1 Introduction

The hermit crab-anemone (HCA) partnership is a facultative mutualism common in the marine environment. The sea anemone protects its host from predators such as cephalopods and shell crushing crabs by its stinging cells. The crab provides the anemone with a hard substrate and increased access to oxygenated water and food. In addition, the hermit crab also protects the anemone from predators by moving away or attacking them with its claws. Hermit crabs establish most partnerships by detaching anemones and placing them on their shells. Initially the HCA partnership was assumed to be simple (Gosse 1860; Cowles 1920; Burton 1969), however, research during the last fifty years revealed that these associations are extremely diverse and complex.

The late D.M. Ross of the University of Alberta, Edmonton, Canada, contributed to our knowledge and understanding of HCA partnerships more than any other researcher. Ross studied HCA partnership formation in a variety of species, mimicry of the crab activity by tactile and electric stimulation of the anemone, neuro-muscular activity in complex anemone behaviors, intra and interspecific competition for actinians, costs and benefits for both partners in these associations and their evolution. Ross findings were summarized in a series of reviews (Ross 1967, 1974a, 1974b, 1983), with the last one published forty years ago. Ten years ago, a review on the HCA partnership was published by Antoniadou et al. (2013) with emphasis given to clarification of the term symbiosis, taxonomy and distribution of associated hermit crabs and sea anemones, costs and benefits of the association for anemones and crabs and the evolution of these partnerships.

My own acquittance with the HCA partnership started while reading many years ago the beautiful studies of D.M. Ross and watching hermit crabs associated with sea anemones during night dives in the Red-Sea. Recently while writing a review on boxer crabs associated with sea anemones (Schnytzer et al. 2022) we often contrasted these associations with HCA partnerships. Following these comparisons, I realized that the complexity and diversity of these amazing partnerships deserve an updated overview for their own merit. The aim of this review is to focus on the behavioral interactions between hermit crabs and sea anemones starting with the studies of D.M. Ross and following up with the more recent studies carried out in the field and laboratory. Here, I included issues such as formation of the associations, the early history of these partnerships, placement of anemones on the hermit crab shell, competition amongst hermit crabs over sea anemones, costs and benefits of the association for crab and anemone and the evolution of these partnerships based on molecular phylogenies using nuclear and mitochondrial markers.

2 General taxonomy and distribution

Sea anemones are the only symbionts among the cnidarians associated with hermit crabs which are actively hosted by them and not haphazardly fixed on their shells during larval settlement (e.g., hydrozoans) (Williams and McDermot, 2004). There are 35 species of symbiotic sea anemones belonging to 14 genera and seven families: Hormathiidae (e.g., Calliactis and Paracalliactis), Sagartiidae (e.g., Carcinactis and Verrillactis), Actinidae (e.g., Stylobates), Actinostolidae (e.g., Antholoba), Aiptasiidae (e.g., Aiptasia), Gonactiniidae (e.g., Gonactinia) and Sagartiomorphidae (e.g., Sagartiomorphe). Sea anemones of the family Hormathiidae which possess acontia loaded with nematocysts represent three quarters of all anemone symbionts associated with hermit crabs. Currently there are 41 hermit crab species that host sea anemones. These species belong to 15 genera and three families – Diogenidae (e.g., Dardanus), Paguridae (e.g., Pagurus) and Parapaguridae. (e.g., Parapagrus). Caution is required in appraising reports of occasional occurrence of actinians on shells of certain hermit crabs not usually found associated since such a host might have acquired the anemones by moving into a shell formerly occupied by another pagurid with the symbiotic habit. Associated HCA are common in temperate and tropical waters of the Atlantic, Indian, and Pacific oceans excluding the poles (see Antoniadou et al. 2013 for a review on HCA taxonomy and distribution in temperate seas). HCA partnerships are found from the shallow infra littoral to deep waters. Five symbiotic Stylobates spp. anemones for example are found over a wide range of depth approximately from 300 m to 900 m in the Indo-West Pacific (Crowther et al. 2011). Two Pacific abysal hermit crabs Tylaspis anomala and Probeebei mirabilis which range in depths of about 4000 m have completely abandoned the use of gastropod shells. However, at least Probeebei mirabilis has been observed carrying on its pleon a so-far not identified anemone (Lemaitre 1998; Cuvelier et al. 2023).

A new type of facultative association between hermit crabs and sea anemones was recently discovered in a shallow bay off the Yucatan peninsula, Mexico (Colombara et al. 2017). Clusters of blue legged hermit crabs Clibanarius tricolor were observed clinging to the stalks of the giant sun anemone, Stichodactyla helianthus. More recently, juvenile hermit crab assemblages including Clibanarius tricolor were observed in Belize in the vicinity of the pedal disc of Stichodactyla helianthus, however, not on the anemone stalk (Brooker et al. 2019). Two additional hermit crab species in the Mediterranean, Clibanarius erytrophus and Eupagurus cuanerisis were similarly observed in the vicinity of the pedal disc of Anemonia sulcata (Patzner 2004; Calado et al. 2007). So far, in the partnerships between hermit crabs and sea anemones the crab has always been the host to a smaller sized anemone whereas in these cases the hosts are the much larger sea anemones.

