310 NATURAL SCIENCE. May, 



Series of forms whose successive links correspond with succes- 

 sive stages of development in their youngest, still existing representa- 

 tives are the only ones that furnish us with an incontestable picture 

 of the path along which any given assemblage has evolved. These 

 are the kind of genealogical trees that form the worthy goal of 

 palaeontology. From them a natural system will arise of its own 

 accord. But from this goal we are unhappily still far removed. As 

 a rule our palseontological trees lack an ontogenetic foundation, and 

 that the foundation itself may be constructed in an arbitrary manner 

 is best shown by the unsatisfactory condition of our ammonite- 

 literature. 



The time, it seems to me, has not yet arrived for the thorough 

 reform of zoological classification on a phylogenetic basis. Among 

 Protozoa and Ccelenterata there are absolutely no satisfactory fixed 

 points for the phyletic arrangement of the various groups. Among 

 Echinodermata it is proved that the correspondence in embryonic 

 development between Asterozoa and Echinozoa is evidence of a 

 common origin ; but the classification of the various Classes is as yet 

 affected only to the smallest extent by the facts of ontogeny and 

 phylogeny. The union of Bryozoa and Brachiopoda into a special 

 phylum — the Molluscoidea, and their connection with the worms, 

 depend entirely on embryological comparison : in their later develop- 

 ment the two Classes go so far apart that we can find no further 

 parallel between them ; and although the beautiful researches of 

 Beecher, Clarke, and Schuchert on the phylogeny and ontogeny of 

 the Brachiopoda will furnish a solid foundation for a new and better 

 classification of the Class, the first adumbration thereof is still 

 somewhat doubtful. On the other hand, researches on the develop- 

 ment of the shell in Mollusca are, without doubt, full of promise. 

 What results we have to expect in this field is shown by the labours 

 of Jackson, Hyatt, and Branco, though it must be confessed they 

 still afford no sufficient basis for a classification of the pelecypods and 

 cephalopods. 



Palaeontology has made its deepest mark in the classification of 

 the Vertebrata. Here we frequently come across firmly-rooted 

 genealogical trees. Phylogenetic and ontogenetic facts have effected 

 the removal of the order Solipedia and the natural grouping of the 

 ungulates. The discovery of the fossil Condylarthra and Creodontia 

 has brought to light unlooked-for relationships between ungulates 

 and carnivores. The remarkable fauna of the Puerco Beds contains, 

 according to Cope, almost completely indifferent mixed types, which 

 cannot be considered either as true ungulates, beasts of prey, or 

 rodents, nor can they even be regarded as typical Condylarthra, 

 Creodontia, or Lemuria : scarcely can they be dovetailed into the 

 framework even of a geological classification, since they show 

 relationships in the most varied directions. 



And here we approach an important question of principle. The 



