Short Report
Evidence of density-independent mortality in a settling coral reefdamselfish, Chromis viridis
David Lecchini1,2*, Yohei Nakamura2,3, Julien Grignon1, and Makoto Tsuchiya2
1 CRIOBE, Centre de Recherches Insulaires et Observatoire de l’Environnement, Moorea, BP 1013, French Polynesia(e-mail: DL, [email protected])2 Laboratory of Ecology and Systematics, University of the Ryukyus, 1 Senbaru, Nishihara, Okinawa 903-0213, Japan3 Department of Global Agricultural Sciences, Graduate School of Agricultural and Life Sciences, The University of Tokyo,Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan
Received: August 14, 2005 / Revised: February 3, 2006 / Accepted: February 8, 2006
Abstract To know if the variation in the number of settling fish larvae can be dampened by density-dependent postsettlement mortality, we investigated the relationship between settler density andpredator-induced mortality of a coral reef damselfish, Chromis viridis. Totals of 2, 3, 5, 8, 10, 12, 14, 16,18, and 20 fish of 10 or 20 mm total length were released in experimental cages enclosing a coral headof Porites rus (to provide settlement habitat) and five predators. The results showed that the mortalityrate of both 10- and 20-mm fish was density independent.
Key words Coral reef fish · Settlement · Predation · Density independence
as a patchily distributed, highly aggregated planktivore (i.e.,a resource competitor with other planktivorous species,including some corals) and as an important prey species formany reef-based predators.
Materials and Methods
Experimental design and methods.—Experiments in cageswere conducted in the Moorea Lagoon (French Polynesia)in 2000. Each cage was made of a circular frame, 3m indiameter, with fine mesh net (stretched mesh, 4mm). Thenet was suspended 15cm below the water surface by aninflated circular buoy and was sealed to the sandy lagoonfloor by heavy metal sheets (see Juncker et al., 2005).
A total of 336 Chromis viridis larvae were collected insitu on Porites rus colonies. Larvae were kept 24h in aquariabefore being introduced into cages. We distinguished twofish size classes [total length (TL) = 10mm ± 0.7 SE and20mm ± 1.3 SE], which corresponded to larvae 3 (±0.9 SE)and 16 (±1.8 SE) days postsettlement. These ages were esti-mated by otolith analysis on 20 individuals of each size. Weused these age classes to distinguish between processes oc-curring just after settlement (3 days) and those affecting fishalready adapted to the reef (16 days).
Five species of sedentary carnivorous fishes were selectedas predators: Epinephelus hexagonatus (Serranidae), Pteroisantennata and Pterois radiata (Scorpaenidae), Sargocentroncaudimaculatum (Holocentridae), and Rhinecanthusaculeatus (Balistidae). For each replication of the experi-ment, one individual of each species was released in eachcage. The individuals of the same species were roughly thesame size (between 8 and 15cm long depending on the
IchthyologicalResearch
©The Ichthyological Society of Japan 2006
Ichthyol Res (2006) 53: 298–300DOI 10.1007/s10228-006-0340-8
Population size is influenced by inputs (births andimmigration) and outputs (deaths and emigration), and
if any of these rates is density dependent, regulation mayoccur in which populations fluctuate within defined limitsover a period of time (Olafsson et al., 1994; Doherty, 2002).Many populations living in marine benthic habitats havedispersive propagules. After spending some time in the wa-ter column, the surviving larvae try to settle on the seafloor.Such benthic populations are characterized as open systems,and inputs (i.e., larval supply) are often highly variablebecause individuals may arrive from distant sources, anddensity-independent processes may be an important sourceof loss or output (Lecchini and Galzin, 2003). Observationaldemographic studies, in which marine larvae were followedto and through the adult state, aimed to determine to whatextent density-dependent postsettlement processes dampenthe variation in the numbers of successful settlers in openpopulations (Olafsson et al., 1994; Doherty, 2002). Some ofthese studies showed only a weak relationship between den-sities of settling larvae and subsequent adult populationsize, but other studies revealed cases where the density ofsettling larvae was reflected in adult population density.However, the importance of density-dependent or density-independent processes (i.e., mortality, growth, and competi-tion) in determining benthic population dynamics is stillcontroversial (Sale and Tolimieri, 2000).
To document the importance of density-dependent ordensity-independent mortality in fish population dynamics,we explored the relationship between the larval supply andpredator-induced mortality of a small planktivorous damselfish, Chromis viridis, which is commonly found on coralreefs throughout much of the Indo-Pacific region. At manysites, C. viridis probably serves an important ecological role
Postsettlement mortality in fish larvae 299
species, with a maximum difference of 2cm between indi-viduals of the same species). Each predator was kept 24hwithout feeding before being introduced into the cage.
Experiment of density effects.—The effect of prey densityon mortality was assessed using a variable density of fishlarvae with a constant density of predators. The experimentinvolved three similar cages, each occupied by a livingPorites rus coral head (15cm in diameter, settlement habitatof C. viridis at Moorea Island; Juncker et al., 2005). In eachcage, five predators were introduced. Successive runs wereconducted respectively using 2, 3, 5, 8, 10, 12, 14, 16, 18, and20 larvae of 10mm TL released in the cage. The larvae werereleased close to shelter and predators at the surface 1hlater. After 24h, the remaining larvae were counted andremoved. Predators were also removed. This protocol wasconsidered one run. Three similar runs were carried out foreach density, each run using fresh predators and larvae ineach cage. A similar complete experiment was conductedusing 20mm TL fish.
