Embeddings of $k$-complexes in $2k$-manifolds and minimum rank of partial symmetric matrices
Extracted source text · arXiv 2112.06636
\documentclass[12pt]{article}
%for dvi ps pdf article
\textwidth=16.5cm\textheight=24.5cm\hoffset=-1cm\voffset=-3cm
% \usepackage[T1]{fontenc}
\usepackage[T1,T2A]{fontenc}
\usepackage[english]{babel}
\usepackage{amsfonts,amsmath,amssymb,amsthm,graphicx,afterpage,url}
\input{xy}
\xyoption{all}
\usepackage{xcolor}
%\usepackage{overpic}
%\usepackage{pgf,tikz}
%\usetikzlibrary{arrows}
%\usepackage[notref,notcite]{showkeys}
%\usepackage{epsfig,epstopdf}
%\newcommand{\journal}[1]{#1}
\newcommand{\talkno}[1]{#1}
\newcommand{\journal}[1]{}
\newcommand{\arxiv}[1]{#1}
%\newcommand{\red}[1]{\begingroup\color{red}#1\endgroup}
\graphicspath{{invasurf_figures/}}
\usepackage{hyperref}
\hypersetup{
colorlinks = true, %Colours links instead of ugly boxes
urlcolor = blue, %Colour for external hyperlinks
linkcolor = blue, %Colour of internal links
citecolor = red %Colour of citations
}
\def\con{\mbox{con}}
\def\st{\mbox{st}}
\def\Cl{\mbox{Cl}}
\def\Sq{\mathop{\fam0 Sq}}
\def\diag{\mathop{\fam0 diag}}
\def\adiag{\mathop{\fam0 adiag}}
\def\Map{\mathop{\fam0 Map}}
\def\Hom{\mathop{\fam0 Hom}}
\def\diam{\mathop{\fam0 diam}}
\def\con{\mathop{\fam0 Con}}
\def\id{\mathop{\fam0 id}}
\def\dist{\mathop{\fam0 dist}}
\def\sgn{\mathop{\fam0 sgn}}
\def\gsgn{\mathop{\fam0 gsgn}}
\def\csgn{\mathop{\fam0 csgn}}
\def\sign{\mathop{\fam0 sign}}
\def\sal{\mbox{sl}}
\def\arg{\mathop{\fam0 arg}}
\def\st{\mathop{\fam0 st}}
\def\rk{\mathop{\fam0 rk}}
\def\lk{\mathop{\fam0 lk}}
\def\delet{\mathaccent"7017 }
\def\rel#1{\allowbreak\mkern8mu{\fam0rel}\,\,#1}
\def\Int{\mathop{\fam0 Int}}
\def\Cl{\mathop{\fam0 Cl}}
\def\Imm{\mathop{\fam0 Imm}}
\def\link{\mathop{\fam0 link}}
\def\im{\mathop{\fam0 im}}
\def\ker{\mathop{\fam0 ker}}
\def\coim{\mathop{\fam0 coim}}
\def\conv{\mathop{\fam0 conv}}
\def\coker{\mathop{\fam0 coker}}
\def\arctg{\mathop{\fam0 arctg}}
\def\t{\widetilde}
\def\Nu{N}
\def\R{{\mathbb R}} \def\Z{{\mathbb Z}} \def\Hc{{\mathbb H}} \def\Ca{{\mathbb Ca}} \def\C{\Bbb C} \def\Q{\Bbb Q}
\let\fiverm\tiny
\let\Bbb=\mathbb
\long\def\comment#1\endcomment{}
\theoremstyle{theorem}
\newtheorem{theorem}{Theorem}[subsection]
% \newtheorem{theoremprime}{Theorem}
\newtheorem{lemma}[theorem]{Lemma}
\newtheorem{corollary}[theorem]{Corollary}
\newtheorem{proposition}[theorem]{Proposition}
\newtheorem{addendum}[theorem]{Addendum}
\theoremstyle{definition}
\newtheorem{remark}[theorem]{Remark}
%\newtheorem{remark}[theorem]{Remark}
\newtheorem{example}[theorem]{Example}
\newtheorem{pr}[theorem]{Assertion}
\newtheorem{problem}[theorem]{Problem}
\newtheorem{conjecture}[theorem]{Conjecture}
\newcommand{\EK}[2]{\colorbox{red}{\vphantom{#2}#1}\colorbox{green}{\vphantom{#1}#2}}
\begin{document}
\title{Embeddings of $k$-complexes in $2k$-manifolds and minimum rank of partial symmetric matrices\talkno{\footnote{We are grateful for useful discussions to A. Bikeev, R. Fulek, T. Garaev, A. Kliaczko and M. Tancer.
Supported in part by Russian Foundation for Basic Research Grant No. 19-01-00169.}}}
\author{E. Kogan and A. Skopenkov\talkno{\footnote{E. Kogan: HSE University.
%Higher School of Economics.
\newline
A. Skopenkov:
Moscow Institute of Physics and Technology, Independent University of Moscow.
Email: \texttt{skopenko@mccme.ru}.
\texttt{https://users.mccme.ru/skopenko/}.}}}
\date{}
\maketitle
%\Large
\begin{abstract}
Let $K$ be a $k$-dimensional simplicial complex having $n$ faces of dimension $k$ and
$M$ a closed $(k-1)$-connected PL $2k$-dimensional manifold.
We prove that {\it for $k\ge3$ odd $K$ embeds into $M$ if and only if there are
$\bullet$ a skew-symmetric $n\times n$-matrix $A$ with $\Z$-entries whose rank over $\Q$ does not exceed
$\rk H_k(M;\Z)$,
$\bullet$ a general position PL map $f:K\to\R^{2k}$, and
$\bullet$ a collection of orientations on $k$-faces of $K$
such that for any nonadjacent $k$-faces $\sigma,\tau$ of $K$ the element $A_{\sigma,\tau}$ equals to the algebraic intersection of $f\sigma$ and $f\tau$.}
\talkno{We prove some analogues of this result including those for $\Z_2$- and $\Z$-embeddability.
%particularly interesting case $k=1$ (when $K$ is a graph and $M$ is a surface).
Our results generalize the Bikeev-Fulek-Kyn\v cl-Schaefer-Stefankovi\v c criteria for the $\Z_2$- and $\Z$-embeddability of graphs to surfaces, and are related to the Harris-Krushkal-Johnson-Pat\'ak-Tancer criteria for the embeddability of $k$-complexes into $2k$-manifolds.}
\end{abstract}
\comment
A general position PL map $h:K\to M$ is called a {\it $\Z_2$-embedding}
if $|h\sigma\cap h\tau|$ is even for any pair $\sigma,\tau$ of non-adjacent $k$-faces.
%Our main result is the following
{\bf Theorem.} The complex $K$ having $n$ faces of dimension $k$ has a $\Z_2$-embedding to $M$ if and only if there are
$\bullet$ a symmetric square $n\times n$-matrix $A$ with $\Z_2$-entries, and
$\bullet$ a general position PL map $f:K\to\R^{2k}$
such that
$\bullet$ $A_{\sigma,\tau}=|f\sigma\cap f\tau|\mod2$ for any non-adjacent $k$-dimensional faces $\sigma,\tau$ of $K$,
$\bullet$ $\rk A$ does not exceed the rank of the modulo 2 intersection form $\cap_M$,
$\bullet$ $A$ has only zeros on the diagonal, if $\cap_M$ is even,
$\bullet$ $A$ has a non-zero diagonal entry, if $\cap_M$ is odd.
A {\it $\Z$-embedding} is defined analogously taking the signs into account.
We present criteria for the $\Z_2$- and $\Z$-embeddability of graphs into surfaces and their
algorithmic corollaries.
We present generalizations of these criteria to higher dimensions, i.e. for the embeddability of
$k$-dimensional complexes (or $(k+1)$-regular hypergraphs) into $2k$-dimensional manifolds.
Thus we give a significantly simplified well-structured exposition of 2019 results by Patak-Tancer.
\endcomment
\talkno{\tableofcontents}
\section{Introduction and main results}\label{s:intr}
\subsection{Introduction}\label{ss:introd}
The study of graph drawings on 2-dimensional surfaces is an active area of mathematical research.
Higher-dimensional generalization is classical and has attracted much attention recently.
A major problem asks if there exists an algorithm for recognizing the embeddability of $k$-dimensional complexes into a {\it given} $2k$-dimensional manifold.
A quick algorithm based on a beautiful mathematical result is (as always) preferable.
If $k=1$, or if the manifold is a sphere (or a ball or $\R^{2k}$),
then there is a classical algorithm\talkno{, see e.g. survey \cite[\S1]{MTW}}.
For $k>1$ and a closed manifold different from $S^{2k}$ no algorithm is known.
%with non-trivial intersection form \cite{IF}
\talkno{We shorten `$k$-dimensional (face, manifold, etc)' to `$k$- (face, manifold, etc)'.}
%\talk
%\newpage
A general position PL map $h:K\to M$ of a graph to a 2-surface (a.k.a. a graph drawing on $M$) is called a
{\bf $\Z_2$-embedding} (a.k.a. Hanani-Tutte drawing) if
$|h\sigma\cap h\tau|$ is even for any pair $\sigma,\tau$ of non-adjacent (a.k.a. independent) edges.
Analogously, a general position PL map $h:K\to M$ of a $k$-complex to a PL $2k$-manifold is called a {\bf $\Z_2$-embedding} if $|h\sigma\cap h\tau|$ is even for any pair $\sigma,\tau$ of nonadjacent $k$-faces.
$\Z$-embedding is defined analogously in \S\ref{ss:z}.
Algorithms for recognizing the $\Z_2$- and $\Z$-embeddability
%(defined below)
of $k$-complexes to given $2k$-manifold are also interesting.
{\bf Our main results} are criteria for
%embeddability,
the $\Z_2$- and $\Z$-embeddability of $k$-dimensional complexes to $2k$-dimensional manifolds\talkno{ (Theorems \ref{t:matrix} and \ref{t:matrixz})}.
%, see also Theorem \ref{t:vkswu}.a).
These criteria
$\bullet$ give embeddability criteria\talkno{ by Theorem \ref{t:vkswu}.a};
$\bullet$ reduce embeddability to finding minimal rank of `partial matrix' (and to related problems); this is extensively studied in computer science\talkno{, see e.g. \cite{Ko21} and surveys \cite{MC, NKS}};
$\bullet$ generalize criteria for the $\Z_2$- and $\Z$-em\-bed\-da\-bi\-li\-ty of graphs to surfaces
obtained in Bikeev's paper \cite{Bi21} using ideas of Fulek-Kyn\v cl-Schaefer-Stefankovi\v c \cite{SS13, FK19};
%those criteria are known particular cases of our criteria for $k=1$;
% (Theorem \ref{c:even}.a on graphs in surfaces is perhaps new for $m$ even);
%Theorem \ref{t:matrix} and \ref{c:ifz}.a are higher-dimensional versions of
%\cite[Theorems 1.1 and 1.4]{Bi21}; the extension from the case $k=1$ is non-trivial.
% \cite[Lemma 6]{SS13}
% generalize the van Kampen-Shapiro-Wu criterion for embeddability of $k$-complexes into $\R^{2k}$, and
$\bullet$ are related to the Harris-Krushkal-Johnson-Pat\'ak-Tancer criteria for the embeddability of $k$-complexes into $2k$-manifolds (as explained in Remarks \ref{r:mot}.b and \ref{r:critsg}).
We also obtain interesting Theorems \ref{c:rkcom}, \ref{c:even}, and Corollaries \ref{p:su-vkhi}, \ref{c:rktype}, \ref{c:su-mohi}, \ref{c:ifz} and \ref{c:oddind}; some of them are easier-to-state and so are presented before the criteria.
The cases $k=1$ of all our results are known\talkno{ (\cite[Theorems 1.1 and 1.4]{Bi21}, \cite[Lemma 1]{SS13}; Theorem \ref{c:even} for $k=1$ follows from \cite[Lemma 3]{SS13})}.
%Our proofs of Theorems \ref{c:rkcom} and \ref{c:even}
%is similar to an improved version of \cite[Lemma 3]{SS13}, with $\lfloor\frac{m}2\rfloor$
%replaced by $\lfloor\frac{m-1}2\rfloor$.
%It would be interesting to know if the improved version is correct.
%Theorems
%\ref{t:crit2sg} and
%\ref{t:matrix}
%s \ref{t:critsg} and
%and \ref{t:matrixz} generalize
%the van Kampen-Hanani-Tutte planarity criterion and
%van Kampen-Shapiro-Wu criterion for embeddability of $k$-complexes into $\R^{2k}$
%(see e.g. survey \cite[\S4]{Sk06} and the references therein).
%(Proposition~\ref{t:pl-vk}).
%\talk
%\newpage
\begin{remark}\label{r:mot}
%\footnote{Remarks are not used later in this text.}
(a) {\it Motivation.} The notions of $\Z_2$- and $\Z$-embedding naturally appeared in the studies of embeddings.
Some proofs of the non-embeddability of $k$-complexes into $\R^{2k}$ actually show that these complexes are not $\Z_2$- or $\Z$-embeddable to $\R^{2k}$.
\talkno{See e.g. surveys \cite[Lemma 1.4.3]{Sk18}, \cite[Theorem 1.4]{Sk14} and \cite[Remark 1.2.d]{KS21}.}
Some constructions of embeddings have constructions of $\Z$-embeddings as a natural intermediate step allowing to structure the proof.
\talkno{See e.g. comments after Theorem \ref{t:vkswu},
%Remark \ref{r:vkshwu},
\cite[\S2]{FKT} and survey \cite[\S4]{Sk06}.}
The property of being a $\Z_2$- or $\Z$-embedding is {\it stable}, i.e. is preserved under small enough perturbation of a map (as opposed to the property of being an embedding).
The notion of $\Z_2$-embedding is most actively studied for the case of graphs drawings on surfaces\talkno{, see survey \cite{Sc13} and \cite{SS13, FK19, Bi21}.
Analogous remarks hold for {\it almost embeddings} (defined e.g. in \cite[\S4]{FKT}, \cite[\S1.1]{PT19}).}
%Here we describe the relation of our results to earlier work (in order to make this description clear we have
%to mention that part of analogous description is missing from \cite{PT19}).
(b) {\it Known higher-dimensional criteria.} Criteria for the embeddability of $k$-complexes in $m$-manifolds for $2m\ge3k+3$ are given in \cite[Theorem 1 and Corollaries 6, 7, 8]{Ha69}, in terms of isovariant maps or cohomology obstructions.
%configuration spaces .
Assume further that $m=2k\ge6$.
Such criteria in a different equivalent form involving intersection cocycles
%instead of cocycles obstructing to the existence of equivariant maps)
are given in \cite{Jo02}; \talkno{see exposition in \S\ref{ss:real}, in} \arxiv{(EH) of Proposition \ref{r:crite}, and in Remark \ref{r:inte}.b;} \journal{\cite[Proposition 2.5.1.EH and Remark 2.5.6.b]{KS21e};} cf. \cite{La70}.