2.1 Hermit crabs and the gastropod shell

There are over 800 species of anomuran hermit crabs which typically possess an asymmetrical soft, not calcified abdomen, protected by an empty gastropod shell. The hermit crab abdomen is twisted to the right to fit the usually left twisted gastropod shell. The crab’s telson and uropods serve as hooks and clamps to hold onto the columella (i.e., the inner lip of the shell aperture) of the gastropod shell and to allow the crab’s rapid withdrawal into the shell in the event of any sudden danger. The possession of a shell serves in addition to protection from predators also in protection from temperature fluctuations, water loss and salinity stress particularly in terrestrial species. Hermit crabs select their shells according to numerous parameters such as shell size, weight, internal space, aperture size and shape and angle of columella axis. As the crab grows the gastropod shell must be often replaced. Hermit crabs may exchange their old shell for a new empty shell that they find. Inspection and assessment of the suitability of a new shell may take up to 5 min. During this process, the new shell is rolled around until the aperture is available for inspection, then the chelipeds are inserted inside the shell before the first insertion of the abdomen. The crab may alter its position several times between the new and the old shell before deciding to adopt the new shell or reject it as unsuitable. Once a crab decides to occupy a new shell it vacates its previous shell which becomes available to another crab and so on. This process has been termed “the vacancy chain process” in which resources are passed from one individual to another (Chase et al. 1991). More commonly crabs swap shells, often involving intense fighting between crabs over shell ownership, termed shell fights. The fights over the shells cause little damage to the contestants probably due to complex ritualized fighting (Elwood and Neil 1992). Shell size and their size relative to that of the hermit crab have major consequences on the crab’s fitness. In any population, hermit crabs occupying shells of preferred size are rare and most individuals occupy shells that are either too large or too small. A too small shell may result in an actual reduction of size following molting, reduction in egg numbers and greater vulnerability to predators. Too large of a shell also renders the crab more susceptible to predator attacks and may have a negative effect upon reproduction and an additional energetic load imposed by the extra weight. According to Hazlett (1995) hermit crabs modify their preference for shell size according to shell availability. When shells are scarce hermit crabs select larger shells which will permit substantial growth and delay the need to replace their shells (Wada et al. 1997; Gherardi 2006). However, when shells are plentiful the hermit crabs will select the better fitting one. In HCA partnerships, shell preferences by hermit crabs are also affected by their associated sea anemone species. Pagurus prideauxi mainly associated with Adamsia palliata (syn. Calliactis palliata Gusmao et al. 2019) that usually covers the entire shell, selects small and damaged shells. In this case the shell primarily serves as a substrate for the settlement of the anemone whereas the protective function of the shell is less important. In the same locality (i.e., Sicilian Channel, southern Italy) and depth, Paguristes eremita often associated with Calliactis sp. uses larger heavy and well conserved shells which serve a double role, substrate for the anemone and a protective resource for the crab (Caruso et al. 2003). The sea anemones associated with hermit crabs are attached and attracted to their shells and therefore the crab’s shell is considered a crucial component of the anemone-hermit crab symbiosis (see Hazlett 1981a; Lancaster 1988 for overviews of hermit crab-gastropod shell relationships).

Two abyssal hermit crabs, Tylaspis anomala and Probeebei mirabilis (family Parapaguridae), which completely abandoned the use of gastropod shells, share many morphological traits. Their abdomens are membranous (except for weak calcification of the first tergite and second pleura) without the usual twisting seen in typical shell-inhabiting hermit crabs. The two species possess long spinose ambulatory legs and the propodus of their fourth periopod possess a strong ventrodistal spine which at least in Probeebei mirabilis is used for grasping an anemone. The scarcity of calcareous shells which tend more rapidly to dissolve at great depths probably influenced abdomen evolution in these hermit crabs (Lemaitre 1998; Cuvelier et al. 2023).

2.2 Establishing the association

Accorrding to Ross (1983) the earliest accounts on the transfer of anemones to hermit crab shells started as soon as these organisms could be observed in aquaria. These early descriptions (Brunelli 1910, 1913; Faurot 1910; Cowles 1919, 1920) emphasized the leading role of the crab in anemone transfer. Only years later was the important role of the anemone in the transfer process realized.

The current recommended procedure for studying the formation of the HCA partnerships is carried out in aquaria by recording encounters between hermit crabs without anemones and anemones attached to rocks or glass plates. The encounters must be carefully observed aided with time lapse cinematography (one frame every 2 to 3 s) which facilitate the resolution and timing of relative slow and complex behavioral components. The use of surveillance video with infrared lighting at night will probably allow the capture of undisturbed hermit crab-anemone interactions in the field. The three main options for the formation of the HCA partnerships consist of either the anemone or the crab playing the main role or both partners contributing substantially to this task. Ross (1983) found no standard formula among pagurids and actinians species pairs; each species pair seems to have evolved a behavior pattern in which a greater or lesser share of the activity is assigned to the host or the symbiont.

2.2.1 Anemone dominant role

In the Atlantic association between Calliactis parasitica and Pagurus bernhardus the anemone actively attaches itself to the hermit crab shell in five steps, with the crab remaining mainly passive. 1. Attachment of the anemone tentacles and oral disc to the shell. 2. Detachment of the pedal disc. 3. Somersaulting movement of the detached pedal disc to the shell. 4. Settling of the pedal disc on the shell. 5. Release of tentacles and normal posture assumption. During the entire process, the anemone does not cling to the crab legs or chelae but to the shell. Calliactis failed to cling to shells boiled in alkali to remove superficial organic deposits or to shells coated with plastic or plaster of Paris molded from shells, providing additional evidence that tactile contact is not sufficient to evoke the clinging response. The active shell factor could not be extracted from the shells by either acetone, ethyl alcohol or ether; it may be a quinone-tanned protein or a muco polysaccharide. Even shells dry for 18 months are still effective (Ross 1959; Ross and Sutton 1961a). The chemical structure of the shell factor has not yet been determined. Similarly, in the association between Paracalliactis mediterranea and Pagurus variabilis the crab seems to be indifferent to the anemone. This association is again mostly formed by the anemone activity which includes a tongue shaped extension of the pedal disc (Fig. 1) to reach the gastropod shell (Ross and Zamponi 1982).

