Data was gathered experimentally at the Baruch Institute for Marine & Coastal Sciences. Consumption of mussels was recorded daily.

The 3-mo experiment (June 2012 to August 2012) assessed how claw loss influenced the prey consumption rate by stone crabs and whether 1-clawed stone crabs became more efficient at foraging over time. There was difficulty obtaining the large number of legal-size crabs needed to examine the effects of claw loss while concomitantly accounting for other factors that influence mussel consumption, such as crab size, mussel size, and so forth. This experiment was therefore conducted with a modest number of crabs (13 total: 8 females, 5 males; mean CW ± SD, 89.1 ± 6.1 mm). It was expected that the variation in consumption would be greatest for 1-clawed crabs. Therefore crabs were allocated to experimental treatments unevenly, which reflects the desire to maximize the replication of 1-clawed individuals while also attempting to account for differences in size among the available study animals. The considerably lower mussel consumption of 1-clawed crabs relative to 2-clawed crabs (see Results) made the qualitative results unambiguous; however, the low replication means specific quantitative differences in consumption between 1- and 2-clawed crabs in this experiment should be interpreted with caution.

The crabs were housed in individual lobster wire cages (approximate dimensions 152 3 152 3 300 mm), to prevent escape, each within a separate 5-gal bucket that had its own flow-through seawater source, which allowed temperature and salinity to fluctuate with ambient conditions. Stone crabs were divided into 4 different 4.5-mm size classes and were each fed ribbed mussels (Geukensia demissa) ad libitum until declawing. At the start of the experiment (June 4, 2012), the larger, crusher claw was removed from 9 of the 13 stone crabs. Four stone crabs, one from each size class, were not declawed and served as control stone crabs in the experiment.

All stone crabs were provided with a diet of 5 live ribbed mussels daily, and fragments of consumed ribbed mussels were removed after 24 h. Any ribbed mussels not consumed within 1 wk were replaced. The ribbed mussels used in the experiment ranged from 55-75mmin length and were scaled with respect to the 4 crab size classes. The ribbed mussels provided to each crab were consistent within 1 mm for the duration of the experiment. The length of all ribbed mussels provided was measured prior to placement in the aquaria and each was marked with a small dot of nail polish to allow distinction between individual ribbed mussels.

For each day of the experiment, the total number of mussels cracked was recorded to determine whether crabs would improve in their ability to crack mussels over time. This ability to improve was examined using a generalized mixed-effects model (Poisson distribution), with number of mussels cracked daily as the response variable, days in the experiment as a continuous predictor variable, number of claws and sex as categorical predictor variables, and crab identification number as the random variable to control for repeated measures. The interaction between days in the experiment and the number of claws in the initial analysis was included to determine whether 1- and 2-clawed crabs responded differently to the amount of time in the experiment.The interaction was not significant (Z¼0.137,P¼ 0.89) and was therefore removed from the analysis. To examine the overall difference in consumption between 1- and 2-clawed crabs, the total number of mussels consumed throughout the length of the experiment was calculated for each crab. A paired t-test (paired by size class) was used to compare the average number of ribbed mussels consumed during the experiment for 1- and 2-clawed crabs (data from multiple 1-clawed crabs within a single size class were averaged prior to this analysis).