The Detection of Regulation in Animal Populations

If the detection of regulation in a population is attempted by taking a linear regression of log population density against log previous population density, the slope, b, of the line thus obtained is supposed to be less than 1 only if regulatory tendencies are present in the population. However, a nonregulated population will probably give similar values of b < 1 provided only that the successive differences between log population values are not all the same. This spurious indication of regulation becomes more likely and more apparent if the number of census points is small, or if the variation in numbers of animals is large compared to the total population change over the period measured. Complexities which are almost certain to occur in natural situations and which make satisfactory models for regulated and nonregulated populations intractable or indistinguishable probably require that investigations of regulation either use more detailed information that census data alone, or that they adopt an experim...