148 ME. J. A. DOUGLAS ON CHANGES OF PHYSICAL [May 1907, 



examined was first determined by means of a ' "Walker's balance,' 

 two or three readings being taken in every case. 



A portion of the rock, about the size of an acorn, was then 

 crushed in a steel mortar, passed through a fine copper sieve, and 

 subsequently ground to the finest powder in an agate mortar. 



A sample of this powder was then placed in the loop of the 

 platinum-furnace, and subjected to a temperature just below the 

 fusion-point of platinum. After some time, the current was 

 switched off, and the molten rock instantaneously solidified as glass, 

 which, on the platinum-loop being pulled out, peeled off in the form 

 of a flake ; it was then crushed, and examined microscopically under 

 crossed nicols. If found not to be perfectly homogeneous, it was 

 remelted until all traces of crystalline particles had disappeared. 



When the fusion was found to be complete, the glass was gently 

 broken up in a small steel mortar, care being taken not to grind it 

 up into too fine a powder. It was then passed through a small copper 

 sieve, with a mesh of 12 to the square millimetre. The sieve was 

 so constructed, that a fresh piece of copper-gauze could be used for 

 each rock. The specific gravity of the glass was then determined by 

 means of a diffusion- column of heavy fluids. 



The diffusion-columns were made from hard glass-tubing, of 

 f-inch bore, cut into lengths of about 9 inches. It was found 

 extremely important to clean and dry thoroughly each tube before 

 use, as the powdered glass readily adhered to its surface. Strong 

 sulphuric acid, absolute alcohol, and benzol were found quite 

 efficient for this pupose. When the tube had been cleaned, its 

 lower end was sealed up in the blowpipe-flame. The construction 

 of a diffusion-column is too well-known to need description, but it 

 may be noted that the powdered glass was always poured in before 

 the fluids, as the latter, by moistening the sides of the tube, caused 

 the fine particles to stick to the surface of the glass, and allowed 

 only a small proportion to enter the column. 



The fluids employed were carbon-tetrachloride and methylene- 

 iodide, and the indices were small glass beads, the specific gravities 

 of which had been determined to the third place of decimals. These 

 were dropped into the column after the glass had risen and formed 

 a zone. 



The top of the tube was then sealed in the blowpipe-flame, so 

 that little or no evaporation took place and the diffusion was 

 very slow. 



The positions of the beads and glass-zone in each column, which 

 were read by the aid of a horizontal travelling microscope, were then 

 plotted on a curve. When the diffusion was complete the beads, if 

 accurate, lay on a straight line ; a week or more often elapsed 

 before this took place, so that the zone of glass could be carefully 

 watched. In cases where the curve did not eventually become a 

 straight line, the beads were redetermined and subsequent cor- 

 rections made. The powdered glass usually settled into a well- 

 defined zone, the specific gravity of which could be determined by 

 insertion on the curve formed by the indices. 



