\section{Setting}\label{sec:setup} We use the notation and results of \cite{Tom26}, cited by their numbers. Throughout, $m\in(\R_{>0})^5$. \subsection{The slice}\label{ss:slice} A normalized central configuration is a solution $(x,y,\delta)$ of the system \cite[(2.1)]{Tom26} of \cite[p.~540]{AK12}. We write $z=x+iy$, $w=x-iy$, $z_{kl}=z_k-z_l$, $w_{kl}=w_k-w_l$, $p_{kl}=\delta_{kl}$ and $r_{kl}=p_{kl}^{-1}$, so that $p_{kl}^2z_{kl}w_{kl}=1$. The \emph{slice} is the affine variety $Y_m=X_m\cap\{u_1=1\}$, where $X_m\subset\C^{18}$ is the zero set of the eighteen rows \[ D_{kl}=p_{kl}^2u_{kl}v_{kl}-1,\qquad Z_k=F_k(u)-F_5(u),\qquad W_k=F_k(v)-F_5(v)\qquad(k\le4) \] of \cite[(2.2)]{Tom26}, in the unknowns $u=(u_1,\dots,u_4)$, $v=(v_1,\dots,v_4)$ and $p=(p_{kl})_{k0$ at the point of $Y_m$ given by a real central configuration. A real central configuration corresponds to exactly two positive normalized central configurations \cite[Lemma~2.1]{Tom26}. The mass vector $m$ is \emph{settled} if $Y_m$ contains no irreducible curve on which some $r_{kl}^2$ is non-constant and $U\not\equiv0$. A settled mass vector has finitely many central configurations up to similarity \cite[Corollary~2.5]{Tom26}. \emph{The colour exchange.} The substitution $(z,w)\mapsto(w,z)$ maps solutions of the first three equations of \cite[(2.1)]{Tom26} to solutions. Followed by the rotation $(u,v)\mapsto(v_1^{-1}u,v_1v)$ of \cite[Lemma~2.2(a)]{Tom26}, it defines the involution \[ \sigma_c(u,v,p)=(v/v_1,\ v_1u,\ p) \] of $Y_m$. It fixes every $p_{kl}$. \subsection{Places and types}\label{ss:places} Let $C\subseteq Y_m$ be an irreducible curve. We use the places at infinity of $C$, the singular places, the exponents $e(f)$ and the expansions of \cite[\S\S2.4, 3.1]{Tom26}. For a pair $kl$ put $\zeta_{kl}=e(z_{kl})$, $\omega_{kl}=e(w_{kl})$ and $x_{kl}=e(r_{kl})=\frac12(\zeta_{kl}+\omega_{kl})$. The \emph{type} of a singular place $P$ is $x(P)=(x_{kl})_{k0 \end{equation} \cite[(3.1)]{Tom26}. A singular place has one of the sixteen diagrams of \cite[Fig.~11]{AK12}, up to relabelling of the bodies and exchange of the colours \cite[\S2.5]{Tom26}. We use the inputs from \cite{AK12} listed in \cite[\S2.6]{Tom26}. We use directly Proposition~1 of \cite{AK12} and the mass conditions (23), (29), (32), (36) and (40) of \cite[\S8]{AK12}. The others, among them the rules of \cite[\S3]{AK12}, Proposition~4 of \cite{AK12} and the facts of \cite[\S\S8.7 and~8.9]{AK12}, enter through the cited results of \cite{Tom26}. For a body $j$ we put $Q_j=\{1,\dots,5\}\setminus\{j\}$. For a triple $T$ let $P_T$ be the complementary pair, and for a set $A$ of bodies let $\mathrm{tri}(A)$ be its set of pairs. We use the following points of $\R^{10}$ of \cite[\S3.2]{Tom26}. \begin{itemize} \item $\pc a_1(T)$, the type of diagram $2$: $0$ on $\mathrm{tri}(T)$ and on $P_T$, $-2$ on the six other pairs. \item $\pc h(s;D_1|D_2)$, the type of diagram $3$ with centre $s$ and diagonals $D_1$, $D_2$, a partition of $Q_s$ into two pairs: $0$ on the four other pairs of $Q_s$ (the sides), $-2$ on $D_1$, $D_2$ and on the four pairs $ks$. \item $\pc v(s;A|B)$ and $\pc v_7(a,b)$, the types of diagrams $6$ and $7$ with centre $s$ and wings $A$, $B$, triples with $A\cap B=\{s\}$: $\pc v_7(a,b)$ is $a$ on $\mathrm{tri}(A)$, $b$ on $\mathrm{tri}(B)$ and $\tfrac32(a+b)-2$ on the four cross pairs, and $\pc v(s;A|B)=\pc v_7(0,0)$. \item $\pc u_A(c)$, the type of diagram $10$, with wings as above: $1$ on $\mathrm{tri}(A)$, $c$ on $\mathrm{tri}(B)$ and $\tfrac{3c-1}2$ on the cross pairs. \item $\pc u_c(j)$ and $\sigma_k(j)(t)$, the types of diagram $14$ with isolated body $j$, and $\pc b_k(j)$, the type of diagram $15$, for $k\in Q_j$: $\pc u_c(j)$ is $1$ on $\mathrm{tri}(Q_j)$ and $c$ on the four pairs $ij$; $\pc b_k(j)$ is $1$ on $\mathrm{tri}(Q_j)$, $0$ on $kj$ and $-\tfrac12$ on $ij$ for $i\in Q_j\setminus k$; and $\sigma_k(j)(t)=(1-t)\,\pc u_0(j)+t\,\pc b_k(j)$. \end{itemize} At a place of diagram~$5$ with $zw$-triangle $T$, each of the other two bodies $X$ has no stroke to $T$. In each colour, say $z$, either all $\zeta_{Xt}$, $t\in T$, are equal and less than $1$ ($X$ is $z$-far), or all $\zeta_{Xt}\ge1$ and at most one of them is larger than $1$ ($X$ is $z$-near), and $X$ is not near in both colours \cite[Lemma~3.4]{Tom26}. With $\bar\omega<1$ the common far exponent in the other colour and $s$ the exceptional exponent, the \emph{modes} of $X$ relative to $T$ are \textup{F}, with $x_{Xt}=c<1$ for all $t$; \textup{Z1}, with $x_{Xt}=\frac{1+\bar\omega}2$ for all $t$; \textup{Zs}$_{t_0}$, with $x_{Xt_0}=\frac{s+\bar\omega}2$ and $x_{Xt}=\frac{1+\bar\omega}2$ for $t\ne t_0$, where $s>1$; and \textup{W1} and \textup{Ws}$_{t_0}$, the same with the colours exchanged. The parameters satisfy $s<2$, $s+\bar\omega<2$, $2s+\bar\omega<3$ and $2s+3\bar\omega<5$. The colour exchange $\sigma_c$ maps a singular place to a singular place with the same type, and exchanges the colours of its diagram. By Lemma~\ref{lem:allcurves}(c) below, the exponents at every singular place satisfy the rules of \cite[Lemma~3.2]{Tom26}, which we recall. The strokes, circles and $zw$-edges are those of the diagram \cite[\S2.5]{Tom26}. For every pair $kl$: \begin{enumerate}[label=\textup{(O\arabic*)}] \item $\zeta_{kl},\omega_{kl}\ge-2$, $\zeta_{kl}+3\omega_{kl}\le4$ and $3\zeta_{kl}+\omega_{kl}\le4$. Hence $\zeta_{kl},\omega_{kl}\le2$ and $-2\le x_{kl}\le1$. \item There is a $z$-stroke $kl$ if and only if $\zeta_{kl}+3\omega_{kl}=4$, and a $w$-stroke if and only if $3\zeta_{kl}+\omega_{kl}=4$. \item On a $z$-stroke, $\zeta_{kl}=3x_{kl}-2$, $\omega_{kl}=2-x_{kl}$ and $0\le x_{kl}\le1$. On a $w$-stroke, $\omega_{kl}=3x_{kl}-2$ and $\zeta_{kl}=2-x_{kl}$. \item $x_{kl}=1$ if and only if $kl$ is a $zw$-edge. A $z$-stroke is maximal \textup{(}its ends are not $z$-close, \cite[Def.~6]{AK12}\textup{)} if and only if $x_{kl}=0$. \item In every triangle $klm$ the minimum of $\zeta_{kl},\zeta_{lm},\zeta_{mk}$ is attained at least twice, and the same holds for $\omega$. \item If $k$ is $z$-circled and $l$ is not, then $\zeta_{kl}=-2$. If $\zeta_{kl}=2$, then $\omega_{kl}=-2$ and $kl$ is a $w$-stroke. \end{enumerate} The results of \cite{Tom26} on singular places that we cite are stated there for \emph{nontrivial} curves, on which $U\not\equiv0$ and some $r_{kl}^2$ is non-constant \cite[\S2.3]{Tom26}. The second condition holds for every irreducible curve. Lemma~\ref{lem:allcurves} lists those that we apply to every irreducible curve, and Appendix~\ref{app:allcurves} checks where $U$ enters their proofs. In their proofs the first condition enters only through \cite[Lemma~3.3]{Tom26}, which says that at a singular place some $x_{kl}$ is at most $0$. We apply the other results of \cite{Tom26} on singular places only to curves with $U\not\equiv0$, in \S\ref{sec:P1} and Appendix~\ref{app:d5P1}. \begin{lemma}\label{lem:allcurves} Let $C\subseteq Y_m$ be an irreducible curve. \begin{enumerate}[label=\textup{(\alph*)}] \item Some $r_{kl}^2$ is non-constant on $C$. A place at infinity at which some $p_{kl}$ has a zero or a pole is singular. The balancing condition \eqref{eq:bal} holds, and $C$ has a singular place. \item Let $P$ be a singular place of $C$, with the bodies numbered as in \cite[Fig.