An exponential growth model with decreasing r captures bottom-up effects on the population growth of Aphis glycines Matsumura (Hemiptera: Aphididae)

1 There is ample evidence that the life history and population dynamics of aphids are closely linked to plant phenology. Based on life table studies, it has been proposed that the growth of aphid populations could be modeled with an exponential growth model, with r decreasing linearly with time. This model has never been tested under field conditions. 2 The soybean aphid Aphis glycines is a new invasive pest to soybean production in the U.S.A. In the present study, we present five datasets on the growth of colonies of A. glycines, monitored during population growth and decline under predator-free conditions in three soybean fields, from 2003 to 2006. 3 We demonstrate that an exponential growth model, with r decreasing linearly with time, gives a much better description of A. glycines dynamics for all datasets (R-2 = 0.94-0.99) than the exponential (R-2 = 0.42-0.98) or logistic growth models (R-2 = 0.77-0.99). Furthermore, it is shown by cross-validation that the exponential model with decreasing r can be used to make population predictions, as shown by the coefficient of prediction, R-pred(2) ranging from 0.55 to 0.97. An improved proved fit of the model was obtained using both aphid (R-pred(2) > 0.79) and soybean pred (R-pred(2) > 0.86) degree-days scales, indicating temperature effects on the phenological time scale for the decrease in r. 4 Our model suggests important bottom-up control of A. glycines population growth, which may interact with other mortality factors. The generality and potential applications of these results are discussed.