A potential mechanism for episodic recruitment of a coral reef fish.
Marine animals generally have complex life cycles with dispersive larvae and sedentary/demersal juveniles and adults. The dispersive larval stage acts to decouple recruitment intensity from local reproductive effort; recruitment should then result from larval production external to each population. Recruitment to these open populations is often extremely variable through time (barnacles - Caffey 1985; bryozoans - Yoshioka 1982; reef fishes - Doherty and Williams 1988). Ultimately, recruitment variation can significantly influence both population size and community structure in marine system (Victor 1983, Menge 1991, Doherty and Fowler 1994).
Despite having dispersive larvae, self-recruitment (i.e., recruitment by larvae returning to the population from which they were produced) may still be important in marine systems (Black et al. 1991, Boehlert et al. 1992, Black 1993, Cowen and Castro 1994, Petersen and Svane 1995). For instance, even though crown-of- thorns starfish (Acanthaster planci) have larvae which are planktonic for [approximately equal to]10 d, self-recruitment has been implicated in outbreaks of this species along the Great Barrier Reef (Black et al. 1995).
Coral reef fishes provide an excellent opportunity to examine the extent of self-recruitment in open populations with dispersive larvae. Coral reef fishes range in spawning patterns (Thresher 1984); their recruitment patterns also display complexity and do not always overlap with spawning patterns (Robertson 1990), thereby reinforcing the potential for open population structure. Reef fish larvae can disperse effectively; species introduced on a single island in the Hawaiian Island chain have spread rapidly throughout the islands (Randall 1987). Due to this high degree of larval dispersal, recruits arriving at a reef should be produced externally to the resident population. However, two recent studies suggest that many families of reef fish larvae are retained near reefs or in lagoons (Brogan 1994, Leis 1994).
If fish populations self-recruit in areas of water retention and slow currents, then there is potential for self-recruitment in Kaneohe Bay, Oahu, Hawaii [ILLUSTRATION FOR FIGURE 1 OMITTED]. The fringing and patch reefs within Kaneohe Bay support large populations of coral reef fishes, and water exchange is limited to tidal transport through two channels. In the southeastern section of the bay, water exchange occurs at [approximately equal to]26% per day and currents average just 0.25 km/h (Bathen 1968). These water exchange rates should be equivalent to (or lower than) the lagoonal sites where self-recruitment of coral reef fish been reported. The 26% water exchange rate integrated over the larval period of most reef fish (2-6 wk) would leave [less than]0.1% of the original larvae remaining, if those larvae act as passive particles. However, active swimming and positioning by larvae may offset the effect of water exchange. For example, in Kaneohe Bay, the larvae of two reef fish, Psilogobius mainlandi and P. sp., actively positioned themselves near reefs (Kobayashi 1989) and other reef fish have shown similar capabilities (Brogan 1994, Leis 1994). In fact, actively swimming larvae could easily maintain themselves at the current speeds reported in Kaneohe Bay (Stobutzki and Bellwood 1994). I therefore hypothesized that the fish populations in the bay would be largely self-recruiting. I chose the endemic Hawaiian domino dam-selfish, Dascyllus albisella, as my study species. D. albisella spawns synchronously among populations (separated by [greater than]1 km) in nonlunar, 6-d cycles (Danilowicz 1995b), and recruits in episodic, nonlunar pulses (Booth 1992). The overlap between actual spawning and the spawning dates of arriving recruits (back-calculated spawning) should reflect the amount of self-recruitment occurring within these Kaneohe Bay populations.
MATERIALS AND METHODS
Validation of daily otolith increments
To determine whether spawning and recruitment were linked, I had to verify that otoliths from Dascyllus albisella recruits accurately reflected when each fish was spawned. Fish collected in this study were numbered, measured (standard length), and their two largest pairs of otoliths (sagittae and lapilli) were removed. The four otoliths were placed in immersion oil on a microscope slide and set aside for a month to enhance ring visibility (Victor 1986). Rings in the lapilli were counted using a video system attached to a 400 x microscope.
