What are hermit crabs? Hermit crabs are decapod crustaceans belonging to the infra-order Anomura. There are over 800 species of hermit crab, the vast majority of which are marine — the only known exceptions being one freshwater species and the twelve coenobitid species that are semi-terrestrial, with females returning to the edge of the sea to release their larvae. This family contains Birgus latro, the ‘coconut’ or ‘robber’ crab, which at up to 5 kg is the largest terrestrial invertebrate. Generally, however, anomurans tend to be relatively small. They are distinct from other decapods in that the fifth pair of appendages is very reduced in size. Examples of anomurans other than hermit crabs include the stone and king crabs (Lithodidae), squat lobsters (Chirostylidae, Galatheidae) and porcelain crabs (Porcellanidae). In contrast to hermit crabs, these show a somewhat carcinised (‘crab like’) body form, with a reduced abdomen that is often tucked under the cephalothorax. In some cases this occurs to the extent that there is a superficial resemblance to ‘true’ brachyuran crabs; this is particularly true of the king crabs (Lithodidae), which, despite their large size, are thought to have evolved from pagurid ancestors. So what makes hermit crabs different? The five hermit crab families are the Paguridae (right-handed hermit crabs), Diogenidae (left-handed hermit crabs), Coenobitidae (land hermit crabs), Parapaguridae (deep-sea hermit crabs) and Pylochelidae (symmetrical hermit crabs). In most hermit crabs the abdomen is not protected by a calcified exoskeleton; instead, they occupy hollow objects, usually empty gastropod shells, which provide physical protection against attack from predators and also offer a degree of buffering against changes in the external environment. These objects are normally carried around by the hermit crab and have been likened to ‘portable burrows’. A characteristic feature of most hermit crabs is an asymmetric body plan. There is a distinct curvature to the abdomen and a clear asymmetry in the size of the chelae (claws). These adaptations allow the crab to fit into the spiral of the empty shell, using the abdominal muscles to grip it. The larger claw can be used as an operculum to close off the aperture of the shell when the crab is threatened, and the ‘handedness’ of pagurids and diogenids refers to which of the two chelae is largest. In some species, objects other than shells are used. The pylochelids or ‘symmetrical’ hermit crabs have a fully calcified abdomen and occupy crevices in wooden detritus or in rocks, and in contrast to other hermit crabs, are thought to leave these refugia when foraging. The pagurid hermit crab Discorsopagurus schmitti uses empty polychaete tubes, and in Pagurus prideaux, protection is provided by a commensal ‘cloak’ anemone, which settles on the gastropod shell and grows with the crab to cover the crab's abdomen. Are hermit crabs choosy about their shells? In a word, yes. As they grow, hermit crabs need shells of increasing size, so that adequate protection is maintained. Crabs in shells that are too small show increased risk of mortality, reduced growth rates and, in females, reduced fecundity. A shell that is too large could also be detrimental because of the costs of carrying it. Shell optima can be established through shell selection experiments where crabs over a range of sizes are allowed a free choice between different sized shells. In this way a relationship between crab size and shell size can be established. For a given size of crab it would then be possible to determine the optimal or ‘preferred’ shell weight, and then to calculate the weight of suboptimal shells as percentages of this. Crabs also show clear preferences for shells of particular species. Extensive studies of shell selection behaviour have been conducted on the European hermit crab, P. bernhardus, and on tropical diogenid species such as Clibanarius vittatus (striped hermit crab) and Calcinus tibicen (orange claw hermit crab). In P. bernhardus, the smallest intertidal crabs show strong preferences for Littorina obtusata shells; large intertidal crabs are found in L. littorea shells; and the largest subtidal individuals occupy Buccinum undatum shells. How do hermit crabs gather information about shells? There will usually be a small proportion of unoccupied empty shells available for crabs to move into. Shells are initially located by vision, with shells that contrast in colour against the substrate being easiest to detect. But high contrast shells are also conspicuous to predators and in most cases shell quality cannot be assessed by vision alone. Once a shell has been located, tactile cues are used to gather more accurate and detailed information. Contact with the antennae allows chemical cues to be detected, which may reveal information about shell condition from the amount of calcium detected. The crab also manipulates the shell using its chelae and walking legs to ascertain the shell size, weight and species and to gather more information on the condition of the shell's surface. The shell is then turned so that the aperture is uppermost, allowing the crab to investigate the interior of the shell by inserting its chelae and walking legs inside the shell. A crab may decide to reject the empty shell at any stage during this series of assessment activities. Only if these exhaustive investigations indicate that the empty shell is of superior quality will the crab decide to switch into this new shell. Before finally rejecting