Published research concerning disease development often shows that the response of fungal development to temperature is non-linear. Unfortunately, this non-linear response has often been ignored when predicting fungal development under varying temperatures using non-linear models and when deriving fungal development models from data collected under fluctuating temperatures. In this paper, the magnitude of non-linear effects on fungal development is shown to depend on the types of non-linear models and on the extent of temperature fluctuations. A method is described, which has been used in other disciplines to fit non-linear models directly to varying temperatures. Hypothetical data were generated to demonstrate the usefulness of this method. With the underlying rate equation being non-linear, models derived from average temperatures underestimate the rates at intermediate temperatures; the greater the temperature fluctuation, the greater this underestimation.