Abstract

The mutualistic nature of cleaning symbioses has long remained unconfirmed because of the difficulty in showing net benefits for clients. We have previously shown that cleaning gobies (Elacatinus spp.) within territories of Caribbean longfin damselfish (Stegastes diencaeus) reduce the number of gnathiid isopod ectoparasites on territory owners. We now investigate whether this benefit of being cleaned has reproductive consequences for male longfin damselfish. The mating success, rate of egg loss, and parental aggression of 40 nest-guarding males were assessed during six consecutive monthly reproductive periods. Ten males had cleaning stations within their territory, 10 males were without cleaning stations, and 20 males initially with a cleaning station had their cleaners removed half-way through the study. Ectoparasite loads on our focal fish were very low; however, damselfish with cleaning stations still had significantly fewer ectoparasites than did fish without cleaning stations. There was, however, no significant difference in the number of eggs, clutches, or area of clutches received, or in the number of eggs lost before hatching between damselfish with and without cleaners. We also found no difference in parental male aggression between damselfish with and without cleaners. We conclude that although ectoparasite removal appears to have no direct consequence for reproduction, at least for the levels of infestations observed on our study site, it may still affect other aspects of damselfish fitness such as survival.

The long-held belief that cleaning interactions among fish are mutualistic (Begon et al., 1990; McFarland, 1985; Thompson, 1994; Trivers, 1971) has long remained unconfirmed because of the difficulty in showing net benefits for clients. Although an early study found that fish became diseased when cleaners were removed from a reef (Limbaugh, 1961), subsequent studies failed to find any effect of cleaner absence on client health or abundance (Gorlick et al., 1987; Grutter, 1996; Losey, 1972; Youngbluth, 1968). Furthermore, there is ample evidence that cleaners often take not only ectoparasites but also scales and mucus from their clients (Arnal and Côté, 2000; Arnal and Morand, 2001b; Gorlick, 1980; Grutter, 1997b).

However, cleaners have recently been shown to reduce ectoparasite loads. Grutter (1999) found significantly fewer ectoparasites on thick-lip wrasse Hemigymnus melapterus clients placed in experimental cages on reefs with access to cleaner wrasses Labroides dimidiatus than on clients without cleaners. Similarly, Cheney and Côté (2001) have shown that longfin damselfish Stegastes diencaeus with cleaning gobies Elacatinus spp. present in their territory visited cleaning stations more frequently and had fewer ectoparasites than did damselfish with less regular access to cleaners.

Ectoparasites can have significant impacts on the fitness of their hosts. They can cause disease (Cusack and Cone, 1986), decrease reproductive output (Adlard and Lester, 1995; Møller et al., 1999), increase predation on weakened hosts (Lafferty and Morris, 1996), and affect behavior and distribution of individuals (Poulin, 1994). Gnathiid isopod larvae, the usual target of cleaner fish predation (see Arnal and Côté, 2000; Grutter, 1996), can cause tissue damage (Honma and Chiba, 1991) and at high densities may cause death of captive fish (Mugridge and Stallybrass, 1983; Paperna and Por, 1977). Given the potential deleterious effects of ectoparasites on host health, their removal during cleaning interactions should lead to increased client fitness, through either enhanced survival or reproduction. Such ultimate consequences of being cleaned have not yet been examined.

In the present study, we investigated the impact of being cleaned on an important aspect of longfin damselfish fitness, namely, reproductive output. Longfin damselfish are among the commonest visitors to cleaning stations operated by cleaning gobies (Elacatinus spp.; Arnal and Côté, 1998), but they restrict their visits to cleaning stations within or very near their territories (less than 2 m away; Cheney and Côté, 2001). Longfin damselfish with a cleaning station within their territorial boundaries harbor fewer ectoparasites than do damselfish further away from cleaners (Cheney and Côté, 2001). We now compare the mating success and parental aggression of male longfin damselfish living with and without cleaning gobies within their territories. We thus provide the first examination of fitness-related benefits of being cleaned.

