34© TRANS. ST. LOUIS ACAD. SCIENCE. 



It will be observed that changing the leveling screws to the 

 utmost limit, which is about one inch, changes the position of the 

 needle half a minute. As we have always taken the precaution 

 to level up the tripod with a good hand level, the leveling screws 

 were not changed sufficiently to introduce any appreciable error. 

 The effect of the side screws turns out to be the most serious, and 

 the effect of all the brass-work is to diminish the easterly declina- 

 tion 2'. 3, and this quantity is added to all determinations where 

 magnet Cg was used during 1880. Previous to that time no 

 correction can be made, as was explained in (he former report 

 (Joe. cit.) 



A similar determination was made with magnet No. i in Decl. 

 No. 3, correction being also made for hourly change. Seven de- 

 terminations showed that the effect of the brass-work of Decl. 3 

 upon magnet No. i, is to throw the magnet to the west 1.04 scale 

 divisions, or I'.gy. The observed declinations made with magnet 

 No. I in Declinometer No. 3, for 18S0, have therefore received 

 a correction of +2'.o- 



INTENSITY. 



The observations of the summer were made in the same man- 

 ner as before, excepting that we found it of great advantage to 

 read watch time, by observing the second-hand with a good 

 pocket lens. The watch lost 2.^ seconds per day during the sum- 

 mer, and correction for rate was therefore omitted. 



The temperature constant q was taken as o.Qoo^g, as deter- 

 mined in 1878.* 



The following table gives the calculation of the value of P for 

 the summer. 



* Trans, vol. iv. No. i, pp. 88-90. 



