494 



SCIENCE. 



[N. S. Vol. XX. No. 511. 



bon atoms often exist in two isomeric forms 

 in which the sequence of the atoms is the 

 same. (5) Derivatives of ethylene often 

 exhibit a similar isomerism. 



Assuming as true that we have acquired 

 a knowledge of the sequence of atoms in 

 carbon compounds, the facts which I have 

 enumerated lead almost inevitably to the 

 corollary that the four atoms attached to 

 a given carbon atom are arranged in ap- 

 proximate symmetry around the center of 

 that atom for their position of most stable 

 equilibrium. The relation between this 

 conclusion and the theory of the sequence 

 of atoms in carbon compounds, or what is 

 ordinarily understood as structure, is very 

 similar to the relation between the atomic 

 theory and Avogadro's law. If we accept 

 the atomic theory, there seems to be no 

 rational escape from the acceptance of 

 Avogadro's law. In a similar manner, if 

 we accept the theory of the sequence of 

 atoms in carbon compounds, there seems no 

 reasonable possibility other than that van't 

 HofE's hypothesis is true in its broad out- 

 lines. 



I hope I may be pardoned here for a 

 brief digression. I am aware that Franz 

 Wald* believes that he can give a satisfac- 

 tory explanation of the laws of fixed and 

 multiple proportion and of combining 

 weights without the aid of the atomic 

 theory, and that Professor Ostwald in his 

 recent Faraday lecturef has accepted and 

 expanded the same thought. I will say 

 frankly that their reasoning does not ap- 

 pear to rae conclusive. Ostwald defines a 

 chemical individual as 'a body which can 

 form hylotropie phases within a finite 

 range of temperature and pressure, ';|: and 

 deduces from this the fact that a given ^ 

 hylotropie phase must have a fixed composi- 

 tion. He appears to forget that the ex- 



* ZtscTir. Phys. Chem., 24, 633, 1897. 

 t /. 0/iem. /Soc. (London) , 35, 506. 

 t76id., p. 515. 



istence of these hylotropie phases implies 

 that the properties of matter are discon- 

 tinuous, or, in other words, that there is a 

 finite number of hylotropie bodies, one of 

 the facts for which the atomic theory gives 

 an explanation. 



There is another characteristic, too, of a 

 chemical compound which all chemists will 

 agree is at least as important as that it 

 shall consist of a 'hylotropie phase.' This 

 is that the compound must not only have a 

 fixed composition, but this composition must 

 bear a definite relation to those numerical 

 quantities which represent the proportion 

 in which each element cf which it is com- 

 posed always combines with other elements. 

 I need hardly add that these numerical 

 quantities are so deeply seated in the prop- 

 erties of matter that, having adopted a 

 unit, all chemists are absolutely agreed in 

 selecting one and only one such quantity 

 for each of the well-known elements. 



In attempting to deduce this law of com- 

 bining weights Ostwald assumes that three 

 elements form the compounds AB, AC, BC 

 and ABC, and adds, ' There shall be but one 

 compound of every [each] kind.' With 

 this assumption, his reasoning may be 

 sound, but I fail to see how it applies when 

 we find ten thousand compounds ABC in- 

 stead of one. The case which he supposes 

 is so far theoretical that I have been unable 

 to find an actual case where the compound 

 ABC can be formed, by the union both of 

 AB with C and of AG with B* But I have 



* It is quite possible that such an illustration 

 may be found, but, in any case, Professor 

 Ostwald's deduction can not be made to apply 

 to those cases in which the compound ABC does 

 not exist, nor to those cases where the com- 

 pound ABC can not, even theoretically, be sup- 

 posed to consist in turn of a known compound 

 AB combined with C and of another known com- 

 pound AC combined with B. Such cases are com- 

 mon because of the fact of valence. In its 

 simplest form the law of combining weights is 

 quite independent of the existence of the com- 

 pound ABC and may be stated thus: If the com- 



