Leprocaulon beechingii is described as new to science based on collections from exposed rock outcrops in the southern Appalachian Mountains in eastern North America. Taxonomic placement in Leprocaulon, and delimitation from other members of the genus with usnic acid, is supported by molecular phylogenetic analyses of ITS and mtSSU sequence data. The species is readily recognized by its occurrence on non-calcareous rocks, normandinoides-type placodioid thallus, and the production of usnic acid and zeorin.

Initial BLAST searches of the sequences of the new taxon against the NCBI nucleotide collection revealed a close match to existing reference sequence of Leprocaulon Nyl. (ITS, e value 0.0, % identity 89.17% to KC184096; mtSSU e value 0.0, % identity 96.36% to KJ66426) Taxon sampling for the present study was based on Lendemer and Hodkinson (2013) which utilized ITS and mtSSU sequence data. As such, data for these loci were assembled for the Leprocaulaceae based on the sampling of Lendemer and Hodkinson (2013), with representatives of Caliciaceae selected as the outgroup (Dirinaria applanata (Fée) D.D. Awasthi, Pyxine sorediata (Ach.) Mont., P. subcinerea Stirt.), and then expanded as follows: 1) sequences of Leprocaulon calcicola Earl.-Benn., Orange, Hitch & M. Powell were included based on Orange et al. (2017); 2) newly generated ITS and mtSSU sequences from one specimen of the potentially new species were included; 3) sequences of L. nicholsiae Lendemer & E. Tripp were included based on Tripp & Lendemer (in press.), 4) a newly generated sequences L. nicholsiae Lendemer & E. Tripp (11, ITS), L. knudsenii Lendemer & B.P. Hodkinson (2, ITS), and an apparently undescribed Leprocaulon species from Chile (3, ITS) were included.

One alignment was prepared for ITS and one for mtSSU, each was then subjected to auto-alignment using the MAFFT online interface and then manually adjusted with Mesquite 3.04 (Maddison & Maddison 2015). Ambiguously aligned regions were defined as an exclusion set in Mesquite. Each dataset was then prepared for maximum likelihood (ML) analyses by transforming uncertainties and polymorphisms to missing data, manually deleting the ambiguously aligned regions, and exporting the alignment as a PHYLIP file. Each resulting PHYLIP file was then input to ML analyses in RAxML 8.2.10 (Stamatakis 2006) with 500 bootstrap replicates, implementing the most complex nucleotide substitution model in that program (GTRGAMMA). The resulting topologies with ML-bootstrap values were visualized in FigTree (Rambaut 2017). As no conflict (defined following Mason-Gamer & Kellogg 1996) was detected between the two datasets, the alignments were concatenated and the combined alignment subjected to ML analyses as outlined above but with the two loci partitioned and the model GTRGAMMA applied to each partition. The results of the combined analyses were also visualized in FigTree.

This dataset includes the following:

1) The single locus DNA alignment for nrITS (NEXUS format)

2) The single locus DNA alignment for mtSSU (NEXUS format)

3) The concatened alignment for both loci (PHYLIP format) and the partitions file for this alignment.

4) A table crossreferencing voucher data, isolate numbers, and NCBI accession numbers.