was estimated by combustion with soda lime, and the result, multiplied by 6.25, was taken for the amount of albuminoids* Here follow the results:

Determination of Cellulose.-The cellulose was estimated by alternate extraction with diluted acid and alkali (Peligot's meth od), after Henneberg's directions, (Versuchs-stationen, vi, 497). As is well known, the methods for cellulose estimation are not entirely accurate, and the results here given are probably somewhat too low.

Another trial, no doubt less accurate, from the fact that the material was allowed to stand for some time in the liquid previous to each decantation, gave

And a trial upon the material left from an unsatisfactory attempt to estimate starch, by Dragendorff's method, described in Johnson's "How Crops Grow," p. 66, gave after extracting starch with dilute chlorhydric acid—

Determination of Aqueous Extract, Gum and Sugar.-Four grams of finely ground material were treated with cold water in a beaker for two hours, and then filtered with the aid of the Bunsen pump. The whole operation required three or four hours for completion. In the case of the sweet corn it was found necessary to make the solution of known bulk, and take aliquot parts of the first portion of the filtrate for the estimations, as the filter became, after a time, so clogged as to forbid the passage of the solution, even when under pressure. filtrate was divided into four equal portions, each representing

The

* The mean factor 6-25 is that adopted by the Versuchs-stationen of Germany. The recent investigations of Ritthausen, Jour. f. Prakt. Chem., cvi, p. 483, give in maize-fibrin 15.58 per cent of nitrogen. Stepf, (same Journal, 76, 90), found 156 per cent of nitrogen. The other albuminoid in maize is very near Ritthausen's Conglutin, and contains 17.72 per cent of nitrogen, but as it forms but 0.5 per cent of the grain, (loc. cit.) we must conclude that 64(100-15.6) is the proper factor for calculating the albuminoids. This gives the following average percentages:

ɔne gram of meal. The first portion was evaporated to dryness in a platinum capsule, weighed, ignited, weighed again, and the difference reckoned as aqueous extract, ash-free. From the last weight was subtracted the weight of the capsule and the difference was reckoned as ash. The results are as follows:

When examined qualitatively, the extract gave with iodine a blue tint, indicative of the presence of a small amount of starch, and when ignited in a glass tube with soda-lime, yielded a slight reaction for ammonia, showing that there was a small amount of albuminoid matters likewise present.

In every instance the solution was somewhat turbid, from presence of a trace of suspended fat (?) This was especially true in case of the sweet corn. On boiling, a white substance, probably a coagula ed albuminoid, separated from the solution. The quantity was however very small, and its estimation was not attempted. Stepf gives the amount of albumin in the maize he examined at 0.62 per cent.

The researches of von Bibra* indicate the absence of dextrin from the cereal grains, a conclusion which is quite in accordance with my own qualitative examinations. I have therefore considered the aqueous extract to consist of sugar and gum exclusively, so far as its non-nitrogenous ingredients are concerned. The second portion of the aqueous solution was also evaporated to dryness at near 100° C., treated with alcohol of 66 per per cent, filtered, the filtrate evaporated to dryness, and the residue treated as before. The following results were obtained:

.0.27

0.00

2.32

4.75

Undissolved, It is probable that the alcohol of 66 per cent carried gum into solution. In B the albuminoid was also dissolved.

More confidence is to be placed in the results obtained with the third portion of the aqueous extract, by the use of Fehling's solution, as described in Fresenius' Quantitative Analysis. The precipitated suboxyd of copper was collected on weighed filters, and the sugar was reckoned as glucose. I obtained:

Sugar,

Sugar and Gum, per cent.
A. II.

B. II.

Gum, by difference, ---4.22

4.64

Want of time cut short my work at this point, and I regard these determinations of sugar and gum as open to revision.

*Die Getreidearten und das Brod, passim.

