rate (compare Allen, Abstr., 1887, 1145). Crampton substitutes phosphoric for sulphuric acid in the distillation, but finds that with care the two give identical results.

Scheffer's test for foreign fats is based on the solvent action of a mixture of amyl alcohol (40 vols.) and ether of sp. gr. 0·715 (60 vols.). 1 gram of butter dissolves in 3 c.c. of this mixture at 28°; 1 gram of lard requires 16 c.c.; 1 gram of stearin 350 c.c.

The following order of value is assigned to the various modes of examining butters for adulteration -(1.) Determination of volatile acids. (2.) Determination of specific gravity. (3.) Determination of saponification equivalent (Koettstorfer). (4.) Determination of the insoluble acids (Hehner, Muter, Blyth, &c.). (5.) Determination of the melting point.

A complete bibliography of butter analysis (up to 1882) is given by Caldwell (Second Ann. Rep. N. Y. S. Bd. of Health, 544-547), and in Sell's Kunstbutter (Arbeit a. d. Kaiserl. Gesundheitsamte).

Milk. For the determination of water, Babcock employs asbestos to absorb the milk before drying; this is much to be preferred to any powder. By placing the asbestos in a tube between plugs of cottonwool, and drawing air through the tube while it is heated at 100°, the desiccation is complete in two hours. The tube can then be transferred to the fat extractor.

Of methods for the determination of the fat, the preference is given to that of Adams (Abstr., 1886, 583), but instead of soaking up the milk with one end of the roll of paper, the plan has been adopted of holding the unrolled strip in a horizontal position, and running the 5 c.c. of milk from a pipette along the middle. The strip is then hung up in a hot chamber, and in two or three minutes is dry and ready for rolling up.

In Soxhlet's aräometric method (Abstr., 1881, 656) great difficulty was encountered in ensuring the separation of the ethereal fat solution, and any long delay in the separation was found to affect the final results. By placing the bottles containing the mixture in à centrifugal machine revolving about 300 times per minute, the time required for separation was reduced to a few minutes, only six samples out of 150 requiring more than quarter of an hour. The fat solution thus separated had, however, a lower specific gravity than that obtained by simple subsidence, so that the percentage of fat as given in Soxhlet's table had to be increased by 0.13, to bring it into agreement with the results of the older method. Cronander also separates the fat by shaking the milk with potash and ether, but evaporates the ether after it has risen to the surface of the milk, and measures the fat by forcing it in a melted state into a graduated tube. Fleischman and Morgen calculate the fat by the formula

where f

percentage of fat, t = percentage of total solids, S = specific gravity of the milk at 15°.

Morse and Piggot add 10 c.c. of milk to 20 grams of dehydrated copper sulphate. The milk becomes dry in a few moments. The fat

is then extracted by light petroleum, and its amount determined (after evaporation) by saponification.

Of lactoscopes, Feser's is said to be the most convenient. It is a graduated glass cylinder, containing in its lower part a smaller cylinder of white glass with black lines on it. 4 c.c. of milk are put in the cylinder, and water is added until the black lines become visible. The reading of the total volume gives at once the percentage of fat.

For the determination of the free acid in koumiss, a bright filtrate was obtained by adding to the koumiss an equal volume of alcohol before filtering.

M. J. S.

Extraction of Fats by Soxhlet's Apparatus. By J. M. MILNE (J. Soc. Chem. Ind., 6, 34).—In using the apparatus for milks, the author proceeds as follows:-About 10 c.c. of the milk is weighed into a tared porcelain basin, and the milk evaporated with frequent stirring in order to render it granular, until on being cooled the residue is semi-solid. The residue is then transferred to a paper cup and placed into the Soxhlet tube, and the fat extracted with ether in the usual way. The author having worked with Adams' paper coils for drying up milk for fat extractions, confirms the fact pointed out by the committee of the Society of Public Analysts, that from 0.3 to 0.5 per cent. more fat is extracted by the coil method. D. B.

