91 
differing slightly but each one constant in size, the micropy lar 
one being the larger one of the two (fig. 2). In the phragmoplast 
no trace of a cell plate could be seen. Vittadinia in this respect 
agrees wholly with Anthemis and Pyrethrum. It is of course 
possible that a cell plate, followed by the formation of a wall 
may develop in the phragmoplast as sometimes happens in 
Tanacetum (see above), but in my material at least, this does 
not seem to occur. Two vacuoles are constantly present at this 
stage, a larger one between the two nuclei and a small one 
above the micropylar nucleus. 
The homoeotypic division shows essentially the same features. 
The micropylar pair of nuclei differs, too, slightly in size, the 
upper nucleus again being the larger of the two. The division 
of the smaller heterotypic nucleus gives rise to a chalazal pair 
of very small nuclei, both, however, of equal size. No walls are 
laid down during the homoeotypic division; the four megaspore 
nuclei are thus arranged ina row, enclosed in a common plasma 
and separated only by means of vacuoles (fig. 3). In a general 
way conditions remind one strongly of Tanacetum and Anthems. 
In the former plant, however, there is practically no difference 
in size in the megaspore nuclei, and in the latter one all of 
the three lower nuclei are, according to Hormeren (1915), of 
about the same volume. | ‘ 
As mentioned above, the four megaspore nuclei at this stage 
lie in a row, occupying the upper two thirds of the entire cell. 
These micropylar nuclei now constitute the embryosac which 
thus develops according to the Scilla type (Parm 1915). 
The basal nuclei generally show no further signs of activity, 
and degenerate slowly, forming two dark staining bodies in the 
antipodal part of the young embryosac (fig. 4—6). In a very 
limited number of cases, however, they have been seen to divide 
simultanously. At present, I am not in a position to tell how 
far this development could possibly proceed. Most likely they 
will degenerate later on; still there are indications that they 
might continue to divide. This very interesting question will be 
taken up again when more material is at hand. Such divisions 
would result in embryosacs with an accessory antipodal region 
