The American Naturalist | 2021
Correction
Abstract
Am. Nat. 2021. Vol. 198, pp. 310 0003-0147/2021/19802-60626$1 DOI: 10.1086/714866 The Editorial Board of The American Naturalist and an independent group of authors recently evaluated the raw data underlying our 2012 article “Behavioral Types of Predator and Prey Jointly Determine Prey Survival: Potential Implications for the Maintenance of Within-Species Behavioral Variation” (American Naturalist 179:217–227). On the basis of their findings and our own reading of the article, several corrections to the article and annotations to the posted data have been implemented. We have identified several issues with the data deposited in the Dryad Digital Repository. First, the posted data do not include the 120 Chlorostoma funebralis activity measures forming the basis of the conclusion that among-individual differences in activity level were not significantly repeatable (“Results,” paragraph 1, line 11). We have not been able to locate this subset of the data. These data likely were not included in the uploaded files because we excluded snail activity levels in all subsequent analyses on the basis that they were not repeatable measures of an individual snail’s behavior. Second, snail shell diameters and fear responses on the posted data sheet titled “raw data” are the normalized values used in our analyses rather than the raw measurements. Thus, although these are the data that formed the basis of our analyses, they were not the original values collected by observers. We are unable to locate the original unnormalized values. Third, the spreadsheet in the Dryad repository labeled “shell size by fear” stems from an incomplete preliminary analysis run near the end of the study and should not be taken as the complete data set. This data set was not used in any of the analyses in the article. Fourth, review by the Editorial Board of the Dryad data on sea star activity level revealed a possible data entry error. One of the sea star movement values taken on day 1 in the raw data is reported as 115 cm. This is three times greater than any other movement value reported. We feel that the value on Dryad may represent a typographical error, given that 115 cm movement, while plausible, is an outlier within this data set. However, original data sheets to verify this point are unavailable. If, in fact, 11.5 cm is the correct value, this would result in a significant relationship between sea star activity level measured on day 1 and the selection gradients they impose on antipredator behavior in snail prey (F1, 16 p 8:07, P p :012). Using the value of 115 cm (which is what was used in the analysis in the article) results in no relationship between these two variables (F1, 16 p 0:18, P p :68). The significant association between sea star behavior and the selection gradients they impose on prey traits was recovered from behavioral data taken from sea stars on day 14, which did not contain any extreme values. The main conclusions of the article were based on day 14 sea star activity level (fig. 3A) and so are unaffected by this anomaly.We have posted a corrected ReadMe file to Dryad to clarify these points.