I'm a little confused by this proof. As far as I can see the idea is that the Rayleigh quotient should tend to $\lambda_1$ as $\epsilon \rightarrow 0$ by the choice of $f$, but I'm a bit confused why exactly that specific variation $f_\epsilon$ was chosen and what exactly is being done in the final step leading to the conclusion.
Thank you in advance.
