282 Messrs. R. G. Parker and A, J. Dalladay on a 



particles of cellulose from the rag used in cleaning the 

 surfaces ; a pressure of many atmospheres can be produced 

 in this way. By careful cleaning it is possible to eliminate 

 them almost entirely. Cells which have comparatively large 

 areas in contact are less free from spots than those in which no* 

 point is far from an edge; the gases which would remain 

 permanently enclosed in the former are able to escape from 

 the latter, and the surfaces then go back into complete 

 contact. 



The optical surfaces of this cell were found to be perfect 

 and quite unchanged after heating, including the unsupported 

 internal surfaces ; we therefore have optically worked sur- 

 faces in the interior of a solid glass cell. En fig. 5 it will 

 be seen that an area at X is completely out of contact except 

 for a narrow arc formed round the drilled aperture. This is 

 clearly shown in fig. 6 (PL V.), under a magnification of 

 six diameters. It is not well known that ground-glass surfaces 

 act as if they were in compression : thus, a thin plate ori- 

 ginally flat and polished on both sides becomes, when finely 

 ground on one side, a concave mirror of long focus on the 

 other. In the case of our cell, the rough surfaces of the 

 drilled holes push up a burr round the edge which is only a few 

 wave-lengths high, but sufficient to prevent complete optical 

 contact. This burr was therefore removed by "retouching" 

 on a polisher, but traces of it evidently remained, and by 

 producing local pressure caused the union to be very complete 

 over its surface. Severe strain was applied by forcing a thin 

 wedge between the two surfaces without breaking this narrow 

 area of complete union, The same effect was also shown by 

 another corner of the same cell which was out of contact 

 after heating and was broken off; the fracture travelled 

 round the edge of the drilled hole instead of passing straight 

 across it (PL V. fig. 7 and fig. 5 at Y). 



A second cell was then made of the same glass and pattern, 

 and heated at the rate of 1°-1J° C. per minute to a tempe- 

 rature nearer the annealing-point — 523° C, — at which it was 

 kept for two hours, then cooled at 4°-5° C. per minute. 

 Pressure was applied by means of a hand screw-clamp. Two 

 brass plates, with surfaces ground flat, and sufficiently thick 

 to eliminate bending (^ in. thick), were used to distribute 

 the pressure ; a few thicknesses of thin paper were placed 

 between the brass surface and glass surface to take up minute 

 irregularities — the carbonization of the paper did not inter- 

 fere with its action. 



The union of the three components was complete except 

 for a few small spots. They could not be separated by any 



