Iln' Ui'chiaii l>i,<char(H' from a Point to a Plane. fi21 



In exnuHiiintr by means of an expansion apparatus the 

 ionization ])ro(liK'ed in a gas by the simple ionizing agents, 

 siK'li as Kiintgen and Becqiierel rays, and also in cases of 

 " s])ontaneous ionization," there is no difficulty in obtaining 

 regular and consistent results, even in the presence of con- 

 siderable impurities in the gas. But when more complicated 

 ionizing processes are employed, such as ultra-violet light 

 or the discharge here investigated, there appear irregularities 

 and inconsistencies which render it at first sight hopeless to 

 obtain a definite result. Slight impurities in the gas, or 

 some other accidents whose nature it is difficult to ascertain, 

 exercise an effi^ct on the results altogether out of pro- 

 portion to their own magnitude. But after some weeks of 

 work it was found that the experiments could be divided 

 into two classes : — 



(1) Regular, 



(2) Irregular. 



(An " experiment " includes all the observations taken with- 

 out refilling the apparatus with gas or changing the point 

 from which the discharge took place.) In the "" regular '* 

 experiments a perfectly definite value could be assigned to 

 the minimum expansion requisite to give a fog with a given 

 discharge : the value remained constant however often the 

 observations were made, in whatever order they were made, 

 or however long a time elapsed between successive observa- 

 tions : the minimum expansion was invariable. It was also 

 found that these regular experiments agreed with each other. 

 For the same gas, the same point, and the same sign of the 

 point, all regular experiments gave the same result. Ex~ 

 jteriments which were regular were also consistent. 



It was entirely different with the "irregular^' experi- 

 ments. Here no definite value could be assigned to the 

 expansion required to give a fog ; of several expansions of 

 the same amount made in succession without stopping the 

 discharge, some would give a fog, some would not. At other 

 times it was thought that definite results had been attained, 

 but on waiting a few minutes they could not be repeated. 

 These irregular clouds were always obtained with expansions 

 less, usually much less, than the minimum value found in 

 the regular experiments. It was occasionally found that on 

 leaving the apparatus? for a few hours without refilling 

 Avith gas, a series \\hich had been irregular would become 

 regular. A change from regular to irregular results was 

 never observed except on the introduction c)f a new point 

 or a new portion of gas. 



