The tubes were sterilised before inoculation with mites and kept in a moist chamber at 22?C. To prevent the mites crawling from one tube to another the tubes were kept in wide flasks containing 5 % KOH on the bottom. In addition, the rims of the flasks were covered with vaseline to prevent the mites from crawling over. The computation of the population was made at fixed intervals, and its technical details have already been described (4). Let us only note here that the contents of a culture was carefully mixed up with the aid of a thin needle, poured out on a glass plate, and divided into several equal parts (2, 4, 8, or 16). In two of such samples the total number of mites was counted (under a Zeiss binocular microscope), and then recalculated for the whole culture. After being counted, all animals were returned to the original culture. With such technique of counting we could separately keep the records for males, females, hexapod and octopod larvae; no attempt to count the total number of eggs was made in the present investigation.The results of our previous experiments, made on wheat flour (4), can be summarised in the following four possibilities of interaction: (1) If we add a small number of predators to a dense population of the prey the predators cannot multiply and so perish. This is apparently due to the specific smell emanating from the prey. (2) If, however, we put the predators into a very sparse population of prey, the latter cannot be discovered and the predators will similarly perish-from starvation. An interaction takes place in the zone lying between these two extremes. (3) When the concentration of predators is considerable, they are able to devour the population of prey completely. The adults and large larvae of the prey are all devoured by predators. On small hexapod larvae and eggs, the attacks of the predators are much weaker and to a certain extent the prey escape destruction. The predators, when numerous, are in a position to devour the remaining prey as they get older. (4) But if the predators reach a fairly low concentration, they will not be able, after eating a large proportion of the prey, to destroy the remainder completely. The prey will begin to increase, this increase being in its turn followed by a new increase of the remaining predators. However, such periodic fluctuations in numbers are observed comparatively seldom in the absence of immigration from outside, and the interaction usually ceases, having followed type (3).
