Apart from the conventional, Mo-containing nitrogenase, the cyanobacterium Anabaena variabilis can express at least two alternative N2-fixing enzyme complexes. This cyanobacterium grows with V in a Mo-deficient medium. The nitrogenase then expressed reduces C2H2 partly beyond C2H4 to C2H6 and produces more H2 than the Mo-enzyme (Kentemich et al., 1988; Yakunin et al., 1991). Genes for this V-nitrogenase, vnfDGK (Thiel, 1993) and nifB (commonly used for both nitrogenases) (Lyons and Thiel, 1995) have subsequently been cloned, mapped and sequenced. It is generally agreed that A. variabilis contains a V-nitrogenase. In addition, there is physiological evidence that A. variabilis also contains an Fe-only nitrogenase (Kentemich et al., 1991). After an extensive state of N-deprivation where the cells totally bleached due to the utilization of phycobilins as N-reserve, A. variabilis restarted to grow slowly. The cultures reduced C2H2 partly to C2H6 and produced H2 with a high rate. Determination by atomic absorption spectrometry showed that the concentration of Mo in the medium was <10 nM and of V <20 nM and likely too low for the expression of a Mo- or a V-nitrogenase. The cells showed an outburst in C2H6- and H2 -formations 3 h after the addition of Mo to the medium which typically happens with the Fe-only hydrogenase from Azotobacter vinelandii (Pau et al., 1989). Hybridizations of the anfH and nifH probes with genomic DNA from A. variabilis gave at least two distinct bands (Kentemich et al., 1991). Supporting evidence for the existence of an Fe-only nitrogenase came from Ni et al., (1990). The gene set coding for this enzyme complex has, however, not yet been found which leaves, of course, doubts about the meanings of these physiological results.