Greeting fellow physicists, I have been having some questions from my AP Physics C: Mechanics students about the direction of friction in rolling. Here are a few situations. 1. For a wheel that is accelerating to the right while rolling smoothly, it must also be rotating clockwise. In order to prevent sliding between the bottom of the wheel and the surface, frictional force must also be pointed to the right. A typical example is a car accelerating from rest. The problem I run into here is since the friction is pointing forward, is causes a positive (ccw) torque, yet the wheel begins rolling in the cw direction?! Do only the magnitudes matter in Newton's second of rotation when considering rolling (net torque = rot. inertia * ang. acc)?! It seems so simple, but I've obviously forgotten something here. 2. For a wheel that is rolling smoothly at a constant velocity to the right (or the left), the frictional force on the bottom of the wheel must be zero (assuming no opposing torques). In this case, friction is not needed to keep the wheel rotating. That is taken care of by the wheel's rotational inertia. No problems here; its what happens when you are driving and hit a patch of ice. No (or very little) friction acts on the tires, but your tires keep spinning at the same rate and the car still moves forward at the same speed (ignoring air resistance). 3. For a wheel that is rolling smoothly to the right with a decreasing velocity, it must still be rotating clockwise. This is where my students loose it. They say since the frictional force must be pointed to prevent sliding (correctly), and the bottom of the wheel would slide left (I think the problem is here), then friction must still be pointed to the right. I use the situation where the car is rolling to the right (no longer on the ice) at a constant speed and a huge gust of wind blows from the front, causing the car to slow down. From observations, we see the CoM of the car slow down and the wheels angular speed also decrease. Without friction, the wheel's angular velocity would not change, even when the car comes to a stop (sure wish I had perfect ice to show this). Thus, friction must act to slow down the angular speed of the wheel, pointing backward to slow it down. This is counter-intuitive again, since the direction of the torque caused by the friction seems to be in the wrong direction. I appreciate you reading all this. Maybe it's just the end of the day, but I'd appreciate some thoughtful feedback on this. Thank you.