Fig. 1
figure 1

Detached Paracalliactis mediterranea climbing on a shell. (A) Tentacles clinging and pedal disc becoming mobile (4 min.). (B) Pedal disc expanded unilaterally with “tongue” directed towards shell (6 min.). (C) Expanded pedal disc close to shell (15 min.). (D) Pedal disc settling on shell, tentacles free of shell and column assuming erect position (32 min.). Times are from beginning to end of anemone activity. (Ross and Zamponi 1982)

2.2.2 Crab dominant role

In the Mediterranean associations between Calliactis parasitica and Pagurus arrosor the crab plays the dominant role in forming the partnership, described by the following four stages. 1. Contact between the crab shell and the anemone tentacles. 2. Tapping and probing the anemone stalk with the crab’s claws and legs leading at first to contraction of the anemone but subsequently to the anemone relaxation. 3. Anemone detachment by insertion of the crab’s claws and legs under the pedal disc. 4. Toppling of the anemone and its transfer to the crab by pressing the pedal disc to the shell. Early studies overlooked the role played by the anemone and ascribed the entire transfer process to the crab, ignoring the crucial importance of the contact of the anemone tentacles with the shell. However, it later appeared that the crab only facilitates the settlement of the anemone to its shell and makes the association more certain. Calliactis parasitica from the Atlantic does behave similarly to this species originating from the Mediterranean, namely being transferred largely by the crab Pagurus arrosor (Ross and Sutton 1961b). Ingeniously, Ross and Sutton managed to mimic the anemone probing and tapping with the crab’s claws and legs by tactile stimulation of the anemone with pipe cleaners, leading to pedal detachment. A similar result was also obtained by electric stimulation of the symbiotic sea anemones (Ross and Sutton, 1969, 1970). In comparative studies around the globe (e.g., Japan, New-Zealand, Puerto Rico and Hawaii) the release of sea anemones from the substrate was commonly achieved by tapping the anemone stalk and base with the claws and legs (Fig. 2) of different hermit crab species (Cutress and Ross 1969; Ross 1975; Ross and Wada 1975; Hand 1975). All the studied sea anemones were members of the family Hormathiidae. Recently the hermit crab Dardanus deformis was found in Japan (Yoshikawa et al. 2018) to use a similar tactile stimulation namely tapping for releasing sea anemones from two different families (e.g., Verrilactis sp. Fam. Sagaritiidae, and Calliactis sp. Fam. Hormathiidae). A much slower tapping rhythm was monitored in Parapagurus dimorphus from New-Zealand, one tap per 8s to 14s. The slower rhythm of this species was ascribed to its lower ambient temperature (i.e., 8-130C) compared with that of the high rate tapping species (i.e., about one tap per 1s) occurring in warmer waters (26-300C) (Hand 1975). Study of the anemone transfer behavior in several species of hermit crabs revealed that only about half of the crabs ”the performers” engage in such behaviors, whereas the rest “the non-performers” ignored the sea anemones. In Pagurus arrosor the performers were mainly females but in other species this behavior was not related to sex (Ross and Sutton 1961b; Cutress and Ross 1969; Hand 1975). Contrast of freshly caught crabs with crabs maintained over a long period in captivity failed to provide any clue to this phenomenon, remaining an enigma.

Fig. 2
figure 2

Dardanus gemmatus detaching Calliactis polypus from a stone and placing the anemone on its shell. (A) C. polypus beginning to relax after first contacts (2 min.); (B) Column extending and pedal disc beginning to lift (4.5 min.); (C) Anemone immediately after detachment (6.0 min.); (D) Anemone held by crab so that pedal disc will settle on right side of shell (7.5 min.). Times are from beginning to end of crab activity. Scale 5 cm. (Ross 1975)

Currently, the anemone and the intact shell are considered cardinal in the formation of the associations, with the hermit crab playing only a minor role. A non-symbiotic sea anemone cannot be induced to associate with a hermit crab despite all its efforts including probing and tapping and exposure of the sea anemones to the crab’s shell. Symbiotic sea anemones cannot be transferred to a shell that is coated with plastic or was boiled in alkali. According to Ross “It is surely one of nature’s finest ironies that these animals (i.e., hermit crab and sea anemone) so often found together are bound by their independent responses to the shells left behind by mollusks no longer living”.

2.3 Placement of sea anemones on hermit crab shells

The spatial distribution of sea anemones on shells of hermit crabs has been described in several associations (e.g., Calliactis parasitica-Dardanus arrosor; C. tricolor-P. pollicaris and C. tricolor-Dardanus pedunculatus) as non-random (Balasch et al. 1977; Brooks 1989; Giraud 2011). The distribution pattern was ascribed to crab activity since anemones alter only little their position after being placed on the shell (Brooks 1991b). It was suggested that the anemones were placed on the shell to maximize their protection from octopuses and to adjust the shell center of gravity. An experimental study for testing these hypotheses was carried out with Pagurus pollicaris and Calliactis tricolor (Brooks 1989). Hermit crabs without anemones were given access to detached anemones under six conditions, in which shell balance was manipulated by adding weight to the right side of the shell, left side of the shell or unaltered shell weight, in the presence and absence of predator (Octopus joubini) chemical cues. In all trials, the anemones were placed on the shell non-randomly. In trials with unaltered shell weight, the crabs placed the anemones mainly on top and right side of the shell probably to counterbalance the natural inclination of the shell to the left. However, in the presence of a predator cue from the octopus the anemones were placed more bordering the shell aperture than in the control. When weight was added to the shell, the anemones were placed in locations furthest from the weight. The predator chemical cue did not affect the anemone placement in trials with added weight. Thus, balance seems to be the most important factor affecting anemone placement in the studied species with predator deterrence playing a secondary role. The effect of the position of the anemone Calliactis tricolor relative to the shell aperture on crab protection was tested with Pagurus pollicaris and Octopus joubini. Anemones positioned within 5 mm of the shell aperture were significantly more effective in protecting the hermit crab from the octopus compared with anemones positioned at least 20 mm from the shell aperture (Brooks 1988).

