218 CLASSIFICATION OF FISHES. 



son is to the sharks ? No one will dispute this. What, 

 therefore, is the logical deduction, but that the Petro- 

 myzonidce should he naturally arranged with those fishes 

 with which, in their general organisation, they have the 

 nearest resemblance ? The structure of the branchial 

 apertures, indeed, in this group, are so variable, that 

 they absolutely become hardly sufficient to characterise 

 a genus ; much less to determine an order. Among the 

 Gymnarchidce, or sea eels, for instance, they are close to- 

 gether and under the throat in Sphagebranchus ; united 

 almost into one in Monopterus ; single, and round, in 

 Synbranchus ; before the pectorals in Gymnotus, and be- 

 hind them in Murcena : in short, they almost assume 

 every possible form and situation within the limits of a 

 few genera, — a clear proof how completely secondary these 

 characters become in the present group. It would seem, 

 indeed, that Nature, upon leaving the annulose circle, 

 and entering that of the fish, intended to show us all the 

 forms of variation in the first group, which she after- 

 wards employs to characterise higher divisions : this 

 she has done in the class Acrita, as Mr. MacLeay has so 

 beautifully illustrated * ; and the same remarks may be 

 made applicable to the group before us : the least or- 

 ganised of all the fishes, as the Acrita are the least 

 among animals. 



(192.) But there is another inference to be drawn 

 from the peculiar structure* of the lamprey, sufficiently 

 important to merit a separate notice. In a former vo- 

 lume, we have stated the proposition that one of the 

 primary laws of the circular succession of all groups is, 

 that the three aberrant divisions constitute a circle of 

 their own, independent of their connection to the other 

 two ; from which it follows, that the primary circles in 

 every group (when that group is perfect in all its parts) 

 are three, although they appear to be five. We have 

 shown that this is more than probable in the union of 

 the monocardian animals ; and even Cuvier confesses 

 the affinity of certain of the cartilaginous order to rep- 



* Hor. Entom. See also Classif. of Animals. 



