NATURAL SCIENCES OF PHILADELPHIA. 59 



understood in my method of treating them, I will state explicitly 

 that I am governed, on this occasion, by the following arbitrary 

 rules or definitions : 



I. I predicate "species" upon specimens presenting any defi- 

 nite, constant, tangible characters whatsoever, that do not, so far 

 as it appears, grade into the characters of other specimens. 



II. I predicate "varieties" upon specimens presenting indefinite 

 and inconstant yet tangible characters that are seen to grade into 

 the characters of other specimens. 



III. I predicate "synonyms" upon specimens presenting in- 

 definite, inconstant, and intangible characters due to individual 

 peculiarities, or to age, sex, season, or locality; as well as upon 

 specimens presenting no special characters at all. 



My present belief is, that there are only four 1 forms of Myi- 

 archus that do not intergrade, and that are differentiated from 

 a common original stock to such degree, or in such manner, that 

 we cannot account for their respective peculiarities according to 

 highly probable laws of geographical variation depending upon 

 differences in food, climate, etc.; but I cannot here enter upon 

 debatable ground. According to the foregoing practically con- 

 venient if not very scientific rules, I find that the specimens ex- 

 amined represent nine species, two of which each present three 

 tangible varieties. 3 



Before proceeding to handle these species and varieties I will 

 recount several propositions that should receive due consideration. 



a. The normal inherent variability in size, of the whole bird 

 and of ifs several members, is at least 12 per cent, of the mean. 

 (This is independent of all extraneous influences.) 



b. Size varies in direct ratio with the latitude of the breeding 

 place. 



c. Size of peripheral parts, as compared with total size, varies 



1 One is the series comprising crinitus, validus, cinerascens, tyrannulus 

 and pTimcepTialus ; the second is lawrencii and nigriceps ; the third, sto- 

 lidus; fourth, tristis. 



2 In defining these varieties, as I do beyond, I must not be taken as 

 meaning that the characters assigned always apply in full force: that 

 would be prima facie evidence of species. On the contrary, I describe the 

 extreme phase of variation, which shades by insensible steps into the 

 "typical" condition of the species. 



1872.] 



