Page images

turpentine, resin oils, and mixtures of these yield products soluble in carbon bisulphide or heavy petroleum, therefore these latter oils can be readily separated from drying oils by sulphur chloride. As excess of sulphur chloride produces a less solid product than when such excess is avoided, this must be taken into consideration when treating unknown oils. In dealing with mixtures of oils in which each individual oil gives a solid product, the results obtained are only approximate, and must be verified by treatment of a known mixture of the oils. The author then indicates the very extensive amount of adulteration and substitution of inferior for superior oils in commerce, which can be detected by means of sulphur chloride. D. A. L.

Estimation of Urea by Titration. By T. PFEIFFER (Zeit. Biol., 24, 336-350). This paper is a lengthy reply to the criticisms recently passed by Pflüger (this vol., p. 201) on Pfeiffer's modifica tion of Liebig's method of estimating urea. W. D. H.

Titration of Pyridine Bases. By K. E. SCHULZE (Ber., 20, 3391).-5 c.c. of pure pyridine (sp. gr. 0.98) was dissolved in 100 c.c. of water; 25 c.c. of the solution was treated with 1 c.c. of 5 per cent. aqueous ferric chloride solution; normal sulphuric acid solution was then carefully added until the precipitated ferric hy droxide redissolved. The amounts actually used were 155, 154, and 155 c.c.; 155 c.c. being the amount required by theory. -Picoline was titrated in a similar manner. N. H. M.

Volatile Alkaloids. By O. DE CONINCK (Compt. rend, 105, 1258 -1260).-A description of colour and other reactions of the pyridic alkaloids.


Improved Method of Estimating Caffeïne in Coffee. E. D. SMITH (Chem. Centr., 1837, 1270-1271; from Zeit. öster. Apoth. Ver., 41, 359).-To determine caffeïne, 0.65 gram of coarsely powdered coffee is mixed with 0.13 gram of magnesia, boiled with 150 c.c. of water for five minutes, filtered, and the filtrate made up to 200 c.c. by percolation; the residue is again boiled for five minutes with 100 c.c. water, filtered, and this filtrate made up to 300 c.c. by percolation. The combined filtrates are evaporated to 20 c.c., the residue treated with 120 c.c. of strong alcohol, the precipitate filtered and washed with alcohol, the alcohol driven off, and the residue dissolved by the gradual addition of small quantities of water. This solution is extracted three times with 25 c.c. of chloroform. On evaporating the chloroform, crystalline caffeïne is obtained. J. P. L.

Estimation of Theïne in Tea. By B. H. PAUL and A. J. COWNLEY (Pharm. J. Trans. [3], 18, 417-419).-The authors recommended the following method of estimating theïne (caffeïne) in tea :-5 grams of the powdered tea is moistened with hot water, mixed with 1 gram of calcium hydroxide, and dried on the water-bath. The residue is then extracted with strong alcohol in a small percolator. The alcohol is removed from the filtrate by evaporation, and the residual aqueous

solution mixed with a few drops of sulphuric acid, filtered, and the filtrate shaken in a separator with about 200 c.c. of chloroform, which is used in six successive portions, the last being tested to The whole of the ascertain that all the theïne has been taken up.

chloroform solution is then shaken in a separator with a very dilute solution of aqueous soda, by which it is completely decolorised; the chloroform extract is evaporated in a tared flask, and the residue weighed.

The authors have found in Indian and Ceylon teas a higher percentage of theïne than has usually been supposed to exist, varying from 3.22 to 4.66 per cent. on the air-dried substance. The amount of theïne had no relation to the commercial value of the 28 samples examined, the prices of which ranged from 7d. to 7s. lb. The numerical results are given in a table. The method used by dealers in testing tea (tasting) is described. The value is not indicated by the amount R. R. of extract obtained by boiling water.


Estimation of the Alkaloïds of Conium. By R. A. CRIPPS (Pharm. J. Trans. [3], 18, 511-512).-In place of the troublesome and inaccurate distillation methods, a process is described based on the extraction of the finely powdered fruits with a mixture of alcohol, chloroform, and a chloroform solution of hydrogen chloride, and the conversion of the bases in the extract into the hydrochlorides, in which form they are weighed. A. J. G.

