Cleaner fish drives local fish diversity on coral reefs

Coral reefs are one of the most diverse habitats in the world, yet our understanding of the processes affecting their biodiversity is limited. At the local scale, cleaner fish are thought to have a disproportionate effect, in relation to their abundance and size, on the activity of many other fish species, but confirmation of this species' effect on local fish diversity has proved elusive. The cleaner fish Labroides dimidiatus has major effects on fish activity patterns and may indirectly affect fish demography through the removal of large numbers of parasites. Here we show that small reefs where L. dimidiatus had been experimentally excluded for 18 months had half the species diversity of fish and one-fourth the abundance of individuals. Only fish that move among reefs, however, were affected. These fish include large species that themselves can affect other reef organisms. In contrast, the distribution of resident fish was not affected by cleaner fish. Thus, many fish appear to choose reefs based on the presence of cleaner fish. Our findings indicate that a single small and not very abundant fish has a strong influence on the movement patterns, habitat choice, activity, and local diversity and abundance of a wide variety of reef fish species.     

Long-term effects of the cleaner fish Labroides dimidiatus on coral reef fish communities

Cleaning behaviour is deemed a mutualism, however the benefit of cleaning interactions to client individuals is unknown. Furthermore, mechanisms that may shift fish community structure in the presence of cleaning organisms are unclear. Here we show that on patch reefs (61–285 m²) which had all cleaner wrasse Labroides dimidiatus (Labridae) experimentally removed (1–5 adults reef⁻¹) and which were then maintained cleaner-fish free over 8.5 years, individuals of two site-attached (resident) client damselfishes (Pomacentridae) were smaller compared to those on control reefs. Furthermore, resident fishes were 37% less abundant and 23% less species rich per reef, compared to control reefs. Such changes in site-attached fish may reflect lower fish growth rates and/or survivorship. Additionally, juveniles of visitors (fish likely to move between reefs) were 65% less abundant on removal reefs suggesting cleaners may also affect recruitment. This may, in part, explain the 23% lower abundance and 33% lower species richness of visitor fishes, and 66% lower abundance of visitor herbivores (Acanthuridae) on removal reefs that we also observed. This is the first study to demonstrate a benefit of cleaning behaviour to client individuals, in the form of increased size, and to elucidate potential mechanisms leading to community-wide effects on the fish population. Many of the fish groups affected may also indirectly affect other reef organisms, thus further impacting the reef community. The large-scale effect of the presence of the relatively small and uncommon fish, Labroides dimidiadus, on other fishes is unparalleled on coral reefs.     

The effect of the cleaner fish Labroides dimidiatus on the capsalid monogenean Benedenia lolo parasite of the labrid fish Hemigymnus melapterus

The cleaner fish Labroides dimidiatus affected the pigmented monogenean parasite Benedenia lolo on the fish Hemigymnus melapterus (Labridae) held in aquaria. The effect of cleaner fish varied with the size class of fish; only small fish [ a posteriori size class <11.5 cm standard length ( L _S)] exposed to cleaner fish had fewer monogeneans compared with fish not exposed to cleaner fish. The abundance of monogeneans on large fish ( a posteriori size class <11.5 cm L _S) was not affected by cleaner fish. The size-frequency distributions of monogeneans on both size-classes of H. melapterus were affected by cleaner fish. Fish exposed to cleaner fish had fewer large (>3 mm) and more small (<1 mm) monogeneans than fish not exposed to cleaner fish, suggesting cleaner fish selectively removed larger monogeneans. This difference was more pronounced on large fish. In the absence of cleaner fish, small fish had almost as many monogeneans as large fish; they also had more small monogeneans than the large fish, suggesting small fish were more vulnerable to infection by monogeneans than larger fish. This suggests that the cleaner fish L. dimidiatus has the potential to control benedeniine monogeneans on captive fish and highlights the importance of taking into account fish size in studies of the effect of cleaner fish on ectoparasites.     

Size-selective predation by the cleaner fish Labroides dimidiatus

A comparison of the size-frequency distribution of parasitic gnathiid isopod larvae in the diet of the cleaner fish Labroides dimidiatus and on six host fish species ( Chlorurus sordidus, Ctenochaetus striatus, Hemigymnus melapterus, Scolopsis bilineatus, Siganus doliatus, Thalassoma lunare ) was made on one occasion. The comparison was repeated with Hemigymnus melapterus on three occasions and between two islands in Australia. L. dimidiatus selected larger gnathiids at all times at Lizard Island but not at Heron Island. Size-selective predation by L. dimidiatus suggests any potential effect of cleaner fish on parasites may vary according to the size of parasite. However, this effect appears to vary spatially.     

