Define $f:\mathbb R \rightarrow \mathbb R$ by $f(x)=\frac{1}{1+3\cos^{2}x}$.

Since $f$ is continuous, $f$ has an antiderivative. That is, there exists some function $F:\mathbb R \rightarrow \mathbb R$ such that $F^{\prime}=f$.

Using the substitution $u=\tan x$ and doing some work, we find that a possible antiderivative of $f$ is the function $g$ defined by $g(x)=\frac{1}{2}\tan^{-1}\frac{\tan x}{2}$.

However, the problem with $g$ is that it's not defined at odd multiples of $\pi/2$, while the $F$ we're seeking is defined on $\mathbb R$.

So, what is $F$ and how do we find it?

(Or is there a mistake somewhere in my argument above?)

  • $\begingroup$ I won't mark this one as a duplicate since the question statements are rather different, but the issue is exactly the same as the one in this question, which has good* answers taking various approaches, involving a similar integral: math.stackexchange.com/questions/3275770/… (*One of the answers is mine, I make no claims about its goodness.) $\endgroup$ – Travis Willse Feb 23 at 4:21

For each odd integer $k$, define $f_k:((k-2)\pi/2,k\pi/2) \rightarrow \mathbb R$ by $f_k(x)=\frac{1}{1+3\cos^{2}x}$.

As you've found, an antiderivative of $f_k$ is $g_k:((k-2)\pi/2,k\pi/2) \rightarrow \mathbb R$ defined by $g_k(x)=\frac{1}{2}\tan^{-1}\frac{\tan x}{2}$.

And as you've also correctly stated, we know by the FTC that $f$ must have an antiderivative $F$.

So, we guess that such a function $F$ can be constructed by "moving up or down" the functions $g_k$ and "stitching" them together, and also "filling in the holes" at each odd multiple of $\pi/2$.

By examining the graphs of each $g_k$, we come up with this guess:

For each odd integer $k$, define $F$ by $$F(x)=\begin{cases} \left(k-1\right)\frac{\pi}{2}, & \text{ for }x=\left(k-2\right)\frac{\pi}{2},\\ \frac{1}{2}\tan^{-1}\frac{\tan x}{2}+k\frac{\pi}{4}, & \text{ for }x\in\left(\left(k-2\right)\frac{\pi}{2},k\frac{\pi}{2}\right)\\ k\frac{\pi}{2}, & \text{ for }x=k\frac{\pi}{2}. \end{cases},$$

To show that the above guess is correct, we can verify that $F^{\prime}=f$ (this will involve more work).

Similar questions: 1, 2, 3. See also Jeffrey and Rich (1994, "The Evaluation of Trigonometric Integrals Avoiding Spurious Discontinuities") and Jeffrey (1994, "The Importance of Being Continuous").

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.