Questions Most vegetation descriptions tacitly assume that floristic composition and physiognomy are tightly linked. However, the two vegetation properties may not respond in a similar way to environmental and disturbance gradients, leading to uninformed management planning and difficulties when attempting to restore degraded ecosystems. In this context, we addressed two main questions: (1) how close are relations between floristic and physiognomic types as defined by numerical vegetation classification in mountain ecosystems; and (2) how are floristic and physiognomic types distributed along the elevation gradient? Location Central mountains of Argentina, between 500 and 1700 m a.s.l. Methods We selected 437 sites where we performed complete floristic and physiognomic relevés. We classified eight physiognomic and eight floristic types. We tested the relationship between the two classifications through a chi square analysis. We tested the association between elevation and each physiognomic and floristic type with random permutations. Results In general, floristic types were significantly and positively associated with more than one physiognomic type and vice versa. Physiognomic and floristic types responded differently to the elevation gradient. Floristic types were restricted to different sections of the gradient, although having large overlap among them. In contrast, seven out of the eight physiognomic types did not show elevation restriction, being distributed along the complete elevation gradient. The open low woodland with shrubs was the only restricted physiognomy, significantly absent from the upper part of the gradient. Conclusions We highlight the importance of considering the two vegetation properties independently when characterizing vegetation patterns in heterogeneous systems, since they show decoupled responses to environmental gradients. We note that the assumption of a direct link between floristic composition and physiognomy may induce bias into the understanding of vegetation patterns and processes. Hence, we encourage managers and restoration practitioners to consider the complete range of possible physiognomic types under each floristic type.

No abstract

We studied the genetic diversity and structure of the Tillandsia capillaris complex, a morphologically diverse group of highly specialized epiphytes, across the distribution range in arid mountain regions of central Peru, Chile, Bolivia and central Argentina. To elucidate the phylogenetic relationships in the complex and to explore the taxon boundaries among populations, we used three plastid markers (rpoB-trnC-petN, trnK-matK-trnK, ycf1, c. 8100 bp in total) and one single-copy nuclear gene (PHYC, c. 1200 bp) for 69 populations and 96 individuals of the T. capillaris complex plus 16 outgroup taxa. Bayesian inference of plastid DNA data indicates the existence of two evolutionary core lineages, which can be recognized as two distinct species: T. capillaris and T. virescens (as proposed previously on the basis of morphological characters). Tillandsia capillaris is a monophyletic and homogeneous group, widely distributed and less genetically variable, whereas T. virescens (including T. kuehhasii) is genetically more divergent with most of the forms growing at high elevation in arid areas, except for the small clade including T. virescens s.s. (= T. cordobensis), which grows in lower, more humid habitats. The nuclear analysis resulted in a polytomy with some individuals showing incongruent positions between plastid and nuclear topologies. The high haplotype diversity, consisting of 63 plastid DNA haplotypes in 69 populations, was resolved as two core lineages occurring from north to south, allowing us to establish a preliminary view of the genetic variation overlapping between the two taxa. The results suggest that the genetic differentiation into two main clades is consistent with morphological variation in this Andean complex.