Effect of the removal of cleaner fish on the abundance and species composition of reef fish

The ecological significance of cleaner fish on coral reefs was investigated. I removed all cleaner fish, Labroides dimidiatus, from eight small reefs, measured the subsequent effect on the abundance and species composition of all reef fish after 3 and 6 months, and compared it with eight control reefs with cleaner fish. The removal of cleaner fish had no detectable effect on the total abundance of fish on reefs and the total number of fish species at both times. Multivariate analysis by non-metric multidimensional scaling and ANOSIM pairwise tests based on 191 fish species revealed no effect of cleaners on the community structure of fish. Similar results were obtained using principal components analysis on subsets of the data using the 33 most common fish species and the 15 most abundant species (≥5 individuals per reef ) with both log10 (x + 1) transformed data and with fish numbers standardized for abundance. This study demonstrates that the removal of cleaner fish for 6 months did not result in fish suffering increased mortality nor in fish leaving reefs to seek cleaning elsewhere. Received: 28 October 1996 / Accepted: 7 February 1997     

Does client size affect cleaner fish choice of client? An empirical test using client fish models

To determine whether the choice of client fishes in the cleaner fish Labroides dimidiatus was influenced by client size, cleaner fish were given a choice of equal amount of food spread on large and small client redfin butterflyfish Chaetodon trifasciatus models. All large models received bites from cleaners compared to 27% for small models. Seventy‐nine per cent of cleaners took their first bite from the large fish model. The results suggest that client size may affect cleaner fish choice.     

Multiple cleaner species provide simultaneous services to coral reef fish clients

Cleaning symbioses on tropical coral reefs are typically documented between two species: a single client fish and one or more conspecific cleaners. However, multiple cleaner species living sympatrically in the Caribbean have been anecdotally reported to simultaneously clean the same client. Nothing is known about the patterns and processes driving these interactions, which may differ from those involving a single cleaner species. Here, we used remote underwater videography on three reefs in Honduras to record simultaneous cleaning interactions involving Pederson''s cleaner shrimp (Ancylomenes pedersoni) and cleaner gobies (Elacatinus spp.). A pilot study on adjacent shrimp and goby stations found interactions were always initiated by shrimp. A larger, multi-year dataset shows cleaner gobies joined 28% of all interactions initiated at A. pedersoni cleaning stations with cleaner gobies residing nearby. Client body size significantly predicted simultaneous cleaning interactions, with 45% of interactions simultaneous for clients greater than 20 cm total body length compared with only 8% for clients less than 20 cm. We also found that simultaneous cleaning interactions lasted over twice as long as shrimp-only interactions. We propose these novel multi-species interactions to be an ideal model system to explore broader questions about coexistence, niche overlap and functional redundancy among sympatric cleaner species.     

Cleaning up the biogeography of Labroides dimidiatus using phylogenetics and morphometrics

Cleaner fishes are some of the most conspicuous organisms on coral reefs due to their behaviour and prominent body pattern, consisting of a lateral stripe and blue/yellow colouration. All obligate cleaner fishes share this body stripe pattern, which is an important signal for attracting client fishes. However, variability in the cleaning signal of the cleaner fish Labroides dimidiatus has been documented across its range. Here, we investigate the geographic distribution of cleaner signal polymorphisms in L. dimidiatus and contrast this to phylogeographic variation in mitochondrial (mt) DNA. We used samples from 12 sites for genetic analyses, encompassing much of L. dimidiatus’ range from the Red Sea to Fiji. We obtained morphometric measures of the cleaner signal body stripe width from individuals among six of the sites and qualitatively grouped tail stripe shape. mtDNA control region sequences were used for phylogenetic and population genetic analyses. We found that body stripe width was significantly correlated with tail stripe shape and geographical location, with Indian Ocean populations differing in morphology from western Pacific populations. L. dimidiatus haplotypes formed two reciprocally monophyletic clades, although in contrast to morphology, Japanese cleaner fish fell within the same clade as Indian Ocean cleaner fish and both clade types were sympatric in Papua New Guinea. An additional novel finding of our research was that the inclusion of two closely related cleaner fish species, Labroides pectoralis and Labroides bicolor, in the phylogenetic analysis rendered L. dimidiatus polyphyletic. Overall, the findings suggest the diversity within L. dimidiatus is underestimated.     

