PRODUCED BY HOT I'LATINUM IN DIFFERENT GASES. 



IS 



eliminated from the experiments which they represent, still an inspection of the actual 

 experimental numbers shows that such effects did occur. It was often observed, for 

 instance, that if the wire had l>een heated for a time to a temperature higher than 

 that at which olxservations were being taken, the value of the leak was too great for 

 some time afterwards. For instance, in an experiment at 816 C., where the pressure 

 was l - 4 millims. and the steady leak = 30, the unit being 6 x 10~" ampere, after 

 heating for a few minutes to about 1100 0. the wire was found to give the following 

 values of the leak at the times in minutes stated. 



It is evident from the above numbers that it took alxmt 20 minutes for equilibrium 

 to be established at this temperature, a fact which gives some indication of the 

 prodigious labour required in taking the observations. 



Another hysteretic effect, which is more likely to be a source of trouble in 

 experiments on the effect of change of pressure, is a time lag of the change in the 

 leak behind the change in the pressure. In testing for an effect of this kind it is 

 evident, from the preceding paragraph, that it is necessary to keep the temperature 

 constant while the pressure change is made. This is very difficult, since with a 

 constant current through the wire its temperature is a function of the pressure of 

 the surrounding gas, but by watching the galvanometer spot of the Wheatstone's 

 bridge circuit, and manipulating the rheostats which control the heating current as 

 quickly as possible, the temperature can be restrained from varying very much whilst 

 the pressure change is l>eing made. The following numbers represent the effect at 

 809 C. of suddenly reducing the pressure from T8 to 0'21 millim., the P.D. being 

 + 80 volts. The steady leak at t'8 millims. had the value 18 (1 = T9 x 10~ 1S ampere) ; 

 the leak at 0'21 millim. had the values given at the times stated. 



In a similar way, on increasing the pressure, keeping the temperature constant, it 

 was found that the leak at the higher pressure was too small at first and only 

 gradually rose to its final steady value. 



It is evident from what has been said that the lag in the leak behind pressure 



