306 THE CROWFOOT FAMILY 
A monkshood is like a columbine except for irregularity of 
sepals and staminodes, absence of inner staminodes, indeterminate 
inflorescence, simplicity of leaves, and sometimes fewer carpels. 
All the above genera agree in having numerous ovules, all of 
which may. become seeds, contained in several or many carpels 
which become dry and dehiscent in fruit. 
In Actzea the carpels are reduced to one, which becomes fleshy 
and indehiscent in fruit; the staminodes may be fewer, both they 
and the sepals are regular; and the leaves are ternately decompound: 
otherwise the genus resembles Aconitum. 
Passing now to Anemone we find its most striking differences 
from Caltha and the other genera already described to be the im- 
perfect development of several of the ovules in each carpel, the 
ripening of only one ovule, the indehiscence of the fruit, and the 
possession of an involucre of two or three bracts. In these respects 
it forms a link between our type genus and Clematis where the 
rudimentary ovules are commonly fewer, and all the leaves (like 
the bracts in some species of Anemone) are opposite. 
A still further divergence in Clematis appears in the occasional 
imperfection of the flowers, the valvate xstivation of the sepals, 
the ternate or pinnate nervation of the leaves, and the climbing 
habit and woody stem sometimes developed. 
In Ranunculus we find a still further reduction of the ovules; 
an invariable presence of both essential organs and staminodes; 
imbricate eestivation of the sepals: alternate, palmate, simple leaves; 
and sometimes annual duration: thus being in some respects more 
nearly like Caltha, while in others it is more divergent. 
Finally, an extreme of divergence by reduction or simplification 
is reached in the mouse-tails which may be regarded as annual crow- 
foots with only about five stamens, staminodes, and sepals, bractless, 
solitary flowers, and leaves with unbranched or obscure nervation. 
It may seem a long way from such plants to peonies; but, as we see, 
there are intermediate links binding them pretty closely together. 
As the student examines other members of the same family he 
will find that they may be readily interposed as links in the same | 
chain with those already studied. Indeed, the transitions will 
appear less abrupt than between the few examples to which we 
have confined ourselves. His experience will be much like that of 
a botanist with forms newly discovered. He compares them with 
the forms already known and links them with those which they most 
nearly resemble. Thus link by link are family chains forged in 
botanical systems. As in the present case, the chain may branch, 
and it might be questioned whether it would not be better to regard 
the branches as separate families. That depends upon how close 
the linkage appears to be, and as to that the judgment of experts | 
may differ. In any event the definition of any family properly 
follows the attempt at natural grouping, and may require revision 
with advancing knowledge or change of view. Such changes in 
