Uptake rates of several combined N sources, NZ fixation, intracellular glutamate (glu) and glutamine (gln) pools, and glutamine synthetase (GS) activity were measured in natural populations and a culture of Trichodesmium IMSlOl grown on seawater medium without added N. In cultured populations, the ratio of GS transferase/biosynthetic activity (an index of the proportion of the GS pool that is active) was lower, and intracellular pools of glu and gln and the ratios of gldglu and glnlaketoglutarate (glnlakg) ratios were higher when NZ fixation was highest (mid-day). There was an excess capacity for NH,' assimilation via GS, indicating that this was not the rate-limiting step in N utilization. In natural populations of Trichodesmium spp., the gldglu ratio closely approximated the glnlakg ratio over the die1 cycle. High gln/glu and gln/akg ratios were noted in near-surface populat i o n~. These ratios decreased in samples collected from greater depths. Natural populations of Tnchodesmium spp. showed a high capacity for the uptake of NH4+, glu, and mixed amino acids (AA).Rates of NO3-and urea uptake were low. NH,+ accumulated in the culture medium during growth and rates of NH4+ uptake showed a positive relationship with the NH4+ concentration in the medium. Although rates of NZ fixation were highest and accounted for the majority of the total measured N utilization during mid-day, rates of NH4+ uptake exceeded rates of NZ fixation throughout much of the die1 cycle. In exponentially growing cultures, only 23% of the total daily N utilization was due to N, fixation while NH,+ uptake accounted for more than 70%. Based on N2 fixation alone, the N turnover time for this culture during exponential growth was on the order of 9 d. This is consistent with the observed chlorophyll-based growth rates for these cultures suggesting that N, fixation was responsible for net growth. Our results contrast with the view that natural populations of Trichodesmium spp. acquire their cell N exclusively through N, fixation. C productivity may overestimate N demand for net production if regenerated production is significant in these populations.