NOTES AND CONCLUSIONS. 



371 



are the same bul modified in their applications in a rtain 

 important respects. In both investigations the histo- 

 logic properties, the iodine and aniline reactions, and 

 the gelatinization reactions with heat and various chemi- 

 cal reagents were studied, the chief differences bein 

 in the method of recording the reactions with the chemi- 

 cal reagents, and in the kinds and i lentrations of the 



reagents. In the former research the quantitative dif- 

 ferentiations by means of the chemical reagents were 

 made by determining the time of the occurrence of com- 

 plete or practically complete gelatinization, and the 

 preparations of starch with the reagenl wit.' uol ade 

 quately protected from the air and evaporation. It was 

 found during the progress of this work that fictitious 

 values may be recorded owing to the existence in nearly 

 ever] form of starch of different kinds of grains which 

 vary in proportions and gelatin izabili ties, together with 

 varying degrees of influence of the air (probably chiefly 

 or solely differences in oxidation), and effects that are 

 due tn varying rapidity and degrees of evaporation. 

 Such sources of fallacy have been practically eliminated 

 in the present research by making records of the progn 

 of gelatinization in regard to both the entire number of 

 grains completely gelatinized and the percentage of the 

 total starch gelatinized at definite time-intervals; and 

 by the prevention of oxidation and evaporation by seal- 

 ing the preparations. In nearly every form of starch 

 there are grains, usually very small, and also parts of 

 grains, that are quite resistant to reagents. The former 

 commonly represent much less than 5 per cent of the 

 total quantity of starch, and it has been assumed that 

 gelatinization is practically complete when 95 per cent of 

 ill.- total starch bus been gelatinized. The methods used 

 and their values in the differentiation of starches have 

 been se1 forth in full in the preceding memoir on pages 

 305 to 313, and supplementary statements are to be 

 found in the present memoir in Chapters II, IV, and V. 

 The histologic method employed in this research is 

 the same in all respects as in the preceding investigation, 

 in the report of which it has been discussed with suffi- 

 cient fulness (page 307). Its value has not only been sub- 

 stantiated but accentuated by the results of the present 

 study of the starches of parent- and hybrid-stocks. 



The polariscopic, iodine, and aniline methods are so 

 crude that the personal equation enters largely into the 

 determination of the values recorded, and while they 

 have proved of unquestionable usefulness they are so 

 inferior to the gelatinization method that they should 

 be given a very subordinate place. The polarization and 

 aniline miethod arc by far the least useful of all of those 

 used, yet the anilines will be found of much value in the 

 differentiation of different lamellre of individual grains, 

 as has been shown by I he work of Dennistnn (see pre- 

 vious Memoir, page 56). Iodine, like (he anilines, can 

 be used to gnat advantage in the study of the structure 

 of the starch grain. It is also of usefulness by showing 

 by variations in the color rent ion- differences in the 

 constitution of starches from different sources; of dif- 

 ferent kinds of grains of the same starch; of the capsular 

 and intracapsular parts of the grains; and of the cap- 

 Bules themselves. The method used in determining the 

 temperature of gelatinization is practical!] exact, as has 

 been shown by the fact that when the experiments are 

 made with proper care the figures recorded are quite as 



uniform as those obtained in the determination of the 

 melting-points of various Bubstan 



! . ionization method by means of 



chemical n i here pur ned a i to be so 



uear ej ai t thai the r cord - •■ lents 



have, c ;cep1 very rarely, be n found to bi i ta tl 

 bically exactly the same, even though made al 

 different pi-no, is and with varying temperature and 

 humidity. Very rarely, for on, a 



more or less marked! ord has bi .In 



everj in tancc this error was detected because of the 

 ab ence of agreement with what was positively indii 

 c i onditions. In fact, as was fou ad as 



will be obvious by the context, tic tions 



obtained by means of the various methods employed are 

 in the case of each ag n and reagent, and of all collec- 

 tively, in a very large mea are - ich other. 

 In other words, the values for tie of a given spe- 

 cies serve as prototype i r tandard with which 

 the records of all other spi i ieties of the genus 

 must conform, unless there arc represented mem 

 of subgenera or otb leric divisions. The cl 



botanically the - or th arieties t ] c will 



the records collectively agree witli the given standard. 

 Varieties of a species exhibit remarkable i I ■■- n 38, and 

 their values represent a specie; type. When members 

 of subgenera or other form of subgeneri i are 



represented they may exhibit differences that are as 

 marked, and even more marked, than those of members 

 of closely related genera. 



It is to be borne in mind that the method of classi- 

 fication of the systematist is of an arbitrary character, 

 as is evident, for instance, in the shifting of species from 

 one to another genus, the remodeling of genera, families, 

 etc. This classifying and reclassifying that has been in 

 progress for generations continues at the present time, 

 and even now the most generally -sification 



can not be accepted as being more than tentative. If, 

 therefore, the results of these invest] eem t i be 



or are not in accord in isolated instances with the classi- 

 fication of the systematist it does not follow that the 

 former are wrong. As evidence of the mutual checking 

 of the records one need examine only the very similar 

 curves of the starches of the closely related members of 

 Iris (('harts E 30 to E 33) and Richardia (Chart E 10) : 

 the dissimilar curves of the starches of members of 

 subgeneric divisions, such as the hardy and te 

 species of Crinum (Charts E 1 to E9); milaT 



curves of the starches of members of subgenera i 

 gonia (Chart- E 36 to E 39 1 ; the 

 the starches of the closely related genera Amaryllis and 

 Brunsvigia ( Chart El), am : ^>niii 



(Charts E 34 and E35); and the dissimilar curves, 

 usually highly characteristic, of the starches of various 

 genera of the same and different families that are shown 

 in tin of charts (E I to E 16), >le. These 



similarities and dissimilarities are in degree varia 

 accordance with what in general should be expected, or 

 what is at least in accord i th unques tanical 



classification. 



The differentiation of starches by heat, as in the 

 temperature of gelatinization method, is to be recom- 

 mended as heme- of much value, both quantitativel] 

 qualitatively. It was shown in the preceding in 



