The dinitrogen (N 2) fixed by diazotrophs and released as dissolved nitrogen (DN) has been compared in batch cultures of four marine diazotrophic cyanobacteria: the colony forming Trichodesmium IMS101 and the unicellular strains Cyanothece ATCC51142, Crocosphaera watsonii WH8501 and WH0003. Two approaches were conducted for this purpose. The first approach consisted in the comparison of the total accumulation of fixed N 2 in the culture medium (both in the dissolved and particulate pools) with the net N 2 fixation rates (i.e., the fixed N 2 incorporated only in the particulate fraction after 15 N 2 incubation). The difference between the two measures accounted for the fixed N 2 released as DN. The second approach consisted in the direct measure of the 15 N-enrichment of ammonium (NH +) and dissolved organic N (DON) following 15 N 4 2 incubations. The fixed N 2 released as DN accounted for ∼0–20% and ∼1% of the fixed N 2 after 24 h in the first and second approach, respectively. We show that the recent methodological improvements in the net N 2 fixation determination applied in this study tend to reconcile the two approaches that formerly led to contrasted values. However, the large analytical uncertainties of the first approach limit its reliability. Thus, the direct determination of the 15 N-enrichment of the dissolved pool remains the best tool to assess the fixed N 2 released in the DN pool, in particular as it allows shorter incubation times. There were no clear patterns detected between the filamentous Trichodesmium and unicellular strains, neither in terms of the amount of fixed N 2 released as DN nor in terms of the proportion of NH + relative to DON. This suggests that the release of fixed N 4 2 is a process shared among the filamentous and free living diazotrophs.