%This form is essentially equivalent to Lemma \ref{l:zstar}. unpublished
A criterion in terms of solvability of a system
%($R_{f,I,\Z}$)
of quadratic Diophantine equations
%(i.e. in terms of realizability)
is the conjunction of \cite[Theorems 4, 6 and 15]{PT19}
(cf. Theorem \ref{t:matrixz} and Remark \ref{r:critsg}.a).
A new formula for the cohomology obstruction is given implicitly in \cite[Theorem 3.2]{Kr00} and explicitly in \cite[Theorem 4]{PT19}\talkno{, see }\arxiv{Remark \ref{r:invar}.b and Lemma \ref{l:omega}}\journal{\cite[Remark 2.5.2.b]{KS21e} and \cite[Lemma 2.5.5]{KS21e}}.
An algorithmic criterion for $\Z_2$-embeddability in terms of the cohomology obstruction is \cite[Theorem 10.i and Proposition 21]{PT19}.
For the K\"uhnel problem see \cite[\S1.1]{PT19}, \cite{KS21} and the references therein.
\end{remark}
%Observe that \cite[Theorem 3.2]{Kr00} is formally not used later in \cite{Kr00},
%only Lemma 3.4 uses a step from the proof of Theorem 3.2.
% and Remark \ref{r:inte}.d.
%The paper \cite{PT19} also contains a new formula for the cohomology obstruction
%(giving a new criterion \cite[Theorems 1 and 4]{PT19}),
%See \arxiv{footnote \ref{f:pt}}\journal{\cite[Footnote 3]{KS21e}}.
%(e; exposition) Our exposition is hopefully more accessible than \cite{PT19} (e.g. we do not mention the %Harris-Johnson cohomology obstruction in \S\ref{s:intr}, although we use it in the proof in \S\ref{s:proofs} and %we expose an invariant version of the Patak-Tancer cohomology obstruction in \S\ref{ss:coho}).
%and the rest of Lemmas \ref{star} is implicit in \cite{Jo02, PT19}.
%\talk
%\newpage
{\bf Notation and conventions.}
Denote by $|S|_2\in\Z_2$ the number of elements modulo 2 in a finite set $S$.
Denote by $\Delta_N^k$ the union of the $k$-faces of the $N$-simplex.
A PL manifold $M$ is called {\bf $(k-1)$-connected} if for any $j=0,1,\ldots,k-1$ any continuous map
$S^j=\Delta_{j+1}^j\to M$ extends to a continuous map $\Delta_{j+1}^{j+1}\to M$.
We work in the piecewise-linear (PL) category.
Unless otherwise indicated, $k\ge1$ is any integer, $K$ a (simplicial) $k$-complex having $n$ faces of dimension $k$, and $M$ is any compact $(k-1)$-connected PL $2k$-manifold (possibly, with boundary).
%{\bf Sketches of known definitions.}
In this paragraph $R$ is $\Z_2$ or $\Z$.
Let $H_k(X;R)$ be the $k$-dimensional {\bf homology group} of a complex $X$ with coefficients in $R$,
defined to be the group of {\it homology classes} of {\it $k$-cycles} in $X$ with coefficients in $R$.
For a simple accessible to non-specialists definition
%of $H_k(X;G)$ and of $\alpha\cap_{M,\Z}\beta$
see \cite[\S2]{IF}, \cite[\S6, \S10]{Sk20}.
Let $H_k(M;R):=H_k(X;R)$ for any triangulation $X$ of $M$.
We omit $\Z_2$-coefficients from the notation of (co)homology groups.
%Further subsections of this section are independent on each other and so could be read in any order.
%\newpage
\subsection{$\Z_2$-embeddings}\label{ss:z2z}
\begin{corollary}\label{p:su-vkhi}
%Let $M$ be a $(k-1)$-connected PL $2k$-manifold.
(a) There is an algorithm checking the $\Z_2$-embeddability of $k$-complexes to $M$.
(b) For any $s$ there is a $k$-complex having no $\Z_2$-embedding (and so no embedding) to the connected sum of $s$ copies of $S^k\times S^k$.
As an example one can take the disjoint union
%$(s+1)\Delta_{2k+2}^k$
of $s+1$ copies of $\Delta_{2k+2}^k$.
\end{corollary}
\talkno{Part (a) follows from Theorem \ref{t:matrix} and Lemma \ref{star} for $M=\R^{2k}$.
% (cf. Proposition \ref{p:coho}.$CH$);
Part (b) follows from Theorem \ref{c:rkcom}.}
%\talk
%\newpage
\smallskip
{\bf Definition of the modulo 2 intersection form.}
For $k$-cycles $\alpha,\beta$ modulo 2 in (any triangulation of) $M$ take close (or homotopic) `general position' modulo 2 $k$-cycles $\alpha',\beta'$.
Let
$$\alpha\cap_M\beta:=|\alpha'\cap\beta'|_2\in\Z_2.$$
%\text{ for }R=\Z_2\quad\text{and}\quad\alpha\cap_{M,\Z}\beta:=\alpha'\cdot\beta\in\Z\text{ for }R=\Z,$$
%where the definition of $\cdot$ is recalled at the beginning of \S\ref{ss:z}.
This is well-defined, i.e., is independent of $\alpha',\beta'$ for given $\alpha,\beta$.
This is also independent of the choice of $\alpha,\beta$ within their homology classes.
%the parity of the number of the intersection points intersection of curves shifted to general position
So we denote this by $[\alpha]\cap_M[\beta]\in\Z_2$, where
%the algebraic intersection of
$[\alpha],[\beta]\in H_k(M)$,
%By $\alpha\cap_M\beta\in\Z_2$ or $\alpha\cap_{M,\Z}\beta\in\Z$ we denote the (modulo 2 or integer) algebraic %intersection \cite[\S2]{IF} of $\alpha,\beta\in H_k(M)$ or $\alpha,\beta\in H_k(M;\Z)$.
and denote by $\cap_M:H_k(M)\times H_k(M)\to\Z_2$ the modulo 2 intersection form of $M$.
% defined by $I_R(M)(\alpha,\beta):=\alpha\cap_{M,\Z}\beta$.
For details and references see \cite[\S2]{IF}.
%We shorten $\cap_{\Z_2,M}$ to $\cap_M$, and $\cap_{\Z,M}$ to $\cap_{M,\Z}$.
%Denote by $r(M)$ is the rank of the modulo 2 intersection form $\cap_M$ of a PL $2k$-manifold $M$.
(If $M$ is closed, then $\rk\cap_M=\rk H_k(M)$ by Poincar\'e duality.)
\smallskip
For $R=\Z$ or $\Z_2$ a bilinear form $q:V\times V\to R$ on a $\Z_2$-vector space or $\Z$-module $V$ is called
\textbf{even} if $q(v,v)$ is even for any $v\in V$, and is \textbf{odd} otherwise.
A symmetric matrix with $\Z_2$- or $\Z$-entries is \textbf{even} (for the case of $\Z_2$ a.k.a. alternate) if its diagonal contains only even entries, and is \textbf{odd} otherwise.
%$\bullet$ \textbf{odd} if its diagonal contains at least one odd entry.
The {\bf type} of a bilinear form or a symmetric matrix is its being even or odd.
\begin{corollary}\label{c:rktype}
The $\Z_2$-embeddability of a given $k$-complex to $M$ depends only on the rank and the type of $\cap_M$.
\end{corollary}
\talkno{This follows from Theorem \ref{t:matrix} and Remark \ref{r:crit2}.a.}
%classification of symmetric bilinear forms over $\Z_2$ \cite[Theorem 6.1]{IF}.
%Most of our results reduce embeddability to the following kind of property.
The complex $K$ is called {\bf compatible modulo 2} to a symmetric $n\times n$-matrix $A$ with $\Z_2$-entries
if there is a general position PL map $f:K\to\R^{2k}$ such that
$$(C_f)\quad A_{\sigma,\tau}=|f\sigma\cap f\tau|_2\quad\text{for any non-adjacent $k$-faces \talkno{(for $k=1$ edges)} $\sigma,\tau$ of $K$}.$$
Compatibility modulo 2 is algorithmically decidable\talkno{ by Lemma \ref{star} for $M=\R^{2k}$}.
\begin{theorem}\talkno{[proved in \S\ref{ss:proof}, \S\ref{ss:real}]}\label{t:matrix}
%For a $(k-1)$-connected PL $2k$-manifold $M$ The following are equivalent: ($E$)
There is a $\Z_2$-embedding $K\to M$ if and only if
%($C$)
$K$ is compatible modulo 2 to a matrix $A$ such that $\rk A\le\rk\cap_M$
and $A$ is even/odd if $\cap_M$ is even/odd.
\end{theorem}
\talkno{See more equivalent conditions in}
\arxiv{Proposition \ref{r:crite}.}
\journal{\cite[Proposition 2.5.1]{KS21e}.}
\talkno{
\begin{remark}\label{r:crit2} (a) Recall that
$\bullet$ there are two isomorphism classes of non-degenerate symmetric bilinear forms over $\Z_2$ of a given rank: odd forms and (for an even rank) even forms \cite[Theorem 6.1]{IF};
$\bullet$ $(k-1)$-connected smooth $2k$-manifolds with odd forms exist only for $k=1,2,4$
(see a folklore proof in \cite[footnote 1]{KS21}; the PL analogue presumably holds).
(b) Is there a polynomial (in $n$) algorithm for checking the $\Z_2$-embeddability of $k$-complexes to a fixed $M$
%under assumptions of Corollary \ref{p:su-vkhi}.a
for $k>1$?
By (a) and Corollary \ref{c:rktype} it suffices to answer this question for $M$ being a connected sum of several copies of $S^k\times S^k$ (for $\cap_M$ even) or
%of $\R P^2$,
of $\C P^2$, of $\Hc P^2$ (for $\cap_M$ odd).
For $k=1$ see \cite[Remark 1.2.b]{Bi21}.
%, of $\Ca P^2$
\end{remark}
}
%\talk
%\newpage
\subsection{$\Z$-embeddings and embeddings}\label{ss:z}
%Assume that $K$ is a graph, $M$ is an orientable 2-surface (i.e. $M$ is a sphere with handles),
%and $f:K\to M$ is a general position PL map.
In this section we assume that $M$ is orientable\talkno{ (until Corollary \ref{c:ifz} there is no need to assume that $M$ is $(k-1)$-connected)}.
Let $h:K\to M$ be a general position PL map.
Then preimages $y_1,y_2\in K$ of any double point $y\in M$ lie in the interiors of $k$-faces.
%edges.
By general position, $h$ is `linear' on some neighborhood $U_j$ of $y_j$ for each $j=1,2$.
Given an orientation on the edges, we can take a basis of 2 vectors formed by oriented $hU_1,hU_2$.
The {\it intersection sign of $y$} is the sign $\pm 1$ of the basis.
The {\bf algebraic intersection number} $h\sigma\cdot h\tau\in\Z$ for non-adjacent oriented $k$-faces
%edges
$\sigma,\tau$ is defined as the sum of the intersection signs of all intersection points from $h\sigma\cap h\tau$.
%, see e.g. \cite[\S1.5.4]{Sk18}.
% \S\ref{0vkamz}.
\begin{figure}[ht]
\centerline{\includegraphics[width=4.5cm]
{intersection_sign_graphs.eps} \qquad
\includegraphics{curves.eps}}
\caption{Two curves intersecting at an even number of points the sum of whose signs is zero (left) or non-zero (right).}
\label{f:curves}
\end{figure}
%A general position PL map $f:K\to M$ is called a {\bf $\Z$-(almost) embedding} if
%$f\sigma\cdot f\tau = 0$ for any pair $\sigma,\tau$ of non-adjacent edges.
An example of a $\Z_2$-embedding which is not a $\Z$-embedding is shown in Fig. \ref{f:curves}, right.
%, then the algebraic intersection number $f\sigma\cdot f\tau \in\Z$ for non-adjacent oriented $k$-faces %$\sigma,\tau$ is defined analogously. (almost)
The map $h$ is called a {\bf $\Z$-embedding} if $h\sigma\cdot h\tau=0$ for any pair $\sigma,\tau$ of non-adjacent $k$-faces.
\talkno{(The sign of $h\sigma\cdot h\tau$ depends on an arbitrary choice of orientations for $\sigma,\tau$
and on the order of $\sigma,\tau$, but the condition $h\sigma\cdot h\tau=0$ does not.)}
\begin{theorem}\label{t:vkswu} (a) A
%connected
complex of dimension $k\ge3$ is PL embeddable into a simply connected PL $2k$-manifold $M$ if and only if the complex is $\Z$-embeddable to $M$.
(b) For each $k\ge2$ there is a 2-complex $\Z_2$-embeddable but not $\Z$-embeddable to $\R^{2k}$.
(c) There is a 2-complex $\Z$-embeddable but not embeddable to $\R^4$.
\end{theorem}
%\begin{remark}\label{r:vkshwu}
%\newpage
\begin{proof}[Comments on the proof]
\talkno{Part (a) is a convenient intermediate step (both for this paper and for \cite{PT19}) allowing to construct embeddings from $\Z$-embeddings, and to describe relation of the proof to known results and methods.}
Part (a) is essentially classical.
\talkno{For $M=\R^{2k}$ it is proved in \cite{vK32, Sh57, Wu58}; for a simple exposition see \cite[\S2]{FKT}, \cite[\S4]{Sk06}.
The general case is proved in the same way, just note that in
\cite[Lemma 4, 5 and application of the Whitney trick in the proof of Theorem 3]{FKT} $\R^{2k}$ could be replaced by $M$.
See details in \cite[Corollary 2 and Theorem 4]{Jo02}.
The proofs of \cite{Jo02} mentioned here and in
%comments after Lemma \ref{star}
Remark \ref{r:hajolem} are written for `smooth' maps of complexes but work for PL maps.
The analogue of (a) for $k=1$ is correct (because a compact simply connected 2-manifold is a sphere, and by the Hanani-Tutte Theorem, see e.g. \cite[Theorem 1.5.3]{Sk18}).}
Part (b) is proved in \cite[Example 3.6]{Me06}.
Part (c) is proved in \cite{FKT} (although it is stated there in a different equivalent form), see a simpler proof in \cite[\S2.2]{AMSW}.
\end{proof}
%\end{remark}
%\talk
%\newpage
%It would be interesting to know if $\Z$-embeddability of a graph to a 2-surface implies embeddability.
We conjecture that {\it the analogue of Corollary \ref{p:su-vkhi}.a is correct for $k\ge2$, $M$ orientable and `$\Z_2$-embeddability' replaced by `$\Z$-embeddability' (or, equivalently for $k\ge3$, by `embeddability')}.
%For this problem Theorem \ref{t:matrixz} could be useful.
%If either $k$ is odd or the intersection form of $M$ is indefinite, then the conjecture should be easier.
%If $k$ is even and the intersection form is definite, then the conjecture should be more complicated
%(because non-degenerate definite symmetric bilinear forms over $\Z$ are not classified).
For some $k,M$ an integer analogue of Theorem \ref{t:matrix} is obtained by replacing residues modulo 2 by integers.
