Okay, I can see how that could be now. The thing that helps me understand is slerping to 0.5 gives an even shorter angle, but 0.25 again gives the same angle as 0.75. So the point is travelling from (-1 0 0) -> (1 0 0), but around some other axis. I think the calculation I just did on getting the difference between q0 -> q1 shows that the axis is not the y axis like I thought it would be, but it must be correct because I'm getting the same answer as the slerp method.
Thanks for your help.