668 
Telopea Vol. 6(4): 1996 
the megasporogenesis of Oenothera and Epilobitim in relation to polarity ai^ the 
competWon between megaspores in a tetrad, or their megagametogenesis to Jo™ 
tnudeate Oenothera typf e^ryo sac. UnHl recently relatively little 
paid to other embryological characters, and many of more than 50 characters that 
discuss in this paper have remained unstudied, including such features as t 
development of anthers, ovules (the integuments and nuceUus in particular) ^ee 
We have presented analyses of certain embryological characters in the whole fanu y, 
specifically, on the histogenesis of integuments (Tobe & Raven 1985 and on the divi 
(or septate) sporogenous tissue of anthers (Tobe & Raven 1986a). Subsequently, 
examined some 40 embryological characters in 11 species of Liidivigia, representing 
seven of its 23 secHons (Tobe & Raven 1986b). As a result, except for the 
of the anther and integuments, we now can summarize the present level of knowledge 
as follows: Ludzmgia isCoughly known; Oenothera, Clarkia (now including 
heterandra; Lewis & Raven 1992), Chamerion and Epdobium (now mdudm^ Boisdiwaha 
Hoch & Raven 1992) are relaHvely well known; Circaea, Lopezia, Camtssonia, Gayophytw 
and Stenosiphon are known to a hmited degree; and Fuchsia, 
Xylonagra, Cahyhphus and Gaum are little known or unknown (for references to individual 
genera, see footnotes in Tables 2-4). 
The purpose of this paper is to clarify embryological attributes of all genera of Ona^^eae 
in order to provide a basis for comparison with other families and within the fam y, 
and, on the basis of additional embryological evidence, to discuss familial and generic 
relationships. For this purpose we have investigated one or more species of 1 
onagraceous genera (Table 1). Subsequent to the specimen examination and data collection 
for this study, it has been demonstrated that Chamerion should be segregate 
Epilobium (Baum, Sytsma & Hoch 1994; Hoch, unpublished data). No collechon of 
Chamerion was included in our specimens, however, there are several reporLs for it in the 
literature (Lebegue 1948b, among others). Because there is no evidence of embryological 
differences between these two groups in the available 
Chamerion separately in this report. Our previous study of Ludzoigia (Jobe & Rave 
1986b) indicated that most embryological features (except for those of embryogenesis 
and seed coat anatomy) do not vary within a genus. This relationship allows us to use 
one or a few species as representative of the general embryological features of each 
genus in the aLence of other information. We have also incorporated previously 
published information about embryology of Onagraceae, evaluating it and presenting it 
along with our own results. 
Materials and methods 
Tliirty-three species representing 14 genera were investigated. All three major components 
of embryology — i.e., anthers, ovules and seeds — were examined in each sp^^s for 
which sufficient material was available (Table 1). Samples of flower buds and fruits in 
various stages of development were fixed in FAA (5 parts stock formalin; 5 parts glacial 
acetic acid; 90 parts 70% ethanol). Observations were made using serial microtome 
sections, except that the number of cells in a mature poUen grain was observed using 
whole pollen grains stained with 1% aceto-carmine (Tobe & Raven 
preparing microtome sections are presented elsewhere (Tobe & Raven 1986b). 
We have made a sufficient number of observations to determine the type of 
embryogenesis in only a few genera, although some features of embryogenesis are 
reported for most of our samples. Likewise, we report here only hnnted informc^ion 
about seed coat structure, which nearly always vanes within individual 
Oenothera, see Tobe, Wagner & Chin 1987; Ludzuigia, see Tobe & Raven 1986b; Tobe, 
Raven & Peng 1988), because our focus in this paper is on generic relationships. 
