Discrimination of lichen genera and species using element concentrations

Abstract

Abstract:The importance of organic chemistry in the classification of lichens is well established, but inorganic chemistry has been largely overlooked. Six lichen species were studied over a period of 23 years that were growing in 11 protected areas of the northern Great Lakes ecoregion, which were not greatly influenced by anthropogenic particulates or gaseous air pollutants. The elemental data from these studies were aggregated in order to test the hypothesis that differences among species in tissue element concentrations were large enough to discriminate between taxa faithfully. Concentrations of 16 chemical elements that were found in tissue samples from Cladonia rangiferina, Evernia mesomorpha, Flavopunctelia flaventior, Hypogymnia physodes, Parmelia sulcata, and Punctelia rudecta were analyzed statistically using multivariate discriminant functions and CART analyses, as well as t-tests. Genera and species were clearly separated in element space, and elemental discriminant functions were able to classify 91–100 of the samples correctly into species. At the broadest level, a Zn concentration of 51 ppm in tissues of four of the lichen species effectively discriminated foliose from fruticose species. Similarly, a S concentration of 680 ppm discriminated C. rangiferina and E. mesomorpha, and a Ca concentration of 10 436 ppm discriminated H. physodes from P. sulcata. For the three parmelioid species, a Ca concentration >32 837 ppm discriminated Punctelia rudecta from the other two species, while a Zn concentration of 56 ppm discriminated Parmelia sulcata from F. flaventior. Foliose species also had higher concentrations than did fruticose species of all elements except Na. Elemental signatures for each of the six species were developed using standardized means. Twenty-four mechanisms explaining the differences among species are summarized. Finally, the relationships of four species based on element concentrations, using additive-trees clustering of a Euclidean-distance matrix, produced identical relationships as did analyses based on secondary product chemistry that used additive-trees clustering of a Jaccard similarity matrix. At least for these six species, element composition has taxonomic significance, and may be useful for discriminating other taxa.