Technical Comments

Response to Comment on “Eocene Fagaceae from Patagonia and Gondwanan legacy in Asian rainforests”

See allHide authors and affiliations

Science  15 Nov 2019:
Vol. 366, Issue 6467, eaaz2297
DOI: 10.1126/science.aaz2297


Denk et al. agree that we reported the first fossil Fagaceae from the Southern Hemisphere. We appreciate their general enthusiasm for our findings, but we reject their critiques, which we find misleading and biased. The new fossils unequivocally belong to Castanopsis, and substantial evidence supports our Southern Route to Asia hypothesis.

We recently (1) reported two Castanopsis rothwellii fossil infructescences from the early Eocene [52 million years (Ma) ago] of Argentine Patagonia. These are (we maintain) the oldest fossils assigned to the genus by ~8 Ma (2, 3), and they co-occur with hundreds of fagaceous leaves indistinguishable from those of living Castanopsis. The same fossil beds contain numerous taxa whose close living relatives characteristically associate with Castanopsis in New Guinea and elsewhere, including Papuacedrus, Agathis, Araucaria Sect. Eutacta, Dacrycarpus, a Phyllocladus relative (4), Podocarpus, Retrophyllum, Ripogonum, Eucalyptus, Ceratopetalum, Gymnostoma, engelhardioid Juglandaceae, and Todea, as cited (1). Nearly all these lineages are well-known examples of the Southern Route to Asia confirmed by fossil evidence from one or more of Antarctica, Australasia, and Asia (57), and we concluded that Castanopsis most likely had similar biogeographic history. Castanopsis thrives on the Australian plate today in New Guinea, and its southern range is only a short distance over shallow water from Australia, with which New Guinea had frequent past land connections and biotic interchanges (8).

In short, we presented a suite of positive evidence for our Southern Route hypothesis that led us to favor the idea. Most notably, we reported the first remains of Fagaceae trees that grew on Gondwana, clearly identifiable as Castanopsis and found in a fossilized New Guinea–type association. However, Denk et al. (9) assert that “evidence for such a pathway is currently missing.”

First, we reject Denk et al.’s appeal-to-authority argument: “…the southern route hypothesis would require that generations of palynologists had overlooked the characteristic pollen of Castaneoideae in Gondwanan records.” This statement appears biased regarding both South America, an integral part of Gondwana 52 Ma ago (10), and those who work there. The current “generation” has found fossil Castaneoideae in Gondwana (1), even though previous highly skilled colleagues had not.

Second, Denk et al. invoke a moving-the-goalposts argument. Even though we just reported Fagaceae fossils several thousand kilometers south of any previous occurrences (1), Denk et al. hold that the family did not range any farther south (their South American “dead end”). Theirs is a perilous position because there was no oceanic separation of South America and Antarctica 52 Ma ago, and thus no “end” to South America (10); instead, abundant austral biotic interchange took place at that time (11). The South American “dead end” would also require that Castanopsis went extinct in the Southern Hemisphere, then rejoined the same New Guinea–type lineages at a much later date via an entirely different Holarctic path after crossing several climate zones. This scenario seems far less likely than the Southern Route, which is supported by our finding the oldest Castanopsis already in a perhumid New Guinea–type rainforest association 52 Ma ago in West Gondwana (1). It is irrelevant that Castanopsis occurs in other plant associations in its living range (9), as we discussed (1).

Third, Denk et al. wrongly assert that molecular data from living Fagaceae can be used to reject a hypothesis about the affinities of specific early Cenozoic fossils determined from paleobotany (“molecular data reject the notion that…”). In so doing, Denk et al. also mistakenly state that Castanea “lacks a fossil record outside Eurasia.” The oldest fossil Castanea, as we cited (1), is from the middle Eocene of Tennessee, not Eurasia. Thus, shared sequences from the sister genera Castanea and Castanopsis in all likelihood reflect their common ancestry in the ancient New World, not the Old World as Denk et al. argue.

Fourth, Denk et al. erroneously contend that Castanopsis rothwellii, a fossil with so many diagnostic characters preserved that it could only be assigned to Castanopsis if “found alive” today (1), has plesiomorphic features and cannot be placed confidently in the extant genus. Their idea rests on a misleading phylogenetic argument (see below), and it is unacceptable at face value because it ignores basic botany and our detailed taxonomic treatment (1). The diagnostic characters of Castanopsis in the fossils are in no way generalized for all Fagaceae, including the spike-like infructescence axes of numerous solitary, asymmetrical, valved, and sutured lateral cupules that entirely enclose the single nut, which retains three short, linear, “castaneoid” styles with unexpanded stigmas. These features match Castanopsis precisely and definitely exclude the fossils from placement with Quercus, Fagus, and the trigonobalanoids. Within the remaining castaneoid genera, C. rothwellii only matches Castanopsis, and thus there is no basis whatsoever for separating this fossil from Castanopsis. Denk et al. also pose an invalid syllogism by arguing that because Castanopsoidea and C. rothwellii have some similar features, and the former is an extinct genus, then our fossils do not belong in an extant genus.

Denk et al.’s phylogenetic conclusions from their emended tree and matrix are misleading, in that any morphological matrix includes characters that are relevant only for the taxa included in the analysis. Because the fossils are castaneoid in all features, we did not include all Fagaceae in our original analysis (1) and likewise did not include all characters relevant to non-castaneoid fagaceous taxa. Denk et al. added several genera to their analysis without adding any morphological characters to resolve these additional taxa, then used their uninformative result to criticize our phylogenetic interpretation as uninformative. In their framework, even a living Castanopsis would not resolve in Castanopsis! By adding just three relevant characters to the Denk et al. scaffold to accommodate the genera they added (Table 1), the fossil Castanopsis rothwellii is placed only with Castanopsis in the single most parsimonious tree (Fig. 1). We note that even when the same morphological data are used alone, without any scaffold, the fossil resolves with the Castanopsis fissa group. We acknowledge our miscoding of Fagus for flower number, a typographic error that does not affect the outcome of any of our analyses. The other character re-codings by Denk et al. only make the morphological data less precise.

Table 1 Additional character scores for phylogenetic analysis.

Scores are shown in the following order (left to right). Stigma: expanded = 0; unexpanded = 1. Nut in cross section: triangular/flattened = 0; generally rounded = 1. Cupule symmetry: symmetrical = 0; asymmetrical = 1.

View this table:
Fig. 1 Phylogenetic analysis.

Consensus of the two most parsimonious trees, based on the Denk et al. (9) scaffold and emended morphological matrix with the addition of the three characters listed in Table 1, generated using the same analytical methods described previously (1). See text for discussion.

We expected vigorous debate regarding the biogeographic implications of our Gondwanan Castanopsis fossils, which hold importance for understanding and conserving the imperiled southern-sourced associations that survive in Asian rainforests (1, 6). Unfortunately, Denk et al. do not advance the discussion. Only time and many more fossils, not negative evidence and misleading assertions (9), will tell where else the Fagaceae occurred.


Acknowledgments: Funding: Supported by NSF grants DEB-1556666, DEB-1556136, EAR-1925755, and EAR-1925552 (P.W. and M.A.G.). Author contributions: All authors contributed to researching and writing this response.
View Abstract

Navigate This Article