CHOSSINGS IN MELANIUM-VIOLETS 



257 



varied iiult'pendently so lar as I tuuld see. These tads do not speak 

 ill favour of the supposition of a monohyhrid segre{^ation. 

 » No classification was made of the direction of tlie stigmatic 



chamber and of tlie pollen-magazine (o])en or closed). Thest' charac- 

 ters could hardly he measured and an ocular estimation would he too 

 uncertain. 



In order to state the variation in size of the lahellum it» width 

 was measured. The shape of the lahellum is shown in fig. 1. The 

 distance a — b was measured. It was necessary to do the measure- 

 ments with the aid of the microscope. Great difficulties were met 

 with in the efforts made to adjust the lahellum in a perfect profile 

 position. There were some additional difficulties to overcome. The 

 stigma was sometimes more or less eaten by insects; further it was 

 often difficidt to rid the stigma from adherent pollen. Therefore, I 

 could only determine the means of 66 individuals; 5 — 15 flowers were 

 measured on each plant. 



The means of the breadth of the lahellum was 3,.'j scale divisions in 

 V. arucnsis, 8,2 in V. tricolor and 4,.') in Fi. In a previous paper (1916) 

 I have stated the breadth in ^^o ni. m. This statement is not 

 correct — the lahellum is much broader — and depends on an er- 

 roneous calculation of the magnification. In this paper the breadth 

 is stated in divisions of an ocular scale and therefore only the relative 

 variation in size is shown. However, onlv this one is of real interest here. 



TABLE 2. The variation of the breadth of the labeUiim in Fj 



It is doubtful whether or not the segregation is transgressive. 

 The modification amplitude is rh 1 scale division in the arvensis-ïorm 

 and a little higher in the tricolor-ïorm. Thus, the variation may very 

 well keep within the limits of the parents. Further, the table shows that 

 most of the F^-individuals group themselves around the mean of 

 V. arvensis. This may depend on the presence of an inhibiting factor 

 in V. arvensis. which I will try to show later. Whether this segrega- 

 tion is monohybrid is not easy to say. 



An adequate statement of this seemingly monohybrid segregation 

 of herkogamy and autogamy is not easy to give. 