METHODS

Study site and species

The study was performed on a fringing reef off the Bellairs Research Institute, on the west coast of Barbados (13°10′ N, 59°30′ W), West Indies. The study area was located in the spur and groove section of the north (approximately 50–100 m from shore) and south divisions of the reef (approximately 100 m from shore) at depths ranging from 3–7 m.

Longfin damselfish are abundant at this site. They are sexually monomorphic, and both sexes defend year-round 1-m2 territories, which provide algal food resources, shelter, and, for males, nest sites. Spawning occurs throughout the year but peaks during the summer months (Robertson, 1990). Each spawning period lasts for approximately 20 days/month (Robertson et al., 1990), peaking around the new moon, with spawning occurring only during the first 2 h of daylight. Males prepare nests by clearing algae from a suitable vertical surface, and then they court visiting females, which lay a discrete monolayer of eggs (a clutch) in the nest. The males guard alone until hatching 5 days later. A male may accept clutches on consecutive days; hence, eggs of different ages (a brood) may be guarded simultaneously. Clutches can be aged by color as they change from white on day 1, to light grey on day 3, to silvery on day 5.

Two species of cleaning goby, Elacatinus prochilos and E. evelynae, occur on Barbadian reefs. Both are found on coral, but E. prochilos is also found on sponge. Only gobies on coral were considered in the present study because they are the most active cleaners (Whiteman and Côté, 2002). Longfin damselfish aggressively defend their territory from all intruders and are therefore rarely cleaned by facultative cleaners such as juvenile bluehead wrasse (Thalassoma bifasciatum) and Spanish hogfish (Bodianus rufus). Cleaning gobies represent the most common cleaners of longfin damselfish.

Damselfish nest monitoring

Forty active longfin damselfish nests were located haphazardly on the study site and monitored every 1–2 days from the beginning of April until the end of September 2001. Thirty territories had cleaning stations at the beginning of the study, whereas 10 did not have a cleaning station within 2 m of the territory. Each month, we recorded the number of clutches laid in each nest, as well as clutch area, on the day of laying (day 1). To estimate the number of eggs in a clutch, we measured egg density by counting eggs in three 0.25-cm2 squares located haphazardly in the clutch. Egg density counts were averaged and multiplied by clutch area. This method proved to be reasonably accurate (i.e., ±4.23% for clutch area, ±11.45% for total egg number) compared with counts of a sample of entire nests derived from video images (Cheney KL, unpublished data). To assess the percentage of eggs lost before hatching owing to filial cannibalism by the guarding male or to egg predators such as bluehead wrasse (Petersen, 1990), we remeasured the number of eggs in each clutch on day 4 or 5 of development.

Cleaner fish removal

Three months after the onset of nest monitoring, cleaners were removed from 20 territories, randomly chosen among those with cleaning stations. Cleaning gobies were anesthetized with clove oil, captured with a hand net, and then released onto other parts of the reef, away from the study area. There were thus 10 damselfish without cleaners throughout the study, 10 damselfish with cleaners throughout the study, and 20 damselfish that had cleaners for the first 3 months and subsequently lost them.

Any cleaner found relocating near the manipulated territories was promptly removed. This was an uncommon occurrence, however, because cleaners have small territories that they rarely leave. Moreover, because both cleaners and damselfish are diurnal, cleaning activities only occur during daylight periods (Cheney KL, personal observation). We are therefore confident that our manipulations successfully prevented damselfish from gaining access to cleaners.

Ideally, our cleaner removals should have been mirrored by cleaner addition experiments. Unfortunately, these manipulations could not be performed because of the reluctance of translocated cleaning gobies to remain in place.

