Intrauterine growth retardation detected in several species by non-normal birthweight distributions.

The statistical distribution of birthweights in 64 litters of pigs, 48 litters of dogs, 59 litters of rabbits, 130 litters of rats and 46 litters of mice was examined. Birthweight followed a Gaussian or normal distribution in about two-thirds of the litters inspected, as judged visually using a normal probability plot, and by the W-test statistic. In the remainder, a single outlying observation in an otherwise normal sample was detected by Dixon's test, and/or a separate sub-population of low birthweight animals could be identified by fitting two normal populations to the observed samples. In the non-Gaussian litters, the average proportion of affected members was 40% of the litter. These results suggest that growth-retarded neonates should not be regarded simply as the tail-end of a normal distribution.