148 



TRE AMERICAN NATURALIST 



[Vol. LIII 



gave rise to both the Annelida and the Hemichordata 

 are quite likely to appear in both Annelida and Hemi- 

 chordata (or in forms descended from them, such as the 

 Arachnida and the Vertebrata).^ 



One of the chief difficulties in the way of reaching a 

 proper understanding of the mutual interrelationships 

 of the different lines of development is the attempt to 

 arrange these lines in the form of a dichotomously 

 branching tree drawn in one plane— which is almost as 

 impossible as the attempt to arrange all animals in a 

 single linear developmental series; for it must be borne in 

 mind that these different lines of development frequently 

 approach one another from different directions, so that 

 it would be necessary to represent their relationships by 

 a figure drawn in three dimensions, rather than in a 

 single plane. If this is done, it becomes easier to under- 

 stand that the line of development of the "Annelida," 

 for example, is paralleled (on different sides^ bv those 

 of several other gi-oui)s, and that all of these lines ol 

 descent mav lend back to a common ancestry, or that 

 tlieir ]>oints ot oimiii ni;iv be near the pomt at which the 

 Imc ot (Icscciir (it llic ••Aiim^lida" arose. 



Ill (liscii>>inii' I he iiltiiiinte relationships of the Verte- 

 l)rata, I'icliinodci iiiatn. Mollusca, etc., the lack of inter- 

 mediate forms annectent between the different develop- 

 mental series, or connecting them with the siip])osedlv 

 ancestral forms, has made the subiect of their affinities 

 extrcKuMv s])('('ulative : and it is not until we come to the 