Behavioral observations

During egg guarding, we quantified parental aggression by placing a model longfin damselfish adjacent to the nest of each of the 40 focal damselfish. The model was constructed by gluing a photograph of a longfin damselfish (standard length: 90 mm; average length of longfin damselfish males on Bellairs reef: 85.4 ± 5.7 mm; Cheney KL, unpublished data) on either side of a 5-mm-thick Perspex outline. The model was waterproofed with boat resin. The response of focal male damselfish to the model was tested on three occasions: immediately before and after cleaners were removed and again at the end of the study. On each occasion, the damselfish model was presented to each male for 1 min on each of five consecutive days, between 0900 and 1200 h. During these observations, we recorded the number of aggressive bites given to the model by the resident damselfish.

Damselfish ectoparasite load assessment

To assess ectoparasite loads, we removed six damselfish with cleaners and eight damselfish without cleaners after 3 months (i.e., just before cleaner removal), and 10 damselfish with cleaners and 11 damselfish without cleaners at the end of the study. Each individual was herded into a barrier net (4-mm mesh size), captured with a hand net, and quickly placed into a sealable plastic bag. Clove oil was sprayed into the bag causing the fish to die within 10 s with little evidence of pain or distress. Fish were then taken to the laboratory, where they were placed in a 0.4% chloretone bath (BDH Chemicals) for 1 h to remove ectoparasites as in Grutter (1995). In the analyses, we focussed on gnathiid isopod larvae because they are the main reported prey item of cleaner fish (see Arnal and Côté, 2000; Arnal and Morand, 2001a; Grutter, 1997a). Caligid copepods were also found on longfin damselfish but were rare. Their inclusion in the analyses did not change the results.

Statistical analyses

We first performed an overall comparison of damselfish reproductive output for males with and without cleaners, regardless of experimental treatment, with a general linear model using STATA version 7.0 (StataCorp). The following model was estimated:  

formula
where yit is the number of eggs, number of clutches, or total clutch area for fish i in month t. xit is the vector of explanatory variables pertaining to fish i in month t. β is the vector of regression parameters corresponding to xit, and ui is the random effect term, which allows for differences between fish. The other error term, ϵit, allows for differences between months for a given fish. Total number of eggs and total clutch area were log-transformed to correct the positive skews in these variables. Two fish did not receive eggs in 1 month; these were therefore omitted from the analysis.

We also conducted before-and-after paired comparisons on the 20 damselfish that had their cleaner experimentally removed. To remove the marked seasonal variation in reproductive parameters (see Results), we standardized data in these paired analyses by dividing the value for each damselfish by the average for that month.

RESULTS

Ectoparasite loads in relation to cleaner fish presence

Before cleaner fish removal, damselfish with a cleaning station within their territory had significantly fewer gnathiids than did damselfish without cleaning stations (Mann-Whitney U = 8.0; n1 = 6, n2 = 8, p =.043) (Figure 1). At the end of the study, that is, 3 months after cleaner fish removal, damselfish that had retained cleaning stations also had fewer ectoparasites than did damselfish that had their cleaning station removed experimentally (Mann-Whitney U = 11, n1 = 9, n2 = 10, p =.004) (Figure 1).

Spawning activity in relation to cleaner fish presence

There was a marked seasonal pattern in the number of eggs per clutch, the number of clutches received, and the area of each clutch (Table 1). These variables peaked in July (Table 1), when 93% more eggs were received for all 40 males than in April (z = 7.2, p <.001).

The presence of a cleaner had a slightly negative effect on mating success, reducing the number of eggs received by 9%, but this effect was not significant (z = −0.83, p =.40). Similarly, the number of clutches received was reduced by 28% and the area of each clutch by 20% in the presence of a cleaner, but again these reductions were not significant (clutch number: z = −1.23; p =.19; clutch area: z = −1.19, p =.23). The number of eggs lost before hatching varied from 1.2 % to 22.4 % (mean ± SD: 14.3% ± 6.2%). There was no difference in the percentage of eggs lost whether a cleaner was present or absent (z = −1.1; p =.25).

