28 Rhodora [FEBRUARY 
repeated many times, so that by universal consent the compound 
habit is not regarded as a mark of relationship. 
Certain leaf characters are common to entire and to compound 
leaves, and are widely spread throughout the series. Such are 
stipules, and absciss-layers of cells at the bases of petioles, causing 
defoliation. These characters may be considered older than the sev- 
eral cases of compounding. 
Certain other characters are found only in compound forms, and 
since they cannot exist in entire-bladed leaves, must be of later his- 
toric development than the cases of compounding where they sever- 
ally occur. Such features are stipels and absciss-layers of cells in 
petiolules. 
These characters, being Jater than the compounding which they 
accompany, are later than the stipules and petiolar absciss-layers of 
the same species, which features are older than the compounding. 
Moreover stipels and petiolular absciss-layers occur in widely separated 
families and must therefore have arisen several times independently 
in the case of stipels, and many times independently in the case of 
absciss-layers. 
Stipels are found in Staphyleaceae — e. g., Staphylea trifoliata, S. 
pinnata, S. Bumalda, Euscaphis staphyleoides, Turpinia occidentalis ; 
in Rutaceae — e. g., Zanthoxylum carolinianum, Z. americanum; in 
many Leguminosae (Papilionaceae) — e. g., species of Phaseolus, 
Robinia, Bradburya, Desmodium, Dolicholus, Galactia, Vigna, Amor- 
pha; and in Caprifoliaceae — e. g., Sambucus canadensis and S. race- 
mosa. In looking over these examples one is struck by two general 
relations. Zirst, in all these cases stipels have been preceded by 
stipules. Sambucus canadensis is usually exstipulate it is true, but S. 
racemosa is stipulate, and on young vigorous shoots of S. canadensis 
itself I have found stipules; the stipulate habit is in the genus and 
occasionally in the species, therefore, I may add that the stipels are 
by no means constantly present. Secondly, in all cases observed 
there is a close resemblance of stipules and stipels in the same spe- 
cies. Thus in Staphylea the stipules are relatively long, linear, pale, 
and membranaceous ; the stipels, though smaller, answer to the 
same description. In Zurpinia occidentalis the stipules are much 
reduced, and the stipels also. In Zanthoxylum carolinianum and 
americanum the stipules are spiny and rather stout ; the stipels agree. 
In Robinia viscosa and other species of the genus, the stipules are 