Determination of Tannin. By F. GANTTER (Zeit. anal. Chem., 26, 680-682).—The author has made comparative determinations of tannin in a variety of tanning materials by the method of the German Tanuers' Association (known as the 1 c.c. method), using on the one hand hide powder, and on the other the N/10 solution of ferric acetate recommended by E. B. (Abstr., 1887, 311). In almost every case, the latter gave the lower result, and from the fact of the quantity of substance taken being varied according to its richness, the differences are inversely proportional to the weights employed. In a nearly pure tannin, the difference amounted to 10 per cent. Unless, therefore, it can be shown that the precipitation by hide powder gives incorrect results, that by ferric acetate cannot be substituted for it.

M. J. S.

Hoppe-Seyler's Soda-test for Carbonic-Oxide-Hæmoglobin. By E. SALKOWSKI (Zeit. physiol. Chem., 12, 227-228).-The following method of performing the test is suggested:-The blood in question is diluted 20 times, and to some of this in a test-tube an equal volume of aqueous soda of sp. gr. 134 is added. In a few seconds, carbonic oxide blood becomes whitish, then red; on standing, red flocculi separate, and finally rise to the surface of a faintly rosecoloured liquid. In normal blood, all that is produced by the addition of the alkali, is a dirty-brown coloration. After standing 20 hours, the precipitate in both cases is redissolved, and the clear red liquid shows almost the same absorption-bands as those of oxyW. D. H. hæmoglobin.


General and Physical Chemistry.

Refraction of Liquids between wide Limits of Temperature. By E. KETTELER (Ann. Phys. Chem. [2], 33, 353-381 and 506-534).— As none of the expressions hitherto proposed for the refraction of liquids at different temperatures are found satisfactory, the author has from theoretical considerations deduced a relation between refraction and temperature, and tested the same experimentally. The first part of the paper is devoted to a description of the apparatus used and the method of working, in the second are the results of the experiments and a discussion of the same. The liquids examined were water and alcohol. The refraction in every case was measured for the sodium, lithium, and thallium lines.

The refraction of water between 0° and 20° having already been measured by Rühlmann, the present experiments were conducted for the interval 20° to 95°, a large number of observations at intermediate temperatures being made. The formula proposed by Rühlmann for the refractive index, v = Vo at+bt, is found to be in fair agreement with the observed values up to 85°. A much better agreement is, however, obtained by the use of the author's proposed formula (n2 — 1)(v — ß) = C(1 + xe ̄kt) = M,

[ocr errors]

where C (n2 — 1) v = 0.62035, is taken from Lorenz's measurements for water-vapour and the sodium line. Also, since it is now proved that the dispersion for liquids follows the same law as that n'a - 1 for gases, = const., where for the same density and ß n's - 1 represent two different coloured rays; the constant C will alone alter with the nature of the light of which the other constants will be independent. For the lithium and thallium lines, C has the values 0.61574 and 0.62428. The equation is also satisfied by two series of values for the other constants, ẞ, a, and k.

a = 0.00246

a = 0·05617

k = 0.02290
k =

(1) B = 0.20271 or (2) B = 0·15999 Both these series give values agreeing very well with the experiments, and the two agree well together up to 100°. For the latter series, k is so small that its higher powers may be neglected, and we get M C{1a(1 kt)}, or the law of refraction takes a linear form with regard to t: (n2 - 1)(v — B) = c — yt.


[ocr errors]

For alcohol, the experiments extend from -7.85° to +76·34°. A law similar to that in the case of water holds good. It is, however, not continuous, and two portions of the series have to be distinguished. Between t = 33.69° and t = co we get M = C (1 + ae kit − t)), where the value of C for the lithium, sodium, and thallium lines is 0-84198, 0.84750, 0·85249; t。 = 33·69°, ß = 0·200, a = 0·07748, and



k = 0.002215.

From t = -10° to t = 33 69°, we get the simpler law M = C(1a), the value of which is 0.91317 for the sodium


In each case, the experiments prove that the dispersive power for the liquid and gaseous states is the same.

H. C.

A New Method for Determining the Rotatory Dispersion of an Active Substance and a Case of Anomalous Dispersion. By G. H. v. Wyss (Ann. Phys. Chem. [2], 33, 554-569).—The spectrum from ordinary white light is directed into a polarising apparatus by means of a collimator, the slit of which is adjusted to allow successively the passage of rays from each portion of the spectrum. The wave-lengths of the rays are calculated by means of Cauchy's formula. In this way, the rotation for different portions of the spectrum and for known wave-lengths is obtained.