Paired comparisons of (standardized) reproductive data for males that had a cleaning station in the first trimester and subsequently lost it also revealed no differences in any reproductive parameters (number of eggs: paired t19 = 0.57, p =.58; clutch area: paired t19 = 0.42, p =.68; number of clutches: paired t19 = 0.83, p =.41; percentage of eggs lost: paired t19 = 0.17, p =.87). Moreover, there was no reproductive difference between June, just before cleaning goby removal, and September, 3 months after removal (number of eggs: t19 = 0.45, p =.66; clutch area: t19= 0.40, p =.69; number of clutches: t19 = −0.39, p =.70; percentage of eggs lost: t19 = −0.13, p =.90).

There was no correlation between spawning activity (number of eggs, number of clutches, area of clutches) or loss of eggs and ectoparasite load for damselfish examined in June (rs <.1; n = 14; p >.73 in all cases) or damselfish examined in September (rs <.24; n = 21; p >.36 in all cases).

Parental aggression

Male longfin damselfish varied greatly in parental aggression toward the damselfish model, with the average number of attacks ranging from 0.3–22.3 bites/min. The average brood age ranged from 1–4.3 days (mean ± SD = 2.2 ± 0.9). There was no correlation in any month between the average brood age and the number of bites given to the model (rs <.12; n = 20; p >.60 in all cases).

Aggression rate did not differ before cleaner removals, immediately after removals, or at the end of the study (repeated-measures ANOVA: F = 0.12, df = 2, p = 0.88). There was, however, a positive correlation between male reproductive success, measured as the total number of clutches obtained in June and July, and aggression (June: rs =.71, n = 20, p =.01; July: rs =.78, n = 20, p <.001; June and July combined: rs =.76, n = 20, p <.001). There was no correlation between aggression rate and the percentage clutch area lost in June, July, or September (rs <.30; n = 20; p >.21 in all cases). The number of gnathiids found on damselfish was not correlated with the number of bites given to the model in the month the damselfish was removed for ectoparasite examination (rs =.09, n = 20, p =.72).

DISCUSSION

In the present study, we considered for the first time a fitness-related consequence of being cleaned for client fishes. We found that longfin damselfish with cleaners present in their territories have significantly lower ectoparasite loads than did damselfish without cleaners nearby (see also Cheney and Côté, 2001). However, the benefit of ectoparasite removal did not appear to translate directly into a reproductive benefit or a parental care advantage. We found no difference in the number of eggs, number of clutches, or area of clutches received by damselfish with and without cleaners in their territory. No detrimental effects on reproduction were apparent for up to 3 months after cleaners were experimentally removed from damselfish territories. We also found no difference in parental male aggression between nest-guarding damselfish with and without cleaners.

Damselfish mating success depends on a wide variety of factors. For example, body size (Schmale, 1981), frequency of courtship (Thresher and Moyer, 1983), presence of eggs in the nest (Sikkel, 1989; Goulet, 1998), and nest site quality (Sikkel, 1995) can all contribute to mating success and could have confounded comparisons between males with and without cleaning gobies within their territories. However, there was no difference in total length between males with and without cleaners (t33 = −1.29, p =.21). Moreover, our experimental design in which the success of individual males was compared before and after removal of cleaning stations controlled for most variables, leaving access to cleaners, or lack thereof, as the only difference between males.

One possible reason for the lack of effect of being cleaned on damselfish reproduction is the small reduction in ectoparasite load experienced by cleaned clients. Male damselfish without cleaning stations, whether naturally or as a result of removal, had on average only two more gnathiids than did males with cleaners on their territories. Although gnathiids have been shown to harm clients in high numbers (Paperna and Por, 1977), these 1-mm-long ectoparasites probably have minimal deleterious effects to hosts at low densities. Levels of ectoparasites are naturally low on Barbadian longfin damselfish, reaching a maximum of only 12 gnathiids per fish (Cheney KL, unpublished data). In fact, all potential client species appear to have low ectoparasite loads in Barbados (Arnal et al., 2000; Sikkel et al., 2000). The scope for finding large benefits of being cleaned, both in terms of ectoparasite reduction and associated fitness consequences, may therefore be limited at this site. However, in other geographical areas where ectoparasite loads are higher, the benefits of being cleaned in terms of ectoparasite reduction are substantially greater (Cheney KL and Côté IM, unpublished data), and reproductive benefits may be measurable. Moreover, mating success and parental care ability are but two aspects of fitness. It remains possible that slightly lowered ectoparasite loads affect other components of fitness, such as survival, if, for example, gnathiids are vectors of disease (Honma and Chiba, 1991)

