Mr. F. A. Bather on British Fossil Crinoids. 385 



Series 2. Scytalecrinites. Decadocrinus, Scytale- 

 crinus. 



Scries 3. Graphiocrinites. Graphiocrinus, Phialo- 

 crinus } Ceriocrinus ' (Bursacrinus } Synypho- 

 crinus) ?. 



Series 4. — (a). Erisocrinites. Erisocrinus,Stemmat- 

 ocrinus. 



(b). Encrinites. Dadocrinus, Encrinus. 



Series 5. Cromyocrinites. Eupachycrinus, Tribrach- 

 iocrinus } Cromyocrinus, Agassizocrinus. 



Fam. 5. Cyathocrinid^e. 



Series 1. Cyathocrinites. Cyathocrinus, Strepto- 

 crinus, Aracknocrinus, Gissocrinus, Lecytho- 

 crinus. 



Series 2. Codiacrinites. Codiacrinus, Lecythio- 

 crinus. 



Series 3. Achradocrinites. Achradocrinus } Hypo- 

 crin us. 



Fam. 6. Belemnocrinid^e. Belemnocrinus, Holocrinus. 



Incertae sedis — Edriocrinus. 



It seems unnecessary to indicate the diagnostic characters 

 of these families and series more clearly than has been 

 already done. The more natural a classification is, the more 

 difficult must be any attempt at definition : as knowledge 

 grows, delimitations vanish. Not that this classification is 

 either natural or complete; of its imperfections in those 

 directions no one can be more sensible than myself. The 

 facts, indeed, on which the argument has been based have 

 been verified as far as possible ; the labour of collecting them, 

 which would otherwise have been too great, has been lightened 

 by the previous work of Messrs. Wachsmuth and Springer, to 

 whom all students of palaeozoic crinoids owe a debt of grati- 

 tude. That I differ from these authors is due partly to differ- 

 ence in the interpretation of the facts, partly to a difference of 

 principle. The principles that govern the above classification, 

 though not yet universally accepted, are steadily gaining 

 ground among biologists. It is in the interpretation of the 

 facts that the real difficulty lies, and here I am as likely as 

 not to be wrong. Increased knowledge alone can solve our 

 problems. The proposed classification should be regarded as 



