308 BOTANICAL GAZETTE | NOVEMBER 
closing the cavity of the pistil. The suture along which the 
laminae meet is like an inverted Y, for before they close they 
are separated at the bottom by the axillary placenta(see fig. 79, 
R. eremogenes) . 
If we now compare with this the conditions found in Poten- 
tilla and Fragaria, we see the following modifications. There is 
no axillary mass of cells developed, for the ovules have their 
origin on the edge of one or the other lamina. Probably this 
originated as follows: in some plant whose ovules were borne 
as in Ranunculus, and in which the suture along which the lam- 
inae met formed an inverted Y, a variation appeared by which 
the axillary placenta lost its median position, one arm of the 
suture becoming elongated and the other shortened. This 
resulted in a placenta attached to one lamina and free from the 
other, which is precisely what we find in Potentilla. Even in 
Ranunculus it occasionally occurs that the axillary placenta is 
not strictly median but slightly shifted to one side or the other. 
This suggests that the mode of origin indicated above is not 
improbable. 
In the uniovulate pistils of Fragaria as well as of Potentilla 
the ovule is borne sometimes on one lamina and sometimes on 
the other, being very variable in this respect. Under such con- 
ditions when neither lamina is especially modified for the pro- 
duction of ovules it is probable that sometimes an ovule might 
be borne on each lamina, as happens in Fragaria. It seems 
probable that it was by such a variation that the majority of the 
genera of the Rosacee became biovulate. When once each 
lamina began to be ovuliferous it would be but a short step to 
the condition in which several ovules are borne, instead of only 
one. In this way the multiovulate pistils may have arisen. 
Probably in this way, too, the multiovulate Ranunculacee wefe — 
developed from the uniovulate Ranunculus, possibly through a 
biovulate form close to Callianthemum, or even Hydrastis. 
From the description of their development as given above it 
must be evident that Sagittaria and Alisma are quite similar to 
Ranunculus. Of these two genera, however, Alisma is much 
