\section{Index of notation} \label{app:notation} The tables list the symbols that are used in more than one place, with the place where each is introduced. Letters not listed here have a meaning local to one proof or one subsection, stated there. A few letters are deliberately reused in different parts for unrelated objects, as indicated. \renewcommand{\arraystretch}{1.12}\setlength{\tabcolsep}{4pt}\raggedbottom \subsection*{General, and Part~\ref{part:geometry}} \begin{longtable}{@{}p{.23\linewidth}p{.52\linewidth}p{.19\linewidth}@{}} \toprule Symbol & Meaning & Introduced\\ \midrule \endhead $\T$, $\av f$, $R_\vartheta$ & circle $\R/2\pi\Z$; angular average; rotation by $\vartheta$ & \S\ref{subsec:conventions}\\ $e_j$ & $\sqrt2\cos(j\theta)$ & \S\ref{sec:rosen-reference-packet}\\ $w$, $h$, $E$, $g$ & profile (slope), radial height with $h'=w$, $E=1+w^2$, metric $E\,dr^2+r^2d\theta^2$ & \eqref{eq:intro-metric}\\ $z$, $u$, $q$ & height $z=h+qu$; small parameter & \S\ref{subsec:normalization}\\ $\bar g$ & remainder metric $g-dz\otimes dz$ & \S\ref{sec:geometric-closure}\\ $F_\tau$, $\tau$, $\varrho_\vartheta$ & the bending, its phase, intrinsic rotation & \eqref{eq:phase-family}\\ $k$, $p$, $D$, $Q$ & $z_r$, $z_\theta/r$, discriminant, $k+Ep_\theta$ & \eqref{eq:height-jet-notation}\\ $\alpha$ & connection form of the remainder metric & \eqref{eq:conn-one-form}\\ $\mathcal N_h(z)$ & curvature residual & \eqref{eq:cart-curvature-scalar}\\ $\mathcal F_q(w,u)$, $\mathcal A$ & normalized residual & \eqref{eq:cart-removable-equation}\\ $a$, $b$, $c$ & coefficients of the linearization & \eqref{eq:linear-coefficients}\\ $\mathscr L_w$, $L_j$, $L'_{2,w}$ & linearization at the radial height; its $j$th mode; derivative in $w$ & \eqref{eq:intro-linearization}, \eqref{eq:rad-eq-modal-operator}\\ $K_0$, $K_g$ & $ww'/r$; curvature of $g$ & \eqref{eq:cart-determinant-identity}, \eqref{eq:radial-curvature}\\ $B$, $\bar c$, $\bar L$ & first residual correction & Lemma~\ref{lem:type-completion}\\ $d$, $v$, $c_\pm$ & drift, speed and slopes of characteristics & \eqref{eq:completed-characteristics}\\ $\mathfrak F_w$ & connection period & \eqref{eq:cart-mean-path-flux}\\ $T$, $T_m$ & travel angle of a negative component & \S\ref{subsec:layout}\\ \bottomrule \end{longtable} \subsection*{Part~\ref{part:reference}} \begin{longtable}{@{}p{.23\linewidth}p{.52\linewidth}p{.19\linewidth}@{}} \toprule Symbol & Meaning & Introduced\\ \midrule \endhead $X$, $\mathcal P$, $\rho$, $V_0$, $\lambda$ & model coordinate, density, potential data, amplitude of a model & Lemma~\ref{lem:liouville}\\ $H_{j,c}$, $H_{j,M}$ & Schr\"odinger operators of the central and Morse models & \S\ref{subsec:rad-central-singlepoint}, \S\ref{subsec:morse}\\ $\upsilon$ & $r_X/r$ & \eqref{eq:rad-eq-sc6}\\ $\varpi$, $d\mu$, $B_j$ & $12w/r^2$; $\varpi\,dr$; $-\varpi^{-1}L_j$ & \eqref{eq:rad-eq-ha2}\\ $A_c$, $a_c$, $c_c$, $\alpha_c$, $\phi_0$ & constants and seed of the central model & \eqref{eq:rad-eq-sc1}, \eqref{eq:rad-eq-sc2}\\ $g_0$, $\ell_0$ & temporary border & \S\ref{subsec:rad-central-singlepoint}\\ $A$, $B$, $d$ & Morse parameters (in \S\S\ref{subsec:morse}--\ref{sec:radial-tails}) & \eqref{eq:rad-eq-rm5}\\ $d_n$, $A_n$, $B_n$ & parameters of cell $n$; $A_n\asymp e^{-n^2}$ & \S\ref{sec:radial-tails}\\ $\lambda_n$, $\sigma_n$ & amplitude of cell $n$; amplitude ratios & \eqref{eq:rad-eq-js2}\\ $S_n$, $H_n$, $q_n$, $\mathcal E_n$ & $\tfrac23n^3$, $e^{-S_n}$, $n^{5/2}\log(e+n)$, $e^{q_n}$ & \eqref{eq:rad-eq-endpoint-weights}\\ $M_n$ & $\exp\{1+n^{5/2}\log^2(e+n)\}$ & \eqref{eq:M-weight}\\ $N$, $C_*$, $c_0$ & first cell index; band constant; angle constant & \S\ref{sec:radial-tails}\\ $X_c$, $X_L$, $X_R$, $\mathfrak r_n$ & cut and roots of a cell; radius factor & \S\ref{subsec:one-cell}, \S\ref{subsec:joins}\\ $\tau_{n,j}$ & action $j\sqrt{A_n}$ of the ramps of cell $n$ & \eqref{eq:cap-action}\\ $\mathsf p_j$, $\Lambda$ & gauge density $wr^{2j^2-1}$; $j^2(j^2-1)$ (in \S\ref{sec:radial-tails}) & \eqref{eq:rad-eq-rm13}\\ $I_n$, $J_n$; $\mu_n,\nu_n,R_n^I,R_n$ & mass and resistance intervals; masses and resistances & \eqref{eq:mass-resistance-def-new}\\ $a_n$, $y_n$, $b_n$, $c_n$ & node values and coefficients of the recurrence & Lemma~\ref{lem:radial-deep-inverse}\\ $\jn_n$, $\Band_n$ & band edges; frequency bands & \eqref{eq:rad-eq-rm21}\\ $\zeta_j$, $\Xi_n$, $\Upsilon_n$, $\delta_n$ & state variable on a join; amplitude contrast; matching angle & \S\ref{subsec:height-selection}\\ $w_*$, $\phi_*$, $\chi_*$, $g_*$, $\ell_*$ & reference profile, seed, metric control, border & Prop.~\ref{prop:rad-fixed-theorem}\\ $\mathbb L_{w_*}$, $\mathbb G_{w_*}$ & bordered reference operator and its inverse & \eqref{eq:rad-eq-bordered-reference}, Lemma~\ref{lem:rad-green-completion}\\ $Y_{\gamma,s}$, $\ea$, $\eb$, $\el$ & source norm; exponents of the coefficient allowance & \eqref{eq:rad-eq-source-norm}, Lemma~\ref{lem:inner-inverse}\\ $\vartheta_m$, $\chi_L$ & partition of unity; endpoint step function & \S\ref{sec:radial-what}\\ $G_h$, $R_o$, $G$ & inner, outer and matched inverses & \eqref{eq:rad-eq-hn9}, \eqref{eq:outer-robin-new}, \eqref{eq:rad-eq-tm11}\\ $\mS$, $\mK$, $\mA$, $\mB$, $D_0$, $D_b$ & matching matrices & \S\S\ref{subsec:rad-localized-hilbert-assembly}--\ref{subsec:terminal-matching}\\ $\Pi_n$ & transfer bound $e^{Cn^2+C_o\tau_{n,j}}$ & \eqref{eq:transition-inputs-new}\\ $\mathcal D^{\rm rad}_\gamma$ & weighted graph domain & Def.~\ref{def:radial-graph-domain}\\ \bottomrule \end{longtable} \subsection*{Part~\ref{part:nonlinear}} \begin{longtable}{@{}p{.23\linewidth}p{.52\linewidth}p{.19\linewidth}@{}} \toprule Symbol & Meaning & Introduced\\ \midrule \endhead $x_\partial$, $H$, $E_0$, $M$ & continuous versions of the endpoint scales & \eqref{eq:readable-physical-weights}\\ $\ell$, $\varrho$, $\Dr$ & unit-scale radial variable and derivative & \S\ref{subsec:spaces}\\ $\beta$, $\epsilon$ & weight exponent $\tfrac34$; gap between exponents & \eqref{eq:beta-choice}, \S\ref{subsec:two-exponents}\\ $\mathcal X_\gamma$, $\mathcal X_\gamma^0$, $\mathcal Y_\beta$, $\mathcal Z_\gamma$, $\mathcal K_\gamma$, $\Qint$ & spaces of heights, sources, circle data; integral functional & Def.~\ref{def:readable-physical-grading}\\ $\bx=(u,\nu,\kappa,h_L)$, $\Phi_*$ & state; reference height $\phi_*e_2$ & \eqref{eq:rewrite-state}\\ $\chi_m$ & travel-angle controls & Lemma~\ref{lem:quantitative-optical-controls}\\ $\mathfrak n_s$, $\mathfrak g_{s,m}$; $b_0$, $\epsilon_0$, $d_*$ & seminorms of the state; the neighbourhood; fixed loss & \S\ref{subsec:neighbourhood}\\ $f$, $h_R$, $\mathfrak e$ & minus return map, right chart, return residual & \eqref{eq:full-height-one-chart}\\ $\mathcal M_T$, $\mathcal Q_h$ & derivative of $\mathfrak e$ in the chart & \eqref{eq:full-height-chart-derivative}\\ $\widehat L$, $\mathscr E_{\bx}$ & corrected linearization; its error & Prop.