README file * Citation to the publication associated with the datasets and code? Title: Early filial cannibalism in fish revisited: endocrinological constraint, costs of parental care, and mating possibility. Journal: The American Naturalist * Author names, contact details Authors: Takeshi Takegaki, Yosuke Nakatake, Yukio Matsumoto, Koushiro Suga and Noriko Amiya Corresponding author: Takeshi Takegaki Contact details: Graduate School of Fisheries and Environmental Sciences, Nagasaki University, Nagasaki, 852-8521, Japan E-mail address: takegaki@nagasaki-u.ac.jp * A brief summary of what the study is about Offspring desertion by parents generally occurs at an early stage of parental care, which is considered to minimize the time and energy costs of parental care prior to desertion and ensure an opportunity and time for future reproduction. In addition to these conventional reasons, we investigated the effects of endocrinological constraint on early total filial cannibalism in fish in which care-giving males eat all eggs and restart reproduction when tending a small number of eggs. The occurrence of total filial cannibalism in male Rhabdoblennius nitidus, a paternal brooding blennid fish with androgen-dependent brood cycling, was strongly associated with decreased plasma 11-ketotestosterone levels, suggesting that males in which it was difficult to exhibit courtship behavior began to cannibalize eggs. Moreover, the additional mating probability at the beginning of parental care may delay the timing of total filial cannibalism; at the same time, the costs of parental care may shift the timing earlier than expected based on androgen decline. These results suggest the potential importance of considering endocrinological constraints with respect to the adaptivity of behavioral traits. * Identify who is responsible for collecting data and writing code Takeshi Takegaki, corresponding author * Software: R version 3.5.2 R Development Core Team. 2018. R: a language and environment for statistical computing. R Foundation for Statistical Computing. Retrieved from http://www.r-project.org * Data files 1) Cannibalism_data.csv group: correspond to the "status" 1: courtship phase 2: cannibal males 3: non-cannibal males 4: parental care phase status: male status courtship: courtship phase parental: parental care phase cannibal: cannibal males non-cannibal: non-cannibal males timing: timing of male sampling 1: on the day of the first egg acquisition (cannibal and non-cannibal males). before egg acquisition (males at courtship phase) 2: on the day following the first egg acquisition 3: 4 days after the first egg acquisition kt: 11-KT (keto-testosterone) level (ng/ml) eggs: the number of eggs tended before removal experiment size: male standard length (SL, mm) 2) EggDNA_data.csv status: male status cannibal: cannibal males non-cannibal: non-cannibal males DNA: amount of DNA /egg (µg/mL) 3) Remating_data.csv timing: timing of egg removal experiment (on the spawning day, 4 to 5 days after spawning) remating: male remating success at one day after egg removal (yes, no) * R source and code 1) Tukey HSD test data<-read.csv("Cannibalism_data.csv") Comparison of male body size among male status (courtship phase, parental care phase, cannibal males, non-cannibal males) Code: TukeyHSD(aov(data$size~data$status)) 2) Steel Dwass test data<-read.csv("Cannibalism_data.csv") Source: source("http://aoki2.si.gunma-u.ac.jp/R/src/Steel-Dwass.R", encoding="euc-jp") Code: Steel.Dwass(data$kt, data$group) 3) t test Welch Two Sample t-test data<-read.csv("Cannibalism_data.csv") Comparison of the number of eggs among male status (cannibal and non-cannibal) Code: t.test(eggs ~ status, data = data) Two Sample t-test data<-read.csv("EggDNA_data.csv") Comparison of the DNA content per egg among male status (cannibal and non-cannibal) Code: t.test(DNA ~ status, data = data, var=T) 4) χ2 test data<-read.csv("Remating_data.csv") Comparison of remating success rate between males whose eggs were removed on the spawning day and males whose eggs were removed 4 to 5 days after spawning. Code: chisq.test(data$timing, data$remating)