Hermit crabs differ in their placement of anemones of different sizes and species on their shells. Dardanus venosus placed large Calliactis tricolor on top of the shell whereas small ones bordered the shell aperture. This hermit crab in Japan placed the small sea anemone Verrillactis sp. on the columella of its shell aperture, and larger Calliactis sp. on the dorsal surface of the shell (Yoshikawa et al. 2018). Crabs also prefer large anemones for placement while small ones were often eaten (Brooks 1991b). Diogenes edwardsi positions a single Sagartia paguri on the outer surface of its claw (Fig. 3), while the larger not yet described actinia A is placed on its shell. The crab detaches actinia A by stroking, tapping and lifting movements and settles it on its shell. However, in case of Sagartia paguri the crab taps it only once or twice with its claw and then positions the claw on the anemone’s tentacular crown which caused the anemone to detach and to cling to the claw where it soon settled. The crabs performed complex discriminatory behavior treating differently the two sea anemones in both detachment and placement (Ross 1975; Ross and Wada 1975). A further twist in shell settlement by anemones is due to a temporary avoidance of locations on shells previously occupied by anemones (Ross and Wada 1975). Calliactis polypus newcomers avoid previously occupied locations for at least 2 h following their abandonment probably to avoid overcrowding. However, the effect of hermit crabs on anemone preferred or non-preferred locations on the shell is so far unknown.

Fig. 3
figure 3

Diogenes edwardsi and Sagartia paguri. A & B. anemone on chela of crab emerging from shell. C. Anemone on claw of crab withdrawn into shell. D. Chela of crab showing smooth area usually occupied by the anemone. Scale: A, B, & C. 0.5 cm; D. 0.5 cm (Ross 1975)

2.4 The early formation of the partnership

Very little was known for a long time concerning the early establishment of the HCA partnerships. Collecting specimens in shallow water by snorkeling and SCUBA diving, and in deeper water with trawls, revealed large to medium sized associated HCA and no associated small individuals of anemones and crabs. Some information on the early formation of the partnership between Pagrus prideauxi and Adamsia palliata was already provided more than a hundred years ago by Faurot (1910) who described in the laboratory how the hermit crab picks up a tiny cylindrical anemone and settles it on its shell. More insight on the early formation of the HCA partnerships was gained by a field study carried out in the Aegan Sea by trawling at depths of 20 to 40 m, aiming at the size/age of populations of Calliactis parasitica and the symbiotic hermit crabs Paguristes erenita and Pagurus excavatus (Christidis et al. 1997). Anemones with a diameter of less than 5 mm and hermit crabs with a cephalothorax of less than 7 mm were not found living symbiotically at the studied area. It was suggested that as soon as this size threshold is crossed a change occurs in anemone and crab behavior, actively seeking to initiate symbiosis. This drive may be related to greater nutritional needs of the anemone and greater needs for defense from predators of the hermit crabs.

In a comprehensive survey, populations of several species of hermit crabs and their associated epibionts were monitored on Dog Island, Florida, over a period of ten years (Sanford 2003). In this study 173 hermit crabs were mainly associated with Calliactis tricolor of relatively large size (i.e., pedal disc diameter exceeding 1 cm).

The early formation of the associations between anemones and hermit crabs may involve frequent “stealing” of anemones. Small Calliactis parasitica (i.e., 10 mm p.d.) were found only on the shell of the submissive hermit crab species, Paguristes oculatus and P. alatus but none on the dominant species Dardanus arrosor. The shells of the submissive hermit crabs probably provide a preferred location for growth to maturity of Calliactis parasitica. D. arrosor steals the anemones only after they approach adult size (Fig. 4) when they provide a more effective defense against predators (Ross 1979).

Fig. 4
figure 4

Dardanus arrosor (Da) removing Calliactis parasitica from Paguristes oculatus (Po). Drawn from frames of 16 mm film. Times between A and B: 6 min.; B and C: 8 min.; C and D: 11 min. (Ross 1983)

The earliest bond between hermit crabs and sea anemones was discovered unexpectedly during Indo-West Pacific deep water plankton hauls. At great depths between 2000 and 5000 m every giant glaucothoe larva (young crab stage) of the hermit crab Paguropsis typica had one attached non-identified sea anemone belonging to the order Actiniaria, suborder Nynantheae, possibly a new species and genus. The bell-shape anemone was attached to the thorax of the glaucothoe allowing free movement of the first pleopods and the three periopods. The collected anemones were sexually mature producing sperm and ova in separate individuals and probably having a short life span. Following metamorphosis at depths around 200 m the anemones were probably lost, whereas the hermit crab acquired a colonial zoanthid Epizoanthus paguropsides for its protection (Schafer et al. 1983).

2.5 Hermit crab immunity to sea anemones

Little is known about the immunity of hermit crabs to the anemone nematocysts. This is surprising given the intimate relationship of these partners and their frequent observation in aquaria. Ross and Sutton (1961b) observed Dardanus arrosor without a shell attempting to transfer a Calliactis parasitica to the shell he did not possess. However, as soon as its non-calcified naked abdomen was exposed to the anemone tentacles, the crab retreated, sometimes fatally affected by this contact.

The following three mechanisms allowing crustaceans to live in close contact with sea anemones without being stung were reviewed by Karplus (2014). 1). The crustacean covers its body with the anemone secretions becoming chemically camouflaged (Levine and Blanchard 1980). 2). The crustacean upon contact with a sea anemone may secrete a chemical that acts on its own and inhibits nematocyte discharge by binding to a receptor site of the nematocyte (Crawford 1992). 3). Like the previous hypothesis except that the chemical secreted by the crustacean combines with an anemone derived substance (Crawford 1992). The association between the hermit crab Clibanarius tricolor which cling to the stalk of Stichodactyla helianthus was investigated to better understand hermit crab immunity. The four treatment groups tested on the giant sea anemone included groups of crabs removed from the anemone, crabs collected from the substrate at least 5 m from the anemone, crabs collected from the anemone whose surface was wiped underwater with a cloth and small stones of the same size as the crabs, collected at least 5 m from the anemone. The results were clear, showing that the anemone tentacles did not adhere to crabs removed from the anemone, whereas the anemone tentacles adhered to the subjects of the remaining three groups. Clearly, the mucus that covered the hermit crab rendered it immune. However, knowledge of the source of this mucus, its nature, and the way it is spread on the crab’s body require further investigation.

