310 



SUMMARIES OP THE HISTOLOGIC CHARACTERS, ETC. 



included in this phase; (2) those in which two phases are 

 definitely dominant, but which may be quite different in 

 value; (3) those in which three phases are dominant, 

 but which may have different values; and (4) those in 

 which the parental relationships of the hybrid seem to 

 be directed largely indifferently to the several phases. 

 Among the starches that were studied in all of the 26 

 reactions it is rare, as, for instance in /rife dorak, to 

 find that the assignment is not unmistakable. Where 

 the number of reactions is restricted to 10 to 13 the 

 classification is often indefinite. The grouping in 

 accordance with the foregoing is as follows : 



* Number of reactiona when less than 26. 



The distribution of the hybrids among the four 

 classes is fairly uniform except in the third class, there 

 being 13 (26 per cent) in the first class, 14 (28 per cent) 

 in the second class, 8 (6 per cent) in the third class, 

 and 15 (30 per cent) in the fourth class. In the first 

 class, 4 of the hybrids are characterized by the con- 

 spicuousness of intermediateness, this phase of parental 

 relationship being noted in one hybrid in 18 of the 26 

 reactions, in another in 16 of 26 reactions, in another 

 in 7 of 10 reactions, and in another in 11 of 13 reactions. 

 In 4 hybrids the characterization is especially in de- 

 velopment in excess of parental extremes, this phase 

 being recorded in one in 21 of the 26 reactions, in 

 another in 20 of the 26 reactions, in another in 8 of 10 

 reactions, and in another in 17 of 26 reactions. In 5 

 hybrids the characterization is especially by development 

 in deficit of parental extremes, this being found in one 

 in 13 of 26 reactions, in another in 14 of 26 reactions, 

 in another in 16 of 26 reactions, in another in 17 of 26 

 reactions, and in another in 20 of 26 reactions. In the 

 second class, the dominant figure of the couple is found 

 in 1 hybrid under the phase the same as the seed parent, 

 in 5 under intermediate, in 2 under highest, and in 3 

 under lowest; in 1 there is duplication of the figures 

 under the phases the same as the pollen parent and inter- 

 mediate, and in another under intermediate and high- 

 est. This coupling is more marked in the instances 

 where 26 reactions were studied than when the number 

 is 10 or 13. In the third class there is not only less 

 tendency to a very marked degree of characterization 

 as regards any one or more of these phases, but also to 

 the characterization being present in three phases usually 

 with slight gradation, as, for instance, in Nerine dainty 

 maid where the values are 7, 6, and 8 under same as 

 both parents, intermediate, and highest, respectively; 

 and in Nerine glory of sarnia, where the values are 6, 8, 

 and 10 under same as pollen parent, same as both parents, 

 and lowest, respectively. Or there may be some dupli- 

 cation, as, for instance, in Lilium dalhansoni, where the 

 values are 4, 9, and 9 under same as seed parent, same 

 as both parents and intermediate, respectively, etc. 



Prom this limited data one may expect that further 

 studies will elicit various combinations of both phases 

 and values. In one hybrid the highest number of the 

 triple is found under same as seed parent, in two under 

 intermediate, in two under highest, and in one, under low- 

 est. In one there is duplication of the highest values 

 under same as both parents and intermediate ; and in an- 

 other under same as both parents and highest. In the 

 three hybrids with which in each only 10 reactions were 

 recorded the grouping of the phases in triplets does not 

 yield the striking comparisons that are observed when 

 the reactions number 26, or 21/2 times larger. In the 

 fourth class, with 7 of the 15 hybrids only 10 reactions 

 were recorded in each, and in these instances the values 

 are (with possibly two exceptions, Narcissus pyramus 

 and N. madame de graaff) so distributed among the dif- 

 ferent phases that there is not the convincing evidence 

 of a well-defined inclination of the hybrids in their 

 parental relationships that was found in corresponding 

 cases in the preceding classes. Among the remaining 

 8 there is marked dominance of 1 phase of the 6 in a 

 single hybrid (Iris doralc) in which 11 of the 26 reac- 

 tions fall under highest, the other values being 5, 3, 2, 1, 



