The Food of Birds. 
93 
less completely. The methods of this spontaneous resto- 
ration of the unsettled balance of natural forces, are, of 
course, worthy of the most careful study. It is only by 
working in harmony with them that we ourselves can 
help to readjust the disturbed order. A fuller treatment 
of this matter may best be postponed until the general 
discussion of results obtained by the investigation. 
Enough has been said to show that the subject, although 
complicated and difficult, will richly repay the study nec- 
essary to its mastery. A full and accurate knowledge of 
the mutual relations of the various forms of organic life 
of a region, both normal and abnormal, is certainly 
quite as essential to the general welfare as a knowledge 
of the chemistry and geology of its soils, the peculiarities 
of its meteorology, or any other part of the inorganic en- 
vironment. 
Concerning the special subject of this paper the knowl- 
edge we need is such that we shall be able to afford for 
every species a tolerably correct answer to the questions, 
What would be the main consequences if this species were 
exterminated! if it were reduced to half its present num- 
bers! What if it, were doubled in number! if it were 
quadrupled! When this is known, we shall evidently be 
able to act wisely and with the best results. That these 
questions are not unanswerable, I shall undertake to 
prove by answering them in substance, for several spe- 
cies, in this paper, and by demonstrating the sufficient 
accuracy of the answers. 
Methods. 
Three methods are possible in determining the food of 
birds. The birds may be fed in confinement, and the kinds 
of food apparently preferred and the amount eaten may 
be noted. This evidently shows only what the bird will 
eat when restrained of its liberty, of such food as may be 
placed before it, and furnishes few data which we can use 
with safety in making up an account of its food in free- 
dom, when foraging for itself. The state of confinement 
is so abnormal for a bird that on this account, also, we 
can rarely reason from its habits in "that state to its or- 
dinary habits. This method is, therefore, available only 
