J'.i-j Ml!. C. T. R. WILSON ON THE KKFICIHNCV AS O>N1>KNSAT1<>.\ 



the expansion apparatus were larger than before, the former being 37 centims. in 

 internal diameter. The arrangements for connecting tin- air-space "nelow the piston 

 with the vacuum in F have been improved in form, and a self-acting iiuliarubl)er 

 valve, V, has l>een substituted for the stop-cock, formerly used to cut off communi- 

 cation with tin- atmosphere immediately before making this connection. The addition 

 of this valve was found to be a great convenience. 



The method of arranging for an expansion of any desired amount is described in 

 the paper to which reference has just been made. In the new appvratus, however, an 

 error is introduced into the measurements of expansion by the yielding of the india- 

 rubl)er stoppers closing the top of the expansion cylinder and the bottom of the 

 vessel, A, in which the clouds are observed, as well as to some extent prolxibly by the 

 momentum acquired by the air in the connecting tube. Both sources of error tend to 

 make the actual maximum expansion of the air in A greater than that obtained by 

 calculation from its pressure before and after expansion. 



Now the older experiments,* made with apparatus suitable for absolute measure- 

 ments, showed that there are two well-defined critical values of the expansion, which 

 we may use as fixed points to standardise an expansion apparatus unsuited for 

 absolute measurements. At the first point the expansion (measured by v-,/t-,, the 

 ratio of the final to the initial volume) is equal to 1/25 ; it is the minimum expansion 

 required to make condensation in the form of drops begin in dust-free air initially 

 saturated, whether the air be exposed to an ionising agent, such as Rontgen rays, or 

 not. At the second point the transition from rain-like to cloud-like condensation (in 

 the absence of ionising agents) takes place; v t /v lt is equal here to T38. 



On testing the apparatus in this way the following results were obtained : 



(1.) The air being exposed to Rontgen rays, condensation was first observed when 

 the apparent value of v-Jvi = 1/22 instead of 1'25. Error = '03. 



(2.) The change from rain-like to cloud-like condensation (in the absence of ionising 

 agents) took place when the apparent value of v 2 /v t = 1'35 instead of 1'38. 

 Error = '03. 



Thus the correction to be added to the apparent values of r.j/r, is the same at both 

 jxnuts and equal to + '03. I have, therefore, assumed that the same correction holds 

 for intermediate values .of v z /Vi. Throughout the paper the corrected values of 

 v.j/t'i are given. 



The vessel, A, in which the ions were produced and the clouds formed upon them 

 observed, consisted of a wide glass tube, 4 '2 centims. in diameter, closed above by a 

 brass plate, cemented to the ground top of the tube with sealing wax ; 2 centims. 

 below this plate was a smaller circular brass plate, 37 centims. in diameter, fixed 

 horizontally within the tube by means of three projecting tags cemented to the sides 

 of the tube. In the side of the tube a horizontal slot was cut, 3 centims. long and 

 3*5 millims. in diameter, the lower edge of the slot being on a level with the upper 



* 'Phil. Trims.,' A, vol. 189, ]>. -JC5, 1897. 



