EASTMAN: FISHES FROM UPPER EOCENE OP MONTE BOLCA. 323 



" The taste for minute prey to be sought by poking about in odd corners may 

 have determined some of the peculiar specialization of the Sea-horse order. We 

 may imagine these to have continually sought smaller and smaller food until the 

 tiny particles came to be sucked up by the elongate muzzle. After probably 

 passing through a stage somewhat like Syngnathus, but less eel-like, the ancestral 

 Sea-horse did not need the quick-darting form of the body to capture its food or to 

 escape enemies; hence the fan-like tail was suppressed (in Hippocampus), and the 

 rapidly vibrating pectoral and dorsal fins enabled the fish to poise, humming-bird 

 fashion, while sucking food through the tubular beak." (Ann. N. Y. Acad. Sci., 

 Vol. XVII, 1907, p. 495). 



Without proceeding further into the question of origins, we wish to lay stress 

 on the fact that among sticklebacks two distinct lines of specialization are traceable 

 as regards the conformation of the region between the orbit and the mouth. Pro- 

 gressive modification in one direction leads to the pushing forward of this whole 

 region, the jaw parts being carried along in this facial elongation and the mouth 

 retaining its terminal position. An evolutionary series of Gasterosteiformes, 

 showing gradual transformation with respect to these characters, was first worked 

 out by Dr. Theodore N. Gill. And it is to be noted that at the same time that the 

 snout was becoming elongated into a slender tube, scales over the body were 

 becoming progressively superseded by dermal armor. The armament is first 

 indicated in the form of bony scutes arranged in rows along the back and fianks, 

 and finally culminates in the cuirass of Amphisile, which is fused with the enlarged 

 ribs and other portions of the endoskeleton (c/. Gregory, I. c, p. 493). 



The second line of progressive modification culminates in the formation of a 

 rostral beak recalling that found in certain Xiphiiformes ( Histiophorus, Blochius, 

 etc.), though the mouth is situated ventrally and but little in advance of the orbits. 

 This series may also be supposed to begin with Gasterosteus or its immediate pro- 

 totype, and leads through stages which are not recorded in paleontology up to the 

 longirostrate type of which Rhamphosus is the only known example. 



Now the interesting thing to note is that this second evolutionary series is 

 not only specialized in the direction of acquiring a sword-like rostrum, but it also 

 exhibits the unfolding of characters which are progressively displayed in the parallel 

 evolutionary series. For convenience we may distinguish these as (1) the tube- 

 snout, and (2), the rostrate series, both having Gasterosteus as a common starting- 

 point. 



The evolutionary changes that have taken place appear to have proceeded in 

 the following manner. A generalized or " synthetic type," to use Agassiz's phrase, 



