86 MERISTIC VARIATION. [PABT I. 



become differentiated to form members of such non-coincident or 

 independent Meristic series. 



Some years ago 1 , in the course of an argument that Balanoglossus 

 should be considered as representing some of the ancestral characters 

 of Chordata, I had occasion to refer to some of these difficulties, and 

 especially to the different characters of the two kinds of segmentation ; 

 that of the Annelids, in which the repetitions of the organs belonging 

 to the several systems are coincident, and, on the other hand, that 

 of the Chordata, for example, in which this coincidence may be 

 irregular or partial. At that time I was of opinion that these two 

 sorts of segmentation may, in certain cases, have had a different 

 phylogenetic history, and have resulted from processes essentially 

 distinct. It appeared to me that we should recognize that, in the 

 Annelids on the one hand, segmentation of the various systems of 

 organs had been coincident from the beginning, while in the Chordata 

 the segmentation had been progressive and had arisen by segmentation 

 or repetition of the organs of the several systems independently. The 

 reasons for this view were derived chiefly from the fact that it is 

 possible to arrange the lower Chordata in order of progressive segmen- 

 tation of the several systems. In particular such treatment was shewn 

 to be applicable to the central nervous system, the vertebral column 

 and the mesoblastic somites, and in these cases it was maintained that 

 the evidence of the lower forms of Chordata goes to shew that segmen- 

 tation had occurred in these systems one after another, and that their 

 segmentation was not derived from a form having a complete repetition 

 of each part in each segment : that these forms, in fact, shewed us tin- 

 history of this progress from a less segmented form to one more fully 

 segmented. 



The views then set forth have met with little acceptance. Those 

 who are occupied with the search for the pedigree of Vertebrates still 

 direct their inquiries on the hypothesis, expressed or implied, that in 

 the ancestral form there was a series of complete segments, each 

 containing a representative of each system of those organs which in 

 the present descendants appear in series. It is thus supposed that each 

 segment of the primitive form must have been a kind of least common 

 denominator of the segments of its posterity. The possibility that the 

 segmentation of Vertebrates may have arisen progressively is, indeed, 

 scarcely considered at all. 



Though in the light of the study of Variation, it now seems to 

 me that the discussion of these questions must be indefinitely post- 

 poned, and that there are radical objections to any attempt to interpret 

 the facts of anatomy and development in our present ignorance of 

 Variation, I have seen no reason to depart from the view expressed 

 in the paper referred to : that interpreted by the current methods of 

 morphological criticism, the facts go to shew that the segmentation of 

 the Chordata differs essentially from that of the Annelids drc, and 

 that it has arisen by progressive segmentation of the several systems of 

 an originally unsegmented form. T<> those who hold as Dohrn, Gaskell. 

 Marshall and others have done, that the evolution of Vertebrates has 



1 Quart. Jour. Micr. Set, 1886. 



