718 
Telopea Vol, 6(4): 1996 
may be an adequate interpretation for stamen fascicles in some flowers, possibly even 
for other myrtalean genera such as Lagerstroemia (Lythraceae; Ronse Decraene & Smets 
1991), it does not appear to be particularly appropriate for the genera of Myrtaceae 
discussed here. The independent development of the pre-staminal bulge prior to stamen 
initiation, the dissimilarity of its size and shape to typical stamen primordia, the 
phyllotactic independence of stamens on one pre-staminal bulge from stamens on 
other pre-staminal bulges, and the maintenance of the resulting fascicle as a discrete 
unit in the mature flower, predicates the pre-staminal bulge and resulting fascicle as a 
morphogenetic structure distinct from the other floral organs. A flexible framework of 
floral construction is needed to allow the recognition, where appropriate, of 
morphological units in addition to the four classical organ categories; in this case, 
fascicles resulting from PSBs are a novel feature for these myrtaceous flowers. 
According to our interpretation the pre-staminal bulge is not a primary stamen 
primordium, so it is inappropriate to describe the morphology of these myrtaceous 
androecia in terms of conventional haplostemony/diplostemony. Haplostemony and 
diplostemony are simply the results of certain spatial and temporal factors that 
determine the positions of petals and stamens in a regular, consistent, acropetal 
phyllotaxy, directly on a usually convex floral meristem. They are not architectural 
groundplans to which patterns of all flowers must conform, nor immutable types of 
which all flowers must be avatars. If the spatial and temporal relations change, then 
so will the morphology and phyllotaxy of the flower, and there is no need to invoke 
a direct correspondence between every organ and positional relationship in such 
different flowers. Although the plesiomorphic condition for Myrtaceae may well be 
diplostemonous (Schmid 1980, Ronse Decraene & Smets 1991), with subsequent 
changes to apparent obdiplostemonous, haplostemonous and obhaplostemonous 
conditions in some taxa (Johnson & Briggs 1984), it is not really appropriate to 
describe flowers such as Lophostemon, Tristaniopsis, Xaufliostcinon and Lysicarpus as 
modifications of these conditions. These flowers, with their deeply concave apical 
meristems and non-acropetal order of floral parts, have different spatial and temporal 
parameters to (ob)haplostemonous and (ob)diplostemonous flowers, and they 
manifest with a different form and phyllotaxy. The antepetalous fascicles of stamens 
exhibited by Lophostemon, Welchiodendron and Tristaniopsis represent a special type of 
polyandry; it is not complex obhaplostemony sensu Ronse Decraene & Smets, because 
each fascicle does not result from the proliferation of a single stamen primordium. 
Similarly, in Xanthostemon and Lysicarpus polyandry results from the prolonged 
inception on the floral apex of numerous individual stamen primordia, not the 
proliferation of a few antepetalous or antesepalous precursors. In fact the term 
proliferation, which is used by Ronse Decraene & Smets (1991) and Johnson & 
Briggs (1984) to explain both of these polyandrous conditions, is best avoided in 
these instances. In both Lophostemon and Xanthostemon, polyandry is beyond the 
concepts of haplostemony and diplostemony in both their ontogenetic and 
phylogenetic senses. 
Acknowledgements 
For collecting specimens used in this study, we thank Peter Wilson, Royal Botanic 
Gardens Sydney; Richard Johnson, Royal Botanic Gardens, Mount Annan; Ian Smith, 
Maranoa Gardens, Melbourne; Gregor Calvert and Chris Roberts, Pajinka Wilderness 
Lodge, Queensland; Bernie Hyland, Tropical Forest Research Centre, Atherton, 
Queensland; Anthony Vadala, The University of Melbourne; Peter Neish, Royal 
Botanic Gardens, Melbourne and Andrew Rozefelds, Tasmanian Herbarium, Hobart. 
