Colour patterns influence symbiosis and competition in the anemonefish–host anemone symbiosis system

Abstract

Colour patterns in fish are often used as an important medium for communication. Anemonefish, characterized by specific patterns of white bars, inhabit host anemones and defend the area around an anemone as their territory. The host anemone is used not only by the anemonefish, but also by other fish species that use anemones as temporary shelters. Anemonefish may be able to identify potential competitors by their colour patterns. We first examined the colour patterns of fish using host anemones inhabited by Amphiprion ocellaris as shelter and compared them with the patterns of fish using surrounding scleractinian corals. There were no fish with bars sheltering in host anemones, although many fish with bars were found in surrounding corals. Next, two fish models, one with white bars and the other with white stripes on a black background, were presented to an A. ocellaris colony. The duration of aggressive behaviour towards the bar model was significantly longer than that towards the stripe model. We conclude that differences in aggressive behaviour by the anemonefish possibly select the colour patterns of cohabiting fish. This study indicates that colour patterns may influence not only intraspecific interactions but also interspecific interactions in coral reef ecosystems.

1. Introduction

Colour patterns in fish are known to be shaped by a variety of selective pressures, including predators, prey, competitors and mate choice (e.g. [1–6]). In particular, fish inhabiting coral reefs have the most diverse pigment cell types of any vertebrate, resulting in a wide variety of colour patterns, such as bars, stripes and spots [7]. In coral reefs with high water transparency, visual signals are effective communication tools, and the colour patterns of fish play important roles in determining the behaviour between individuals and species, such as, camouflage/mimicry [8–14], species/individual identification [15–17], courtship [10,18–20] and other social interactions [21,22]. Competition and symbiosis can be established using such visual information, and as a result, there must be rules regarding the composition of colour patterns within the local community. However, there have been no studies focusing on the role of visual information in determining the species composition of fish communities.

Anemonefish (Amphiprioninae, Pomacentridae) have conspicuous white bars against a background colour of orange, red or black, and the number of white bars varies depending on the species [7,23–25]. The evolutionary function of colour patterns in anemonefish is poorly understood, but at least three adaptive hypotheses have been proposed [7,26–29]. The first hypothesis is that the number of white bars has a recognition function since anemonefish have species-specific numbers of white bars. This hypothesis is supported by the fact that the differences in the number of white bars between species inhabiting the same area are significantly greater than would be expected at random [27]. In addition, anemonefish patterns change during ontogeny in some species [7,23–25,27], and the different patterns of juveniles from adults may be a dishonest signal to conceal their presence and reduce agonistic interactions [26,27,30]. The second hypothesis is that the contrast of the bright base colour and white stripes is disruptive and functions to hide the fish silhouette. In Amphiprion ocellaris and Amphiprion percula, in particular, an indentation on the dorsal fin and the white bars extending across the dorsal fin appear to have a high disruptive colour effect [27]. The third is an aposematism hypothesis, that is, the conspicuous colour patterns serve to advertize the toxicity of the host anemone. Phylogenetic analyses have revealed that host venom strength and tentacle length are correlated with colour patterns of anemonefish, supporting their function as warning colours [29].

Anemonefish inhabit host anemones for their whole life except for their pelagic larval stage (e.g. [31–33]). Host anemones have numerous nematocytes that sting most fish, but anemonefish can use host anemones as shelter thanks to their mucus, which prevents stinging [31,34,35]. Anemonefish defend their host anemone as territories and express aggressive behaviour towards other anemonefish and towards other fish species [36–39]. Despite this, the host anemone inhabited by anemonefish pair may be used as a temporary refuge by other species of fish, such as damselfish (Pomacentridae), cardinalfish (Apogonidae) and wrasses (Labridae) [40–42]. In the Ryukyu Archipelago, Japan, 16 species from three families (Apogonidae, Labridae and Pomacentridae) have been observed inhabiting host anemones, and Labridae come to host anemones to clean anemonefish [42].

Dascyllus trimaculatus (Pomacentridae), which is the most frequentspecies using anemonefish colonies, has a black ground colour with three small white spots. When a model of this pattern was presented to six species of anemonefish, aggressive behaviour was observed in response to the models, but there were species differences in the frequency of this behaviour [41]. Dascyllus trimaculatus tended to use colonies of Amphiprion sandaracinos, which showed relatively less aggressive behaviour, but not colonies inhabited by Amphiprion frenatus or Amphiprion polymnus, which showed more aggressive behaviour [41]. Further, during our studies on anemonefish–host anemone symbioses in the Ryukyu Archipelago, we noticed that fish species other than anemonefish that use host anemones have stripes and spots but not patterns with vertical bars [42]. This may be because the aggressive behaviour of anemonefish is less frequent towards fish without bars or stripes, such as D. trimaculatus, and towards fish with stripes, such as Labroides dimidiatus (Labridae), but more frequent towards fish with bars. Therefore, we hypothesized that anemonefish may decide whether to tolerate or exclude potential intruders based on their colour patterns, and that their behaviour may influence fish community structure.

