Plant and Soil 137: 157-160, 1991. © 1991 Kluwer Academic Publishers. Printed in the Netherlands. PLSO NL24 Are bacteria the third partner of the Azolla-Anabaena symbiosis? FRANCISCO CARRAPI~O Departamento de Biologia Vegetal, Faculdade de Ci6ncias de Lisboa, Campo Grande, Bloco C2, P-1700 Lisboa, Portugal Key words: Azolla, bacteria, megasporocarps, symbiosis Abstract The development of the prokaryotic colony in Azolla filiculoides indicates that Anabaena azollae is maintained through the life cycle of the fern and present in the leaves and megasporocarps. The same biological pattern is applied to the bacteria that are also present in these structures and seems to follow a development pattern identical to the cyanobacteria and probably can be considered the third partner of this symbiotic association. Introduction The presence of bacteria in the Azolla leaves were previously described by Grilli (1964) and later referred by several authors, with some re- serve, as an organism present in the leaf cavity of the fern (Forni and Grilli Caiola, 1986; Gates et al., 1980; Peters et al., 1978). In spite of this, organism having been identified by the first time by Bottomley in 1920 as Pseudomonas, the cor- rect identification was done by Wallace and Gates (1986) and included in the genus Arth- robacter. Based on this identification Petro and Gates (1987) suggest that these bacteria may be the third component of the symbiosis. Using different methods Carrapigo and Tavares (1987, 1989) have shown, by electron microscopy studies, that these bacteria were always present as a permanent member of the prokaryotic col- ony at different stages of leaf development of Azolla filiculoides and based on these data they have also suggested that these bacteria were the third partner of the symbiosis. At the same time Forni et al. (1987, 1989) have identified by bio- chemical tests several species of Arthrobacter present in different species of Azolla, suggesting the same hypothesis. The first references of the presence of bacteria in sporocarps was made by Peters and Calvert (1982) and by Tavares and Carrapigo (1988). In this work, further research was made on this symbiotic association, trying to integrate and develop the understanding of the relationship between the bacteria and the life cycle of the fern. Materials and methods The Azolla filiculoides plants were collected in Alcochete (70 km from Lisbon) and maintained in cultivation chambers. The megasporocarps were collected in April of 1987 and 1988 directly from the wild and prepared for further studies. Light and electron microscopy research were made on dorsal lobe leaves and in mega- sporocarps. For electron microscopy studies, leaves were submitted to a double step fixation process by 4% glutaraldehyde in 0.05 M cacody- late buffer (pH 7.2) and in 1% OsO 4 in 0.05 M cacodylate buffer (pH 7.4) with or without ruthenium red (Carrapi~o and Tavares, 1989). The leaves were dehydrated in acetone series and embedded in Epon-Araldite (Mollenhauer, 1964). The megasporocarps were fixed in a fixa- tive consisting of formaldehyde-glutaraldehyde (2%-2.5%) in 50 mM PIPES buffer (pH 7.4) for