IV PREFATORY NOTES. 



methods far less exacting in technical skill and time. Hence, the histological ex- 

 aminations have been limited to the grosser characters of the grains, with especial 

 reference to their application to the sciences, arts, and trades. It would have been 

 well worth the time and labor, as was discovered rather late in the research, had 

 these examinations been carried out in extenso, because there are very many more 

 or less inconspicuous characteristics pertaining to the form of the grain, the size 

 and form and exact location of the hilum, the fissuration, the primary lamellse, 

 the positions and the general and special characters of the secondary lamellse, the 

 characters of the grains as regards isolation and aggregation and compoundness, 

 etc., which are of undoubted importance not only in the identification of the source 

 of the starch but also in expressing biological peculiarities of the starch-forming 

 plastids, and hence of the peculiar constitutional form of the starch. 



HISTOLOGICAL AND STEREOCHEMIC TYPES. 



Notwithstanding the very superficial character of the records included under 

 Histological Characteristics, it will be found that definite relationships can be traced 

 between the histology of the starch-grain and the curve of reaction-intensity, or 

 in other words between histological type and stereochemic type. Thus, in the 

 starch of a member of a given genus (such as Convallaria, page 616), having two 

 entirely different types of starch-grains, there are corresponding differences in the 

 types of reaction-intensities. The starches obtained from members of a genus, 

 such as Lilium (page 474), have the same histological and reaction types; likewise 

 starches from closely related genera, such as those of the tribe Tulipece (page 613), 

 have histological and reaction types which are in entire harmony. On the other 

 hand, where starches of the same gross histological type occur in genera of unrelated 

 families, as in Arum (plate 13, fig. 77, chart 92), Colchicum (plate 51, fig. 301, 

 chart 206), and Crocus (plate 75, fig. 441, chart 289), or as in Canna (plate 83, fig. 

 497, chart 338) and Solanum (plate 100, fig. 595, chart 389), the reaction curves 

 are so different that the starches could not be confounded, although histologically 

 it might be difficult to differentiate one positively from the other. 



Moreover, when grains of the same histological type are found in different 

 genera which at present are assigned by the systematic botanist to a given family, 

 as Tigridia (plate 68, fig. 405, chart 306) and Gladiolus (plate 68, fig. 407, chart 

 307), both of which are classed among the Iridaceoe, the reaction-curves may 

 exhibit such marked differences as to indicate misclassification on the part of the 

 one, if there is a correspondence of one with the family prototype, which we assume 

 to be that of Iris. 



Then again, when starches of different genera of a family, as constituted upon 

 the data of the systematic botanist, exhibit different histological types, as in Aroi- 

 dece, where the Dieffenbachia type is entirely different from that common to Arum, 

 Arisccma, Dracunculus, and Richardia, corresponding differences may not be found 

 in the reaction types, there being merely a modification of a family prototype, so 



