102. Mr. W. Duddell : Znstrwments for the Measurement 
receiving wire in spark telegraphy, and thus permit some 
quantitative measurements to be made which are so much 
wanting in the published literature on this subject. 
The controlling force in the thermogalvanometer is ex- 
tremely small, and it seemed to me for some time a very 
drastic procedure to pivot such an instrument and apply to 
it ordinary controlling springs so as to develop a switch- 
board instrument from it. Measurements of the power 
obtainable from different designs of thermojunctions and 
heaters, led me to conclude that a pivotted instrument might 
be constructed. I have now made two such instruments. 
In order to save time these two first instruments were made 
by converting ordinary moving-coil direct-current instru- 
ments. So that it is probable that improvements will be 
effected when the instruments are manufactured from the 
commencement specially for the purpose. These two first 
instruments have turned out very successful. Both instru- 
ments have ordinary pointers about 10 em. (4 inches) long, 
and deflect through anangle of about 75° for their full scale- 
deflexion. No. 1 is cased in wood with a mirror under the 
pointer. No. 2 is iron-cased of the ordinary switchboard 
type. The works are the same except that one has a heater 
designed for a small current and the other for a small P.D. 
The data of the instruments is as follows :— 
Wo. I. No. 2. 
Current for full scale-deflexion 0°] ampere 2°0 amperes. 
Fee): § ee Hi 2 volts 0°15 volt, 
These instruments attain to within 1 per cent. of the 
proper deflexion in 10 seconds after starting the current, so 
that they are fairly quick in action. They have the property 
of reading slightly too high (about 4 per cent.) if the current 
be suddenly brought to its full value, when the instrument is 
quite cold and has been out of use for some time, the reading 
decreasing to within 4 per cent. of the proper value during . 
the first 30 to 40 seconds after starting the current. 
The creep of zero is small; in an ordinary room it is 
practically nil. If the back of the iron-cased instrument 
where the thermo-junction is situated is heated in front of the 
fire till hot to the touch while the front remains quite cold, 
a zero creep of about 1° can be produced. If this great 
difference in temperature should occur in practice, which is 
highly improbable, it will produce an error in measuring the 
full current of under 1 per cent. 
The power required to work the instrument, namely 0:2 
to 0'3 watt, is comparatively low for pivotted thermal instru- 
ments. The low current, 0°1 ampere, taken by No. 1, and 