The complex $K$ is called {\bf compatible} to an $n\times n$-matrix $A$ with $\Z$-entries if there is a general position PL map $f:K \to\R^{2k}$ such that for some collection of orientations on $k$-faces (for $k=1$ edges) of $K$ we have
$$(C_{f,\Z})\quad A_{\sigma,\tau}=f\sigma\cdot f\tau\quad\text{for any non-adjacent $k$-faces $\sigma,\tau$ of $K$}.$$
It is not clear whether compatibility is algorithmically decidable (in spite of Lemma \ref{l:zstar} for $M=\R^{2k}$).
%Since $M$ is $(k-1)$-connected, we have $H_k(M;\Z)\cong\Z^s$ for some $s$; this $s$ is called $\rk H_k(M;\Z)$$
{\it In this paper by the rank
%of an abelian group,
of a matrix with integer entries, or of a bilinear form on a $\Z$-module, we mean its rank over $\Q$.}
\begin{corollary}\label{c:su-mohi}
(a) A 4-complex embeds into
%the connected sum of $r$ copies of
$\Hc P^2$ if and only if the complex is compatible to a rank 1 symmetric $n\times n$-matrix $A$ with $\Z$-entries, %having rank 1 and
whose any diagonal element is the square of an integer.
(b) A 2-complex has a $\Z$-embedding to $\C P^2$ if and only if the complex is compatible to a matrix as in (a).
\end{corollary}
\talkno{Corollary \ref{c:su-mohi} follows by Theorem \ref{t:matrixz} and Lemma \ref{l:rank1} (for (a) we also need Theorem \ref{t:vkswu}.a).
Corollaries \ref{c:su-mohi}.a, \ref{c:ifz}.a and Theorem \ref{t:matrixz} could perhaps be useful for
\cite[Problem 26]{PT19}.}
%\talk
%\newpage
\smallskip
{\bf Definition of the integer intersection form.}
For integer $k$-cycles $\alpha,\beta$ in (any triangulation of) $M$ take close (or homotopic) `general position' integer $k$-cycles $\alpha',\beta'$.
Let
$$\alpha\cap_{M,\Z}\beta:=\alpha'\cdot\beta'\in\Z.$$
%where the definition of $\cdot$ is recalled at the beginning of \S\ref{ss:z}.
This is well-defined, i.e., is independent of $\alpha',\beta'$ for given $\alpha,\beta$.
This is also independent of the choice of $\alpha,\beta$ within their homology classes.
So we denote this by $[\alpha]\cap_{M,\Z}[\beta]\in\Z$, where $[\alpha],[\beta]\in H_k(M;\Z)$,
and denote by $\cap_{M,\Z}:H_k(M;\Z)\times H_k(M;\Z)\to\Z$ the integer intersection form of $M$.
For details and references see \cite[\S2]{IF}.
Clearly, the product of a torsion element of $H_k(M;\Z)$ with any element is zero.
Hence $\cap_{M,\Z}$ can be regarded as a form $H_k(M;\Z)/T\times H_k(M;\Z)/T\to\Z$, where $T$ is the torsion subgroup.
By $\rk H_k(M;\Z)$ and $\rk \cap_{M,\Z}$ we denote the rank over $\Q$ of $H_k(M;\Z)/T$ and of $\cap_{M,\Z}$ defined on this quotient.
(If $M$ is closed, then $\rk\cap_{M,\Z}=\rk H_k(M;\Z)/T$ by Poincar\'e duality.
A standard fact is that $H_k(M;\Z)$ is a free abelian group, i.e., $T=0$, when $M$ is $(k-1)$-connected.)
\begin{corollary}\label{c:ifz}
(a) For $k$ odd and $M$ closed there is a $\Z$-embedding $K\to M$ if and only if $K$ is compatible
to a skew-symmetric matrix $A$ such that $\rk A\le\rk\cap_{M,\Z}$.
%For $k$ odd (then $\rk\cap_{M,\Z}$ is even) the complex $K$ admits a $\Z$-embedding to $M$ if and only if
%$K$ is realizable by the matrix $H_{\rk\cap_{M,\Z}/2}$.
(b) The $\Z$-embeddability of given $k$-complex to $M$ depends only on $\cap_{M,\Z}$.
\end{corollary}
\talkno{Corollary \ref{c:ifz} follows by Theorem \ref{t:matrixz}; for (a) we also need Lemma \ref{l:Z_Hg} and the fact that any unimodular skew-symmetric bilinear form over $\Z$ is isomorphic to the symplectic form \cite[Proposition 6.2]{IF}.}
%\talk
%\newpage
For general $k,M$ an integer analogue of Theorem \ref{t:matrix} is more complicated.
Let $V$ be an abelian group (or, equivalently, a $\Z$-module).
%whose every non-zero element has order two).
The complex $K$ is called {\bf realizable by a bilinear form $I:V\times V\to\Z$ and a collection of elements $y_\sigma\in V$ indexed by $k$-faces $\sigma$ of $K$}
%homomorphism $\psi:\Z^n\to V$
if there are a general position PL map $f:K\to\R^{2k}$ and
%a collection of
orientations on $k$-faces of $K$ such that
$$
(R_{f,I,\Z})\quad I(y_\sigma,y_\tau) = f\sigma\cdot f\tau\quad\text{for any non-adjacent $k$-faces $\sigma,\tau$ of $K$}.
$$
%Here by $\sigma\in\Z^n$ we denote the `indicator'? vector of a $k$-face $\sigma$.
\talkno{
%\begin{remark}\label{r:realma}
{\bf Comment.} Here is the matrix form of the above definition.
The complex $K$ is called {\it realizable by an $s\times s$-matrix $I$ of integers and
a collection of vectors $y_\sigma\in\Z^n$ indexed by $k$-faces of $K$} if there are
a general position PL map $f:K\to\R^{2k}$, and orientations on $k$-faces of $K$ such that
$$
(R_{f,I,\Z})\quad y_\sigma^TIy_\tau = f\sigma\cdot f\tau\quad\text{for any non-adjacent $k$-faces $\sigma,\tau$ of $K$}.
$$
Since the system ($R_{f,I,\Z}$) of Diophantine equations with coefficients $I,f\sigma\cdot f\tau$ and variables $\{y_\sigma\}$ is quadratic, it is not clear whether its solvability is algorithmically decidable.
%\end{remark}
}
\begin{theorem}\talkno{[proved in \S\ref{ss:proof}, \S\ref{ss:real}]}\label{t:matrixz} There is a $\Z$-embedding $K\to M$ if and only if
%for some (or equivalently, for any) basis in the module $H_k(M;\Z)/\mbox{Torsion}$
%$K$ is realizable by the matrix of $\cap_{M,\Z}$ in the basis.
$K$ is realizable by $\cap_{M,\Z}$ and some collection of homology classes $y_\sigma\in H_k(M;\Z)$.
%[$n$] [from]
\end{theorem}
\begin{corollary}\label{c:oddind} Assume that $k$ is even, $M$ is closed, and $\cap_{M,\Z}$ is odd indefinite with positive and negative ranks $r_+$ and $r_-$.
There is a $\Z$-embedding $K\to M$ if and only if $K$ is realizable by a collection of vectors $y_\sigma\in\Z^{r_++r_-}$ and the square matrix of size $r_++r_-$ having $r_+$ units on the diagonal, $r_-$ elements `$-1$' on the diagonal and zeros outside the diagonal.
\end{corollary}
This follows by Theorem \ref{t:matrixz} and classification of symmetric unimodular odd indefinite bilinear forms over $\Z$ \cite[Theorem 6.3.a]{IF}.
\talkno{
\begin{conjecture}\label{cj:ifzm}
(a) For $k$ odd there is a $\Z$-embedding $K\to M$ if and only if $K$ is compatible to a skew-symmetric
matrix $A$ such that $\rk A \leq r:=\rk\cap_{M,\Z}$ and every $r\times r$ minor of $A$ is divisible by
the determinant of the matrix of $\cap_{M,\Z}$ in some basis of $H_k(M;\Z)$.
(This would follow from (b) and Theorem \ref{t:matrixz}.)
(b) Suppose that $A, H$ are skew-symmetric matrices with $\Z$-entries, $H$ is nondegenerate,
$\rk A\leq\rk H$, and every $\rk H\times\rk H$ minor of $A$ is divisible\arxiv{\footnote{The divisibility is automatic when $\rk A < \rk H$.}}
by $\det H$\journal{(see \cite[footnote 2]{KS21e})}.
Then there is a matrix $Y$ with $\Z$-entries such that $A = Y^THY$.
(c) Both conditions of Corollary \ref{c:oddind} are equivalent to $K$ being compatible to an odd indefinite
symmetric matrix $A$ whose positive and negative ranks do not exceed those of $\cap_{M,\Z}$.
%such that $\rk_{\Q,+}A\le r_+(M)$ and $\rk_{\Q,-}A\le r_-(M)$.
%This conjecture is motivated by the classification of indefinite unimodular symmetric bilinear forms
%over $\Z$ \cite{IF}.
\end{conjecture}
}
%\talk
%\newpage
\begin{remark}[Relation to known results]\label{r:critsg}
(a) For $k\ge3$ and $M$ closed
$\bullet$ Corollary \ref{p:su-vkhi}.a is \cite[Theorem 10.i]{PT19};
%it is a simple corollary of essentially known Lemma \ref{star}; NO! need calculation of $\nu$
$\bullet$ Corollary \ref{c:ifz}.b is essentially the same as \cite[Proposition 7]{PT19}, cf. (e);
$\bullet$ Corollary \ref{c:rktype} is easily deduced from \cite[Proposition 21]{PT19}, while Theorem \ref{t:matrixz} is easily deduced from Theorem \ref{t:vkswu}.a and the conjunction of
\cite[Theorems 4, 6 and 15]{PT19}. (The cited statements of \cite{PT19}, as opposed to Corollary \ref{c:rktype} and Theorem \ref{t:matrixz}, involve cumbersome definition of a cohomology obstruction.)
%We have to warn the reader that this definition is based on meaningless formulas %$\xi(\sigma\times\tau)=(-1)^k\xi(\tau\times\sigma)$ and $\nu_f(\sigma\times\tau)=f(\sigma)\cdot f(\tau)$
%\cite[p. 2 and 7]{PT19}.
%See an explanation why they are meaningless in \cite[Remark 3.3, Critical remarks, (6)]{KS21}; see a meaningful %exposition in \arxiv{Remark \ref{r:inte}.a referring to \S\ref{ss:real}} \journal{\cite[Remark 2.5.6.a referring %to \S 2.3]{KS21e}}.
%The authors of \cite{PT19} prepared an improved version in 2020 (according to critical remarks by A. Skopenkov).
%When that (or even more improved) version will be publicly available, critical remarks on \cite{PT19} here
%will be replaced by remarks on the update (i.e., hopefully will be just deleted).}
%\journal{See \cite[Footnote 3]{KS21e}.}}
Our statements and proofs are simpler than those of \cite{PT19}.
This is mostly because we do not use Pat\'ak-Tancer version of the cohomology obstruction,
but present a simpler equivalent statement of the necessary condition\talkno{(compare the definition of realizability to the definition of the cohomology obstruction in \cite[\S1, \S3]{PT19} exposed in }\arxiv{\S\ref{ss:coho})}\journal{\cite[\S2.5]{KS21e})}.
This simpler equivalent statement is an algebraically invariant (i.e., a basis-free) version of the conjunction of \cite[Theorems 4, 6 and 15]{PT19}.
Having it explicitly stated allows us obtain stronger results with simpler proofs.
%Instead of stating the conjunction and giving its direct proof, the paper \cite{PT19}....
%This allows to use simpler Lemma \ref{l:zstar}.
%See Remarks \ref{r:mot}.b, \ref{r:hajolem}. and \ref{r:invar}.b.
%The equivalence $(E\Leftrightarrow R)$ of Proposition \ref{r:crite} (and its integer version, cf. Remark \ref{r:inte}.b)
%is a user-friendly reformulation
%essentially proved in unpublished paper simplified accessible to non-specialists versions of
%of \cite[Theorems 1, 3 and Proposition 16]{PT19} (or rather of $(i)\Leftrightarrow(ii)$ part of that proposition;
%the $(ii)\Leftrightarrow(iii)$ part easily follows by known methods which is clear from its proof).
%is a version for $\Z$-embeddings (and so not restricted to $k\ge3$) of
(b) The notion of compatibility (modulo 2) appeared in \cite{FK19, Bi21}.
The definition of realizability (modulo 2) appeared implicitly in \cite[\S2.2, equality (3)]{SS13}, \cite[\S3.1, equality (1)]{FK19}, \cite[equality (11) in \S3]{PT19},
\cite[\S2, Proof of the implication ($\Leftarrow$) of Theorem 1.4]{Bi21}, and explicitly in our discussions with A. Bikeev.
%Observe that
\talkno{The collection of vectors $y_\sigma$ from \cite{SS13, FK19} and this paper is essentially the same
as the homomorphism $\psi$ from \cite{PT19}.}
(c) Our constructions of $\Z_2$- and $\Z$-embeddings (in Theorems \ref{t:matrix} and \ref{t:matrixz}) use
the known relation between compatibility and realizability (algebraic Lemma \ref{l:main})
and the known construction of a map inducing a given homomorphism in homology\talkno{ (cf. }\arxiv{Lemma \ref{l:reahom})}\journal{\cite[Lemma 2.5.3]{KS21e})}.\arxiv{\footnote{For this reason a part of our constructions is similar to \cite[\S5, proof of Theorem 6, step 1]{PT19} (whose authors presumably rediscovered this known construction, and so did not refer to it).}} \journal{See \cite[footnote 4]{KS21e}.}
Our construction of matrix $A$ and vectors $y_\sigma$ from $\Z_2$- and $\Z$-embeddings (in Theorems \ref{t:matrix} and \ref{t:matrixz}) use Lemmas \ref{star} and \ref{l:zstar} on the Harris-Johnson cohomology obstructions\talkno{ (see Remark \ref{r:hajolem})}, and the modification of a given map as in
\cite[Lemma 12]{PT19}\talkno{ (cf. }\arxiv{Lemma \ref{l:omega}).}\journal{\cite[Lemma 2.5.5]{KS21e}).}
%, and a simple representation of application of this modification (compare \cite[proof of Lemma 13?]{PT19}
% to equalities (1)-(3) in the proof that $(EH\Rightarrow RH)$).
\talkno{
(d) Geometric proofs (closer to \cite{Bi21}) of Theorems \ref{t:matrix}, \ref{t:matrixz}, and \ref{c:even} could perhaps be obtained using a handle decomposition of $M$.
%A direct geometric proof of Theorem \ref{c:even}.a would give an alternative proof of
%\cite[Theorem 1.1.b]{Bi21}.
%\footnote{
(e) It would be interesting to know if {\it for $k\ge3$ any two $(k-1)$-connected PL $2k$-manifolds are PL homeomorphic if they have isomorphic intersection forms and boundaries PL homeomorphic to $S^{2k-1}$}
(cf. classification of `almost closed' manifolds \cite[p. 170]{Wa62}; pre-H-spaces and H-spaces are defined on p. 168 and 169).
This statement implies Corollary \ref{c:ifz}.b for $k\ge3$ and $M$ closed (so it implies
\cite[Proposition 7]{PT19}).}
\end{remark}
%So proof of Theorem \ref{t:vkswu}.a in \cite{PT19} is dispersed in proofs of
%\cite[Theorems 4 and 6 and Proposition 16]{PT19}.
