Ionization from Certain Hot Salts. 241 



earlier part of this paper, gave very erratic results. It was 

 felt that this might be caused by the drop of potential along 

 the molybdenite strip owing to its higli resistance. This 

 makes the value of Y to be substituted in the formula for 

 e\m uncertain even when, as in most of the experiments, the 

 potential is applied at the middle of the strip by means of a 

 sliding contact to an external high resistance shunt. Even 

 in this case the correct effective potential will only be that 

 at the centre of the strip provided the emission is symmetrical 

 about the centre. This difficulty might have been overcome 

 by measuring the magnetic deflexions with the heating- 

 currents successively in opposite directions, but this method 

 was not found to be practicable. The methods finally 

 adopted were (1) heating the molybdenite strip directly by 

 an alternating current, and (2) heating the molybdenite on 

 an auxiliary strip of platinum. 



On searching for a positive emission from the strip a very 

 slight amount was observed in all cases when the strip was 

 new at a point considerably below the break. This was 

 transient and could not be measured. Permanent emission 

 commenced just below or just at the start of the break, and 

 it is here that measurements were commenced. The sub- 

 stance at red heat is very active in its emission, so that as 

 the temperature was increased the sensitiveness of the 

 electrometer bad to be decreased by adding capacities to 

 the pair of quadrants in order to get anything like accurate 

 readings. Except in a few instances the balance method 

 was the only one by which good results were obtained, as 

 the substance was too erratic in its behaviour to give at all 

 satisfactory curves. Other repeated attempts by the slit 

 method failed, principally for the reason that constancy of 

 resistance, and hence temperature, could not be maintained. 

 Even when the temperature apparently remained constant 

 there was often a sudden or a gradual shift in the location 

 of the maxima. The measurements were all made in a 

 mercury vacuum, with the McLeod gauge showing less than 

 5 X 10 ~ 5 mm. of mercury. 



In the measurements of e/m the centre of the molybdenite 

 strip was maintained at a fixed potential V with reference 

 to the opposite plates by means of a connexion from a high 

 potential battery to a point between two megohm resistances 

 in series shunted across the strip outside the apparatus. The 

 strip was then heated by an alternating current obtained by 

 the commutation of a direct current. The use of an alter- 

 nating current makes the effective potential entering into 



Phil. Mag. S. 6. Yol. .33. No. 195. March 1917. R 



