1921] MACPHERSON—CUSCUTA AND CONVOLVULUS 303 
The material was collected in September. The Cuscuta ma- 
terial was killed, part in a chromo-acetic solution and part in 
a picro-acetic solution, and the germinated seeds in a fixing 
solution consisting of 60 cc. of 95 per cent alcohol, 35 cc. of 
water, 5 cc. of formaldehyde, and 2 cc. of glacial acetic acid. All 
the Convolvulus material was killed in this latter fixing solution. 
The Cuscuta and Convoloulus were treated in the usual way and 
imbedded in paraffine, sectioned, and stained, some in haema- 
toxylin, some in haematoxylin and eosin, and the larger part in 
iron-alum haematoxylin. The sections of Cuscuta were cut about 
25 w in thickness and those of Convolvulus about 10 wu in thickness. 
The ovule in both species is anatropous (figs. 1, 20). The 
development of the embryo sac up to and including the 8-celled 
stage both for Cuscuta and Convolvulus follows the usual method. 
In the 2-celled stage of the embryo sac of Cuscuta Gronovii (fig. 2) 
there is a dense mass of protoplasm lying between and surrounding 
the two nuclei which are located at the ends of the sac. In the 
4-celled stage (fig. 3) the sac is much larger and the protoplasm less 
dense. In the 8-celled stage (fig. 4) the sac is somewhat larger, 
and the protoplasm is most abundant immediately surrounding 
the polar nuclei. The synergids and egg are large, prominent, and 
somewhat irregular, but conform to the usual type. The anti- 
podals are much smaller, but very distinct and well defined. The 
polars are oval, intermediate in size between those of the egg appara- 
tus and the antipodals, and vacuolate. At this time the cells of 
the nucellus show the presence of a large quantity of starch grains 
(fig. 4), which, on account of their taking the stain so much more 
readily, render the nucellus very conspicuous. The cells of the 
nucellus are rich in starch, and those immediately surrounding 
the embryo sac show indications of rapid disintegration, accom- 
panying the enlargement of the sac. 
As previously stated, CouLTER and CHAMBERLAIN describe 
the endosperm of Cuscuta as arising by free nuclear divisions and 
not by continuation of the process of cell division, with the forma- 
tion of walls. This is contrary to the usual course of development 
of the endosperm in saprophytic and parasitic dicotyledonous 
