772 
Telopea Vol. 6(4): 1996 
systems in the latter. All Grevilleoideae (except Neorites) have a morphological and 
anatomical modification of the style-end that is termed a pollen-presenter because it 
holds and displays the dehisced pollen from the anthers. There are several 
developmental and morphological features in the flowers associated with the pollen- 
presentation systems in Grevilleoideae including lobes around the upper portion of 
the filament that clutch the basal portion of the pollen-presenter, differential cell 
enlargement and growth in the style and a highly synorganized tepal-stamen-carpel 
enlargement process (Douglas, unpub.). It is evident from developmental studies of 
other grevilleoid taxa that the pollen-presentation systems have become increasingly 
elaborate and involve numerous facets of a flower and its development, thus the 
synorganization has increased. Given the interdependence of floral organs 
developmentally, structurally and functionally in highly synorganized flowers 
(Endress 1990,1994) it seems unlikely that a pollen-presenter and all of the associated 
floral features could be lost. Although loss of a pollen-presenter system is possible, 
it seems dubious based on the various developmental processes involved in other 
parts of the flower among Grevilleoideae. A phylogenetic analysis of the family is 
necessary however before Venkata Rao's (1971) hypothesis can be entirely dismissed. 
There are two other possibilities concerning the origin of the Carmrvonia inflorescence; 
one being the amplification of a single flowered system as found in non-grevilleoid 
taxa via the delay of commitment to flowering of axillary meristems or that the 
Carnarvonia inflorescence is a plesiomorphic condition for Proteaceae derived from a 
common ancestor that had an anthotelic or blastotelic panicle. 
Developmental studies here illustrate that there are several unique features to the 
genus not found in other Proteaceae (as well as those characters cited in Johnson & 
Briggs 1975), hence providing little support for allying the taxon with any one 
proteaceous group. In a phylogenetic analysis of the family (Douglas in prep), 
Carnarvonia is in a basal position within the family or as a basal branch leading to 
Persoonioideae/Proteoideae, but the branch length of Carnarvonia is invariably long 
relative to other branches and hence there is a bit of topological plasticity. Until a 
thorough phylogenetic analysis has been undertaken, the combination of apparently 
autapomorphic features and plesiomorphic and homoplastic features (sensu Johnson 
& Briggs 1975) within Carnarvonia are interpreted here to imply a long time isolation 
from other Proteaceae and perhaps the opinion that it occupies a basal position 
relative to extant taxa is justifiable. 
Acknowledgements 
I am grateful to the Pacific-Dunlop Post-doctoral Fellowship at the Royal Botanic 
Gardens, Melbourne for supporting this research. Scanning electron microscopy was 
carried out at the School of Botany, University of Melbourne. I am also indebted to 
Jim Grimes, Barbara Briggs, Peter Weston, Alistair Hay and Peter Wilson for critically 
evaluating, and offering valuable insights into earlier copies of this manuscript. 
References 
Abbe, E.C. (1974) Flowers and inflorescences of the 'Amentiferae'. Bot. Rev. 40: 160-261. 
Armstrong, J.E. & Douglas, A.W. (1989) The ontogenetic basis for corolla aestivation in 
Scrophulariaceae. Ball. Torr. Bot. 116(4): 378-389. 
Barlow, P.W. (1989) Meristems, metamers and modules and the development of shoot and root 
systems. Bot. ]. Linn. Soc. 100: 255-279. 
Bravais, L. & Bravais, A. (1837) Essai sur la disposition symmetrique des inflorescences. Ann. Sci. 
Nat., sen 2, Bot. 7. 
