164 



REACTION-INTENSITIES OF STARCHES. 



With these exceptions, the figures for the several 

 members of each group and each genus tend to be distrib- 

 uted among the several divisions in case of each genus 

 with remarkable uniformity, in some genera a conspicu- 

 ously large number falling among the very high, or the 

 very high and high reactions, or the very low, or the very 

 low and low reactivities, and so on. Such differences, of 

 themselves, are usually quite definite in making distinct 

 groups which upon comparison will be found to agree 

 remarkably with botanical classification. Thus Hippeas- 

 trum, Nerine, Gladiolus, and Tritonia are characterized 

 particularly by the relatively large number of reactions 

 that are very low (the number varying in the different 

 genera) and the fairly uniform distribution of the re- 

 maining reactions among the other divisions, chiefly 

 among the moderate and low. In Lilium, Phaius, and 

 Cymbidium the characterization is by the very large 

 number of very high reactions and the fairly uniform 

 distribution of the other reactions among the other 

 divisions, especially generally among the high and mod- 

 erate. In Amaryllis-Brunsvigia, Crinum, Hcemanthus, 

 Iris, Begonia, and Musa variations from these systems 

 may be observed because of certain subgeneric peculiari- 

 ties that have already been referred to. 



These data indicate quite clearly that peculiarities in 

 the distribution of these reaction-intensities are inti- 

 mately related to generic and subgeneric divisions, and 

 that when the distributions in the case of members of a 

 set or of a genus may be alike or nearly alike there may be 

 differences in the sums and averages that are more or less 

 definitely distinctive. For instance, the distribution in 

 Brunsdonna sanderce alba and B. sanderoa is identical, 

 but the sums and averages differ sufficiently to differ- 

 entiate these hybrids. In Nerine, the distributions dif- 

 fer very little ; in some cases the sums and averages are 

 absolutely or practically identical, and in others they 

 differ within small to very narrow limits. Under such 

 conditions positive identification of different members 

 of the group can not satisfactorily be made. Correspond- 

 ing conditions are found in relation to intergeneric dif- 

 ferentiation. Thus, the distributions in Flippeastrum 

 and Nerine are closely the same, and were dependence 

 placed upon this feature to distinguish genera it would 

 naturally be concluded that the genera are alike; but 

 upon a careful examination of the two sets of figures it 

 will be found that in Hippeastrum there is a manifest 

 tendency for a shifting of the reaction-intensities toward 

 the very low reactivity end, and in Nerine in the same 

 direction, but to a slightly less degree, so that in the final 

 summing up the sums and averages in the former fall 

 lower than in the latter in Hippeastrum, ranging from 

 748 to 925 and 29 to 36, respectively; and in Nerine 

 from 869 to 1199 and 33 to 46, respectively. In Glad- 

 iolus and Tritonia, very closely related genera, the dis- 

 tribution closely corresponds to the preceding groups in 

 the several respects referred to. On the other hand, 

 Lilium and Cymbidium, while in general very closely alike 

 in distribution, sum, and average are very markedly 

 different from all other groups. Phaius values bear a close 

 resemblance to the figures of Lilium and Cymbidium. 

 Iris in its first three sets stands apart from all other 

 genera in the manner of distribution of the reaction- 

 intensities, yet the sums and averages are close to but 



somewhat less than in Nerine. In other word?, different 

 genera may or may not exhibit distinctive peculiarities in 

 the distribution, sum, and average of the reaction-inten- 

 sities. The value of suuh data seems to lay particularly 

 in showing that members of a genus that are not ?o 

 differentiated as to fall into subgeneric divisions tend to 

 exhibit a method of distribution of the reaction-intensities 

 according to a definite system, which system is composed 

 of the averages of the number of very high, high, moder- 

 ate, low, and very low reaction-intensities, of the average 

 of the sum of the reaction-intensities, and of the average 

 of the latter. For comparative purposes the system repre- 

 sented by Hippeastrum, Iris (first three sets), and Lilium 

 may be taken because they show different types : 



If the figures for any given member of any one of the 

 genera represented be compared with the figures for the 

 genus, it will be found that those for the corresponding 

 columns differ, if at all, only within narrow limits. Thus), 

 in case of Hippeastrum the figure in the first column 

 of this table and chart is 2.8, while the figures for the 

 nine starches represented in this genus vary between 2 

 and 5 ; in the last column the figure is 12.8, while the 

 range for all of these starches is from 11 to 14. The sum 

 is 836, and the range from 748 to 925. The average is 

 31, and the range from 29 to 36. And so on with 7ns 

 and Lilium. When, however, there are subgeneric 

 groups there may be as many types as there are groups, 

 as is well illustrated by instances referred to. 



Obviously, the method of differentiating genera, sub- 

 generic groups, species, hybrids, and varieties by such 

 a system has its limitations, not because of the failure 

 of the data per se, but because of the faultiness of the 

 method of formulating the data. This is manifest, for 

 instance, in Hippeastrum and Nerine, in which the data 

 as tabulated indicate very closely related genera or even 

 subgenera, yet these genera, although belonging to the 

 same family, are well separated and are no't confounded 

 by the botanist. When, however, the data are presented 

 in other forms, as in other tables and charts, the genera 

 are as markedly differentiated from each other, and the 

 members of each genus from each other, as they are by 

 the data of the systematist. Finally, it is of interest 

 to note that in summing up these averages intermediate- 

 ness of the hybrid is not the rule, the tendency beinsj 

 more frequently for the hybrid values to exceed or fall 

 below those of the parents than to be intermediate. 



AVERAGE TEMPERATURES OF GELATINIZATION COMPARED 

 WITH THE AVERAGE REACTION-INTENSITIES. 



(Table B 2, Chart B 42) 



During the progress of the research it was found that 

 the temperatures of gelatinization bore varying relation- 

 ships to the average reaction-intensities, as a whole, of 

 different members of certain sets, different sets, and dif- 