~\ref{prop:optical-residual-completion}\\ $\sigma$, $\varphi$; $\xi_L$, $\xi_R$, $\xi_c$ & travel coordinates; root coordinates & \eqref{eq:fixed-strip-coordinates}\\ $W_i$, $h_i$, $\alpha_i$, $V_i$ & canonical unknown, conjugating factor, density, potential & \eqref{eq:fixed-strip-airy-equation}\\ $y$, $\bd$, $P_{\bx}$, $P_*$ & positive coordinate; $y(1-y)$; positive canonical operators & \S\ref{subsec:positive-crossing}\\ $\mathscr C_{\bx}$, $\mathscr S_{\bx}$, $\mathbb L_{\bx}$ & transformations to canonical form & \S\ref{subsec:canonical-problem}\\ $\mathbb V_{t,s}$ & spaces of canonical sources & \eqref{eq:vol-source-spaces}\\ $\mathbb P_{\bx}$, $\bW=(W,a_L,\mu,\lambda)$ & augmented operator and unknown & \eqref{eq:vol-abstract-complete-operator}\\ $k_{\bx}$, $s_{\bx}$, $c_*$ & columns of the augmented operator & \S\ref{subsec:canonical-problem}\\ $\mathbb G_*$, $\mathbb D_*$, $L_0$ & reference inverse, its domain, its loss & Lemma~\ref{lem:two-weights-closed-domain}\\ $|v|_{\mathcal F}$ & $L^1$ norm of the Fourier transform & Lemma~\ref{lem:one-high-factor}\\ $\mathfrak g_{s,m}$, $\mathfrak n_{s,m}$ & localized seminorms of the deformation and of the state & \S\ref{subsec:neighbourhood}\\ $\delta_n^{\rm root}$ & width of the root zones & \eqref{eq:geometry-root-collar-width}\\ $\mathscr R_{\bx}$ & relative operator & \S\ref{subsec:relative}\\ $G_{\bx}$ & inverse at a nearby state & Prop.~\ref{prop:rewrite-volume-inverse}\\ $\mathbf R$, $\mathbf V$, $\mathbf Q$ & residual, approximate inverse, quadratic error & \S\ref{sec:solving-the-equation}\\ $\mathfrak X$, $\mathfrak Y$ & spaces of increments and residuals & \S\ref{sec:solving-the-equation}\\ $C^\circ$, $R^\circ$, $\mathcal D$, $\mathsf J$, $K_{\bx}$, $U_{\bx}$ & blocks of the linear system & \eqref{eq:joint-circle-linear-rows}\\ $(\eta,\mu,\varkappa,\psi)$ & increments of $(u,\nu,\kappa,h_L)$ & \eqref{eq:corrected-coupled-linear-system}\\ $B_\beta$ & space of control increments & \S\ref{sec:solving-the-equation}\\ $\mathcal B_{w,U}$, $\mathcal B^\circ_{\mathbf v}$, $\mathcal C$, $\mathscr P$, $\mathcal U$ & remainders, mean return map, primitive, mean of the height & \S\ref{sec:final-proof}\\ \bottomrule \end{longtable} \paragraph{Words with a fixed meaning.} \begin{itemize} \item \emph{Height} always refers to the height function $z=h+qu$ of the surface or to one of its parts ($h$, $u$, increments). The positive constant by which a piece of the profile is multiplied is its \emph{amplitude} ($\lambda_n$, with \emph{amplitude ratios} $\sigma_n$). \item \emph{Cap} is the stationary part of the closed surface (Section~\ref{sec:geometric-closure}). The two positive parts of a cell are its short and long \emph{ramps} (Section~\ref{subsec:one-cell}). \item \emph{Collar} is a neighbourhood of a boundary circle of the annulus on which the surface is planar. The neighbourhood $|x|<8\delta_n^{\rm root}$ of a turning circle is the \emph{zone} of that root (Lemma~\ref{lem:quantitative-root-collars}). \item A \emph{coefficient power} is a bound $C\mathcal E_n^{C}$, and \emph{coefficient bounds} are bounds $C_s\mathcal E_n^{C_s}$ for derivatives of order $s$; see \eqref{eq:coefficient-power}. A \emph{reference function} is a function of $r$ determined by $w_*$ alone, with coefficient bounds. \item The \emph{conjugating factor} $h_i$ or $h_{\bx}$ is the positive function by which a height is divided to obtain its canonical unknown (Section~\ref{subsec:root-coordinate}). \item \emph{Component} means a maximal interval on which $w_*'$ has one sign; \emph{cell}, \emph{join}, \emph{connector} and \emph{tail} are the building blocks of the profile (Section~\ref{subsec:layout}); a \emph{piece} is one of the cylinders of Section~\ref{subsec:canonical-problem}. \end{itemize} \paragraph{Letters with several meanings.