1909] WIELAND—WILLIAMSONIAS OF MIXTECA ALTA 437 
although the point of stamineal attachment has become strongly 
depressed and indeed is virtually basal now, the flower has arisen 
very similarly to that shown in fig. 4. The two single stamens 
alternating with the two pairs thus form an opposing analogy to the 
Suppression of megaspore regions in the Agave. But.other explana- 
tions are also possible. For we believe that in the angiosperms 
petals have developed from bracts, that they have resulted from the 
complete sterilization of sporophylls, and that, as explained above, 
they result in vast numbers of instances from the apical expansion 
of a sporophyll, which, though greatly reduced, may still bear either 
Megaspores or microspores or both.s And furthermore, we regard 
the stamens of Liriodendron as final reduction stages of individual 
sporophylls that were once large. 
Of further interest in the cruciferous flower is the fact that little 
or no difference in the development of the petals accompanies the 
Staminate change. But, on the contrary, in the diagram of a honey- 
suckle (jig. 8), while the stamens are of normal number and size, 
3 There also appear to be various instances of more or less completely salver- 
shaped corollas, in which what are usually called the sepals are in reality fertile or; 
alternate with the petals, between which they press and fuse up to the beginning of 
the lobes of the corolla, and there bear stamens just as do the petals. Take for instance 
those forms with four “sepals,” four petals, and eight stamens borne four on the 
petals and four alternately between them, but plainly scaling off with the members 
of the calyx. The cherry (fig. 9 from GRayY) 
xa t one 
retaining more stamens, these not being so 
distinctly determinate in number. (They may 
be of the inner row.) 
Perhaps as great a difficulty as we meet 
with discrete hypogynous stamens and a free 
multiovulate ovary with an apparently distinct 
axial relation, for here there are several pos- 
sibilities. Naturally that first coming to mind is that of a central carpellary whorl, 
followed below by a staminate whorl, and then by the members of the floral envelope. 
But another method needs to be reckoned with, namely, that which may be seszind 
characterized as a completed supero-axial shortening and consolidation of the bi- 
Sporangiate fronds of an apical whorl, in such manner J 
Pinnules assumed a vertical position and fused to form the ovary, style, and stigma, 
While the more apical microspore regions of these same fronds produced the stamens, 
the frond tips finally forming the corolla. 
