Fuel economy is measured under controlled conditions in a laboratory using a standardized test procedure specified by federal law and implemented by the Environmental Protection Agency (EPA). Manufacturers test their own vehicles— usually preproduction prototypes —and report the results to the EPA. The EPA reviews the results and confirms about 10–15 percent of them through their own tests at the National Vehicles and Fuel Emissions Laboratory. Vehicle is driven on a roller dynamometer (the vehicle remains stationary while the driven wheels are allowed to roll) and a hose is connected to the tailpipe to collect engine exhaust. The carbon in the exhaust is measured to calculate the amount of fuel burned during the test. This is more accurate than using calculations based on distance traveled and the amount of fuel consumed. Five separate test schedules are driven to simulate the driving conditions of city, highway, and hot weather with air conditioning on, high-speed driving and acceleration, and cold weather environments. The latter three tests are new for 2008 model year vehicles and will likely result in a decrease in mpg versus the same model tested under the pre-2008 test procedure. Driving cycles: City — 23 stops with accelerations to 20–60 mph. Highway — approximately 13 minutes of mixed rural and interstate driving with speeds ranging from30 to 60 mph. High Speed — aggressive accelerations to speeds of up to 80 mph with six minutes of driving 60–80 mph. Air Conditioning — 10 minutes of accelerations to varying speeds up to 60 mph with the air conditioning on while the vehicle is in an environment of 95 degrees and approximately 40 percent relative humidity. Cold — represents urban driving in 20-degree weather. The vehicle is started with the engine cold and driven in simulated stop-and-go rush hour traffic.