Direct association between wheat roots and an ammonia-excreting mutant of the cyanobacterium Anabaena variabilis, strain SA-1, was required for maximal enhancement of growth of wheat plants in nitrogen (N)-free, hydroponic medium. Over 85% of the cyanobacterial mutant SA-1 inoculated to the roots were adsorbed under non-saturating conditions. The adsorption process of SA-1 to wheat roots was biphasic : an initial rapid adsorption was followed by a slow phase with about 10% of the initial adsorption rate. The maximal adsorption rate of filaments observed was 1.6 mg dry wt. SA-1 adsorbed.plant(-1) h(-1). Bypassing CO2 fixation and sugar formation, the C-14 label from [C-14]sucrose was directly applied to leaf blades to study sugar translocation. The C-14 label from this treatment appeared in the wheat culture medium within an hour. Nitrate-grown plants excreted about 30% of the C-14 label into the medium, compared to only 10% excreted by wheat/Anabaena co-cultures. SA-1 assimilated 27% of all C-14 translocated from [U-C-14]sucrose applied to wheat leaves, and C-14 label from this treatment was recovered from strain SA-1 after 30 min. Roots and cyanobacteria accounted for 51% of all radioactive label recovered in the plants cocultured with SA-1 vs 20% for nitrate-grown plants. We studied the activity of beta-fructosidase (invertase) in wheat of variety Yecora rojo. Roots from nitrate-grown wheat plants produced high levels of invertase activity, which converted over 85% of 3 mM sucrose into glucose and fructose in 24 h. The rate of sucrose disappearance in the medium of co-cultures using A. variabilis SA-1, was 70% of that of nitrate-grown plants, but the levels of glucose and fructose in these cultures were always very low during sucrose conversion, suggesting hexose assimilation. To study the role of diffusible metabolites, a dialysis membrane was employed to separate the ammonia-excreting SA-1 from the wheat roots. Containing SA-1 in a dialysis bag away from direct root contact, severely limited leaf growth to less than one-third of the growth rate of nitrate control cultures. Ammonia produced by mutant SA-1 in dialysis bag cultures was excreted into the medium at 0.4 mM vs 1.2 mM in free-living cultures, but ammonia was not detectable in co-cultures with or without the dialysis bag containing the mutant. The nitrogenase activity derepressed in the mutant and responsible for ammonia excretion was always higher in the association co-cultures than in either free cells or in dialysis-bag cultures. The nitrogenase activity of strain SA-1 was highest (200 mu mol ethylene formed.mg(-1) Chl.h(-1)) when the cyanobacterium was associated with the root tips. Dialysis membrane separation of plant and cyanobacterium severely inhibited growth of wheat during a complete growth cycle of 2 months. Total biomass and grain yield were very similar for control cultures without inorganic N or SA-1, and for diffusion cultures containing SA-1, kept in a dialysis bag around the roots. Total biomass of the association coculture attained 75% of the biomass of the nitrate-grown control. It is proposed that wheat roots supplied fructose derived from sucrose for growth and nitrogen fixation of SA-1 in the light, and that ammonia excreted by SA-1 was utilized by the wheat plant for its own growth.