Range of R squared after combining two predictors

An OLS regression of $y$ on $X_1$ has $R^2=0.15$, and one of $y$ on $X_2$ has $R^2=0.2$. If $[\min,\max]$ is the tightest interval that must contain the $R^2$ of the regression of $y$ on both $X_1$ and $X_2$, what is $\max/\min$?

$5$. Adding a regressor can never lower $R^2$ (OLS may set its coefficient to zero), so the lower bound is $0.2$. The two predictors can together explain $y$ exactly (e.g. $y=X_1+X_2$ with suitable dependence), so the upper bound is $1$, giving $1/0.2=5$.