A method for assaying N2-ase activity of algal crusts in the field was used to evaluate the effect of inoculation with liquid or dried sand cultures of three species of algae on field plots sown to winter wheat, with and without added N and irrigation. Fixation was increased by all inoculants but the most effective was a liquid application of Nostoc ellipsosporum. Anabaena cylindrica and Nostoc punctiforme were most effective when applied as dry sand cultures. Inoculation had no effect on crop yield. The spring application of 80kg N ha-1 as nitro-chalk reduced the N fixation on plots inoculated with Nostoc ellipsosporum early in the season, but stimulated fixation later when much of the applied N had been used by the crop and leached by rainfall. This increase was a result primarily of the denser crop canopy reducing soil surface desiccation. Air drying crusts of N. ellipsosporum grown on soil in the laboratory caused extensive lysis of the vegetative cells when crusts were re-wetted, followed after a few days, by a substantial increase in NH4+ concentration. All N2-ase activity was lost. Viable akinetes remained after desiccation but the soil surface was recolonised by Lyngbya which seemed to be favoured by the high soluble N concentration. Very similar effects were noted when crusts were cooled to -3°C. Cooling to 0°, 0.5 and -2.0°C induced some cellular leakage and a decrease in N2-ase activity but this recovered after 2-4 days to a rate exceeding that of the control. Water extractable N in unstressed algal crusts represented less than 2% of the total whereas 9-35% of the total N was lost from the cells in liquid culture. © 1979 Martinus Nijhoff, The Hague/Kluwer Academic Publishers.