Introduction: Variation. 75 



yielding pliancy of the tip making it difficult to remove the web from this 

 point without cutting off the end of the feather altogether. In Nature any feather 

 of sufficient stiffness, prolonged so as to stand out beyond the other feathers, will 

 be liable to such a process as this, attrition against the twigs of trees, the walls 

 of their nesting holes etc. , supplying the place of the knife. Assumed that the 

 two middle tail-feathers of Prioniturus were originally a little longer than the 

 rest'), the ends, if sufficiently prolonged, are liable to attrition; and a narrowing 

 of the tips, such as is now seen in the young birds (pi. V, fig. 1), will result. 

 The friction at the ends of the feathers causes irritation to the roots; an in- 

 creased supply of blood ensues there, with the result of an increased size of these 

 feathers. These longer feathers are more liable to attrition, and half- formed 

 rackets (pi. V, fig. 3) take shape; the increased irritation and consequent 

 lengthening of the feather results in the production of other stages (pi. V, fig. 4), 

 up to the most advanced development of the present time (pi. VI, fig. 1). Yet 

 the striking features shown in the plates were not obtained in one generation, 

 as has been proved; on the other hand this appears to have been a process 

 of ages, more and more advanced results being obtained in successive 

 generations and transmitted by heredity. The simplest stages of this for- 

 mation are displayed by young birds in first plumage which in respect of the 

 tail probably resemble the first ancestors of the genus (pi. V, fig. J ) ; the second 

 moult, when the webs are often qiiite absent on the shafts of the rackets, which 

 are about half the full length in old birds, seems to show a later period in the 

 history of the race ; while the highest development of these feathers, as seen in 

 old birds (especially males) of the present day, is probably the most recent stage 

 in the evolution of the genus. 



The following are the arguments in proof that these rackets are the inherited 

 effects of attrition: 



1 . It has been shown that such can easily be formed artificially by scraping, 

 the size of the spatule depending upon the stiffness of the feather. 



2. Where the shafts are not exposed to attrition they are not bare. 

 It is only on the projecting part of the middle tail-feathers that the shafts are 

 bare; and as far as the ends of the lateral feathers, by which the middle ones 

 are protected from attrition, they are fully webbed. If the bareness were due 

 to something else, it might' be expected that the naked shaft would not in 

 every species^) arise just at this point of the tail, but sometimes much higher 

 up, or sometimes much lower down^). 



3. Rackets do not occur on unexposed feathers sheltered from attrition. 



•) A very common condition in birds. 



2) We have examined P. luconensis, cyaneiceps, suluensis, discums, flavicans, platitrus, and the plate of 



t^cvt'tc (t ll s 



3 Genera in which the racket-feathers are longer and consequently heavier (e. g. Bhringa] usually have 

 larger spatules and the attenuation of the webs on the shaft continued towards the base of the tail — a result 

 of friction upon the other feathers. 



10* 