Cleaning behavior is riskier and less profitable than an alternative strategy for a facultative cleaner fish

Cleaning symbioses on coral reefs involve small cleaner fish or shrimps picking ectoparasites from the exterior surfaces of larger client organisms. These mutualisms are thought to evolve in part because the cleaner receives a reliable source of profitable prey items and immunity from predation. However, the benefits of cleaning behavior have never been measured relative to those of alternative, non-cleaning strategies. This study examined these costs and benefits in the sharknose goby, Elacatinus evelynae, a facultative cleaner fish, at the Caribbean island of St. Croix. Sharknose gobies are found on coral heads, where they maintain cleaning stations, and on basket sponges, where they spend little time cleaning and feed predominantly on nonparasitic copepods. For immature gobies that are not allocating energy to reproduction, hindcast otolith growth rates (a reliable proxy for somatic growth) were significantly higher for non-cleaning sponge-dwellers than for coral-dwellers. Furthermore, tagging large, competitively dominant gobies on sponges and corals revealed that mortality rates were higher for coral-dwelling gobies. These unexpected results provide further evidence that cleaning mutualisms are context dependent: far from being a uniformly profitable life history strategy, cleaning may be a suboptimal choice at some times and places.     

Arginine vasotocin regulation of interspecific cooperative behaviour in a cleaner fish

In an interspecific cooperative context, individuals must be prepared to tolerate close interactive proximity to other species but also need to be able to respond to relevant social stimuli in the most appropriate manner. The neuropeptides vasopressin and oxytocin and their non-mammalian homologues have been implicated in the evolution of sociality and in the regulation of social behaviour across vertebrates. However, little is known about the underlying physiological mechanisms of interspecific cooperative interactions. In interspecific cleaning mutualisms, interactions functionally resemble most intraspecific social interactions. Here we provide the first empirical evidence that arginine vasotocin (AVT), a non-mammalian homologue of arginine vasopressin (AVP), plays a critical role as moderator of interspecific behaviour in the best studied and ubiquitous marine cleaning mutualism involving the Indo-Pacific bluestreak cleaner wrasse Labroides dimidiatus. Exogenous administration of AVT caused a substantial decrease of most interspecific cleaning activities, without similarly affecting the expression of conspecific directed behaviour, which suggests a differential effect of AVT on cleaning behaviour and not a general effect on social behaviour. Furthermore, the AVP-V1a receptor antagonist (manning compound) induced a higher likelihood for cleaners to engage in cleaning interactions and also to increase their levels of dishonesty towards clients. The present findings extend the knowledge of neuropeptide effects on social interactions beyond the study of their influence on conspecific social behaviour. Our evidence demonstrates that AVT pathways might play a pivotal role in the regulation of interspecific cooperative behaviour and conspecific social behaviour among stabilized pairs of cleaner fish. Moreover, our results suggest that the role of AVT as a neurochemical regulator of social behaviour may have been co-opted in the evolution of cooperative behaviour in an interspecific context, a hypothesis that is amenable to further testing on the potential direct central mechanism involved.     

Simple ecological trade-offs give rise to emergent cross-ecosystem distributions of a coral reef fish

Ecosystems are intricately linked by the flow of organisms across their boundaries, and such connectivity can be essential to the structure and function of the linked ecosystems. For example, many coral reef fish populations are maintained by the movement of individuals from spatially segregated juvenile habitats (i.e., nurseries, such as mangroves and seagrass beds) to areas preferred by adults. It is presumed that nursery habitats provide for faster growth (higher food availability) and/or low predation risk for juveniles, but empirical data supporting this hypothesis is surprisingly lacking for coral reef fishes. Here, we investigate potential mechanisms (growth, predation risk, and reproductive investment) that give rise to the distribution patterns of a common Caribbean reef fish species, Haemulon flavolineatum (French grunt). Adults were primarily found on coral reefs, whereas juvenile fish only occurred in non-reef habitats. Contrary to our initial expectations, analysis of length-at-age revealed that growth rates were highest on coral reefs and not within nursery habitats. Survival rates in tethering trials were 0% for small juvenile fish transplanted to coral reefs and 24-47% in the nurseries. As fish grew, survival rates on coral reefs approached those in non-reef habitats (56 vs. 77-100%, respectively). As such, predation seems to be the primary factor driving across-ecosystem distributions of this fish, and thus the primary reason why mangrove and seagrass habitats function as nursery habitat. Identifying the mechanisms that lead to such distributions is critical to develop appropriate conservation initiatives, identify essential fish habitat, and predict impacts associated with environmental change.     

