34 Prof. Threlfall and Messrs. Brearley and Allen. [May 24, 



resistance and capacity. The method of producing films of pure 

 sulphur between aluminium electrodes is also explained. It is neces- 

 sary to perform the melting, &c., in a gold vessel. 



Section 4 of the paper deals with the method of constructing gal- 

 vanometers of high sensitiveness and resistance. In order to observe 

 as small currents as possible, advantage was taken of every circum- 

 stance, both of observation and construction, likely to lead to 

 enhanced sensitiveness. For instance, instead of observing the 

 steady deflection, we habitually observed the throw of the needle on 

 reversal of the current through the galvanometer. The steady de- 

 flection was only observed as a check. 



The highest degree of sensitiveness we ever found it necessary to 

 use, was such that the throw on reversal was 1 micrometer division 

 for a current of 1'5 X 13~ 13 amperes, with a period of vibration of 

 about 25 seconds. 1 micrometer division is divisible into five parts, 

 so that the sensitiveness for least observable throw on reversal is 

 3 X 10~ u amperes. This sensitiveness, however, cannot be taken ad- 

 vantage of, except with very elaborate contact keys, and under rare 

 conditions of magnetic steadiness. We adopted the Kelvin type of 

 instrument. We consider that the problem of sensitive galvano- 

 meter building has not hitherto been approached in the proper 

 manner. Almost any first-rate instrument will give enormous sensi- 

 tiveness on occasion : but this sensitiveness is, in general, accom- 

 panied by instability, and is useless in practice, on account of zero 

 changes. The really important matter is to ensure the presence of 

 high sensitiveness with ease and certainty, not after hours of adjust- 

 ing, but immediately on the necessity arising ; in this we have been 

 perfectly successful. Success in this matter depends entirely on a 

 large number of details, for a discussion of which the paper must be 

 consulted. Exact drawings are also provided, both of the instrument 

 as a whole, and of the more important subsidiary parts. The follow- 

 ing notes must suffice here. 



1. It is essential that the coils shall be adjustable to the magnetic 

 system after the latter is mounted. 



2. Astaticism of sufficient perfection can only be secured by the 

 simultaneous magnetisation of all the members of the magnetic 

 system when they occupy their final relative positions. This necessi- 

 tates special appliances. 



3. If copper wire is used for the coils, no other metal must be 

 included in the circuit or connections of the instrument, otherwise 

 thermo-electric effects cannot be avoided. 



4. The instrument has four tiers of coils and magnets, whereby 

 improved electromagnetic conditions are obtained. 



5. The most important part of the instrument is that which 

 belongs to the adjusting of the magnetic control. This must be 



