2.16500 units; that of three, 13500 + 2.16500 units; and that of four, 2.13500 +2.16500 units. Again, among the alcohols, the primary are found to have the highest, and the isomeric tertiary the lowest heats of combustion, the secondary occupying an intermediate position. These results are attributed by the author to the effect of the local accumulation of negative elements in the molecule. W. P. W. Note on the Foregoing Communication. By S. U. PICKEring (Phil. Mag. [5], 23, 109-112).-The author points out that if Thomsen's value of f(C1) be erroneous, the only meaning then attaching to his values of v1, v2, and v, will be that they are the differences between their actual values. Adopting Armstrong's view that f(C1) = 135340 + x, it is shown that the value for r will become r + 4 while those of v1, v2, and v ̧ will be converted into v1 + V2 + x, and 2' 3a 13 + respectively. If, as is most probable, a represents some 2' number considerably larger than 14000 units, the heat developed in the union of two carbon-atoms will be very nearly, though not quite, proportional to the number of bonds by which they are united. Thomsen's argument that the aldehydic radicle consists of hydroxyl rests on the fact that the heat of formation of HCO in the aldehydes, together with that of CO in the ketones, was found to be equal to that of O: COH in the acids; the introduction of a, however, destroys this equality, for the sum of the heat of formation of the aldehydic and ketonic radicles will exceed that of carboxyl by and this excess renders the results entirely in accordance with 2' the generally accepted views concerning the constitution of these subW. P. W. stances. Criticism of Thomsen's Theory of the Heat of Formation of Organic Compounds. By J. W. BRÜHL (J. pr. Chem., 35, 181— 204, and 209-236).-Starting with the premise that conclusions drawn from experiments undertaken with the object of ascertaining the molecular structure of compounds from physical considerations require, for the present at least, a general confirmation from chemical data, since this question has been studied for a longer time and to a greater degree by chemical than by physical methods, the author submits that such conclusions must be received with caution when they lead to constitutional formula which are inexplicable, having regard to the methods of formation and the chemical properties of the compounds concerned. The views on the constitution of carbon-compounds put forward by Thomsen (Thermochem. Unters., Bd. iv) are of this nature (compare preceding Abstracts), and the author criticises these views with the object of demonstrating that they are untrustworthy and based on fallacious arguments. With regard to the calculation of the fundamental constant v2 (loc. cit.), it is noted that Thomsen assumes without any experimental proof that the union of gaseous atomic carbon with two atoms of oxygen is attended with the development of twice as much heat as when combination ensues with one atom; and further that the value of v2 = 142505000, if the extreme values be substituted for the mean deduced from 14 comparisons of the heats of combustion of saturated and unsaturated compouds. Moreover, the values of the fundamental constants, v, and r, and the conclusion v1 = v2, deduced in the first instance from the heat of formation of hydrocarbons, are assumed by Thomsen to hold for all classes of compounds, and on these values he bases his theories of the constitution of carbon-compounds, although the author shows that assuming v, and r constant, v2 has among others the values 7500 units, 9590 units, and 11110 units (deduced from the heats of formation of allyl and ethyl alcohols, the ethers, and formates respectively). The method adopted by the author in his criticism is well illustrated in the section of the paper devoted to the consideration of Thomsen's views of the constitution of aldehydes, ketones, acids, and ethereal salts. Thomsen deduces the heat of formation of COH (65400 units), CO (54250 units), and COOH (119960 units), by subtracting from the heat of formation of the aldehyde, ketone, or acid the amount of heat assumed to be developed in the formation of the contained hydrocarbon radicle, it being assumed that in these compounds v1 = 14200, and r = 15000 units; and since 65400 + 54250 119650, he argues that the aldehyde radicle contains hydroxyl, thus C.