To establish when the first ring in the otolith was deposited, larvae were reared following the methods of Danilowicz and Brown (1992). Briefly, eggs were brought into the laboratory and hatched; larvae were then placed in rearing tanks with a combination of algae, rotifers, and fractions of progressively larger wild zooplankton. The only deviation from the methods of Danilowicz and Brown (1992) was that rearing was conducted in a 20:4 light: dark cycle (larvae raised in continual light failed to show any increments in their otoliths; B. S. Danilowicz, personal observation). Twenty larvae were collected from the rearing tanks both 7 and 8 d after hatching and placed in buffered 95% ethanol. Rings were counted "blind" (i.e., the number of expected rings was not known prior to counting). If the first ring was deposited on the night of hatching, the number of rings present should equal the number of days since a larva hatched.
To confirm daily deposition of otolith increments in the field, 50 recruits were collected from Kaneohe Bay. Collected fish were immersed for 8 h (at night) in a 250 mg tetracycline/L solution. The recruits were then returned to isolated coral heads on the reef and subsequently recollected. Twenty-seven of the tetracycline-marked recruits were recollected: 0 (of 6 initially released) at 6 d; 4 (of 9) at 7 d; 0 (of 7) at 8 d; 13 (of 14) at 10 d; and 10 (of 14) at 12 d. The fish were placed in buffered 95% ethanol until later analysis. The number of rings (including and after the fluorescent ring) was counted using an epifluorescent microscope. The number of rings expected was known prior to counting. If rings are deposited daily, their number should equal the number of days since the tetracycline treatment.
Observations of spawning were made on nine plots within Kaneohe Bay (three 15 x 10 m plots on Reefs 2, 3, and 5, as in Roy (1970; Fig. 1). Observations were made on lunar weeks (full, last quarter, new, first quarter) from 19 January 1992 through 1 December 1992, and daily from 28 May through 12 October 1992. Daily and weekly spawning observations are fully described in Danilowicz (1995b) and Danilowicz (1995a), respectively.
The Hawaii Institute of Marine Biology weather station provided maximum daily surface water temperature. The average weekly maximum water temperature was calculated for comparison to weekly spawning and recruitment data.
Larvae that have ended their dispersive stage and have begun their reef-associated stage are referred to as recruits. Recruits are typically the smallest size class of any fish that can be observed on a reef. To monitor the daily and weekly arrival of recruits, two isolated grids of corals were used at each of four locations: adjacent to Sampan Channel; Reef 2; Reef 3; and Reef 5 [ILLUSTRATION FOR FIGURE 1 OMITTED]. Each grid consisted of 18 coral heads (6 coral heads of each of 3 coral species: Porites compressa, Pocillopora meandrina, Montipora verrucosa). The Sampan Channel had a uniform sandy bottom at 2-3 m depth; the three reef locations consisted of silted slopes at 6-9 m depth with occasional outcroppings of coral rubble. The coral heads at each location were separated from each other by at least 5 m. To examine patterns of recruitment with high-temporal resolution, recruits were counted and removed daily from one coral grid at each of the four locations between 1 July-2 November 1992. To examine seasonal and lunar patterns of recruitment, recruits were counted and removed on lunar weeks from the other grid at each of the four locations between 19 January and 23 December 1992. Recruits were removed using the anesthetic Quinaldine and hand nets, and then preserved in buffered 95% ethanol for later otolith analyses.
Each recruit's hatching date was calculated by subtracting the number of rings in its otolith from the date that it was collected. Since D. albisella eggs hatch 3.5 d after they are spawned, and the first ring is deposited on the night of hatching (shown below in the results), three additional days were subtracted from the back-calculated hatching date to provide the spawning date of each recruit.
I examined the number of increments deposited in labeled and reared fish with regression analyses, and a second-order regression was used to examine the age/length relation of arriving recruits (Sokal and Rohlf, 1981).
Prior to the analysis of time series, spawning and recruitment data were square-root transformed to reduce temporal variance (Meekan et al. 1993). To remove seasonal autocorrelations from each data set, a second-order regression (simulating a seasonal trend; Priestly 1981) was fit to each time series and the residuals were used in subsequent analyses. Data were fit with a Hanning data window, then spectral analyses were performed to explore dominant cycles within the time series (Bloomfield 1976). Data use and degrees of freedom were maximized for spectral analyses: the daily recruitment data (n = 125) were split into two 64 point ensembles with three overlapping points (providing 2 df); the back-calculated recruitment data (n = 136) were split into three 64 point ensembles with 28 overlapping points (providing df = 4); the weekly recruitment data (n = 44) were placed into 32 point ensembles with 20 overlapping points (providing df = 2).