Determining the fitness consequences of being cleaned remains an important goal in the study of cleaning symbioses, because it will elucidate the nature of cleaning interactions. At present, two alternative perspectives of cleaner–client interactions exist. Cleaning symbioses are either mutualisms based on ectoparasite removal (Poulin and Grutter, 1996; Trivers, 1971) or parasitisms in which cleaners exploit the tendency of clients to seek tactile stimulation and thus gain access to food, including scales and mucus (Losey, 1979, 1987). The fact that cleaners do significantly reduce ectoparasite loads in some cleaner–client interactions (Grutter, 1999; Cheney and Côté, 2001) and that clients visit cleaning stations more and solicit cleaners for longer when they have higher ectoparasite loads (Arnal et al., 2000; Grutter, 2001; Sikkel et al., 2000) supports the mutualistic interpretation. However, we have yet to identify a clear fitness-related benefit of ectoparasite removal to firmly establish cleaning symbioses as mutually beneficial for its fish protagonists.

Figure 1

The number of gnathiid ectoparasites on individual longfin damselfish with and without cleaning stations in their territory after 3 months (i.e., before cleaner fish removal) and after 6 months (i.e., 3 months after cleaner fish removal). Mean parasite numbers are shown ±1 SE. Sample sizes are shown in parentheses

Figure 1

The number of gnathiid ectoparasites on individual longfin damselfish with and without cleaning stations in their territory after 3 months (i.e., before cleaner fish removal) and after 6 months (i.e., 3 months after cleaner fish removal). Mean parasite numbers are shown ±1 SE. Sample sizes are shown in parentheses

Table 1

Range and mean (± SD) monthly values for reproductive parameters measured for 40 male longfin damselfish over 6 months.

  Estimated no. of eggs per clutch
 
 No. of clutches per nest
 
 Area of individual clutch (cm2)
 
 
 No. of clutches Range Mean Range Mean Range Mean 
April 108 620–10060 4160 ± 2090 1–7 3.0 ± 1.6 5–82 31 ± 9.2 
May 128 670–12020 4630 ± 1890 1–8 3.5 ± 1.6 5–74 30 ± 7.8 
June 175 960–18000 6330 ± 2460 1–11 4.9 ± 2.2 8–150 34 ± 12.4 
July 198 900–12640 5600 ± 1360 1–10 5.1 ± 2.4 7–132 38 ± 11.7 
August 185 560–14030 4360 ± 1980 1–12 4.6 ± 2.7 5–174 35 ± 13.5 
September 162 340–11020 4020 ± 1870 1–10 4.5 ± 2.0 9–152 33 ± 10.7 
  Estimated no. of eggs per clutch
 
 No. of clutches per nest
 
 Area of individual clutch (cm2)
 
 
 No. of clutches Range Mean Range Mean Range Mean 
April 108 620–10060 4160 ± 2090 1–7 3.0 ± 1.6 5–82 31 ± 9.2 
May 128 670–12020 4630 ± 1890 1–8 3.5 ± 1.6 5–74 30 ± 7.8 
June 175 960–18000 6330 ± 2460 1–11 4.9 ± 2.2 8–150 34 ± 12.4 
July 198 900–12640 5600 ± 1360 1–10 5.1 ± 2.4 7–132 38 ± 11.7 
August 185 560–14030 4360 ± 1980 1–12 4.6 ± 2.7 5–174 35 ± 13.5 
September 162 340–11020 4020 ± 1870 1–10 4.5 ± 2.0 9–152 33 ± 10.7 

We thank everyone at the Bellairs Research Institute, especially its director, Dr. Bruce Downey, for logistical support. Thank you to Peter Moffatt and Douglas Yu for help with statistical analyses, to John Reynolds for comments on the manuscript, and to Véronique Binette for help in the field. K.L.C. was supported by a UK Biotechnology and Biological Sciences Research Council studentship, and we are grateful for the financial support of the John and Pamela Salter Charitable Trust and the University of East Anglia.

