Indeed, the evidence for the individuality and continuity of inheritance units is based entirely upon such segregation and recombination, so that the entire Mendelian theory may be said to rest upon the principle of segregation. If there are cases in which such segregation does not take place they belong to other forms of inheritance than the Mendelian: if segregation occurs in every instance there is no other type of inheritance than that discovered by Mendel. Are there cases which do not segregate according to Mendelian expectation?
Fig. 64. Diagram of Inheritance of Color-blindness through Female. A color-blind female transmits her defect to all her sons, to half of her granddaughters and to half of her grandsons. Corresponding distribution of sex chromosomes on right. (After Morgan.)
When the Mendelian theory was new it was generally supposed that there were forms of inheritance which differed materially from the Mendelian type; indeed, it was supposed that the latter was one of the less common forms of heredity and that blending of parental traits and not segregation was the rule. All cases in which the characters of the parents appeared to blend in the offspring, or in which there was not a clear segregation of the parental types in the F2 generation or in which the ratio for a monohybrid differed from the well known 3:1 ratio, were supposed to be non-Mendelian.
However, further work has shown that some of these are really Mendelian. Sometimes offspring are intermediate between their parents owing to incompleteness of dominance, rather than to incompleteness of segregation; in such cases the parental types reappear in the F2 generation as in the cross between red and white four-o'clocks. Sometimes departures from the 3:1 ratio are caused by the fact that two or more factors of the same sort are involved in the production of a single character. Nilsson-Ehle found that when oats with black glumes were
