1144 The American Naturalist. [December, 
erate simplicity of structure, or (where it occurs in the less simple 
members of the class) appears as the almost mechanical separation 
of part of the body, tending to destroy the general equilibrium by 
the collection of sexual products. The formation of a new head 
in this process can only be regarded as a phenomenon of regen- 
eration that has been thrown back into earlier periods in conse- 
quence of the regular repetition of the process of fission from 
generation to generation, and that finally begins in a period when 
the separating part is still connected with the parent. 
Moreover, if metamerism in segmented animals had originally 
arisen from the formation of buds, then, in the first place, the 
production of new segments must always take place at the extreme 
end of the body, or, in other words, the terminal section should be 
always the most recently formed metamere; and, in the second place, 
the entire body must consist of completely equivalent, consecutive 
pieces. In place of this we see, however, that the new segments 
arise from a formative zone anterior to the telson, that the telson 
belongs to the oldest part of the animal, and that all parts of 
the body do not correspond, since neither the cephalic lobe 
with the buccal zone and foregut, nor the terminal piece with the 
hindgut, can be compared with the metameres lying between 
them. If we derive the origin of metamerism from the kind of 
strobilation found in Acalephs, then the youngest segments follow 
immediately upon the first body segment, which (though occur- 
ring in the segmented Cestodes) is not at all the case in all other 
segmented anrmals from the Annelids up. 
The heteronomy of prostomium, trunk segments, and telson, be- 
comes intelligible, nay necessary, when we regard the metameres 
as arising from the segmentation of the trunk between the head 
and tail pieces,—that is, as arising in situ. How may this process 
have been effected in phylogeny ? 
The acquisitions of recent years indicate with considerable cer- 
tainty that all segmented animals, omitting the Cestodes, have 
descended directly or indirectly from Annelids, and thus in them 
or in their immediate ancestors the formation of metameres had its 
origin. As long as these ancestors are regarded as trochospheres 
or medusa-like creatures (Hatschek, Kleinenberg), or if the Anne- 