The purpose of this study was to examine the relationships between the frequency of anemonefish aggressive behaviour and the colour patterns of fish by clarifying the following two questions: (i) Within the same habitat, is there a difference in the colour patterns of fish that live in host anemones (with anemonefish) and in scleractinian corals (without anemonefish)? and (ii) Do anemonefish differ in the frequency of their aggressive behaviour toward fish with vertical bars and horizontal stripes? The answers to these two questions will help clarify the effects of anemonefish behaviour on fish communities in the anemonefish–host anemone symbiosis system.

2. Material and methods

(a) . Colour patterns of fish using scleractinian corals and host anemones

We conducted field surveys at five study sites (two study sites on reefs around Miyakojima Island, one around Ishigakijima Island and two around Iriomotejima Island) in the Ryukyu Archipelago from September 2020 to October 2021 (figure 1). Amphiprion ocellaris and their host Stichodactyla gigantea were targeted in each study area, which ranged from 0 to 2 m in water depth. The fish fauna of a total of 49 individuals of S. gigantea and that of the 49 scleractinian corals that were closest (2–10 m away) to each of the S. gigantea were recorded (Miyakojima: n = 18, Ishigakijima: n = 6, Iriomotejima: n = 25). The scleractinian corals studied ranged from ca 20 to 60 cm with a mean of 35 cm (s.d. = 6.38) in diameter, and the host anemones ranged from 20 to 47 cm with a mean of 31 cm (s.d. = 10.14). The areas within the outermost range of the host anemone's tentacles and scleractinian coral colony branches were defined as the target area, including spaces under the tentacles/branches, between the tentacles/branches, and above the tentacles/branches [42]. To minimize the influence of the observer on the behaviour of anemonefish, based on the results of [43], we slowly approached the host anemone/scleractinian coral from ca 2 m and started recording at a distance of 0.5 m. In order to compare the fish species using the host anemone and scleractinian coral as shelter, we recorded around the target area for 3 min using a video camera (Olympus TG-6). Fish species swimming within the target area for a minimum of 2 min were identified as fish using the host anemone/scleractinian coral, thereby excluding from the analysis fish species that briefly entered the host anemone/scleractinian coral [44]. The colour patterns of fish using host anemones/scleractinian corals were divided into three categories: bars (vertical bars) (figure 2a), stripes (horizontal stripes) (figure 2b) and others (figure 2c). Fishes that had both bars and stripes, such as butterflyfish, were counted as ‘others’. Differences in the frequency of colour patterns of resident fish between host anemones and corals were compared using a χ-squared test of independence conducted in IBM SPSS Statistics v. 28.

Figure 1.

Map of the study site in the Sakishima Islands. Arrows indicate the study sites. (Online version in colour.)

Figure 2.

Photographs and colour pattern classifications of the various fish species observed in scleractinian corals and host sea anemones in the present study. Species indicated with asterisks were found in both corals and sea anemones. (Online version in colour.)

(b) . Aggressive behaviour of anemonefish toward bar and stripe patterns

We conducted behavioural experiments on 45 A. ocellaris colonies (Miyakojima: 15 colonies, Ishigakijima: 6 colonies, Iriomotejima: 24 colonies). After 3 min recording fish species using host anemones as described above, we started behavioural experiments by setting the model fish. Two types of fish models were prepared: one with two white vertical bars on a black background (bar model) and the other with two white horizontal stripes on a black background (stripe model). Fish-shaped plastic toys of 5 cm in total length were painted with a permanent marker (Mitsubishi Paint Marker PX-20, Mitsubishi Pencil Corporation, Tokyo, Japan) in black for the base colour and white for the bars and stripes (electronic supplementary material, figure S1). The reflectance spectra of colours were determined by means of a Konica Minolta CM-700D spectrophotometer every 10 nm from 400 to 700 nm using a 30 mm diameter illumination area (electronic supplementary material, figure S1). The three attributes of colour (L: lightness, C: chroma, H: hue) of the models were measured by a colorimeter (NR-11A; Nippon Denshoku Industries, Tokyo, Japan). Measurements were conducted three times and the mean value of the base colour was 2.1, 1.9, 312.0 (L, C, H), and that of bar/stripe was 71.6, 6.7, 111.7 (L, C, H).

