biomass of large reef fishes far outweighs that of cryptobenthics, although adult abundances are approximately even. (E) Despite higher gamete output from large reef fishes, cryptobenthic larvae dominate the near-reef ichthyoplankton. This restarts the “crypto-pump” through rapid replenishment of consumed individuals. Uncertainty estimates in (C) and (D) are based on 100 iterations of the full model.

This role of cryptobenthics is empirically reflected in their extreme mortality [up to 70% per week (20, 21)] and their consumption by virtually any predator capable of eating them (21, 22). Although the community-wide representation of cryptobenthics in fish diets frequently appears lower than the ~58% identified herein (23), their true contribution to coral reef trophodynamics may be obscured by (i) rapid digestion, precluding reliable visual identification (24); (ii) predation on cryptobenthics by invertebrates (22), which are fed on by larger fishes; and (iii) predation on cryptobenthics by juvenile predatory fishes [e.g., cryptobenthics comprise up to 88.6% of fish prey for juvenile groupers (25, 26)], which are rarely included in community-wide dietary analyses.

The key to the unique demographic dynamics of cryptobenthics and their productivity might be a shift away from long-range dispersal toward retention of larvae in the immediate vicinity of natal (home) reefs. Four lines of evidence, along with our findings, support this hypothesis. First, larval dispersal models show that larval supply easily maintains adult populations in large-bodied, long-lived reef fishes (7). Conversely, small-bodied, short-lived taxa appear unable to sustain local populations even when active swimming by larvae is considered (7), yet cryptobenthic populations persist. Near-complete retention of cryptobenthic larvae close to natal reefs may solve this paradox. Second, driven by olfactory and auditory cues, cryptobenthic larvae show stronger natal homing than similarly sized large reef fishes (27) and can have very short dispersal distances (28). Third, cryptobenthic larvae have limited yolk sacs and ingest prey immediately after hatching, indicating dependence on resource-rich, near-reef environments (8, 29). Finally, remaining close to natal reefs during development should result in fine-scale genetic structuring. With few exceptions (30), this is observed in cryptobenthic fishes (10, 27, 31). Collectively, this suggests that the “pelagic” larvae of most cryptobenthics remain close to their natal reefs (32), resulting in prodigious near-reef abundances and a unique role of cryptobenthics in coral reef ecosystem functioning.

Larval retention may lead to two evolutionary consequences: (i) rapid speciation through micro-allopatry arising from restricted gene flow among populations and frequent reproductive isolation (10, 33, 34) and (ii) a higher risk of extinction than commonly assumed for small marine fishes (35) because populations can easily become ephemeral and disappear after stochastic environmental changes (36). If these processes scale up to macroevolutionary levels, then cryptobenthics should be phylogenetically rare but species rich and this is indeed the case. Few reef fish families have successfully adopted cryptobenthic lifestyles and major diversification events are restricted to two lineages (Fig. 4): the larger Blenniiformes and the Gobiaria (gobies and apogonids). Nevertheless, these lineages are among the most rapidly diversifying clades of actinopterygian fishes (37) and cryptobenthics collectively account for almost half (44.5%) of total reef fish biodiversity (Fig. 4) (13).

Fig. 4 Phylogenetic positioning and species richness in cryptobenthic and large reef fishes.

(A) Cryptobenthic reef fishes (blue) have few independent origins but (B) account for almost half of all reef fish species (2799 species). Black, reef fish taxa; gray, non–reef fish taxa [as in Brandl et al. (13)]. Bubble sizes (A) and bars (B) represent species richness within all taxa (all fish species) and cryptobenthic families (reef-affiliated species only). Arrow in (B) indicates cumulative richness in the Blenniiformes.

In summary, through their extraordinary larval dynamics, rapid growth, and extreme mortality, the hyperdiverse consortium of abundant, tiny, and short-lived cryptobenthic species appears to be a critical functional group on coral reefs. The “dark productivity” provided by cryptobenthics underpins reef fish biomass production and supports the characteristic fast-paced dynamics of modern coral reefs.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S6

Tables S1 to S6

References (3995)

Raw Data and Code

References and Notes

  1. Materials and methods are available in the supplementary materials.
Acknowledgments: We thank V. Huertas, S. Degregori, A. Mercière, and the staff of CRIOBE, Lizard Island, and Carrie Bow Caye for field support and S. Vignieri, K. Wilson, and R. Streit for editorial support and comments. Funding: Funding was provided by the BNP Paribas Foundation (V.P.), the Agence National de la Recherche (V.P., S.J.B., D.R.B.), the Smithsonian Institution’s TMON (S.J.B.), the National Science and Engineering Research Council of Canada (I.M.C., S.J.B.), and the Australian Research Council (D.R.B.). This is contribution 35 from the Smithsonian’s MarineGEO Network. Author contributions: S.J.B., C.H.R.G., L.T., and D.R.B. conceived the study; S.J.B., J.M.C., C.H.R.G., V.P., R.A.M., C.C.B., and N.M.D.S. collected data; S.J.B., R.A.M., N.M.D.S., and L.T. analyzed the data; S.J.B. wrote the first draft and all authors contributed to writing the paper. Competing interests: None declared. Data availability: Data and code are available on Zenodo (38).

Stay Connected to Science

Navigate This Article