Volumes of n-spheres


By James Foster :

Just wondering if anyone knew the general formula for the volume of an n-sphere (that is a sphere/hypersphere in n dimensions)

James


By James Foster :
Okay it's not perfect, but say the diagram below is that of a circle with radius R and at any point x it has a length perpendicular to the x-axis of 2z. Consider the following:


z =   ______
ÖR2 - x2
 

z = Rcos(t)

x = Rsin(t)

dx/dt = Rcos(t)

The area of the circle can be worked out by the integral of the lengths of the lines at position x with respect to x between -R and R.

This is

ò-RR 2z dz = ò-RR 2Rcos(t) dx

as x goes from -R to R, t goes from -p/2 to p/2 so

ò-RR 2Rcos(t) dx=òp/2-p/2 2Rcos(t)(dx/dt)dt

=òp/2-p/2 2(r2)cos2(t) dt

since cos2(t)=(cos(2t)+1)/2

area =òp/2-p/2 (R2)(cos(2t)+1) dt

since sin180=0 the cos(2t) cancels and leaves

area =òp/2-p/2[R2 t]=pR2

the notorious formula for a circle's area using the same notation as above. The volume of a sphere can be worked out by the integral of the area's of its slices at position x.

Therefore,

volume =ò-RR(p) z2 dx=òp/2-p/2(p)R2 cos2(t) (dx/dt) dt

=òp/2-pi 2(p)R3 cos3(t) dt

=òp/2-p/2[(p)R3(sin(t)-sin3(t)/3)]

=(p)(R3)2(1-1/3)

=(4/3)(p)(R3) the formula for the volume of sphere

If the theory is right then we can work out the hypervolume of a 4-ball by integrating the volume of the spherical slices.

So ò-RR (4/3)(p)z3 dx

=òp/2-p/2(4/3)(p)R3cos3(t)(dx/dt)dt

=òp/2-p/2(4/3)(p)R4cos4(t) dt

since cos4(t)=cos(4t)/4+cos(2t)/2+3/8

and as before sin180=0 and sin360=0 then

cos(4t) and cos(2t) integrate to 0

The hypervolume =òp/2-p/2(4/3)(p)R4(3/8)

=(1/3)(p2)(R4)

Let C(n)=òp/2-p/2cosn(t) dt

C(1)=2

C(2)=(1/2)p

C(3)=4/3

C(4)=(3/8)p

C(5)=16/15

C(6)=(5/16)p

in general Vn(R)=R(Vn-1(R))(C(n))

The two formulas I gave can be proved using mathematical induction to be the above.


[Ed: indecipherable image omitted here -- John if you'd like to upload a gif or jpg then please do! ]


By Wallace Home on Thursday, December 17, 1998 :
Try starting with 2p, 8p, 32p,...... integrate up to the required dimension and see what constants must be multiplied in. ignore constants of integration, i think (read hope). This might provide a better idea of how the expressions for hypervolumes of hyperspheres can be generated.

I must point out (no pun intended) that 2p is not a point: it has zero dimensions and is just a number. it doesn't do anything, it has no spatial place in relation to r or anything else. dimensionless numbers cannot be points. 2pPr is 1-dimensional expression, or point. (For comparison, imagine the question "How are you today?" Answering it with "42" or any other number is meaningless unless you give the number a dimension).

this whole thing is irritating me, actually, since everyone seems to think that "2000" is significant. OK, it's a good excuse to have a party, but I think people read too much into just an empty quantity. Is 2048 more significant than 2000? why? I'd like to start a discussion about the whole meaning or numbers thing, cos it's interesting how people perceive them.

Ed.