A sample of turpentine thus examined was found to be feebly lævorotatory. It gave for λ = 661 a rotation of 3.0815°. As the wave-lengths decreased, the rotation increased and reached a maximum of 3.3688° for λ = 565. The rotation now, however, began to decrease on approaching the blue end of the spectrum, and for λ = 494 was only 2.9976°.

This behaviour is contrary to Biot's law that the rotation of a ray is inversely proportional to the square of the wave-length, an exception to which, however, had already been observed by Biot and Arndtsen in the case of tartaric acid, the latter concluding that the sample examined was probably a mixture of lævo- and dextro-rotatory acid. As it was quite possible that the present anomaly might arise from a similar cause, mixtures of dextro- and lævo-turpentines in different proportions were examined, and it was indeed found that a mixture containing about 68 per cent. of dextro-turpentine behaved in a manner exactly similar to the above. H. C.

Comparing Spectra. By E. F. J. Love (Phil. Mag. [5], 25, 1-6). In order to discriminate coincidences between lines of different spectra, a method based on the law of error is made use of, in which the differences between the wave-lengths of the compared lines are divided into groups according to the magnitude of the difference. A curve is then plotted having the number of observations in a group for an ordinate, and the average error of the group for an abscissa. The curve is compared with that given by the equation y = ae-c2x2, and any large discrepancies noted. The above law supposes the errors to be of every magnitude and infinite number, hence the observed curve will be steeper than the theoretical one. Comparisons by this method of spectra of the same element obtained by different observers give curves which agree closely with the curve of error, whilst that obtained by a comparison of iron with nickel and titanium is widely divergent. A comparison between the arc-spectrum of cerium and the widened lines of a sun-spot spectrum gave a curve closely resembling the above test curves. Grünwald's comparison (Abstr., 1887, 1070) between the wave-lengths of the water-spectrum as deduced

by him from those of hydrogen and their values as obtained by observation, is examined. The curve obtained agrees almost exactly with the theoretical curve except at four points. Finally, a comparison between the spectrum of Grünwald's constituent "b" of hydrogen and the nearest solar lines gives a curve which agrees with theory except at two points. The discrepancies in the two last comparisons seem to point to a small systematic difference, probably between the scale of measurement of the hydrogen spectrum and Ångström's scale.

H. K. T.

Modifications in the Absorption-spectrum of a Substance. By F. STENGER (Ann. Phys. Chem. [2], 33, 577-586).-The absorption of light by any substance of definite chemical composition is assumed to depend on the size of the physical molecules of that sub-stance in the condition in which it is examined. Hence as long as the molecular aggregation remains the same, whatever condition the substance be in, the absorption-spectrum will be the same, but an alteration in the molecular aggregation will bring about an alteration in the absorption-spectrum.

Iodine dissolves in carbon bisulphide to a violet, in alcohol to a brown solution. As the colour of the carbon bisulphide solution approaches more nearly to that of iodine vapour than that of the alcoholic, it may be assumed that in the first the molecules are in a simpler state of aggregation, and this view is supported by Wiedemann's observation that when the carbon bisulphide solution is cooled by means of ether and solid carbon dioxide, the colour changes to brown.

Magdala-red has a very different absorption-spectrum in alcohol, in which it is very soluble, to that which it shows in water, benzene, toluene, xylene, turpentine, or carbon bisulphide, in which it is only sparingly soluble. By mixing any of the latter solvents, however, with alcohol in different proportions, the absorption-spectrum of the one may gradually be converted into that of the other. The molecular aggregation in the first case where the substance is readily soluble is probably simpler than in those in which it is sparingly soluble.

Vogel (Berl. Monatsber., 1878, 409) has shown that when a solution of a coloured substance or salt is evaporated on a glass plate, the absorption-spectrum of the solid obtained is very different from that of the solution. This alone greatly supports the view as to the influence of molecular aggregation. The author shows that if the solution be mixed with collodion, gelatin, or starch-paste and then evaporated, the character of the absorption-spectrum does not alter with the change from the liquid to the solid state. The above substances would appear to keep the molecules apart when drying, and prevent further aggregation taking place.

The above also holds good of those substances which show fluorescence in the liquid but not in the solid state. If solutions of such substances in gelatin be evaporated, the film fluoresces strongly even after it has been kept for some time. As fluorescence probably depends on the smallness of the physical molecules of the fluorescent substance, this further supports the author's views.

H. C.

« PreviousContinue »