I used cross-correlations to determine whether significant lags were present among the time series (Priestley 1981). The same transforms used for spectral analyses were also used prior to cross-correlation analyses, except the data were not fit with a Hanning window.
Validation of daily otolith increments
Growth rings were clearly visible only in the lapillar otoliths. Therefore, only increment counts of lapillar otoliths were used. Of the 27 fish that were recaptured after immersion in tetracycline, 15 had discernible fluorescent rings in their otoliths. Of these, 7 rings were counted in the otoliths of two fish recaptured after 7 d, 9 in one fish (10-d recapture), 10 in three fish (10-d recapture), 11 in two fish (12-d recapture), 12 in six fish (12-d recapture), and 13 in one fish (12-d recapture). The slope of the regression (m = 0.99) was not significantly different from 1 (n = 15, SE = 0.075, P [greater than] 0.9), indicating that one increment was deposited in the otolith each day.
Of the 40 laboratory reared larvae, only 6 possessed otoliths with clearly visible rings. The other 34 larvae had no rings or the increments were too light to provide reliable counts. Of the four 7-d-old larvae, two had six increments and two had seven. The two 8-d-old larvae each had eight increments. Thus, each larvae had approximately one increment per day after hatching in both its lapillar otoliths. The slope of a regression through the data (m = 1.5) was not significantly different from 1 (n = 6, SE = 0.289, P [greater than] 0.2), and the intercept (b = -4) was not significantly different from 0 (SE = 2.85, P [greater than] 0.2). These results suggest that the first ring was deposited on the night of hatching (due to the low sample size, this interpretation should be viewed with caution).
Daily spawning and recruitment patterns
Dascyllus albisella spawning ([ILLUSTRATION FOR FIGURE 2A OMITTED]; from Danilowicz 1995b) and recruitment [ILLUSTRATION FOR FIGURE 2B OMITTED] varied over time in Kaneohe Bay. Spectral analysis detected no significant frequency in daily recruitment, indicating that recruitment was not temporally predictable at this resolution (e.g., not on a lunar cycle; [ILLUSTRATION FOR FIGURE 3A OMITTED]). When spawning was back-calculated from the recruitment pattern using otoliths [ILLUSTRATION FOR FIGURE 2C OMITTED], spectral analysis again found no significant frequency to be present [ILLUSTRATION FOR FIGURE 3B OMITTED]. Therefore, no cycle was apparent in either recruitment or recruit production in the daily time series.
If the arriving recruits were produced within Kaneohe Bay, back-calculated spawning should resemble the distribution of spawning in the bay (i.e., Fig. 2C should look like Fig. 2A). However, the temporal pattern of back-calculated spawning differed markedly from actual spawning, and a cross-correlation (lag = 0) between these data was not significant (r = -0.09, n = 136). Evidently, when most of the recruits were produced, reproduction was occurring at a low rate in Kaneohe Bay [ILLUSTRATION FOR FIGURE 2 OMITTED]. In addition, when the reproduction was sustained at a high rate, few arriving recruits were produced.
A plot of recruit standard length against age revealed a significant nonlinear relationship (Fig. 4; y = -0.001[x.sup.2] + 0.083x - 0.359, n = 926, [r.sup.2] = 0.723, P [less than] 0.0001). Average age at the time of settlement was 24.8 d posthatch (or 27.8 d postspawn; cv = 14.0), while the average standard length was 1.02 cm (CV = 11.2). If larvae hatch at [approximately equal to]0.2 cm, growth averaged 0.33 mm/d during the pelagic larval phase. The age and length frequencies of recruits are skewed to slightly younger and smaller distributions (Fig. 4).
Weekly reproduction and recruitment patterns
D. albisella spawned throughout the year [ILLUSTRATION FOR FIGURE 5A OMITTED]; from Danilowicz 1995a), and recruitment fluctuated over the same period [ILLUSTRATION FOR FIGURE 5B OMITTED]. Water temperature oscillated by 2 [degrees] C at [approximately equal to]6-wk intervals in the first half of the year (January through June), and spawning was significantly correlated with water temperature during these oscillations (Danilowicz 1995a). Spectral analysis detected no significant cycle to recruitment [ILLUSTRATION FOR FIGURE 6 OMITTED], indicating that recruitment at this resolution was not temporally predictable.
