
OF PARENTS BY THOSE OF THE CHILDREN. (207) 15 
T. TAMMEs had already remarked (1911, p. 242) that such combin- 
ations of factors must be expected. | 
By means of the formulae which are composed of these factors, 
NILSSON-EHLE explains the results of his experiences and especially 
the presence of exemplars which surpass the length of the internodiae 
of both parents. From the formulae the constitution of the different 
F,-forms is to be retraced and to be tested to the further segregation 
in F,. Though these experiments have not been so far worked out as 
some of 1909, yet in the results NILSSON-EHLE sees a confirmation of 
his hypothesis. MAYER-GMELIN (1917) found deviating results in the 
sense that compactum would not be dominant. 
The striking of the cases of the tables [Ib and IIIb (p. 12) is that the 
index of so many children surpasses that of both parents. According 
to the proportions of numbers of the formulae this number is always 
but a moderate one. If both parents are heterozygotes and if there is 
only one pair of heredity-factors, then, for every four children the 
property of one child will surpass that of both parents. If there are two 
pairs of polymere factors, then, for every 16 children, the property of 
5 children can surpass that of both parents and if there are 3 pairs of 
polymere factors and if both parents are heterozygotes (A]a,A545A 323), 
then it is easily to be calculated that for 64 children the property of 22 
children can surpass that of both parents. So in the extreme case the 
property of 4 of the number of children can surpass that of both par- 
ents. That this should be applicable to our cases (p. 12), that PA 
the index of all 8 or of all 9 children surpasses that of both parents, 
will, I suppose not be admitted by chance (PEARSON 1900, p. 157). 
If, for our material, we simply take into consideration whether there 
are phenomena of segregation, then the value of the peculiar cases of 
the tables II and III of p. 12. can be eliminated by considering it as 
the effect of non-heredity variability. If the cases in which the indices 
of parents differ little and the indices of children form a mutually little 
differing series, it may f. i. be accepted that both parents are non-hered- 
itary minus-variations (table II), resp. plus-variations (table III). 
- Accepting f.i. for family 269 b, that the parents are non-hereditary 
factors. One L-factor has a weaker working than both other ones together ; the second 
however may perhaps dominate the other one, so that the working of the second 
factor alone would be the same as that of both L-factors together (See also FRETS 
1920 De 2) 
