26
$\begingroup$

I tried proving the formula presented here by integrating the circumferences of cross-sections of a right circular cone: $$\int_{0}^{h}2\pi sdt, \qquad\qquad s = \frac{r}{h}t$$ so $$\int_{0}^{h}2\pi \frac{r}{h}tdt.$$ Integrating it got me $\pi h r$, which can't be right because $h$ isn't the slant height. So adding up the areas of differential-width circular strips doesn't add up to the lateral surface area of a cone?

EDIT: I now realize that the integral works if I set the upper limit to the slant height - this works if I think of "unwrapping" the cone and forming a portion of a circle. The question still remains though: why won't the original integral work? Won't the value of the sum of the cylinders' areas reach the area of the cone as the number of partitions approaches infinity?

$\endgroup$
3

7 Answers 7

30
$\begingroup$

You seem to be ignoring the fact that s and r vary as the segment you consider varies. By using the same variable names it appears that you are confusing them to be constants...

Anyway, for a derivation, look at the following figure:

alt text

This is a cross-section of the cone.

The area of the strip of width $\displaystyle dh$ that corresponds to $\displaystyle h$ (from the apex) is $\displaystyle 2\pi r \frac{dh}{\cos x}$

Now $\displaystyle r = h \tan x$

Thus $\displaystyle dA = 2 \pi h \frac{\tan x}{\cos x} dh$

Thus the total area

$$= \int_{0}^{H} 2 \pi h \frac{\tan x}{\cos x} dh = \pi H^2 \frac{\tan x}{\cos x} = \pi (H \tan x) \left(\frac{H}{\cos x}\right) = \pi R S$$

Hope that helps.

$\endgroup$
2
  • $\begingroup$ Can you please explain how you find the area of the strip? I can't seem to understand. $\endgroup$
    – Stephan
    May 1, 2020 at 17:05
  • 1
    $\begingroup$ @Stephan: The assumption is that it is same as a rectangular strip of width $dh/\cos x$ and length equal to circumference $2 \pi r$ and so area is the product. $\endgroup$
    – Aryabhata
    May 7, 2020 at 23:07
7
$\begingroup$

The fundamental issue with using the integral stated in the original question is that is the wrong shape to use. It is using small cylinders to approximate what are in reality frustums (except the first one, which is a small cone). As the cylinders become smaller and smaller, as we approach the limit of the height of each becoming 0, those heights remain proportionally smaller than the slant height of the frustums. As those heights get smaller, the number of cylinders proportionally increases. When you sum it all up, the error between the approximated surface area to the actual surface area, remains unchanged.

The easiest way to see this is to approximate through numerical integration. Go ahead and start with one cylinder to approximate the cone. If you use the average radius for the cone, you will get a total surface area = $\pi h r$. Then split it into $2$ cylinders and add it up. You will get the same value for the total surface of the two cylinders. As you continue to break it into smaller and smaller cylinders, the total surface area remains constant. It remains the value of the integral stated above. In essence, this is similar to multiplying too numbers together, where one number of approaching 0 and the other is approaching infinity and the rates perfectly offset each other.

$\endgroup$
1
  • $\begingroup$ The "wrongness" of the shape may also be suggested by a one-dimensional example, where one attempts to approach the arc length of a curve $y=f(x)$ through summing a large number of horizontal segments. Even though more segments may be used to approximate the curve, the lengths of these will always sum to the same horizontal distance. $\endgroup$
    – hardmath
    Feb 20, 2016 at 18:17
5
$\begingroup$

Ok, so I've been thinking about this for a few days, and I asked the same question on physicsforums.com. And luckily, someone posted an answer that explained the problem. Here is the question that I posted: http://www.physicsforums.com/showthread.php?p=4031752. If you scroll all the way to the second to last post, you will see a person who says,

"The problem I think is that you wouldn't get the surface area if you used cylinders whose sides aren't parallel to the sides of the shape. For example, consider trying to "square" the perimeter of a circle. For illustration: http://qntm.org/trollpi

"The sides of the square are cut into many pieces, and then "steps" are created from it, but you can always combine the steps back into the side of the square. Consider the fourth picture in the link, look at the upper half of the circle. The whole upper side of the square is there, it is just in pieces. So you can jag them all you want, the perimeter is the same and doesn't start approximating a circle.

"In the cross section of the cone, considering it as a 2-d object, you can also try to calculate the perimeter of the slice. If you add up the sides of the rectangles, they will always add up to 2H, no matter how small you make them, however you can check using Pythagoras that it should be more. because its twice (for two sides of the slice) the square root of H^2+R^2"

That answer made a lot of sense to me, and I hope that it makes sense to everyone else here :)

$\endgroup$
3
$\begingroup$

I appreciate the thought process and did something similar as did by Mr G.P Burdell. To add to the confusion (I apologise!) let me put my point forward as this:

Consider a Cone with height H and radius R. Let dh and dr be the respective changes in H and R. Therefore
$$Volume = \int_0^H \pi{r^2} dh$$

Now “r” is a function of “h” so $$h = H- \frac{H}{R}r$$

$$dh = - \frac{H}{R}dr$$

Substituting the same in the above equation we get the integral as

$$\int_R^0 \pi{r^2} \frac{(-H)}{R}dr$$ $$Volume = \frac{\pi}{3}R^2H$$

If you do the same process for the surface area you will end up getting

$$Surface Area = \pi RH$$

Which is not true.

