REPORT OF THE KEW COMMITTEE. 



H 



riments made at Kew upon the first Barograph. These experiments consisted 

 in subjecting the instrument to a very considerable range of temperature 

 artificially produced, while frequent comparisons of its indications with those 

 of a Standard Barometer gave the means of determining approximately 

 what ought to be the position of the fulcrum. It may be presumed that the 

 determination thus arrived at cannot be wrong more than one-tenth of the 

 whole, and assuming this to be the case, the next point is to find what is the 

 actual daily temperature range at the various observatories. 



The following Table exhibits both the mean and the maximum daily range 

 for each mouth for each of the seven observatories up to the end of 1868. In 

 all these, with the exception of tStonyhurst, a night observation is made of 

 the temperature of the Barograph at 10 o'clock, but the result will show that 

 in Stouyhurst such an observation is unnecessary. It ought here to be borne 

 in mind that from the system adopted in these instruments, namely, constant 

 reference each day to a standard, it is only the daily range of temperature 

 that we have to consider. 



Daily range of temperature, in degrees Fahrenheit, as given by the 



observation hours. 



From the res.ults of this Table it would appear that, assuming the tempera- 

 ture adjustment to be one-tenth wrong, the greatest error introduced from 

 this cause into any of the observations during the year 1868 would be about 

 0-0024 in., while the mean monthly error would bo inappreciable in all 

 cases. 



"We may therefore with confidence presume that in these Barographs the 

 method of tabulation exemplified in the Report for 1867 and now practised is 

 sufiiciently accurate to obviate all effects of changes of temperature, and that it 

 is unnecessary to resort to that more complicated but perfect system of reduc- 

 tion alluded to in the same Report, by which the influence of temperature 

 may be completely eliminated. The near correspondence between the simul- 

 taneous Standard and Barograph readings, as exhibited in page Ivii of this 

 Report, is another proof that the temperatiu-e correction is practically perfect. 



8lu(/gishness of Mercurij. — As the Barograph tube is always in perfect 



c?2 



