PRITCHAKD S WEDGE PHOTOMETER. 



311 



It is obvious from a comparison of the numbers (c) in this table with those in 

 Table I. that there is a very close agreement between the results of the two days, 

 and that the variations in the ti-ansmissibility for a unit thickness are dependent on 

 the numbers of unit thickness observed which are systematic and not accidental. 

 To make this fact still more evident, we add in Table II a. the results of the nine 

 preceding series, reduced to a uniform quantity of heat, that is, with the varied 

 atmospheric absorption, etc. between the series eliminated. 



TABLE II a. 



The above are the measurements of transmission of the total solar beam by 

 the Pritchard wedge, each series being reduced separately. 



Taking the mean of all the observations in Tables I. and II., we have for the 

 final values : 



It will be observed that the probable errors above given represent, not only 

 those peculiar to the apparatus, but, what is much more important, the effect of 

 all changes in the sky during some hours' observation. If each series were separately 

 reduced to the form (J), the probable error of the mean would be much less. 



