THEOR 7 OF NA TUBA L SELECTION. 21 7 



alone. But how to obtain the beginning of such useful 

 development?'^ In answer, it may be asked, why should 

 not the early progenitors of the whales with baleen have 

 possessed a mouth constructed something like the laniel- 

 lated beak of a duck? Ducks, like whales, subsist by sift- 

 ing the mud and water; and the family has sometimes 

 been called Onblatores, or sifters. I hope that I may not 

 be misconstrued into saying that the progenitors of whales 

 did actually possess mouths lamellated like the beak of a 

 duck. I wish only to show that this is not incredible, and 

 that the immense plates of baleen in the Greenland whale 

 might have been developed from such lamellae by finely 

 graduated steps, each of service to its possessor. 



The beak of a shoveller-duck (Spatula clypeata) is a 

 more beautiful and complex structure than the mouth of a 

 whale. The upper mandible is furnished on each side (in 

 the specimen examined by me) with a row or comb formed 

 of 188 thin, elastio lamellae, obliquely bevelled so as to be 

 pointed, and placed transversely to the longer axis of the 

 mouth. They arise from the palate, and are attached by 

 flexible membrane to the sides of the mandible. Those 

 standing toward the middle are the longest, being about 

 one-third of an inch in length, and they project fourteen 

 one-hundreths of an inch beneath the edge. At their 

 bases there is a short subsidiary row of obliquely transverse 

 lamellae. In these several respects they resemble the plates 

 of baleen in the mouth of a v/hale. But toward the ex- 

 tremity of the beak they differ much, as they project in- 

 ward, instead of straight downward. The entire head of 

 the shoveller, though incomparably less bulky, is about one- 

 eighteenth of the length of the head of a moderately large 

 Balaenoptera rostrata, in which species the baleen is only 

 nine inches long; so that if we were to make the head of 

 the shoveller as long as that of the Balaenoptera, the lam- 

 ellae would be six inches in length, that is, two-thirds of 

 the length of the baleen in this species of whale. The 

 lower mandible of the shoveller-duck is furnished with 

 lamellae of equal length with these above, but liner; and in 

 being thus furnished it differs conspicuously from the 

 lower jaw of a whale, which is destitute of baleen. On 

 the other hand, the extremities of these lower lamella3 are 

 frayed into fine bristly points, so that they thus curiously- 



