Problem in an example of Introduction to stochastic processes by Lawler page 25

Question

Example. page 25: Consider the two-state Markov chain with $S=\{0,1\}$ and P= $\begin{pmatrix} 1-p & p \\ q & 1-q \\ \end{pmatrix}$ where $0< p,q< 1 $
Asuume the chain starts in state 0 and let $T$ be the return time to 0. It has shown that $\bar{\pi}=(q/(p+q),p/(p+q))$ and hence
$$E(T)=\frac{1}{\pi(0)}=\frac{p+q}{q}$$ In this example we can write down the distribution for T explicitly and verify the above equation For $n>1$
$P(T \geq n)= P(X_1=1,X_2=1,\dots X_{n-1}=1| X_0=0)=p(1-q)^{n-2}$
I couldn't understand the last equation. Please explain why the $P(T \geq n)=p(1-q)^{n-2}$