24
boost speeds are nonrelativistic, there is essentially no rotation of the frame in following the
two geodesics.
Next consider a reference frame in real space undergoing circular motion in the counter-
clockwise (positive) direction with a constant, nonrelativistic speed. Classically, this situa-
tion is treated by representing it as the limiting case of a set of small, non-colinear boosts.
That is, circular motion is approximated as motion along a polygon with an ever increasing
number of sides. Because the boosts involved all have the same nonrelativistic speed, all
the circular arcs representing them in rapidity space are essentially indistinguishable from
straight lines. Hence, as long as the speed of the frame in circular motion is nonrelativistic,
it undergoes essentially no change in orientation upon its return to the origin (as we would
expect).
Now suppose the reference frame undergoing circular motion has a constant speed that is
relativistic. The geodesics in rapidity space that form the polygon are now small arcs that
lie on circles whose intersection with the edge of the disk is orthogonal, as shown in Fig..
Consequently, the frame being boosted undergoes a definite change in orientation with each
boost. Thus, when the reference frame returns to its starting point, it will have undergone
a clockwise (negative) rotation with respect to its initial orientation. This sum of all the
rotations experienced along the way is the Thomas precession. Furthermore, we see from
the geometry of rapidity space that the amount of rotation will be a function of the speed of
the circular motion (that is, the rapidity space distance from the rapidity space origin), and
will increase without bound as this speed approaches the speed of light, as shown in Fig..
We also can use the geometry of rapidity space to show that non-colinear, relativistic
boosts are also (in general) noncommutative. Suppose we first boost a reference frame from
speed zero to a speed close to the speed of light, and then boost the frame through a second
non-colinear velocity. From the rapidity space diagram in Fig., it is clear that we will end up
at a completely different point if we do the same boosts in reverse order. However, if we look
closely at the rapidity space diagram, we see that although the two resultant velocities are
represented by different points in rapidity space, they both are the same (rapidity space)
distance from the rapidity space origin. Consequently, they both have the same rapidity
(and hence speed), but not the same direction. The result that the final speed resulting
from two successive non-colinear boosts is independent of the order in which the boosts are
applied is usually proved by a somewhat long algebraic calculation.