314 PALEONTOLOGY 



in the mental workings of an Aristotle, a Galen, a Harvey, 

 and a Cuvier, — to admit the instinctive impression of a design 

 or purpose in such structures as the valves of the vascular 

 system and the dioptric mechanism of the eye. In regard to 

 the few intellects, — they have ever been a small and unfruitful 

 minority, — who do not receive that impression and will not 

 admit the validity or existence of final causes in physiology, 

 the writer has elsewhere expressed his belief that such intel- 

 lects are not the higher and more normal examples, but rather 

 manifest some, perhaps congenital, defect of mind, allied or 

 analogous to "colour-blindness" through defect of the optic 

 nerve, or the inaudibleness of notes above a certain pitch 

 through defect of the acoustic nerve. 



The truth of a physiological knowledge of the condition 

 of a correlated structure, and of the application of that know- 

 ledge to palaeontology, is not affected by instances adduced 

 from that much more extensive series of coincident structures 

 of which the physiological condition is not yet known. Nor 

 is the power of the application of the physiologically interpreted 

 correlation the less certain because the merely empirically re- 

 cognized coincidences have failed to restore, with the same 

 certainty and to the same extent, an extinct form of animal. 



Certain coincidences of form and structure in animal bodies 

 are determined by observation. By the exercise of a higher 

 faculty the reason, or a reason, of these coincidences is dis- 

 covered, and they become correlations ; in other words, it is 

 known not only that they do exist, but how they are related 

 to each other. In the case of coincidences of the latter kind, 

 or of "correlations" properly so called, the mind infers with 

 greater certainty and confidence, in their application to a fossil, 

 than in the case of coincidences which are held to be constant 

 only because so many instances of them have been observed. 



Because the application of the latter kind of coincidences 

 is limited to the actual amount of observation at the period 



