Megahertzia, A. S. George & B. Hyland (George and Hyland, 1995

Carpenter, Raymond J. & Rozefelds, Andrew C., 2023, Leaf fossils show a 40 - million-year history for the Australian tropical rainforest genus Megahertzia (Proteaceae), Australian Systematic Botany 36 (4), pp. 312-321 : 314-318

Notes and justification for assignment to Megahertzia in Proteaceae

Christophel et al. (1987) confirmed their placement of the fossil leaves in Proteaceae by reporting the presence of paracytic stomata, an apparent synapomorphy for the family ( Carpenter et al. 2005), and complex trichome bases.

Amplexicaul (i.e. stem-clasping) and lobed leaves are a highly distinctive feature of the single extant species of Megahertzia . To our knowledge, no other species of Proteaceae have such leaves; amplexicaul but highly dissimilar leaves occur only among a few species of nonrainforest habitats, perhaps most notably Hakea amplexicaulis R.Br. of Western Australia. This species and closely related taxa in the Prostrata group of Hakea ( Barker et al. 1999) , which is strongly supported by molecular analyses ( Cardillo et al. 2017), all show amplexicaul or auriculate leaves ( Barker et al. 1999), a state that is unusual in Hakea and which must be considered synapomorphic for the Prostrata group. Amplexicaul leaves are not known in the putative Hollandaea + Helicia sister group of Megahertzia , which has over 100 species, nor elsewhere in Roupaleae . We are therefore confident that amplexicaul leaves in M. amplexicaulis are autapomorphic among at least the extant members of Roupaleae . The presence of this state in the fossils not only allows close comparison with M. amplexicaulis , but also that this trait can be regarded as diagnostic at genus level, not just for M. amplexicaulis as treated by George and Hyland (1995).

All the fossil Megahertzia leaves are three (rarely five)- lobed, and entire or with sparse and minute teeth, exactly as for many juvenile leaves of M. amplexicaulis , although

the modern foliage may be larger and may have more prominent teeth ( Fig. 3). However, the smaller leaf size of the fossils may in part be due to collection biases; Rowett and Christophel (1990) noted that leaves were particularly brittle and difficult to recover as wholly or mostly intact specimens, but that some were observed to exceed 30 cm in length. This is approximately twice the length of specimens examined in the current study, and consistent in size with some extant specimens held at BRI.

The cuticle of M. amplexicaulis ( Fig. 3) (as ‘ Orites code #752’) was described by Carpenter (1994). It shows quite rare trichome bases that are associated with numerous (commonly eight) basal epidermal cells, sinuous to buttressed anticlinal cell-wall cuticle, prominent and finely striated outer cuticle surfaces and granular inner cuticle surfaces. As has been demonstrated ( Fig. 2), the same features are found in the fossil cuticles, although there is typically less evidence of adaxial striations and inner granulations, and some specimens show many more cells associated with the trichome bases (compare these features in Fig. 2 and 3). There is also variation in epidermal pavement cell anticlinal wall outlines, with some specimens showing more-or-less straight walls, and others sinuous and slightly buttressed walls. This variation is likely to reflect litter input to depositional sites from different light levels within the local Anglesea vegetation, because it is well known that, in shade, these cells will have more undulated outlines than do those exposed to full sun ( Watson 1942; Kürschner 1997; Dunn et al. 2015). In conclusion, we regard all the above differences as minor, and moreover note that among species of all Proteaceae genera studied by the first author, only M. amplexicaulis has a cuticular morphology that is consistent with that of the fossils. On this basis, referral of the fossil leaves to the extant species is possible, but this option was rejected in consideration of the massive time separation and absence of supporting evidence. Of possible future relevance here is that Christophel (1984, p. 184) mentioned that two proteaceous inflorescences of unknown affinities had been recovered from the Anglesea clays, as well as a Grevillea -like follicle. Unfortunately, the fate of these fossils is unknown, but it is noted that the fruits of M. amplexicaulis are readily comparable with those of at least the larger-fruited species of Grevillea ( Makinson 2000) , being 30–35 mm long dehiscent follicles ( George and Hyland 1995).

In Proteaceae , the foliage of many taxa changes throughout development, including transitions from lobed juvenile and new shoot forms to simple adult leaves ( Johnson and Briggs 1975). However, beyond the seedling stage, only a few extant species, in subfamily Grevilleoideae , are known to produce types of pinnately lobed leaves with occasional teeth similar to those in Megahertzia , notably Orites excelsus R.Br. (tribe Roupaleae ) and Athertonia diversifolia (C.T.White) L.A.S.Johnson & B.G.Briggs ( Johnson and Briggs 1975) of tribe Macadamieae subtribe Virotiinae . None of these species has remotely similar leaf cuticular morphology to that of Megahertzia , with for example O. excelsus and A. diversifolia having obviously waxy surfaces ( Carpenter 1994). The fossil-leaf species Maslinia grevilleoides D.T.Blackburn from the Eocene Maslin Bay site in South Australia is of a size similar to Megahertzia leaves and is also lobed with sparse, small teeth ( Blackburn 1981). Unfortunately, the single specimen was described as lacking a base, and its cuticular features were not clearly illustrated. However, Blackburn made no mention of large trichome bases, and instead described ( Blackburn 1981, p. 19) ‘indistinct’ trichome bases ‘with one or two hair base support cells’ that are ‘frequently associated with stomates’. Such trichome bases do not occur in Megahertzia . Another similar lobed and toothed taxon from the Early to Middle Eocene of New Zealand is ‘cf. Orites excelsa ’ ( Pole 1994) . These specimens cannot be closely compared with Megahertzia because they lack cuticle, but they appear to be distinct in having acute leaf bases. The Early Eocene Patagonian Lomatia occidentalis (Berry) Frenguelli also lacks cuticular preservation, but is anyway distinct in having imparipinnate or pinnatisect, clearly serrate leaves with lateral leaflets or lobes incised nearly to the midvein ( Frenguelli 1943; Gonzalez et al. 2007).