THE CAUSATION OF DISEASE. 337 



sooner or later occur. Inasmuch, however, as the most stable E 

 does change from generation to generation, the species also 

 must of necessity undergo modifications in course of time, in 

 order to maintain the standard of adaptation ; yet, since changes 

 in E are for the most part slow and steady, they are compelled 

 to maintain a certain degree of fixity. They are, in fact, kept 

 within certain limits in accordance with the great biological 

 principle that equilibrium between S and E must be maintained. 

 But when we turn to pathological classes, do we find that 

 there is any principle at work tending in like manner to 

 keep them fixed ? On the contrary, pathological variations 

 place the individual out of harmony with his surroundings, 

 and, far from there being any factors at work tending to fix 

 and preserve such variations, there are strong forces for ever 

 crushing them out of existence — above all, that most potent 

 agency, Natural Selection. Hence the wonder is that diseases 

 admit of any classification except of a very crude order, and 

 that any tendency whatever to definite types is manifested ; for 

 it must be allowed that there is such a tendency, although we 

 should be greatly erring in regarding the types as fixed and un- 

 alterable. The explanation is to be sought for in the fact that phy- 

 siological species are necessarily kept within certain limits, and 

 consequently all individuals in a species are more or less alike. 

 The S's, then, being more or less alike, and the various mal-E's 

 having a certain degree of stability in them, it follows that 

 diseases tend towards similarity — i.e., towards definite types. 

 It is, however, only permissible to assume fixed types for pur- 

 poses of classification, and with the limitation that we allow for 

 wide deviations from such types. Unless we are thoroughly 

 alive to the extent of the possible deviations, and are, more- 

 over, willing to acknowlege that in many cases they render 

 classification utterly impossible, we shall be constantly falling 

 into error. 



The following two hypothetical cases will illustrate this 

 statement. We will first suppose the S to be constant, the E 

 varying ; and, secondly, the specific mal-E to be constant, 

 the S varying. 



I . Where the S is constant, the JE varying. — Even suppos- 

 ing all individuals to be exactly alike, it would still be 



