edges were found to be thin, so that it had to be rejected for the subsequent experiments.

The result marked † is also a little doubtful. After washing, the surface of the zinc was found to be darker and duller than usual. It strongly suggested surface oxidation, but we could not test the point. We might have been justified in leaving these two results out, but we prefer to show the extreme differences obtained by the method.

The general regularity of the other results leads us to think there was little or no disturbance from the two sources of error named. Experiment showed that the second was ordinarily inoperative. We tried several times to see if repeated washing and drying had any influence on the weight of the zinc, but got only minute changes which would not affect the ratio.

It may be useful here to say that the first five experiments (involving 10 ratios) were made with special redistilled zinc obtained from Hopkin and Williams. The last experiment giving the two ratios 3.2982 and 3.2961 was made with two sheets cut from a piece of common zinc, bought at the nearest zinc-workers. It will be noticed that the two ratios with this zinc agree very well, although the current density is twice as great in one case as in the other.

The mean ratio of the equivalents of silver and zinc, as given by our experiments, is 3.298 ± 0.0008. If we take 107.93 as the atomic weight of silver, and the atomicity of zinc as 2, the atomic weight of zinc is

If, as seems preferable, silver be taken as 107.66, then zinc is

We give below the comparison between zinc and copper. The density of current for the copper receiving plate varied comparatively little from 1 ampère per 50 square centimetres of receiving surface, the lowest density being 1 ampère for 70 square centimetres. It is known that the solution of copper sulphate acts on copper, but when slightly acid, and with these densities of current, the solvent action is small. Still the ratio obtained is not so reliable as that given by the preceding table for silver. As a matter of fact, the copper voltameter had only been used in the early experiments to give a ready check on the others. The ratios between silver and copper, which vary between 3.400 and 3.408, indicated a very fairly good agreement with the already known value, and afforded strong presumption that the ratio Zn: Cu would be sufficiently useful as a confirmation of the other results.

The mean ratio for Zn: Cu is, therefore, 1.0322. As the solvent action of the solution, already alluded to, is likely to give figures for copper a trifle lower than they ought to be, the ratio will be affected in an inverse sense-that is, the figure 1·0322 is very probably a trifle too high.

The values given for the atomic weight of copper vary somewhat. Three determinations made in recent years agree fairly well in giving a higher value than that formerly accepted. W. N. Shaw, as the result of many electrolytic experiments (B. A. Report, 1886) and a careful discussion, suggests 63:33.

T. W. Richards (Amer. Chem. J., 10, 187–191) has used a simple and apparently accurate method, namely, the precipitation of silver from a solution of its nitrate by known weights of copper. The mean result was 63:43.

Baubigny has given a value very nearly the same as Shaw's, and, as the latter was obtained by a method like that followed by us, we use it for our calculation. From it we get

63.33 × 1.0322 = 65.37,

as the atomic weight of zinc.

But this number (65·37) is obtained by means of a ratio which we have already seen to be too high, probably by something like one part in a thousand.

Reviewing the whole of the results, and attaching greater weight to the value derived from silver, the atomic weight of zinc would appear to be very close to 65.3. This is slightly higher than the figures given by Marignac, Baubigny, Morse and Burton, but somewhat lower than the value given by Reynolds and Ramsay.

449

XLIV.-The Amylodextrin of W. Nägeli, and its relation to Soluble Starch.

By HORACE T. BROWN, F.R.S., and G. HARRIS MORRIS, Ph.D., F.I.C.

In the year 1874, W. Nägeli described, under the name of amylodextrin (Beiträge zur Kentnniss der Stärkegruppe, Leipzig), a substance which he obtained by the long-continued action of dilute mineral acids upon ungelatinised starch in the cold. The starch-granules, whilst retaining for some considerable time their original structure, were, in the course of several weeks, completely disintegrated, a portion of their substance going into solution in the acid, whilst the iodine reaction of the insoluble residue gradually changed from blue through violet to reddish-yellow. The residue consisted of crude amylodextrin, which was purified by solution in hot water and subsequent precipitation, either by freezing the solution, or by the addition of alcohol.

Nägeli describes amylodextrin as separating from its solutions in the form of crystalline spherules which are made up of minute needles arranged radially. The drawings which he gives of these crystalline aggregates suggest a close resemblance to the well-known spherules of inulin.

Amylodextrin is said to have a rotatory power of [×] = 175° to 177° when observed in the Wild's instrument with ordinary light; and it is also stated to be non-diffusible.

