222 MESSRS. W. R. BOUSFIELD AND W. ERIC BOUSFIELD 



of moisture through the holes in the obturator as the vapour pressure rises. This 

 uncertainty seemed to show that the method, involving as it does a free surface of 

 water, is not likely to yield accurate results beyond 80 C. In a point-to-pomt 

 experiment the rise of temperature from 75 C. to 80 C. takes 5 minutes, so that, 

 untuning an uncertainty of 0'05 C. in the rate of rise of temperature per hour there 

 would thence arise an uncertainty of 0'005 C. in the temperature interval from 55 C. 

 to 80 C., i.e., an uncertainty of 1 in 5000. 



Obturator Heating in Cmtinuous-flow Experiments In the j continuous-flow 

 experiments a further element comes into play. In order that the cold water (at 

 13 C.) might flow through the obturator without sensible heating, the flow tube 

 was jacketed, as previously described, with a larger tube through which flowed water 

 at the same temperature in much larger volume. This cold-water jacket terminated 

 at a point near g (see fig. 3, p. 207) on a level with the bottom of the obturator, and 

 formed a small cold area on which steam condensed. To determine the total effect 

 of obturator heating with this new element, experiments were made at a temperature 

 of about 54 '5 C., similar to those before described, but with the new element of a 

 flow of cold water through the water jacket. Most of these experiments were made 

 after the two first series of continuous-flow experiments were concluded, as it was 

 not discovered till later that a deposition of moisture had taken place which varied in 

 each experiment. The earlier results were consequently erratic, and it was not till a 

 much later period that the cause was traced to the fact that the lower surface of the 

 obturator was so much cooled in the neighbourhood of the cold-water jacket as to 

 permit of the condensation of water upon it, the evaporation or fall of which 

 produced irregular results. 



It was therefore decided to make a third series of continuous-flow experiments. 

 In these the end of the inflow water-casing was surrounded by a short piece of 

 rubl>er tubing y l , which was shaped to fit it and which was cemented into the end 

 of the opening through the obturator with rubber solution. The under surface of 

 the obturator, just round the opening g, was also painted with a coat of rubber 

 solution. It was found that these precautions entirely prevented the deposition of 

 moisture on the obturator, and reduced the whole correction to so small an amount 

 that variations in its amount were of small importance. 



Experience showed that the circumstances which determined the amount of the 

 correction differed in different experiments from some obscure cause. . In the third 

 series of continuous-flow experiments the determination was therefore made at the 

 conclusion of each run, so that the apparatus was in the same condition in all respects 

 as during the run. 



As regards the first and second series of continuous-flow experiments, there were 

 fortunately two determinations of the correction which were also made at the end of 

 runs. The other determinations, which were made under conditions in which there 

 had been no accumulation of moisture on the obturator, were rejected. These two 



