Roark's table has "Formulas for curved beams loaded normal to plane of curvature". The table shows values of for BETA =1.3. But BETA = (E/G) x (I/K) = 2.5 x (I/K). Or in other words K = 1.92(I). How in the world, we have sections, whose K value almost 2 times the value of I? I believe, the booK is not wrong. Where am I making the mistake?

Please help