%These results involve `van Kampen obstruction for embeddability of $K$ into $M$' whose definition in
%\cite{PT19} is very technical.
%Analogous remarks hold for \cite{Ha69}.
%So it might be easier for the reader to prove Theorem \ref{t:vkswu}.a using the methods for $M=\R^{2k}$ as
%exposed in \cite{FKT}, \cite[\S4]{Sk06}, rather than to dig the proof out of \cite{Ha69, PT19}.
%\talk
%\newpage
\section{Proofs}\label{s:proofs}
\subsection{Linear algebraic lemmas}\label{ss:alemmas}
\input{invasurf_alg}
%\talk
%\newpage
\subsection{Construction of $\Z_2$- or $\Z$-embeddings}\label{ss:proof}
%of the main results
\begin{lemma}\label{l:realcl} Let $M$ be a $(k-1)$-connected PL $2k$-manifold, possibly with boundary.
Then any homology class in $H_k(M)$ or in $H_k(M;\Z)$ is represented by a general position PL map $S^k\to M$.
For $k\ge3$ this map can be assumed to be an embedding.
\end{lemma}
\talkno{
\begin{proof}[A well-known proof]
For $k=1$ the lemma is obvious, so assume that $k\ge2$.
Then $M$ is simply-connected.
Hence by Hurewicz theorem $H_{k-1}(M;\Z)=0$.
So by the coefficient exact sequence the reduction modulo 2 $H_k(M;\Z)\to H_k(M)$ is epimorphic.
Hence it suffices to prove the lemma for $\Z$.
By Hurewicz theorem the map $\pi_k(M)\to H_k(M;\Z)$ is epimorphic.
Hence any homology class $\alpha\in H_k(M;\Z)$ is represented by a continuous map $S^k\to M$.
Then by a small shift we obtain a general position PL map $S^k\to M$ representing $\alpha$.
For $k\ge3$ the latter map is homotopic to an embedding by the Penrose-Whitehead-Zeeman-Irwin Theorem \cite{Ir65}, cf. \cite[Theorem 2.9]{Sk06}.
\end{proof}
}
\begin{proof}[Proof of the implication $(\Leftarrow)$ of Theorem \ref{t:matrix}]
Since $K$ is compatible to the matrix $A$, by \talkno{(the implication $(\Leftarrow)$ of)} Lemma \ref{l:main} $K$ is realizable modulo 2 by $\cap_M$ and a collection $y_\sigma\in H_k(M)$.
So there is a general position map $f:K\to\R^{2k}$ such that ($R_{f,\cap_M}$) holds.
Take a $2k$-ball $B\subset\Int M$.
We may assume that $fK\subset B$.
By Lemma \ref{l:realcl} for any $k$-face $\sigma$ the class $y_\sigma$ is represented by a general position PL map $\widetilde\sigma:S^k\to M$.
By general position $\widetilde\sigma(S^k)\cap B=\emptyset$.
We may assume that the maps $\widetilde\sigma$ for different $\sigma$ are in general position to each other.
Take a general position PL map $h:K\to M$ obtained from $f$ by connected summation of $f|_\sigma$ and
$\widetilde\sigma$
%-\cup_\tau\widetilde\tau$
along an arc $l_\sigma$,
%disjoint from $f(K-\Int\sigma)$,
for every $k$-face $\sigma$.
(*) For $k\ge2$ by general position we may assume that
$l_\sigma\cap(f\tau\cup\widetilde\tau S^k\cup l_\tau)=\emptyset$ for $\sigma\ne\tau$.
For $k=1$ we may assume that for the tube $U_\sigma$ along which we make connected summation for $\sigma$,
each of the intersections of $U_\sigma$ with $f\tau$, with $\widetilde\tau S^k$, and with any other tube, consists of an even number of points.
Now the map $h$ is a $\Z_2$-embedding because for any non-adjacent $k$-faces $\sigma, \tau$ we have
$$
|h\sigma \cap h\tau|_2 \stackrel{(1)}=
|f\sigma \cap f\tau|_2+|\widetilde\sigma S^k \cap \widetilde\tau S^k|_2 \stackrel{(2)}=
|f\sigma \cap f\tau|_2+y_\sigma\cap_M y_\tau \stackrel{(3)}= 0.
%|f\sigma \cap f\tau|+y^T_\sigma Iy_\tau \stackrel{(4)}{\equiv}0\mod2.
$$
Here
$\bullet$ equality (1) holds because (*) holds,
$\widetilde\sigma S^k\cap f\tau \subset \widetilde\sigma S^k\cap B=\emptyset$ and analogously
$\widetilde\tau S^k\cap f\sigma=\emptyset$;
$\bullet$ equality (2) holds because $\widetilde\sigma S^k,\widetilde\tau S^k$ represent
$y_\sigma,y_\tau$, respectively;
$\bullet$ equality (3) is $(R_{f,\cap_M})$.
\end{proof}
\talkno{
\begin{proof}[Proof of the implication $(\Leftarrow)$ of Theorem \ref{t:matrixz}]
The proof is obtained from the proof of the implication $(\Leftarrow)$ of Theorem \ref{t:matrix} by the following changes.
Replace the first paragraph by the following sentence.
{\it Since $K$ is realizable by $\cap_{M,\Z}$ and a collection $y_\sigma\in H_k(M)$, there are a general position map $f:K\to\R^{2k}$ and orientations on $k$-faces of $K$ such that ($R_{f,\cap_{M,\Z}\Z}$) holds.}
Replace $\Z_2$ by $\Z$, `an even number of points' by `some points the sum of whose signs is zero',
%vector space $H_k(M)$ by the free $\Z$-module $H_k(M;\Z)/Torsion$,
$|A\cap B|_2$ by $A\cdot B$, and $\cap_M$ by $\cap_{M,\Z}$.
\end{proof}
}
%\talk
%\newpage
\subsection{Deduction of realizability from $\Z_2$- or $\Z$-embeddability}\label{ss:real}
{\it The implication $(\Rightarrow)$ of Theorem \ref{t:matrix}} follows by \talkno{(the implication $(\Rightarrow)$ of)} Lemma \ref{l:main} and Lemma \ref{l:realtree}.
\begin{lemma}\label{l:realtree} If $M$ is $(k-1)$-connected and there is a $\Z_2$-embedding $h:K\to M$, then
$K$ is realizable modulo 2 by $\cap_M$ and some collection of homology classes $y_\sigma\in H_k(M)$.\arxiv{\footnote{The proof shows that the $(k-1)$-connectedness assumption can be weakened to `every map $K^{(k-1)}\to M$ extendable to $K$ is null-homotopic'.}} \journal{(See \cite[Footnote 6]{KS21e}.)}
% (or to the equivalent more cumbersome condition (H) of \cite[\S1.1]{PT19})
%If both $K$ and $M$ are 2-connected, then this is equivalent to `the induced homomorphism ...'
%the matrix of $\cap_M$ in the basis.
\end{lemma}
For a proof
%of Lemma \ref{l:realtree} and Addendum \ref{a:retr}
we need the following essentially known definitions and lemma.
%(Outside this proof we do not need to assume that $M$ is $(k-1)$-connected.)
{\it In the rest of this subsection $M$ is a compact connected PL $2k$-manifold (not necessarily closed or $(k-1)$-connected), and $g:K\to M$ is any general position PL map.}
Take any pair of non-adjacent $k$-faces $\sigma,\tau$.
By general position the intersection $g\sigma\cap g\tau$ consists of a finite number of points.
Assign to the pair $\{\sigma,\tau\}$ the residue
$$\nu(g)\{\sigma,\tau\}:=|g\sigma\cap g\tau|_2.$$
Denote by $K^*$ the set of all unordered pairs of non-adjacent $k$-faces of $K$.
The obtained map $\nu(g):K^*\to\Z_2$ is called the (modulo 2) {\bf intersection cocycle} of $g$ \talkno{ (we call it `cocycle' instead of `map' to avoid confusion with maps to $M$).
Maps $K^*\to\Z_2$ are identified with subsets of $K^*$ consisting of pairs going to $1\in\Z_2$.
(Maps $K^*\to\Z_2$ can also be regarded as `partial matrices', i.e., symmetric arrangements of zeroes and ones in those cells of the $n\times n$-matrix that correspond to the pairs of non-adjacent $k$-faces.)}
Let $\alpha$ be a $(k-1)$-face of $K$ which is not contained in the boundary of a $k$-face $\sigma$ of $K$.
% (for $k=1$, is not the end of an edge $\sigma$).
An {\bf elementary coboundary} of the pair $(\alpha,\sigma)$ is the subset $\delta_K(\alpha,\sigma)\subset K^*$
consisting of all pairs $\{\sigma,\tau\}$ with $\tau\supset\alpha$.
Cocycles $\nu,\nu':K^*\to\Z_2$ (or $\nu,\nu'\subset K^*$) are called {\bf cohomologous} (modulo 2) if
$$
\nu-\nu'=\delta_K(\alpha_1,\sigma_1)+\ldots+\delta_K(\alpha_s,\sigma_s)
$$
for some $(k-1)$-faces $\alpha_1,\ldots,\alpha_s$ and $k$-faces $\sigma_1,\ldots,\sigma_s$ (not necessarily distinct).
Here $+$ is the componentwise addition (corresponding to the sum modulo 2 of subsets of $K^*$).
%\talk
%\newpage
\begin{lemma}\label{star} Let
%$f:K\to M$ be a general position PL map and
$\nu:K^*\to\Z_2$ be a map.
%whose restriction to the $(k-1)$-skeleton of $K$ is null-homotopic?
%%such that the induced homomorphism $g_*:H_j(K;\Z)\to H_j(M;\Z)$ is zero for any $j<k$.
There is a general position PL map $g':K\to M$ homotopic to $g$ and such that $\nu(g')=\nu$ if and only if $\nu$ is cohomologous to $\nu(g)$.
\end{lemma}
%For any $k$-complex $K$ and PL %$2k$-manifold $M$
%the intersection cocycles of homotopic general position PL maps $K\to M$ are cohomologous.
\talkno{\begin{remark}\label{r:hajolem}}
Lemmas \ref{star} and \ref{l:zstar} are essentially known.
%, see Remark \ref{r:hajolem}, although they were not explicitly stated before.
\talkno{
They were proved by van Kampen and Shapiro for $M=\R^{2k}$, see a proof in \cite[Lemma 1.5.8 and Proposition 1.5.9]{Sk18} (modulo 2 version for $k=1$) and in \cite[Lemma 3.5]{Sh57} \cite[\S2]{FKT} (integer version for any $k$, cf. \cite[Remark 1.6.4]{Sk18}).
The proof for an arbitrary $M$ is analogous (e.g. the finger moves are done in a regular neighborhoods of a path in $M$, which is homeomorphic to the $2k$-ball).
The `only if' part of Lemma \ref{l:zstar} is \cite[Theorem 1]{Jo02}, \cite[Lemma 11]{PT19}, the `if' part of Lemma \ref{l:zstar} is \cite[Theorem 6]{Jo02} (in which the assumptions `$n\ge3$' and `$M$ is 1-connected' are superfluous).
\end{remark}}
\begin{proof}[Proof of Lemma \ref{l:realtree}]
%[Proof that $(EH\Rightarrow RH)$] Let $h:K\to M$ be a $\Z_2$-embedding.
Denote by $K^{(k-1)}$ the union of all those faces of $K$ whose dimension is less than $k$.
Take a $2k$-ball $B\subset\Int M$.
Since $M$ is $(k-1)$-connected, $h$ is homotopic to a map $h':K\to M-\Int B$ such that $h'(K^{(k-1)})\subset\partial B$.
%(keeping $h$ to be a $\Z_2$-embedding?) so as to assume that
By general position we may assume that $h'|_{K^{(k-1)}}$ is an embedding.
% into $\partial B$.
%,and $K\cap \Int B=\emptyset$.
%$B\cap h'\Int\sigma=\emptyset$ for any $k$-face $\sigma$.
%By Lemma \ref{star} $h'$ is a cohomological $\Z_2$-embedding.
%For any $k$-face $\sigma$ let $\overline\sigma$ be the cone over $h'\partial\sigma$ with a vertex in $\Int B$.
%We may assume that these vertices are in general position.
Define a map $f':K\to B$ to be $h'$ on $K^{(k-1)}$, and to be the cone map over $h'|_{\partial\sigma}$ with a vertex in $\Int B$ on any $k$-face $\sigma$.
%, so that $f'\sigma=\overline\sigma$.
We may assume that these vertices are in general position.
Then $f'$ is a general position PL map.
It remains to prove that the complex $K$ is realizable by $\cap_M$ and the collection
$y_\sigma:=[f'\sigma\cup h'\sigma]$.
%By Lemma \ref{star} $\nu(h')$ is cohomologous to $\nu(h)=0$
By Lemma \ref{star} the intersection cocycles $\nu(h)=0$ and $\nu(h')$ are cohomologous.
Hence by Lemma \ref{star} for $M=B$ there is a general position PL map $f:K\to B$ homotopic to $f'$ and such that
$$(*)\qquad \nu(f)=\nu(f')+\nu(h').$$
%Let $I$ be the matrix of $\cap_M$ in some basis of $H_k(M)$. For any $k$-face $\sigma$ denote by
%$y_\sigma$ the coordinate vector of the cycle $f'\sigma\cup h'\sigma$ in this basis.
Now ($R_{f,\cap_M}$) follows because for any non-adjacent edges $\sigma, \tau$ we have
$$|f\sigma \cap f\tau|_2 \stackrel{(1)}=
|f'\sigma\cap f'\tau|_2 + |h'\sigma \cap h'\tau|_2 \stackrel{(2)}=
|(f'\sigma\cup h'\sigma)\cap(f'\tau\cup h'\tau)|_2 \stackrel{(3)}= y_\sigma\cap_My_\tau.
$$
Here
$\bullet$ equality (1) holds by (*) because $\nu(f)\{\sigma,\tau\}=|f\sigma\cap f\tau|_2$ and the same for $f',h'$;
%$\bullet$ the second congruence holds because $f'\sigma=\overline\sigma$;
$\bullet$ equality (2) holds because $\sigma\cap\tau=\emptyset$ and $h'|_{K^{(k-1)}}$ is an embedding, so that
$f'\sigma\cap h'\Int\tau\subset B\cap h'\tau=\emptyset$, and analogously $f'\tau\cap h'\sigma=\emptyset$.
$\bullet$ equality (3) holds by general position and definitions of $y_\sigma$, $y_\tau$ and $\cap_M$.
%Hence by Lemma \ref{l:main} $K$ is compatible modulo 2 to a matrix $A$ such that $\rk A\le\rk I$ and $A$
%is even/odd if $\cap_M$ is even/odd.
\end{proof}
\talkno{
\begin{proof}[Proof of the implication $(\Rightarrow)$ of Theorem \ref{t:matrixz}]
The proof is obtained from the proof of Lemma \ref{l:realtree} by the following changes.
