208 



Prof. Silas W. Holman on the Effect of 



EXPEEIMENTAL BeSULTS. 



The results will be here given as they were for convenience 

 classified during the progress of the work, viz. with dry air in 

 five series, with carbonic acid in two series. The method of 

 reduction of the results is given at page 204, and in my first 

 paper. 



Air. 



First Series. — This consists of the twenty-one measurements 

 made in 1876, as given at page 49 of the article referred to *, 



and gives as a mean result of the ten measurements of i 



as 



Vi 



ith 



shown in Table I. below, the value y= — =1*234, with an 



average deviation f of 0*044, or about 4 per cent. A glance 

 at the page referred to will show, however, that of these ten 

 measurements but six were at the temperatures of 0° and 100° 

 respectively. These six, as shown in Table II., give the mean 



«= 5^0=1-270+ ad. 0*005; 



and this mean I shall use in comparing this series with the 

 later ones, omitting those results obtained at temperatures 

 between 0° and 100°, because the want of certainty in the 

 temperature-measurements would render the labour of the 

 necessary computations fruitless. 



Table I. Table II. 



No. 



2*. 



n 2 



d. 



2 ... 



1-083 



-•151 



7 ... 



1-212 



-•022 



8 ... 



1-206 



-•028 



9 ... 



1-215 



-•019 



11 ... 



1-272 



+•038 



12 ... 



1-267 



+•033 



13 ... 



1-271 



+•037 



14 .. 



1-273 



+ •039 



18 .. 



1-277 



+•043 



21 .. 



1-259 



+•025 





1-234 



0-044 



No. 



Vi, 



d. 



11 ... 



1-272 



+ •002 



12 ... 



1-267 



-•003 



13 ... 



1-271 



+ •001 



14 ... 



1-273 



+ •003 



18 ... 



1-277 



+ •007 



21 ... 



1-259 



-•011 





1-270 



0005 



Second Series. April 1878. — In making a study of the best 

 forms of apparatus, a careful trial was given to capillaries in 



* Phil. Mag. iii. p. 85 (1877). 



•j- Let a v a 2 , a 3y ... an be a series of measurements of the same quantity, 



and 

 a 3 



mean 



let m 



be their arithmetical mean : then a 



m=d v a 



[ '=d„ 



, = d s , &c. will be the deviation of these measurements from their 

 , and ^ + ^+g? 3 +. •■+<?» w iu be the average deviation if, in the 



summation of the numerator, the values of d lf d % , &c. be taken arithme- 

 tically, without regard to algebraic sign. 



