Our dataset includes 115 leaf cuticles related to two species of Pseudotorellia Florin from three stratigraphically similar samples at the Tevshiin Govi lignite mine in central Mongolia (~119.7–100.5 Ma, Aptian–Albian, Cretaceous). We apply a well-vetted paleo-CO₂ proxy based on leaf gas-exchange principles (the Franks model) to those leaves for paleo-CO₂ reconstruction, which requires leaf stomata and carbon isotope analysis. All cuticle measurements are summarized in this dataset. 

The leaves studied here are from the Tevshiingovi lignite mine in central Mongolia, approximately 220 km south-southwest of Ulaanbaatar. We excavated lignite blocks from three stratigraphically similar samples from lignite #2 along the Tevshiin Govi lignite mine wall. The blocks were disaggregated with detergent, 3% hydrogen peroxide, and deionized water. Because chemical treatments can affect δ¹³C measurements (Barral et al., 2015), we measured one leaf (cut in half) both with and without peroxide treatment and found an isotopic difference of 0.22 ‰, which propagates to a small change in estimated CO₂ (~2.6 %). The lithological column was illustrated using SDAR package version 0.9-55 (Ortiz and Jaramillo, 2020).

Leaves were first cut in half, with one half set aside for isotope analysis. The remaining half was cleared following standard practices (Dilcher, 1974; Wang, 2010; Liang et al., 2022). Fragments were soaked in deionized H₂O (dH₂O) for 1 hour. Softened leaf fragments were cut at the margin and then cleaned using dilute household bleach (c. 1% sodium hypochlorite solution (NaClO)). Once the leaves turned transparent (or light yellow), the reaction was immediately terminated by neutralizing with dH₂O before being transferred to a microscope glass slide. A fine dissecting needle was used to separate the leaf into the upper (adaxial) and lower (abaxial) sides, each containing one epidermis and part of the mesophyll tissues. They were then gently washed with dH₂O twice to remove loosened mesophyll tissue. 

Clean abaxial epidermises were observed with a Leica DMLB microscope under transmitted light and epifluorescence with mercury light and a 420–490 nm filter cube. ZStack photos of four different focal planes were taken with a Jenoptik Gryphax digital microscope camera and then integrated into a single composite image. We measured stomatal density mostly at 100× (0.80 mm² field-of-view), but occasionally at 200× (0.20 mm² field-of-view) when lower-quality preservation limited large fields-of-view.  The stomata in both species appear randomly distributed. 

For each leaf fragment, we measured 10 guard cells for their length and width as well as the length and width of the bounded stomatal pore. These images were taken at 400× and measured using ImageJ (Rasband 1997). Both species of Pseudotorellia are hypostomatous and have sunken guard cells. Pseudotorellia resinosa possesses thickened subsidiary cells that partly overlap the underlying guard cells (Shi et al. 2018; see their Fig. 16C-E). These thickened areas are prominent under both transmitted light and epifluorescence, and comprise a smaller footprint than the underlying guard cells (see representative lines for guard cell length [gcl] and single guard cell width [sgw]. 

Bulk-leaf δ¹³C values were measured at the Light Stable Isotope Mass Spec Lab at the Department of Geological Sciences of the University of Florida using a Thermo Electron DeltaV Advantage isotope ratio mass spectrometer coupled with a ConFlo II interface linked to a Carlo Erba NA 1500 CNHS Elemental Analyzer. After combustion in a quartz column at 1000 °C in an oxygen-rich atmosphere, the sample gas was transported in a He carrier stream and passed through a hot reduction column (650 °C) consisting of elemental copper to remove oxygen. The effluent stream then passed through a magnesium perchlorate trap to remove water followed by a 0.7 m GC column at 125 °C to separate N₂ from CO₂. The sample gas next passed into a ConFlo II interface and into the inlet of a Thermo Electron Delta V Advantage isotope ratio mass spectrometer running in continuous flow mode where the sample gas was measured relative to laboratory reference CO₂ gas. All carbon isotopic results are expressed in standard delta notation relative to VPDB. Instrument error across the multiple runs ranged from 0.01-0.19 ‰. We analyzed 36 samples in duplicate, with an average difference of 0.33 ‰ (0.06–1.03 ‰ range).