Replace $(R_{f,\cap_M})$ by $(R_{f,\cap_{M,\Z},\Z})$, \ $\nu$ by $\nu_{\Z}$, \ $|A\cap B|_2$ by $A\cdot B$, \ $\cap_M$ by $\cap_{M,\Z}$,
and $f'\sigma\cup h'\sigma$ by the integer cycle formed by the union $f'\sigma\cup(-h'\sigma)$ of $f'\sigma$ with the orientation coming from the orientation of $\sigma$, and of $h'\sigma$ with the orientation coming from the opposite orientation of $\sigma$ (and make the same replacement for $\tau$ instead of $\sigma$).
Refer to Lemma \ref{l:zstar} below instead of Lemma \ref{star}.
\end{proof}
%vector space $H_k(M)$ by the free $\Z$-module $H_k(M;\Z)/Torsion$,
Assume that $M$ is oriented.
Take any orientations on $k$-faces of $K$.
Assign to any ordered pair $(\sigma,\tau)$ of non-adjacent $k$-faces the integer
$$\nu_{\Z}(g)(\sigma,\tau):=g\sigma\cdot g\tau.$$
Denote by $\t K$ the set of all ordered pairs of non-adjacent $k$-faces of $K$.
The obtained map $\nu_{\Z}(g):\t K\to\Z$ is called the (integer) {\bf intersection cocycle} of $f$.
This cocycle is {\bf super-symmetric}, i.e., $\nu_{\Z}(g)(\sigma,\tau)=(-1)^k\nu_{\Z}(g)(\tau,\sigma)$.
For oriented edge $AB$ of $K$ denote $[AB:B]=1$ and $[AB:A]=-1$ ($AB$ \emph{goes} to $B$ and \emph{issues} out of $A$).
For oriented 2-face $ABC$ of $K$ denote $[ABC:AB]=[ABC:BC]=[ABC:CA]=1$ and $[ABC:BA]=[ABC:CB]=[ABC:AC]=-1$
(this disagrees with \cite[\S2.4.2]{Sk18}, but both definitions work).
Analogously, for oriented $k$-face $A_0A_1\ldots A_k$ of $K$ and the oriented $(k-1)$-face $\alpha$ obtained from $A_0A_1\ldots A_k$
by dropping the vertex $A_l$ denote $[A_0A_1\ldots A_k:\alpha]=(-1)^l$.
For an oriented $(k-1)$-face $\alpha'$ obtained from $\alpha$ by an odd permutation let $[A_0A_1\ldots A_k:\alpha']=-[A_0A_1\ldots A_k:\alpha]$.
% $$a_{A,\sigma,R,R'}= \begin{cases}
%1 & \text{either ($R\ni A$ and $R'\supset\sigma$) or ($R'\ni A$ and $R\supset\sigma$)}\\
%0 & \text{otherwise}\end{cases}.$$
Let $\alpha$ be an oriented $(k-1)$-face of $K$ which is not contained in the boundary of a $k$-face $\sigma$ of $K$.
An (integer) {\bf elementary coboundary} of the pair $(\alpha,\sigma)$ is the map $\delta_K(\alpha,\sigma):\t K\to\Z$ assigning
$$(-1)^k[\tau:\alpha]\text{ to }(\sigma,\tau),\quad [\tau:\alpha]\text{ to }(\tau,\sigma)\quad\text{and \quad 0 to any other pair}.$$
Cocycles $\nu,\nu':\t K\to\Z$ are called (integer) {\bf cohomologous} if
$$
\nu-\nu'=c_1\delta_K(\alpha_1,\sigma_1)+\ldots+c_s \delta_K(\alpha_s,\sigma_s)
$$
for some integers $c_1,\ldots,c_s\in\Z$, \ $(k-1)$-faces $\alpha_1,\ldots,\alpha_s$, and $k$-faces $\sigma_1,\ldots,\sigma_s$ (not necessarily distinct).
Observe that change of the orientation of $\alpha$ forces change of the sign of $\delta_K(\alpha,\sigma)$.
Hence the cohomology equivalence relation does not depend on the orientations of $(k-1)$-faces.
\begin{lemma}\label{l:zstar} Assume that $M$ is oriented and
%$f:K\to M$ be a general position PL map and
$\nu:\t K\to\Z$ is a super-symmetric map.
There is a general position PL map $g':K\to M$ homotopic to $g$ and such that $\nu_{\Z}(g')=\nu$ if and only if $\nu$ is cohomologous to $\nu_{\Z}(g)$.
\end{lemma}
\subsection{Byproducts: additivity and even embeddings}\label{ss:coro}
%Proofs of Theorems \ref{c:even} and \ref{c:rkcom}
%\smallskip
The \textbf{$\Z_2$-rank} $r(X)$ of a $k$-complex $X$ is the minimal number $\rk\cap_N$ such that $N$ is a $(k-1)$-connected PL $2k$-manifold to which $X$ has a $\Z_2$-embedding.
\begin{theorem}[proved in \S\ref{ss:coro}]\label{c:rkcom}
We have $r(K\sqcup L)=r(K)+r(L)$ for any two $k$-complexes $K,L$.
%(Cf. \cite[Lemma 1]{SS13}.)
\end{theorem}
%See more additivity properties in Remark \ref{?r:crite}.c.
A map $f:K\to M$ is called \textbf{even} if $f_*\alpha\cap_M f_*\alpha=0$ for any $\alpha\in H_k(K)$.
For $k=1$ this is equivalent to \emph{orientability} of $f$: for any cycle subgraph $C\subset K$ the orientation of $M$ does not change along $fC$.
\begin{theorem}\label{c:even}
%(a) If a graph has an orientable $\Z_2$-embedding to the disk with $m$ M\"oebius bands (or, equivalently,
%to the sphere with $m$ cross-caps), then the graph has a $\Z_2$-embedding to the sphere with %$\left\lfloor\dfrac{m-1}2\right\rfloor$ handles.
%In particular, if a graph has an orientable $\Z_2$-embedding to the Klein bottle (or to the M\"oebius band),
%then the graph has a $\Z_2$-embedding to the plane. (b)
If a $k$-complex has an even $\Z_2$-embedding to $M$ and $\cap_M$ is odd, then the complex has a $\Z_2$-embedding to the connected sum of $\left\lfloor\dfrac{\rk\cap_M-1}2\right\rfloor$ copies of $S^k\times S^k$.
\end{theorem}
%For k>1 maps from k-dimensional complex to (k-1)-connected complex are homotopic <=>
%their induced homomorphisms in H_k(.,\Z) are the same.
%Part (a) follows by part (b). Part (b) is proved in \S\ref{ss:coro}. See more in Remark \ref{r:criteven}.
For proofs we need the following addendum to Lemma \ref{l:realtree}.
A {\it maximal $k$-forest $T\subset K$} is a maximal subcomplex $T\subset K$ such that $H_k(T)=0$.
% (this is a maximal forest for $k=1$).
For any $k$-face $\sigma\subset T$ set $\widehat\sigma=\widehat\sigma_{K,T}:=0$.
For any $k$-face $\sigma\subset K-T$ there is a unique $\widehat\sigma=\widehat\sigma_{K,T}\in H_k(K)$ coming from $H_k(T\cup\sigma)$.
(For $k=1$ we have that $\widehat\sigma$ is the union of $\sigma$ and the simple path in $T$ joining the ends of $\sigma$.)
\begin{addendum}\label{a:retr} In Lemma \ref{l:realtree} for any maximal $k$-forest $T\subset K$ we can take the collection $y_\sigma=h_*\widehat\sigma$.
\end{addendum}
\begin{lemma}\label{l:hur} If $T$ is a $k$-complex and $H_k(T)=0$, then any map of $T$ to a $(k-1)$-connected space is homotopic to the map to a point.
\end{lemma}
\begin{proof}[Sketch of a well-known proof]
Since $T$ is a $k$-complex and $H_k(T)=0$, we have $H_k(T;\Z)=0$.
Now the lemma is proved either applying the relative Hurewicz theorem to the mapping cylinder of the map, or using obstruction theory.
\end{proof}
\begin{proof}[Proof of Addendum \ref{a:retr}]
Since $M$ is $(k-1)$-connected, by Lemma \ref{l:hur} we can make a homotopy of $h'$ and assume additionally that $h'(T)$ is contained in a small neighborhood of $\partial B$.
%Take a $k$-cycle $\widehat\sigma_0$ representing $\widehat\sigma$ such that $\widehat\sigma_0-\sigma\subset T$.
Then
$$[f'\sigma\cup h'\sigma] = [f'\sigma\cup h'(\widehat\sigma-\sigma)]+[h'\widehat\sigma] =
0 + h'_*\widehat\sigma = h_*\widehat\sigma.$$
\end{proof}
\begin{lemma}\label{l:mrank1} Let $M$ be a $(k-1)$-connected PL $2k$-manifold.
(a) If $\cap_M$ is odd, then there is a $(k-1)$-connected PL $2k$-manifold $N_1$ such that $\rk\cap_{N_1}=1$.
(b) Let $A$ be the Gramian matrix of some homology classes in $H_k(M)$ with respect to $\cap_M$.
Then there is a $(k-1)$-connected PL $2k$-manifold $N$ such that $\cap_N$ has the same rank and type as $A$.
\end{lemma}
\begin{proof}
(a) \arxiv{\footnote{Part (a) follows from the PL analogue mentioned in Remark \ref{r:crit2}.a.
We present a much simpler direct proof.}}\journal{(See \cite[Footnote 7]{KS21e}.)}
If $k = 1, 2$, then $N_1 = \R P^2,\C P^2$ satisfy the requirements.
Assume that $k\geq3$.
Since $\cap_M$ is odd, there is $\alpha\in H_k(M)$ such that $\alpha\cap_M\alpha = 1$.
By Lemma \ref{l:realcl} $\alpha$ is represented by an embedding $f\colon S^k\to M$.
Let $N_1$ be the regular neighbourhood of $f(S^k)$.
Then $N_1$ is $(k-1)$-connected and $\rk\cap_{N_1}\leq\rk H_k(N_1) = \rk H_k(S^k) = 1$.
Since $\alpha\cap_M\alpha = 1$, it follows that $\rk\cap_{N_1} = 1$.
(b) If $A$ is even, then $\rk A$ is even.
% by Lemma \ref{l:even_mat}.
So let $N$ be the connected sum of $\rk A/2$ copies of $S^k\times S^k$.
If $A$ is odd, then $\cap_M$ is odd.
Hence by (a) there is a manifold $N_1$ such that $\rk \cap_{N_1} = 1$.
So let $N$ be the connected sum of $\rk A$ copies of $N_1$.
In both cases $\cap_N$ has the same rank and type as $A$.
\end{proof}
\begin{proof}[Proof of Theorem \ref{c:rkcom}]
The inequality $r(K\sqcup L) \leq r(K) + r(L)$ is clear.
Let us prove that $r(K\sqcup L) \geq r(K) + r(L)$.
It suffices to prove that $\rk \cap_M \geq r(K) + r(L)$ if there is a PL $\Z_2$-embedding
$h:K\sqcup L\to M$.
Let $T$ be a maximal forest of the complex $K\sqcup L$.
Then $T\cap K, T\cap L$ are the maximal forests of $K, L$, respectively.
Denote by $A_{\sigma,\tau}:=h_*\widehat\sigma\cap_M h_*\widehat\tau$
the Gramian matrix with respect to $\cap_M$ of the homology classes $h_*\widehat\sigma$
corresponding to $k$-faces $\sigma$ of $K\sqcup L$.
Denote by $A_K, A_L$ the Gramian matrices of the homology classes $h_*\widehat\sigma$
corresponding to $k$-faces $\sigma$ of $K$ and of $L$, respectively.
Then $A_K, A_L$ are submatrices of $A$.
So
$$
\rk \cap_M\stackrel{(1)}\geq \rk A \stackrel{(2)}= \rk A_K + \rk A_L\stackrel{(3)}\geq r(K) + r(L).
$$
Here
$\bullet$ $(1)$ follows by Lemma \ref{l:gram};
$\bullet$ $(2)$ follows because for any $\sigma\in K, \tau\in L$ we have
$$
A_{\sigma,\tau} = h_*\widehat\sigma\cap_M h_*\widehat\tau
= \sum_{\sigma'\in C_\sigma, \tau'\in C_\tau} |h\sigma'\cap h\tau'|_2 = 0
%\stackrel{\text{(*)}}
$$
where $C_\sigma, C_\tau$ are some $k$-cycles representing $\widehat\sigma, \widehat\tau$, and the last equality holds because $h$ is a $\Z_2$-embedding.
The inequality $(3)$ follows from the inequalities $\rk A_K\geq r(K)$ and $\rk A_L\geq r(L)$.
In the following paragraph we prove that $\rk A_K\geq r(K)$.
The inequality $\rk A_L\geq r(L)$ is proved analogously.
By Addendum \ref{a:retr} the complex $K$ is realizable by $\cap_M$ and the homology classes $h_*\widehat\sigma$ corresponding to $k$-faces $\sigma$ of $K$.
Hence $K$ is compatible to $A_K$ (cf. Lemma \ref{l:main}).
By Lemma \ref{l:mrank1}.b there is a $(k-1)$-connected PL $2k$-manifold $N$ such that $\cap_N$ has the same rank and type as $A_K$.
By Theorem \ref{t:matrix} there is a $\Z_2$-embedding of $K$ to $N$.
Then $\rk A_K = \rk\cap_N\geq r(K)$.
\end{proof}
\arxiv{{\bf Comment.}
Analogues of Theorem \ref{c:rkcom} hold for {\it even}, for {\it odd} and presumably for {\it closed} $\Z_2$-ranks, which are defined analogously.
The \emph{$\Z$-rank} $r_{\Z}(X)$ of a complex $X$ is defined to be the minimal number $\rk\cap_{N,\Z}$ such that $N$ is a $(k-1)$-connected PL $2k$-manifold to which $X$ has a $\Z$-embedding.
We conjecture that $r_{\Z}(K\sqcup L)=r_{\Z}(K)+r_{\Z}(L)$ for any two $k$-complexes $K,L$.}
\begin{proof}[Proof of Theorem \ref{c:even}]
Take an even $\Z_2$-embedding $h:K\to M$ and a maximal $k$-forest in $K$.
Since $\cap_M$ is odd, by Lemma \ref{l:matrix_HgI} there is a basis $H_k(M)$ in which the matrix $I$ of $\cap_M$ is diagonal, with only units and zeroes on the diagonal.
Let $Y$ be the matrix formed by the coordinates $Y_\sigma$ of homology classes $y_\sigma:=h_*\widehat\sigma$ in this basis.
Denote $A := Y^TIY$.
Since $h$ is even, we have
$A_{\sigma,\sigma}=Y_\sigma^T IY_\sigma=h_*\widehat\sigma\cap_Mh_*\widehat\sigma=0$, so $A$ is even.
We have $I=I_1^TI_1$, where $I_1$ is a $\rk\cap_M\times\rk H_k(M)$ diagonal matrix with units on the diagonal.
Then $A=Y^TI_1^TI_1Y=(I_1Y)^TI_1Y$.
%$Y_1^TY_1$, where $Y_1 =I_1Y$.
Hence by Lemma \ref{l:alternate} we have $\rk A\le \rk \cap_M-1$.
