I am studying compact embedding in Brezis' books and I faced the following problem:

I have done the problem with $p > N$ and $p < N$, but i don't really know why the case $p = N$ reduces to the case $p < N$?
Thanks for your ideas.
Copyright © 2021 JogjaFile Inc.
For all $s \in [1, p)$, you have $W^{1,p}(\Omega) \hookrightarrow W^{1,s}(\Omega)$. Thus, $W^{1,p}(\Omega) \hookrightarrow L^{s^*}(\Omega)$ and this gives you the desired embeddings.