Correction

Previous articleNext article FreeCorrectionJonathan N. Pruitt, John J. Stachowicz, and Andrew SihJonathan N. Pruitt1. McMaster University, Hamilton, Ontario L8S 4L6, Canada Search for more articles by this author , John J. Stachowicz2. University of California, Davis, California 95616 Search for more articles by this author , and Andrew Sih2. University of California, Davis, California 95616 Search for more articles by this author Original articleBehavioral Types of Predator and Prey Jointly Determine Prey Survival: Potential Implications for the Maintenance of Within-Species Behavioral Variation.Full TextPDF Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinked InRedditEmailQR Code SectionsMoreThe 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=8.07, 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=0.18, 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.Sample sizes and degrees of freedom required some corrections. The article mistakenly reported that 2,638 C. funebralis snails were used. However, the actual number of snails used was 2,366. The error stems from a mistaken transposition in snail counts (2,368, not 2,638) and a miscount (2,366, not 2,368). The denominator degrees of freedom on line 6 of the results section and at the bottom of the first paragraph of that section should be 2,366 minus model terms. The denominator degrees of freedom on line 9 should be 99, not 101. The degrees of freedom on sea star behavioral traits and their associated relationships with the selection gradients they impose on prey appear to be in order.Review by the Editorial Board found a relatively small number of duplicated sequences of values in C. funebralis shell diameter data (the “N Max” column on the “raw data” tab of the Dryad file), which amounted to 2%–3% of these morphological data. However, these sequences are an expected outcome of experimental design and data management in this study. This is because snails were assigned nonrandomly to mesocosms on the basis of their shell diameter size. Thus, a permutation-based test to generate a null distribution of the length of strings of duplicates is not appropriate because that null model assumes completely random assignment of snails to mesocosms by size. Moreover, the data were then sorted by size within a given cohort, which further inflates the odds of observing repeated strings of numbers. However, these procedural details were not included in the original methods and are thus noted here now. Splitting individuals of like traits between cohorts was done to prevent the odds of assigning groups of snails of similar or identical traits to the same test cohort, which would eliminate the trait variation necessary to test for selection on prey traits within each cohort. Notably, the column labeled “N Anti-pred response,” which forms the basis for the study’s major conclusions, does not contain a statistical excess of repeat sequences. The data for the antipredator response were checked by the Editorial Board, and no excess of repeated values was found beyond that which would be expected by chance.The Editorial Board also identified an overrepresentation of values ending in .5 and of whole values (those ending in .0) in the snail shell diameter measures. In contrast, values ending in .4 and .9 are underrepresented. Although we cannot be sure, the excess of values ending in .5 and .0 could stem from observers’ tendency to round values to whole and half digits. Unfortunately, we do not know which observers recorded which values to evaluate whether this tendency was attributable to one observer or was an overall pattern. Regardless, the terminal-digit bias suggests a level of observer involvement or subjectivity in the values assigned, but this level of bias is unlikely to affect qualitative conclusions.Finally, an independent anonymous author conducted a randomization analysis that found that some pairs or mesocosms contained greater overlap in snail shell diameter values than is anticipated by chance. No such anomalous overlap in values was present in the article’s focal trait, which was snail antipredator behavior. Although greater than expected overlap in shell diameter is present only in a small number of mesocosm pairs, we detail here some of the mechanisms likely to be responsible for this pattern. First, as mentioned above, we deliberately split snails of identical size between mesocosms, which means that snails were not completely randomly assigned, and greater between-mesocosm overlap in snail size is expected on the basis of this procedure. Second, snails used in this study came from multiple collection trips. Because of variation in tide levels during collection and a tidal gradient in snail size (Doering and Phillips 1983), not all snail sizes were equally available before creating each mesocosm. This too is predicted to increase between-mesocosm resemblance in snail body sizes in mesocosms established immediately following the same collection trip. Finally, mesocosms established by the same observer are likely to be subjected to similar terminal-digit biases when estimating snail size, and this too can increase the level of overlap in shell size diameter values in mesocosms established by the same observer. Taken together, there are several procedural explanations for the observation that a small number of mesocosms share more shell diameter values in common than expected on the basis of a null model.We regret the errors and omissions in our original article and are glad for the opportunity to correct them. In light of our analysis and that of the Editorial Board, we conclude that these corrections do not alter the major conclusions of the article.Literature CitedDoering, P. H., and D. W. Phillips. 1983. Maintenance of the shore-level size gradient in the marine snail Tegula funebralis (A. Adams): importance of behavioral responses to light and sea star predators. Journal of Experimental Marine Biology and Ecology 67:159–172.First citation in articleCrossrefGoogle Scholar Previous articleNext article DetailsFiguresReferencesCited by The American Naturalist Volume 198, Number 2August 2021 Published for The American Society of Naturalists Article DOIhttps://doi.org/10.1086/714866 Views: 1301 HistorySubmitted January 21, 2021Published online June 01, 2021 © 2021 by The University of Chicago. All rights reserved. Crossref reports no articles citing this article.Related articlesBehavioral Types of Predator and Prey Jointly Determine Prey Survival: Potential Implications for the Maintenance of Within-Species Behavioral Variation.17 Jul 2015The American Naturalist

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.