Finding the Power Series Representation for \(\tan^{-1} {\bf x}\)

Show that the power series representation for \(\tan ^{-1}x\) is \[\bbox[5px, border:1px solid black, #F9F7ED]{\bbox[#FAF8ED,5pt]{ \tan ^{-1}x=x-\dfrac{x^{3}}{3}+\dfrac{x^{5}}{5}-\dfrac{ x^{7}}{7}+\cdots +(-1)^{n}\dfrac{x^{2n+1}}{2n+1}+\cdots =\sum\limits_{k\,=\,0}^{\infty }\dfrac{(-1) ^{k}x^{2k+1}}{2k+1} }} \]

The power series representation for \(\tan^{-1}x\) is called Gregory’s series (or Gregory–Leibniz series or Madhava–Gregory series). James Gregory (1638–1675) was a Scottish mathematician. His mother, Janet Anderson, was his teacher and taught him geometry. Gregory, like many other mathematicians of his time, was searching for a good way to approximate \(\pi\), a result of which was Gregory’s series. Gregory was also a major contributor to the theory of optics, and he is credited with inventing the reflective telescope.

Find the radius of convergence and the interval of convergence.

Solution If \(y=\tan ^{-1}x\), then \(y^\prime =\dfrac{1}{1+x^{2}}\), which is the sum of the geometric series \(\sum\limits_{k\,=\,0}^{\infty}(-1) ^{k}x^{2k}\). That is, \[ \dfrac{1}{1+x^{2}}=\sum\limits_{k\,=\,0}^{\infty }(-1) ^{k}x^{2k}=1-x^{2}+x^{4}-x^{6}+\cdots\, \]

This series converges when \(\vert x^{2}\vert < 1\), or equivalently for \(-1 < x < 1\). We use the integration property of power series to integrate \( y^\prime =\dfrac{1}{1+x^{2}}\) and obtain a series for \(y=\tan ^{-1}x\). \[ \begin{eqnarray*} \int_{0}^{x}\frac{dt}{1+t^{2}} &=&\int_{0}^{x}(1-t^{2}+t^{4}-\cdots)\, dt \nonumber \\ \tan ^{-1}x &=& x-\frac{x^{3}}{3}+\frac{x^{5}}{5}-\frac{x^{7}}{7}+\cdots +(-1)^{n}\frac{x^{2n+1}}{2n+1}+\cdots = \sum\limits_{k\,=\,0}^{\infty }(-1) ^{k}\dfrac{x^{2k+1}}{2k+1} \end{eqnarray*} \]

The radius of convergence is \(1\). To find the interval of convergence, we check \(x=-1\) and \(x=1\). For \(x=-1\), \[ -1+\frac{1}{3}-\frac{1}{5}+\frac{1}{7}-\cdots \]

For \(x=1\), we get \[ 1-\frac{1}{3}+\frac{1}{5}-\frac{1}{7}+\cdots \]

Both of these series satisfy the two conditions of the Alternating Series Test, and so each one converges. The interval of convergence is the closed interval \([-1,\,1] \).