1. Introduction

The amber-embedded flower Tropidogyne pikei was previously described from the mid-Cretaceous Noije Bum 2001 Summit Site in Myanmar and tentatively assigned to the family Cunoniaceae, a largely Southern Hemisphere family of 27 genera distributed from South America and islands of the Caribbean to southern Africa, Madagascar, Australia, New Guinea, the Malay Peninsula, and islands of the South Pacific (Bradford et al. 2004; Chambers et al. 2010). Tropidogyne pikei has five spreading, epigynous sepals with five major veins and branched, anastomosing veinlets. Its inferior ovary has ten distinct ribs, each of which shows a darkly pigmented, possibly glandular area at the apex. On top of the inferior ovary is a flat, three-lobed nectar gland from which project three short, arched styles (Chambers et al. 2010). The present paper describes another species of this genus, T. pentaptera sp. nov., based on seven newly discovered fossil flowers from the Noije Bum 2001 Summit Site. The new species has spreading, veiny sepals, a nectar disc, and a ribbed inferior ovary like the previous fossil, but it differs in having only two elongate, slender styles and in the ribs of its ovary lacking darkly pigmented terminal glands. A relationship of these two fossil species to the family Cunoniaceae is proposed, in particular to Ceratopetalum D. Don, an extant Southern Hemisphere genus known from fossils of the Tertiary Period (Barnes et al. 2001). An explanation is proposed for the fossils' disjunction from the Cretaceous continent of Gonwanaland.

2. Materials and methods

The specimen originated from the Noije Bum 2001 Summit Site amber mine excavated in the Hukawng Valley and located southwest of Maingkhwan in Kachin State (26°20′N, 96°36′E) in Myanmar. Based on paleontological evidence, this site was dated to the late Albian of the Early Cretaceous (Cruickshank & Ko 2003), placing the age at 97 to 110 Ma. A more recent study using U-Pb zircon dating determined the age to be 98.79 ± 0.62 Ma, at the Albian/Cenomanian boundary (Shi et al. 2012). Nuclear magnetic resonance (NMR) spectra and the presence of araucarioid wood fibers in amber samples from the Noije Bum 2001 Summit Site indicate an araucarian tree source for the amber (Poinar et al. 2007a). Observations and photographs were made with a Nikon SMA-10R stereoscopic microscope at 80x and a Nikon Optiphot microscope with magnifications up to 600x. Helicon Focus Pro X54 was used to stack photos for better clarity and depth of field.

Fig. 1.

Five-sepal specimens of Tropidogyne pentaptera sp. nov. examined in this study. A: Holotype. Bar = 1.0 mm. B: Paratype B Bar = 1.2 mm. C: Paratype C. Bar = 1.0 mm. D: Paratype D. Bar = 0.9 mm. E: Paratype E. Bar = 0.9 mm. F: Paratype F. Bar = 1.0 mm

3. Description

Fig. 2.

Flower of Tropidogyne pentaptera sp. nov., paratype C, showing basal fusion of sepals (arrow) and connection between lateral veins (arrowhead). Bar = 0.9 mm. Insert shows six-sepaled flower (paratype G). Bar = 0.6 mm.

Fig. 3.

Margin, venation and pubescence of a portion of a sepal of Tropidogyne pentaptera sp. nov. Paratype B. Bar = 0.1 mm.

Fig. 4.

Adnate hypanthium (arrow) showing 5 of the 10 longitudinal ribs of Tropidogyne pentaptera sp. nov., paratype F. Bar = 0.7mm. Insert: Sepal of Ceratopetalum succinibrum, showing venation similar to the fossil.

Fig. 5.

Holotype (B-An-5A) of Tropidogyne pentaptera. sp. nov. showing styles and stamen. Arrow points to filament. Anther shows distal connective protrusion. Bar = 0.5 mm. Insert shows pollen grain with verrucose exine. Bar = 8 µm.

Fig. 6.

Fruit of Tropidogyne pentaptera sp. nov. showing rounded upper portion of semi-inferior ovary. Paratype E. Bar =1.0 mm.

4. Discussion

In the publication of Tropidogyne pikei by Chambers et al. (2010), there is a detailed discussion of the similarities of this flower to those of present-day Ceratopetalum of the Cunoniaceae. The comparison works equally well for T. pentaptera, with the added features of a pair of elongated styles and a semi-inferior fruiting ovary by which these fossils resemble Ceratopetalum but differ from T. pikei. Only in its five rather than ten stamens and the slight basal connation of the sepals does T. pentaptera differ from the modern genus. Although the comparison is entirely floral, since no vegetative parts are preserved in the amber, we are emphasizing the morphological relationship to Ceratopetalum more positively here than in the earlier paper, in particular because we have evidence from another Burmese amber fossil, discussed below, that Mid-Cretaceous fossils from this amber may bear a surprising and unsuspected similarity to modern angiosperm groups that were thought to be tens of millions of years younger. Another point to be noted is that fruits of Ceratopetalum are already well known in the Cenozoic fossil record, as discussed by Gandolfo & Hermsen (2017), who found fossils referable to this genus in the Early Eocene of Patagonia and who cite previous reports of Eocene to Miocene fossils of Ceratopetalum from Australia (Barnes & Hill 1999).

