PHYSICS. 105 



weight, at the lower end of which was a hardened steel point 

 7.068 square millimeters in area. A flat anvil of Bessemer 

 steel was employed as a support for the nitroglycerin, which 

 was placed on it in a thin layer, and the weight dropped on 

 it from different heights. The mean height of fall necessary 

 to cause explosion of the liquid was 0.78 meter, whereas the 

 frozen nitroglycerin did not explode till a height of 2.13 me- 

 ters was reached. The author has also determined some of 

 the constants of the solid substance. The heat of fusion was 

 found to be 33.54 units as a mean of three experiments. 

 The density was found to be 1.735 at 10 C. a temperature 

 near its melting-point. The density of the liquid being 1.599, 

 it follows that, in crystallizing, nitroglycerin contracts about 

 tVt of its original volume. 



Wilson has proposed a simple mode of showing convection 

 currents in liquids. A glass cell with flat sides has a brass 

 tube in a depression in the bottom, which communicates with 

 a steam supply. The tube is surrounded with a jelly con- 

 taining aniline red, which is insoluble in cold Avater. On 

 filling the cell with water, and blowing steam through the 

 tube, the jelly dissolves, and colored currents stream up from 

 below. 



Guthrie has made a series of experiments to determine the 

 effect of a crystalloid on a colloid when in the presence of 

 water. Two or three lumps of rock-salt were added to a 

 jelly of size, and the whole was hermetically sealed in a 

 glass tube. The colloid parted with its water readily, a sat- 

 urated solution of the salt was obtained, and the size became 

 perfectly white and opaque, having undergone a structural 

 change. Experiments were also made in which a more hy- 

 grometric salt, calcium chloride, was employed. The author 

 thinks that it might be possible to fix the existence of a 

 point at which the jelly does not give up its water to the 

 hygrometric substance, and points out the analogy between 

 a jelly and a mass of small bags filled with liquid. 



The same author has also observed the curious fact that 

 while a crystal of alum or a saturated solution of salt, when 

 introduced into the Torricellian vacuum, depresses the mer- 

 curial column to a less extent than water, a solution of size, 

 gum-arabic, or of any colloid depresses it to precisely the 

 same extent. Hence water has different vapor densities in 



E2 