To verify that the disappearance of a larva from a cagewas the result of predation mortality rather than naturalmortality or emigration, control experiments withoutpredators and with 10 larvae per cage were conducted at thebeginning and the end of the density experiment.
Results and Discussion
Of the 60 fish of 10mm TL and the 60 fish of 20mm TL thatwere introduced into cages without predators (control ex-periment), only one 10 mm TL fish died (natural mortality),and no emigration was observed 24h later. Based on theseresults, disappearance of a larva during the density experi-ment was attributed to predation mortality. In the densityexperiment, the number of larvae, both small (10mm TL)and large (20mm TL), that remained in the cages wassignificantly correlated with the number of individuals re-leased 24h earlier (linear regression: r2 = 0.96, P = 0.0001 for10mm TL; r2 = 0.97, P = 0.0001 for 20mm TL). The mortalityrate was thus independent of the initial abundance of 10mmfish (r2 = 0.03, P = 0.64), with a mean mortality value of 37%and means ranging between 17% and 50% (Fig. 1). For20-mm fish, the mean mortality rate was 20%, with meansvarying from 0% to 33% (Fig. 1). No pattern of density-dependent mortality has been detected across a large rangeof larvae densities (2–20 larvae per coral head of 15cmdiameter).
In marine fishes, the mortality of new settlers in the fewdays following settlement may eliminate up to 60% of thetotal population within a day and certainly contributes to adecrease of about 90%–95% of the total population within2–3 months (Doherty, 2002). Predation is an importantsource of mortality for new settlers. To explore the effects ofsettling fish density on postsettlement mortality by preda-tion, we chose to look at the instantaneous mortality rate(during 24h) rather than mortality over several weeks,which is the usual window of interest in population regula-tion studies (e.g., Doherty and Fowler, 1994; Schmitt andHolbrook, 2000). Because of this nontraditional approach,
we highlighted that the adult population size of Chromisviridis was only determined by density-independent mortal-ity across a large range of larvae densities (2–20 larvae percoral head 15cm in diameter). Additional treatments withmore individuals (>20 larvae introduced into cages) mostprobably would be needed to detect density-dependentmortality. Overall, the processes that generate density-dependent or density-independent mortality in animalpopulations are not well understood. Some authors notethat investigating behavioral trade-offs of prey at risk ofpredation may be a key to understanding population-levelpatterns in prey growth and survival as a function of density(Luttbeg and Schmitz, 2000; Schmitz, 2001). The trade-offbetween growth and mortality rates, mediated by foragingactivity, may be a viable behavioral mechanism to explaindensity-dependent or density-independent mortality(Sinclair, 1989). For example, if prey animals increase activ-ity in response to lower food abundance resulting fromdensity-dependent depletion of food, then this may leaddirectly to higher mortality as a result of increased visibilityand encounter rates with predators. Thus, much of the his-tory of population ecology has focused on understandingthe causes of density-dependent or density-independentgrowth and survival in animal populations (Sinclair, 1989), afocus that stems from the significant impacts of these factorson population regulation and population dynamics.
Acknowledgments This research was supported by the CRIOBE anda C.O.E. Fellowship awarded to D. Lecchini. We declare that any ex-periments comply with the current laws of French Polynesia. The firstauthor wishes to thank Albert and Jean-Claude Rivière (<<à ces deuxcévenoles d’exception>>).
Literature Cited
Doherty PJ (2002) Variable replenishment and the dynamics of reeffish populations. In: Sale PF (ed) Coral reef fishes: dynamics in acomplex ecosystem. Academic Press, San Diego, pp 327–358
Fig. 1. Variation in mortality rate (mean and standard error) of 10and 20mm TL Chromis viridis as function of initial abundances ofindividuals released 24 h earlier in the cages
300 D. Lecchini et al.
Doherty PJ, Fowlert T (1994) Empirical test of recruitment limitation ina coral reef fish. Science 263:935–939
Juncker M, Wantiez L, Lecchini D, Galzin R (2005) Effet de l’habitatsur le comportement et la mortalité des juvéniles de Chromis viridis.Cybium 29:3–12
Lecchini D, Galzin R (2003) Influence of pelagic and benthic, biotic andabiotic, stochastic and deterministic processes on the dynamics ofauto-recruitment of coral reef fish. Cybium 27:167–184
Luttbeg B, Schmitz OJ (2000) Predator and prey models with flexibleindividual behavior and imperfect information. Am Nat 155:669–683
Olafsson EB, Peterson CH, Ambrose WG (1994) Does recruitmentlimitation structure populations of macro-invertebrates in marinesoft sediments. Oceanogr Mar Biol Annu Rev 32:65–109
Sale PF, Tolimieri N (2000) Density dependence at some time andplace? Oecologia (Berl) 124:166–171
Schmitt RJ, Holbrook SJ (2000) Habitat-limited recruitment of coralreef fish. Ecology 81:3479–3494
Schmitz OJ (2001) From interesting details to dynamical relevance:toward more effective use of empirical insights in theory construc-tion. Oikos 94:39–50
Sinclair ARE (1989) Population regulation in animals. In: Cherret JM(ed) The contribution of ecology to an understanding of the naturalworld. Blackwell, Oxford, pp 197–241
Recommended
