214 SEXUAL SELECTION: BIRDS. [Part II. 



the tropics, among forests that never lose their foliage, that 

 we find whole groups of birds whose chief color is green." 

 It will be admitted by every one, who has ever tried, how 

 difficult it is to distinguish parrots in a leaf-covered tree. 

 Nevertheless, we must remember that many parrots are 

 ornamented with crimson, blue, and orange tints, which 

 can hardly be protective. Woodpeckers are eminently 

 arboreal, but, besides green species, there are many black 

 and black-and-white kinds — all the species being appar- 

 ently exposed to nearly the same dangers. It is there- 

 fore probable that strongly-pronounced colors have been 

 acquired by tree-haunting birds through sexual selection, 

 but that green tints have had an advantage through nat- 

 ural selection over other colors for the sake of protection. 

 In regard to birds which live on the ground, every one 

 admits that they are colored so as to imitate the surround- 

 ing surface. How difficult it is to see a partridge, snipe, 

 woodcock, certain plovers, larks, and night-jars, when 

 crouched on the ground ! Animals inhabiting deserts offer 

 the most striking instances, for the bare surface affords no 

 concealment, and all the smaller quadrupeds, reptiles, and 

 birds, depend for safety on their colors. As Mr. Tristram 

 has remarked, 51 in regard to the inhabitants of the Saha- 

 ra, all are protected by their " isabelline or sand-color." 

 Calling to my recollection the desert-birds which I had 

 iseen in South America, as well as most of the ground- 

 birds in Great Britain, it appeared to me that both sexes 

 in such cases are generally colored nearly alike. Ac- 

 cordingly, I applied to Mr. Tristram, with respect to the 

 birds of the Sahara, and he has kindly given me the fol- 

 lowing information: There are twenty-six species, belong- 

 ing to fifteen genera, which manifestly have had their 

 YDluniage colored in a protective manner ; and this coloring 

 is all the more striking, as with most of these birds it is 



51 'Ibis,' 1859, vol. i. p. 429, el seq. 



