maltodextrin-a of Ling and Baker, 18 the achroodextrin II of Lintner, 19 and the maltodextrin of Brown and Morris. 20. This protodextrin was further hydrolyzed by malt extract; and, by a similar process, another dextrin was isolated from the hydrolytic products. This dextrin was termed y-maltodextrin. By analysis and depression-of-the-freezing-point-measurements, its elementary constitutional formula was determined to be C24H42021. Synkiewski claims that this dextrin is identical with the maltodextrin-ẞ of Ling and Baker, and the achroodextrin of Prior. 21 Its formation from protodextrin II is represented by the equation

C36H62O31+ H2O → C24H42O21 + C12H22O11

Hydrolysis of this maltodextrin yields the isomaltose of Lintner. Protodextrin I was prepared and isolated in a manner similar to that for protodextrin II, with the exception that the malt extract acted in the cold. The molecular formula was found to be C72H124062 or just twice that of protodextrin II. Hydrolysis of this substance by malt yielded a sugar with the same formula as maltose, but, because of its higher optical rotation and lower reducing power than maltose, it was believed to be a polymer of the latter sugar and received the name "dextrinose."

The next part of this author's investigations consisted in the study of soluble starch (prepared by heating starch paste in an autoclave under pressure) or, as he called it, amylodextrin. He determined its formula to be (C54H90O44)n + H2O; and, by means of acetylation, found it to contain thirty hydroxyl groups. Years ago Schützenberger and Naudin22 analyzed an acetyl derivative of ordinary starch, with results, according to Synkiewski, that demonstrated the presence in ordinary starch of twenty-seven hydroxyl groups. Since the formation of maltose from this thirty-hydroxyl amylodextrin is complete, and since the hydrolysis of ordinary starch stops before all of it is converted to maltose, he concludes that the amylogen residues of the starch molecule (each of which contain three maltose residues), under the influence of malt extract, first split off these maltose residues as maltose molecules, provided they have been previously provided with hydroxyl groups. For each hydroxyl that the amylogen nucleus takes on, one maltose residue is ready to be split off by the diastase.

Upon the configuration of the amylogen residues rests the structure of the starch molecule. Synkiewski holds that the amylogen residue contains three different kinds of carbonyl linkings: the one which is readily hydrolyzed by malt, and yields maltose and protodextrin I, is called an a bond; the second, which is broken only by long action of malt diastase and finally yields glucose from the protodextrin, is called β; and the third, which connects the glucose residues in maltose, is called y. According to this scheme the amylogen nucleus may be represented by

(C6) -а- (C6) -- (C6)

β

(C6) -а- (C6) -- (C6)

β

(C6) -а- (C6) -- (C6)

Since starch consists of n amylogen residues connected by anhydride carbinol linkings, then, when an a-carbonyl hydrolysis takes place, 3n molecules of maltose are produced and n molecules of the protodextrin I. A β-carbonyl hydrolysis splits the starch so that each amylogen complex is divided into three similar portions, which can be termed protodextrin-residues II. These residues each consist of three glucose molecules. Since protodextrin II contains six glucose molecules, it must consist of two protodextrinresidues II, and its constitution can be schematically arranged as follows:

(C6) -- (C) —— (C) — (C) — — (C6)——(C6) From this formula it is apparent that the molecule of this substance contains two intact y bonds and two maltose groups connected to a C6-Ce radical by means of a linkings. Since a linkings are easily attacked it can readily be seen why protodextrin II is so easily converted to sugar. This saccharification can be represented in two stages by the equations

1. (С6) -у- (C) —— (C6) - (C) —— (C6) —— (C6) Protodextrin II

→ (C6) -- (C) —— (C6) - (C6) + (C6)——(C6)

Synkiewski points out that C54, the number of carbon atoms in the amylogen group, is not a multiple of C36, the number of carbon atoms in protodextrin II, hence the molecule of protodextrin II cannot be formed from a single amylogen group. Both amylogen, and the last named dextrin, have in common the C1s group, so that C54=3(C18) and C36 = 2 (C18). Since, however, these substances come from the starch molecule quantitatively he feels that there can be no doubt that the 2 (C18) residues of protodextrin II originate from two amylogen groups according to the expression,

2[3(C18)]=3[2(C18)]

or, since the starch molecule consists of n amylogens, n[3(C18)]=[2(C18)]

From the fact that the protodextrin-II molecule is made up of two different amylogens bound together by only a carbinol linking, it follows that the linking which joins the two C18 groups of protodextrin II is a carbinol bond (the bond in the dextrinose molecule).

The amylogen nucleus consists of nine glucose groups, containing nine carbonyl radicals; but, since amylogen does not reduce Fehling solution, all the carbonyl groups must take part in uniting these glucose groups.

Synkiewski thinks that there are two possible ways in which these glucose residues may be joined together: between the nine glucose molecules there are eight carbonyl bonds, which causes the assumption of one di-carbonyl linking; or, there are nine monocarbonyl bonds, which necessitates the assumption of a ring

structure.

Let us consider the first possibility, for which a di-carbonyl group is assumed. It is evident that none of the y bonds is a carbonyl (there is no di-carbonyl bond in maltose). It is also easy to see that none of the ẞ bonds in the dextrin residue is di-carbonyl. Free protodextrin I reduces Fehling solution, which indicates that it contains a free carbonyl radical. At the two linkings between the three glucose molecules of this dextrin, only two carbonyl radicals are formed upon hydrolysis: these are mono-carbonyl.

If a di-carbonyl linking were present in the amylogen, it could be only one of the a bonds. By use of the sign (<) to denote a carbonyl bond, and putting di-carbonyl bonds in the place of a linkings, the formula of amylogen may be written as follows:

(C6) > (C12)

(C) > (C12)

(Co) >< (C12)

Under this scheme, hydrolysis would give no maltose; but, on the other hand, two different kinds of dextrin, one possessing no reducing power. This is not the case and, therefore, there is no di-carbonyl bond in the amylogen nucleus.

This conclusion naturally suggests the other possibility. The nature of the hydrolytic products of amylogen that have been studied would make it difficult to imagine an a or a y linking in a ring structure. Such a structure, Synkiewski claims, may be easily made up of ẞ bonds, and there are two possibilities :

As shown before, the rational formula for the amylogen group is some multiple of

C18H27012 03. (C12H23011)3

and by making use of Fischer's formula for maltose, Synkiewski writes the graphic formula for amylogen as follows:

HO

CH(CHOH)CH2-O-CH(CHOH)2CH-CHOH-CH2OH

-CH(CHOH)-CH

CH(CHOH)CH2-O-CH(CHOH)2CH-CHOH-CH2OH

0—

OCH(CHOH)CH2-O-CH(CHOH)2CH-CHOH-CH2OH

Synkiewski suggested that this formula represents a half acetal of the ten-hydroxyl alcohol of maltose, which is analogous to the alcoholates of chloral; and that it might be named "tri-malto tri-glucosate," if the not impossible but as yet undiscovered triglucose sugar were known to exist.

In order to combine all facts, including a ring combination of the three glucose molecules, Synkiewski shows that the formula for the amylogen nucleus could be written in accord with the arrangement shown in Fig. 3, on p. 392.