Patterns of Local Distribution of Labroides Dimidiatus in French Polynesian Atolls

Cleaning behaviour by fishes has been described to be either facultative or specialised. If being specialised in cleaning is more advantageous than to be a facultative cleaner, cleaner fishes would prefer to settle on reefs were ecological conditions promote specialisation. To test this hypothesis, we looked at the influence of physical and ecological factors on cleaner wrasse abundance, studying variations in L. dimidiatus density between 10 atolls of French Polynesia. We used a multiple regression method based on permutations of distance matrices. Our study reveals that L. dimidiatus are significantly more numerous at sites where ecological factors, such as the species richness of the fish community, may promote their cleaning activity. Moreover, they were significantly more abundant in atolls presenting a large number of sedentary fish species, few predators and few fishes living in large groups. Finally, physical factors, such as the distance between atolls, the lagoon surface, the mean percentage of hard substrate, the mean current index, and the mean depth, play a minor role in the occurrence of L. dimidiatus.     

Mutualistic cleaner fish initiate trait-mediated indirect interactions by influencing the behaviour of coral predators

1.?Indirect interactions resulting from changes in organismal traits such as behaviour [i.e. trait-mediated indirect interactions (TMIIs)] are widespread in biological communities, yet few studies have explored the potential for mutualisms to initiate TMIIs. 2.?This study used a combination of behavioural observations and manipulative field experiments to investigate potential TMIIs resulting from a mutualism between specialized cleaner fish and the 'clients' that visit cleaners for the removal of ectoparasites. 3.?Behavioural observations indicate that the bluestreak cleaner wrasse, Labroides dimidiatus, increases local predation pressure on corals at cleaner stations by attracting corallivorous butterflyfish to their territories. 4.?Observations of the ornate butterflyfish, Chaetodon ornatissimus, suggest a trade-off between seeking cleaning and foraging; individuals decreased their foraging rate at cleaner stations and shifted their diet to include a greater proportion of less preferred prey items. Nonetheless, predation pressure on corals was higher at cleaner stations because the spatial response of butterflyfish to cleaners more than compensated for their lower foraging rates. 5.?The results of a field experiment suggest that the greater predation pressure observed at cleaner stations may be sufficient to reduce the growth rate of the unpreferred coral Porites rus. 6.?Together, these results emphasize the need to consider mutualists as potential initiators of TMIIs and highlight the importance of integrating individual movement into conceptual analyses of TMIIs.     

Relationship between roving behaviour and the diet and client composition of the cleaner fish Labroides bicolor

Diet analyses and observations of cleaning behaviour of two cleaner fishes revealed that Labroides bicolor fed more on client mucus, but Labroides dimidiatus fed more on ectoparasites, and that L. bicolor interacted with fewer species (36 species) compared with L. dimidiatus (44 species). The client species which contributed most to the dissimilarity between cleaner species were the dusky farmerfish Stegastes nigricans and bicolor chromis Chromis margaritifer damselfishes, which L. dimidiatus interacted with more often than L. bicolor, and the striated Ctenochaetus striatus and brown Acanthurus nigrofuscus surgeonfishes, which L. bicolor interacted with more; L. bicolor interacted with all parrotfishes (Scaridae) more. These results confirm the importance of repeated interactions and partner choice in determining the nature of interactions in mutualisms.     

