If $\chi_G(k)$ is the chromatic polynomial of graph $G$ on $n$ vertices, we want to prove the following statements:
- The only real root $x<1$ for $\chi_G(k)$ except for $x=0$.
- There is no real root $x>n$ for $\chi_G(k)$.
Things that I have tried:
- I can show that the terms of the chromatic polynomial alternate in sign. This means that there is no negative root. But still cannot handle the interval $(0,1)$.
- By definition of the chromatic polynomial, it is obvious that there is no integer root. But hard to say something about the real roots.
Any help is appreciated.
For the first question: if for some $y$ in (0, 1), we have $P(G, y)=0$, then we know $P(G\setminus e, y)= P(G/e, y)$. Since all the efficients of chromatic polynomial are integers and $\deg P(G\setminus e, x)=n$, $\deg P(G/e, x)=n-1$, we can compare the number of digits after demical and conclude that it is impossible.