Regression and correlation coefficient

Either. Given $R$, you can find $R^2$, and vice versa (given that you know whether the slope is positive or negative).