TRANSACTIONS OF SECTION B. 787 
Following in order these results came the Welsbach light, in which the gas 
heats to whiteness a tube or mantle, composed of a filmy thickness, of the oxides 
of Zirconium and Thorium, the mantle being surrounded by a glass tube similar 
to that used in some paraffin oil lamps; in this case the unburned carbon exceeded 
in amount the unburned hydrogen, there being 15:486 of the former and 3-794 of 
the latter per 1,000 of completely burned carbon. 
Three experiments were made with a Marsh-Greenall’s heating stove, in which 
three Bray’s luminous burners were employed. 
The first was made with a consumption of 5:62 cubic feet of gas per hour, when 
12°6 and 3:0 parts of carbon and hydrogen respectively were registered per 1,000 
parts of carbon completely burned. 
The second experiment, with a consumption of 5°74 cubic feet per hour, gave 
37°6 and 11:8 respectively of carbon and hydrogen unburned. 
The third experiment, with an increased consumption of gas (7'1 cubic feet per 
hour), gave 97°4 and 12:1 of carbon and hydrogen respectively unburned. 
Two experiments were made with one of T. Fletcher’s heating stoves, in which 
eight Bunsen burners play upon some fancy metal-work (iron coated with magnetic 
oxide) ; the one experiment, in which the amount of gas passing was not measured, 
gave a of carbon and 24°6 of hydrogen unburned per 1,000 of carbon completely 
burned. 
In the second experiment, where 6°81 cubic feet of gas were burned per hour, 
66°3 and 20:0 respectively of carbon and hydrogen unburned were registered. 
One experiment was made with one of T. Fletcher’s stoves in which twenty 
Bunsen burners play on asbestos projecting from a fire-clay back; with a consump- 
tion of 8:14 cubic feet of gas, 138°9 and 11:7 parts respectively of carbon and 
hydrogen per 1,000 parts of completely burned carbon were formed. 
8. Contributions to the Analysis of Fats.) 
By J. Lewxowirscu, Ph.D., F.I.C., £.C.8. 
The author gave in brief outline a review of the methods for the chemical 
analysis of fats. He recommended for the estimation of glycerol in fats, as the 
most exact method, the combination of the alcohol-ether extraction with Benedikt 
and Cantor's acetic method, which he has shown to give concordant results. The 
various methods for determining the nature of the various fats, and especially of 
the fatty acids, were shown by means of an analytical table, and the methods of 
Hehner, Reichert, Kéttstorfer, Hiibl, and Hazura were referred to. 
The author took objection to Benedikt’s method of determining the ‘ acetyl 
value ’—z.e., to give a value for hydroxylated acids present in a fat. Benedikt 
assumed that hydroxylated fatty acids, on being boiled with acetic anhydride, were 
acetylated, and transformed into acetyl hydroxylatids. On titrating these products 
in alcoholic solution with standard alkali, Benedikt obtained a certain acid value, 
due to the COOH group (as he thought) of the (supposed) acid, and, on saponifica- 
tion, when the acetyl was split off, a larger saponification value; the difference 
between the two values yielded the ‘acetyl value.’ The two reactions that were 
to take place may be expressed by the following equations for ricinolic acid :— 
1. C,,H,,(00,H,0)COOH + 1KOH =C,,H,,(0C,H,0)COOK + H,0. 
2. C,,H;,(0C,H,0)COOH + 2KOH =0,H,0, OK + C,,H,,(OH)COOK + 2H,0. 
The author, however, has shown that on boiling fatty acids with acetic anhy- 
dride the acids are transformed into their anhydrides, and he has proved this for 
capric, lauric, palmitic, stearic, cerotic, and oleic acids. A hydroxylated acid, eg., 
dihydroxystearic acid, which was prepared from oleic acid, undergoing this opera- 
tion will, of course, become acetylated, but at the same time anhydrated, so that the 
resulting product is nothing else than diacetylhydroxystearic anhydride. These 
anhydrides give no acid value, and all the ‘acid values’ which Benedikt has found 
in his experiments are only due to the fact that he dissolved the products of the 
1 Journ. of the Society uf Chemical Industry, 1890, p. 842. 