2.6 Partner specificity

Only certain species of hermit crabs form partnerships with specific species of sea anemones. In Japan, non-symbiotic crabs paired with symbiotic sea anemones failed completely to show any signs of activity towards these anemones. Moreover, pairing 18 symbiotic hermit crabs with non-symbiotic sea anemones ended with the same result. In five of the trials, the crabs showed no activity towards the anemones. In the remaining trials, the crabs tapped and prodded the anemones for only about 10–15 s, seemingly testing the anemone before rejecting it. The crabs seem to distinguish anemones that are not potential partners using the anemone response to these first contacts. These trials provide further evidence for the cooperative formation of the crab-anemone bond (Ross 1975).

An overall lack of partner specificity was expressed by Ross and Zamponi (1982) stating that symbiotic actinians can associate with any pagurid that is suitable in size, habits (i.e., symbiotic) and location. Sixty-eight different types of anemone-hermit crab partnerships have been reported in the literature. No quantitative analysis of the specificity was so far carried out with only the occurrence or non-occurrence of certain combinations noted. The hermit crab Dardanus arrosor hosts the largest number of sea anemones, i.e., seven species, followed by Pagurus alatus, P. bernhardus and P. cuanensis and Paguristes eremita each associated with three different species of sea anemones. The anemone Calliactis polypus is associated with eight hermit crab species, followed by C. parasitica associated with seven hermit crabs. C. tricolor and Adamsia palliata associated with six species and Verrillactis paguri with five species (Antoniadou et al. 2013). Calliactis tricolor in addition to its partnership with hermit crabs is also associated with the callapid crab Hepatus epheliticus and the spider crab Stenocionops furcate, each using a different tactile behavior to stimulate the anemone during transfer (Cutress et al. 1970). According to Ross and Sutton (1970) this diversity of anemone stimulation by differently structured hosts renders it susceptible to detachment almost at will and made afterwards to settle almost anywhere rendering these species generalists vs. specialist species associating only with a single partner. Indeed, the rest of the hermit crabs and sea anemones have been reported associated with one or two species. The actual number of anemone-hermit crab pairs may be much higher, possibly revealed in future careful field studies. Cautious is however necessary as a crab might have acquired an anemone by moving into a shell formerly occupied by a crab with the symbiont habit.

The associations between the deep-sea anemones Stylobates spp. and hermit crabs were suggested to exhibit partner specificity and an obligatory relationship regarding the anemones. Each of the five described species of Stylobates was found in association with a single species of hermit crab. Stylobates aeneus occurs with Sympagurus dofleini, S. cancrisocia and S. birtlesi both occur with Sympagurus trispinosus, S. losietteae occurs with sympagurus brevipes and S. calcifer occurs with Pagurodofleinia doederleini. Partner specificity is probably related to the allopatric distribution of Stylobates, with each species inhabiting a separate deep Ocean basin of the Indo-West Pacific. (Crowther et al. 2011; Yoshikawa et al. 2022). The behavioral and ecological mechanisms regulating partner specificity in the genus Stylobstes are still unknown. According to Dunn et al. (1980) the relationship with this actinian is probably facultative for the crustacean but obligatory for the anemone. The carcinoceum (i.e., a chitinous structure secreted by the anemone) must be continuously occupied by a hermit crab for the actinian to survive. In the absence of the symbiont, the sea anemone will probably be under nourished, and the light shell may tumble about, and thereby abrade the anemone.

Little is known about the attraction between hermit crabs and sea anemones. Preference for shells occupied by Calliactis tricolor over non-occupied shells has been demonstrated for Pagurus pollicaris (Conover 1976). Choice tests showed that Dardanus venosus had preference for large Calliactis tricolor anemones over small ones (Brooks 1991b). Hermit crab preference for larger anemones may be due to their greater efficiency in deterring predators and greater nutritional value. In contrast, Calliactis parasitica showed no preference for shells occupied by crabs vs. non-occupied ones (Ross and Sutton 1961b). The sensory modalities involved in partner selection in the hermit crab-anemone partnership are still unknown. Currently we only know that the hermit crab Dardanus venosus without sea anemones is chemically attracted (i.e., tested with a Y-trough olfactometer) to its associated anemone Calliactis tricolor while the crab Pagurus pollicaris without sea anemones failed to do so. This species usually carries fewer sea anemones on its back compared with Dardanus venosus. Dardanus venosus with sea anemones on its shell was as well not attracted to Caliactis tricoclor possibly due to saturation of its chemoreceptors resulting in random movement of the crab (Brooks 1991a). The chemical compound underlying the chemical attraction to Calliactis tricolor is still unknown.

2.7 Intra and interspecific competition for actinians

The sea anemone is a highly valuable resource for the hermit crab since it provides the crab with protection against predators such as octopuses and shell crushing crabs. The acquisition of this resource in a competitive set up is related to social status in the hermit crab Dardanus arrosor. In a laboratory study (Mainardi and Rossi 1969), pairs of D. arrosor deprived of their anemones formed a dominance hierarchy. These crabs were presented with several Calliactis parasitica one after the other. In all trials the dominant and usually the larger crab took all the anemones and placed part of them on its shell. The subordinate crab received none even while some anemones remained lying on the substrate. Only after the dominant was removed, the subordinate acquired sea anemones. A different result was obtained for two hermit crabs from the Gulf of Mexico Pagurus pollicaris and P. impressus associated with Calliactis tricolor. Groups of four individuals of these two species differing in size and deprived of anemones were presented with eight sea anemones attached to plastic plates. In this case, the number of acquired anemones was not affected by crab size (Brooks and Mariscal 1986).

