One of the main problems in ecology is identification of processes that determine the pattern of population dynamics. For example, many insect population periodically increase in numbers to the outbreak level and cause intensive damage of crops or forests. Predicting and preventing of outbreaks requires the knowledge of processes that determine this pattern of population dynamics. The first quantitative approach to solve this problem was suggested by Varley and Gradwell who proposed the 'key-factor analysis' of life tables. Their assumption was that the most important mortality process should have a high contribution in terms of ¿-value to the total ¿-value for the entire life cycle, K, and its change over the years should closely correlate with K.
According to Varley and Gradwell, the mortality process called 'winter disappearance' appeared to have the maximum contribution (¿1) to the total K in the dynamics of winter moth (Operophtera brumata) in Great Britain (Figure 2). Also the pattern of change in ¿1 from year to year closely resembled the pattern of change in K. Thus, winter disappearance was considered to be the 'key factor' that drives the population dynamics of the winter moth. Although this method is definitely useful for understanding the mechanisms of population dynamics, it has several weak points. First, winter disappearance was estimated by subtraction of population numbers in the fall and in spring; hence, Varley and Gradwell failed to identify the immediate mortality cause (they hypothesized that it was predation). Second, the meaning of a 'key' factor was not explicitly defined. Thus, it is not clear what predictions can be made from the knowledge that some mortality process is a key factor. For example, it does not help us to develop strategies of pest control or conservation of endangered species. The key-factor analysis was often considered as a substitute for modeling. It seems easier to compile time series of ¿-values and to find key factors without developing models of ecological processes. However, reliable predictions require models in addition to life tables.
Was this article helpful?