Premium
Distinguishing the Brazilian mangrove species Avicennia germinans and A. schaueriana (Acanthaceae) by elliptic Fourier analysis of leaf shape
Author(s) -
Nascimento Maria Gracelia Paiva,
Mayo Simon Joseph,
Andrade Ivanilza Moreira
Publication year - 2021
Publication title -
feddes repertorium
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.24
H-Index - 25
eISSN - 1522-239X
pISSN - 0014-8962
DOI - 10.1002/fedr.202000025
Subject(s) - acanthaceae , biology , botany , taxon , mangrove , allometry , principal component analysis , avicennia , ecology , mathematics , statistics
The diagnostic value of leaf shape for mangrove species Avicennia germinans and A. schaueriana was examined in four areas within the Rio Parnaíba delta, Maranhão, Brazil. Elliptic Fourier analysis of 627 leaves, followed by principal component analysis, provided shape and size data; the first seven principal component shape variables expressed 99.2% total variation. Discrimination between twelve combinations of species, location and disturbance levels was studied using k‐nearest‐neighbour classification. The species differed significantly in leaf shape but determination was reliable only in the tallest undisturbed forest ( germinans narrowly elliptic or ovate, base acute, apex acute to acuminate; schaueriana broadly obovate, base attenuate, apex rounded). Symmetric variables (PC1 and PC3, 84.5% total variation) showed strongest differences. Predominant variation between species was in relative width and base shape (PC1), and within species in ovateness (PC3) between conserved and regenerating areas. Leaves were significantly smaller in A. schaueriana . Vernacular taxa siriba branca and siriba preta were inconsistent with taxonomic species but often significantly different in leaf shape. Avicennia germinans varied more under different geographical and ecological conditions. Shape‐size correlation needs further allometric study. Species delimitation needs morphometric study of a broader character range. Vernacular taxa recognition should be studied with ethnobiological techniques.