130 COLLIGATWE RELATIONS AND SCIENTIFIC LAWS 



soluble, destroyed by corrosive acids, etc." Apparently a definition of 

 what we mean by "metals," this is a genuine colligative relation hav- 

 ing the form of the sulfur relation: If a substance has several of the 

 following properties . . . , theri probably it is a metal and has the 

 other properties. This relation antedates the rise of systematic chem- 

 istry, and survived for millennia without theoretic accommodation. 

 Early in the 18th century the phlogiston theory proposed that metals 

 be viewed as compound bodies with one component, phlogiston, 

 common to all of them. The relation is then rationalized: the presence 

 of the common component "explains" the occurrence of properties 

 common to all. Toward the end of the 18th century, however, the 

 ad\^ent of Lavoisier's new chemical theory destroyed the phlogiston 

 theory. In Lavoisier's system all the metals are distinct elements, ap- 

 parently having nothing in common. The relation is then no longer 

 rationalized, but of course it survives the loss of its theoretic ac- 

 commodation. 



For more than a century the relation of metals survived, amply 

 useful but an enigma. Only in the last half century has it acquired 

 what we conceive to be an enduring rationalization. Metals, we now 

 suppose, do contain something in common: the company of labile 

 electrons that constitute the metallic bond. Once again we "under- 

 stand " the relation— indeed if we read "electrons" for "phlogiston" 

 we understand it, and many other such relations, in much the same 

 way they were "understood" in the phlogiston theory. But through 

 all such clouds of theoretic speculation the relation itself looms up 

 and endures solid as a rock. 



Colligative Relations not Wholly 

 Independent of Theories 



So far in this chapter, and particularly in the last section, we have 

 considered colligative relations as completely independent of their 

 theoretic affiliations. This simplification facilitates discussion, but 

 must not be pressed too far. Typically, a colligative relation is doubly 

 connected. By the denotations of its conceptual terms it is linked 

 widi experience; by its accommodation in a theory it is linked de- 

 ductively to the postulates thereof. Either link may be formed first. 

 The relation may first materialize as an empirical regularity, as did 

 Boyle's law, and only subsequently acquire theoretic accommoda- 



