The Core Eudicot Boom Registered in Myanmar Amber

A perfect flower in a mid-Cretaceous (early Cenomanian) Myanmar amber is described as Lijinganthus revoluta gen. et sp. nov. The fossil flower is actinomorphic and pentamerous, including calyx, corolla, stamens, and gynoecium. The sepals are tiny, while the petals are large and revolute. The stamens are dorsifixed, filamentous, and each has a longitudinally dehiscing bisporangiate anther. The gynoecium is in the centre of the flower, composed of three fused carpels with a stout style. Lijinganthus revoluta gen. et sp. nov. demonstrates a great resemblance to the flowers of Pentapetalae (Eudicots), adding new information to the enigmatic early evolutionary history of Pentapetalae and Eudicots.

Several clumps of a single type of pollen grains are present in the same amber block and closely associated with the flower (Figs 3i,m and S1a-q). Pollen grains are tricolpate, 14-21 μm long and 10-12 μm in diameter ( Fig. 3i-l). Nectary surrounds the base of the ovary (Fig. 3g-h). The gynoecium is situated centrally, including an ovary that tapers distally with no obvious style (Figs 3g,h and 4a,b). The ovary appears to be composed of three fused carpels, probably with axile placentation (Fig. 3g). The ovary is approximately 0.77 mm in diameter, with its tip only 0.13 mm in diameter, and round triangular in cross view (Figs 2b,c, 3g,h, 4b and 5).
Etymology: The generic name, Lijinganthus, is dedicated to poetess Ms. Jing Li  for her talent and poetry, and "anthus" from the Latin "anthos" (flower). The specific epithet revoluta is for the revolute form of the petals. Remarks: Although only eight stamens are preserved in the flower, the spatial relationship between stamens and petals (Fig. 2d) and the presence of five petals (Fig. 2a-c) suggest that the total number of original stamens in Lijinganthus should be ten. Further research need to be done to confirm this.
The character combination of the fossil does not allow us to assign it to any known fossil or extant genus of angiosperms, thereby justifying a new genus.

Discussion
Comparison with extant angiosperms. The occurrence of over 800 single type of tricolpate pollen grains closely associated with Lijinganthus revoluta gen. et sp. nov. (preserved in the same block, and their distances from the flower range from 2.5 to 0 mm (Figs 3i,m and S1a-q)) strongly suggests a eudicot affinity for Lijinganthus, as tricolpate pollen grains are a characteristic feature of Eudicots, which are frequently termed Tricolpates. Among Eudicots, the occurrence of distinct calyx and corolla in Lijinganthus distinguishes it from the basal eudicots (with undifferentiated perianth) 30 and Gunnerales (lacking perianth) 35 , suggesting that Lijinganthus belongs to the Pentapetalae in Core Eudicots. Several features of Lijinganthus are also seen in Crassulaceae (Saxifragales), but the latter has a basifixed anther, >3 more or less free carpels, decurrent stigma, and parietal placentation 36,37 and is thus distinct from Lijinganthus with dorsifixed anther, fused carpels, capitate stigma, and axile placentation. Although the presence of distinct sepals and petals suggest that Lijinganthus is most likely related to the Superrosids, and its tricarpous gynoecium suggests a possible affinity to Malpighiales, we think it is premature to assign Lijinganthus to any group within Pentapetalae, at least for the time being.
Comparison with the Early Flowers of Core Eudicots. Previously, the earliest record of a flower with distinct calyx and corolla was marked by a fossil flower named "Rose Creek flower" from the Albian-Cenomanian 30 , which was recognized by Basinger and Dilcher in 1984 29 and reworked on and renamed as Dakotanthus cordiformis by Manhcester et al. (2018) with additional specimens (especially of fruits) using CT technology 27 .
Lijinganthus is similar to Dakotanthus cordiformis in distinct pentamerous symmetry (5 sepals and 5 petals) and bisexuality, but it differs from the latter in long (rather than ovate) petals, a slender (rather than stout) filament, a bisporangiate (rather than tetrasporangiate) anther, tricolpate (rather than tricolporate) pollen grains, a trimerous (rather than pentamerous) gynoecium, and 1 (rather than 5) style 27,29 . Dakotanthus cordiformis is initially interpreted as approximately 94 Ma old 29 , but recent study, with more specimens (especially of fruits) from strata other than the original locality (Rose Creek), suggests that the age of Dakotanthus cordiformis may be extended to 105 Ma 27 . Dakotanthus cordiformis was considered to be "the first fossils with unequivocal features of core eudicots" from the Albian-Cenomanian 30    ) and Lijinganthus at about the same time 19,25-28 seems to suggest the Core Eudicots underwent a rapid diversification ("Core Eudicot Boom") at the very beginning of the Late Cretaceous (Table 1).
It is intriguing to investigate whether there is a coupling between this important plant event and the Upper Albian OAE 1d event (including rapid CO 2 concentration rising) 38 as well as the decline of Gnetales and Bennettitales. The presence of nectary disk in Lijinganthus suggests that insects may have begun interacting with flowers by the Cenomanian. Whether such an interaction is a major driving force for the diversification of Core Eudicots is apparently a question deserving further investigation.
Earlier Origins. Various molecular clock studies have indicated that Eudicots and angiosperms originated much earlier than formerly assumed [31][32][33] . The discovery of Lijinganthus from the late Albian-Early Cenomanian  (98.79 Ma) adds to the diversity and abundance of early Core Eudicots and helps to reconcile the conflicts between different schools and studies. Compatible with the molecular studies, the early age of Lijinganthus, together with recently found earlier-than-recorded fossils of Poaceae 39 and Solanaceae 40 as well as various contemporaneous Core Eudicots 19,25-28 points to a cryptic and unexpectedly longer history of angiosperms and Eudicots. It is noteworthy that this conclusion is compatible with previously documented pre-Cretaceous traces of angiosperms [41][42][43][44][45][46][47][48] .

Conclusions
Lijinganthus revoluta gen. et sp. nov. and other contemporaneous fossil flowers suggest a Core Eudicot Boom at the very beginning of the Late Cretaceous. Increasing number of reports of early fossil flowers seem to converge earlier origins of various lineages that have been predicted by molecular clocks.

Methods
The specimen was collected from Noije Bum 2001 Summit Site, Hukawng Valley, Kachin, Myanmar (26°20′N, 96°36′E) (Fig. 1). Paleontological studies indicate that the specimen belongs to the earliest Cenomanian-latest Albian, Early Cretaceous (98.79 Ma) 34 , which is generally agreed on by Poinar et al. 28 and Xing et al. 49 . Two parallel planes were made on the amber sample before observations. Observations and photographs were made with a Nikon SMZ1500 stereoscopic microscope at the Nanjing Institute of Geology and Palaeontology, Nanjing, China. Micro-CT was performed using a Zeiss Xradia 520 versa X-ray microscope at the Nanjing Institute of Geology and Palaeontology, Nanjing, China. The 3D reconstruction and virtual sections were generated using VGStudio MAX 3.0. All figures were organized for publication using Photoshop 7.0.

Data Availability
The holotype (PB22841) is accessible in the palaeobotanical collection of the Nanjing Institute of Geology and Palaeontology, Chines Academy of Sciences, 39 Beijing Dong Road, Nanjing 210008, China.