Spawning dates were not precisely calculated for recruits collected weekly due to the low variation of recruit ages found in fish collected daily [ILLUSTRATION FOR FIGURE 4 OMITTED]. Spawning dates were therefore estimated for recruits collected weekly by subtracting 4 wk from their recruitment time-series (i.e., 28 d were subtracted from the collection date of recruits rather than the individually specific ages [which averaged 27.8 d] of the daily collected recruits). A cross-correlation of spawning and recruitment (lag = 4 wk) should be significant if spawning was related to recruitment episodes. However, a non-significant negative correlation existed (r = -0.27, n = 40), indicating that spawning and the production of recruits arriving in Kaneohe Bay did not coincide.
Although there were no significant cycles to weekly recruitment (or to back-calculated spawning), recruitment does appear to be related to water temperature. The four water-temperature oscillations associated with spawning pulses in January through June also have an associated pulse of recruitment [ILLUSTRATION FOR FIGURE 5B OMITTED]. Each pulse of arriving recruits coincided with a peak in water temperature. Additionally, when water temperature reached minimums between the temperature oscillations, there were no apparent recruitment pulses. This relation seems to hold from January through June. It remains unclear whether the recruitment pulses in July and August are associated with water temperature. In the latter part of the year (September through December) there does not appear to be an association of recruitment and temperature.
Daily reproduction and recruitment
In July through October, when recruits were collected daily, no self-recruitment was detected. I draw this conclusion from two aspects of the results. First, daily spawning in Kaneohe Bay did not coincide with the back-calculated spawning dates of the recruits. This indicates that larvae produced in Kaneohe Bay during this period did not contribute noticeably to recruitment in the bay. Second, arriving recruits were not produced on a 6-d cycle. Dascyllus albisella synchronize spawning among local populations, thereby resulting in a synchronous 6-d spawning cycle ([ILLUSTRATION FOR FIGURE 2A OMITTED]; Danilowicz 1995b). If a significant portion of arriving recruits was produced within Kaneohe Bay, the 6-d cycle in the spawning data should be reflected in the back-calculated spawning dates of recruits. Since a significant spawning cycle was not evident in recruit production, and bouts of daily spawning did not overlap with recruit production, I conclude that most D. albisella recruits were produced elsewhere during these 4 mo. These D. albisella populations are therefore functionally open.
The possibility that a significant relationship between spawning and back-calculated spawning was masked by my methods is unlikely. Extremely high larval survival during times of reduced spawning could lead to an apparent non-overlap between spawning and recruitment. However, survival would need to be at least 11 times greater than during periods of reduced reproduction. A large change in survival would imply density-dependent mortality of larvae in the plankton, and competition for food (a potential cause of density-dependent mortality) has not been shown to be important in marine plankton communities (Morgan 1995). Another possibility is that slight density-dependent mortality of larvae in the plankton may act to "smooth" the 6-d spawning cycle, masking a spawning/back-calculated spawning relationship. However, if self-recruitment contributed a significant proportion of recruits, then the back-calculated spawning dates of recruits should still reflect the original spawning cycle (this argument focuses on tracing spawning to settlement patterns, which follows the arguments of Holm  on tracing density-dependent settlement to recruitment patterns).
Although my methods could not detect any self-recruitment within the bay, this does not mean it did not occur. Some larvae may successfully self-recruit, but their numbers may simply be swamped by those recruiting from elsewhere.
Weekly spawning and recruitment
During the year of this study, spawning in Kaneohe Bay generally did not coincide with the estimated spawning dates of recruits. This further supports my conclusion that these reefs were not self-recruiting, and that D. albisella populations in Kaneohe Bay are functionally open. Changes in seasonal growth patterns and density-dependent mortality of larvae probably do not mask self-recruitment. Dascyllus albisella larvae may take longer to develop during the winter than during the summer (only summer-collected recruits were aged in this study). If winter recruits developed much more slowly than the summer average of 24.8 d (developing over a 38-d period instead), then the observed pattern of reproduction could result in the observed pattern of recruitment during these months. This seems unlikely because several species of Abudefduf (also in the family Pomacentridae) collected in winter and summer have shown no differences in expected development time (Wellington and Victor 1989). The possibility that extreme density-dependent mortality of planktonic larvae would mask self-recruitment seems unlikely for the same reasons discussed with daily recruitment.
