3. A new pollution control device for automobiles is claimed not only to reduce carbon monoxide emission but also to increase a car's miles per gallon. To test the latter claim, the mileage per gallon of 10 randomly selected cars is measured first without and then with the device, yielding the following data:

Miles per gallon

Car Without Device With Device

1 28 32

2 20 21

3 44 36

4 7 10

5 20 20

6 11 14

7 11 9

8 34 32

9 9 14

10 48 46

Typically, the variance of the mileage per gallon of cars not fitted with the device is 175. Is the variance of the 10 cars in this study significantly less than this when they were driven with the device ( .05)? Is their variance significantly greater than 175 when they were driven without the device? ( .05)?