184 Annals of the Carnegie Museum. 



and sides of head. The natural inference would be, therefore, that 

 maculatus and virescens approximate each other in their characters! 

 Through the courtesy of Mr. Bangs I have before me the two adult 

 specimens from Martinique upon which Mr. Oberholser has based his 

 comparisons. They prove to be very poor skins, with necks unduly 

 stretched — a feature which makes the colors of the feathers of this 

 part appear less intense. The measurements (which I have confirmed), 

 it will be noted, are fully up to those of true virescens, and, although 

 the neck and sides of the head are slightly paler, it is true, than the 

 average specimen of that form, it is easy to match their colors in a 

 series. Even on the assumption that additional material from 

 Martinique would bear out the trifling differences indicated, formal 

 separation would seem scarcely justifiable. Admitting that Green 

 Herons from the West Indies (collectively considered) average slightly 

 smaller and paler than those from eastern North America, and are 

 thus possibly worthy of subspecific recognition, it is confusing to find 

 the only specimens from the type-locality of maculatus, the earliest 

 name available for the birds of this region, differing so little. Two 

 male examples from Porto Rico, also, are larger, instead of smaller, 

 than the average. In short, the variation in size seems so great, on 

 the whole, and the range in color so subtle and inconstant in character, 

 even in specimens from the same locality, that it is only provisionally, 

 and with great reluctance, that I here recognize the Cuban and Porto 

 Rican bird as distinct subspecifically. In any case, it is certainly far 

 less trenchantly defined than the Bahaman race, contrary to Mr. 

 Oberholser's implication. After a study of his paper it is difficult to 

 avoid the impression that he has carried subdivision too far. Surely 

 the use of trinomials, for which the American school of ornithologists 

 has contended so long, was never designed to cover such a case as 

 this. Even admitting that this is largely a matter of individual opin- 

 ion, it would nevertheless seem true as a general proposition that a 

 certain amount of variation ought to be allowed a given species 

 without thereby subjecting it to formal division. There are limits 

 in refinement beyond which it does not seem profitable to go, and while, 

 as previously remarked, I am not now in a position to go into further 

 details in this particular case, I suspect that the acquisition of fresh 

 material may eventually necessitate a revision of present conclusions.^ 



* In this connection I may add that Dr. Thomas Barbour has advised me that 

 it is absolutely certain, from repeated and extensive observations, that neither the 



