APPLICATIONS TO SYSTEMATIC BOTANY. 341 



sias have also been confounded, and, in fact, among the earliest illustrations of Freesia 

 are pictures of Tritonia refracta and T. odorala. While the starch-grams of these genera 

 ai-e so ahke in shape that it would be difficult or impossible to distinguish one from the 

 other with certainty, the reaction-curves differ distinctly, and those of Freesias bear much 

 closer resemblances to T. pottsii and its hybrid than to T. crocata and T. securigera. 



Finally, the systematic classification of plants is of an arbitrary character, as is evi- 

 dent in the large number of classifications that have been used frum time to time, in the 

 continual changes that are going on at the present time, and in the recognition that there 

 is no system at present that is uni\'ersally accepted or which is regarded as having more 

 than a tentative value. In view of these facts, coupled with the specificities of starches 

 which have been shown to exist in relation to species and genera, and the concordance 

 of these specificities with the classification of the botanist whenever the assignment of the 

 latter has been accepted as being absolutely or reasonably permanent, it would seem to 

 follow, as a corollary, that where doubt exists as to classification that is based upon botan- 

 ical characters, such doubt can be confirmed or set aside and revisions suggested by the 

 reaction method employed in this investigation. It has been conclusively shown that 

 all members of a genus have a certain tj'^pe of reaction, and that each may be distinguished 

 from the others by variations of the reaction-curves; that in the case of all genera of a 

 given family the genera when closely related show striking resemblances in the types of 

 their reaction-curves; and that the farther the members of a genus or of a family are 

 separated, the greater the departures from the corresponding given types of curves that 

 may be taken to be typical of the genus or family. If, therefore, in a given recognized 

 genus we find among its members one, for instance, that has a reaction-curve that does 

 not correspond with those of the others, it may be taken for granted that it either belongs 

 to another genus or is a hybrid resulting from a cross with a member of another genus, 

 or some aberrant form, etc.; and again, if the reaction-curves of different genera assigned 

 to a family do not correspond in general characters, it may be likewise concluded that one 

 or more of the genera are inisclassified. 



The systematic botanist, by the conventional methods of research, is continually 

 finding e\'idence of misclassification, as is evident in the continual shifting of species and 

 genera in the rebuilcUng of broken-up famifies, and in the breaking-up of families by assign- 

 ing certain genera to other families or establishing entirely new fanfifies. Thus we find in 

 very recent years that ConvaUaria, Trillium, and Asparagus have not, by some botanists, 

 been included among the Liliacece, but have been set apart as a distinct family, Conval- 

 lariacece; and Aletris has been transferred to Hoemodoraceoe. Musacece, Zingiberacea, Can- 

 naccw, and Marantaccce are by some grouped in one family, but by others as separate families. 

 We are therefore deafing not with a stable classification, but one that is most unstable. If 

 therefore, as stated, we find in the reaction-cm-ves evidence of misclassification, it may be 

 taken for granted that we have the strongest kind of evidence to suggest modifications. 

 For instance, it will be seen that reaction-curves of Lilium, Fritillaria, Calochortus, Tulipa, 

 Scilla, Chionodoxa, Ornithogalimi, Puschkinia, and Enjthronium are all in accord with a given 

 type; that in Hyacinthiis and Muscari there is a marked modification of tliis type wliich is 

 manifested chiefly in the much lower temperature of gelatinization and the low reactive in- 

 tensity with Purdy's solution; that in Galtonia there is another kind of mocUfication, and of 

 a character which suggests that Galtonia stands, as it were, between Hyacinthus and Muscari 

 on the one hand and the fii'st group (which constitutes almost wholly the tulip tribe) on the 

 other; that in Brodimi and Triteleia there are types of curves wliich bear close resemblances 

 to each other, but which cfiffer distinctly from those of the other groups noted. According 

 to tliis data, without going into further detail, all of wliich is in accord with the data of the 

 botanist, it would seem that there is a logical basis for a separation of the Liliacew into a 

 number of families, each of which can be distinguished by the sum of the botanical and 



