o 



2S 



CLASSIFICATION OF FISHES 



as to be nearly five times the length of the body : the 

 pre-opercule is serrated, but in all other parts of their 

 structure the Paradise-fish betray the closest affinity to 

 the mullets, 



(29.) The MuLLiD.E, or surmullets^ have always been 

 placed close to the TriglidcE. or gurnards, probably on ac- 

 count of a certain resemblance in the physiognomy 0^ 

 the two ; for when we compare their structure, no two 

 gi'oups can hardly be more dissimilar. In the sur- 

 mullets the scales are large, deciduous^ and smooth ; in 

 the gurnards they are small^ tenacious, and rough : the 

 head in one is compressed, in the other depressed. 

 Nevertheless, the distant resemblance between them is 

 of much importance when viewed as a relation of 

 analogy only ; and such we may safely pronounce it. 

 The surmullets, in fact, are clearly the chironectiform 

 type of the tribe, representing the TriglidcB by their large 



head, vertical eyes, and small 

 mouth ; and the Gadidce by 

 the fleshy cirrus or barbels on 

 the lower j aw (^fig. 4.). The 

 species are fev/, and consti- 

 tute two genera only. The 

 surmullets, hke the Holo- 

 centrincB, are. nearly all of 

 different shades of red, varied 

 with yellow stripes : it was the death of these fishes, so 

 common in the ^lediterranean, which gave delight to 

 the effeminate and luxurious RomanSj in the closing 

 years of their power. 



(30.) The last, and not the least interesting, of the 

 aberrant families of the spine-finned order, is that of the 

 SpiROBRANCHiD.a:, which corresponds to Cuvier's '^laby- 

 rinthiform pharyngeals." This is perhaps the most 

 natural of all the groups characterised by that admir- 

 able naturalist ; and this not merely by possessing certain 

 peculiarities of internal structure found in no other fishes, 

 but also because they are equally distinct in their exter- 

 nal conformation ; while they present a beautiful grada- 



