The seven special integrals listed above are each proved by a trigonometric substitution, chosen according to which quadratic surd or quadratic denominator is present (see the substitution table).
Formula 1: ∫x2+a2dx=a1tan−1ax+c. Put x=atanθ, so θ=tan−1ax and dx=asec2θdθ. Then x2+a2=a2sec2θ, so the integral becomes ∫a2sec2θasec2θdθ=a1∫dθ=a1θ+c=a1tan−1ax+c. (This can also be checked directly: differentiating a1tan−1ax gives a1⋅1+x2/a21⋅a1=x2+a21, confirming the formula without any substitution at all.)
Formulas 2 & 3.∫x2−a2dx: factor x2−a2=(x−a)(x+a) and split into partial fractions, x2−a21=2a1(x−a1−x+a1); integrating each piece gives 2a1logx+ax−a+c. The companion result ∫a2−x2dx=2a1loga−xa+x+c follows by the identical partial-fraction split with the sign of x2 reversed.
Formula 4: ∫a2−x2dx=sin−1ax+c. Put x=asinθ, so θ=sin−1ax, dx=acosθdθ, and a2−x2=acosθ. The integral collapses to ∫dθ=θ+c=sin−1ax+c.
Formula 5: ∫x2−a2dx=logx+x2−a2+c. Put x=asecθ, dx=asecθtanθdθ, and x2−a2=atanθ; the integral reduces to ∫secθdθ=log∣secθ+tanθ∣+c1. Converting back to x (using secθ=x/a, tanθ=x2−a2/a) and absorbing −loga into the constant gives the stated formula. Formula 6 (∫x2+a2dx=log∣x+x2+a2∣+c) is obtained the same way using x=atanθ.
Formula 7: ∫xx2−a2dx=a1sec−1ax+c. Put x=asecθ as before; xx2−a2=asecθ⋅atanθ, and dx=asecθtanθdθ, so the secθtanθ factors cancel, leaving a1∫dθ=a1θ+c=a1sec−1ax+c.
These formulae are most powerful once combined with a substitution that first turns a more complicated integrand into one of these seven shapes:
∫3sin4x−4sin2x+1sin2xdx. Put sin2x=t, so 2sinxcosxdx=sin2xdx=dt: integral becomes ∫3t2−4t+1dt. Completing the square (see 3.2.4) gives 31∫(t−32)2−(31)2dt, a Formula-2 shape, and the final answer works out to 21log3sin2x−13sin2x−3+c.
∫e−x−exex/2dx. Simplify the integrand to 1−e2xex (multiplying inside the root by e2x/e2x and pulling ex out front); putting ex=t (so exdx=dt) gives a Formula-4 shape, ∫1−t2dt=sin−1(ex)+c. …