Modelling seasonal increments in size to determine the onset of annual growth in fishes S. GAGNE ´ AND M. A. RODRI ´ GUEZ* D epartement de chimie-biologie, Universit  e du Qu ebec a`Trois-Rivie`res 3351 boul. des Forges, Trois-Rivie`res, Qu ebec, G9A 5H7 Canada (Received 5 September 2007, Accepted 4 April 2008) A new method based on modelling of seasonal growth increments (G SI ) in total length was found useful for assessing the date of onset of annual growth for 16 fish species in a temperate fluvial lake. Model comparisons indicated that polynomial (linear or quadratic) functions provided better fits to seasonal growth and were more likely to avoid convergence problems than alternative non-linear models. There was little evidence for differences in the date of onset of growth between years, nor among age classes within individual species. The onset of growth also was to some extent synchronized among species and was concentrated within a narrow temporal window of c. 2 weeks, between 18 May and 2 June, which corresponded to mean water temperatures between 161 and 173° C. There was no apparent relationship between date of onset and species’ thermal preferenda or thermal preferences. By producing a point estimate along with appropriate 95% CI, the G SI method provides useful information on the onset of growth and the uncertainty about that estimate. The G SI analyses can contribute to a better understanding of environmental influences on the onset of growth and the length of the growing season, and of thermal thresholds for growth, including their use in calculation of degree-day metrics. # 2008 The Authors Journal compilation # 2008 The Fisheries Society of the British Isles Key words: age classes; Akaike information criterion; backcalculation; hard structures; Lake St Pierre; model comparisons. INTRODUCTION Growth is a fundamental aspect of fish ecology and a crucial element in fish- eries research and management (Hoar et al., 1979; Summerfelt & Hall, 1987). Growth patterns can be compared among sexes, cohorts or populations, and can be related to environmental variables to help understand biotic and abiotic factors affecting growth rate (Weatherley & Gill, 1987; Elliot, 1994). Water temperature and the duration of the growing season are reliable predictors of fish growth and size (Conover, 1990; Power & van den Heuvel, 1999). Growth and fish size in turn influence various ecological characteristics of pop- ulations, such as predator–prey interactions, reproduction and vulnerability to fishing (Neuheimer & Taggart, 2007). The timing and duration of seasonal *Author to whom correspondence should be addressed. Tel.: þ1 819 376 5011, ext. 3363; fax: þ1 819 376 5084; email: marco.rodriguez@uqtr.ca Journal of Fish Biology (2008) 73, 153–168 doi:10.1111/j.1095-8649.2008.01922.x, available online at http://www.blackwell-synergy.com 153 # 2008 The Authors Journal compilation # 2008 The Fisheries Society of the British Isles