Examination of Wines and Oils. By P. SPICA (Gazzetta, 17, 304-312). The author at the outset makes the oft-repeated complaint that the conditions, such as the variation of concentration, or even the nature of the acids used, required for the successful application of test-reactions, are not defined with sufficient exactness in original papers. Various processes have been proposed to recognise the colouring matters, whether natural or artificial, of wines; but preference is given by the author to the methods proposed by Cazeneuve (Abstr., 1886, 397), Arata (Gazzetta, 17, 44), Blarez and Deniges (Abstr., 1886, 1084), and Girard and Gautier. In the lastnamed process the substitution of tablets of Magnesia alba for those of plaster of Paris is suggested; these are immersed in egg albumin for a short time and dried. A drop of the wine to be examined is let fall on such a prepared tablet, and the colour of the stain produced is observed. The natural colouring matters of wine give a yellowishbrown, those containing rosaniline or "vinoline" a reddish-yellow, those with indigo an azure-green, with orchil a violet red, and those with amaranth a greyish-violet stain. Another method suggested consists in shaking up the wine with baryta-water and amyl alcohol, when the latter extracts the colouring matter from the wine. It appears that certain preparations containing coal-tar colouring matters, called "vinoline," are sold by druggists in Padua. An examination of such a preparation, called "maroon vinoline," was found to consist of about 40 per cent. of mineral matter, in which arsenic was present in considerable quantities.

As regards the method of examination of oils proposed by Maumené, which is founded on the rise of temperature when the

sample is mixed with concentrated sulphuric acid, it appears that a confusion has arisen between grams and cubic centimetres. Concordant results are obtained with mixtures of 50 grams or 55 c.c. of oil with 10 c.c. of acid.

Bechi has proposed a method for the recognition of cotton-seed oil in olive oil, which consists in adding to the oil an alcoholic-ethereal solution of silver nitrate in presence of an amyl alcohol solution of petroleum; it is here shown that this method gives fallacious results. A method of the greatest practical value is that proposed by Hubl, which has given satisfactory results in the hands of Moore, Allen, Oglialoro, and other observers. V. H. V.

con

Gravimetric Estimation of Tannins. By H. R. PROCTER (J. Soc. Chem. Ind., 6, 94-96). The process described by the author is a combination of the methods published by Müntz and Simand, and depends on the fact that in filtration through a column of dry hide powder the upper layers absorb most of the tannin, a very complete and rapid separation being obtained from the large surface exposed. The author utilises the lamp chimneys employed in the common round-wicked German petroleum lamps, which are tracted just above the base of the flame and are cylindrical for the remainder of their length. A perforated disc of cork is made slightly cup-shaped on its two faces. A piece of linen is then stretched over it, and it is pressed down the chimney until it rests on the contracted neck. Five grams of hide powder is weighed into the tube, and when shaken down will occupy a space of about 50 c.c. The tube is now cut off, allowing only length for the insertion of a cork, which may press slightly on the powder, as it contracts in volume when wet. This cork is perforated and hollowed like the first, and after being covered with linen is pressed into the tube. A short piece of quill tubing passes through the cork, and is fitted by a second cork into a flask. The filtering tube is inverted, broad end downwards, into a beaker of 100 c.c. capacity, which is filled with the liquid to be filtered until it rises into the hide powder. The tube is left in this position for one or two hours, after which it is reversed, and the enlarged end filled with the solution, when the filtration will be found to proceed evenly and steadily. The filtrates thus obtained are perfectly free from tannin, and tested by the Löwenthal method show a lower result for "non-tannin" than those by any other method of absorption. The method is, however, inapplicable in the presence of gallic acid, the latter being freely absorbed by hide powder. The author hopes to overcome the difficulty either by some method of removing the gallic acid or of preventing its absorption by the hide.

D. B.

97

General and Physical Chemistry.

Relation between the Composition and Absorption-spectra of Organic Dyes. By H. W. VOGEL (Ber. Akad. Ber., 1887, 715— 718).-Experiments made with azo-dyes gave the following results:The substitution of methyl for hydrogen in diazobenzene gives rise to change of position of the absorption-bands towards the red end of the spectrum. The increase of wave-length is 10 millionths mm. when the substitution takes place in the ortho-position, and 14 millionths mm. in the case of the para-position. (2.) The substitution of B-naphtholsulphonic acid S or B-naphtholdisulphonic acid R for B-naphtholsulphonic acid B, causes a shifting of the bands which, in the case of B-naphtholsulphonic acid S amounts to 4 to 5 millionths mm., and in the case of B-naphtholdisulphonic acid R to 6 millionths mm. (3.) In the substitution of methyl, the space between the two bands becomes clearer, and the bands become more equal in intensity and in width. The substitution of B-naphtholsulphonic acid S or B-naphtholdisulphonic acid R, in the place of the acid B, acts similarly on the character of the bands.