These models were dangled by a transparent fishing line and placed in close proximity to each A. ocellaris colony. As aggressive behaviour of anemonefish toward intruders is known to decrease gradually over 2 min [41,43], we continued recording for 3 min. To prevent habituation to the model, each colony was tested only once, and one of the two different models was randomly presented to each A. ocellaris colony. Following previous studies, both ‘chasing’ and ‘biting’ were collectively defined as aggressive behaviour towards the model fish [41,43,45]. To calculate the difference in frequency of aggressive behaviour between bar and stripe models, and sexual differences in the frequency of aggressive behaviour, a non-parametric Mann–Whitney U-test conducted in IBM SPSS Statistics v. 28 was used.

3. Results

(a) . Differences in colour patterns of fish using host anemone and scleractinian corals

Throughout all study sites, 19 out of 49 individuals (39%) of S. gigantea anemones were used by other fish species (table 1a). Three species of cardinalfish and three species of damselfish were observed during the course of our study (table 1a). Throughout all study sites, all 49 individuals (100%) of scleractinian corals were used by other fish species (table 1b). A total of 26 species were observed in scleractinian corals, consisting of 4 species of cardinalfish, 3 species of wrasse, 12 species of damselfish, 2 species of butterflyfish, 2 species of surgeonfish, 1 species of goby, 1 species of blenny and 1 species of snapper (table 1b). Five of 26 species (Ostorhinchus properuptus, Cheilodipterus quinquelineatus, D. trimaculatus, Chrysiptera cyanea and Pomacentrus chrysurus) used both host anemones and scleractinian corals (table 1, figure 2). D. trimaculatus and O. properuptus tended to use host anemones rather than scleractinian corals (table 1). Most fish observed in host anemones and scleractinian corals were immature fish.

Table 1.

List of fish species inhabiting (a) host anemones and (b) scleractinian corals (Pocilloporidae, Acroporidae and Poritidae) at each island.

family of coexisting fish species	coexisting fish species	colour patterns	number of host anemones/scleractinian corals inhabited by each fish species			total	percentage in 49 host anemones/scleractinian corals (%)
			Miyakojima	Ishigakijima	Iriomotejima
(a) host anemones
Apogonidae	Apogon nigrofasciatus	stripes	1	0	0	1	2.0
	Cheilodipterus quinquelineatus	stripes	4	0	0	4	8.2
	Ostorhinchus properuptus	stripes	4	0	2	6	12.2
Pomacentridae	Chrysiptera cyanea	others	2	4	0	6	12.2
	Dascyllus trimaculatus	others	4	2	3	9	18.4
	Pomacentrus chrysurus	others	1	1	0	2	4.1
(b) scleractinian corals
Apogonidae	Cheilodipterus quinquelineatus	stripes	2	0	0	2	4.1
	Ostorhinchus sp.	stripes	2	0	2	4	8.2
	Ostorhinchus ishigakiensis	others	0	0	1	1	2.0
	Ostorhinchus properuptus	stripes	2	0	1	3	6.1
Labridae	Hemigymnus melapterus	bars	0	0	1	1	2.0
	Stethojulis strigiventer	others	1	0	1	2	4.1
	Thalassoma hardwicke	bars	1	1	0	2	4.1
Pomacentridae	Abudefduf sexfasciatus	bars	0	2	0	2	4.1
	Chromis viridis	others	2	0	1	3	6.1
	Chrysiptera biocellata	bars	0	0	2	2	4.1
	Chrysiptera cyanea	others	13	5	23	41	83.7
	Dascyllus aruanus	bars	15	1	8	24	49.0
	Dascyllus trimaculatus	others	0	1	1	2	4.1
	Dischistodus prosopotaenia	bars	1	0	0	1	2.0
	Neoglyphidodon nigroris	stripes	1	0	1	2	4.1
	Pomacentrus amboinensis	others	0	0	1	1	2.0
	Pomacentrus chrysurus	others	1	2	8	11	22.4
	Pomacentrus moluccensis	others	4	3	3	10	20.4
	Stegastes punctatus	others	2	0	4	6	12.2
Chaetodontidae	Chaetodon auriga	others	0	0	7	7	14.3
	Chaetodon lunulatus	others	0	0	1	1	2.0
Acanthuridae	Acanthurus triostegus	bars	0	0	1	1	2.0
	Zebrasoma velifer	bars	1	0	2	3	6.1
Gobiidae	Asterropteryx semipunctata	others	0	0	1	1	2.0
Blenniidae	Meiacanthus kamoharai	stripes	0	0	1	1	2.0
Lutjanidae	Lutjanus gibbus	others	0	0	2	2	4.1

