THE FAMILY CHAIN 397 
classification the history of the science illustrates; yet progress is 
in the direction of stability, and certain chains, having held from the 
first, bid fair to endure. The integrity of the Ranunculacee, for 
example, seems assured in spite of the wide divergence of its extreme 
forms and in spite of the difficulty of defining its limits. 
We have now to define the family as best we can. The generic 
formulas will help us to a formula for the family and this in turn will 
lead us to our definition. Taking the prevailing characteristics 
of each part as typical for the family, and neglecting the less sig- 
nificant exceptions to the general rule, we may express a generalized 
view of the salient features as shown in the formula of Ranunculaceze 
on pages 404, 405. 
The only invariable features here expressed are the anatropous 
ovule and the uncoiled embryo surrounded by albumen, and these 
as we shall see are common to a number of other families. But, as 
we shall also see in comparing the Ranunculaceze with other groups, 
it lacks features which they possess. 
Taking into account all the facts we ie learned, the 
crowfoot family may be described as consisting of herbaceous 
or rarely woody plants, never trees, without milky juice, oil 
or other secretions in special reservoirs, but with a mostly 
colorless and odorless sap which is generally acrid, and in 
some cases renders the plant poisonous to eat or to touch; 
leaves mostly palmately branched, or at least palmately 
ribbed; flowers mostly regular and perfect with the parts 
free and distinct (with rare exceptions); sepals commonly 
five, generally petaloid; petals rarely present, often replaced 
by more or less petaloid staminodal nectaries of widely 
differing forms; stamens generally numerous; anthers de- 
hiscing by slits; pistils almost always simple, numerous, 
few, or rarely solitary; ovules anatropous, many, few or 
solitary, sometimes rudimentary; fruit follicular, capsular, 
achenial, or rarely fleshy; the seeds with hard albumen sur- 
rounding a minute uncoiled embryo. Or, if we disregard all 
that is untypical, it may be said that whenever we find an 
herb with the juice colorless and scentless, the flowers having all 
their parts distinct and free, sepals about five, and essential 
organs numerous, we may be tolerably sure that our plant is 
one of the crowfoot family, although some departure from 
these characteristics would not necessarily exclude it from 
the group. 
