EASTMAN: FISHES FROM UPPER EOCENE OF MONTE BOLCA. 321 



There remains to be considered the genus Rhamphosus, which is placed by 

 A. Smith Woodward among the Centriscidse, and made by Gill the type of an inde- 

 pendent family. Only two species are known, R. rastrum (Volta) and R. biserratus 

 Bassani, both from the Eocene of Monte Bolca and both very rare. Nearly all 

 writers who have noticed this genus have recognized its close agreement in the 

 majority of structural characters with the modern Centriscus and Amphisile, as 

 these terms are commonly used (not, however, in the sense employed by Jordan and 

 Gilbert, Proc. U. S. Nat. Mus., 1883, Vol. V., p. 575). The differences which it 

 presents were pointed out in the first instance by Agassiz, with the exception of 

 one very important feature. He failed to emphasize the fact that in the fossil 

 form the mouth is not borne at the end of an elongate, tubiform snout. And yet, 

 at the very close of his diagnosis of the genus he writes: " Les machoires s'ouvrent 

 peu et sont placees immediatement au dessous de I'orbite." 



This observation of Agassiz, which is undoubtedly correct, appears to have 

 been overlooked by subsequent writers, all of whom ascribed to Rhamphosus, either 

 directly or by implication, a character which it does not possess, namely, that of 

 having a " snout produced in a long tube, with small, terminal, toothless mouth." 

 In reality the condition is very different from that which is common throughout 

 the order, and resembles that occurring in modern sword- and sail-fishes, or in the 

 extinct Aspidorhynchus, Hemirhynchus, Blochius, &c., where the snout is produced 

 in a sharp, spear-like rostrum. These are all forms in which a prominent beak 

 results from a forward extension of the upper jaw only, but a parallel modification 

 is found in the " Half-Beaks " or Hemiramphs, in which it is the lower jaw only 

 that is produced. A still different modification is that observed in the African 

 family of Mormyrids, where the pore-like mouth is at the extremity of a long, taper- 

 ing proboscis. Regarding the latter group the following remark by Boulenger is 

 of interest to us in the present connection: 



" Some species of Mormyrops show how a form like Gymnarchus may have 

 evolved out of a more typically-formed fish. Nothing is more striking than the 

 variation in the shape of the snout within one and the same genus, and the names 

 given to some of the species {ovis, caballus, elephas, taniandua, numenius, ibis) are 

 suggestive of resemblance with the heads of various animals." (Cambridge 

 Natural History, Fishes, p. 550.) Similar modifications of the snout are to be 

 observed in the family Gymnotidse. 



Just as a series of stages in the formation of a tubiform snout can be traced in 

 the sticklebacks leading from Gasterosteus through Aulorhynchus up to the flute- 

 mouths, so in the same manner a series is traceable from the non-elongate snout of 



