93 
Below the big central vacuole are the lower polar nucleus and 
the three antipodal nuclei to be seen (fig. 6). The remains of 
the degenerated megaspore nuclei are here still found in the 
basis of the sac. 
As in most members of the Compositae there are originally 
two antipodal cells, the upper one containing two nuclei. In 
some genera as f.i. Cosmos, Cosmidium, Helianthus (‘TAcKHoLM 
1916), and Bidens (Dauteren 1919) the antipodals constantly 
remain two in number while in the other genera of the family 
as rule three or more cells with one or more nuclei are ulti- 
mately formed. Vittadinia is no exception to this common con- 
dition, as in later stages a wall is developed between the two 
nuclei in the upper antipodal cell. This wall formation is generally 
later followed by one division of some or all of the three anti- 
podal nuclei. The develop- 
ment of the antipodals may, 
however, in some cases ‘ 
proceed still further by 
means of successive cell 
and nuclear divisions until 
a slender “tissue” of ten or 
twelve antipodal cells is 
formed. Fig. 7 shows the 
initial stage of such an 
antipodal growth. 
At this stage no trace of 
the degenerating chalazal 
megaspore nuclei can be 
seen, 
Fertilization seems regu- 
larly to occur according to a 5 
snd ecual; the earliest Vittadinia triloba. Fig. 7. Antipodals of scion 
endosperm stages, however, embryosac (Leitz. Ok. 4, obj. tam 1/12). — Fig. 8. 
mbryosac with embryo and cellulair endosperm. (Leitz 
were not seen. Neverthe- ox. 1, obj. 6) 
less, it seems without doubt ae herel 
that the exidospermis’ calluist from the beginning t so 
confirming the recent work of Danicren (1919) who has desert 
