Road transport is a major source of both greenhouse gas emissions and local air pollution. Large vehicles cause disproportionate damages. Research and policy prioritize improving vehicle quality rather than changing driver behavior, even though driving techniques substantially affect fuel consumption and emissions. I conducted a field experiment in Karnataka, India, randomly assigning public sector bus drivers to two interventions: a training program on safe and fuel efficient driving, and a financial incentives scheme for achieving fuel efficiency targets. The training program increased fuel efficiency in the short term for four months and had no effect thereafter. The incentives scheme increased fuel efficiency for a twelve month period. I find no evidence of any complementarities between training and incentives. Training increased fuel efficiency by a marginally significant 0.0186 kilometers per liter for four months, which saved 0.19\% of baseline fuel consumption over twelve months, and had a return on investment of Rs. 3.12 per rupee spent. Incentives increased fuel efficiency by a statistically significant 0.0168 kilometers per liter for twelve months, which saved 0.35\% of baseline fuel consumption, and had a return on investment of Rs. 4.22 per rupee spent. Along with the high return on investment from fuel savings, the interventions generated positive externalities from reduced vehicle emissions.