If these reefs received most of their recruits externally, the timing of recruitment episodes may reveal their origin (assuming recruitment intensity is positively correlated with spawning intensity of the source population). Recruitment episodes are coupled to water-temperature oscillations from January through June. In January through June, currents generally run to the northwest at 10 km/d, parallel to the Hawaiian islands (Barkley et al. 1964). Water-temperature oscillations of [approximately equal to]6-wk periods were apparent in this study during the first half of 1992, and also in at least 8 of 14 other years on Oahu (Seckel and Yong 1970). These water-temperature oscillations should move parallel to the islands with average current speed and direction. Each temperature oscillation in 1992 coincided with a change in D. albisella reproductive output, and temperature changes are thought to control spawning in this species (Danilowicz 1995a).
My findings, coupled with available oceanographic information, lead to the following hypothesis. When water temperature increased at the island of Hawaii (the first island in the chain), spawning was initiated. Eggs produced would hatch [approximately equal to]1 wk after the temperature increase (allowing a few days for adults to spawn and the 3.5-d egg incubation period), and the larvae would then enter the NW current. The warmer water in the northwest current would include and be "followed" by a large patch of D. albisella larvae. As the warm water oscillation passes over each of the down stream islands, it would both initiate spawning and deposit competent larvae (as recruits) originating from upcurrent islands. The distances between the islands, current velocities, average age of recruits, and temperature data support this hypothesis. The distances between Kaneohe Bay, Oahu, and the closest point of Maui and farthest point of Hawaii are 120 and 360 km, respectively. Given average current speed of 10 km/d, water passing Maui and Hawaii would take between 12 and 36 d to reach Kaneohe Bay. Since the average age of recruits (in fact all of the recruits but one) fit within this time period, and self-recruitment was not detectable, the recruits arriving in Kaneohe Bay from January through June were likely derived from these upcurrent islands. Additional information on spawning, ichthyoplankton distribution, and water currents are needed to test this hypothesis.
In contrast with the period of January through June, recruitment episodes for July through December were apparently unrelated to water temperature. During this later period, currents generally run in a westerly to southwesterly direction perpendicular to the islands at 10-18 km/d (Barkley et al. 1964, Lobel 1978). Larvae might then be advected away from the islands and not contribute substantially to recruitment (Lobel 1978).
In instances where no recruit source is upcurrent from a population, recruitment should be less intense (Cowen 1985). Since no islands are upcurrent of Oahu during July through December, I expected recruitment to show a corresponding decline. Two large recruitment episodes, one in July and one in August, do not fit this prediction. In fact, of the 9 yr that D. albisella recruitment has been monitored in the Hawaiian Islands, recruitment has occurred predominately during June through August (5 of 6 yr in Kaneohe Bay; Stevenson 1963, Groll 1984, Booth 1992, this study; and 2 of 3 yr at Midway Island; Schroeder 1985), although major recruitment episodes also were reported during other seasons. Overall, approximately one-third to one-half of D. albisella recruitment occurs during the summer months when there is no upstream recruit source.
There are several possible explanations as to why the summer recruitment peak does not correspond with expected times of favorable currents or to the observed reproductive peak. The most likely explanation is that this study only examined spawning of D. albisella in Kaneohe Bay, and this species as a whole may predominantly spawn in late spring through summer months in locations outside of the bay. The summer recruitment peak would then be coupled with the reproductive peak. In support of this possibility, Stevenson (1963), using a combination of gonosomatic indices and observations of nest preparation by males, reported that D. albisella exhibited peak spawning in May through August in Kaneohe Bay and Waikiki (south shore of Oahu) over a 2-yr period.
Other possible explanations for the observed decoupling of spawning and recruitment include mesoscale eddies, which may retain larvae and deposit them back on their natal islands or along the chain after they become competent to recruit. These eddies have been implicated as important to fish recruitment in the Hawaiian Islands (Sale 1970, Lobel 1978, 1989). Additionally, the currents mentioned in this paper represent the average direction of water movement through the islands. At any given time, water masses may be moving in any direction (Lobel 1978), which may result in episodic recruitment during any season. However, mesoscale eddies and random water movement are both extremely variable by their nature and should not result in the consistency of summer recruitment observed in this species.
