146 
James Small. 
(6). Sac from all four megaspore nuclei and sixteen- 
nucleate, as in Pyrethruiu parthenifolium var. aureuin 
(Palm, 85). 
Winge’s interpretation of his figures is rather unsatisfactory 
and has been shown to he wrong hy Carano (47). In his English 
summary he states that three megaspores degenerate but in the 
text he figures and describes three large bi-nucleate megaspores 
and three very minute, degenerate antipodals in the region of the 
egg-apparatus. He considers that the micropylar megaspore 
develops into the embryo sac, but his Pig. 3 shows the chalazal 
megaspore slightly larger than the others, and his Fig. 4 shows 
three small megaspores and a large one, the row being isolated so 
that the micropylar or chalazal position of the large megaspore 
cannot be determined. 
Winge’s figures, apart from the very problematical degenerate 
antipodals, are similar to those given here and are best intei preted 
in a similar way, as Carano (46-47) and others maintain. 
Palm’s interpretation of the phenomena in Aster and Solidago 
has already been more or less discredited by Chamberlain (49). 
There remain the anomalies of types 3-6; these all occur in the 
Anthemideae and it is possible that further investigations will 
confirm the presence in this tribe of anomalous embryo-sac develop¬ 
ment, as in Plumbagella (51), Lilium , etc. 
The disappearance of the upper megaspores is very clear in 
Senecio and Silphium. Even in the interesting case of an aposporic 
embryo sac described by Rosenberg (90) in Hieracium, the tetrad 
was formed and the chalazal megaspore was the last to disappear 
before the nucellus was pushed on one side by the growth of the 
abnormal sac. The origin of the chalazal haustorium from the 
antipodals can be taken as without any properly authenticated 
exception. 
In the mature embryo sac there is the normal arrangement of 
synergids, oosphere, endosperm or fusion nucleus, and three 
antipodals (Pig. 63). The nucellus disappears completely, leaving 
the embryo sac surrounded by the epithelial layer of the integu¬ 
ments, or a few remnants may be left to form the so-called nucellar 
cap (Fig. 67). 
Antipodal Cells. In Senecio vulgaris the basal or chalazal 
antipodal elongates (Fig. 63) and divides (Figs. 64-67), the basal 
portion showing as many as four extra antipodal cells in one 
longitudinal section, each cell having one or more nuclei (Pig. 67) 
This is the structure of which it is said that “the antipodals of many 
of the Composite are organised into an aggressive haustorium 
