determined over a much wider range, Wüllner has found Helmholtz's formula, and Ketteler his own formula, to give excellent results in the case of feebly dispersive substances, although, as the author points out, the two are fundamentally different, and cannot both be correct. From the author's calculations, it appears that both are equally insufficient when applied to highly dispersive media. In proof, a table is given of the differences between the observed values of n for the ordinary and extraordinary spectra of calcspar, and the values calculated by Helmholtz's and Ketteler's equations respectively. The observed values are by Mascart for lines from A to R, and by Sarasin for the ultra-violet (cadmium) spectrum. The constants were determined for lines over the whole spectrum. The results show that for the extraordinary spectrum, in which the dispersion is low, Helmholtz's formula is satisfactory, Ketteler's somewhat better. But when applied to the ordinary spectrum, in which the dispersion is 23 times greater, the discrepancies are very serious, extending in the case of Helmholtz's formula to the third place of decimals. Ketteler's equation proved somewhat better. From his investigations, the author draws the following important conclusions:-That all the formulæ connecting wave-length and refrangibility hitherto proposed are of purely empirical character, and applicable only to media of low dispersive power: that at present it is not certain that the refractive index reaches a limiting value, unity or otherwise, either for co or for any other wave-length; and, finally, that in investigations bearing on chemical constitution and refractive power, dispersion must either be taken into account, or be eliminated in some empirical way. = Сн. В. Supposed Influence of Multiple Bonds of Union on the Molecular Refraction of the Hydrocarbons. By JULIUS THOMSEN (Ber., 19, 2837).—In 1862-64, Landolt (Ann. Phys. Chem., 117, 122, and 123) proved beyond doubt that a connection exists between molecular refractive power and chemical composition, by showing that in any homologous series the refractive power increases from member to member by a constant amount, and is independent of isomerism and metamerism. Since then Brühl (Abstr., 1880, 295, 781; 1882, 445, and preceding Abstract) has endeavoured, with apparent success, to establish a relation between molecular constitution and refractive power, by showing that the refractive power of a compound, although independent of the number of single bonds of union between carbon-atoms in its molecule, is decidedly influenced by double and treble bonds. The following considerations, however, make such an influence at least doubtful: Brühl's results are summed up in the formula n.c + 2m.h + aV1 + ßV2 + yV3 2 = Ꭱ . [1] in which R is the molecular refraction of a hydrocarbon C, Ham, c and h the refractions of a carbon and hydrogen atom respectively, a, B, and y the numbers of single, double, and treble bonds of union, and V1. V2, and V3 the increments of refractive power due to them respectively. Brühl now puts V1 = 0; but this assumption may be avoided, and V, eliminated as an independent constant by combining equation [1] with the necessarily true relation 2n Multiplying the latter by V1, and adding, we get - m- α 23 : 3y = 0. n(c + 2V1) + m(2h — V1) + ß(V2 − 2V1) + 7(V3 − 3V1) = R, [3] and finally substituting x, y, p, and q for the constant quantities within brackets, we get which is of the same form as [1], and identical with it when V1 = 0, but in which R is not necessarily independent of the number of single bonds of union between carbon-atoms. In either case, x + y = c + 2h + V1 represents the increase in refractive power for each addition of CH2. Experimentally, x + y has been found to vary from 4.85 in the benzene series to 4.525 in the naphthalene-group. No values of x and y can therefore exist which will in all cases satisfy equation [4], and this must also be true of the constants p and 9. The following example illustrates this : Taking the experimental value of the refraction of two molecules. of benzene, 51.86 = 12c + 12h + 6V1 + 6V2, and of one molecule of naphthalene, 43-93 10c + 8h + 6V1 +5V, and subtracting, we have 2c+4h + V2 = : 7.93. From this and the relation c + 2h+ V1 = 4.85 or 4525 we find V2 2V1 = constant p in equation [4] =-1.77 or - 1.12. Now Brühl (preceding Abstract) has found the following values for the constants of equation [4]: x = 2·48, y = 2.08, p = 1.78, q = 1.97 [5] The above calculation shows the negative influence of p to be as great as Brühl assumes it to be positive. Again the difference between the molecular refractions of naphthalene and benzene, calculated with Brühl's constants = 42:02 26.46 = 5.56, whereas the actual difference = 8·00. Brühl attributes the discrepancy to the greater dispersive power of naphthalene. The author rather attributes it to the incorrectness of Brühl's hypothesis, which he proceeds to show is unnecessary. = Assuming either that double and treble bonds of union have no influence on molecular refraction, or that V2 2V1, V3 = 3V1, equation [3] reduces to R = nx + my. The author has now determined the constants x and y from the observed molecular refractions of a series of five hydrocarbons of the general formula C.H2m, containing 1, 2, 3, and 4 double bonds of union, and thus arrives at the equation R = n. 4·014 + m. 0.840. A table is then given showing the molecular refractions of these five hydrocarbons and of three others, containing from 0 to 5 double bonds, calculated by the author's formula and by equation [4] with Brühl's constants; and the agreement with experiment is seen to be much better in the former case (mean error = 0-37) than in the latter (mean error = 0.8). The author therefore concludes that the mode of union of carbonatoms has no influence on molecular refraction. In the author's formula, the values of x and y can only be regarded as approximate. However, the range of variation is not great. A table is given of the values of a calculated for 22 hydrocarbons, roughly classified according to the number of double bonds in their molecules. Throughout y is taken = 0.84. In no case does the value of x differ much from the mean value given above. For one group only, of which hexahydronaphthalene may be considered typical, it falls as low as 375; but even here x + y = 459 is within the observed limits of variation. Сн. В. Thomsen's Supposed Explanation of Molecular-refraction Relations. By J. W. BRÜHL (Ber., 19, 3103-3108).