Rearing experiments (REESE, 1962) showed that the behavior of hermit crabs toward the shells which they inhabit is fully and completely expressed the first time it is released, and therefore that the form of the behavior is not dependent upon previous experience with an adequate releasing object. The present investigation is an analysis of the sense organs and the properties of the shells, the sign stimuli, which release the fixed motor patterns comprising the shell exploration and selection behavior. Hermit crabs orient visually to objects, such as shells and pebbles, which contrast with the background and are of an appropriate size relative to the crabs.There is no behavioral evidence for visual discrimination of shape. After the initial orientation, vision plays no further role in the behavior. Likewise, olfaction is not involved in the behavior. Once the crabs have made contact with shells, all subsequent sensory information regarding the shells is supplied via mechanoreceptors, both tactile and proprioceptors of the appendages and probably of the abdomen and cephalothorax. Crabs with only a single intact appendage (cheliped or walking leg) are able to enter and ultimately select the preferred species of shell. Observations of normal behavior of crabs lacking their abdomens suggested that proprioceptors and perhaps tactile receptors of the abdomen feed-back only negative information to the higher centers controlling the behavior. Thus "no news is good news" and with ablation of the abdomen no feed-back can occur and the behavior proceeds normally. The ablation of appendages revealed compensatory adjustments by the intact appendages which take over the functional roles of the ablated appendages. By removing both the chelipeds and the first pair of walking legs, it was possible to bring two functional roles, holding the shell and exploring the aperture, into conflict on the second pair of walking legs. Although this pair of legs was used to explore the aperture, their primary function of holding the shell predominated. The properties of the shells, which are encountered sequentially by the crabs, release distinct and characteristic motor patterns which bring the crabs into the next stimulus situation, where in turn they encounter the sign stimuli which release the next fixed motor pattern. In this way the behavior proceeds step by step until the crab has entered and righted the shell. In the order in which they are encountered by the crabs, the properties of the shells which release the fixed motor patterns are: contrast with the background, movability, surface texture, external shape, aperture presence, aperture free of obstructions, internal size, shell up-side-down, internal configuration and weight. The motor patterns are reviewed in the beginning of the discussion. Although normally the fixed motor patterns follow one another in an ordered sequence, they can be released individually in isolation, that is out of context relative to the other patterns. Therefore, they are not interdependent as are the events in a reaction chain. Upon completion of the final act of the behavior, entering and righting the shell, the crabs continue to show appetitive searching behavior, characterized by a high level of motor activity and responsiveness to shells, depending upon the species and/or weight of the shells entered by the crabs. For example, if the crabs enter Tegula shells the level of the appetitive behavior is low, but if Olivella shells are entered, it remains high. The crabs are able to discriminate between the species of shells used in the investigation on the basis of either the external shape alone or the aperture alone, and although the shape of any shell, of appropriate relative size, releases the motor patterns associated with external exploratory behavior, the shape of the preferred species of shell is more stimulating to the crabs. For example, the crabs spend more time exploring the shape of Tegula shells than Olivella shells. Therefore, the sign stimuli must be thought of as having graded releasing value. An interesting recycling of the behavior was observed in conjunction with shells with their apertures partially obstructed. Cessation of the recycling appeared to be due to adaptation occurring centrally, perhaps of the releasing mechanism itself. Finally, a functional scheme is presented which explains a selection process which is not dependent upon previous experience with the stimulus objects. Two properties of the shell, weight and internal configuration, appear to contribute directly to the determination of the total stimulus value of the shell. The other properties of the shell, besides releasing specific fixed motor patterns, may also contribute to the evaluation of the shell as a stimulus object. An experiment is described to test this hypothesis. The total stimulus value of the shell determines the further responsiveness of the crabs toward shells. Thus crabs select shells by a lack or reduction of responsiveness to shells rather than by any direct comparison between shells.