The garden chafer. Phyllopertha horticola (L.), renews its population every June. Its numbers fluctuate irregularly from year to year. During 1947 75 in the English Lake District, it had two ‘outbreaks’ (with peaks in 1952 and 1975) separated by a long period of low population (1956–70). In nine of the years 1948 75, October population increased significantly over that of the preceding year: in eleven, it did not change significantly; and in ihe remaining eight, it decreased significantly. These fluctuations in numbers were due to fluctuations in the effective environmental factors (weather, enemies and intraspecific competition) influencing sun ival, fecundity and migration. Weather was more effective than enemies in this respect and intraspecific competition became important only when weather and enemies together failed to keep numbers from rising dangerously high for the chafer. Weather was, of course, density-independent in action. Among enemies, interspecific competitors were density-independent while pathogens, parasites and predators were all imperfectly density-dependent. The only perfectly density-dependent factor was intraspecific competition. These findings bear out Milne's theory (1957a et seq.) that for part of the time, natural control of population increase of any species X is due to the combined action of density-independent and imperfectly density-dependent environmental factors but, whenever this combined action fails (as inevitably it sometimes must), increase to the point of collective suicide is prevented by the one and only perfectly density-dependent factor, namely, intraspecific competition of X; on the other hand, decrease of numbers of X to zero is prevented ultimately by density-independent factors alone.