The genus Hohenbergia has 54 recognized species (Luther, 2002; Leme, 2003) divided into two subgenera. Subgenus Hohenbergia includes 33 species and four varieties and occurs exclusively in Brazil (except for H. stellata Schultes f.), with a major center of diversity in northeastern Brazil, mainly in the state of Bahia. Subgenus Wittmackiopsis has the remaining taxa, occurring in the Greater Antilles.
Despite the great diversity of Hohenbergia species concentrated in northeastern Brazil and the uncommon potential for studies in the area, few contributions to a better understanding of the genus have been conducted. However, the low level of publication comparatively to the richness of Brazilian biodiversity concerns the problematic taxonomical status of some key species [e.g., Hohenbergia catingae Ule and its varieties, H. eriantha (Brong. ex Baker) Mez, H. vestita L. B. Sm.] to which most of the newly investigated taxa are apparently related following the identification key provided by Smith & Downs (1979). Some of the important species are still imperfectly known and poorly represented in
herbaria, with much lacking data (e.g., leaf shape, floral morphology, besides the variation in field populations), which make species delimitation a very complex task.
Although, H. flava belongs to the morphologically problematic complex headed by H. catingae, it can be recognized here as a new species closely related to the typical variety (i.e., H. catingae var. catingae), differing from it by the leaf blades broader at base (5-6.5 cm vs. 3-4 cm), margins bearing uncinate-retrorse spines (vs. spines prevailing antrorse, not distinctly uncinate), floral bracts ecarinate (vs. carinate) and densely whitelanate outside (vs. sparsely flocculose), sepals with the large lateral wing exceeding the midnerve (vs. sepal-wing equaling the midnerve), and by the yellow petals (vs. purple). It is important to stress here that Smith & Downs (1979) also reported yellow petals for H. catingae, but the indicated feature is not in accordance to the protologue of H. catingae and possibly represents the misidentification of some herbarium specimen related to the new species described here.
When collected, the specimen Leme 3720 et al. of H. flava was growing along a farm fence in the domain of the Atlantic Forest. However, the general characteristics of the species, especially the succulence of the leaves and the clear adaptation to drier and
exposed habitats suggest that the original source of collection was very probably obtained from any inland dryer area dominated by caatinga vegetation. This is.was not an isolated fact, because we also observed the same species in a public garden at the city of Vitoria da Conquista, Bahia, growing side by side with local native bromeliad species (e.g., Aechmea leucolepis L. B. Sm., Hohenbergia rosea L. B. Sm. & R. W. Read). There are also other sources of H. flava in cultivation, like the specimens introduced many years ago by the late Roberto Burle Marx, but not used in this study. The most informative collection came from Luiz Felipe Nevares de Carvalho. According to Carvalho, his specimen (i.e, the paratype) was found at the county of Santa Terezinha, where a transitional vegetation of deciduous forest to caatinga dominates the scenery. The paratype, was growing epiphytically side by side with Aechmea aff. lingulatoides Leme & H. Luther. —SeeVidalia, Revista da
