385 
species of Crotallaria were examined. In both cases the flower- 
ing peduncles representing the end of the stem are attached di- 
rectly opposite the axil of a leaf, but usually a little below its 
level; one leafy branch arises from the axil of this leaf, and as- 
suming the same direction as the stem apparently terminates it; 
another leafy branch starts from the axil of the leaf below, and is 
evidently axillary ; four nodes intervene between successive flow- 
ering peduncles, so that the order of arrangement of the stem 
after it begins flowering: is two leaves usually without anything in 
their axils, a third leaf with a leafy branch in its axil, and a final 
fourth leaf with a leafy branch in its axil and the terminal flower- 
ing pedicel beyond the fourth leaf. 
C. sagitialis, L., has long, narrow, inversely sagittate stipules, 
whose decurrent margins reach the node below in every case ex- 
cept along the first internode of each sympodial branch, which con- 
tinues the leafy growth of the stem after the flowering peduncle 
has been thrust aside. Along this internode the decurrent margins 
of the stipules extend only half way down the internode, and in 
Tare instances are confined only to the very top of this internode. 
These leafy continuations of the stem being in each case branches, 
though sympodial, of course have the first leaf of their one-half 
phyllotaxy in a lateral position, causing the decurrent margins of 
the stipules of the first internodes to be pressed by the front of the 
Subtending leaf and the back of the flowering peduncles, and this 
May account for the abortion of part of its decurrent stipular 
growth, 
3 C. ovalis, Pursh., offers in this respect a more interesting case. 
Here the two lower nodes show only the merest traces of stipules 
at the node itself. The third node at times has the stipules like- 
Wise almost obsolete, but usually the stipules are present, though 
of small size and decurrent for only half the length of the inter- 
node; much more rarely the stipules of this node nearly equal 
those of the fourth node. The fourth node always has the in- 
Versely sagittate and rather conspicuous stipules fully developed 
and the decurrent margins extending to the next lower node. In 
this case the suppression of the two lower stipules and of the 
lower part of the decurrent margins of the third can be less 
_ Teadily accounted for, and their full retention at the node preced- 
é 