Limited self-recruitment in open marine populations
Although passive/active larval retention may be important to recruitment of fish and invertebrates in some locations (Sammarco and Andrews 1988, Black et al. 1991, 1995), it was not important to Dascyllus albisella recruitment in Kaneohe Bay in 1992. Therefore, my original hypothesis, that D. albisella populations in this sheltered bay would be self-recruiting, was not supported. Similarly, Brogan (1994) found no evidence for the retention of pomacentrid larvae near reefs in the Gulf of California, although other families were retained through their developmental period. Leis (1994) found that some species of pomacentrids dispersed away from reefs, while others were retained. Thus, the importance of self-recruitment within the pomacentrids may need to be determined species by species, and probably from location to location (due to differential retention capacities and current speeds around individual reefs and shore regions).
As this study was conducted in one year and only in Kaneohe Bay, estimating the expected frequency of self-recruitment or its effects upon a D. albisella population is not possible. However, if a species does function almost entirely as an open population, it may not have the opportunity to adapt to local environments. Uniform mixing of larvae within the species' distribution would result in a local population being not specifically adapted to its local environment. Instead, the population would be adapted to the environment that influences the entire species, as the recruits arriving at a location may be from a wide range of other locations over which the species is distributed. Limited larval mixing among regions within the species' range could also occur in open populations with dispersive larvae. Larval dispersal capabilities may then not be strictly correlated with genetic differentiation over the species' range. Regions that are largely self-recruiting may develop different morphs in their populations from those in neighboring regions (e.g., regional color morphs in the damselfishes Amblyglyphidodon leucogaster, Chromis opercularis, and Acanthochromis polyacanthus; Allen 1991, Doherty et al. 1994).
It is important to remember that the genetic/phenetic patterns we see at the present may not be explainable by present-day patterns of dispersal. For instance, during glaciation periods, oceanic currents and species distributions in the tropics should have differed from what we observe today. This might explain why some genetic differentiation in these regions is counter to the present-day patterns of larval dispersal (Benzie and Williams 1995, Doherty et al. 1995).
High larval mixing rates among open populations would promote long-term coexistence of species (Roughgarden et al. 1985). Although a species may be able to competitively exclude another species from a single location at any given time, a competitively dominant species would not necessarily recruit best to a reef (because a reef is not self-recruiting). Thus, after disturbance and/or mortality have made resources available at a location, any species may recruit into that opening (Sale 1977). Species with long life-spans can continually produce potential recruits thereby allowing recruitment to locations where they were formerly excluded (Warner and Chesson 1985, Lertzman 1995). In coral reef fish assemblages, and possibly many other marine plant and animal assemblages, resources are thought by some not to limit either the number of individuals or the number of species that occupy a reef (Victor 1983, Sale 1991, Doherty and Fowler 1994). When self-recruitment is minimal in recruitment-limited assemblages, the chance that any species at any time should be removed from even a single location by competition is further reduced. Therefore, the number of species that inhabit any given location in a recruitment-limited community, where self-recruitment is minimal, would be determined by the physical processes bringing the larvae of different species to an area, as well as other forces, such as predation, that limit the establishment of particular species.
Jeff Mahon, Debbie Gochfeld, Lisa Privetera, Cara Pyle, and Richard Pyle assisted with field collections. Helpful discussion and/or reviews were provided by Richard Barber, Dave Booth, Bob Cowen, Richard Forward, Peter Howd, Steve Lindley, Ann Oliver, Andrea Risk, Peter Sale, Nick Tolimieri, and two anonymous reviewers. This research was supported by a National Defense Science and Engineering Graduate (NDSEG) Fellowship, the Duke University Marine Laboratory, the Hawaii Institute of Marine Biology, a Robert Safrit Fellowship (DUML), and the Lerner-Gray Marine Research Fund. Final preparation of the manuscript was supported by the University of Windsor, Ontario.
Allen, G. R. 1991. Damselfishes of the world. Mergus, Melle, Germany.
Barkley, R. A., B. M. Ito, and R. P. Brown. 1964. Release and recoveries of drift bottles and cards in the Central Pacific. United States Fish and Wildlife Service Special Science Report Number 492:1-31.
Bathen, K. H. 1968. A descriptive study of the physical oceanography of Kaneohe Bay, Oahu, Hawaii. Technical Report, Hawaii Institute of Marine Biology 14:1-353.
Benzie, J. A. H., and S. T. Williams. 1995. Gene flow among giant clam (Tridacna gigas) populations in Pacific does not parallel ocean circulation. Marine Biology 123:781-787.
