Stage-specific density dependencies of blowfly populations: experimental testing of non-parametric model predictions

Periodic fluctuations in abundance is a characteristic feature of many insect populations. It is necessary to understand the underlying mechanisms in order to predict or regulate insect populations. A.J. Nicholson showed that population-intrinsic factors produced sustained oscillations in experimental blowfly populations (Lucilia cuprina (Wied.)). Non-linearities and time-lags in density dependence are of particular importance because they can give rise to complicated population dynamics such as cycles or chaos. We used a non-parametric modelling method, GAM (generalized additive modelling), to estimate density dependences in time-series data from experimental populations of Lucilia sericata (Meig.). This method makes no assumptions about the structure of density dependences and is therefore very useful for identifying non-linearities. The model indicated that 1) larval survival was non-linearly dependent on larval density (highest survival at intermediate densities), 2) adult survival was density-independent, and 3) reproduction was dependent on adult density. These predictions were tested experimentally in order to evaluate the model. The experimental results confirmed the non-linear relationship between larval survival and density, and the density-independence of adult survival. Moreover, delayed density dependence was shown in the reproductive rate: it was dependent on past larval density rather than on present adult density. Because of the correlation between larval and adult density within the same generation their effects can not be separated by time-series analysis, only by experimental testing. The accordance between the model predictions and the experimental results show that the non-parametric modelling method can be a powerful tool for analysing ecological time series, and for designing critical experiments.