(OH). No evidence is offered of so surprising a conclusion, and with the same right that Thomsen argues that the group CO in formic, acetic and propionic acids has the same value as CO in acetone, the author submits that the residue, C(OH), of the acids should be supposed to have the same value as it has in methyl, ethyl and propyl alcohols, that is, 44600 units, whence 119650 = 4460075050, and CO = 75050 units. But even if Thomsen's unproved assertion that the heat of formation of the CO-group in the ketones and acids has a constant value = 54250 units be accepted, it does not follow that because in this case the heat of formation of the COH residue must be 65400, a similar value to that already deduced for the aldehyde residue from independent considerations, that the aldehyde radicle must have an analogous constitution any more than a similar sp. gr., boiling point, or other physical property establishes the identity of two compounds. Adopting Thomsen's value that the heat of formation of the aldehyde radicle = 65400 units, and that r = 15000, the constitution HCO would lead to a value for CO = 50400 units, a value not very different from that of CO in acetone; that the value should be exactly the same even in one and the same series is scarcely to be expected: for example, in the alcohols the heat of formation of C(OH) varies between 41800 and 61600, and in the acids = 65400 units, so that Thomsen's view that the aldehyde radicle contains hydroxyl must be regarded as arbitrary. Passing on to acetic anhydride, Thomsen calculates for the heat of formation of the group O: C·O·C: O, the value 165940 units, and, inasmuch as 165940 = 3.55310, assumes that the heats of formation of the groups COC and CO are identical. But as, in the case of the ethers, Thomsen has deduced for the heat of formation of CO-C the value 31500 units, and since 165940 = 31500 + 2.67220, the heat of formation of C: O must be 67220 units instead of 55310; hence the author urges that the fundamental hypothesis of the thermal constancy of the affinity between carbon and oxygen is disproved, inasmuch as the heat of formation of the group CŎ or C.O.C is shown to be variable. Moreover, comparing Thomsen's value of the heat of formation of the group C:O in ketones and acids, 54250 units, with the value of this group in carbonic oxide and carbonic anhydride which, according to Thomsen's fundamental assumption, must be 67670 units, the difference between these values, 25 per cent., is irreconcilable with the hypothesis on which the whole theory is based. With regard to ethereal salts, the author severely criticises the argument by which Thomsen is led to deduce the value 105000 units for the heat of formation of the group O: COC, and to alter the formulæ of those salts in which there is a marked deviation from this average value; moreover, it is pointed out that the formula proposed for ethyl acetate, OH CHMe CMe O, and calculated from the equation 7r + 3v1 + 54250 + 65400 = 267250 (the actual value being 265910), is based on an entirely wrong conception, inasınuch as Thomsen's proposed formula for the aldehyde radicle is RC(OH). This radicle is not present in the proposed formula for ethyl acetate, and as a compound of the formula given would most probably resemble a secondary alcohol, the value of the heat of formation of C(OH) in isopropyl alcohol, namely, 50710 units, ought to be substituted for 65400, giving a total 252560, which in nowise agrees with the experimental results. Criticising in a similar manner Thomsen's views with regard to the molecular structure of the other classes of carboncompounds, for details of which the paper must be consulted, the author concludes that they cannot be maintained, and finally calls attention to the controversy which has arisen as to the correctness of Thomsen's value for the heat of combustion of benzene. W. P. W. Thomsen's Investigations. By F. STOHMANN (J. pr. Chem. [2], 35, 136-141; compare Abstr., 1886, 812).