In Gavin Stark and Shai Meiri (2018), the authors would like to notify the readers that in the original published version of the article, It has several supplemental tables and figures are missed and published.

We have found that parts of the dataset we recently used in preparing and writing this article were incorrect. This has prompted our immediate reassessment of the entire database. Consequently, we went over the published database (spanning 527 amphibian species) and found multiple errors in our original data with regards to the estimation of sample sizes (but very rarely in the longevity data, that are the main focus of the paper). These errors prevailed in the dataset as a whole (i.e., the sample sizes we used were often not found in the original sources or other numbers used instead of the correct ones). Most issues occurred in data from capture-recapture studies and museum records, where several values of sample size are available and where missing data on specimens’ date of birth or death, making it harder to assess their age. We now standardized the way we tallied sample size in general, and for capture-recapture studies and for museum specimens in particular. We now consider the number of specimens in capture-recapture studies as the number of individuals recaptured at the first recapture attempt as the relevant sample size. For museum specimens we consider only specimens that have a record of one year at least. The corrected dataset is available in the revised Appendix S1 in Supporting Information. Due to this error, we re-run all our analyses with the updated dataset (n = 527).

The PGLS model for the relationship between mean body size and longevity among and within amphibian orders (Table 1) remained qualitatively similar. Amphibian longevity increased to the 0.14 power of mass (log10 transformed) as opposed to 0.12 in the original publication, and body size explains 9%–13% of the variance in amphibian longevity as opposed to 7-14% in the original analyses (Appendix S1.2). Similarly, the minimal adequate model (mam) for the analysis of all amphibian species (Table 2; Figure S2) remained qualitatively similar (Appendix S1.2), except for chemical protection having a marginal effect (p = .01) as opposed to the original analysis where the effect was significant (p = .0002). For anuran species (Table 3), we only found differences in the effect of chemical protection: significant in the original version (p = .0005) and not significant in the corrected one (p = .02). Additionally, sample size did have a significant (positive) effect (p < .0001) as opposed to the original analysis where it did not (p = .10). Moreover, there were no interactions between any variables as found previously between chemical protection and origin of data for anurans (p = .002). The Urodela model (Table 4) showed similar results, however, with marginal effect of activity time (p = .01) and sample size (p = .008) on longevity as opposed to the original analysis where they were significant (at p < 0.005). For the full details of the new models see the new Appendix S2.

The revised sensitivity analysis revealed only minor differences in the overall results, with a non-significant effect of chemical protection (p = .09) on longevity as opposed to the original significant results (p = .005). For the Anura, the model for well-sampled species was the same as the original except for the now insignificant (p = .11) interaction between activity period and annual temperature. For Urodele the new results showed a significant effect of chemical protection (p = .0003) and a marginal one of annual temperature (p = .02). In contrast, activity period that was not significant (p = .50) as opposed to the original results where it was (p = .0002). The rest of the variables were now found to be not significant. See Appendix S3 for the full details on all the models, full or well sampled ones. Finally, the comparison between the longevity data of animals held in captivity and those in the wild (n = 140 originally, and now n = 130), found no significant difference (p = .12) (Appendix S4) as opposed to the original analysis where it had marginal effect (p = .03).

Correcting our errors resulted in no fundamental change in most of our results (i.e., results are almost invariably qualitatively the same), and further support our main conclusions regarding the main drivers that shape the variation in longevity among amphibians.

The author would like to apologize for this error.

在 Gavin Stark 和 Shai Meiri ( 2018 ) 中，作者想通知读者，在文章的原始出版版本中，它有几个补充表格和数字被遗漏和出版。

我们发现我们最近在准备和撰写本文时使用的部分数据集是不正确的。这促使我们立即重新评估整个数据库。因此，我们查看了已发布的数据库（跨越 527 种两栖动物），发现原始数据中关于样本量估计的多个错误（但在寿命数据中很少，这是本文的主要重点）。这些错误在整个数据集中普遍存在（即，我们使用的样本大小通常在原始来源或其他使用的数字中找不到，而不是正确的数字）。大多数问题发生在捕获-再捕获研究和博物馆记录的数据中，其中有几个样本量值可用，并且缺少有关标本出生或死亡日期的数据，这使得评估他们的年龄变得更加困难。我们现在标准化了我们统计样本量的一般方式，特别是用于捕获-再捕获研究和博物馆标本。我们现在将捕获-再捕获研究中的样本数量视为在第一次再捕获尝试中重新捕获的个体数量作为相关样本量。对于博物馆标本，我们只考虑至少有一年记录的标本。修正后的数据集可在支持信息中修订的附录 S1 中找到。由于此错误，我们使用更新后的数据集重新运行所有分析（对于博物馆标本，我们只考虑至少有一年记录的标本。修正后的数据集可在支持信息中修订的附录 S1 中找到。由于此错误，我们使用更新后的数据集重新运行所有分析（对于博物馆标本，我们只考虑至少有一年记录的标本。修正后的数据集可在支持信息中修订的附录 S1 中找到。由于此错误，我们使用更新后的数据集重新运行所有分析（n  = 527）。

两栖动物目之间和内部的平均体型与寿命之间关系的 PGLS 模型（表 1）在性质上保持相似。两栖动物寿命增加到质量的 0.14 次方（log10 转换），而不是原始出版物中的 0.12，体型解释了两栖动物寿命差异的 9%–13%，而原始分析中为 7-14%（附录S1.2)。同样，用于分析所有两栖动物物种（表 2；图 S2）的最小适当模型（mam）在性质上保持相似（附录 S1.2），除了具有边际效应（p  = .01）的化学保护与效果显着的原始分析 ( p = .0002）。对于 anuran 物种（表 3），我们只发现化学保护效果的差异：原始版本显着 ( p  = .0005) 和修正版本不显着 ( p  = .02)。此外，样本量确实具有显着（正）效应（p  < .0001），而原始分析则没有（p  = .10）。此外，之前发现的化学保护和无尾目数据来源之间的任何变量之间没有相互作用（p  = .002）。Urodela 模型（表 4）显示了类似的结果，但是，活动时间的边际效应 ( p  = .01) 和样本大小 ( p = .008) 寿命，而不是原始分析中它们是显着的（p  < 0.005）。有关新型号的完整详细信息，请参阅新的附录 S2。

修订后的敏感性分析仅显示总体结果中的微小差异，与 原始显着结果 ( p  = .005)相比，化学保护 ( p = .09) 对寿命的影响不显着。对于 Anura，采样良好的物种的模型与原始模型相同，除了 活动期和年温度之间现在不显着 ( p = .11) 的相互作用。对于 Urodele 而言，新结果显示化学保护的显着影响 ( p  = .0003) 和年温度的边缘效应( p  = .02)。相反，活动期间，这不是显著（p  = 0.50），而不是原来的结果，其中它是（p = .0002）。现在发现其余变量并不显着。有关所有模型（完整或采样良好的模型）的完整详细信息，请参阅附录 S3。最后，圈养动物的寿命数据与野生动物的寿命数据（ 最初n = 140，现在n  = 130）之间的比较发现与原始数据没有显着差异（p  = .12）（附录 S4）分析它具有边际效应 ( p  = .03)。

纠正我们的错误导致我们的大多数结果没有根本变化（即，结果几乎总是在性质上相同），并进一步支持我们关于影响两栖动物寿命变化的主要驱动因素的主要结论。

作者想为这个错误道歉。