the new shell, the crab tests its fit by rapidly withdrawing into it, and raising it off the substrate, presumably to gauge its internal volume and weight. Even after the switch there is often a quick reinvestigation of the old shell in order to ‘double check’ that the new one is really better. Shell choice is thus a complex and important process, with shell investigation allowing information to be gathered by using a combination of visual, chemical and tactile cues. How do hermit crabs fight? Most suitable shells are already occupied at any given time, so hermit crabs in suboptimal shells often have to resort to initiating a ‘shell fight’ in an attempt to take a better shell from another crab (Figure 1). During these encounters, not only do hermit crabs have to size up the quality of the opponent's shell but they also have to assess the fighting ability of their opponent. The encounter involves two distinct roles, ‘attacker’ and ‘defender’. An attacker in a suboptimal shell will initiate a fight against a usually smaller defender in possession of a better shell. Because shells are critically important for both sexes, males and females engage in agonistic encounters over their ownership. The encounter begins with a period of displays using the chelae performed by both crabs, and then the attacker usually lunges forward and grabs the defender's shell. The defender withdraws tightly into its shell and remains in this position until the end of the encounter. After an initial period of investigation of the occupied shell the attacker performs a series of bouts of ‘shell rapping’ by rapidly and repeatedly striking the surface of its own shell against the shell occupied by the defender. In between bouts of rapping the attacker may attempt to pull the defender out of its shell through the aperture. Fights usually last for between eight and ten bouts, though they can sometimes continue for much longer. The encounter is terminated either by a decision from the defender to relinquish its shell, allowing the attacker to evict it by pulling it out through the aperture; or the attacker may give up, and release the defender, without effecting an eviction. If the attacker is successful, it will perform the usual investigatory activities on the vacated shell and decide whether to occupy it permanently. During this process the evicted defender is left without a shell, but once the attacker has made its decision the defender is free to move into the shell that is eventually discarded, such that the encounter often ends with an ‘exchange’ of shells. Negotiation or aggression? In P. bernhardus, successful attackers rap more vigorously than those that give up, performing more raps per bout, hitting harder and leaving shorter pauses between bouts. The detail of shell rapping has not been examined in other species but in all cases there is the possibility that the defender, as well as the attacker, can benefit from the encounter. For example, if a defender entered the encounter with a shell that was much too large and exchanged it for a smaller shell that the attacker discarded, then the exchange could lead to an improvement in the defender's shell quality as well as that of the attacker. Indeed, it has been shown in C. vittatus that exchanges are more likely to occur if the defender also benefits. It has been suggested that these hermit crabs ‘negotiate’ rather than ‘fight’ over their shells, and that the function of shell rapping is in fact to advertise the quality of the attacker's shell, allowing the defender to base its decision on this factor. The defender could assess this by monitoring the pitch of the raps, which would vary with shell size. But this doesn't explain the difference in vigour between evictions and non-evictions in P. bernhardus. If the rapping advertises shell quality why would attackers need to rap vigorously as shell quality will not vary with the vigour of rapping? In P. bernhardus at least, shell exchanges seem to be primarily agonistic encounters where any benefit to the defender is incidental to the gains made by a successful attacker. Shell rapping seems to be related to the attacker's stamina and advertises its fighting ability or ‘resource holding potential’. What can hermit crabs tell us about decision-making and contests? Theoretical models of fighting are based on the idea that animals gather information about the resource, the opponent and the costs of fighting and use this information to make strategic decisions. Hermit crabs are ideal for investigating contests because resource value can be quantified and manipulated by supplying the crabs with shells of known quality. Differences in resource holding potential can be manipulated by varying the difference in crab weight between attackers and defenders. Another advantage is that the main agonistic shell rapping behaviour can be easily quantified by analysing measures of its vigour. Hermit crabs have been used to investigate the physiological consequences of fighting, such as changes in energetic status and hormones, and to relate these mechanisms to the decisions predicted by theory. These studies on P. bernhardus show that during fights each opponent may make decisions in a different way, basing them on different sources of information and on different costs of fighting. This is relevant to many other types of contest as fighting often involves two distinct roles, where one individual holds a resource that an opponent attempts to take over, for example, when there is an ‘owner’ and an ‘intruder’.