Of the six fish species observed in host anemones, three had stripes, the other three had other types of patterns, and no fish species had bar patterns (table 1a, figures 2 and 3). On the other hand, 8 of the 26 species that used corals had bar patterns, 5 species had stripes and 13 species had other patterns (table 1b, figures 2 and 3). There was a significant difference in the frequency of colour patterns of fish using host anemones and corals (figure 3, chi-squared test, χ² = 14.27, d.f. = 2, p < 0.01). In host anemones, there were no fish species with bar patterns but fish species with stripe patterns were more abundant (figure 3a). On the other hands, in corals, fish species with bar patterns were found more frequently than fish species with stripe patterns (figure 3b). This tendency was similarly observed in sites at Miyakojima, Ishigakijima and Iriomotejima islands (figure 3).

Figure 3.

The number of (a) host anemones and (b) scleractinian corals that were inhabited by each colour pattern of fish. (Online version in colour.)

(b) . Differences in the frequency of aggressive behaviour toward bar and stripe models

In female anemonefish, the duration of aggressive behaviour towards the bar model was 3 s per 3 min (median), 0–10.5 s (25–75% quartiles), 0–40 s (min.–max. range), and that towards the stripe model was 0 s per 3 min (median), 0–4 s (25–75% quartiles), 0–15 s (min.–max. range), and the former was significantly longer than the latter (figure 4a; Mann–Whitney U-test; U = 150, p < 0.05). In males, there was no significant difference in duration of aggressive behaviour between the bar (0 s per 3 min median, 0–2 s 25–75% quartiles, 0–12 s min.–max. range) and stripe (0 s per 3 min, 0–2.5 s 25–75% quartiles, 0–5 s min.–max. range) models (figure 4b; Mann–Whitney U-test; U = 200.5, p = 0.84). No aggressive behaviour was observed in immature fish. The duration of aggressive behaviour toward the bar models was significantly longer in females than males (Mann–Whitney U-test; U = 142.5, p < 0.01). On the other hand, there was no significant difference in the duration of aggressive behaviour toward the stripe model between males and females (Mann–Whitney U-test; U = 174.5, p = 0.63).

Figure 4.

Box plot showing the duration of aggressive behaviour of (a) female and (b) male Amphiprion ocellaris towards the bar model and the stripe model during 180 s videos. Crosses represent means, boxes represent interquartile ranges (25–75%), centre lines represent medians, whiskers represent ranges excluding outliers, and circles represent outliers. Significant differences by Mann–Whitney Utest indicated by asterisks; *p < 0.05. (Online version in colour.)

4. Discussion

The present study showed that fish species apart from anemonefish that use host anemones did not have vertical bar patterns, but that nearby scleractinian corals were used by a variety of fish species with various colour patterns, including bars. The fact that only 39% of anemones were used as shelter by other fish, while 100% of scleractinian corals were used by fish, may be due to the lack of tolerance to the venom of the host anemone by most fishes. However, this observation may also be linked to the aggressive behaviour of anemonefish, which actively chase away many intruding fish. Therefore, we hypothesize that the reason for this difference was the aggressive behaviour of anemonefish, which defend host anemones as their territory. We further hypothesize that the behaviour of anemonefish differs depending on the colour pattern of the intruder fish.

When either the bar or stripe modelwas presented to colonies of anemonefish, we observed that female anemonefish attacked the bar model more persistently and for a longer time than the stripe model. When a similar experiment was conducted using a model with white spots on a black background (imitating D. trimaculatus), the duration of aggressive behaviour by A. ocellaris was 0 s per3 min (median) with 0–2 s (25–75% quartiles) and 0–3 s (min.–max. range) in females and 0 s per 3 min (median) with 0–0 (25–75% quartiles) and 0–1 (min.–max. range) in males [41]. Although comparisons between the current study and this past research should be made with caution owing to the use of different study sites, the duration of aggressive behaviour did not differ between the D. trimaculatus model and the striped model, but was much higher in the bar model. As expected, anemonefish responded more aggressively to intruders with bar patterns and protected the host anemone as their own territory, indicating that only fish species without bar patterns may have access to host anemones. Our results strongly suggest that the fish community around host sea anemones is likely to be biased by the preferences of anemonefish. In order to further develop the present results, future comparisons of aggressive behaviour of anemonefish towards a variety of colour patterns at different sites are needed.

