Answer:
Graph A
Explanation:
The more sample you have under the curve, the more accurate it is.
Graph B is over estimating. You can literally see the corner of the rectangle overshoot the line of the graph.
Graph C is alright, but doesn't have as many rectangles to estimate the curve as graph A.
Graph D is overestimating like graph B. It has few rectangle is and overshooting above the line of the graph.