Spore masses and isolated sporangia, containing laevigate hilate cryptospores attributable to the dispersed taxon Laevolancis divellomedia sensu lato, have been recovered on bulk maceration of Upper Silurian (Pridoli) and Lower Devonian (Lochkovian) deposits from the Welsh Borderland. Detailed morphological, anatomical and ultrastructural analysis, using light microscope, scanning electron microscope and transmission electron microscope techniques, reveals subtle differences between the specimens and they can be grouped into five distinct types. The different groups are distinguished principally by using sporangia–spore mass characteristics, presence or absence of extra–exosporal material and nature of spore–wall ultrastructure. Of the groups, one has a uniformly homogeneous exospore and the other four groups have a bilayered exospore. In the former the spores lack extra–exosporal material and occur in a discoidal sporangium. Of the bilayered groups, two have exospores of homogeneous composition but with the two layers differing in electron density. They occur in discoidal sporangia and spore masses and are distinguished on the presence or absence of extra–exosporal material and differences in the widths of the two layers. Finally, two bilayered groups possess a lamellate inner layer, but vary in presumed sporangial shape. Elongate sporangia have spores with concentric continuous lamellae, lacking further ultrastructure. In contrast, spores from a discoidal spore mass have white–line–centred, presumably tripartite, lamellae which are laterally discontinuous, overlapping and irregularly spaced. These findings, which suggest that morphologically similar spores were produced by a number of plant taxa, have important implications regarding the assessment of early land–plant diversity. The affinities of hilate cryptospore–producing plants are unknown and problematic, particularly as no extant non–angiosperm plants produce dyads, other than through meiotic irregularity, and spore–sporangial characters have no exact counterpart in coeval plants. Studies of specimens with in situ hilate cryptospores suggest that they derive from rhyniophytoids, i.e. plants that resemble the simplest of vascular plants but lack evidence of vascular tissue, although hilate cryptospore–containing examples show no axial branching. It might be argued, based on evidence from spore wall ultrastructure, that some of the plants have more in common with lycopsids and filicopsids than bryophytes, a surprising finding bearing in mind the stratigraphic distribution of hilate cryptospores–dyads and inferences that the producers were bryophyte–like. Detailed studies of wall structure in the hilate cryptospores permit consideration of spore wall development. It is suggested that extra–exosporal material derives from a tapetum and is thus produced by the diploid sporophyte. The white–line–centred lamellae in a single specimen provide the earliest evidence for the presence of such structures in early land plant spores and provide further evidence that sporopollenin deposition on such structures is the most primitive mode of sporopollenin deposition among land plants.