Since $A$ is even,
%by Lemma \ref{l:even_mat}
$\rk A$ is even.
Hence $\rk A\le 2g:=2\left\lfloor\cfrac{\rk \cap_M - 1}2\right\rfloor$.
Thus by Lemma \ref{l:realtree}, Addendum \ref{a:retr} and Lemma \ref{l:main} $K$ is compatible to $A$.
So by Theorem \ref{t:matrix} $K$ has a $\Z_2$-embedding to the connected sum of $g$ copies of $S^k\times S^k$.
% Hence by $(R\Rightarrow E)$ of Theorem \ref{p:coho} $K$ has a $\Z_2$-embedding to $S_g$.
\end{proof}
}
%$$h(K^{(k-1)})\subset B\quad\text{and}\quad B\cap h\Int\sigma=\emptyset\quad\text{for any $k$-face}\quad\sigma.$$
%Moreover, by general position we may assume that $h_{K^{(k-1)}}$ is an embedding into $\partial B$.
%Define a matrix $A$ by the formula
%$$A_{\sigma,\tau} \equiv (\overline{\sigma}\cup h'\tau) \cap (\overline{\tau} \cup h'\sigma) \mod 2.$$
%Here $\cap$ is algebraic intersection of cycles.
%by the formula $A_{\sigma,\tau} \equiv |(\overline\sigma\cap\overline\tau) \cap (\overline\sigma\cap\overline\tau|)| \mod 2$ for any $k$-faces $\sigma, \tau$.
%For any $k$-face $\sigma$ let $y_\sigma$ be the coordinate vector in this basis of the $k$-cycle modulo 2 represented by $\overline\sigma\cup h\sigma$.
%Then $A$ is NOT even/odd if $\cap_M$ is even/odd.
%By Lemma \ref{l:gram} we have that $\rk A \leq \rk \cap_M$.
\arxiv{\begin{remark}\label{r:criteven} The following are equivalent:
(EV) there is an even $\Z_2$-embedding $K\to M$;
(CV) $K$ is compatible modulo 2 to an even matrix $A$ with $\Z_2$-entries such that $\rk A\le\rk\cap_M$;
(RV) $K$ is realizable modulo 2 by $\cap_M$ and a collection of homology classes $y_\sigma\in H_k(M)$ such that $y_\sigma\cap_My_\sigma=0$ for each $k$-face $\sigma$;
{\it Proof.}
The equivalence (CV)$\Rightarrow$(RV) is proved analogously to Lemma \ref{l:main}.
The implication (RV)$\Rightarrow$(EV) is proved analogously to the implication $(\Leftarrow)$ of Theorem \ref{t:matrix} (\S\ref{ss:proof}).
The obtained $\Z_2$-embedding $h:K\to M$ is even because for any $k$-cycle $C$ modulo 2 in $K$ we have
$$h_*[C] = \sum\limits_{\sigma\in C}f\sigma\#\t\sigma S^k =
\sum\limits_{\sigma\in C}f\sigma+\sum\limits_{\sigma\in C}\t\sigma S^k =
\sum\limits_{\sigma\in C}\t\sigma S^k = \sum\limits_{\sigma\in C}y_\sigma,$$
which together with $y_\sigma\cap_My_\sigma=0$ implies that $h_*[C]\cap_M h_*[C]= 0$.
Let us prove the implication (EV)$\Rightarrow$(RV).
Let $h\colon K\to M$ be an even $\Z_2$-embedding.
Take a maximal forest in $K$.
By Addendum \ref{a:retr} we can take the collection $y_\sigma = h_*\widehat\sigma$.
Then $y_\sigma\cap_My_\sigma = h_*\widehat\sigma\cap_Mh_*\widehat\sigma = 0$.
\end{remark}}
\arxiv{
%\newpage
\subsection{Appendix: reformulations of the criteria}\label{ss:coho}
\begin{proposition}\label{r:crite}
The following are equivalent:
(E) there is a $\Z_2$-embedding $K\to M$;
(EH) there is a \emph{cohomological $\Z_2$-embedding} $K\to M$, i.e., a general position PL map
$h:K\to M$ such that $\nu(h)$ is cohomologous to zero;
(C) $K$ is compatible modulo 2 to a matrix $A$ such that $\rk A\le\rk\cap_M$
and $A$ is even/odd if $\cap_M$ is even/odd;
(CM) $K$ is compatible modulo 2 to a matrix $A$ such that $\rk A\le\rk\cap_M$
and $A$ is $\begin{pmatrix} H_g &0 \\ 0 &0 \end{pmatrix}$ or $\begin{pmatrix} E &0 \\ 0 &0 \end{pmatrix}$
%the identity matrix
if $\cap_M$ is even or odd, respectively;
%, banded by zeros.
%(CH) there is a symmetric $n\times n$-matrix $A$ with $\Z_2$-entries such that $\rk A\le\rk\cap_M$,
%\ $A$ is even/odd if $\cap_M$ is even/odd, and $A|_{K^*}$ is cohomologous to $\nu(f)$ for some
%(or, equivalently, for any) general position PL map $f:K\to\R^{2k}$;
(R) $K$ is realizable modulo 2 by $\cap_M$ and a collection of homology classes $y_\sigma\in H_k(M)$;
%there is a collection such that $y_\sigma\cap_M y\tau=v(K)(\sigma,\tau)$.
(R') there is a collection of homology classes $y_\sigma\in H_k(M)$ such that for any $2k$-cycle $C\subset K^*$ we have $\sum\limits_{ \{\sigma,\tau\} \in C} y_\sigma\cap_M y_\tau = v_C(K)$;
(RT) there is a homomorphism $\psi:H_k(K)\to H_k(M)$ such that for any $2k$-cycle $C\subset K^*$ and for some (equivalently, for any) maximal $k$-forest in $K$ we have
$\sum\limits_{ \{\sigma,\tau\} \in C} \psi\widehat\sigma\cap_M\psi\widehat\tau = v_C(K)$;
(RI) there is a homomorphism $\psi:H_k(K)\to H_k(M)$ such that if $\alpha_1,\ldots,\alpha_q,\beta_1,\ldots,\beta_q$ are $k$-cycles in $K$ and $C:=\sum\limits_{j=1}^q[\alpha_j\times\beta_j]$ is a $2k$-cycle in $K^*$, then\footnote{This $C$ is automatically a $2k$-cycle in $K^2/\Z_2\supset K^*$.
Since $K$ is $k$-dimensional, $H_k(K)$ is the group of $k$-cycles in $K$, so $\psi\alpha_j$ and $\psi\beta_j$ do make sense.}
$\sum\limits_{j=1}^q \psi\alpha_j\cap_M \psi\beta_j=v_C(K)$;
(RH) there is a homomorphism $\psi:H_k(K)\to H_k(M)$ such that $\omega(\psi)=v(K)$.
\end{proposition}
%arXiv:1101.0967 approx?
%($RH$) there is an $n\times\rk H_k(M)$-matrix $Y$ with $\Z_2$-entries and a basis in $H_k(M)$
%such that for the matrix $I$ of $\cap_M$ in this basis and some (or equivalently, any) and general
%position PL map $f:K\to\R^{2k}$, cocycles $(Y^TIY)|_{K^*}$ and $\nu(f)$ are cohomologous.
The set $K^*$, the intersection cocycle $\nu(f)$ and cohomology of cocycles are defined in \S\ref{ss:real}.
By Lemma \ref{star} the property of being a cohomological $\Z_2$-embedding is homotopy invariant (as opposed to the property of being a $\Z_2$-embedding).
\smallskip
{\bf Definitions required for (R').}
%Denote by $F$ the set of $k$-faces of $K$.
%Recall that $K^*$ is the set of all unordered pairs of non-adjacent $k$-faces of $K$.
A subset $C\subset K^*$ is called a {\it $2k$-cycle} (modulo 2) if for each $k$-face $\sigma$ and $(k-1)$-face $a$ there is an even number of $k$-faces $\tau\supset a$ such that $\{\sigma,\tau\}\in C$.
For a $2k$-cycle $C\subset K^*$ and a general position PL map $f:K\to\R^{2k}$ the {\it $C$-van Kampen number} is defined as $v_C(K):=\sum\limits_{ \{\sigma,\tau\} \in C} |f\sigma\cap f\tau|_2$.
This depends only on $K,C$, not on $f$.
Cf. \cite[Remark 1.5.5.c]{Sk18}.
\smallskip
{\bf Comment.} Clearly,
(a) for any disjoint $k$-cycles $\alpha,\beta$ in $K$ considered as unions of $k$-faces the following is a $2k$-cycle:
$$[\alpha\times\beta]:=
\left\{\{\sigma,\tau\}\in K^*\ :\ \sigma\subset\alpha,\ \tau\subset\beta\right\};$$
(b) $(K_5)^*$ and $(K_{3,3})^*$ are 2-cycles;
(c) the sum (modulo 2) of $2k$-cycles is a $2k$-cycle.
For $k=1$ any 2-cycle in $K^*$ is a sum of 2-cycles from (a) and (b) above \cite[\S3.4]{Sa91}.
\smallskip
{\bf Definition required for (RI).}
For $k$-cycles $\alpha,\beta$ in $K$ considered as unions of $k$-faces let
$$[\alpha\times\beta]:=
\left\{\{\sigma,\tau\}\in K^*\ :
\ \sigma\subset\alpha,\ \tau\subset\beta,\ \{\sigma,\tau\}\not\subset\alpha\cap\beta\right\}.$$
This is a $2k$-cycle.
\smallskip
{\bf Definitions required for (RH).}
For an assignment $\overline\psi$ of elements of $H_k(M)$ on $k$-faces of $K$
%(and for $K^*$ defined in \S\ref{ss:real})
define a mapping (a cocycle)
$$\omega(\overline\psi):K^*\to\Z_2\quad\text{by}\quad \omega(\overline\psi)(\sigma,\tau):=\overline\psi\sigma\cap_M\overline\psi\tau.$$
For any homomorphism $\psi:H_k(K)\to H_k(M)$ there is an assignment $\overline\psi$ of elements of $H_k(M)$
on $k$-faces of $K$ such that $\psi(\sigma_1+\ldots+\sigma_s)=\overline\psi(\sigma_1)+\ldots+\overline\psi(\sigma_s)$ for any $k$-cycle $\sigma_1+\ldots+\sigma_s$.
(In other words, $\psi$ can be regarded as an element of $H^k(K;H_k(M))$ which is such an assignment $\overline\psi$ up to cohomology.)
Let $H^{2k}(K^*)$ be the group of modulo 2 cohomology classes of cocycles $K^*\to\Z_2$.
It is easy to check that
$$\omega(\psi):=[\omega(\overline\psi)]\in H^{2k}(K^*)$$
is well-defined, i.e. that cocycles $\omega(\overline\psi)$ are cohomologous for cohomologous assignments $\overline\psi$.
%In ($RH$) and below we use the definitions , cf. the end of \cite[\S1.5.3]{Sk18}. \S\ref{s:vkam2-cc}.
By Lemma \ref{star} for $M=\R^{2k}$ the class $v(K):=[\nu(f)]\in H^{2k}(K^*)$, where
%the cohomology class of the intersection cocycle of
$f:K\to\R^2$ is any general position map, is well-defined.
\begin{remark}\label{r:invar}
(a) The properties (RH) and (RI) are topologically invariant, i.e., are stated in terms of group $H_k(K)$ and elements of the group $H^{2k}(K^*)$, which groups are isomorphic for PL homeomorphic complexes $K$, and which elements map naturally under PL homeomorphisms.
In spite of that, the property (RH) is similar to the non-invariant statement of \cite[Theorem 4]{PT19}.
The above explicit statement of (RH) could be more useful to a reader than \cite[footnote 10]{PT19}.
(b) An alternative way to state the above definition of $\omega(\psi)$ (analogous to \cite[before Theorem 3.2]{Kr00}) is $p^*\omega(\psi)=((\psi\times\psi)^*\cap_M)|_{\t K}$, where $p:\t K\to K^*$ is the quotient map and
%Consider the restriction of the inverse image in $H^{2k}(K\times K)$ under $\psi\times\psi$ of the intersection form of $M$
$\cap_M$ is considered as an element of $\Hom(H_k(M)\otimes H_k(M))\cong H^k(M)\otimes H^k(M)$
which is a direct summand of $H^{2k}(M\times M)$ by the K\"unneth formula.
Since $\cap_M$ is symmetric,
%this restriction
$((\psi\times\psi)^*\cap_M)|_{\t K}$ is symmetric, hence this restriction comes from $H^{2k}(K^*)$.
%Observe that $\omega(\psi)$ is the element in $H^{2k}(K^*)$ corresponding to
Theorem 3.2 of \cite{Kr00} concerns the case of embedding onto spines, $k=2$, rational coefficients, and embeddings (not almost embeddings as in \cite[Theorem 4]{PT19}).
Generalizations to arbitrary embeddings, to $k\ne2$, and to integer coefficients, are straightforward.
For $k\ge3$ almost embeddability is equivalent to embeddability by Theorem \ref{t:vkswu}.a (it would be interesting to see whether the obstructions of \cite{Kr00} are obstructions to almost embeddability, not just to embeddability).
The completeness of this obstruction follows by seeing that the obstruction of \cite[Theorem 3.2]{Kr00} equals to the obstruction of \cite{Jo02}; the completeness was explicitly proved only as \cite[Theorem 6]{PT19}.
Since the van Kampen obstruction has order 2, the rational van Kampen obstruction of \cite[Theorem 3.2]{Kr00} is zero.
\end{remark}
\smallskip
Below (except for the proof of Proposition \ref{r:crite}) $M$ need not be closed or $(k-1)$-connected.
A map $g:K\to M$ \emph{induces} the homomorphism $g_*:H_k(K)\to H_k(M)$ by the formula $g_*[\alpha]:=[g\alpha]$ for any $[\alpha]\in H_k(K)$.
The following lemma is simple and well-known.
\begin{lemma}\label{l:reahom} Any homomorphism $H_k(K)\to H_k(M)$ is induced by some map $g:K\to M$.
\end{lemma}
\begin{proposition}\label{l:kunneth} Let $T$ be a maximal $k$-forest of $K$.
(a) Every $k$-cycle $\alpha$ in $K$ equals to $\displaystyle\sum_{\sigma\in\alpha}\widehat\sigma$.
(b) Then $\overline T := T\times K \cup K\times T$ is a maximal $2k$-forest of $K\times K$.\footnote{The product $K\times K$ has a natural structure of a cell complex, the cells being $\sigma\times\tau$ for faces $\sigma,\tau$ of $K$. A maximal $k$-forest and $\widehat{\sigma\times\tau}$ are defined for cell complexes analogously to the simplicial complexes.}
(c) For this maximal $k$-forest and any $k$-faces $\sigma,\tau$ of $K$ we have $\widehat{\sigma\times\tau}=\widehat\sigma\times\widehat\tau$.
(d) For any $2k$-cycle $Z$ in the simplicial deleted product
$$\t K:=\cup\{\sigma\times\tau\subset|K|\times|K|\ :\ \sigma\cap\tau=\emptyset\}$$
symmetric w.r.t. the involution $(x,y)\leftrightarrow(y,x)$ we have
$Z = \sum\limits_{\{\sigma,\tau\}\colon(\sigma,\tau)\in Z}
(\widehat\sigma\times\widehat\tau+\widehat\tau\times\widehat\sigma)$.
