There were several errors in J. Exp. Biol. (2020) 223, jeb217075 (doi:10.1242/jeb.217075).
There was a discrepancy between the terms used to describe the statistical model outputs in the paper and the column names for the raw data published in the supplementary material. Specifically, the terms ‘class’ and ‘time point’ in the paper referred to ‘subject’ and ‘hours’ in the supplementary material, respectively. The paper and supplementary material have now been updated to correct for these discrepancies.
As noted in Materials and Methods, ‘Tank’ and ‘male ID’ were included as random effects in the behavioural analyses. However, the statistical output values for random effects were not included in the original publication. These are now shown as part of Table 1.
There were also three typographical errors in the statistical outputs in Table 2. For the maximum dominance analysis, the partial eta squared (ηp2) value (0.059) for body mass as a main effect was mistakenly entered as the P-value for body mass as a main effect; this should have been 0.252. The total SS value (5307) was mistakenly entered as the residual SS value; this should have been 3925. For the liver LDH analysis, the residual SS value should read 4.510, rather than 4.541. These typographical errors have now been corrected.
The assumption test for ‘affiliation’ failed normality of residuals. These data were reanalysed using a log(source+1) transformation but note that ‘Male ID’ was removed as a random component because of a singularity in the data. The Materials and Methods section and Table 1 have been updated to reflect this.
Both the online full text and PDF versions of the paper and the supplementary material have been corrected. The authors apologise to the readers for these errors, which do not impact the results or conclusions of the paper.
