Hermans, R.M. orcid.org/0000-0003-1675-191X, Gallaher, T.J. orcid.org/0000-0002-5781-6874, Brightly, W.H. orcid.org/0000-0002-0012-5183 et al. (3 more authors) (2026) Reconstructing Cenozoic grasslands in North America and western Eurasia: evidence for cold‐season grasses from 3D phytolith morphology. New Phytologist. ISSN: 0028-646X
Abstract
Pooideae, the largest grass subfamily, radiated in the Northern Hemisphere from the Paleocene onward and today dominates cold-season and high-altitude grasslands. Fossil phytolith assemblages suggest pooid-rich ecosystems during the Oligocene–Miocene, but the lack of quantitative phytolith identification has limited reconstruction of early grassland composition and ecology.
We combined 3D geometric morphometrics of grass silica short-cell phytoliths (GSSCPs) with machine-learning classification and phylogenetically informed climate modelling. A dataset of modern GSSCPs was used to classify Eocene–Miocene phytoliths from Türkiye and the US Great Plains and reconstruct paleoclimate.
3D GSSCP morphology distinguished modern grass subfamilies and tribes. Fossil assemblages contained mainly phytoliths assigned to Bambusoideae and Pooideae (including Poeae I and II). In modern grasses, Crenate and Polylobate phytolith dimensions correlated with temperature and precipitation. Fossil Crenates and Polylobates yielded climate estimates warmer and wetter than today. Ancestral state reconstructions indicated milder winter conditions at early Pooideae nodes, with cold tolerance arising early.
Major temperate grass lineages (Poeae I and II) were established earlier than previously recognized, and Pooideae-rich grasslands were present by the Oligo-Miocene under warmer, wetter climates than today. These findings demonstrate the value of 3D phytolith morphometrics for interpreting early grassland evolution.
Metadata
| Item Type: | Article |
|---|---|
| Authors/Creators: |
|
| Copyright, Publisher and Additional Information: | © 2026 The Author(s). New Phytologist © 2026 New Phytologist Foundation. This is an open access article under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits use, distribution and reproduction in any medium, provided the original work is properly cited. |
| Keywords: | 3D geometric morphometrics; Eocene–Miocene; Pooideae; grass evolution; grasslands; machine learning; paleoecology; phytoliths |
| Dates: |
|
| Institution: | The University of Sheffield |
| Academic Units: | The University of Sheffield > Faculty of Science (Sheffield) > School of Biosciences (Sheffield) |
| Date Deposited: | 03 Nov 2026 16:32 |
| Last Modified: | 03 Nov 2026 16:32 |
| Status: | Published online |
| Publisher: | Wiley |
| Refereed: | Yes |
| Identification Number: | 10.1111/nph.71687 |
| Related URLs: | |
| Open Archives Initiative ID (OAI ID): | oai:eprints.whiterose.ac.uk:247472 |

CORE (COnnecting REpositories)
CORE (COnnecting REpositories)