4 PROCEEDINGS OP THE NATIONAL MUSEUM vol.67 



the tongue ends in two main branches, tipped with heavy spines 

 many times larger than the marginal spines. In this type of tongue 

 there is a definite line of demarcation between the fleshy body of the 

 tongue and the translucent cornified tip which is frayed. 



In studying such a tongue for factors that are constant, one is 

 forced to conclude that beyond the general shape and appearance 

 there is nothing that can be accepted as invariable. The number 

 of posterior spines is inconstant within the species, although they 

 may be counted on to be in a single row (in contradistinction to the 

 multiple rows as seen in owls, for example). 



The length depends on the amount of wear. In a series of meadow- 

 lark tongues collected in one day in South Dakota the length varies 

 from 16 mm. to 20.5 mm. The birds were feeding almost entirely on 

 grasshoj)pers that were at that time a serious pest. 



The main posterolateral or heaAdest spines are not invariable 

 in arrangement ; while always present they may be bifid or in birds, 

 as some of the sparrows, where they are normally split into two, 

 there may be three or four subdivisions. 



The curling, splitting, and fraying is also variable within the 

 species and shows individual modifications, although, as will be seen 

 later, these characters serve as very important adaptive features and 

 undergo extensive variations in certain families. 



Bearing in mind these inconstant factors it is of interest to trace 

 the modifications that may be found of this fundamental pattern. 



With slight differences in curling, splitting, length, and arrange- 

 ment of spines this tongue is to be found in a large number of pas- 

 serine birds, as the warblers, vireos, thrushes, thrashers, crows, fly- 

 catchers, shrikes, wrens, bulbuls, drongos, and the like, with Glareola 

 closely simulating it. The divergence from the type, however, is 

 most marked and comes to its greatest development in the flower- 

 frequenting forms. 



The typical tongue has an inherent tendency to curl, split, and 

 fray, and any one or all of these tendencies may be combined to make 

 up the tongues of the flower frequenters. 



Thus splitting alone with little tendency to curl and no fraying 

 is exemplified by the tongue of the floAverpeckers or Dicaeidae, which 

 is deeply split, forming very slender long forked tips, two in Dicaeum 

 and four in Prionochilus. 



On the other hand marked curling is seen in the Old World sun- 

 birds (Nectariniidae), where it may be a complete tube for the 

 greater part of its length, without fraying of the margins of the 

 tube and with splitting into two tips either absent or very slight. 

 Whether the tongue be a relatively short one, as in Hevniotimia (fig. 

 141), or very long as in Arachnothera^ this perfect tubular arrange- 

 ment exists in the anterior two-thirds of each. Splitting is not 



