Why Does Leg Selection in Trig Substitution Affect the Integral Result?

  • Context: Graduate 
  • Thread starter Thread starter FatalFlare
  • Start date Start date
  • Tags Tags
    Integration
Click For Summary
SUMMARY

The discussion centers on the impact of leg selection in trigonometric substitution on the results of integrals. The user initially derived the expression sqrt(x^2-9)+3*arcsin(3/x), which was incorrect, while the correct expression is sqrt(x^2-9) - 3*arctan(sqrt(x^2-9)/3). Both expressions yield the same derivative, indicating they differ by a constant, yet produce different numerical results for specific values of x. The conversation highlights the importance of understanding the relationships between inverse trigonometric functions and the implications of choosing different substitutions in integration.

PREREQUISITES
  • Understanding of trigonometric identities and functions
  • Familiarity with integration techniques, specifically trigonometric substitution
  • Knowledge of derivatives and their properties
  • Basic algebraic manipulation skills
NEXT STEPS
  • Study the properties of inverse trigonometric functions and their derivatives
  • Learn about trigonometric substitution techniques in calculus
  • Explore the relationship between different forms of antiderivatives
  • Investigate the geometric interpretation of trigonometric functions using right triangles
USEFUL FOR

Students and educators in calculus, particularly those focusing on integration techniques, as well as mathematicians interested in the nuances of trigonometric identities and their applications in solving integrals.

FatalFlare
Messages
1
Reaction score
0
I used trig substitution and got sqrt(x^2-9)+3*arcsin(3/x) which seems to be incorrect when I check it in my calculator and the textbook. I made a right triangle where one of the legs was sqrt(x^2-9) and it so happens that if you switch the leg the answer becomes sqrt(x^2-9) - 3*arctan(sqrt(x^2-9)/3)) which is the correct answer. It shouldn't matter which leg I choose but it does why?
 
Physics news on Phys.org
I moved this thread, which was originally posted in the HW section. It is not so much a homework problem as a question about why an answer can appear in two different forms. (Also, it might have drawn a warning in the HW section, as it was not posted using the homework template.)
 
FatalFlare said:
I used trig substitution and got sqrt(x^2-9)+3*arcsin(3/x) which seems to be incorrect when I check it in my calculator and the textbook. I made a right triangle where one of the legs was sqrt(x^2-9) and it so happens that if you switch the leg the answer becomes sqrt(x^2-9) - 3*arctan(sqrt(x^2-9)/3)) which is the correct answer. It shouldn't matter which leg I choose but it does why?

Both ##\sqrt{x^2-9}+3\arcsin(3/x) ## and ##\sqrt{x^2-9} - 3\arctan(\sqrt{x^2-9}/3))## have the same derivative, namely, your original integrand (assuming ##x>3##).
 
Ray Vickson said:
Both ##\sqrt{x^2-9}+3\arcsin(3/x) ## and ##\sqrt{x^2-9} - 3\arctan(\sqrt{x^2-9}/3))## have the same derivative, namely, your original integrand (assuming ##x>3##).
But when we substitute something for x, why do they not give the same result? And also they don't have the same derivative..!? Only the derivate of sqrt(x^2-9)-3arctan(sqrt(x^2-9)/3) results in the original integrand
 
ssc said:
But when we substitute something for x, why do they not give the same result? And also they don't have the same derivative..!? Only the derivate of sqrt(x^2-9)-3arctan(sqrt(x^2-9)/3) results in the original integrand
Please show your work!
 
PeroK said:
Please show your work!
Screenshot 2024-12-17 132218.png
Screenshot 2024-12-17 132349.png



If we allow the x to b an integer like 4 that is greater than 3, sqrt(x^2-9)-3arctan(sqrt(x^2-9)/3) results in 0.477 while sqrt(x^2-9)+3arcsin(3/x) gives 5.18

1734461145479.png
 

Attachments

  • 1734459970001.gif
    1734459970001.gif
    43 bytes · Views: 51
  • 1734460198453.gif
    1734460198453.gif
    43 bytes · Views: 66
  • Like
Likes   Reactions: PeroK
For ##x > 3##, the functions differ by a constant, hence have the same derivative. Note that the first term in both functions is the same, so you can focus on the ##\arcsin## and ##\arctan## terms.

Note that for the derivative of the ##\arctan##, in your final expression, ##\frac{x^2 - 9}{9} + 1## simpifies to ##\frac{x^2}{9}##. And, you can see that the two derivatives are equal.
 
PS the derivative of the ##\arcsin## function simplifies as well:
$$\frac{x}{\sqrt{x^2 - 9}} - \frac{9}{x^2\sqrt{1-\frac 9{x^2}}} = \frac{x}{\sqrt{x^2 - 9}} - \frac{9}{x\sqrt{x^2-9}} = \frac{x^2 - 9}{x\sqrt{x^2 - 9}} = \frac{\sqrt{x^2 - 9}}{x}$$Which is the same as the ##\arctan## derivative.

Remember that an antiderivative has an arbitary constant of integration, ususally denoted by ##C##. In this case, the two functions differ by some constant (see your graph for ##x > 3##): hence, have the same derivative. That wasn't so easy to see without doing some algebra on the resulting derivatives.
 
