I'm reading through chapter 15 on Algebraic Graph Theory by Godsil and Royle.
The $\mathit{rank \ polynomial}$ for a matroid $M$ is defined as
$$R_M(x,y)=\sum_{A \subseteq\Omega} x^{rk(\Omega)-rk(A)}y^{|A|-rk(A)}, $$
where $rk(X)$ is the the rank function of a subset $X$ of the ground set $\Omega$ of the matroid $M$.
A $\mathit{loop}$ is defined as an element $e$ of the ground set such that $rk(e)=0$ and a $\mathit{coloop}$ is defined as an element $f$ of the amtroid such that $f$ is a loop in the dual matroid $M^*$. The dual matroid $M^*$ is defined as the matroid with the same ground set as $M$ and with rank function
$$ rk^*(A)= |A| +rk(\Omega \setminus A) - rk(\Omega). $$
There follows this theorem which I'm trying to prove:
Let $M$ be a matroid on $\Omega$ and let $e \in \Omega$. Then:
$$ R_M(x,y) = (1+y)R_{M \setminus e}(x,y), \ \ \mathrm{if \ e \ is \ a \ loop;}$$
$$ R_M(x,y) = (1+x)R_{M / e}(x,y), \ \ \mathrm{if \ e \ is \ a \ coloop;}$$
$$ R_M(x,y) = R_{M \setminus e}(x,y) + R_{M / e}(x,y), \ \ \mathrm{otherwise.} $$
I'm okay with the first statement:
\begin{align} R_M(x,y) &= \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A)}y^{|A|-rk(A)} + \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A \cup e)}y^{|A \cup e|-rk(A \cup e)} \\ &= R_{M \setminus e} (x,y) + \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A)}y^{|A|+1 -rk(A)} \\ &= R_{M \setminus e} (x,y) + yR_{M \setminus e} (x,y) \\ &= (1+y)R_{M \setminus e} (x,y). \end{align}
by observing that since $e$ is a loop it will not contribute to the rank of any subset.
I don't see the second statement.
First, I'll restate the definition of a contraction. Let $M$ be a matroid with ground set $\Omega$ and let $T \subset \Omega$. Define $ \rho$ on $\Omega \setminus T$ as $\rho(A)=rk(A \cup T)- rk(T)$. Then $M/T$ is the matroid with rank function $\rho$ defined on $\Omega \setminus T$.
Since $e$ is a coloop:
$$\rho(A)=rk(A \cup e)- rk(e)=rk(A) +rk(e)- rk(e)=rk(A)$$
As well, for loops and coloops, $M/e = M \setminus e$, so
\begin{align} R_M(x,y) &= \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A)}y^{|A|-rk(A)} + \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A \cup e)}y^{|A \cup e|-rk(A \cup e)} \\ &= R_{M / e} (x,y) + \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A \cup e)}y^{|A \cup e| -rk(A \cup e)} \\ &= R_{M / e} (x,y) + \sum_{A \subseteq\Omega \setminus e} x^{rk(\Omega \setminus e)-rk(A) -1}y^{|A|+1 -rk(A) -1}\\ &= (1+x^{-1})R_{M /e} (x,y) \\ &\neq (1+x)R_{M / e} (x,y). \end{align}
I understand the third statement for a graph matroid but it's not clear to me why for a general matroid $R_{M/e}$ accounts for all subsets that contain e.
Also, I'd appreciate some feedback regarding the statement of the question. Was it too detailed?