%there are $k$-cycles $\alpha_1,\ldots,\alpha_q,\beta_1,\ldots,\beta_q$ in $K$
%such that $Z=\sum\limits_{j=1}^q(\alpha_j\times\beta_j+\beta_j\times\alpha_j)$.
(e) For any $2k$-cycle $C\subset K^*$ we have
$C=\sum\limits_{\{\sigma,\tau\}\in C} [\widehat\sigma\times\widehat\tau]$.
%there are $k$-cycles $\alpha_1,\ldots,\alpha_q,\beta_1,\ldots,\beta_q$ in $K$
%such that $C=\sum\limits_{j=1}^q[\alpha_j\times\beta_j]$.
\end{proposition}
%(d) is incorrect for $C$ in $K\times K$: take K=S^k and C=S^k\times S^k.
%For the 2-cycle $C=(K_5)^*$ (or $C=(K_{3,3})^*$) there are 1-cycles %$\alpha_1,\ldots,\alpha_q,\beta_1,\ldots,\beta_q$ in $K_5$ (or in $K_{3,3}$) such that %$C=\sum\limits_{j=1}^q[\alpha_j\times\beta_j]$.
\begin{proof}
{\it Part (a) follows} because the difference is a $k$-cycle, lies in $T$, and hence is trivial.
{\it Proof of (b,c).}
Using the Mayer-Vietoris sequence and K\"unneth formula, we see that $H_{2k}(\overline T) = 0$.\footnote{Here is an independent proof.
Take a $2k$-cycle $Z\subset\overline T$.
\newline
If $Z\supset T\times T$, then $Z=\emptyset$ because by K\"unneth formula
$H_{2k}(T\times T)=H_k(T)\otimes H_k(T)=0$.
\newline
If $Z\not\supset T\times T$, then w.l.o.g. for some $k$-face $\sigma\in K - T$ the set
$Z_\sigma := \{\tau \ :\ (\tau, \sigma)\in Z\}$ is non-empty.
This set is a $k$-cycle in $T$.
A contradiction.}
For any pair $\sigma\times\tau\in K\times K$ we have a nontrivial $2k$-cycle
$\widehat\sigma\times\widehat\tau\subset \overline T\cup\{\sigma\times\tau\}$.
{\it Part (d) follows} because
$$ Z \stackrel{(1)}= \sum_{\sigma\times\tau\in Z}\widehat{\sigma\times\tau}
\stackrel{(2)}= \sum_{\sigma\times\tau\in Z}\widehat\sigma\times\widehat\tau
\stackrel{(3)}= \sum_{\{\sigma,\tau\}\colon(\sigma, \tau)\in Z}
(\widehat\sigma\times\widehat\tau + \widehat\tau\times\widehat\sigma). $$
Here (1) holds by (a), (2) holds by (c), and (3) holds because $Z\in\widetilde K$ and $Z$ is symmetric.
%This is the required representation of $Z$.
%Let $Z := \{(\sigma, \tau)\ :\ \{\sigma, \tau\}\in C\}$.
%By (c) there are $k$-cycles $\alpha_j, \beta_j, j = 1, \ldots, q$ such that
% $\displaystyle Z = \sum_{j = 1}^q(\alpha_j\times\beta_j + \beta_j\times\alpha_j)$.
% Then $C = \sum_{j = 1}^q[\alpha_j\times\beta_j]$.
{\it Part (e) follows by (d)} because the $\Z_2$-quotient of
$\widehat\sigma\times\widehat\tau + \widehat\tau\times\widehat\sigma$ is $[\widehat\sigma\times\widehat\tau]$.
\end{proof}
{\bf Comment.}
Propositions \ref{l:kunneth}.de describe simple generators of $H_{2k}(K^*),H_{2k}(\t K)$ which are not elements of $H_{2k}(K^*),H_{2k}(\t K)$.
This is useful for our applications, perhaps could be useful elsewhere, but properly considered to be not very natural.
For this reason Propositions \ref{l:kunneth}.de can be new, although homology of the deleted product have been much studied, for complexes see e.g. \cite{FH10, MS17} and the references therein.
%BF09,
%G. Yo, Cohomology mod p of deleted cyclic product of a manifold, Sci. Sinica, 12 (1963), 1779--1794.
%For $k=1$ Proposition \ref{l:kunneth}.d also follows from the cases $C=K_5^*,K_{3,3}^*$ and \cite[\S3.4]{Sa91}.
%For $K=\Delta^k_{2k+2}$
%the k-skeleton of the (2k+2)-simplex
%and $C=K^*$ Proposition \ref{l:kunneth}.d also follows analogously to \cite[Lemma 1.5]{KS21}.
%E.g. for $k=1$ take the 3 splittings into pairs of the 4 edges issuing out of a vertex of $K_5$.
%The 3 splittings define 3 pairs of triangles in $K_5$: $a_1,b_1,a_2,b_2,a_3,b_3$.
%Then $K_5^*=\sum\limits_{j=1}^3 [a_j\times b_j]$.
\begin{lemma}\label{l:omega}
Let $g:K\to M$ be a map whose restriction to the $(k-1)$-skeleton $K^{(k-1)}$ of $K$ is null-homotopic.
Take a $2k$-ball $B\subset M$ and a general position map $g':K\to M$ homotopic to $g$ such that
$g'(K)\subset M-\Int B$ and $g'(K^{(k-1)})\subset\partial B$.
For each $k$-face $\sigma$ of $K$ take a $k$-disk $\overline\sigma\subset B$ spanning $g'(\partial\sigma)$.
For any non-adjacent $k$-faces $\sigma,\tau$ let
$$\omega(g)(\sigma,\tau):=\left|(\overline\sigma\cup g'\sigma)\bigcap(\overline\tau\cup g'\tau)\right|_2.$$
(a) The cocycle $\omega(g)$ is well-defined, i.e. is independent on $B,g'$ for fixed $g$.
(b) The cohomology class of $\omega(g)$ is $\omega(g_*)$.
%The cohomology class of $\Omega(g)$ is $\Omega(g_*)$.
%some representative $\overline{g_*}$ of
(c) The cohomology class of $\nu(g)$ is $\omega(g_*)-v(K)$.
\end{lemma}
Lemma \ref{l:omega} (and its integer version of Remark \ref{r:inte}.d) is essentially proved in
\cite[Lemmas 12 and 13]{PT19}.
Parts (a,b) are simple.
Part (c) follows from
$$\omega(g)(\sigma,\tau) = |g'\sigma\cap g'\tau|_2+|\overline\sigma\cap \overline\tau|_2 = (\nu(g)+\nu(f_0))(\sigma,\tau),$$
where $f_0:K\to B$ is the map coinciding with $g'$ on $K^{(k-1)}$ and such that $f_0(\sigma)=\overline\sigma$ for any $k$-face $\sigma$.
% (cf. sketch of a proof of Remark \ref{r:g}.d).
%Part (b) shows that $\nu(g)$ depends on $g$ (or rather on $g_*$) in a `quadratic' way.
\begin{proof}[Proof of Proposition \ref{r:crite}]
The equivalence (E)$\Leftrightarrow$(C) forms Theorem \ref{t:matrix}.
The implication (E)$\Rightarrow$(EH)
%, (C$\Rightarrow$CH) and (R$\Rightarrow$ RH) are
is clear.
The implication (E)$\Leftarrow$(EH) is the particular case of Lemmas \ref{star} for $\nu=0$.
%$(C\Leftarrow CH)$ and $(R\Leftarrow RH)$ by Lemma \ref{star} for $M=\R^{2k}$,
We have (C)$\Leftrightarrow$(R) by Lemma \ref{l:main}, and (C)$\Leftrightarrow$(CM) by Remark \ref{r:crit2}.a.
The implication (E)$\Rightarrow$(R) is Lemma \ref{l:realtree}.
The implication (R)$\Rightarrow$(E) is essentially proved in the proof of $(\Leftarrow)$ of Theorem \ref{t:matrix} (\S\ref{ss:proof}).
The implications (R)$\Rightarrow$(R')$\Rightarrow$(RI) are clear.
The implication (RH)$\Rightarrow$(EH) follows from Lemmas \ref{l:reahom} and \ref{l:omega}.
% $(EH\Rightarrow RH)$ follows from Lemma \ref{l:omega}.
So it remains to prove the implications (RI)$\Rightarrow$(RT)(any) and (RT)(some)$\Rightarrow$(R')$\Rightarrow$(RH).
%{\it Proof that (E)$\Rightarrow$(RT).}
%Since $C$ is a 2-cycle, the sum is preserved under homotopy of $g$ \cite{IF}.
%Hence we may assume that $g:K\to M$ is a $\Z_2$-embedding, andNO!!! there is a disk $B\subset M$ such
%that $g(T)\subset B$ and $g(K-T)\subset M-\Int B$.
%Take a map $f:K\to B$ coinciding with $g$ on vertices.
%Then the required equality follows because
%$$g\widehat\sigma\cap g\widehat\tau = (g\sigma\cup f\sigma)\cap (g\tau\cup f\tau) =
%|g\sigma\cap g\tau|_2+|f\sigma\cap f\tau|_2 = |f\sigma\cap f\tau|_2.$$
The implication (RT)(some)$\Rightarrow$(R') follows because we can take $y_\sigma=\psi\widehat\sigma$.
In this paragraph we prove the implication (R')$\Rightarrow$(RH).
Define the homomorphism $\psi\colon H_k(K)\to H_k(M)$
by the formula $\displaystyle\psi[C] = \sum_{\sigma\in C}y_\sigma$.
Take a general position map $f\colon K\to\R^{2k}$.
We have $\omega(\psi) = [\omega(\overline\psi)] = [\nu(f)] = v(K)$.
Here the first and the third equalities are obvious.
The second equality holds because the scalar product $H^{2k}(K^*)\times H_{2k}(K^*)\to\Z_2$ is nondegenerate
and for any $2k$-cycle $C\subset K^*$ we have
$\displaystyle\sum_{\{\sigma, \tau\}\in C}(\omega(\psi) - \nu(f))(\sigma, \tau) = 0$.
% $(RI \Rightarrow R')$ follows from Proposition \ref{l:kunneth}.e.
In this paragraph we prove the implication (RI)$\Rightarrow$(RT)(any).
%(R').
%Take any maximal $k$-forest $T\subset K$.
Denote $y_\sigma = \psi\widehat\sigma$.
By Proposition \ref{l:kunneth}.e there are $k$-cycles $\alpha_j, \beta_j, j = 1, \ldots, q$
such that $\displaystyle C = \sum_{j = 1}^q [\alpha_j\times\beta_j]$.
Then
\begin{multline*}
\sum_{\{\sigma, \tau\}\in C} y_\sigma\cap_M y_\tau
\stackrel{(1)}= \sum_{j = 1}^q\sum_{\{\sigma, \tau\}\in[\alpha_j\times\beta_j]} y_\sigma\cap_M y_\tau
% = \sum_{j = 1}^q\sum_{\sigma\in\alpha_j, \tau\in\beta_j, \{\sigma, \tau\}\not\subset\alpha_j\cap\beta_j} y_\sigma\cap_M y_\tau
\stackrel{(2)}=\\\stackrel{(2)}= \sum_{j = 1}^q\sum_{\sigma\in\alpha_j, \tau\in\beta_j} y_\sigma\cap_M y_\tau
\stackrel{(3)}= \sum_{j = 1}^q\psi\alpha_j\cap_M\psi\beta_j \stackrel{(4)}= v_C(K).
\end{multline*}
Here
$\bullet$ $(1)$ follows by definition of $C$;
$\bullet$ $(2)$ follows because for every pair $\{\sigma, \tau\}\subset \alpha_j\cap\beta_j$
we have $y_\sigma\cap_M y_\tau + y_\tau\cap_M y_\sigma = 0$;
$\bullet$ $(3)$ follows by Proposition \ref{l:kunneth}.a;
$\bullet$ $(4)$ follows by (RI).
\end{proof}
%The equivalence $(R\Leftrightarrow R')$ also follows from Remark \ref{r:g}.c and Lemma \ref{l:reahom} (we take $y_\sigma=\psi\widehat\sigma$).
%The equivalence $(R\Leftrightarrow RI)$ also follows from Remark \ref{r:g}.d and Lemma \ref{l:reahom}.
%We do not spell out the equivalent condition $RH$ because we do not use it.
%Thus the implication $(E\Rightarrow C)$ of Theorem \ref{t:matrix} follows by the implication
%$(EH\Rightarrow RH)$.
\begin{remark}[integer versions]\label{r:inte}
In this remark $M$ is an oriented PL $2k$-manifold.
(a) Let $H^{2k}_s(\t K;\Z)$ be the group of cohomology classes of super-symmetric cocycles $\t K\to\Z$ (defined in \S\ref{ss:real}).
Analogously to the above one defines $v_{C,\Z}(K)\in\Z$, $\omega_{\Z}(\overline\psi)$ and
$\omega_{\Z}(\psi),v_{\Z}(K)\in H^{2k}_s(\t K;\Z)$ for a homomorphism $\psi:H_k(K;\Z)\to H_k(M;\Z)$.
(b) Integer analogues of the equivalences of Proposition \ref{r:crite} not involving (C) and (CM) hold.
E.g. the integer analogue of (E)$\Leftrightarrow$(R) is Theorem \ref{t:matrixz}; the integer analogue of (RH) is as follows:
{\it there is a homomorphism $\psi:H_k(K;\Z)\to H_k(M;\Z)$ such that $\omega_{\Z}(\psi)=v_{\Z}(K)$.}
(c) {\it The complex $K$ is realizable by the matrix $H_{g,\Z}$ and some collection of vectors $y_\sigma$
if and only if $K$ is compatible to a matrix $A$ such that $\rk A \leq 2g$.}
Cf. (CM)$\Leftrightarrow$(R) of Proposition \ref{r:crite}.
Let us deduce this statement from Lemmas \ref{l:gram} and \ref{l:Z_Hg}.
%Let $I\colon \Z^{2g}\times \Z^{2g}\to \Z$ be a bilinear form with .
Proof of the implication $(\Rightarrow)$ is obtained from
the proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main} for $I$ even
by replacing `even' with `skew-symmetric' and `$\rk A \leq \rk I$' with `$\rk A \leq 2g$'.
Let us prove the implication $(\Leftarrow)$.
By Lemma \ref{l:Z_Hg} the number $r:=\rk A$ is even and there is an $r\times n$ matrix $Y$ such that $A = Y^T H_{r/2} Y$.
Denote by $Y_1$ the $2g\times n$ matrix obtained from $Y$ by adding $2g - r$ zeroes below each column of $Y$.
Then $A = Y_1^TH_{g}Y_1$.
Denote by $(e_i)$ the basis in which $I$ has the matrix $H_{g,\Z}$.
Denote by $y_\sigma$ the vector whose coordinates in the basis $(e_i)$
form the corresponding column of matrix $Y_1$.
Then ($R_{f,H_{g,\Z}}$) holds.
