The Mathematical Theory of RelativityEddington, Arthur Stanley, Sir
Science
The Mathematical Theory of Relativity
Eddington, Arthur Stanley, Sir
Relativity (Physics)
Let a ray travelling with velocity~$v$ traverse a distance~$R$ in a direction~$\theta$,
so that
\[
dt = R/v,\quad
dx = R \cos\theta,\quad
dy = R \sin \theta.
\]
\PageSep{20}
Let the relative velocity of~$S$ and~$S'$ be small so that $u^{2}/c^{2}$~is neglected. Then
by~\Eq{(5.3)}
\[
dt' = dt + u\, dx/c^{2},\quad
dx' = dx + u\, dt,\quad
dy' = dy.
\]
Writing $\delta R$, $\delta\theta$, $\delta v$ for the change in $R$,~$\theta$,~$v$ when a transformation is made
to $S'$'s~system, we obtain
\begin{align*}
\delta(R/v) &= dt' - dt = uR \cos \theta/c^{2}, \\
\delta(R \cos\theta) &= dx' - dx = uR/v, \\
\delta(R \sin\theta) &= dy' - dy = 0.
\intertext{Whence the values of $\delta R$, $\delta\theta$, $\delta (1/v)$ are found as follows:}
\delta R &= uR \cos \theta/v, \\
\delta\theta &= -u \sin \theta/v, \\
\delta\left(\frac{1}{v}\right) &= u \cos\theta \left(\frac{1}{c^{2}} - \frac{1}{v^{2}}\right).
\end{align*}
Here $\delta(1/v)$ refers to a comparison of velocities in the directions~$\theta$ in~$S$'s
system and $\theta'$~in $S'$'s~system. Writing $\Delta(1/v)$ for a comparison when the
direction is~$\theta$ in both systems
\begin{align*}
\Delta\left(\frac{1}{v}\right)
&= \delta\left(\frac{1}{v}\right) - \frac{\dd}{\dd\theta}\left(\frac{1}{v}\right) · \delta\theta\Add{,} \\
&= \frac{u}{c^{2}} \cos\theta - \frac{u}{v^{2}} \cos\theta + \frac{u\sin\theta}{v}\, \frac{\dd}{\dd\theta}\left(\frac{1}{v}\right) \\
&= \frac{u}{c^{2}} \cos\theta + \tfrac{1}{2} u \sin^{3}\theta\, \frac{\dd}{\dd\theta} \left(\frac{1}{v^{2} \sin^{2}\theta}\right).
\end{align*}
Hence
\[
\Delta\left(\frac{1}{v(\theta)} + \frac{1}{v(\theta + \pi)}\right)
= \tfrac{1}{2} u\sin^{3}\theta\, \frac{\dd}{\dd\theta}\left\{\frac{1}{\sin^{2}\theta}\left(\frac{1}{v^{2}(\theta)} - \frac{1}{v^{2}(\theta + \pi)}\right)\right\}.
\]
By~\Eq{(6.3)} the left-hand side is independent of~$\theta$, and equal to the constant
$C' - C$. We obtain on integration
\begin{align*}
\frac{1}{v^{2}(\theta)} - \frac{1}{v^{2}(\theta + \pi)}
&= \frac{C' - C}{u} (\sin^{2}\theta · \log\tan \tfrac{1}{2}\theta - \cos\theta), \\
\intertext{or}
\frac{1}{v(\theta)} - \frac{1}{v(\theta + \pi)}
&= \frac{C' - C}{C} · \frac{1}{u} (\sin^{2}\theta · \log\tan \tfrac{1}{2}\theta - \cos\theta).
\end{align*}
It is clearly impossible that the difference of~$1/v$ in opposite directions should
be a function of~$\theta$ of this form; because the origin of~$\theta$ is merely the direction
of relative motion of $S$ and~$S'$, which may be changed at will in different
experiments, and has nothing to do with the propagation of light relative to~$S$.
Hence $C' - C = 0$, and $v(\theta) = v(\theta + \pi)$. Accordingly by~\Eq{(6.3)} $v(\theta)$~is independent
of~$\theta$; and similarly $v'(\theta)$~is independent of~$\theta$. Thus the velocity of
light is uniform in all directions for both observers and is therefore to be
identified with the fundamental velocity.
Public-domain text, read in full here on John Shaqi.
Reviews
Reviews
No reviews yet
Be the first to share your thoughts on this work.
Join the Discussion
Join the discussion
Sign in to leave a comment or review.
Sign InorCreate an account