1863.] 691 



as the instruments have been left standing with different amounts of 

 torsion on the fibres at different times, the zero of torsion may have 

 changed. This cause would principally affect the Portable Electro- 

 meter, whose zero of torsion has always been assumed to be 30, ex- 

 cept on May 2nd, when it was very carefully tested, and found to be 

 30'l. The zero of torsion for the Station Electrometer has been so 

 frequently tested, that no sensible correction can be needed for it 

 beyond what has been made. That the Portable Electrometer is 

 affected by some such disturbing cause will be seen by reference to 

 the comparisons on February llth, where the earth-readings of the 

 "Portable" are 51-0, 51-8, 51-2, the increase from the first reading to 

 the second apparently depending on the increased torsion which was 

 applied in the interim, and the subsequent fall being assisted by the 

 diminution which was made in the torsion between the second read- 

 ing and the third. The " Portable" also seems to hang a little (see 

 comparisons on November 4th). 



16. Besides errors from want of exactness in reducing all the 

 observations, except those taken with Station II., to units of that 

 instrument, and those arising from changes in the Electrometers 

 themselves, it is probable that others have been produced by circum- 

 stances affecting the collection of the electricity from the air, the 

 place where the water-stream breaks into drops being subject to 

 small variations, depending on the wind and the head of water in the 

 can. In like manner, when burning matches are used, the place 

 where the smoke breaks away depends upon the wind and the length 

 of the match, the latter being sometimes six and sometimes only 

 three inches at the commencement of an observation. Rain or snow 

 falling on the pipe has also doubtless a disturbing effect, equivalen 

 to a shortening of the pipe ; but the effect cannot be great ; for I 

 have tested that during a shower of rain, when the water in the can 

 was not allowed to flow out, a strong artificial charge given to the 

 can and pipe, though of the opposite kind to that of the rain, was 

 dissipated with extreme slowness ; but when the water was turned on 

 and a fresh charge given, the loss was extremely rapid say thirty 

 times as fast as before. This observation has been confirmed by 

 observing that even when the electricity of the air during heavy snow 

 or rain was excessively strong, as collected by water-dropping, very 

 feeble indications were obtained if the water was turned off. It is 



