.~>70 DR EDWIN H. BARTON ON THE TEMPERATURE VARIATION 



galvanometer remained in the same place throughout, and with its controlling magnet 

 at a definite height. The diurnal variations of the earth's magnetic field were considered 

 negligible. 



12. Thermo-electric Effects. — Seventh. When the magnetite was strongly heated 

 considerable trouble was at first experienced owing to slight fluctuations in the tempera- 

 ture of the secondary coil encircling the specimen. Indeed, the variations thus set up 

 in the thermo-electric current flowing round the secondary circuit often caused an 

 irregular movement of the galvanometer needle greater than the kicks to be obtained 

 by induction. 



Advantage was then taken of the fact that, although the time-integral of the thermo- 

 current often equalled or exceeded the time-integral of the induced current, yet the 

 E.M.F. of the induced current was much the greater of the two. Hence, any method 

 of exposing the galvanometer to this double effect for a space of time so short as to be 

 at all comparable with the duration of the induction phenomena would practically 

 eliminate the disturbance due to the thermo-current. To accomplish this, some 

 automatic double-contact maker was needed. Professor C. V. Boys, F.R.S., then 

 suggested the use of a pendulum working on a geometrical hinge, and carrying two 

 platinum points passing through troughs containing mercury. The use of this pen- 

 dulum, however, showed that the high temperatures at first contemplated would not 

 be required, for at about 600° C. the permeability of the magnetite was found to have 

 already become unity. It was, therefore, not necessary to explore the permeability of 

 the specimen beyond this point. Moreover, for the comparatively moderate tempera- 

 tures needed it was possible so far to moderate the variations in the thermo-currents as 

 to make the use of the Polil's reversing commutator quite admissible. This was effected 

 by covering the coils P and S, fig. 1, with asbestos wool and cloths and carefully 

 excluding all draughts of air. The reversing commutator was accordingly used through- 

 out the observations in preference to the pendulum, which, in the simple form rigged 

 up, only made or broke the current. 



13. Magnetite unequally heated. — Eighth. Errors might arise through non- 

 uniformity of temperature throughout the specimen of magnetite. 



These were to a great extent obviated by turning on the gas of the stove to a certain 

 amount, and leaving the coils and specimen exposed to this constant flame for some 

 time. The temperature was read at intervals by the thermo-couple and potentiometer ; 

 and only when further rise of temperature had almost ceased was a set of readings for 

 the kick taken. 



14. Magnetite crumbles on heating. — Of the original specimen of magnetite about 

 half crumbled away on the first heating. 



As it is difficult to get large pieces of magnetite which can bear exposure to high 

 temperatures, the remnant from the first piece was next tried. This small piece bore, 

 without any sign of damage, repeated heatings, and was employed throughout the 

 determination for the kick at various temperatures. 



