We log transformed (log10[x + 1]) the innovation, social
learning and neophobia variables collected during the captive
study because of non-normal distributions, and thus were
able to use parametric statistical procedures. All statistical
tests used were two-tailed tests.