In some arid regions, rehabilitation of whole system N-fixation may be strongly facilitated by the recovery of populations of the lichen genus Collema. Identification of the limits to recovery of Collema in apparently suitable habitat should inform selection of rehabilitation techniques. We simultaneously tested the relative importance of three hypothetical limits to Collema recovery: active erosion, resource limitation, and propagule scarcity. We found that in our experimental system, active erosion had no effect on short-term establishment of Collema, whereas propagule addition did enhance recovery and microhabitat (a resource availability gradient) also exerted a strong influence. It is possible that attempts to improve N cycling via re-establishment of Collema might be best served by developing economical means of simulating moister, cooler microhabitats, e.g., sloping soil or creating partial shade, which would favor the establishment of naturally dispersed propagules, rather than introducing propagules.