EXPERIMENTAL KNOWLEDGE OF THE PBOPERTIES OF MATTER. 509 



It is noticeable, too, that tlie errors fall in both directions. 



Objections by MM. Bartoli and Stracciati. 



In the course of a study of the physical properties of American petro- 

 leum (from Pennsylvania) MM. Bartoli and Stracciati ' criticise the pro- 

 positions of MendelejefF on the volumes of liquids and their expansion by 

 heat, and the deductions of Thorpe and Riicker from Van der Waals' and 

 Mendelejeff's conclusions. 



Mendelejeff's ' very interesting memoir ' is resolved by these authors 

 into three propositions : — 



(A) Any t-wo liquids -which dilate equally between (say) 0° and 10° 

 ■will dilate equally between 0° and t° ; and if the dilatation' of one liquid 

 is greater or less than the other liquid between 0° and 10°, it -will also be 

 greater or less between 0° and t°. 



(B) Let Vq, Vi, Vo) • . • • , "V„ be the volumes of a liquid at tem- 

 perature 0°, ^,, <2) • • • -i^nj then, if another liquid has the same volumes 

 at temperatures 0°, i/, (2', . . . . , /„', -we shall have — 



^i Co 



Js'- 



(C) The volume V, of a liquid at temperature t is expressed (in terms 

 of the volume at 0° as unit) by the equation Y,=l/1 — A:;. 



They proceed to say that proposition (A) can only be taken as an 

 approximation, and that ' it is easy to see ' that it is deduced from the 

 f oi'mula of Van der Waals. 



Now Mendelejeff says,^ 'Just as for gases the expression V=l + ^-f 

 applies only to a first approximation, or to a so-called ideal gas, so does 

 the expression y={l — Jd) "^ apply for liquids only to a first approximation, 

 namely, to 'ideal liquids,' giving in a foot-note his idea of the characteristics 

 of an ideal liquid. Mendelejeff has not claimed more for his formula and 

 his general conclusions than that they are approximate. 



Again, 'it is easy to see ' that Mendelejeff's result, as embodied in A, is 

 deduced from the formula of Van der "Waals may mean that it is not easy 

 to sJioiv ; if these aathoi's had shown this they -would have given a theore- 

 tical confirmation of a result of Mendelejeff's experience. 



Again, they object that if proposition (B) is true, then in the formula 

 ■which expresses accurate determinations of expansions of many liquids, viz., 

 Y^=l + at + U- + ct^ + clt* we shoxddhsire a=b/ a =c/b=d/c=k ; relations 

 ■which are not verified : this would be a fatal objection if Mendelejeff 



' Ann. Ckem. Ph. (G), 7, 1886 o. 384. 



= a S. J. Trans. 45, 1884, p. 131. 



