Leaves of <i>Podozamites</i> and <i>Pseudotorellia</i> from the Early Cretaceous of Mongolia: stomatal patterns and implications for relationships

Strap-shaped, parallel-veined leaves of <i>Podozamites</i> and <i>Pseudotorellia</i> are among the most common fossils in Late Triassic to Early Cretaceous floras from the Northern Hemisphere. <i>Podozamites</i> is considered to be the leaf of a voltzian conifer, but its cuticle is poorly known, whereas <i>Pseudotorellia</i>, with a thick cuticle and haplocheilic stomata, is commonly considered to be the leaf of a ginkgoalean. Here we describe the leaf morphology, cuticle and stomata of <i>Podozamites</i> and <i>Pseudotorellia</i> based on excellently preserved material from the Early Cretaceous of central Mongolia. <i>Podozamites harrisii</i> sp. nov. has transversely oriented, paracytic (probably syndetocheilic) stomata that are regularly arranged in longitudinal files. <i>Pseudotorellia resinosa</i> sp. nov. and <i>Pseudotorellia palustris</i> sp. nov. have scattered, longitudinally oriented stomata in which the two guard cells are sunken and surrounded by 2–5 specialized lateral subsidiary cells and 1–3 unspecialized polar cells. Association evidence and similarities in cuticular structure suggest that <i>Podozamites harrisii</i> was produced by the same plant as the seed cone <i>Krassilovia</i>. The distinctive stomatal pattern of <i>Podozamites harrisii</i> and <i>Krassilovia</i> is also seen in some species of <i>Swedenborgia</i> and <i>Cycadocarpidium</i>, suggesting these plants may all belong to the same natural group. <i>Cycadocarpidium</i>, <i>Krassilovia</i> and <i>Swedenborgia</i> have previously been treated as conifers, but their transversely oriented, paracytic stomata hint instead at a possible relationship with Bennettitales and Gnetales.