REFERENCES

Adlard RD, Lester RJG,
1995
. The life-cycle and biology of Anilocra pomacentri (Isopoda, Cymothoidase), an ectoparasitic isopod of the coral-reef fish, Chromis nitida (Perciformes, Pomacentridae).
Aust J Zool
 
43
:
271
-281.
Arnal C, Côté IM,
1998
. Interactions between cleaning gobies and territorial damselfish on coral reefs.
Anim Behav
 
55
:
1429
-1442.
Arnal C, Côté IM,
2000
. Diet of broadstripe cleaning gobies on a Barbadian reef.
J Fish Biol
 
57
:
1075
-1082.
Arnal C, Côté IM, Sasal P, Morand S,
2000
. Cleaner–client interactions on a Caribbean reef: influence of correlates of parasitism.
Behav Ecol Sociobiol
 
47
:
353
-358.
Arnal C, Morand S,
2001
. Cleaning behaviour in the teleost, Symphodus melanocercus: females are more specialized than males.
J Mar Biol Assoc UK
 
81
:
317
-323.
Arnal C, Morand S,
2001
. Importance of ectoparasites and mucus in cleaning interactions in the Mediterranean cleaner wrasse Symphodus melanocercus.
Mar Biol
 
138
:
777
-784.
Begon M, Harper JL, Townsend CR,
1990
. Ecology: individuals, populations and communities. London: Blackwell.
Cheney KL, Côté IM,
2001
. Are Caribbean cleaning symbioses mutualistic: costs and benefits of visiting cleaning stations to longfin damelfish.
Anim Behav
 
62
:
927
-933.
Cusack R, Cone DK,
1986
. A review of parasites as vectors of viral and bacterial diseases of fish.
J Fish Disease
 
9
:
169
-171.
Gorlick DL,
1980
. Ingestion of host fish surface mucus by the Hawaiian cleaning wrasse, Labroides phthirophagus (Labridae), and its effect on host species preference. Copeia 1980:863–868.
Gorlick DL, Atkins PD, Losey GS,
1987
. Effect of cleaning by Labroides dimidiatus (Labridae) on an ectoparasite population infecting Pomacentrus vaiuli (Pomacentriade) at Enewetak Atoll. Copeia 1987:41–45.
Goulet D,
1998
. Spawning success in the damselfish Amblyglyphidodon leucogaster: the influence of eggs in the nest.
Anim Behav
 
55
:
651
-664.
Grutter AS,
1995
. Comparison of methods for sampling ectoparasites from coral reef fishes.
Mar Fresh Res
 
46
:
897
-903.
Grutter AS,
1996
. Experimental demonstration of no effect by the cleaner wrasse Labroides dimidiatus (Cuvier and Valenciennes) on the host fish Pomacentrus moluccensis (Bleeker).
J Exp Mar Biol Ecol
 
196
:
285
-298.
Grutter AS,
1997
. Size-selective predation by the cleaner fish Labroides dimidiatus.
J Fish Biol
 
50
:
1303
-1308.
Grutter AS,
1997
. Spatiotemporal variation and feeding selectivity in the diet of the cleaner fish Labroides dimidiatus. Copeia 1997:346–355.
Grutter AS,
1999
. Cleaner fish really do clean.
Nature
 
398
:
672
-673.
Grutter AS,
2001
. Parasite infection rather than tactile stimulation is the proximate cause of cleaning behaviour in reef fish.
Proc R Soc Lond B
 
268
:
1361
-1365.
Honma Y, Chiba A,
1991
. Pathological changes in the branchial chamber wall of stingrays, Dasyatis spp., associated with the presence of juvenile gnathiids (Isopoda, Crustacea).
Fish Pathol
 