Black, K. P. 1993. The relative importance of local retention and inter-reef dispersal of neutrally buoyant material on coral reefs. Coral Reefs 12:43-53.
Black, K., P. Moran, D. Burrage, and G. De'ath. 1995. Association of low-frequency currents and crown-of-thorns starfish outbreaks. Marine Ecology Progress Series 125: 185-194.
Black, K. P., P. J. Moran, and L. S. Hammond. 1991. Numerical models show coral reefs can be self-seeding. Marine Ecology Progress Series 74:1-11.
Bloomfield, P. 1976. Fourier analysis of time series: an introduction. J. Wiley & Sons, New York, New York, USA.
Boehlert, G. W., W. Watson, and L. C. Sun. 1992. Horizontal and vertical distributions of larval fishes around an isolated oceanic island in the tropical Pacific. Deep-Sea Research 39:439-466.
Booth, D. J. 1992. Larval settlement patterns and preferences by domino damselfish Dascyllus albisella Gill. Journal of Experimental Marine Biology and Ecology 155:85-104.
Brogan, M. W. 1994. Distribution and retention of larval fishes near reefs in the Gulf of California. Marine Ecology Progress Series 115:1-13.
Caffey, H. M. 1985. Spatial and temporal variation in settlement and recruitment of intertidal barnacles. Ecological Monographs 55:313-332.
Cowen, R. K. 1985. Large scale pattern of recruitment by the labrid, Semicossyphus pulcher: causes and implications. Journal of Marine Research 43:719-743.
Cowen, R. K., and L. R. Castro. 1994. Relation of coral reef fish larval distributions to island scale circulation around Barbados, West Indies. Bulletin of Marine Science 54:228-244.
Danilowicz, B. S. 1995a. The role of temperature in spawning of the damselfish Dascyllus albisella. Bulletin of Marine Science 57:624-636.
-----. 1995b. Spatial patterns of spawning in the coral reef damselfish Dascyllus albisella. Marine Biology 122:145-155.
Danilowicz, B. S., and C. L. Brown. 1992. Rearing methods for two damselfish species: Dascyllus albisella (Gill) and D. aruanus (L.). Aquaculture 106:141-149.
Doherty, P. J., and T. Fowler. 1994. An empirical test of recruitment limitation in a coral reef fish. Science 263:935-939.
Doherty, P. J., P. Mather, and S. Planes. 1994. Acanthochromis polyacanthus, a fish lacking larval dispersal, has genetically differentiated populations at local and regional scales on the Great Barrier Reef. Marine Biology 121:11-21.
Doherty, P. J., S. Planes, and P. Mather. 1995. Gene flow and larval duration in seven species of fish from the Great Barrier Reef. Ecology 76:2373-2391.
Doherty, P. J., and D. McB. Williams. 1988. The replenishment of coral reef fish populations. Oceanography and Marine Biology Annual Review 26:487-551.
Groll, P. E. 1984. Recruitment ecology of Dascyllus albisella (Gill), (Pisces, Pomacentridae): patterns, predation, and social interactions. Master's thesis. University of Hawaii, Honolulu, Hawaii, USA.
Holm, E. R. 1990. Effects of density-dependent mortality on the relationship between recruitment and larval settlement. Marine Ecology Progress Series 60:141-146.
Kobayashi, D. R. 1989. Fine-scale distribution of larval fishes: patterns and processes adjacent to coral reefs in Kaneohe Bay, Hawaii. Marine Biology 100:285-293.
Leis, J. M. 1994. Coral Sea atoll lagoons: closed nurseries for the larvae of a few coral reef fishes. Bulletin of Marine Science 54:206-227.
Lertzman, K. P. 1995. Forest dynamics, differential mortality and variable recruitment probabilities. Journal of Vegetation Science 6:191-204.
Lobel, P. S. 1978. Diel, lunar, and seasonal periodicity in the reproductive behavior of the pomacanthid fish, Centropyge potteri, and some other reef fishes in Hawaii. Pacific Science 32:193-207.
-----. 1989. Ocean current variability and the spawning season of Hawaiian reef fishes. Environmental Biology of Fishes 24:161-171.
Meekan, M. G., M. J. Milicich, and P. J. Doherty. 1993. Larval production drives temporal patterns of larval supply and recruitment of a coral reef damselfish. Marine Ecology Progress Series 93:217-225.
