PREFACE. xli 
This is only one of many instances of an attempt to 
reduce effects to causes. 
We cannot explain every thing, in the present state 
of knowledge it is indispensable that effects should be 
referred to causes only when there is a manifest and a 
steady relation between two points of structure, which 
may be made to enter into the explanation. As 
To pressure for instance, many things are ascribed 
which are not due to it, and many things both our 
ignorance and knowledge are interested in ascribing to 
unknown causes. 
A confession of ignorance cannot lead to error, 
while a subtle hypothesis especially by an eminent 
Savant may not only be false, but is very apt to check 
enquiry, from the assumption of knowledge contained 
in it. | 