-A partial reply to Thomsen (ibid., 19, 2837, see previous Abstract). The author points out that the constant y in Thomsen's formula for molecular refraction, R = nx + my, has been calculated from observations on a series of hydrocarbons of the formula C,H2m, in which the increments of refractive power for each addition of H2 are very variable, sometimes positive, sometimes negative, a variability which can only be attributed to chemical constitution. Like other empirical formulæ, Thomsen's is no doubt applicable to the observations from which it was deduced, and possibly to others besides. But in many cases (quoted by the author) it is seriously at fault, whilst the author's formula applies to all hydrocarbons except those having a high dispersive power, to which the author attributes a special influence. Referring to his previous work (see preceding Abstracts) for a full discussion of the question, he here contents himself with quoting examples of isomeric compounds whose molecular refractive powers differ more or less. In all the compounds selected, the dispersive power is low. Inspection of the table shows that when isomerides are equally saturated, that is, contain the same number of doubly or trebly bound atoms (isovaleric acid, propylic acetate, methylic butyrate), their molecular refractive powers differ very slightly; but when the saturation is unequal (allyl ethyl ether and valeral, cymene and hexahydronaphthalene) the differences are considerable, and beyond. the possible limits of experimental error. The influence of saturation on refractive power cannot therefore be ignored. Сн. В. Absolute Electrodynamometer. By H. PELLAT (Compt. rend., 103, 1189-1190).-A description of a new electrodynamometric balance. New Apparatus for Electrochemical Investigations. By N. V. KLOBUKOFF (J. pr. Chem. [2], 34, 539-547).-A description, with plates, of a "universal commutator" which fulfils the following con That with a given large number of circuits in which the measurement of the strength of the currents has to be determined with one and the same measuring instrument, the making and breaking contact with the latter can be easily and quickly effected. By PIONCHON (Compt. rend., 103, 1122—1125). cally introduced equal to that of the instrument. = 0.0748. 14.375 +0.0612931t 0.00001047412 + 0·00000001034483. = 0.0612931 0.0000209482t + 0·0000000310344ť. =" %= 011012t + 0-00002533337 + 00000000546666. = = = 14 0.218ł -39. t = 0·218. = ob +0.00002233ť. +0.00004466t. = 0·183493, −0·000282ť2 + 0·00000046666ť. 0.099 0.00006175t. 0.10584t+ 0.0000228667 +0.0000000219427ť. 0.10584 0.0000457334t+0·0000000658281ť2. 0.124t0.000042 14.8. 0.124 0.00008t. specific heats before and after fusion. Great care was taken to prevent Silver affords a further illustration of the identity between the the absorption of oxygen by the silver, and it was found that under these conditions the metal melts at 907°, a temperature much lower than that given by previous observers. Molten tin differs from all ordinary liquids in that its specific heat changes very slowly. Experiments with gas-carbon of which the vessels used to hold some of the metals were made, confirm Weber's statement that at high temperatures the specific heats of the different varieties of carbon are identical. The variations in the specific heats of the magnetic metals, iron, nickel and cobalt, show that they exist in several allotropic modifications. It is worthy of note that the changes take place at very different temperatures in the three cases. Experiments which are not yet completed, show that these changes of state are intimately connected with the magnetic properties of the three metals. The author's results confirm Berthelot's criticisim of the law of Dulong and Petit. This law amounts simply to a statement that there is a certain interval of temperature (between 0° and 100°) in which the values of the products of the specific heats of the elements into their combining weights are approximately equal. C. H. B. Thermochemistry of Bibasic Phosphates and their Congeners. By A. JOLY (Compt. rend., 103, 1197-1199).-The heats of formation of the bibasic phosphates from the dissolved acids and dissolved oxides were determined both directly and indirectly. Calcium hydrogen phosphate, CaHPO1, cryst. cryst... Barium hydrogen arsenate, BaHASO, gelat. cryst.... Cal. + 26 9 + 26.6 + 27.8 + 28.4 + 25.2 + 22.73 .... Strontium hydrogen phosphate, SrHPO, cryst... .... + 24-3 + 35-2 + 28.2 The differences between the heats of formation of the colloïdal and crystallised varieties is sufficient to show that the reactions involved in their formation are of a complex character. C. H. B. Ammonium Magnesium Phosphate. By BERTHELOT (Compt. rend., 103, 966-970).—The heat of formation of magnesium ammonium phosphate was determined by measuring the thermal disturbance which takes place when a solution of a magnesium salt is mixed with sodium phosphate and afterwards with ammonia. The numbers thus obtained for the heat of formation of the crystallised double phosphate were +407; +410; +40-6; +41.9. The last value is probably the most accurate, since the ammonia was added to the crystalline magnesium hydrogen phosphate, and the latter was converted directly into the crystalline double salt. A series of experiments in which a mixture of ammonia and sodium |