280 Messrs. R. G. Parker and A. J. Dalladay on a 

 Experimental. 



The temperature selected for our experiments on the union 

 of glass surfaces was considerably below the annealing-point, 

 in accordance with the facts already discussed. Two pre- 

 liminary experiments w T ere made, using a metal box having 

 7 in. sides, filled with sand and heated by gas-burners. 

 Pieces of crown glass, 5 x 5 x 1*5 mm. were put into contact, 

 wrapped in asbestos-wool and held together by a clamp; they 

 were then heated at the rate of 1° C. per minute to a probable 

 maximum of 400°-450° C, maintained at this temperature 

 for 1 hour, and slowly cooled. 



They were found to have united, but not completely. 

 When torn apart with considerable force each piece carried 

 with it glass from the other. All the surfaces were optically 

 perfect after heating (fig. 4, PI. V.) . 



In the second experiment larger pieces (1 in. in diam., 

 ^ in. thick) of different kinds of glass were employed — crown 

 and heavy flint. Heating was as before, but rather more 

 vigorous ; probable maximum temperature 450°-500° C. The 

 glass was clamped between pieces of brass plate, with asbestos 

 packing. The disks united at one place only; here the joining 

 was perfect, and when they were forced apart the flint carried 

 with it a portion of the crown over the whole of this area. 

 The external surface of the flint glass was spoiled, as it had 

 softened slightly; the crown glass was almost uninjured. 



These experiments having proved that joining would take 

 place, more careful ones were conducted with a view to con- 

 structing apparatus by the method. 



A platinum-wound electric-tube furnace (internal dimen- 

 sions 3^ in. x 2^- in., 14 in. long) was fitted with double silica 

 windows at each end. It was controlled by two resistances 

 in series. One, the main resistance, was used to compensate 

 for the very great changes in the resistance of platinum which 

 take place between 20° C.and 600° C. ; the second permitted 

 of much more gradual changes of resistance, and formed the 

 fine adjustment in this respect. The temperature at any 

 instant was known by means of a platinum resistance thermo- 

 meter, allowance being made for any lag in its indications ; 

 and, further, the thermometer was connected to a Callendar 

 recorder, giving a permanent record of the temperatures 

 used in any experiment, and enabling the rate of change of 

 temperature to be seen at a glance. 



The heating and cooling were therefore completely under 

 control, and a continuous temperature change as small as 

 -|° C. per minute could be maintained indefinitely. 



An arrangement by which polarized light could be passed 



