I think I get it now, maybe.
Since theta = pi/3, you check multiples of that in the equation, not multiples of 2pi like I thought.
So I plug in npi/3 for n = 1, 2, 3, 4, 5, ... and find all the values of n that make the equation 0 in the interval specified. Then I have my roots as 5pi/3, 7pi/3...