44 THE ANALYST. THE MICROSCOPIC EXAMINATION O F AXERZCAN COTTONSEED CAKE. BY A. L. WINTON. IN a recent issue of this journal, Dr. Voelcker has described a number of methods, partly chemical, partly physical, and partly mechanical, for discriminating between Egyptian and Bombay cottonseed cake. This paper was of special interest to me, its I have attempted to solve a similar problem-namely, the discrimination of upland from Sea Island cottonseed cake as produced in the Southern States of my'own country. Although my work has been largely microscopic, the conclusions reached bear a close relation to those of Dr. Voelcker. The upland or short-staple cotton growing in the United States is commonly c1a.ssed as Gossypium herbaceurn, although quite possibly some of the varieties were obtained by crossing with other species.It yields the great bulk of the American cotton fibre. After ginning, the seed is still clothed, like Bombay seed, with a dense mat of ground-fibre, which can be removed only with great difficulty. This fibre, however, is of no serious detriment to the cake, plrovided, as has been almost universally the custom, the seed is decorticated before expressing the oil. In this connection it may be interesting to note that the cotton hulls thusTHE ANALYST. 45 separated were long used as fuel under the boilers of the oil mills, and the ashes regarded as waste. Almost twenty years ago, however, the Connecticut Agricultural Experiment Station pointed out that these ashes contained something like 25 per cent. of potash (&O), largely in the form of carbonate, and that they were admirably adapted for tobacco fertilizers.A lively traffic in these ashes at prices ranging from 25 to 50 dollam per ton was at once created, and continued until recently, when the recognitioq of the value of cotton hulls for feeding led to the abandonment of the former wasteful practice. Upland cottonseed cake has been placed on the home market in the form of cottonseed meal, a rich yellow product of high value, both as a cattle food and fertilizer. I t has usually been guaranteed to contain 7 per cent. of nitrogen (equiva- lent to 43.75 per cent. of protein), and often has contained 7.50 or even 8 per cent. Of late, however, meal containing a large amount of hulls and much less nitrogen (as little as 4 per cent.), has appeared on the market, and in some instances has been fraudulently substituted for prime meal.While it is a well-recognised fact that cotton hulls can be fed to advantage in conjunction with the meal, and that undecorticated meal is preferred by English feeders, it must be admitted that meal containing or adulterated with a considerable amount of hulls should command a lower price than the concentrated product. This is especially true when, as is often the case, the meal is purchased solely as a source of nitrogen for fertilizing purposes. Frequently manufacturers and jobbers in these inferior and adulterated products have chimed that the product was Sea Island meal, and the hulls were normal constituents. Sea, Island cotton, G. barbadense, is grown on the coast of South Carolina and Georgia, and yields the highly-prized long-staple cotton. The seed is dead black, without ground-fibre, and for mechanical reasons is not decorticated.It is stated that a large share of this Sea Island cake finds a, market in England. From Dr. Voelcker's description, I infer that the Egyptian cake mentioned by him was from this or a related species. Although Sea Island cake, because of the absence of ground-fibre, is superior to undecorticated upland cake, it is obviously inferior to prime decorticated cake from the latter variety. I t was partly with the hope of finding some means of refuting the claims that low grade meals were Sea Island products that I made a comparative study of the microscopic structure of the two seeds.The most striking elements are the thick-walled epidermal cells with dark con- tents, the colourless cells, the palisade cells, each with a globular cavity one-third the distance from the outer to the inner end, and the fringed cells of the perisperm. The outer brown layers and the inner testa also, with brown contents, are not so clearly defined, but of great importance in diagnosis. In the Sea, Island I found that the outer epidermis was not only free from hairs, but the epidermal cells proper were narrower and higher than in the upland seed; also that the colourless cells of the third layer were thicker, and more often in two rows or divided by tangential walls, and that the cells of both the outer brown layer and the inner testa were thicker, more strongly developed (or, rather, less completely My first work was with single samples of both varieties of seed.46 THE ANALYST.obliterated), and contained mwe abundant brown contents. Not being content to accept these results as final, I procured from a seedsman and others in the South samples of eight leading varieties of upland cotton and three of Sea Island. On examining these, it appeared that all of the distinctions named could not be depended on in diagnosis; but those based on the presence of ground-wool in upland and its absence in the Sea Island varieties, and on the difference in the thickness and amount of coloured substa,nce in the outer brown layer and the inner testa, are of considerable value. These are much the same distinctions as Dr. Voelcker has brought to notice in his paper.As for the ground-fibre, while all the seeds in upland cotton are clothed with a, dense wool after passing through the gin and linter, some are largely deprived of this coat in grinding, so that the presence or absence of hairs on the ground husk is not an absolutely sure criterion as to the origin of the seeds. On the other hand, all Sea Island seeds are not entirely denuded of fibres, and the meal made from them usually contains a little fibre, although never so much as an' upland meal ground with the hulls. The means of distinction based on the amount of brown colouring matter in the two layers appeared to be most valuable, but was not found infallible. While in all the Sea Island cotton-seeds I have examined both brown layers contained a pro- nounced amount of brown matter, some upland varieties exhibit the same charac- teristics, although in a lesser degree.As a rule, however, the distinction was well msrked, and should prove of no little value when, as in the cake on the English market, the hulls are in large fragments. But the meal on the American market-noti only the prime article, but especially that containing large amount of hulls-is sold in a finely-ground condition. The manufacturers appear to have learned that by fine grinding they secure a brighter yellow appearance, not only because of the fine division of the particles-although this unquestionably plays an important part-but because the light-coloured sides of the palisade cells are brought into view, whereas in the commly-ground hulls only the brown outer and inner coats are exposed.This fine pulverization both deceives the purchaser and renders the microscopic test described by Dr. Voelcker valueless ; it also embarrasses the microscopist, as he is unable to find fragments large enough for cutting sections. The chemical test found efficient by Dr. Voelcker does not aid us, as the hulls cannot be separated from the meal. Furthermore, I find that our Sea Island cotton hulls give a more distinct colour by this test than the upland. It thus appears that in many cases the problem in question is more difficult than that encountered by the English analyst, and cannot be solved with the mme degree of certainty. So far as concerns the mere detection of excessive amounts of hulls or starchy adulterante, mimoscopic examination, especially if coupled with determinations of nitrogen and fibre, is all that could be desired. The following results are interesting as showing the extremes in chemical composition :THE ANALYST. 47 Prime Cotton- seed Per Cent. 8.98 6.18 48.19 5.54 22.46 8.65 ------ 100.00 / Water . . . ... ... . . I Ash ... ... Fibre ... ... ... ... Fat ... ... ... ... Protein (N-x 6.25). . ... ... Nitrogen-free extract . . . . . I Cottm-seed Meal Adulterated with Cotton Hulls. Per Cent. 10-48 4-70 25-50 18-95 34.27 6.10 100*00 Total Cotton Hulls, it Average of Four Analyses. Per Cent. 10.41 2.59 4.04 44-42 36.52 2.02 100*00 CONNECTICUT AGRICULTURAL EXPERIMENT STATIOK, NEW HAVEN, CONK., U.S.A.