Larval growth history determines juvenile growth and survival in a tropical marine fish

Processes that occur around the transition between larval and juvenile life‐stages can have a major effect on the population dynamics of organisms with complex life cycles. We explore the roles of larval history and selective post‐settlement mortality in determining the growth and survival of newly settled individuals of the damselfish, Pomacentrus amboinensis (Pomacentridae). Specifically, we determine whether the direction and intensity of selection on the recruits differs among various size‐classes of predators. A mark‐recapture study showed that individuals who survived 9 or more days were significantly larger at settlement than those that died within the first day (12.3 vs 11.9 mm SL), when mortality was highest (25% d^?1). Censuses revealed that the species and size composition of piscivores differed markedly between two reef habitats where P. amboinensis was common. A cage experiment, conducted in both habitats, manipulated the sizes of predators that could access newly settled P. amboinensis to determine whether the resulting mortality of the recently settled fish was influenced by larval growth history or size at settlement. Ten days after the start of the experiment individuals that grew slowly in the second half of their larval life had been lost from most of the experimental treatments. Small fish were also selectively lost from the coarse‐mesh cage on the reef base. Significant positive relationships between pre‐ and post‐settlement growth rates were found in both habitats for the fine mesh cages, cage controls and open patch reefs. This relationship was reversed in the coarse mesh cages in both habitats. This growth compensation was facilitated through the action of a particular size range of predators, whose impact was disrupted or masked in the open treatments by the action of a diverse predator pool. The present study underscores the complexity of the processes that influence the early post‐transition growth and survival in organisms with complex life‐histories.     

Small‐scale demographic variation in the stoplight parrotfish Sparisoma viride

Age‐based analysis of the stoplight parrotfish Sparisoma viride was used to examine whether observed differences in their abundance and size structure among reefs in a cross‐shelf portion of the upper Florida Keys could be explained by variation in demographic rates. Annual and daily sagittal otolith increments were enumerated for 176 individuals collected from replicates of reefs in two strata, inshore and offshore reefs (2–6 m depth). von Bertalanffy growth functions fitted to size‐at‐age plots for each site were similar between reefs within each stratum (inshore and offshore), but differed between strata. Sparisoma viride on offshore reefs attained greater average standard length (L_S) at age, greater mean asymptotic size and were longer lived than fish from inshore reefs. Fish on inshore reefs attained only half the maximum age observed on offshore reefs (4 v. 8 years, respectively). No terminal phase fish >4 years of age were found on either reef type. Estimates of mortality rates from age‐frequency data of collected fish revealed higher mortality on inshore reefs. Demographic variables obtained in this study were similar to published values for S. viride from Caribbean reefs but differed significantly from published values from reefs at a similar latitude (Bahamas), reflecting high demographic plasticity on both local and regional scales.     

Consequences of the founder effect in the genetic structure of introduced island coral reef fish populations

Between 1955 and 1961, the Division of Fish and Game of the State of Hawaii (USA) undertook an introduction program which brought 11 species of families Serranidae and Lutjanidae (Pisces) from French Polynesia to the coral reefs off Oahu and Big Island in Hawaii. Only three— Cephalopholis argus, Lutjanus fulvus and Lutjanus kasmira —for which we have records of locations and number offish released, are known to have become established. Comparison of allozyme distribution of C. argus and L. kasmira between individuals collected in Hawaiian and Polynesian populations provided a unique opportunity to estimate the impact of genetic drift and selection processes caused by a founder event. We used temporal variance of allelic frequencies to estimate and validate effective population size within marine fish populations. Despite the fact that only 571 C. argus and 2435 L. kasmira were released, we did not observe major changes in polymorphism and heterozygosity. Using different models to estimate effective population size from temporal variance in allelic frequencies and the number of generations, we estimate that the effective population size is about 1–5% of the total population size. Such reduced effective population size explains why most of the species introduced in Hawaii (8/11) failed to become established. Our results have implications for conservation biology because they emphasize that in spite of the fact that only a few individuals bequeathed their characteristics to subsequent generations, no significant change in genetic diversity was observed; success of introduced species is therefore limited by the number of fish released.     