Menge, B. A. 1991. Relative importance of recruitment and other causes of variation in rocky intertidal community structure. Journal of Experimental Marine Biology and Ecology 146:69-100.
Morgan, S. G. 1995. Life and death in the plankton: larval mortality and adaptation. Pages 279-321 in L. R. Mc-Edward, editor. Ecology of marine invertebrate larvae. CRC Press, Boca Raton, Florida, USA.
Petersen, J. K., and I. Svane. 1995. Larval dispersal in the ascidian Ciona intestinalis (L). Evidence for a closed population. Journal of Experimental Marine Biology and Ecology 186:89-102.
Priestley, M. B. 1981. Spectral analysis and time series. Academic Press, San Diego, California, USA.
Randall, J. E. 1987. Introductions of marine fishes to the Hawaiian Islands. Bulletin of Marine Science 41:490-502.
Robertson, D. R. 1990. Differences in the seasonalities of spawning and recruitment of some small neotropical reef fishes. Journal of Experimental Marine Biology and Ecology 144:49-62.
Roughgarden, J., Y. Iwasa, and C. Baxter. 1985. Demographic theory for an open marine population with space limited recruitment. Ecology 66:54-67.
Roy, K. 1970. Changes in bathymetric configuration, Kaneohe Bay, Oahu, 1882-1969. Hawaii Institute of Geophysics Report Number 70-15.
Sale, P. F. 1970. Distribution of larval Acanthuridae off Hawaii. Copeia 1970:765-766.
-----. 1977. Maintenance of high diversity in coral reef fish communities. American Naturalist 111:337-359.
-----. 1991. Reef fish communities: open nonequilibrial systems. Pages 564-598 in P. F. Sale, editor. The ecology of fishes on coral reefs. Academic Press, San Diego, California, USA.
Sammarco, P. W., and J. C. Andrews. 1988. Localized dispersal and recruitment in Great Barrier Reef coral: the Helix experiment. Science 239:1422-1424.
Schroeder, R. E. 1985. Recruitment rate patterns of coral-reef fishes at Midway Lagoon (Northwestern Hawaiian Islands). Proceedings of the Fifth International Coral Reef Congress, Tahiti 1:379-384.
Seckel, G. R., and M. Y. Y. Yong. 1970. Harmonic functions for sea-surface temperatures and salinities, Koko Head, Oahu, 1956-69, and sea surface temperatures, Christmas Island, 1954-69. Fisheries Bulletin 69:181-214.
Sokal, R. R., and F. J. Rohlf. 1981. Biometry. Second edition. W. H. Freeman, New York, New York, USA.
Stevenson, R. A. 1963. Life history and behavior of Dascyllus albisella Gill, a pomacentrid reef fish. Dissertation. University of Hawaii, Honolulu, Hawaii, USA.
Stobutzki, I. C., and D. R. Bellwood. 1994. An analysis of the sustained swimming abilities of pre- and post-settlement coral reef fishes. Journal of Experimental Marine Biology and Ecology 175:275-286.
Thresher, R. E. 1984. Reproduction in reef fishes. TF.H. Publications, Neptune City, New Jersey, USA.
Victor, B. C. 1983. Recruitment and population dynamics of a coral reef fish. Science 219:419-420.
-----. 1986. Delayed metamorphosis with reduced larval growth in a coral reef fish (Thalassoma bifasciatum). Canadian Journal of Fisheries and Aquatic Sciences 43:1208-1213.
Warner, R. R., and P. L. Chesson. 1985. Coexistence mediated by recruitment fluctuations: a field guide to the storage effect. American Naturalist 125:769-787.
Wellington, G. M., and B. C. Victor. 1989. Planktonic larval duration of one hundred species of Pacific and Atlantic damselfishes (Pomacentridae). Marine Biology 101:557-567.
Yoshioka, P. M. 1982. Role of planktonic and benthic factors in the population dynamics of the bryozoan Membranipora membranacea. Ecology 63:457-468.
|Printer friendly Cite/link Email Feedback|
|Author:||Danilowicz, Bret S.|
|Date:||Jul 1, 1997|
|Previous Article:||Variation of trophic fractions and connectance in food webs.|
|Next Article:||Reproductive success of smallmouth bass estimated and evaluated from family-specific DNA fingerprints.|