Ronald van Luijk

Part (2) of the main theorem (Thm. 1.1) takes two rational numbers

sigma_0, sigma_1 > 1.

Unfortunately, the proof uses that if these are different, then the ratios

tau_i = (sigma_i-1) / ( sigma_i*(sigma_i+1) )

are different. This is clearly a false statement, as the counterexample

sigma_0 = 2 and sigma_1 = 3

shows. Indeed, for n=1, the triangles found in Remark 1.2 for these

sigma_i are similar.

One way to fix the theorem is to take sigma_i > 1 + sqrt(2).

I thank Martin Bråtelund for noting this and suggesting the fix.