The Surface Area of a Cone is $$= \pi RS$$ where S is Slant Height of the Cone.

If we try to argue on this explanation of Surface Area then our explanation for Volume is in contradiction.

$\endgroup$
2
$\begingroup$

When ever you use calculus you need to start with the correct mathematical relationships. You further need to ensure that you integrate over the correct limits and over the correct variables.

You can see integration as the summation of a lot of intesimal small items to give the final answer.

All correctly setup integrals will result in the same answer.

Before one start or rush into solving a problem, think what would be the easiest way to address the problem.

I thought approaching it as a triangle would be quite easy to solve the problem. So this solution will use a triangle as the basic mathematical relationship.

We all know the area of a triangle is:

$$Area = \frac{1}{2}Base * Height$$

I will thus setup the problem as follows. (Solution 1)

Define a triangle with a height of S (defined as the length from point to side of base) and a base of an infinite small length of the base and call it dbase.

As we know that radians is defined in such a way that the circumference of a circle will yield exactly $$2.\pi$$ for a circle of unity (r=1). This relationship helps us to define dbase then as a plain ratio of an infinitely small angle multiplied by the radius. In this case the radius of the cone. Thus we can say:

$$dbase = R. d angle$$

The area of my triangle would then become:

$$darea = \frac{1}{2} S. R. dangle$$

We will integrate between the limits 0 and 2pi to include the total circumference. In degrees this would have been 0 to 360, however the relationship we use is defined for radians. (Note that I couldn't get the upper limit as $$2\pi$$ and thus integrate over half the circumference and multiply the result by 2. I am not familiar with the syntax. The answer however is the same. In some instances we use this as a trick to reduce the effort in calculating the result of integrals etc. )

$$Area = 2\int_0^ \pi \frac{1}{2} S. R. dangle$$

$$Area=2.(\frac{1}{2} S. R . \pi)-2.(\frac{1}{2} S. R . (0))$$

$$Area =\pi S R$$

And the result is exactly as the rest of the world believe the answer should be.

Important is to be always mathematically correct.

In answers above you made basic mathematical errors and end up with wrong answer.

If you follow maths you can't go wrong. If you went wrong you made a mistake, search your assumptions and relationships and try gain.

Try the volume yourself. I did this on my phone and it is somewhat difficult.

Solution 2 - integrate from 0 to H, where H is height of the cone and R the radius.

h =height, dheight is very small increase in height and r is the radius at h.

Angle is the angle between the vertical and side length S.

The mathematical relationship for the area is the circumference multiple by the small increase in the length of S. Note that the length is not dh but do because of the angle. Be very careful here.

$$darea = 2\pi . r. ds$$

To setup the integral correct we need to express r and ds in terms of h to ensure consistency. The relationship is as follows:

$$tan(angle) = r /h $$ $$r = h. tan(angle) $$

And

$$cos(angle) = dh/ds $$ $$ds = dh. Cos(angle) $$

darea becomes then

$$darea =2\pi.tan(angle). cos(angle) .h.dh$$

We just need to integrate this as:

$$Area = 2\pi. tan(angle). cos(angle) \int_0^ H h. dh$$

$$Area = 2\pi. tan(angle).cos(angle) (( \frac{1}{2} H^2)-(\frac{1}{2} (0)) $$

$$Area = \pi. tan(angle).cos(angle) H^2$$

We only need to get rid of the tan and cos and then we should have the same answer. From the above we can substitute back to get:

$$Area = \pi. S. R$$

It worked again, no magic.

Solution 3 - integrate from 0 to S. Same variables as in solution 2.

Area definition same. This time we need to express r in terms of s.

$$darea = 2\pi . r. ds$$

$$sin(angle) = r /s $$ $$r = s. sin(angle) $$

Replace r in darea, then integrate

$$Area = 2\pi. sin(angle) \int_0^ S s. ds$$

$$Area = 2\pi. sin(angle) (( \frac{1}{2} S^2)-(\frac{1}{2} (0)) $$

$$Area = \pi.sin(angle) S^2$$

Get rid of the sin term and can you believe, once again the same result as the rest of the world.

$$Area = \pi. S. R$$

I love maths.

$\endgroup$
1
$\begingroup$

See the discussion to a previous question here which might help

$\endgroup$
1
$\begingroup$

The diagram makes sense to me. The area around our slice clearly seems to me to be the circumference at that point multiplied by $\mathrm dS$ instead of $\mathrm dh$ as we might be tempted to do just because $\mathrm dh$ is easily defined in terms of $r$. Handily $\mathrm dS$ is just a constant times $\mathrm dr$: $\mathrm dS=\dfrac{\mathrm dr}{\cos(x)}$.

$\endgroup$
1
  • 1
    $\begingroup$ Welcome to MSE! It helps readability to format questions/answers using MathJax (see FAQ). Regards $\endgroup$
    – Amzoti
    Apr 30, 2013 at 17:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .