Numerical Methods: Euler to Runge–Kutta · Challenge problem

An adaptive RK4 code models the local error of a step of size $h$ as $Ch^5$. A trial step with $h = 0.2$ has estimated local error $3.2 \times 10^{-5}$. The code accepts a step only if its local error is at most $10^{-5} \cdot h$ (error per unit step at most $10^{-5}$). Using the model, what is the largest step size it can accept? Express your answer as a decimal.