334 Agassiz's Contributions to the 
times requires the dividing off of a family group, when external 
form would not seem to demand it? 
We throw out these queries without attempting to give them 
a formal answer. We do not imply any doubt as to the truth- 
falness of the grand idea laid down for the Branches, or of that 
for the Classes, or of the importance of ordinal relations, or of 
tm, as characteristic of the Family group, or of superfici 
structural details as the basis of Genera. Instead of attributing 
less importance to these ideas we are led by the views them 
selves, from which we have derived profound instruction, to.sus- 
even a more comprehensive meaning and use of them th 
he idea of ordinal relations seems to Ts? 
as well as below; 
d in the scale 
eee ne 
are not Orders under the Branches; and why the firs Ad 
of Orders under Classes should be a primal division and not - 
it is an ideal element even as r gards a species, Si ith 
there is a range of variations more or less wide. But the ese 
in extremely small, compared with the — j 
parts; but as we descend, we come to a range UJ 
constant, and then the variation is in the relations of the Pia 
at this grade of subdivisions then, form may be of tha - 
which strikes the mind of one but little accustomed to pee 
eralities, and this is most prominently the Family grade, altho la- 
degree If then we give to peingane 
* h: t 
all the more strongly the view #2 
