Behaviour of Metallic Vapours in Flames. 



27 



to compensate for the decreased diffusion, and consequent 

 narrowing of the streak at high pressure, the effective size 

 of the loop was increased by turning its plane towards the 

 horizontal at high pressure, it being vertical at atmospheric 

 pressure. 



The results are shown in fig. 4, where the energy radiated 



Fig. 4. 



1 



■3 











X 



2 









**5 



--" 



1 





-f X 



o,V 



£ 





Mass vaporized (mg. per hour)-^ 



(obtained from the pyrometer readings in the way mentioned) 

 is plotted against the weight of SrO (in mgs.) vaporized 

 per hour. The energy is in arbitrary units, as only relative 

 numbers can be obtained. The crosses refer to the measure- 

 ments made at 1 atmosphere, the circles to those made at 

 4 atmospheres. The inaccuracies of measurement with such 

 small weights of salt are necessarily large (10 to 20 per 

 cent.), and unfortunately for the larger weights, where we 

 hope for increased accuracy in the weighings, the radiated 

 energy varies very rapidly with the pyrometer reading, 

 introducing compensating uncertainty. We may, however, 

 from these experiments safely conclude that at fourfold 

 pressure the light-energy radiated by a given amount of 

 vaporized salt is not changed by 25 per cent. ; it is probable 

 that it does not vary with the pressure. Guoy * has found 

 that for small amounts of salt the luminosity is proportional 

 to the amount of salt vaporized. A straight line is 

 therefore drawn through the origin and the mean of the 

 points, to enable us to judge roughly what weight of metal 



* M. Guoy, Ann. de Chim. et Phys. (5) xviii. p. 1 (1870). 