There have been occasional observations on theft of anemones amongst hermit crabs. The transfer of anemones during theft is different from the behavior regularly practiced by the crab. Instead of tapping, the anemones are forcefully detached, often with the release of acontia by the anemone and laceration of the pedal disc. Experimental studies on stealing of anemones (Calliactis tricolor) by hermit crabs (Dardanus pedunculatus) revealed the magnitude of this phenomenon (Giraud 2011). To examine stealing, two crabs of different size, one without anemones and one with a single anemone, were placed in the same tank for 12 h. The larger crab without anemones stole an anemone from the small crab in 82% of the trials. The small crab without an anemone never stole an anemone from the large crab. Cheliped size was the most important factor in establishing dominance in competitive interactions. Even when the cheliped of D. pedunculatus was only bigger by a few mm, dominance was clear and stealing of anemones from the smaller crab occurred. To test whether Calliacts tricolor is a limited resource the coverage of small and large crabs with anemones in Moorea lagoon was compared to that in the laboratory achieved under surplus of anemones (Atkinson 2012). Crabs carry between 0 and 6 anemones. Large crabs carried in the field more anemones with a greater volume and percentage of coverage of their dorsal side than small crabs. Under conditions of excess of Calliactis tricolor in the laboratory, the larger crabs showed similar anemone collecting patterns as in the field while smaller crabs covered a greater percentage of their dorsal side with anemones. These findings indicate that Calliactis parasitica is both a valuable and limited resource for this population of hermit crabs.

Three species of symbiotic hermit crabs, Dardanus arrosor, Paguristes oculatus and Pagurus alatus co-occur in the western Mediterranean with Calliactis parasitica on their shells. Stealing was studied in these species by staging encounters between a crab with no anemones and a crab with a single anemone on its shell. The frequency of intraspecific stealing of anemones was only 6% for D. arrosor and even lower in P. oculatus and P. alatus. In interspecific encounters D. arrosor stole from P. oculatus in every trial (Fig. 4) and in more than 80% of the trials from P. alatus. Neither of these species stole any anemone from D. arrosor. Stealing of Calliactis parasitica from P. oculatus and P. alatus was not only more frequent than from conspecifics, but it happened sooner and was more visible. It seems that intraspecific stealing in D. arrosor was inhibited, deterred from stealing from members of its own species. The submissive behavior of P. oculatus and the flight behavior of P. alatus particularly in a large open space seem to be expressions of their subordinate roles in the acquisition and retention of Calliactis parasitica. Indeed, D. arrosor in the field, carries usually two to three times the number of anemones found on either P. oculatus or P. alatus. D. arrosor is clearly a dominant species in the competition for actinians among pagurids, which may explain its cosmopolitan distribution (e.g., Mediterranean, South Atlantic and Indo-Pacific) (Ross 1979). Interspecific theft of Calliactis tricolor by two hermit crabs, Pagurus pollicaris and P. impressus from the Gulf of Mexico was relatively infrequent but also unequal. P. pollicaris transferred anemones from P. impressus much more frequently (i.e., 86% of the transfers) than vice versa. P. pollicaris seems to be similar to Dardanus arrosor in being dominant in interspecific anemone theft. Similar, to the Mediterranean hermit crab species, intraspecific theft by Pagurus pollicaris and P. impressus was minimal (Brooks and Mariscal 1986).

Little is known about the effect of associated sea anemones on the intra and interspecific aggressive behavior of symbiotic hermit crabs. At least in Pagurus arrosor large crabs carry more Calliactis parasitica than smaller ones and in a competitive set up the dominant crab acquires anemones, and the subordinate obtains none (Mainardi and Rossi 1969). Thus, it seems likely that the possession of several large C. parasitica could by itself signal the high social status of the crab. This hypothesis has not been tested yet. According to Hazlett (1970), the increase of the shell size of Clibanarius vittatus due to black attached plastic tags and increase of its weight resulted in winning of contests among crabs. Likewise, the attached anemone increases the size of the shell and its weight, probably also increasing the chances of winning contests. Finally, in a comparative study (Hazlett 1981b), the intraspecific aggressive behavior of two hermit crabs, Pagurus prideauxi associated with Adamsia palliata and P. bernhardus living in empty shells was contrasted. P. prideauxi which depends less on shells since its anemone increases in size with the crab, fights much less intra specifically compared with P. bernhardus which frequently must replace its shells during growth.

2.8 Costs and benefits for hermit crabs and sea anemones

In the early descriptions of the HCA partnerships mutual benefits were suggested, the anemone protects the crab through camouflage and its nematocysts while the crab provides the anemone with a mobile substrate. Only years later, part of these suggestions was tested, and more complex components of the costs and benefits were revealed.

Most studies on predator deterrence by sea anemones were carried out with cephalopods and particularly with octopuses. In an early study Octopus vulgaris was described retreating after contact with Calliactis parasitica (Boycott 1954). Two groups of Dardanus spp. one covered with Calliactis parasitica and the other without sea anemones were singularly introduced into a tank with an Octopus vulgaris. All crabs lacking anemones were extracted from their shells and consumed, whereas none of the crabs with anemones were damaged. The octopus did not hold crabs with anemones longer than 15s before dropping them. On meeting the anemone’s tentacular crown, an octopus arm was always pulled back sharply as if stung (Ross, 1971). In a subsequent study carried out by McLean (1983) a hermit crab with a single Calliactis tricolor was not effective in preventing predation by Octopus joubini. These contradictory results may be due to the different number of sea anemones involved, the tested species or the experimental design. The effect of the number of attached anemones on hermit crab protection was tested in an experiment carried out with the hermit crab - Pagurus pollicaris, the anemone Calliactis sp. and Octopus joubini (Brooks 1988). Three crabs, one without anemones, one with a single anemone and one with three anemones were introduced at the same time into a tank containing an Octopus joubini. Hermit crabs with a larger number of attached anemones survived longer. Several hermit crabs with three anemones survived more than 40 days in the presence of an octopus.

