674 
Telopea Vol. 6(4): 1996 
its relative frequency, for example, in Epilobium (Michaelis 1925), Lopezia (Tackholm 
1914) and Oenothera (Hulbary & Rao 1959; O'Neal 1923; Subramanyam & Govindu 
1948). In his study of Oenothera tetraptera (‘Hartmannia letraplera'), Johansen (1929: 
289) corrected his early notes on the archesporium, from 'probably several 
archesporial cells [exist]' to 'a single archesporial initial [exists],' stating that 'cells 
adjoining the archesporial initial may often simulate the functional appearance of 
the latter.' However, all samples of Onagraceae we examined had plural archesporial 
cells in conformity with a few earlier reports of the occasional occurrence of plural 
archesporial cells (e.g., Hulbary & Rao 1959; Michaelis 1925; O'Neal 1923; 
Subramanyam & Govindu 1948), as well as of plural megaspores (derived from 
different archesporial cells) and embryo sacs (e.g., Langendorf 1930; Renner 1914; 
Tackholm 1915). These earlier reports also seem to suggest that the multi-celled 
archesporium is prevalent. We cannot confirm the existence of a one-celled 
archesporium in any Onagraceae. 
Usually a sporogenous cell derived from one of the plural archesporial cells increases 
its volume and becomes a megaspore mother cell. The megaspore mother cell 
undergoes meiosis, forming nearly always a linear tetrad of megaspores (Figs 9-12) 
and very rarely an oblique linear tetrad. A triad of megaspores may also occasionally 
occur, resulting from the suppression of homotypic division in the lower cell of the 
dyad (for the frequencies of triads in certain genera, see Rodkiewicz & Sniezko 
A 
Fig. 9. Diagrams illustrating two contrasting ovules with respect to the thickness of parietal 
tissue (pt) lying above a tetrad of megaspores (tm). a. Ovule with a thin parietal tissue (5 cells 
thick), b. Ovule with a thick parietal tissue (22 cells thick). 
