In [TY90] the volume growth and curvature decay rates of the $n$-dimensional Calabi metric $\omega_{\mathcal{C}}$ are estimated as $O(r^{\frac{2n}{n+1}})$ and $O(r^{-\frac{2}{n+1}})$, respectively. When $n=2$, this suggests that $|{\rm Rm}|$ is borderline not in $L^{2}$. However, while the volume estimate is sharp for all $n$, the curvature estimate is sharp only for $n\geq 3$. For $n=2$, the leading term in the asymptotic expansion of the curvature vanishes because the Calabi-Yau metric on an elliptic curve is flat, and the true curvature decay rate of $\omega_{\mathcal{C}}$ for $n=2$ is $O(r^{-2})$. This was also pointed out by R. Kobayashi in [Kob90] but is perhaps most easily seen in the Gibbons-Hawking picture.