Several species of hermit crabs were observed in aquaria to have fewer attached anemones, the longer the time they spent in captivity, frequently damaging and eating them. However, the introduction of an octopus to these aquaria or chemical cues from three cephalopod species including a cuttlefish resulted in renewal of the crab behavior of transferring anemones to their shells. However, this activity did not return when only visual cues were provided (Balasch and Mengual 1974; Ross and Boletzky 1979; Brooks and Mariscal 1986). These findings suggest that predation in the natural habitat is one of the factors responsible for the acquisition of anemones by the hermit crabs. In a subsequent study two hermit crab species, Dardanus venosus and Pagurus pollicaris avoided a water current in a Y- trough olfactometer carrying chemical cues from Octopus joubini (Brooks 1991a). This chemically mediated avoidance is consistent with the chemically recognition of octopus substances enhancing anemone transferring behavior. In a field study (Brooks and Mariscal 1986), the effect of predator pressure on hermit crab-anemone interactions was evaluated. Two localities in the Gulf of Mexico, one with a very high density of Octopus joubini and the other with a low density were contrasted regarding shell coverage with anemones. Pagurus pollicaris collected from the high octopus density carried significantly more sea anemones than P. pollicaris collected from the low octopus locality. Moerover, hermit crabs from the high octopus density locality transferred in aquaria more anemones to their shells compared with hermit crabs collected from the locality with low octopuses’ presence (Brooks and Mariscal 1986).

Studies on hermit crab protection by sea anemones were not limited to octopuses. Another tested predator of hermit crabs was the shell crushing crab Calappa flammea. Two Pagurus pollicaris, one with Calliactis tricolor and one without anemones were simultaneously introduced for 45 min to a tank containing a Calappa flammea not fed for 24 h. The crab dropped the hermit crab with anemones about 30 times during each trial but only 0.4 times per trial the crab without anemones. Hermit crabs with anemones survived in 91% of the trials whereas only 17% of the crabs without anemones survived (McLean and Mariscal, 1973). A complimentary field study (Bach and Herrnkind 1980) revealed that hermit crabs carried more anemones in areas with high predation pressure measured by percentage of recently damaged shell (probably caused by shell crushing crabs) compared with areas of low predation pressure. According to Brooks and Mariscal (1986) there may be several additional predatory species that cause hermit crabs to be active towards sea anemones. These species may include a variety of crustaceans such as Menippe mercenaria, Hepatus epheliticus, Callinectes sapidus and Panulirus argus and certain fish species such as cod and dogfish that consume hermit crabs by crushing their shell (Jackson 1913) or biting off the anterior portion of the crab (Brightwell 1952).

The cloak anemone Adamsia palliata (syn. Calliactis palliata Gusmao et al. 2019) associated with Pagurus pridauxi produces from its glandular pedal disc ectoderm a chitinous “living cloak” (i.e., carcinecium”) in the shape of two wings that cover the entire crab abdomen (Fig. 5) and seal potential holes. The cloak expands the crab living space. The necessity to change to bigger shells is possibly spared since the anemone and crab may grow at a similar pace. Moreover, there is no need to transfer anemones from the old to the new shell since the crab remains with the same shell and anemone. Pagurus prideauxi is exceptionally mobile since it is protected by only a light shell compared with other pagurids with large and heavy shells. The symbiosis with Adamsia is less effective against Octopus vulgaris than that of Dardanus with Calliactis. Yet Adamsia serves as a deterrent and when size relations are favorable an octopus can be repulsed (Ross 1984). Several species of hermit crabs associate with species of Stylobates which replace the gastropod shell with a chitinous “carceinoecium” shaped like a trochoid shell secreted by the anemone. So perfect is the likeness, that the “shell” was originally thought to belong to an undescribed gastropod and was classified as such. According to Gusmao et al. (2020) the ability to produce chitin in substantial amounts has evolved at least twice within the Actinaria, with the carceinoecium representing a distinct convergent morphological adaptation. It seems likely that a single anemone is involved in carcinoecium formation, since it seems highly improbable that two or three anemones could coordinate efforts to produce a single shell of such precise geometry (Dunn et al. 1980). The carcinoecium of Stylobates sp. collected at a depth of about 300 m in association with Pagurodofleinia doederleini consisted of 13% chitin, 13% HCL soluble components, 38% NaOH soluble components and 39% unidentified remnants (Dunn and Liberman 1983; Yoshikawa et al. 2019). The formation of the carcinoecium is essential for the hermit crab since species of Stylobates occur in deep water (100–1000 m) in areas scarce with large shells which tend more rapidly to dissolve at these depths. The associations between the deep-sea Stylobates and hermit crabs were suggested to represent obligatory partnerships and to exhibit partner species specificity (Dunn et al. 1980; Ross 1984; Yoshikawa et al. 2019; Gusmao et al. 2020).

Fig. 5
figure 5

Pagurus prideauxi and Adamsia palliata viewed from side (A) and from below (B) to show the tentacles of the anemone in relation to the claws and oral region of the pagurid. In (C) the anemone has seized food (f) from the pagurid. In (D) the pagurid has caused the anemone to retract after the food had been taken. Drawn from frames of 16 mm film. (Ross 1984)

Hermit crabs were rarely observed feeding in captivity on associated sea anemones. However, in an experimental study starved Dardanus pedunculatus were feeding on Calliactis polypus, removing anemones from their own shells and from conspecifics. The crabs gained some weight from consuming anemones which indicated some value of the later as food (Imafuku et al. 2000). These authors also suggested that feeding on free living sea anemones followed by feeding on anemones temporarily carried on the shell as a food reserve paved the evolutionary way for continuously carrying sea anemones by the hermit crab. The main cost for the hermit crab from its association with sea anemones is the extra weight and particularly drag reducing the crab mobility (Conover 1976).

