INTRODUCTION. 7 



These groups were found by Cross and Bevan to "pass by imperceptible gradations 

 into a Iieterogeneous class of natural products which, while possessing some of the char- 

 acteristics of the celluloses proper, are so readily resolved by hydrolytic treatment that 

 they must represent a very different constitutional type or types. To this group of com- 

 plex carboliydrates the name hemicellulose has been assigned," which group is stated to 

 be readily resoh-etl into crystalline monoses. From the foregoing data it is obvious that 

 there are certain forms of cellulose that are not strictly speaking isomeric, and that there 

 are also forms that are isomerides. Cellulose is optically active, and therefore contains 

 one or more asynmietric carbon atoms, and, as a corollary, each isomer may have a number 

 of stereoisomeric forms, the number varying with the number of asymmetric carbon 

 atoms. Reasoning from this, starch may likewise exist as groups which are not strictly 

 speaking isomeric, yet in which the molecules differ so little in their molecular and struc- 

 tural formulae as to have the essential characteristics of a given prototj^je; secondly, 

 each group may be made up of a mmiber of stereoisomers; and thirdly, the starches of 

 nature, as observed in the starch-grains of a given plant, may, as it seems certain, be vari- 

 able mechanical mixtures of two or more different chemical forms. 



The number of possible starch stereoisomers is entirely problematical. Miescher has 

 estimated that the serum albumin molecule having 40 carbon atoms may have as many as 

 a thousand million stereoisomeric forms. If we assume that the molecular weight of starch 

 is as low as 15,000, and that the molecular formula is nCCoHjoOs)?!, the total number of 

 carbon atoms in the molecule is at least 550. What proportion are asymmetric is unknown, 

 but judging from the relatively high percentage in such compa^ati^ely simple substances 

 as the aldohexoses, and the striking tendency for protoplasm during the synthesis of 

 organic substances to form bodies with asymmetric carbon and asymmetric nitrogen 

 atoms, it is probable that nearly all are asymmetric. Moreover, if we conceive, as we 

 should upon the present basis of our knowledge, that the molecule is not a poljoneride of 

 preformed atomic groups, but an aggregate of labile groups of ionic units, the possible 

 number of stereoisomers is absolutely inconceivable. 



It has already been pointed out that a trifling transposition of elements, groups, or 

 masses attached to an as;ymmietric carbon atom may cause a marked change in crystalUne 

 form, and in optical, chemical, and physiological properties; and also that in stereoisomers 

 changes in the configuration of the molecule, however sUght, may give rise to greater dif- 

 ferences than may be shown by isomerides which have entirely different structural formulie. 

 This is a matter of the most profound fundamental importance in connection with proto- 

 plasmic processes, and in it we seem to have the key to unlocking many baffling problems 

 of physiology, toxicology, and pathology not to speak of those of general biology; nor is 

 it necessary to enter into speculation for illustrations of such applications, because many 

 instances in literature, dating from Pasteur's experunents, are at one's disposal. 



In earlier pages, experunents of Pastevu* were referred to which showed the marked 

 differences in the dextro-, laevo-, and racemic tartaric acids m relation to Penicillium 

 glaucum, the dextro form being consimied, the racemic form being split into the dextro 

 and laevo forms, and the dextro form being used, but the laevo form discarded and re- 

 maining in solution. It has since been shown tliat the dextro forms of glucose, mannose, 

 galactose, and fructose are fermentable, or, in other words, consumed by various kinds 

 of micro-organisms as foodstuffs; while the laevo forms are not at all or but inappreciably 

 affected. Again, when mandelic acid (racemic) is subjected to the action of Penicillium 

 glaucum the compound Ls split, the Itevo form is consumed, but not the dextro form; 

 whereas, in the presence of Saccharomyces elUpsoideus, the dextro form but not the laevo 

 form disappears. With glyceric acid treated with Penicilliimi and Bacillus elhaceticus, 

 respecti\'ely, the primary splitting process is the same as with mandelic acid, but the 

 Penicillium uses the dextro form and the Bacillus the laevo form. 



