Azolla sporophytes with attached sporocarps from the early Paleocene (Danian) Salamanca Formation, Patagonia, Argentina

JUD NATHAN A.; DE BENEDETTI FACUNDO; HERMSEN ELIZABETH J.; GANDOLFO MARÍA A.

Athens, Ohio

Encuentro; 35th Midcontinent Paleobotanical Colloquium; 2018

Azolla Lam. are heterosporous floating ferns found in calm freshwater habitats worldwide. The genus comprises c. 5?7 extant species in two sections, Euazolla and Rhizosperma and approximately 70 fossil species. Most of the fossil species are based on dispersed megaspore apparatuses and microspore massulae; some of these are assigned to one of five extinct sections. Only a handful of fossil species are known from both vegetative and reproductive structures, but whole plant concepts are key to understanding patterns of morphological evolution and diversification of the genus. We describe a new Azolla from the early Paleocene (Danian) Salamanca Formation at the Palacio de los Loros locality in Chubut Province, Argentina. The fossils consist of leafy vegetative and fertile shoots, as well as vegetative shoots with attached roots that are organized in fascicles. The Salamanca fossils are most like the extant African species Azolla nilotica in section Rhizosperma because they share fascicled roots, leaves with a wide hyaline margin, sporocarps in homosporangiate or heterosporangiate groups of 3-4, the presence of a collar and indusial cap on the megaspore apparatus, and microspore massulae with hair-like glochidia. We are investigating the number and organization of floats under the indusial cap. Despite these similarities with A. nilotica, the Salamanca specimens are easily distinguished from that species and other Azolla by the sculpture of the excrescences on the megaspore complex and by the venation of the leaves. The phylogenetic relationships among extant Azolla and those extinct species known from both vegetative and reproductive structures is explored using a ?total-evidence? matrix of morphological and molecular characters analyzed using Maximum Parsimony and Bayesan Inference techniques.