Seascape-scale trophic links for fish on inshore coral reefs

It is increasingly accepted that coastal habitats such as inshore coral reefs do not function in isolation but rather as part of a larger habitat network. In the Caribbean, trophic subsidies from habitats adjacent to coral reefs support the diet of reef fishes, but it is not known whether similar trophic links occur on reefs in the Indo-Pacific. Here, we test whether reef fishes in inshore coral, mangrove, and seagrass habitats are supported by trophic links. We used carbon stable isotopes and mathematical mixing models to determine the minimum proportion of resources from mangrove or seagrass habitats in the diet of five fish species from coral reefs at varying distances (0–2,200 m) from these habitats in Moreton Bay, Queensland, eastern Australia. Of the fish species that are more abundant on reefs near to mangroves, Lutjanus russelli and Acanthopagrus australis showed no minimum use of diet sources from mangrove habitat. Siganus fuscescens utilized a minimum of 25–44 % mangrove sources and this contribution increased with the proximity of reefs to mangroves (R ² = 0.91). Seagrass or reef flat sources contributed a minimum of 14–78 % to the diet of Diagramma labiosum, a species found in higher abundance on reefs near seagrass beds, but variation in diet among reefs was unrelated to seascape structure. Seagrass or reef flat sources also contributed a minimum of 8–55 % to a fish species found only on reefs (Pseudolabrus guentheri), indicating that detrital subsidies from these habitats may subsidize fish diet on reefs. These results suggest that carbon sources from multiple habitats contribute to the functioning of inshore coral reef ecosystems and that trophic connectivity between reefs and mangroves may enhance production of a functionally important herbivore.     

Interactions between gnathiid isopods,cleaner fish and other fishes on Lizard Island,Great Barrier Reef

The rate of emergence of micropredatory gnathiid isopods from the benthos, the proportion of emerging gnathiids potentially eaten by Labroides dimidiatus, and the volume of blood that gnathiids potentially remove from fishes (using gnathiid gut volume) were determined. The abundance (mean ±s.e .) of emerging gnathiids was 41·7 ± 6·9 m^?2 day^?1 and 4552 ± 2632 reef^?1 day^?1 (reefs 91–125 m²). The abundance of emerging gnathiids per fish on the reef was 4·9 ± 0·8 day^?1; but excluding the rarely infested pomacentrid fishes, it was 20·9 ± 3·8 day^?1. The abundance of emerging gnathiids per patch reef was 66 ± 17% of the number of gnathiids that all adult L. dimidiatus per reef eat daily while engaged in cleaning behaviour. If all infesting gnathiids subsequently fed on fish blood, their total gut volume per reef area would be 17·4 ± 5·6 mm³ m^?2 day^?1; and per fish on the reefs, it would be 2·3 ± 0·5 mm^?3 fish^?1 day^?1 and 10·3 ± 3·1 mm³ fish^?1 day^?1 (excluding pomacentrids). The total gut volume of gnathiids infesting caged (137 mm standard length, L_S) and removed from wild (100–150 mm L_S) Hemigymnus melapterus by L. dimidiatus was 26·4 ± 24·6 mm³ day^?1 and 53·0 ± 9·6 mm³ day^?1, respectively. Using H. melapterus (137 mm L_S, 83 g) as a model, gnathiids had the potential to remove, 0·07, 0·32, 0·82 and 1·63% of the total blood volume per day of each fish, excluding pomacentrids, caged H. melapterus and wild H. melapterus, respectively. In contrast, emerging gnathiids had the potential of removing 155% of the total blood volume of Acanthochromis polyacanthus (10·7 mm L_S, 0·038 g) juveniles. That L. dimidiatus eat more gnathiids per reef daily than were sampled with emergence traps suggests that cleaner fishes are an important source of mortality for gnathiids. Although the proportion of the total blood volume of fishes potentially removed by blood‐feeding gnathiids on a daily basis appeared to be low for fishes weighing 83 g, the cumulative effects of repeated infections on the health of such fish remains unknown; attacks on small juvenile fishes, may result in possibly lethal levels of blood loss.     

Cleaner wrasse mimics inflict higher costs on their models when they are more aggressive towards signal receivers

Aggressive mimics are predatory species that resemble a 'model' species to gain access to food, mating opportunities or transportation at the expense of a signal receiver. Costs to the model may be variable, depending on the strength of the interaction between mimics and signal receivers. In the Indopacific, the bluestriped fangblenny Plagiotremus rhinorhynchos mimics juvenile cleaner wrasse Labroides dimidiatus. Instead of removing ectoparasites from larger coral reef fish, fangblennies attack fish to feed on scales and body tissue. In this study, juvenile cleaner wrasse suffered significant costs when associated with P. rhinorhynchos mimics in terms of reduced cleaning activity. Furthermore, the costs incurred by the model increased with heightened aggression by mimics towards signal receivers. This was apparently because of behavioural changes in signal receivers, as cleaning stations with mimics that attacked frequently were visited less. Variation in the costs incurred by the model may influence mimicry accuracy and avoidance learning by the signal receiver and thus affect the overall success and maintenance of the mimicry system.