THE BOOK OF MIGRATORY BIRDS 117 



fly with such an addition of weight before its wings, which 

 would throw it out of the centre of gravity. We see the 

 heron and many other birds obliged to extend their legs 

 behind and contract their necks when flying, in order to 

 balance themselves on the wing. Seven quarts of water 

 (the quantity mentioned by Dr. Douglas) are nearly equal 

 to fourteen pounds weight, and certainly more than the 

 bird could carry in that situation.'* 



It would appear, indeed, from the observations of Sir 

 Everard Home, that Montagu's objections are valid, for 

 in the Adjutant (Ciconia argala), which has a bag precisely 

 similar, he found that it contained "nothing but air, which 

 the bird has the power of expelling and filling the bag 

 again at pleasure." In the adjutant the bag communi- 

 cates with the large air-cells at the back of the neck, and 

 therefore we may fairly conclude it is intended to render 

 the birds light and buoyant for running, since they are too 

 heavy to fly without considerable difficulty. 



These birds are remarkable for the length of their legs, 

 which must be very advantageous for swiftness of 

 running ; but it would be wrong to infer as a general 

 principle that all birds with long legs are swift-footed. 

 On the contrary, the Wading Birds (Grallatores), which 

 have proportionately much longer legs than the ostrich or 

 the bustard, are not well adapted for walking on land. 

 Amongst these the Flamingo (Phcenicopterus ruber) is 

 one of the longest legged birds; yet it is in this respect 

 far exceeded by the Stilt (Himantopus melanopterus), and 

 the legs in the latter are, besides, slender, and even "so 

 flexible," as Wilson says of the American stilt, "that they 

 may be bent considerably without danger of breaking," 

 as if, in accordance with Pliny's name (Himantopus), they 

 had been cut out of a thong of leather. 



The reasoning of naturalists, indeed, respecting the con- 

 formation of the feet of birds is, when not derived from 

 living specimens, as frequently wrong as right. It has 

 been usual, for example, since the time, if I mistake not, 



