( 229 ) 



plane denotes a combination {6. rr), to which every time a certain 

 pair of substances will answer. 



In the said figure the line C'APB denotes the corresponding 

 values of 6 and rr from <9 = to 6» = 9,9. For C' <9 = 0, .t = 9, 

 for .4 6* = 1, rr = 7,5 ; with 6 — 2,22 corresponds n = 4,94. (Case 

 71 = 19' or a, = aj; for P rr = 6» = 2,89 (Case rr = ^ or b^ = b,); 

 forB <9 = 9,9, rr = 1. For values of ^ > 9,9 the double point would 

 lie on the side of the line v = b, where v <^b. It appears from the 

 figs. 23, 24 and 25 of the said paper, that then the line BB (-t ^ 1) 

 forms the line of demarcation between type I and II (III). For 

 starting from a point, where -t <^ 1 (however little) and ^ is com- 

 paratively low% wiiere therefore we are undoubtedly in region II (III), 

 we see clearly that we cannot leave this region, when with this 

 value of -T that of 6> is made to increase. For we can never pass 

 to type I, when not for realizable values of v (so <^b) a. double 

 point is reached, and now a simple consideration (see the paper 

 cited) teaches, that for n <^1 a double point would always answer 

 to a value of v <^b. 



Now it is clear that ^ = 0, .t = 9 is the same as ^ = oo, rr := ^Z,; 

 that 8 = :t = 2,89 is identical with 6 = n = 1/2,89 = 0,35 ; etc., etc. 

 (the two components have simply been interchanged), so that the 

 line CA' will correspond with the line C' A, wiiile A' B' corresponds 

 with AB. If we now consider only values of 6 which are ]> 1, if 

 in other words we always assume T, ^ T^, we may say that the 



16 



Proceedings Royal Acad. Amsterdam. Vol. IX. 