Territorial and social fish, like the Ambon damsel (Pomacentrus amboinensis) or cichlid species (Cichlidae), have been shown to use colour patterns to distinguish between competitors and mates [1,15–17,46]. According to [1], dorsoventral bars, especially in the case of melanin pigmentation, are associated with intraspecific aggression in various species of fish. For example, the Lake Tanganyika cichlid Tropheus sp. has a yellow bar. Tropheus sp. tends to less frequently attack individuals with wider yellow bars [46]. Also, the cichlid Neolamprologus pulcher has been shown to attack individuals with vertical lines lighter and shorter than its own [47]. Most previous studies have focused on the significance of bar patterns in intraspecific competition. In this study, we suggest that aggressive behaviour of anemonefish towards intruders with bar patterns (presumably toward anemonefish) may also have a secondary effect on interspecific interactions.

By contrast, cranio-caudal stripes are often associated with intraspecific cooperation, such as shoaling in cichlids, zebrafish and rainbow fish [1]. Moreover, stripes can be a signal of interspecific cooperation, such as cleaning behaviour [44,48,49]. In fact, five species of wrasse that clean anemonefish in the Ryukyu Archipelago have a striped pattern [42]. In the cleaner wrasse, L dimidiatus, a longer striped pattern is known to attract comparatively more fish [49]. Zhokhov et al. [50] reported that 11 parasite species (two species of copepods, six digeneans, two nematodes and one acanthocephalan) have been found in anemonefish in the South China Sea. Thus, the presence of cleaner fish, which feed on body surface parasites, should be beneficial to anemonefish. As all the cleanerfish in the study site have horizontal stripes, it was not possible to confirm whether the anemonefish recognized the cleanerfish species or judged them solely on the basis of their stripe patterns, which remains an issue to be examined in the future. It is also necessary to focus on how cleanerfish behave in response to aggressive behaviour by anemonefish.

In this experiment, females were more aggressive than males toward bar models. In most species of anemonefish, females are larger than males and are often more aggressive [41,43]. However, in laboratory conditions, female A. ocellaris displayed fierce aggressive behaviour toward female intruders, whereas males were more often aggressive to male intruders [39]. Thus, we theorize that if the size of the models was reduced to 3 cm, equivalent to the total length of male A. ocellaris, males may have become more aggressive. Iwata & Manbo [39] reported that immature intruders were rarely attacked under laboratory conditions, suggesting that size as well as colour pattern is involved in aggressive behaviour in intraspecific competition.

At least seven species of anemonefish are known to use S. gigantea as a host [35]. In the present study site, the Ryukyu Archipelago, Amphiprion clarkii and Amphiprion perideraion are known to use S. gigantea, but only A. ocellaris inhabited the S. gigantea examined in the present study. Moreover, A. ocellaris does not co-inhabit with other species of anemonefish [51,52]. Thus, A. ocellaris may be aggressive not only against competitors of the same species, but also against other anemonefish species that may use the same host species. In this experiment, we used a model with two white bars on a black background, which is similar to A. clarkii. However, A. ocellaris has a pattern of three white bars, and A. perideraion has one white bar. Since anemonefish are known to recognize the same species by visual cues [28], it is possible that they can identify the numbers of white lines and alter their behaviour based on such information.

The present results, together with previous reports [26,27,30], suggest that the white bars of anemonefish have two different effects: (i) they allow the regulation of the relationship within anemonefish in the case of bispecific colonies [53] or when immature fish recruits enter the host, and (ii) they act as a selection filter to tolerate other species that use the sea anemone as a shelter. Future work needs to more precisely examine the aggressive behaviour of anemonefish towards various bar patterns and sizes, including patterns of the same and different species of anemonefish, as well as fish species other than anemonefish. Results from these future studies may allow us to better understand how the various patterns are recognized and to pinpoint the basic rules controlling this inter- and intraspecific signalling.

On tropical and subtropical coral reefs, the presence of ‘shelters’ such as zooxanthellate scleractinian corals is known to be important for the coexistence of many species of fish in these ecosystems (e.g. [54–58]). Similarly, host sea anemones inhabited by anemonefish play an important role in maintaining the diversity of coral reef ecosystems [42]. Investigating the rules of colour patterns that determine the behaviour of anemonefish will help us to understand the community structure in the anemonefish–host anemone symbiosis.

Ethics

All of the experimental procedures described were approved by University of the Ryukyus and Okinawa Institute of Science and Technology (OIST: FWA-2021-003).

Data accessibility

All data are available as part of supplementary material and in table 1 [59].

Authors' contributions

K.H.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, resources, visualization, and writing—original draft; K.T.: project administration, supervision, and writing—review and editing; J.D.R.: project administration, supervision, and writing—review and editing; V.L.: funding acquisition, project administration, supervision, and writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

The present study was partly supported by a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (grant no. 20J11845) and a research scholarship from the OIST Kicks programme.