It is well established that many characters in mammals are affected by naturally occurring variation in the uterus or, to use a more inclusive term, the prenatal maternal environment. Although this factor is regarded as environmental as far as the embryo is concerned, it is itself presumably subject to both environmental and genetic factors (Russell'). The work of Wright, 2 Wright and Chase, 3 Green 4 and others,'who made statistical analyses of variation within inbred strains of mammals, clearly shows that the characters which they studied, polydactyly, white spotting, number of vertebrae, etc., were affected by environmentally determined variation in the prenatal maternal environment. The effect of genetically determined variation in the prenatal maternal environment on character variation has, however, received little attention, perhaps because of a lack of easy methods of evaluating it. Evidence that this factor may be important is provided by the increasing number of cases in which a difference has been observed between reciprocal F1 hybrids obtained by crossing inbred strains. Russell and Green6 have reported one-such case for number of lumbar vertebrae in mice. Similar data from other crosses have been obtained and are in preparation for publication. The mere discovery of such a difference does not, however, furnish proof of a difference between the maternal environments of the parental strains: without further analysis the difference between the hybrids could equally well be attributed to the eggcytoplasm or, in the heterogametic sex, to sex linked genes.

Two methods have beendeveloped which can be used to test directly for differences between prenatal maternal environments. The first has so far been utilized for this purpose on a large scale only by Fekete and Little. 7 In their work, fertilized ova are removed from the oviduct of one inbred