From seven of the eight studied coalified ovules (Trigonocarpus grandis: Sydney Coalfield, Canada) sufficient material could be macerated (Schulze's process) for histochemical investigation. This encompasses histological identification of the ovular structure/tissue components by methods of Nomarski phase-contrast microscopy, and determination of the chemical make-up by Fourier transform infrared (FTIR) spectroscopy. The generated data are then input for principal component analysis (PCA), based on the chemometric approach. Not included in PCA, but complementary to it, are data from pyrolysis gas chromatography/mass spectrometry (Py-GC/MS), powder X-ray diffraction, carbon 13 magnetic resonance analyses ((CNMR)-C-13), and introducing mass spectrometric data of selected epidermal/nucellar and vitrain samples. Addressed amongst other questions are evolution of ovular chemical grouping which includes vitrain and cutin; if coalified ovules reflect optimally original Carboniferous seeds, and why; and can chemotaxonomy/systematics of medullosalean ovules be advanced through histochemistiy? Demonstrably preserved in T. grandis are outer and inner integumentary epidermises, a double-walled nucellus with nucellar cuticles, and endospermous tissue. These structures are protected by tecta or nucellar cuticles. Molecular structures for epidermises and nucellii are probably not the same which is suggested by initial mass-spectrometric experiments. These "hard" parts are most resistant to diagenetic influences, correlating with aliphatic (lipid) composition, but facies changes influenced fossilization as in ovular molds/casts vitrain lost all its otherwise preserved tissues. This collectively suggests a narrow window of fossilizing conditions by coalification. Inferred from PCA are transitional changes, rather than sharp delineation, where the nucellus occupies a chemical composition intermediate between epidermis/cutin and the vitrain. Integumentary fibers, tectum, inorganic replacement of an epidermis, and some nucellar specimens are difficult to group by PCA. Nucellar material is probably suited for chemotaxonomic/systematic research because of the lipid chemistry. (C) 2013 Elsevier B.V. All rights reserved.