Similarities of Tropidogyne to Ceratopetalum are in the regular calyx of five valvate sepals that are lanceovate, firm, and spreading, the pattern of anastomosing sepal venation (Figs. 1–4, 4 insert), the connection between lateral veins of adjacent sepals (Fig. 2, arrowhead), the lack of petals (only C. gummiferum has petals), the diplostemonous androecium (in T. pikei), the short stamens with dorsally inserted filaments (Rozefelds & Barnes 2002, fig. 5), the anthers with a conspicuous connective protrusion (Rozefelds & Barnes 2002: 656, fig. 6), the well-developed epigynous nectar disc, resembling C. gummiferum in its lack of the puberulent zone described for all other species of Ceratopetalum (Rozefelds & Barnes 2002: 656, fig. 5), the bicarpellate pistil with elongate, recurved styles (in T. pentaptera), and the inferior ovary with ten-ribbed adnate hypanthium, becoming semi-inferior in fruit (Figs. 4, 6).

The oldest fossil flower assigned to Cunoniaceae is Platydiscus peltatus from the Santonian/Campanian boundary, 83.5+/-0.7 Ma, in Sweden (Schönenberger et al. 2001). However, figures of this fossil in Friis et al. (2001: 331,438) show five superior, separate carpels, quite unlike Australian Ceratopetalum. Gandolfo & Hermsen (2017) were reluctant to attribute a relationship of Tropidogvne pikei to Ceratopetalum because of the fossil's age and its occurrence in Burma, far from Australia. However, a recently submitted paper (Poinar & Chambers, submitted 2017) sheds important new light on the Burmese amber fossils. In it, six capitula of Asteraceae, assigned by the authors to subfamily Barnadesioideae, the basal clade of the family, were found in amber from the Noije Bum Summit Site locality. This subfamily occurs only in the Southern Hemisphere and is recognized as the sister group to all other Asteraceae (Funk et al. 2009, fig. 44.3). The earliest record of the subfamily was reported by Barreda et al. (2015), based on pollen from a site in Antarctica dated to ca. 76–66 Ma. This record is itself much older than the previously estimated basal date for Asteraceae of 41–50 Ma (Funk et al. 2009, fig. 44.8), and the date of 98.79 +/- 0.62 Ma for Barnadesioideae in Burmese amber is much greater yet.

The geological history of the Malaysian region provides a possible explanation of how the unexpectedly old Burmese amber flowers of Barnadesioideae and Tropidogyne became separated from related Asteraceae and Cunoniaceae of the Southern Hemisphere. Malaysia, including Burma, was formed during the Paleozoic and Mesozoic Eras by subduction of terranes that successively separated from Gonwanaland and were transported by continental drift northward across the Tethys Sea (Hall 2012; Metcalfe 2013). The last of these parts of Gonwanaland to move northward was the India block, including Greater India Parts 1 and 2, which had been adjacent to western Australia (Hall 2012, fig. 26; Metcalfe 2013, fig. 17). When this terrane reached Asia in the Middle Eocene, the northeastern part was adjacent to Burma (Metcalfe 2013, fig. 28). We suggest that if the Noije Bum Summit Site amber deposit was part of Greater India when it sutured to Burma, both Tropidogyne and the Barnadesioid fossils would have originated in the late Early Cretaceous, when this fragment of Gonwanaland was adjacent to Australia. This proposal of a Southern Hemisphere origin explains the morphological similarity and, we believe, the taxonomic relationship of Tropidogyne to Ceratopetalum, leading to the conclusion that Ceratopetalum-like species of Cunoniaceae were present in Gonwanaland in the late Early Cretaceous. Our proposal is similar to one by Prasad et al. (2005) for the grass family Poaceae. Based on a diversity of grass phytoliths in Late Cretaceous strata in India, these authors suggest that “crown-group Poaceae had diversified and spread before India became geographically isolated,” that is, when it was part of Gonwanaland.

Despite the previous lack of fossil evidence, an early occurrence of major groups such as Asteraceae and Cunoniaceae must now be accepted because of the wellpreserved specimens discovered in Myanmar amber. Other examples may include previously described fossils with affinities to the families Monimiaceae (Poinar & Chambers 2005), Cornaceae (Poinar et al. 2007b) and Lauraceae (Poinar 2017).