362 HENRY A. ROWLAND 



were crooked, we could yet turn it into such a position as to nearly 

 eliminate error. 



It was used at a distance of about 110 cm. from the object, and no 

 difficulty was found after practice in setting it on the column to j\ mm. 

 at least. The cross hairs made an angle of 45 with the horizontal, as 

 this was found to be the most sensitive arrangement. 



The scale was carefully calibrated, and the relative errors c for the 

 portion used were determined for every centimeter, the portion of the 

 scale between the and 100 points of the air thermometer being 

 assumed correct. There is no object in determining the absolute value 

 of the scale, but it should agree reasonably well with that on the 

 barometer; for let H , H t , and H 1QO be the readings of the barometer, 

 and Ti , h t , and /t 100 the readings of the cathetometer at the temperatures 

 denoted by the subscript. Then approximately 



(.#100 + /? 100 ) (fft> + ^o) ^100 HQ + h lw A 



As the height of the barometer varies only very slightly during an 

 experiment, the value of this expression is very nearly 



"100 "0 



which does not depend on the absolute value of the scale divisions. 



But the best manner of testing a cathetometer is to take readings 

 upon an accurate scale placed near the mercury columns to be meas- 

 ured. I tried this with my instrument, and found that it agreed with 

 the scale to within two or three one-hundredths of a millimeter, which 

 was as near as I could read on such an object. 



In conclusion, every care was taken to eliminate the errors of this 

 instrument, as the possibility of such errors was constantly present in 

 my mind; and it is supposed that the instrumental errors did not 

 amount to more than one or two one-hundredths of a millimeter on the 

 mercury column. The proof of this will be shown in the results 

 obtained. 



The Barometer 



This was of the form designed by Fortin, and was made by James 

 Green of New York. The tube was 2-0 cm. diameter nearly on the 

 outside, and about 1-7 cm. on the inside. The correction for capillarity 

 is therefore almost inappreciable, especially as, when it remains con- 



6 These amounted to less than -016mm. at any part. 



