F. TWYMAN 113 



To anneal glass, then, it is necessary to keep it within the annealing 

 range till the stresses have died out, and then to cool it with sufficient 

 slowness. No kind of heat treatment which does not raise a badly 

 annealed sample of glass to within this range will greatly affect its 

 condition of internal stress, w^hether for good or ill, in any reasonable 

 time. 



The method developed in our laboratory for determining the 

 annealing temperature will be found in the paper cited above. The 

 steps in the argument may be briefly summarised. 



The degree of annealing to be attained, and the time in which it 

 is to be accomplished must be defined. For glassware we have laid 

 down the condition that at the annealing temperature 95 per cent, 

 of the original stress must disappear in three minutes. For optical 

 glass appropriately modified stipulations are adopted. 



The case of the disappearance of stress in a viscous body was con- 

 sidered by Maxwell,* who gave an exponential expression applicable 

 to such cases connecting stress with time, thus 



F-ESe-^ 

 where S is a distortion or strain of some kind produced in the body 



by displacement, F is the stress thus excited, E is the co-efficient 

 of elasticity for that particular kind of strain, t the time, and 

 T a time named by Maxwell the time of relaxation, which depends 

 on the nature of the body. 



The product ET he calls the co-efficient of viscosity, since in the 

 case of steady progressive strain or distortion produced by constant 

 stress the rate of strain multiplied by this product gives the stress. 



It is obvious, then, that by defining the annealing temperature 

 in the way we do, we have at the same time defined a viscosity. 

 All we have to do, then, is to find the temperature at which the 

 glass has the viscosity so defined, and we have the annealing tem- 

 perature. 



For details of apparatus and method the paper mentioned above 

 must be referred to; but one point may be of interest. 



It was found by us that in the neighbourhood of the annealing 

 range most glasses examined double in mobility for every 8° C rise 

 of temperature, approximately. If then an attempt were made to 

 anneal at 500° C. a glass whose annealing temperature (as defined 

 above) is 580° C, the glass would require to be left one thousand 

 times as long in the former case as in the latter. 



The Eesearch Laboratory, 

 Adam Hilger, Ltd. 



* Phil. Mag. S4, Vol. 35, Feb., 1868, p. i2g. 