Some HCA partnerships occur only within a limited range of their overlapping distribution but not outside this range. For example, Pagurus variabilis lives with Paracalliactis mediterranea in the Mediterranean but not in the eastern Atlantic where both species co-occur. Similarly, Pagurus bernhardus and Calliactis parasitica associate only within one section of the pagurid range. According to Ross and Zamponi (1982) this situation possibly indicates that the benefits to the pagurids in this case are only marginal if at all.

The main benefit to the anemone from the association is the possession of a hard mobile substratum in an otherwise barren land made of mud and sand. The mobile substratum provides access to areas rich with oxygen, suitable temperatures, and food. The sea anemone Calliactis parasitica attached to gastropod shells often twists its body so that their entire expanded oral disc is parallel to the substrate allowing their tentacles to sweep the ocean floor (Chintiroglou and Koukouras, 1991). On a stable substrate this anemone was able to exploit food supplies from a limited area (i.e., 0.5 m2/day) whereas it covered a much larger area (i.e., 20 m2/day) due to symbiosis (Stachowitsch 1979, 1980). There is currently a debate whether hermit crabs feed associated sea anemones. This was particularly suggested for juvenile Pagurus prideauxi associated with Adamsia palliata (Fox 1965). However, the messy feeding of hermit crabs and the proximity between the anemone tentacles and the crab feeding appendages (Fig. 5) allows some food transfer from crab to anemone (Ross 1984; Chintiroglou and Koukouras, 1991). In addition, the water currents generated by the antennae and mouth parts of the crab which are independent of local water currents supply the anemone with food particles and as well with oxygen rich water (Balss 1924; Jenson 1970).

About thirty years ago the protection of sea anemones from predators by their hermit crabs’ hosts was described in the laboratory, providing the first evidence for mutualism in predator deterrence. Sea stars (Echinaster spinulosus) and fire worms (Hermodice carunculate) consume Calliactis tricolor sea anemones. Pagurus pollicaris provided protection to anemones by moving away from sea stars. Likewise, Dardanus venosus protected the anemones from fire worms by moving away and by attacking the worms, ripping their bodies into pieces and consuming them (Brooks and Gwaltney 1993).

Other benefits to sea anemones associated with hermit crabs include avoidance of smothering by sand which often happens to organisms attached to empty shells. Furthermore, being elevated above the sediment surface protects the anemones from surface dwelling predators (Conover 1975; Stachowitsch 1980).

The cost of being associated with hermit crabs could be predation particularly when the hermit crabs are starved (see previous section on benefits to the hermit crabs), and the removal of food captured by the anemone such as by the insertion of the crab’s claws deep into the anemone gastric cavity (Burton 1969; Lancaster 1988).

2.9 The evolution of the anemone-crab partnership

A shared evolutionary history was suggested for all anemones associated with hermit crabs based on shared morphology and behavior (e.g., pedal disc shape and anemone orientation on the shell). However, recently based on molecular phylogenies based on three mitochondrial markers (i.e., 12 S, 16 S and COIII) and two nuclear markers (18 S and 28 S) it became evident that the bond between anemones and hermit crabs evolved at least four times independently. Two separate origins were found within the family Hormathiidae. Members of the genera Calliactis and Paracalliactis were each monophyletic nested amongst non-symbiotic genera (Gusmao and Daly 2010). Two additional independent origins of the symbiosis with hermit crabs, includes anemones of the genus Carcinactis within the metridioidean clade and members of the genus Stylobates within the actinioideans clade (Gusmao et al. 2020). Concerning the hermit crabs, propensity to initiate activity towards anemones is common to unrelated species and genera, reflecting the frequent and independent establishment of this bond amongst the crabs.

Two alternative hypotheses for the drivers of the crab-anemone partnership, (1) the “crab driven” hypothesis and (2) the “shell response” hypothesis were suggested by Ross (1974a) and are still valid to day (Gusmao et al. 2020). According to the “crab driven” hypothesis, the hermit crab initiates the establishment of the crab-anemone symbiosis by its behavior of placing sea anemones on its shells, gaining camouflage and predator deterrence. According to this hypothesis a clear fitness benefit for the crab is gained. According to the shell response hypothesis, the main driver is the anemone’s attraction to gastropods by a chemical cue from the shell/periostracum, resulting in shell mounting, whereby the anemone receives a substrate, transport and access to food and oxygenated water. Subsequently, settling on shells occupied by hermit crabs, which are even more mobile than the gastropods, occurs. The strongest support of the first hypothesis is the fact that most associations between crabs and anemones are initiated by the crab activity. Furthermore, sea anemones are more frequently found on shells occupied by crabs than on living gastropods even in areas with dense gastropod populations. The second hypothesis is supported by the observation of the exclusive presence of some anemones such as Allantactis parasitica and Hormanthia digitata on living gastropods (Ates, 1995). In deep water-trawls (i.e., 400–1400 m) in the north Atlantic, Allantactis parasitica was collected attached to eight species of living gastropods. However, in these same hauls hermit crabs dwelling in empty shells of symbiotic gastropods did not harbor any of these anemones (Mercier and Hamel 2008). Further support of the “shell response” hypothesis is found in the equal presence of some species such as Calliactis conchiola on both living gastropods and shells occupied by hermit crabs (Hand 1975). Irrespective of the initial behavior stimulating the start of the anemone-crab symbiosis, hermit crabs evolved specialized behavior such as tapping to enhance anemones colonization of their shells and the anemones evolved positive responses to their stimulation by the crabs.