EXPERIMENTAL MAMMALIAN TERATOLOGY 243 



malformation may be classified as individual factors, affecting littermates differently 

 (local variations in blood supply, implantation, and so forth) and background factors, 

 affecting all embryos within a litter (maternal physiology, variations in size of uterine 

 artery, and so forth). Within inbred strains, it may be possible to detect such back- 

 ground factors by comparing the frequencies with which defective offspring occur 

 within litters with the distribution expected if the probability of being malformed is 

 constant from embryo to embryo. If there is a tendency for clustering, that is, for an 

 excess of litters with many defective offspring and of litters with none, a maternal 

 variable is suggested. A tendency for some mothers to have more defective offspring 

 than others (in the absence of genetic segregation) suggests that the maternal factor is 

 a relatively permanent constitutional factor, whereas fluctuations in the maternal 

 environment may be implicated if the variance is greater between litters than between 

 mothers. 



6. What are the intra-uterine variables related to the production of malformations? — One of 

 the most difficult problems to attack experimentally in teratology is the fact that, when 

 a teratogen is applied to a genetically homogeneous litter in the same uterus, some 

 embryos may have no malformations and among the malformed ones there may be 

 considerable variation in nature and severity of the defect from one littermate to 

 another. In some cases this can be accounted for on the basis of quasicontinuous 

 variation — a continuous distribution separated by some developmental threshold 

 into discontinuous parts. 501 In the case of cleft palate, for instance, the time of closure 

 of the embryonic palate can be considered a continuously distributed variable. If move- 

 ment of the shelves from their vertical position on either side of the tongue to the 

 horizontal position above the tongue is sufficiently delayed, the width of the head will 

 have increased so much that the shelves cannot meet in the midline when they do reach 

 their final position. Thus the point at which they can no longer meet separates the 

 continuous distribution of times of palatal closure into a discontinuous one- — cleft, 

 or not cleft. 405 However, this interpretation only pushes the problem back one 

 step. What are the factors that determine where on the continuous distribution an 

 individual embryo will lie ? Why does shelf movement occur so late in one embryo 

 that the teratogenic procedure pushes it over the threshold, while in its genetically 

 similar littermate shelf movement occurs early enough to keep it on the side of 

 normality? Is some other variable perhaps involved, such as variation in shelf width? 

 So far, no relevant factors have been identified. In addition, there may be variations 

 within a litter in the amount of teratogen that actually reaches the embryo, and this 

 aspect of the question remains almost entirely unexplored. 



The problem of intra-uterine variability is perhaps even more important in the 

 case of the spontaneous malformation that appears with a low but characteristic 

 frequency in certain inbred strains, for instance, the cleft lips that occur in about 10 

 per cent of newborns in the A/Jax strain, or the microphthalmia that occurs in some 

 sublines of the C57BL strain. Here the situation may be more analogous to the situation 

 in humans ; a particular grouping of genetic factors increases the probability that an 



