4. Discussion
Our overall aim was to quantify inter-individual variation in the stress response in captive roe deer and to explore the link between this variation and individual condition or quality. Firstly, we showed that the four stress response parameters, rectal temperature, behavioral score, hematocrit level and N/L ratio, were all highly repeatable across seasons, indicating that the stress response of individuals appears to be consistent over time. From a meta-analysis, Bell et al. (2009) reported that behavioral traits are generally repeatable, with an average value of 0.37, whereas the repeatability of our behavioral score was substantially higher (r = 0.61, 95% confidence of interval = [0.365, 0.790]). Published information on the repeatability of other stress parameters is scarce (Biro and Stamps, 2010; Szafranska et al., 2007). For instance, although a stress-induced rise in temperature has been extensively studied on model organisms in controlled environments (Cabanac and Guillemette, 2001; Cabanac, 1975; Carere and van Oers, 2004; Moe and Bakken, 1997), consistency in this parameter was only clearly demonstrated recently in eastern chipmunks (Tamias striatus, Careau et al., 2011). I