The complete (mean) analyses stand as follows, taking the sugar and gum as estimated with Fehling's solution and reckoning starch by difference:

Subjoined are the mean percentages of the proximate elements calculated upon dry substance:

In these analyses the percentage of water and ash alone can be regarded as quite correct. The percentage of albuminoids must be considered as a close approximation to the truth. The figures given for "sugar" and "gum" (aqueous extract) are likewise probably not far from right. They include, however, a little of some soluble albuminoid and possibly soluble starch -the "amiduline" of F. Schulze and "amylogen" of Jessen. The cellulose is much lower than most chemists have found. Thus Poggiale, Payen and Polsen obtained respectively 4:5, 9 and 14.9 to 204 per cent of fiber. That these figures are erroneous must be concluded from simple ocular investigation. Fresenius, using Peligot's method obtained 1:58 per cent. Bibra rejects Peligot's method as yielding too little cellulose, but the numerous analyses executed by the German Agricultural Chemists within recent years indicate that its results do not vary widely from the truth. The error in the percentage of starch (insoluble), which is determined by difference, cannot probably exceed 2 or 3 per cent.

The only detailed investigations upon Indian corn by Amer ican chemists with which I am acquainted, are those of J. H

Salisbury, Trans. N. Y. State Ag. Soc., 1848, p. 678, and those of C. T. Jackson, U. S. Patent Office Report, 1857, p. 160, and Geology of New Hampshire, pp. 256 et seq. These investigations, the former of which, especially, was quite extended, were unfortunately made for the most part by methods too imperfect to yield valuable results. I append an analysis by Salisbury, above report, p. 779, of "Yellow Corn," the same variety as B of the present article, also one of "Rhode Island Sweet Corn, same report, p. 784.

Yellow Corn.
11.18

R. I. Sweet Corn.

10.22

Sugar,..

Dextrin or gum,

Starch,

The following is an analysis by Jackson, Patent Office Report,

1857, p. 161, of King Philip Corn.

As seen above, Dr. Salisbury found in Sweet Corn 12:32 per cent of dextrin or gum and 13.00 per cent of albumen, and in the article referred to attributes the shrivelling of the kernels to the loss of water in the drying of these bodies. It cannot be doubted that the estimations were erroneous.

In "The Geology of New Hampshire," page 259, Dr. Jackson reports the amount of oil in different varieties of corn as varying from 6 to 11 per cent, the latter amount being found in the Canada Corn (B of this paper). In the same article, p. 257, he states that the oil in this variety as in several others, is con

fined chiefly to the endosperm, whereas Lenz and myself find it nearly all in the chit. On page 258 Dr. Jackson "is able to prove that the popping of corn" is due entirely to the decomposition of the oil and the formation of carburetted hydrogen gas. But it is a matter of experience that sweet corn which is remarkable for its large content of oil, expands only very slightly in "popping." On page 257, Dr. Jackson mentions "a beautiful and new application of chemistry" by which the use of qualitative tests, "iodine," "sulphate of ammonia," &c., “we may easily cause any grain to point out the extent and precise limits of each of its ingredients and by the eye we can form a pretty correct estimate of their relative proportions in different seeds." It would be ungracious to notice these statements, were it not that they are frequently quoted as a part of our standard knowledge on this subject.

It will be seen from my analyses that the three first varieties of corn which are quite similar in general appearance and in the consistency and structure of the kernel, are almost identical in composition, while the sweet corn is much poorer in starch and richer in sugar and gum as well as in fat. I hope to have an opportunity in the future of extending my investigations in this direction so as to accomplish more perfect separations and also to examine a large number of varieties of American maize.

ART. XXXIV.-On the Magnetite in the mica of Pennsbury, Pa., in reply to Prof. G. Rose; by JAMES D. DANA and GEORGE J. BRUSH.

1. Note by J. D. Dana.

PROFESSOR ROSE, of Berlin, in a paper published in the Proceedings of the Royal Academy at Berlin (Monatsberichte für 1869, p. 339), has pronounced the statement in my Mineralogy (p. 150), that the markings in the Pennsbury mica are due to thin films of magnetite, an error, and has instead called the mineral hematite. The first reference of the species to magnetite was made by Prof. Brush, and I therefore leave it to him to mention the observations upon which the conclusion was based. They afford full proof, I feel assured, that the mineral is not hematite, and make it scarcely less certain that it is magnetite.

Professor Rose's arguments are mostly opinions rather than facts or demonstrations. He infers from his having seen very minute grains of magnetite, even those obtained in blowpipe reductions, which were not transparent, that the excessively thin transparent crystalline films in the mica are not magnetite.