Breeding experiments which show that hybridisation and mutation etc. 24I 



each other, but the departure from a i : i ratio in the totals is too 

 great to be accounted for by chance. 



A comparison of the totals in sections VI and III is instructive, 

 for it shows that (grandiflora x rubricalyx) x grandiflora yields^ (with 

 373 plants) a ratio for R:r of 1.39 : i; while the reciprocal crossed 

 back with the same parent, i. e., (rubricalyx x grandiflora) x grandi- 

 flora gives, in a total of 300 plants, practically the same ratio, namely 

 1.33 : I. The coincidence of these two ratios adds greatly to the signi- 

 ficance of the results, for not only does it show that reciprocal crosses 

 57ield the same ratio, but that six families, five of which are derived 

 from separate seed-parents, give ratios which are not significantly 

 different from each other. In other words, these results are not only 

 self -consistent but they mutually strengthen the whole interpretation. 

 This simple unit-character is apparently not inherited in the way 

 Mendelian characters are supposed to be inherited. 



It will be seen that there is a varying degree of excess of R's over 

 Mendelian expectation, in all the cultures. The consistency of the 

 results may be tested by comparing the results of various crosses in 

 which the same individual takes part. I have made a number of such 

 tests and have found no inconsistencies. These tests have brought 

 out several facts which I shall mention here which make still clearer 

 the meaning of the whole mass of data. Thus in culture 49 plant No. VI. 6 

 of grandiflora :■ rubricalyx gives a ratio for R : r oi 142 : 15 or 9.5 : i ; 

 yet with such a large excess of i?'s, when No. VI. 6 is crossed back with 

 grandiflora (cultures 74 and 78), the ratio is only 180 : 131 or 1.37 : i(i), 

 i. e., the excess over the expected Mendelian equality is much less than 

 the excess of 9.5 : i over a 3 : i ratio. This cross was made on August 17, 

 while the selling was done on August 14. It is conceivable that some 

 change in the meteorological conditions might hav-e influenced the 

 result, but I have no data to show this. I am inclined to believe, 

 however, that the cause is of a different nature; for the same result, 

 i. e., corresponding decrease in the precentage of i?'s, on crossing back 

 with grandiflora, is seen in all the other cases. Thus the offspring of 

 No. IX. 4 (culture 48) gives a ratio of 4.25 : i, and similarly the off- 

 spring of No. VI. I (culture 62) yield the ratio 5.15 : i; yet when the 

 former plant is crossed with grandiflora (culture 73) the ratio R : r 



(1) Since this is the only plant which gave a ratio approaching 10 : i, and 

 since when crossed back it gave practically the same ratio as the plants which in 

 F2 produced a 5 : i ratio, it seems possible that some unknown source of error may 

 have produced the aberrant ratio 9.5 : i. 



Induktive Abstämmlings- und Vererbungslehre. XI. 1 ^ 



