In the mite Macrocheles matrius (Hull), like in most Macrochelida species, the males arise from unfertilized eggs, while the females develop from fertilized eggs (arrhenotoky). This kind of reproduction should lead to oscillatory variations in the population sex ratio. In fact, in the case of a population with a 1 : 1 sex ratio or an excess of males, most (or the whole) of the eggs should be fertilized, determining in the following generations a considerable prevalence of females. In this new situation the probability of fertilization should be much lower, with a consequent increase in the percentage of non fertilized eggs. This should evidently introduce a new sex ratio inversion, with a numerical prevalence of males. Such sex ratio oscillations do not seem, however, to occur in nature. Collections repeated in various times of the year have indeed always shown a large excess of females (spanandry). In order to understand the causes of this situation we have studied for many successive generations the trend of sex ratio in M. matrius experimental populations reared under strictly constant environmental conditions and founded by an equal number of freshly moulted adult males and of female deutonymphs, next to moult (and so surely virgin). The main results obtained are as follows: a) experimental population sex ratio rapidly shifts from an initial 1 : 1 condition to a considerable pravalence of females. Already 20 days after the beginning of the culture the percentage of females among the adults reaches 90%; this value generally remains constant for the duration of the rearing, realizing a situation similar to the one observed in nature; b) the numerical prevalence of females among the adults is only due to a lower longevity and competitivity of the males, less mobile and smaller than the females, and often devoured by the latter, especially in overcrowding conditions; c) sex ratio at hatching very regularly remains next to 50%; this is not due to the existence in the culture of an equal number of virgin and inseminated females, but to the fact that the majority of the progeny consists of individuals of both sexes; d) in spite of the low number of adult males, relatively few females remain virgin: each male, in fact, normally inseminates many females; e) inseminated females lay both unfertilized eggs, that will give rise to males, and fertilized eggs, that will give rise to females; in particular, in the first 12 hours of egg-laying all the eggs are unfertilized; afterwards, and up to the 8th day from the beginning of egg-laying, fertilized eggs prevail (in some progeny between the 3rd and the 8th day all the eggs are fertilized); between the 8th and the 10th day the situation reverses again and almost all the laid eggs are virgin; after the 10th day of oviposition, and up to the end of the reproductive period, only unfertilized eggs are laid. This peculiar reproductive behaviour can be explained assuming: 1) that during copulation only a small number of spermatozoa, part of which probably still immature (prosperms), insufficient to fertilize all the mature eggs are transferred to the female; 2) that in the females spermatozoa are exhausted in a few days (or that they lose after a few days the capability of fertilizing the eggs); 3) that about a week after the beginning of adult stage females are no more inseminable. Only indirect evidences are available for points 1 and 2, while point 3 has been confirmed by us through experiments in which virgin females of different age were put together with one or more males. Virgin females adult since a week or more have not resulted receptive to insemination and have given rise exclusively to male progeny, even after the immission into the breeding container of one or more males. Finally it is to be noted that our data seem to exclude the possibility of a direct or indirect control from the mother of egg fertilization. The percentage of fertilized eggs would therefore depend only on the number of mature sperm present in the female receptacle. This percentage can vary from 0 to 100%, but our data clearly show that the latter value is reached only exceptionally and for very short periods of time (a few days). The average percentage of fertilized eggs in our experimental populations has always stayed next to 50%, with only very moderate shifts.