This is known only from rocky riverbanks, generally on calcareous material, in disjunct localities of Tenn., Ky., Ind., ?Ohio, ?W.Va., Md. and Pa. (K; Weakley et al. 2023). Colls. need to be checked from Ohio (e.g. C. Weishaupt 16Sep1961 from Clermont Co. at OS) and W.Va. (e.g. E.S. Steele in 1898 from Preston Co. at NY), The species is locally abundant but has often been overlooked. Some colls. are difficult to distinguish from gigantea var. serotina, which often occurs in the same habitats, and intergradation has been suspected. In southeastern regions both taxa may be largely diploid (2n = 18), but polyploids also occur across their ranges (Cr, FNA 20). In rupestris, flowering heads tend to be smaller than in gigantea: involucres average ca. 2-3 mm high (versus 3-4 mm); ray florets number 2-7 (versus mostly 7-12). Flowers tend to open later in Ky. (usually late Aug-early Sep versus mid-late Aug), and flowering branches tend to be longer, the primary ones mostly 10-20 cm (versus 5-10 cm). Without flowers, rupestris can still usually be distinguished from gigantea by its leaves, which are narrower (mid-cauline ones mostly 1-1.5 cm wide versus 1.5-3 cm) and deeper green. Also, its stems are more clustered (with much shorter rhizomes), and they tend to lack the glaucous bloom or more matured purplish hue of typical gigantea; but they can become reddish with age.