-In determining the heat of combustion of liquid ethyl ether the mean 8805 cal. per gram or 651570 per gram-mol. was obtained. Thomsen obtained the number 660200 cal. as the mean of the series of experiments in which the ether was burnt partly as vapour mixed with air, partly in a universal burner as vapour at 22°. Thomsen's number when corrected gives 652830 cal. A second series of experiments is described in which the ether is burnt as gas at 17° in a current of oxygen by the method employed for burning benzene vapour (J. pr. Chem. [2], 33, 256). The mean number 8921 cal. per gram, equal to 660175 cal. per gram-mol., was obtained. Thomsen found the heat of combustion of ether at 18° to be 659600 cal. These experiments show that the author and Thomsen obtain concordant results when working under the same conditions. The author still maintains that the higher numbers obtained by VOL. LII. Thomsen are due to heat carried over from the burner to the calorimeter (Abstr., 1886, 409), and suggests that the error in the case of benzene being 11,400 cal., the errors in the case of substances which have to be heated at 116° and higher must be still greater. N. H. M. Alcoholates of Sodium Glyceroxide. By DE FORCRAND (Compt. rend., 104, 291-294).-If a concentrated solution of sodium methoxide in absolute methyl alcohol is mixed with an equivalent quantity of glycerol, the mixture deposits an alcoholate of sodium glyceroxide, CH(OH)2 ONa + MeOH, in very deliquescent, white needles, which lose the whole of their methyl alcohol when heated at 120° in a current of dry hydrogen. It dissolves in methyl alcohol to the extent of 120 grams per litre at 15°. Heat of dissolution in water at 16° -2.00 cal. = C,H,(OH), liq. + NaHO solid + MeOH C3H5(OH)3 + MeONa diss. in nMeOH liq. It is evident that as in the case of the ethyl alcoholate the formation of the alcoholate is exothermic whilst that of the glyceroxide alone would be endothermic. The thermal disturbances which accompany the four corresponding reactions in the case of the ethyl alcoholate are respectively, +16.60 +4.58 +3.92 -0.63. The ethyl alcoholate of sodium glyceroxide dissolves in ethyl alcohol to the extent of 13 grams per litre at 15°. The propyl alcoholate, C3H5(OH), ONa, PrOH, is obtained in a similar manner; one litre of propyl alcohol dissolves 7 grams at 15°. Heat of dissolution in water = -0.57 cal. The thermal disturbances corresponding with the four reactions already given are respectively, +16.69 +4.67 +6.10 1.65 cal. The isobutyl alcoholate is obtained in the same way; one litre of isobutyl alcohol dissolves 4-6 grams at 15°. Heat of dissolution in water = +1.23 cal. The thermal disturbances corresponding with the four reactions are respectively, where a is the heat of dissolution of sodium isobutyl oxide in excess of isobutyl alcohol, which has not been determined, but is in all probability between +10 and +12 cal. The amyl alcoholate dissolves in amyl alcohol to the extent of 28 grams per litre at 15°. Heat of dissolution in water +0.99 cal. The thermal disturbances corresponding with the four reactions are +12.41 - x, +2.80 +15.02 x +14.82 where x has the same signification as in the case of the isobutyl compound. C. H. B. Alcoholates of Potassium Glyceroxide. By DE FORCRAND (Compt. rend., 104, 361-364).—These alcoholates are compounds of potassium glyceroxide, C,H,O,K, with one molecule of an alcohol of the paraffin series. They are obtained by adding one molecule of glycerol to a solution of one molecule of the potassium alkyl oxide in the corresponding alcohol. The following table gives the heats of dissolution of the alcoholates in water, the thermal disturbances corresponding respectively with the following reactions: (1.) CHO, liq. + HOK solid + ROH liq. H2O solid + C3H,O,K,ROH solid. (2.) C2H,O,K solid + ROH liq. = C,H,O,K,ROH solid. (3.) CHO, liq. + ROK diss. in nROH liq. = C,H,O,K,ROH solidnROH liq. (4.) CHO, liq. + ROK diss. in nROH liq. C,H,O,K solid + (n + 1)ROH liq. and the amount of the alcoholate dissolved by one litre of the corresponding alcohol. Isobutyl alcohol forms no alcoholate with potassium glyceroxide; with this exception, all the monhydric alcohols behave in the same way. The formation of the alcoholate is exothermic, but the quantity of heat developed, with the same metallic derivative, decreases as the molecular weight of the alcohol increases. For the same alcoholate, the heat developed diminishes as the atomic weight of the metal in the glyceroxide increases. The solubility of the alcoholates in the corresponding alcohol also diminishes as the molecular weight increases. C. H. B. Heat Equivalents of the Homologues of Benzene. By F. STOHMANN, P. RODATZ, and W. HERZBERG (J. pr. Chem. [2], 35, 40-42).—These investigations presented considerable difficulties |