Mild stress during development affects the phenotype of great tit Parus major nestlings: a challenge experiment WILLEM TALLOEN 1 , LUC LENS 2 , STEFAN VAN DONGEN 1 , FRANK ADRIAENSEN 1 and ERIK MATTHYSEN 1 * 1 Evolutionary Ecology Group, Department of Biology, Groenenborgerlaan 171, University of Antwerp, B-2020 Antwerpen, Belgium 2 Terrestrial Ecology Unit, Department of Biology, Ledeganckstraat 35, Ghent University, B-9000 Ghent, Belgium Received 13 July 2009; accepted for publication 17 December 2009 Conditions experienced during early development may affect both adult phenotype and performance later during life. Phenotypic traits may hence be used to indicate past growing conditions and predict future survival probabilities. Relationships between phenotypic markers and future survival are, however, highly heterogeneous, possibly because poor- and high-quality individuals cannot be morphologically discriminated when developing under good environmental conditions. Sub-optimal breeding conditions, in contrast, may unmask poor-quality individuals in a measurable way at the morphological level. We thus predict stronger associations between phenotype and performance under stress. In this field study, we test this hypothesis, experimentally challenging the homeostasis of great tit (Parus major) nestlings by short-term deprivation of parental care, which had no immediate effect on nestling fitness. The experiment was replicated during two subsequent breeding seasons with contrasting ambient weather conditions. Experimental (short-term) stress affected tarsus growth but not residual mass at fledging, whereas ambient (continuous) stress affected residual mass but not tarsus growth. Short-term stress effects on tarsus length and tarsus fluctuating asymmetry were only apparent when ambient conditions were unfavourable. Residual mass and hatching date, but none of the other phenotypic traits, predicted local survival, whereby the strength of the relationship did not vary between both years. Because effects of stress on develop- mental homeostasis are likely to be trait-specific and condition-dependent, studies on the use of phenotypic markers for individual fitness should integrate multiple traits comprising different levels of developmental complexity. © 2010 The Linnean Society of London, Biological Journal of the Linnean Society, 2010, 100, 103–110. ADDITIONAL KEYWORDS: developmental instability – fitness – fluctuating asymmetry. INTRODUCTION Morphological traits, comprising the phenotypic end- points in organisms with determinate growth, reflect environmental conditions experienced during growth and may predict long-term fitness consequences of suboptimal developmental conditions (Lindström, 1999; Naguib & Gil, 2005). Morphological changes as a result of adverse conditions during development are considered to be primarily caused by a disruption of developmental homeostasis (i.e. an organism’s prop- erty to adjust itself to variable conditions such that phenotypic variation is decreased) (Lerner, 1954). Canalization decreases phenotypic variation between environments (Waddington, 1959), and compensatory growth and developmental stability decrease varia- tion within an environment in trait size and develop- mental noise, respectively (Nijhout & Davidowitz, 2003; Sibly & Calow, 1986; Van Valen, 1962). Rela- tionships between phenotypic indices of homeostasis and measures of environmental stress and individual quality are typically heterogeneous (Bjorksten, Fowler & Pomiankowski, 2000), possibly because developmental rate and stability may remain unchal- lenged under (near-)optimal conditions (Lens et al., 2002; Van Dongen, 2006). Under this hypothesis, *Corresponding author. E-mail: erik.matthysen@ua.ac.be Biological Journal of the Linnean Society, 2010, 100, 103–110. With 2 figures © 2010 The Linnean Society of London, Biological Journal of the Linnean Society, 2010, 100, 103–110 103