An established system for filling 50-gallon drums has a known standard deviation of 10 ounces. A new process was tested on 17 barrels and yielded a standard deviation of 8.59. Is there evidence to conclude that the new process has a lower variance than the old (let α=.05) ?