} The following letters have more than one meaning. Each meaning is confined to the sections named, and where two meanings meet in one section they differ by a subscript or an argument, not by the typeface alone. \begin{itemize} \item $q$ is the small parameter of the height; $q_n$ is the exponent of the coefficient allowance. \item $\lambda$ is the amplitude of a model in Part~\ref{part:reference} and the auxiliary scalar of the augmented operator in Section~\ref{sec:stable-inverse}. \item $\mu$ is a mass in Section~\ref{sec:radial-tails}, the measure $d\mu=\varpi\,dr$ in Section~\ref{sec:radial-reference-inverse}, and the increment of $\nu$ in Part~\ref{part:nonlinear}; $\nu_n$ is a mass in Section~\ref{sec:radial-tails} and $\nu$ the metric parameter elsewhere. \item $\kappa_i$ is a boundary coefficient in Section~\ref{sec:radial-reference-inverse}; $\kappa_m$ are the controls in Part~\ref{part:nonlinear}. \item $\tau$ is the phase of the bending; $\tau_{n,j}$ is an action. \item $A,B$ are Morse parameters; the matrices $\mA,\mB$ and the coefficient $B=E/(rQ)$ are distinguished by type or by context. \item $\Lambda$ is $j^2(j^2-1)$ in Section~\ref{sec:radial-tails}, a matrix of outward derivatives in Section~\ref{sec:radial-reference-inverse}, and $|D_\varphi|$ in Section~\ref{sec:stable-inverse}. \item $G$ with a subscript or in blackboard type ($G_h$, $G_o$, $\mathbb G_*$, $G_{\bx}$) is an inverse operator; $G(t,\theta)$ is the speed of a flow on the circle (Lemma~\ref{lem:circle-flow}, Section~\ref{sec:final-proof}); and $G(x,\zeta,\theta)$ is the function of the regularized flow at a root (Section~\ref{subsec:root-coordinate}). \item $D$ is the discriminant; $D_u$, $D_\nu$, $D_{\bx}$ are derivatives; $D_0$, $D_b$, $D_h$, $D_\kappa$ are matrices or operators of Section~\ref{sec:radial-reference-inverse}. \item $E=1+w^2$; $\mathcal E_n$ is the coefficient allowance; $E_b$ is an error matrix in Section~\ref{sec:radial-reference-inverse}; $\mathscr E_{\bx}$ is the error of the return correction; $\Even$ is the even projection. \item $H_n$ is the decay scale and $H^s$ a Sobolev space; $\mathscr H_j$ and $H_{j,c}$ are operators of Part~\ref{part:reference}. \item $T_m$ is a travel angle and $\T$ the circle; $T$ and $T_h$ are matrices in Section~\ref{sec:radial-reference-inverse}. \item $a$ is the leading coefficient of the linearization; $a_L$ is the endpoint constant of a height; $\ea$ is a coefficient exponent; $a_p$ are the constants of Lemma~\ref{lem:ang-base-inverse-regularity}. \item $c_\pm$, $\widehat c_\pm$ are characteristic speeds and $c_0$, $c_i$ root factors in Part~\ref{part:nonlinear}; $\bar c$, $\widehat c$ are coefficients; $c$ is the scalar of a source pair and $c_0$ the constant of $\delta_n$ in Part~\ref{part:reference}. \item $d$ is the drift $(c_++c_-)/2$; $d_i$ is a flux ratio in Section~\ref{sec:radial-reference-inverse}; $d_k$, $d_*$, $d_{\rm geo}$, $d_{\rm tot}$ are losses of derivatives. The boundary-defining function $y(1-y)$ is written $\bd$. \item $h$ is the radial height; $h_L$, $h_R$ are charts of the circle; $h_i$, $h_{\bx}$ are conjugating factors; the subscript $h$ marks the inner problem of Section~\ref{sec:radial-reference-inverse}. \item $R_\tau$ is a rotation; $R_n$, $R_n^I$ are resistances; $R_o$ is the outer resolvent; $R^\circ$ and $\mathscr R_{\bx}$ are operators of Part~\ref{part:nonlinear}. \item $\mK$ is the matching matrix of Section~\ref{sec:radial-reference-inverse}; $\mathcal K_\gamma$ is the space of controls and $K_{\bx}$ the derivative with respect to the controls in Part~\ref{part:nonlinear}. \item $I_n$, $J_n$ are the mass and resistance intervals of a cell; the frequency bands are written $\Band_n$. \end{itemize}