TRANSACTIONS OF SECTION B. 837 



volume of water. The cooling curves obtained in this way are of considerable 

 interest. The fatty acids investigated have been pahnitic, .stearic, lauric, and 

 myristic acids. The cooling curve of a pure fatty acid turns sharply round when 

 the solidifying point is reached, runs straight up the plate till the solidifying is 

 finished, and then turns sharply off again. One per cent, of another fatty acid quite 

 perceptibly alters the shape of the curve, so that the character of a cooling curve 

 eeems a good test of purity. When a larger portion of a fatty acid is introduced a 

 second latent heat-point is developed, the curve showing a discontinuity below the 

 solidifying point of the mixture. As the solidifying point is lowered by introducing 

 more and more of the second fatty acid, this discontinuity is gradually merged in 

 the common melting-point of the mixture, thus reproducing the phenomena 

 observed by Professor Roberts Austen in the case of certain alloys. This discon- 

 tinuity can hardly be due to the formation of a compound, but is probably caused 

 by the presence of the ' eutectic alloy ' in the mass. We are now repeating our ex- 

 periments with synthetically prepared organic bodies of known constitution, and 

 studying also the cooling curve of water. 



7. On the More Exact Determination of the Densities of Crystals. 

 By the Earl of Berkeley. 



A comparison of the several values found by different observers for the density 

 of one and the same crystallised salt shows variations amounting in some cases to 

 10 per cent. As the density is assumed to be a physical constant, independent of 

 the manner in which the crystals have been produced, these variations are probably 

 due to errors of experiment. 



The chief sources of such errors are (1) imperfect measurement of temperature 

 and volume, (2) occlusion of mother liquor, (3) adhesion of air, (4) hygroscopic 

 nature of the salts. 



In the paper are described the methods devised by the author ior reducing the 

 amount of these errors. 



1. Two conical pyknometers, of about 7 c.c. capacity, with thermometer 

 stoppers and calibrated capillary side tubes, were used. One served as a counter- 

 poise, and was treated externally exactly like the other. They were repeatedly 

 heated to 130° C, and allowed to cool, in order to bring the glass into a con- 

 dition of molecular rest. For determining capacity the flask filled with water 

 was placed in a desiccator, from which the air was exhausted till, on tapping, the 

 water boiled. After it had thus been kept boiling for some time, the flask was 

 removed, and the stopper inserted in a position and under a pressure which were 

 observed. After the two pyknometers had been wiped dry they were placed on 

 the balance pans, and when the temperature had become steady and the level of 

 the liquid in the capillary had, in consequence of evaporation round the neck, 

 fallen below the highest graduation, the weight, the level in the capillary, and the 

 temperature were noted. The greatest ditference between any two out of eight 

 estimations of capacity thus made was 0-00029 c.c. 



The liquid in which the crystals were weighed was carbon tetrachloride. 

 Owing to the high coefficient of expansion of this liquid, special means, which are 

 described, were devised for keeping constant the temperature of the balance case. 

 The evaporation between the neck and the stopper was at the rate of about 00001 

 gram per minute. 



2. To remove occluded mother liquor the crystals may be reduced to powder 

 and then dried, the presumption being that the crystals will break across the 

 cavities containing the mother liquor, and that the latter on evaporating will 

 deposit crystals of the same kind. But the crushing process may produce change, 

 as in the familiar case of mercuric iodide ; and it is better to form small crystals in 

 a solution kept at a constant temperature and constantly stin-ed. An apparatus 

 was planned and constructed for this purpose. 



3. To prevent adhesion of air, the pyknometer with the weighed crystals at 

 bottom, together with a bulb holding a charge of carbon tetrachloride, were so 



