676 
Telopea Vol. 6(4): 1996 
1978). A micropylar megaspore in the tetrad is functional and enlarged (Fig. ), 
while the three remaining megaspores degenerate. A chalazal megaspore, instead of 
the micropylar one, or both the micropylar and chalazal megaspores may rarely 
appear to be functional, but we have never observed the chalazal megaspore to 
develop into an organized embryo sac. The nucleus of the functional micropylar 
megaspore always is located on the micropylar side of the cell and, following its 
mitotic divisions, develops into a two- (Fig. 19) and a four-nucleate (or -celled) 
Oenothera type embryo sac (Fig. 20). An organized embryo sac has one egg cell, two 
synergids and one polar nucleus. 
Some genera of Onagraceae differ markedly in the thickness of their nucellar tissue, 
particularly in the parietal tissue derived from archesporial cells (for two contrasting 
ovules, see Figs 9a, b). The primary parietal cell (see pp in Figs 7, 8) denved from the 
archesporial cell divides periclinally, and the two daughter cells continue to divide 
periclinaUy. Consequently, above the linear tetrad of megaspores lies a relahvely thin 
layer of parietal tissue, about five to eight cells thick (Fig. 9a), as found m Fuchsm 
(Hg. WX Circaea, Lopezia (Fig. 11), Epilobium (Fig. 12) and G<ii/oplf uw (Fig. 13), or a 
relatively thick layer, about ten to 20 cells thick (Fig. 9b), as found in Clarkia (Fig. 14), 
Hamm (Fig. 15), Gongylocarpus, Xylonagra, Camissonia, Calylophus, Gaura, Oenothera and 
Steiiosiphon. In most genera the tetrad of megaspores is deeply buried in the nucel us 
and, except in Glarkia heterandra (‘Heterogaiira’) and Go»^^/ocnrpi(S, is positioned at or a 
little above the bottom of the nucellus. Compared to other species of Clarkia, C. heterandra 
is unusual in that the divisions of the parietal cells are retarded and that a tetrad o 
megaspores is positioned in the middle of the nucellus (Fig. 16). At this stage, C. heterandra 
has a thin layer (about five to eight cells thick) of parietal tissue above and a somewhat 
thicker nucellar tissue below the megaspores (Fig. 16). In later stages of development, 
the cells of both tissues above and below divide rapidly, increasing their respective 
thickness, so that the embryo sac comes to be positioned in the center of the nucellu^ m 
the two-nucleate embryo sac stage, the parietal tissue is up to 16 to 20 cells thic 
(Fig. 17). Gongylocarpus is similar to most other genera of the tribe Onagreae in having 
thick parietal tissue (18 to 20 cells thick) above the tetrad of megaspores, but less nucellar 
tissue than the other genera below the tetrad, so that the tetrad is positioned between 
the bottom and the center of the nucellus (Fig. 18). 
A nucellar cap derived from the nucellar dermal cell by its periclinal division is two 
or three cells thick if present, but is relatively insignificant and poorly defined. For 
these reasons, we do not consider it to be a useful feature for making comparisons 
between genera. 
Starch grains, whose functional aspects have sometimes been discussed in relation 
to the polarity of megaspores in a tetrad (e.g., Rodkiewicz & Bednara 1974; Rodkiewicz 
& Sniezko 1978; Sniezko & Harte 1984a), are abundant in the nucellar cells, 
megaspores and embryo sac before fertilization, particularly on the micropylar side 
(Fig. 20). They become less conspicuous in the post-fertilization stages, however. 
Ishikawa (1918: 311) gives a comparison among some genera, noting that starch 
grains in the nucellus are 'many' in species of Oenothera, Gaura and Circaea,^ very 
few' in Chamerion angustifoliuni ('Epilobium') and 'none in Clarkia (Codetia sp.) and 
Fuchsia macrostemnuh However, we have confirmed the presence of starch grains in 
the nucellus and the embryo sac in all species examined of all genera, including 
Epilobium, Clarkia and Fuchsia. 
A hypostase, which is distinguished by accumulation of densely staining tannin-like 
substances in its cells and the thickening of those cell walls, is always formed although 
its differentiating stage is different from species to species. According to Johansen 
(1928), Oenothera, Gaura, Clarkia and Circaea possess a definite hypostase, whereas 
