On Mr Wallace's TJieory of Birds' Nests 2'j() 



are consistent and closely connected with " blind instinct." In as 

 far as the nature of a bird's nest is determined by its own structure, 

 and by hereditary habit, and by surrounding conditions, in so far 

 it is determined by conditions over which the bird itself has no 

 control, and which belong to it as part of its very constitution and 

 nature. 



We now come to the first sentence which foreshadows the 

 coming theory. Specifying the structure of a bird, and its en- 

 vironment as to the most important elements in determining its 

 kind of nest, Mr ^^^allace proceeds thus : — 



"If, therefore, we find less important and more easily modified characters than 

 these correlated with peculiarities of nidification, we shall be justified in con- 

 cluding that the former are dependent on the latter, and not vice versa." 



The obscure wording of this sentence makes it rather a hard one 

 to construe or to follow : but, taken with the context, I believe the 

 argument to be this — When two things, or two sets of things, are 

 correlated together, the one being more fixed or less changeable in 

 its nature than the other, we may conclude that the most change- 

 able is " dependent on " the least changeable (as on its physical 

 cause ?). The reasoning, then, as apj^lied to the question in hand, 

 may be stated thus : — 



' ' The structure and habitat of birds we find to be correlated with certain 

 peculiarities in their nesting ; but sti'ucture and habitat are both comparatively 

 fixed and difficult of change ; the peculiarities of nests are therefore dependent 

 on the peculiarities of structure and of habitat in birds. But if, on the same 

 principle, we can find any other circumstance about birds which also is correlated 

 with peculiarities of nesting, but which is more easily capable of change, then 

 we may conclude that this circumstance is one dependent on the nature of the 

 bird's nest and not vice versa." 



The fallacies which lie hid in this argument are about as nume- 

 rous as the words which it contains. In the first place, the defini- 

 tion of the peculiarities which are selected as correlated together may 

 be altogether fanciful and arbitrary; in the second place, things which 

 are really correlated together, whether always or only in general, 

 may have no "dependence on" each other as physical cause and 

 effect, but may be and often are the result of some cause or causes 

 which lie above and behind them both ; in the third place, the 

 changeability of any peculiar character may be an assumption 

 as arbitrary as the definition and conception of the peculiarity 

 itself. 



To test these fallacies we may take a case. Sandpipers, 



