I want to calculate the length of the circle $\theta=\pi/4$, where $\theta$ is the latitude, on the unit sphere.
I know that the length of a curve $\gamma (t), t \in [0,T]$ on a manifold is given by $\int_{0}^{T} \sqrt{g_{ij} \frac{d\gamma ^i}{dt}\frac{d\gamma ^j}{dt}} dt$ (using Einstein's summation convention).
I also know that the metric on spherical coordinates is given by $g=\begin{pmatrix} 1 & 0 & 0\\ 0 & r^2 & 0 \\ 0 & 0 & r^2 \sin{\theta} \end{pmatrix}$
Using this information the only non zero term in the root is $r^2 \sin^2{\theta} \cdot 1 \cdot 1 = \frac{1}{2}$ and the whole integral gives $\sqrt{2} \pi$ as the result for the length.
Is this correct? I have the feeling that the result should be slightly bigger. Where did I get it wrong?
You can prove that your result is correct simply noting that the radius of the circle at latitude $\pi/4$ is $\cos (\pi/4)=\sqrt{2}/2$. So the leght is $\sqrt{2}\pi$.