BY DR. E. P. RAMSAY AND J. DOUGLAS OGILBY. 13 



The free part of the tail is much depressed and compressed, 

 forming an almost perfect square, its height being equal to or 

 but little more than its breadth immediately before the base of 

 the caudal fin, and all four sides being slightly concave. On each 

 side of the posterior part of the tail there are two strong keels, 

 which converge behind. 



The scales are set in regular oblique rows, each of which is 

 slightly curved backward at its dorsal extremity, and strongly 

 forwards on the lower half of the body, thus forming a gracefully 

 sigmoidal band. They are small and firmly adherent laterally, 

 each scale slightly overlapping that one which lies next below 

 it on its own row, and to a less extent its corresponding scale on 

 the succeeding row. Each scale bears seven or eight visible 

 longitudinal keels, and each of these is again armed with four or 

 five denticles, and terminates in a strong spine. 



The lateral line is slightly curved along its anterior half, and 

 consists of one hundred and twelve to one hundred and fifteen 

 scales, which are slightly raised above those adjoining them, and 

 about twenty of which are minute and lie between the caudal 

 keels.* The transverse series which rises at the base of the first 

 dorsal ray bears thirty scales, counting obliquely forwards to the 

 middle of the abdomen, ten of these being above the lateral line. 



The color is of an uniform dark brown : the roof of the mouth 

 is Vjlack, the inside of the jaws bright yellow. 



Before comparing these two specimens with the descriptions of 

 previous authors, and with our own Mediterranean example, we 

 must call particular attention to the fact that the two above 

 described are small, measuring but a fraction over seven inches 

 each, while Lowe's was nine and a quarter, and the specimen in 

 spirits examined by Dr. Giinther in 18G1 twelve and a half 

 inches, our example from the same locality as that last mentioned 

 being thirteen and a half inches, and it is possible that these 

 differences in size may account for the undoubtedly great dis- 

 crepancies in tlie comparative measurements. 



• Hence the ninety lateral line scalcH of Ix»wc and Mncleay. 



