282 X REVIEW OP THE TRILOBITES. 



torn of classification very commonly adopted, is founded on the num- 

 ber of the thoracic segments or body-rings. This was first proposed 

 by Professor Quenstedt, of Tubingen, in 1837. Of the importance 

 of this character there can be no doubt, more especially when we 

 take into consideration its constancy throughout one entire group of 

 crustaceans (the malacostraca), and the results of Barrandk's re- 

 searches, shewing its definite nature with respect to adult trilobitic 

 forms. It can only be looked upon, however, as possessing a specific 

 value ; for there are several well-known types which differ from one 

 another in the number of the body-rings, but which can be readily 

 shewn to be generically alike. Hence, by the adoption of this sys- 

 tem, without regard to other characters, many unnatural separations 

 necessarily arise. 



In Barrande's classification, the trilobites are arranged in two 

 sections, each comprising various families. These sections are found- 

 ed respectively on the presence of raised or furrowed pleurae, a cha- 

 racter often of difficult recognition even in perfect specimens (illoenus 

 nileus), and one that appears at the best to be of questionable 

 value. The divisions founded upon it, like all indeed based upon a 

 single character, break through many natural analogies, and place in 

 distant parts of the system, forms which are evidently akin to one 

 another. A single example may suffice to corroborate this assertion. 

 Respecting the existence of a close relationship between phacops and 

 ceraurus, there can be, I think, but one opinion. The peculiar 

 direction of the facial suture ; the anterior expansion of the glabel- 

 la ; the (at least in normal cases) eleven thoracic segments — and 

 other characters — render this sufficiently evident. But in Bar- 

 rande's system, the two are placed widely apart. At the same time 

 it is not intern] ed to deny that ceraurus is also related to acidaspis 

 (with which it is placed by Barrande). It holds undoubtedly a mid- 

 dle place between phacope and thi3 latter genus, and such is the or- 

 der in which it occurs in the classification given below; whereas, by 

 adopting Barrande's subdivisions, various unrelated genera would 

 necessarily intervene. Between bronteus and illanus again, evidently 

 allied types if regard be paid to all their characters, a wide separa- 

 tion occurs in Barrande's system. 



In the classification now proposed, the trilobites (omitting the ag- 

 nosti) are arranged in thirteen families. Some of these divisions 

 might be thought perhaps, on a first consideration of the subject, to 

 possess simply a generic value ; but their adoption as true families 

 may be sustained, I believe, on really satisfactory grounds. If cer- 

 tain of them exhibit but few genera at present, that need be no ob- 



