282 THE EVOLUTION OF SCIEXTIFIC THEORIES 



ease of the tobacco plant, which is not so. This is awkward, but so 

 long as there is only one disease, and so long as there is, in a broad 

 sense, only one method of study, nothing much is going to be done 

 about it. As soon as other cases are found, and as soon as other 

 methods find them, active work starts up again, and the great chances 

 are that it will have new directions which may set the main stream of 

 research for quite a long time. 



Here is initiated a new endeavor, with its own characteristic orienta- 

 tions and potentialities. Here is born a new theory— probably rather 

 crude and narrow, but promising— to enter upon the cycle I have 

 sketched. Sometimes, as in the case of the virus theory, the new 

 theory will merely supplement the old. But when instead it bids fair 

 to supplant the old, the crisis of the old order is at hand. 



The death struggle. The challenge of a new theory at once im^ests 

 with new significance anomalies perhaps long disguised, or dismissed 

 as unimportant. Precisely those anomalies, having in most cases sup- 

 plied the primary stimulus to creation of the new theory, are likely 

 to be highlighted at the center of its focal region. For the older 

 theory to lea\'e them still unexplained becomes then an overt mark 

 of its insufficiency. Yet such explanation will ordinarily require one 

 or more supplementary ad hoc postidates. The need for these will, if 

 the new theory pro\'es viable, be ever increased. That theory will lead 

 us to a multitude of new phenomena. These remain still to be dis- 

 covered precisely because the older theory has never gi\'en any hint 

 of their possibility, and these are consequently just the discoveries 

 likely to tax most cruelly that theory's correlative capacity. Now we 

 see wide before us the way in which a heuristically powerful theory 

 will, in quite short order, completely re\'erse a balance of correlati\'e 

 efficiency and explanatory appeal initially unfavorable to it. 



Every time Mendeleev left a space in his periodic table he had to 

 postulate an undiscovered element. W\\h the subsequent discovery 

 of those elements the ad hoc postulates are completely eliminated. 

 Such new discoveries, the correction of apparently contradictory 

 data {e.g., Priestley's "stubborn facts"), and the kind of conceptual 

 refinements earlier noted will, all together, tend substantially to re- 

 duce the body of postulates required to constitute the new theory. 

 Beyond such reductions in the denominator, the correlative index of 

 that theory will be still further increased by complementary changes 

 in the numerator. To the extent that some of the original exclusions 



