Precision Method of uniting Optical Glass. 283* 



of the means indicated, and when subjected to severe mecha- 

 nical strain a portion of the cell broke away, showing the 

 conchoids I fracture of a solid glass block. Xo bending of 

 the fracture occurred at the interfaces, except at one small 

 portion where a spot showed that there was no contact 



(fig. 3, PI. v.). 



The optical surfaces had all remained perfect as regards 

 plastic deformation. The outer surfaces were slightly cor- 

 roded by the paper-ash, and had to be repolished. When 

 examined by means of a Babinet compensator (for the effect 

 on polarized light), very little internal stress was observed, 

 in spite of the unnecessarily rapid cooling ; the maximum 

 stress indicated was about 150 lb. per square inch. This 

 amount could easily be decreased, but is already negligible 

 for most purposes. 



A third cell or' rather different construction (fig. 8, PI. V.) 

 showed very few imperfections of contact. Heating was 

 carried out exactly as in the previous case, i. e. at 90° C 

 below the annealing-point. The areas in contact are all 

 narrow, and no point is more than 2^ mm. from an edge. 

 Any centres of gas-pressure developed in the way already 

 described are therefore likely to disappear owing to escape 

 of the gas. and we find in general that these narrow areas give 

 the most perfect joining. 



The glass of tie cell showed very little .^tra'n, although 

 no special precautions, such as annealing, were taken. By 

 avoiding temperature-differences when putting the com- 

 ponents into optical contact, and by subsequent slow cooling 

 from the annealing temperature, all traces of strain can be 

 eliminated. 



The internal surfaces of the cell-windows were parallel to 

 1 second both before and after union . The degree of parallelism 

 remained quite unaltered, and it therefore appears that the 

 accuracy of parallelism is determined only by the initial 

 accuracy of the central component. 



Union of Two Different Kinds of Glass. — The components 

 of the ceils described above were ad made of one kind of 

 glass, and the joining of different glasses is complicated 

 by at least three factors — differing expansion coefficients, 

 differing annealing-points of the glasses, and the degree of 

 ease with which they will fuse together in the blowpipe-flame. 

 The first will certainly cause the surfaces to come apart while 

 the temperature is rising, but if they are under pressure at 

 about '60 lb. per square inch this should not greatly matter, 

 as they will go into contact when the temperature is constant 

 at its highest limit, and they will then join perfectly. Strain 



