Single‐Factor Analysis in Population Dynamics

Abstract

In the study of natural populations it is desirable btit not always feasible to measure the effects of all mortality factors. This requires frequent population sampling, supported by data on natural enemies, climate, and other factors, and leads to the compilation of detailed life tables for successive generations. These tables must be compiled simultaneously for a number of differentt environments if it is desired to learn why population density varies in place as well as in time. It is important to know, for example, why a forest insect periodically develops high populations in certain well-defined stand types but not in all types where it occurs. This multi-factor approach therefore demands much time and effort and, for species that are difficult to sample, is more suitable for a team of investigators than for the individual. A preliminary examination of rather extensive life-table data for the spruce budworm, Choristoneura fumniferana (Clem.), suggested that the factors affecting this species in any one place are of two types-those that cause a relatively constant mortality from year to year and contribute little to population variation, and those that cause a variable, though perhaps much smaller, mortality and appear to be largely responsible for the observed changes in population (Morris 1957). A factor of the latter type will here be called a 'key factor,' meaning simply that changes in population density from generation to generation are closely related to the degree of mortality caused by this factor, which therefore has predictive value. This leads to some interesting questions: Do these key factors exist for other species or are most populations affected by so many interacting factors that they behave like random numbers? If a key factor is suspected in a population, can its existence be demonstrated effectively by a very limited 'single-factor' approach in which only population density and the mortality caused by this factor are measured in each generation? If so it may be possible, with the expenditure of only a few man-days each season, to discover relationships of value for the prediction of population changes. This would be particularly useful for certain species of economic importance which, be-