Abstract
Recent work on the relative abundance of species in samples taken from animal and plant populations in the field shows that there is nearly always evidence of a regular pattern of frequency distribution. This has usually been expressed in terms of logarithmic series (which fit best to smaller samples) or log-normal (which fits better to larger samples). A speculative attempt is made to extend these 2 series to the insect population of the world. It is suggested as a working basis that at any moment the insect population of the world is 1018 individuals and that there are perhaps as many as 3 million species. On these assumptions the log-series indicates a large number of rare species, with the median number of individuals per species just under two millions. With the log-normal it is necessary to make a further assumption of a value for the standard deviation of the distribution and from evidence of large samples it is suggested that this might be approximately e4. In this case the median number of individuals per species will be (for 1018 individuals and 3 x 106 species) approximately 1.2 x 108. The probable number of individuals in the most abundant species is also considered. For the log series there is not yet any reliable formula which can be extrapolated so far. With the log normal, it would seem that 1 or 2 species out of 3 million might contain 3% or perhaps 5% of the total population, or 3 - 5 x 1016 individuals.