~11]{AK12}. If the diagram of $P$ is not $5$, then $x(P)$ is as in \cite[Table~1]{Tom26}, except for four diagrams. For diagram $7$, $x(P)=\pc v_7(a,b)$ with $00$ has a solution for exactly six values of $\nu$, none of them in $\{1,\frac14,\frac19\}$. \begin{lemma}\label{lem:lex} Let $C\subseteq Y_m$ be an irreducible curve, and let $f$, $g$ be linear forms on $\R^{10}$. It is impossible that $f(x(P))\ge0$ at every singular place $P$ of $C$ and $g(x(P))>0$ at every singular place with $f(x(P))=0$. \end{lemma} \begin{proof} By \eqref{eq:bal}, $\sum_Pd_Pf(x(P))=0$ is a sum of nonnegative terms, so $f(x(P))=0$ at every singular place. Then $\sum_Pd_Pg(x(P))>0$, since $C$ has a singular place by Lemma~\ref{lem:allcurves}(a). This contradicts \eqref{eq:bal}. \end{proof} \subsection{Branches and their lifting}\label{ss:lifting} In \S\S\ref{sec:d3} and~\ref{sec:d6} we will expand the coordinates at places of diagrams $3$ and $6$ in a weighted system. Let $P$ be a singular place of an irreducible curve $C\subseteq Y_m$. After the rotation of \cite[Lemma~3.1]{Tom26}, each coordinate $q\in\{u_k,v_k,p_{kl}\}$ is $\vartheta^{d\,e(q)}y_q(\vartheta)$ with $y_q(0)\ne0$. Suppose that the \emph{weights} $w_q=\frac12e(q)$ are integers, and put $T=\vartheta^{2d}$, so that $q=T^{w_q}y_q$. In each row substitute $q=T^{w_q}y_q$ for every coordinate and divide by the lowest power of $T$. This yields \[ G_r(T,y)=\sum_{i\ge0}T^iG_{r,i}(y)\qquad(r=1,\dots,18), \] with polynomials $G_{r,i}$. Put $e'_q=w_q-w_{u_1}$, $w_q+w_{u_1}$ or $w_q$ for $q$ a $u_k$, a $v_k$ or a $p_{kl}$, and fix a coordinate $f$ with $e'_f=-2$. A \emph{leading solution} is a zero of $G_0=(G_{r,0})_r$ with $y_{u_1}=y_f=1$ and all coordinates nonzero. We write $J_0$ for the $18\times16$ Jacobian matrix of $G_0$ at a leading solution with respect to the sixteen coordinates other than $y_{u_1}$ and $y_f$. A \emph{branch} is a vector $y(\xi)$ of convergent power series in the eighteen coordinates such that $G(\xi^L,y(\xi))\equiv0$ for an integer $L\ge1$, $y_{u_1}\equiv y_f\equiv1$, and $y(0)$ is a leading solution. We then write $T=\xi^L$. \begin{lemma}\label{lem:branch} After a rotation and a change of the local parameter, the coordinates near $P$ are $q=T^{w_q}y_q(\xi)$, where $y$ is a branch. \end{lemma} \begin{proof} This is the proof of \cite[Lemma~E.1]{Tom26}, which treats diagram $2$ with $f=v_1$. A rotation by $y_{u_1}^{-1}$ makes $y_{u_1}\equiv1$. Replacing $T$ by $T'\phi^L$ and rotating by $\phi^{-Lw_{u_1}}$ turns $y_q$ into $\phi^{Le'_q}y_q$, and the equation $\phi^{-2L}y_f(\xi\phi)=1$ has a convergent solution $\phi(\xi)$ with $\phi(0)^{2L}=y_f(0)$ by the implicit function theorem. The rows vanish on $X_m$, so $G(\xi^L,y(\xi))\equiv0$, and setting $\xi=0$ we obtain $G_0(y(0))=0$. \end{proof} Let $y_0$ be a leading solution with $\rank J_0=16$, and let $R$ be a set of sixteen rows such that $J_0[R]$ is invertible. Define $y_1,y_2,\dots\in\C^{16}$, vectors in the sixteen coordinates other than $y_{u_1}$, $y_f$, by \[ y_n=-J_0[R]^{-1}\beta_n[R],\qquad\beta_n=\big[T^n\big]\,G\Big(T,\ y_0+\sum_{k0}$. The coefficient of $T^q$ in $G(T,y)$ is $J_0c_q$ plus a polynomial in the $c_{q'}$ with $q'