The above results were obtained with sulphuric acid solutions; the results, as well as those obtained with alcoholic solutions, are shown in tables. N. H. M.

Fluorescences with Well-defined Spectra. By L. DE BOISBAUDRAN (Compt. rend., 105, 784-788). When gallium oxide is employed as a solid solvent of other oxides, it gives fluorescences which are usually much less brilliant than those obtained with alumina, but the results are of the same order. Calcination of the oxide at a very high temperature converts the bands in the spectrum into lines, the spectral groups being displaced towards the red. The brilliancy of the spectrum increases with the time of action of the electric discharge, a result contrary to that obtained with alumina.

Measurements are given of the spectra of the fluorescences of gallium oxide with oxides of samarium, Za and Zp. The fluorescence of Zẞ with gallium oxide is very feeble, the difference between Za and ZB being even more strongly marked than when the oxides are mixed with alumina.

A moderately calcined mixture of alumina with praseodymium oxide yields only a very faint trace of a rosy fluorescence, but if the mixture is very strongly heated, it then yields a beautiful fluorescence, the colour of which depends on the time during which the electric discharge has been passing. At first it is violet, but afterwards becomes rose-coloured. The spectrum is complicated, the principal bands, all of which are nebulous, being situated at 6457, 6237, 6162, 6035, and 5212. C. H. B.

Rotatory Power of Solutions of Ammonium Molybdate and Tartaric Acid. By D. GERNEZ (Compt. rend., 105, 803-806).—

VOL. LIV.

h

The experiments previously made with solutions of tartaric acid and sodium molybdate (Abstr., 1887, 540) were repeated with ammonium molybdate. The rotatory power increases regularly as the quantity of ammonium molybdate increases, and is proportional to the quantity of this salt present up to a quarter of an equivalent. Between one-fourth and one-third of an equivalent, the increase of rotation for the same weight of molybdate changes suddenly, and becomes little more than half its original value. The maximum rotation observed is 57 times that of tartaric acid, and corresponds with a proportion of ammonium molybdate equal to one-third of an equivalent, it remains sensibly constant between 42.66/128 and 56/128 of an equivalent of molybdate. With large quantities, the rotatory power diminishes rapidly, and becomes practically constant when one equivalent of molybdate is present.

As in the case of sodium molybdate, the tartaric acid at first combines with the whole of the ammonium molybdate, forming a com pound of the composition 8C,HO, + 3(NH4)20,7M0O,,4H,Ŏ, which is gradually converted into a second compound, containing 6 mols. of tartaric acid and 1 mol. of the molybdate, and this is finally transformed into a third more stable compound,

2C.H.O. + 3(NH4)2, 0,7M0O3,4H2O,

6

In the case of sodium molybdate, the maximum rotation is obtained with one equivalent of the salt, whilst in the case of the ammonium molybdate, the maximum rotation is given by one-third of an equivalent. The ammonium salt, however, contains three equivalents of the alkali in the molecule for each equivalent present in the molecule of the sodium salt, and hence in each case the maximum rotation is given by compounds which contain the alkali and the tartaric acid in equal equivalents. C. H. B.

Influence of Light on the Heat Conductivity of Selenium. By M. BELLATI and S. LUSSANA (Gazzetta, 17, 391-405).-The analogies of heat and electric conductivity induced the authors to study the influence of light on the heat conductivity of selenium, the electric resistance of which, as is well known, is diminished on exposure. The plan of experiment consisted in sprinkling the double iodide of copper and mercury on the disc of selenium, on which a circular figure had been blackened with Indian ink. The selenium was heated by the passage of an electric current, which produced at first a dark spot, owing to the change in colour of the double iodide. This subsequently extended into a fairly regular circular figure, the measurement of the diameter of which afforded a means of determining the heat conductivity of the selenium. This method was found to be more practicable than the usual method of melting wax. In all cases, the diameter of the circle was greater when the selenium was exposed to reflected sunlight from which the greater part of the heat rays had been removed by passage through solutions of alum and of ammoniacal copper sulphate. The relation of heat conductivity without and with exposure to light was found to be in the ratio of 1:11 as the result of several concordant experiments. The same