@ssc Instead of using the triangle method, I would have tried the direct substitution:
$$x = 3\sec \theta = \frac 3 {\cos \theta}$$And used the standard identity:$$\sec^2\theta = \tan^2 \theta + 1$$Hence$$\sqrt{x^2 - 9} = \sqrt{9\sec^2\theta - 9} = \sqrt{9\tan^2\theta} = 3\tan \theta$$And $$dx = (\frac{3\sin\theta}{\cos^2 \theta})d\theta = (3\tan\theta \sec\theta)d\theta$$For the original integral, we have:
$$\int \frac{\sqrt{x^2 - 9}}{x} \ dx = \int \frac{3\tan \theta}{3\sec\theta}(3\tan\theta \sec\theta)d\theta$$$$ = 3\int \tan^2 \theta \ d\theta = 3\int (\sec^2 \theta - 1) \ d\theta$$$$= 3\tan\theta - 3\theta + C$$Note that the derivative of ##\tan \theta## is ##\sec^2 \theta##, which is easy to show, but a very useful thing to remember!

Then, I would use the standard identity to finish this off:
$$= \sqrt{9 \tan^2 \theta} - 3\theta +C = \sqrt{9 \sec^2 \theta - 9} - 3\theta + C$$$$= \sqrt{x^2 - 9} - 3arcsec\big({\frac x 3}\big ) + C$$And we have yet another function that must have the same derivative!

In summary:
$$\int \frac{\sqrt{x^2 - 9}}{x} \ dx = \sqrt{x^2-9} - 3\arctan(\frac{\sqrt{x^2-9}}3)) + C_1 $$$$= \sqrt{x^2-9}+3\arcsin(\frac 3 x) + C_2$$$$= \sqrt{x^2 - 9} - 3arcsec\big({\frac x 3}\big ) + C_3$$
PS You can check those by differentiating them all. Or, you could find a list of inverse trig identities. Here are some, for example:
https://brilliant.org/wiki/inverse-trigonometric-identities/

You can see from that list that a lot of inverse trig functions are related by ##\pm## signs and constants. Which function you get for an integral may depend on the specific technique you use.
 
  • #10
PPS Let ##x = \sec \theta## for ##x \ge 1##. So that ##\theta = arcsec(x)##. Then:
$$\arctan(\sqrt{x^2 - 1}) = \arctan(\tan \theta) = \theta$$So, we have another inverse trig identity:
$$arcsec(x) = \arctan(\sqrt{x^2 - 1}) \ \ (x \ge 1)$$And we can use that to see that the first and third solutions above are indeed identical.
 
Last edited:
  • #11
PPS I think I understand the point of the triangle now. In this case, we have a triangle with adjacent side ##3## and opposite side ##\sqrt{x^2 - 9}## and a hypoteneuse of ##x##.

We are free to pick any trig substitution:
$$\cos \theta = \frac 3 x, \ \sec \theta = \frac x 3, \ \sin \theta = \frac{\sqrt{x^2 - 9}}{x}, \tan \theta = \frac{\sqrt{x^2 - 9}}{3}$$They are all essentially the same in terms of the relationship between ##x## and ##\theta##. It's really just a question of which one looks simplest. I picked ##\sec \theta## without drawing the triangle, because that looked the natural substitution to me.

Whatever you do, you should end up with the integral of ##\tan^2 \theta## and an intermediate answer of ##3\tan \theta - 3\theta + C##.

At this point, you can then use any of the trig functions from the triangle to express ##\theta##. I used ##x = 3\sec \theta##, because that was the substitution I originally used. But, it's equally valid to note that we also know that ##\tan \theta = \frac{\sqrt{x^2 - 9}}{3}##, and use the inverse of that.

This also explains why all the answers are valid. We can express ##\theta## in several different ways in terms of ##x##. Just look at the triangle:
$$\theta = \arccos(\frac 3 x), \ \theta = arcsec(\frac x 3), \ \theta = \arcsin(\frac{\sqrt{x^2 - 9}}{x}), \ \theta = \arctan(\frac{\sqrt{x^2 - 9}}{3})$$This explains the point of drawing the triangle: you have all these different equations at your fingertips.
 
Last edited:
  • Like
Likes   Reactions: SammyS
  • #12
PeroK said:
PPS I think I understand the point of the triangle now. In this case, we have a triangle with adjacent side ##3## and opposite side ##\sqrt{x^2 - 9}## and a hypoteneuse of ##x##.

We are free to pick any trig substitution:
$$\cos \theta = \frac 3 x, \ \sec \theta = \frac x 3, \ \sin \theta = \frac{\sqrt{x^2 - 9}}{x}, \tan \theta = \frac{\sqrt{x^2 - 9}}{3}$$They are all essentially the same in terms of the relationship between ##x## and ##\theta##. It's really just a question of which one looks simplest. I picked ##\sec \theta## without drawing the triangle, because that looked the natural substitution to me.

Whatever you do, you should end up with the integral of ##\tan^2 \theta## and an intermediate answer of ##3\tan \theta - 3\theta + C##.

At this point, you can then use any of the trig functions from the triangle to express ##\theta##. I used ##x = 3\sec \theta##, because that was the substitution I originally used. But, it's equally valid to note that we also know that ##\tan \theta = \frac{\sqrt{x^2 - 9}}{3}##, and use the inverse of that.

This also explains why all the answers are valid. We can express ##\theta## in several different ways in terms of ##x##. Just look at the triangle:
$$\cos \theta = \arccos(\frac 3 x), \ \theta = arcsec(\frac x 3), \ \theta = \arcsin(\frac{\sqrt{x^2 - 9}}{x}), \ \theta = \arctan(\frac{\sqrt{x^2 - 9}}{3})$$This explains the point of drawing the triangle: you have all these different equations at your fingertips.
Which I always found the triangle approach to be confusing to students. So i never taught it that way.

Instead I used the Unit Circle, and divided all terms by Sin^2 or Cos^2.

Down with the triangle!
 

Similar threads

  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 6 ·
Replies
6
Views
3K
  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 29 ·
Replies
29
Views
5K
  • · Replies 4 ·
Replies
4
Views
4K