A luminescence based marker system was developed for detection of genetically modified Pseudomonas fluorescens and E. coli. During batch growth in liquid culture, luminescence measured by luminometry was directly proportional to biomass concentration and enabled detection of 104 - 106 cells ml-1 of P. fluorescens and 101 cells ml-1 of E. coli, in actively growing cultures. Following inoculation into soil, detection levels were reduced ten-fold. After the subsequent utilisation of available nutrients, activity and luminescence decreased and luminometry then provided a direct, non-extractive means of measuring population activity of lux-marked inocula. Potential luminescence, measured as luminescence following amendment with nutrients, enabled assessment of the rate of reactivation of the lux-marked inocula and quantification of the size of the activatable population. Both these techniques, and traditional techniques, were used to investigate the survival of P. fluorescens and E. coli in soil microcosms. The effect of matric potential and indigenous organisms on luminescence and on survival of P. fluorescens was assessed. Matric potential significantly decreased the activity of both introduced and indigenous populations, but the indigenous population also significantly decreased the activity and biomass concentration of the introduced P. fluorescens population. Use of luminometry as a non-extractive measure of biomass concentration provided qualitative correlation with viable cell concentration, suggesting its potential for rapid enumeration of marked inocula. Reactivation of cells at increased matric stress was decreased, but use of high substrate:cell ratios at -30 kPa produced higher levels of luminescence and may, therefore, improve the use of luminometry as an estimate of biomass.