The area under a curve y=f(x), bounded by the X-axis and two vertical lines x=a and x=b, is found by slicing the region into a very large number of extremely thin vertical strips, each of width dx and height y=f(x), and adding up their areas. This sum, in the limit as the strips become infinitesimally thin, is exactly the definite integral ∫abf(x)dx -- so the same tool built in the previous chapter to evaluate integrals now directly measures plane area.
If the curve lies on or above the X-axis throughout [a,b], the integral itself gives the area. But an area can never be negative, so if the curve dips below the X-axis, the raw integral comes out negative and its absolute value must be taken to get the true area. If the curve changes sign somewhere inside the interval -- crossing the X-axis at an interior point -- the interval has to be split at that crossing point, the area of each piece found and made positive on its own, and the two positive areas added; a single integral taken across the whole interval would let the positive and negative parts cancel and understate the actual area.
The same idea works the other way round: when a curve is more naturally written as x=g(y), the area between the curve, the Y-axis, and two horizontal lines y=c and y=d is found by summing horizontal strips instead, giving ∫cdxdy. Deciding whether to integrate with respect to x or y is simply a matter of which form keeps the calculation simplest for the particular curve in the problem. This idea also underlies standard results such as the area of a circle or an ellipse, both of which can be derived by integrating one quadrant and using the curve's symmetry, and the area of a sector cut off from a circle by a straight line.