Of late years, the few chemists who have noticed amylodextrin have taken it for granted that it is identical with soluble starch, an opinion for which Musculus and Gruber (Bull. Soc. Chim., 30, 54, 1878) and Arthur Meyer (Botan. Zeitung, 1886, Nos. 41 and 42) are mainly responsible; whilst Tollens also adopts this view in his recent Handbuch der Kohlenhydrate.

A careful investigation of this little-known substance has, however, convinced us that it is a perfectly well-defined compound and entirely different in its nature from soluble starch. We are now in a position to define its properties more accurately, and to assign it a place in the series of starch-transformation products.

In describing his method for the determination of diastatic power, C. Lintner has stated (J. pr. Chem., 34, 378, 1886) that a comparatively short treatment of ungelatinised potato-starch with hydrochloric acid of 7.5 per cent. renders the granules, when subsequently freed from acid, capable of completely dissolving in hot water without the production of the usual viscid paste. Our experiments fully confirm

the accuracy of this statement. In one case in which the abovementioned strength of acid had been used, the starch had lost all power of gelatinisation within 10 days; whilst in another experiment, in which acid of 12 per cent. had been used, this point was reached in less than 24 hours.

This extraordinary change in the properties of the granules is not accompanied by the slightest change of structure, and the altered granules have exactly the same influence on polarised light as the unaltered starch.

We have very carefully examined the nature of the substance which goes into solution on treatment of the altered starch with hot water, and find that it is in every way identical with the soluble starch described by O'Sullivan and others, and prepared by the limited action of heated malt-extract or of acid upon starch-paste heated to a suitable temperature.

Soluble starch prepared by any of these methods slowly separates out on long standing of its concentrated solutions, or immediately on the addition of alcohol to dilute solutions, as a white substance which in mass is somewhat pasty in its nature, and on washing with alcohol and subsequent drying becomes very friable. No matter how slowly it may be precipitated from its solutions, it is always found under the microscope to be made up of minute particles entirely without structure and without action on polarised light. The substance itself and also its solutions are coloured an intense blue by iodine. Although almost insoluble in cold water, it is readily soluble in water at 60-70°, and is thrown down again on cooling as a white, flocculent, amorphous precipitate.

The specific rotatory power of soluble starch in solution is [a]3-96 2160°, and it has no cupric reducing power.

=

We have shown in our previous papers that soluble starch, which is the first product of the action of diastase on starch-paste, is capable, at temperatures up to 60°, of rapid hydrolysis down to a certain definite point, beyond which progress is relatively very slow. This point of equilibrium is reached before the complete transformation of the starch into maltose, the reaction becoming almost stationary when the mixed products give the following numbers :

and to the equation—

5(C12H20010) + 4nH2O 4nC12H22O11+ (C12H20O10)

Starch.

Maltose.

Dextrin.

In the above-mentioned properties, the soluble starch prepared by the limited action of diastase or acids upon starch-paste, agrees in every particular with the soluble starch produced by the action of dilute acid upon starch-granules in the cold. This will be seen more clearly when we describe the gradual changes which take place in the granules during the last-mentioned process.

Before considering the chemical changes which occur during the continued digestion of starch with dilute acid, it will be well to point out the successive changes in appearance which the granules exhibit under the microscope.

When we employ 12 per cent. hydrochloric acid, it is not until about the twentieth day that any visible action is apparent, the first signs of disintegration of the granules being afforded by their splitting in a direction parallel with their shortest perimeter. Subsequently radial clefts are formed, and at the end of about two months the granules have become more or less disintegrated, especially in the direction of their stratification, this disintegration being complete in from three to four months, when the residual substance retains little or nothing of the origina! form of the starch-grain.

Concurrently with the changes mentioned above, the iodine reaction of the residual starch undergoes considerable modification, even before any change can be detected in the structure of the granule. The original pure blue colour given by iodine gradually changes through purple, reddish-purple and reddish-brown, to a pale yellowish-red, which is not further modified even if the starch remain in contact with the acid for some years.

The residual substance, after prolonged action of acid, is the crude amylodextrin of W. Nägeli and A. Meyer. This, as will be seen from our subsequent experiments, has been derived from the soluble starch first formed, by a gradual process of hydrolysis.

The time required to reach the final result has been stated by other observers to be about 100 days, but our own experiments point to this period as being insufficient for obtaining the maximum effect.

The final product which we will first describe had been in contact with the dilute acid for eight years and a half, the insoluble residue representing about 60 per cent. of the potato-starch originally taken. The residue under the microscope appeared to be made up of the disintegrated lamella of the original starch-granules, and was found to be coloured a light yellowish-red by iodine, and to be almost completely soluble in hot water. It was purified from the small quantity