2 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 65 



The consummation of such an arrangement is not always an easy 

 task, for we too commonly fall into the error of over-estimating the 

 comparative value of, and thereby placing too much reliance upon, 

 some single obvious or exaggerated character, instead of taking into 

 consideration and carefully weighing all of the characters presented. 

 Thus we are prone to place types distinguished by some unique 

 and phylogenetically aberrant feature, though not otherwise remark- 

 able, ahead of others which, more conservative throughout, are except 

 for this single feature more advanced. 



The recent crinoids offer a good illustration of the many difficul- 

 ties in the path of a logical phylogenetic arrangement. The sequence 

 of the families now commonly accepted is, beginning with the most 

 specialized, as follows : 



Order Articulata 

 Pentacrinitidc-e (including the Pentacrinitida 



and the Comatulida) 

 Apiocrinidse 

 Phrynocrinidse 

 Bourgueticrinidas 

 Holopodidae 



Order Inadunata 

 Plicatocrinidje 

 This sequence has been determined not by an exhaustive study of 

 the characters of each type and a subsequent comparison based upon 

 the results of such a study, but rather by a more or less fortunate 

 application of the doctrine of probabilities, based upon general resem- 

 blances. 



It is the aim of the present paper to analyze all of the characters 

 employed in the differentiation of the larger groups of recent cri- 

 noids, and, on the basis of this analysis, to indicate the true linear 

 phylogenetic interrelationships of the recent types. 



THE DETERMINATION OF THE PHYLOGENETIC SIGNIFICANCE 

 OF THE DIFFERENTIAL CHARACTERS EMPLOYED 

 IN SYSTEMATIC WORK 



In the systematic study of organisms the differential characters 

 are always employed in pairs, the two components of each pair being 

 contrasted with each other. 



Within each group individual pairs have ordinarily only a limited 

 application, serving for the differentiation of certain units, but being 

 quite useless for the differentiation of others. 



