300 REPORT OF NATIONAL MUSEUM, 1890. 



Variations in the form of the tail. — The number of tail-feathers is 

 usually, perhaps always, ten. One species, the Marvellous Humming 

 Bird (Loddigesia mirabilis), is said to possess only four ; but, although 

 I have not seen the bird iu question, I can not help thinking that the 

 apparently absent rectrices are aborted and confounded with the 

 coverts.* Certainly in at least one other genus (Myrtis), such an abor- 

 tion of the middle pair of tail-feathers is very evident, these being so 

 much reduced in size as to be entirely concealed by the upper tail- 

 coverts, t On the other hand, another genus (Florisuga) seems to pos- 

 sess twelve tail-feathers ; but a close examination will show that what 

 are apparently the middle pair of rectrices are in reality the two longest 

 upper tail-coverts. 



Fig. 26.— Tail of Myrtis fanny, showing Fig. 27.— Tail af Florisuga mellivora, showing 

 aborted middle rectrices (x x), tbe specialized pair of uuper coverts (a a).. 



upper tail-coverts beina parted to ex- 

 pose tbem to view. 



The shape and development of the tail-feathers in the Humming Birds 

 vary to a degree that has no parallel among other birds, many of the 

 forms assumed being also entirely unique. Much the larger number 

 of species have the tail of a shape not very different from the ordinary 

 types among birds, that is to say, moderately rounded or forked, or 

 nearly even ; but besides tails of a normal shape there are deeply 

 forked or forficate (scissors shaped) tails, graduated or wedge-shaped 

 tails, double-rounded and double-emarginated tails, tails with streamers, 

 tails with racket- or paddle-shaped feathers, and tails whose shape can 

 not be designated by any special term. Putting aside those which 

 depart least from the ordinary shapes, the different types, with their 

 variations, are illustrated by the figures composing Plates xviii-xxiv. 



Variations in development of the tail-coverts. — Excepting the case of 

 Florisuga, referred to above, there are few notable modifications of the 

 tail-coverts. Two genera, however, may be mentioned which have these 

 feathers conspicuously developed, the one as to size, the other as to 



* la Acestrura miorura, according to Mr. Gould, all tbe rectrices are thus aborted, 

 the entire tail being concealed by tbe coverts. 



t Gould also cites as examples Thaumastura cora and Doricha enicura ; but in these 

 species the middle pair of rectrices while very short are not wholly hidden by the 

 coverts, as is the case in Myrtis fanny. (See Plate xxiv, Fig. 1, and Plate xxn, Fig. 1). 



