For some potentially useful and emerging in situ bioremediation technologies it is important to control bacterial attachment to subsurface materials during the injection of microbial cell suspensions. In this study the attachment patterns of Methylosinus trichosporium OB3b were measured after horizontal injections into a two-dimensional miniature aquifier simulator containing a wet homogeneous sand. In preliminary sand column assays, bacterial attachment to the sand was increased nearly 2 orders of magnitude compared to attachment in the presence of distilled water by raising the concentration of a pH 7.0 sodium-potassium phosphate buffer to 10 mM. The maximal concentration of attached cells was ~4¿108 cells/g dry sand with both sand minicolumns and the wet sand aquifier simulator. For the latter this occurred on streamlines directly between the horizontal injector and withdrawal ports, where injection-withdrawal velocities were the highest. The effects of a simulated groundwater cross flow during suspension injection on bacterial attachment to the aquifier simulator sands were also studied, and a peristaltic pumping method to counteract these groundwater flow effects resulted in a more localized pattern, i.e., without extensive downstream skewing of the bacterial attachment zone. Phenol red was utilized as a nonbinding, red-colored tracer compound. It proved to be very convenient for quantitatively measuring the earlier breakthroughs of cells versus an inert tracer during the aquifier simulator and subsequent capillary tube cell injection experiments and also for visualizing the anticipated boundaries of cell attachment in the aquifier simulator. The effect of injection velocity on the observed bacterial attachment patterns in these experiments appears to be accounted for by colloid filtration theory. ¿ American Geophysical Union 1993