1 A long-recognized problem in paleobotanical studies is the fragmentary nature of plant fossils (compression/ impression), which can pose near-insurmountable taxonomic problems (Sakala 2004; Bateman and Hilton 2009). In the Euramerican Carboniferous, detached ovular seed-fern structures are no exception (Arnold 1938, p. 209). In addition, compression/impression/permineralized ovules connected to other plant parts are extremely rare, making taxonomic assignment of isolated ovules a challenge (e.g., Seyfullah et al. 2010). In the Sydney Coalfield in Nova Scotia, fossil-plant collecting by many palaeobotanists and geologists over the past 200 years has yielded one intact 12-mm-long trigonocarpalean compression/impression ovule that is attached to an ultimate neuropterid pinna (Zodrow and McCandlish 1980; Luthardt et al. 2021, p. 5). In contrast, 30 or more detached single trigonocarpalean compression/ impression seeds up to 120 mm long and superbly preserved have been discovered (Zodrow et al. 2014; Zodrow et al. 2013; D’Angelo and Zodrow 2011; Cleal et al. 2010; Zodrow 2002, 2004). These fossils were collected from a rock pile that derives from the ca. 2-m-thick roof shale of the Lloyd Cove Seam, which was mined in an open pit near Point Aconi (Zodrow and Mastalerz 2019, fig. 1; Zodrow 2002). It is noteworthy that equally well-preserved, compressed male fructifications of Dolerotheca Halle 1933, containing the largest-known oval prepollen grains at over 800 μm, were also collected from the same location (Zodrow and Mastalerz 2019). Together, these fossils potentially provide insights into the male and female organs of the same whole plant species.

2 The specimen documented here is a cupule containing an ovular structure that is the only one known from the Sydney Coalfield, and indeed from the Carboniferous of the Maritimes. However, questionably identified cupules without preserved seeds, for example Pterispermostrobus bifurcatus Stopes 1914 from the Upper Carboniferous “Fern Ledges”, Saint John, New Brunswick (Stopes 1914, fig. 15; pls. XVII, 45 and XXV, 69). resemble the presently described specimen.

3 The tiny detached coalified structure described here is approximately 10 mm wide and 8 mm high, and consists of a single 3×2 mm ovule (Fig. 1). It was collected by the author from the rock pile of silty roof rocks of the Harbour Seam, at the old Lingan Mine dump, near Glace Bay, Cape Breton Island, Nova Scotia (46.19°N, 59.95°W). Accordingly, the chronostratigraphic age is middle Westphalian D sensu Zodrow and Cleal (1985), equivalent to the upper Moscovian Stage of the Carboniferous System (Lucas 2018, p. 3; Nelson and Lucas 2021).

4 The mode of preservation was investigated by longitudinal, precision serial grinding to the level of the exposed seed. Through this process, eight ground surfaces, approximately 100 μm apart, were produced. Each surface was studied microscopically. The remaining in situ seed is curated in the Palaeobotanical Collection at Cape Breton University, accession number 78-402.

5 The detached structure is differentially eroded such that it consists partly of a coalified three-ribbed oval-shaped compression devoid of a seed, and partly of an intact, white cupulate seed embedded in situ in the compression. No foliar remains are associated with the fossil.

6 Each of the ground surfaces consisted of only sub-micron calcitic crystallites without cellular structures, and neither the micropyle nor the chalaza is identifiable. A micron-thick coalified layer surrounding the calcified ovule probably represents the partially preserved integument (CL in Fig. 1).

7 Most likely during preservation, initial coalification was followed by biological degradation of the original seed tissue, leaving a cavity. This cavity was subsequently infilled with a calcitic solution that on crystallization contracted to form wrinkled folds on the ovule. The process probably modified the overall dimensions of the seed; however, the undeformed coalified structure resisted the compaction pressure of the sedimentary load. Considering Schopf’s (1975, table I, 3.) classification, the seed is described as a calcitic cast in a cupulate compression, reflecting a dual preservation process during fossilization.

8 Determining the taxonomic affinity of the seed is difficult, particularly because it is not permineralized; hence, its anatomy cannot be established (Seyfullah et al. 2010). Adding to the problem is the absence of foliar remains. Such small seeds are regarded as lyginopterid (or hydrasperman; Meade et al. 2020) forms that had foliage of the sphenopterid–mariopterid type. Macroscopically and in terms of size, the present seed resembles the small seeds described by Van Amerom (1975, p. 25–29, fig. 9, drawings 1, 2, 5, 7); similar forms were described by Gensel and Skog (1977) from the Lower Mississippian, and by Gillespie et al. (1981) from the Late Devonian. Based on the author’s long experience of plant-fossil collecting from the Sydney Coalfield, the paucity of cupulate seeds is probably due to sample bias resulting from their relatively small sizes.