26
:
9
-16.
Lafferty KD, Morris AK,
1996
. Altered behaviour of parasitized killifish increases susceptibility of predation by bird final hosts.
Ecology
 
77
:
1390
-1397.
Limbaugh C,
1961
. Cleaning symbiosis.
Sci Am
 
20
:
42
-49.
Losey GS,
1972
. The ecological importance of cleaning symbioses. Copeia 1972:820–833.
Losey GS,
1979
. Fish cleaning symbiosis: proximate causes of host behaviour.
Anim Behav
 
27
:
669
-685.
Losey GS,
1987
. Cleaning symbioses.
Symbiosis
 
4
:
229
-258.
McFarland D,
1985
. Animal behavior. Menlo Park, California: Benjamin Cummings.
Møller AP, Christe P, Lux E,
1999
. Parasitism, host immune function, and sexual selection.
Q Rev Biol
 
74
:
3
-20.
Mugridge RER, Stallybrass HG,
1983
. A mortality of eels, Anguilla anguilla L., attributed to Gnathiidae.
J Fish Biol
 
6
:
81
-82.
Paperna I, Por FD,
1977
. Preliminary data on the Gnathiidae (Isopoda) on the northern Red Sea, the Bitter Lakes and the eastern Mediterranean and the biology of Gnathia piscivora n.sp.
Rapp Com Intern Med Medit
 
24
:
195
-197.
Petersen CW,
1990
. The occurrence and dynamics of clutch loss and filial cannibalism in two Caribbean damselfishes.
J Exp Mar Biol Ecol
 
135
:
117
-133.
Poulin R,
1994
. Meta-analysis of parasite-induced behavioural changes.
Anim Behav
 
48
:
137
-146.
Poulin R, Grutter AS,
1996
. Cleaning symbioses: proximate and adaptive explanations.
Bioscience
 
46
:
512
-517.
Robertson DR,
1990
. Differences in the seasonalities of spawning and recruitment of some small neotropical reef fishes.
J Exp Mar Biol Ecol
 
144
:
49
-62.
Robertson DR, Petersen CW, Brawn JD,
1990
. Lunar reproductive cycles of benthic brooding reef fishes: reflections of larval biology or adult biology?
Ecol Monogr
 
60
:
311
-329.
Schmale MC,
1981
. Sexual selection and reproductive success in males of the bicolor damselfish, Eupomacentrus partitus (Pisces: Pomacentridae).
Anim Behav
 
29
:
1172
-1184.
Sikkel PC,
1989
. Egg presence and developmental stage influence spawning site choice in an algal-nesting damselfish.
Bull Mar Sci
 
57
:
682
-689.
Sikkel PC,
1995
. Effects of nest quality of male courtship and female spawning-site choice in an algal-nesting damselfish.
Bull Mar Sci
 
57
:
682
-689.
Sikkel PC, Fuller CA, Hunte W,
2000
. Habitat/sex differences in time at cleaning stations and ectoparasite loads in a Caribbean reef fish.
Mar Ecol Prog Ser
 
193
:
191
-199.
Thompson JN,
1994
. The coevolutionary process. Chicago, Illinois: University of Chicago Press.
Thresher RE, Moyer JT,
1983
. Male success, courtship complexity and patterns of sexual selection in three congeneric species of sexually monochromatic and dichromatic damselfishes (Pisces: Pomacentridae).
Anim Behav
 
31
:
113
-127.
Trivers RL,
1971
. The evolution of reciprocal altruism.
Q Rev Biol
 
46
:
35
-57.
Whiteman EA, Côté IM,
2002
. Cleaning activity of two Caribbean cleaning gobies: intra- and interspecific comparisons.
J Fish Biol
 
60
:
1443
-1458.
Youngbluth MJ,
1968
. Aspects of the ecology and ethology of the cleaning fish, Labroides phthirophagus.
Z Tierpsychol
 
25
:
915
-932.