(c) A map $g:K\to M$ \emph{induces} homomorphism $g_*:H_k(K;\Z)\to H_k(M;\Z)$ by the formula $g_*[\alpha]:=[g\alpha]$ for any $[\alpha]\in H_k(K;\Z)$.
It is simple and well-known that {\it any homomorphism $H_k(K;\Z)\to H_k(M;\Z)$ is induced by some map $g:K\to M$.}
Cf. Lemma \ref{l:reahom}.
(d) The analogue of Lemmas \ref{l:omega}.abc hold for $\nu(g),v(K)$ and $\omega(g)$ replaced by $\nu_{\Z}(g),v_{\Z}(K)$ and $\omega_{\Z}(g)$,
where $\omega_{\Z}(g)(\sigma,\tau):=(\overline\sigma\cup(-g'\sigma))\cdot(\overline\tau\cup(-g'\tau))\in\Z$.
\end{remark}
\begin{remark}\label{r:g}
Consider the problems of existence of a $\Z_2$- or $\Z$-embedding $K\to M$ homotopic to given map $K\to M$.
(a) {\it Assume that $k\ge3$ and $M$ is simply connected.
A map $K\to M$ is homotopic to an embedding if and only if the map is homotopic to a $\Z$-embedding.}
This is proved analogously to Theorem \ref{t:vkswu}.a.
The analogue for $k=2$ is incorrect by Theorem \ref{t:vkswu}.c.
The analogue for $k=1$ is correct analogously to the analogue for $k=1$ of Theorem \ref{t:vkswu}.a.
%The analogue for $k=1$ and $M$ not simply connected is incorrect (because there is a degree 3 $\Z$-embedding
%of $S^1$ to the annulus \cite[Example 6.3]{La70}; this example for a different problem was rediscovered
%in \cite[Example 1.6]{RS98}).
%(a) (examples) Denote by $S^1$ the cycle graph.
%$\bullet$ Any map from $S^1$ to a surface is homotopic to a $\Z_2$-embedding (but not necessarily to an embedding).
%$\bullet$ There is a map $S^1\sqcup S^1\to T^2$ to the torus $T^2$ not homotopic to an embedding and not homotopic to a $\Z_2$-embedding.
%RF: Regarding 1.6.6a the best reference I can provide is the Schaefer's survey, Section 1.4.2
(b) (algorithms) Let $T_M$ be a triangulation of $M$ and $g:K\to T_M$ be a simplicial map whose restriction to the $(k-1)$-skeleton of $K$ is null-homotopic.
Then there are polynomial algorithms for checking the existence
$\bullet$ of a $\Z_2$-embedding $K\to M$ homotopic to $g$;
$\bullet$ of a $\Z$-embedding $K\to M$ homotopic to $g$, if $M$ is oriented.
This follows from the criteria in (c,d).
The algorithms are polynomial because the obtained system of equations is linear.
%\begin{proposition}\label{p:su-vkg} (criteria)
(c) Let $g:K\to M$ a general position PL map whose restriction to the $(k-1)$-skeleton of $K$ is null-homotopic.
Then following are equivalent.
\quad(Eg) there is a $\Z_2$-embedding homotopic to $g$;
\quad(EHg) $\nu(g)$ is cohomologous to the zero cocycle;
%\quad(R'g) there is a collection of homology classes $y_\sigma\in H_k(M)$ such that
%for any $\alpha\in H_k(K)$ we have $g_*[\alpha]=\sum\limits_{\sigma\in\alpha} y_\sigma$ and
%for any $2k$-cycle $C\subset K^*$ we have $\sum\limits_{ \{\sigma,\tau\} \in C} y_\sigma\cap_M y_\tau = v_C(K)$.
\quad(RTg) for any $2k$-cycle $C\subset K^*$ and for some (equivalently, for any) maximal $k$-forest in $K$ we have
$\sum\limits_{ \{\sigma,\tau\} \in C} g_*\widehat\sigma\cap_M g_*\widehat\tau = v_C(K)$.
\quad(RIg) for any $k$-cycles $\alpha_1,\ldots,\alpha_q,\beta_1,\ldots,\beta_q$ in $K$ such that
$C:=\sum\limits_{j=1}^q[\alpha_j\times\beta_j]$ is a $2k$-cycle in $K^*$ we have
$\sum\limits_{j=1}^q g_*\alpha_j\cap_M g_*\beta_j=v_C(K)$.
\quad(RHg) $\omega(g_*)=v(K)$.
This (even with more equivalent conditions) is proved analogously to Proposition \ref{r:crite}.
%The equivalence $(\nu g\Leftrightarrow RHg)$ follows by Lemma \ref{l:omega}.c.
%The implication $(Eg\Rightarrow RIg)$ is clear.
%The implication $(RIg\Rightarrow RHg)$ follows analogously to the implication $(RI\Rightarrow RH)$.
%from (d) for $k=1$ and Proposition \ref{l:kunneth}.d
(d) Suppose that $M$ is oriented.
Let $g:K\to M$ a general position PL map whose restriction to the $(k-1)$-skeleton of $K$ is null-homotopic.
Then there is a $\Z$-embedding $K\to M$ homotopic to $g$ if and only if $\nu_{\Z}(g)$
is cohomologous to the zero cocycle.
This is the particular case of Lemma \ref{l:zstar} for $\nu_{\Z}=0$.
Other criteria (c) have also their integer analogues.
\end{remark}
}
%\newpage
%\renewcommand{\refname}{References}
\begin{thebibliography}{RSS95}
\input{references}
\bibitem[Al]{Al} \emph{A. Adrian Albert}. Symmetric and alternate matrices in an arbitrary field, I. Trans. Amer.
Math. Soc., (1938) 43(3):386--436.
\amsw
%\bibitem[BF09]{BF09} \emph{K. Barnett, M. Farber}. Topology of Configuration Space of Two Particles
%on a Graph, I. Algebr. Geom. Topol. 9 (2009) 593--624. arXiv:0903.2180.
%\beet
\bito
%\brst
\bibitem[Bo81]{Bo81} * \emph{N. Bourbaki}. Elements of mathematics. Algebra.
%\ckv
%\clreng
%\bibitem[CKP+]{CKP+} E. Colin de Verdi\' ere, V. Kaluza, P. Pat\' ak, Z. Pat\' akov\' a and M. Tancer.
%A direct proof of the strong Hanani-Tutte theorem on the projective plane. Journal of Graph Algorithms and Applications, 21:5 (2017) 939--981.
\bibitem[FH10]{FH10} \emph{M. Farber, E. Hanbury}. Topology of Configuration Space of Two Particles on a Graph, II. Algebr. Geom. Topol. 10 (2010) 2203--2227. arXiv:0903.2180.
%\fkosc
\fkon
\fktnf
%\fozf
%\fpstz
%\ga
\hasn
\info
\irsf
\jozt
%\kkrot
\koto
\krzz
\ksto
\journal{\kstoe}
\bibitem[La70]{La70} \emph{L.L. Larmore}, The first obstruction to embedding a 1-complex in a 2-manifold,
Illinois J. Math. 14 (1) (1970) 1-11.
%\loot
\bibitem[MC]{MC} \url{https://en.wikipedia.org/wiki/Matrix_completion}
%\mazt
\mezs
%\moss
%\moen
\bibitem[MS17]{MS17} \emph{T. Maciazek, A. Sawicki.} Homology groups for particles on one-connected graphs
J. Math. Phys. 58, 062103 (2017).
\mtwoo
\bibitem[MW69]{MW69} * \emph{J. MacWilliams}. Orthogonal matrices over finite fields. Amer. Math. Monthly, (1969) 76:152--164.
\nkon
%\przse
%\bibitem[PSS]{PSS} \emph{M. J. Pelsmajer, M. Schaefer, D. Stasi}. Strong Hanani-Tutte on the projective
%plane. SIAM J. Discrete Math., 23:3 (2009) 1317â1323.
%\bibitem[Pa71]{Pa71} C.W. Patty, Homology of the deleted product spaces, Math. Scand. 29:2 (1971) 211--231.
\pton
%\rszf
\sano
%\sczf
\scot
\shfs
\skde
\skze
\sks
\skoe
%\skoeo
\skotz
\ssot
%\sktzd
%\stos
%\ta
%\vizteng
\vktt
\walst
%\winum
\wufe
\end{thebibliography}
{\bf Books, surveys and expository papers in this list are marked by the stars.}
\end{document}
(Please do not read the rest of this subsection yet.)
(historical remark) (please do not read this remark yet)
%although the statements there are different and not topologically invariant.
%essentially proved in unpublished paper simplified accessible to non-specialists versions of
%the cohomology obstruction to the embeddability of $k$-complexes into $2k$-manifolds
%of the mod2 version of the construction
%The user-friendly reformulation of the integer version is analogous.
%These results and constructions generalize results \cite{SS13} and \cite{FK19}
%from graph theory known before \cite{PT19}
%The higher-dimensional generalization is not hard.
%Higher-dimensional generalizations seem to be easy modulo known results and methods concerning
%higher-dimensional manifolds and their (co)homology.
%Neglect of simple important particular cases obstructs proper description of relation to previous work,
%and obstructs finding simple user-friendly formulations of main results.
%For a 2-cycle $C\subset K^*$ and a general position PL map $g:K\to M$ define the {\it $C$-van Kampen} %$v_{2,C}(g)\in\Z_2$ analogously to Remark \ref{vector}.c.
%Analogously to Lemma \ref{ratv-vk2} $v_{2,C}(g)$ is independent of the choice of $g$ within a homotopy class.
%\begin{lemma}\label{l:inv} For any connected graph $K$, surface $M$, map $g:K\to M$ and 2-cycle $C\subset K^*$
%we have $v_C(g)=\omega_C(g)+v_C(f)$.
%\end{lemma}
%Perhaps one can prove directly that the cohomology class of $\nu(g)$ depends only on the induced
%homomorphism $g_*:H_k(K)\to H_k(M)$.
%We proved this via Lemma \ref{l:omega}.c calculating $\nu(g)$ in terms of $\omega(g_*)$.
%This calculation yields criteria not involving $\omega(g_*)$: Remarks (d,e).
% {\bf Disclaimers.}
%(c) For simplicity, we call a `theorem' a result stated as a theorem in an arXiv paper or in a computer
%science conference, even if it is not published in a refereed journal.
%This is imperfect (and even publication in a refereed journal is not a guarantee), see a discussion
%in \cite[Remark 1.5]{Sk21d}.
%However, we did check or reinvent proofs of all the results below except Theorems \ref{t:hts}.ab
%This does not mean that we see any problems with these theorems.
%Rather we did not read available proofs because other results exposed in this survey are
%formally independent of these theorems.
%The map $\pi$ to the plane of
%\includegraphics[scale=.9]{figures-konovalenko/Scop.10.eps}
Generalized Proposition 8 for *arbitrary K*, not just for the skeleton.
You should state it.
(Currently you have undefined K in the proof.)
Then you would need that M or K is (k-1)-connected because in the proof you apply C3 to M.
(which is a nicer version of (H); a weaker nicer version of (H) is that f induces zero on \pi_j for j<k.)
Also you need embeddings into `almost closed' manifold M-(a ball).
Thus Proposition 8 should be compared with classification of `almost closed' manifolds:
https://www.maths.ed.ac.uk/~v1ranick/papers/n-1con2n.pdf, p. 170.
Pre-H-spaces and H-spaces are defined in p. 168 and 169.
The classification should be stated and it should be explained why Proposition 8 is not covered by it (if it is not).
In the proof of P3 PT need to evaluate Q on a symmetric 2k-cycle C in K x K.
By K"unneth formula
C = \sum_j ( \alpha_j\times\beta_j + \beta_j\times\alpha_j ) for
some $k$-cycles $\alpha_j,\beta_j$ in $K$.
Define Q(C):=\sum_j Q(\alpha_j,\beta_j ).
This is well-defined, i.e. is independent of the above representation of C, because
Hom ( H_k(K) \otimes H_k(K) ; \Z_2 ) = Hom ( H_{2k}(K) ; \Z_2 ) = H^{2k}(K) .
Here the first isomorphism is K"unneth formula and the other two are obvious.
Your paper contains a less abstract (but non-invariant) interpretation of this argument.
Later:
Sorry, you do not need this.
The sum is independent of the choise of v because the sum is 1 (mod2) by the hypothesis of T3.
Please also look at Lemma 2 of https://arxiv.org/pdf/1808.08363.pdf
Your L2 is a finer version of https://arxiv.org/pdf/1704.00300.pdf, L3.b for d even:
you prove the formula analogous to L3.b for d even but without \sum_A.
So you might want to check references [Ho], [Me11, Theorem 6.5], [Me06, Example 4.7].
The abstract of FK19
"An unpublished result by Robertson and Seymour implies that for every t, every graph of sufficiently large
genus contains as a minor a projective t à t grid or one of the following so-called t-Kuratowski graphs:
K3,t, or t copies of K5 or K3,3 sharing at most 2 common vertices. We show that the Z2-genus of graphs in
these families is unbounded in t; in fact, equal to their genus. Together, this implies that the genus
of a graph is bounded from above by a function of its Z2-genus"
and the discussion below the statement of Problem 1 on page 2.
\begin{example}[cf. Remark \ref{vector}.c]\label{vkamz-ex}
Order the edges of $K_5$ lexicographically, and orient the edge $ij$ from $i$ to $j$, where $i<j$.
Take a linear map $f:K_5\to\R^2$ such that $f(1)f(2)f(3)f(4)f(5)$ is a convex pentagon.
Then the integral intersection cocycle is
$$
\left(
\begin{array}{cccccccccc}
\text{-} & \text{-} & \text{-} & \text{-} & \text{-} & \text{-} & \text{-} & 0 & 0 & 0 \\
\text{-} & \text{-} & \text{-} & \text{-} & \text{-} & -1 & -1 & \text{-} & \text{-} & 0 \\
\text{-} & \text{-} & \text{-} & \text{-} & 0 & \text{-} & -1 & \text{-} & -1 & \text{-} \\
\text{-} & \text{-} & \text{-} & \text{-} & 0 & 0 & \text{-} & 0 & \text{-} & \text{-} \\
\text{-} & \text{-} & 0 & 0 & \text{-} & \text{-} & \text{-} & \text{-} & \text{-} & 0 \\
\text{-} & 1 & \text{-} & 0 & \text{-} & \text{-} & \text{-} & \text{-} & -1 & \text{-} \\
\text{-} & 1 & 1 & \text{-} & \text{-} & \text{-} & \text{-} & 0 & \text{-} & \text{-} \\
0 & \text{-} & \text{-} & 0 & \text{-} & \text{-} & 0 & \text{-} & \text{-} & \text{-} \\
0 & \text{-} & 1 & \text{-} & \text{-} & 1 & \text{-} & \text{-} & \text{-} & \text{-} \\
0 & 0 & \text{-} & \text{-} & 0 & \text{-} & \text{-} & \text{-} & \text{-} & \text{-} \\
\end{array}
\right).
$$
\end{example}
(CVH) there is an even symmetric $n\times n$-matrix $A$ with $\Z_2$-entries such that
$\rk A\le\rk\cap_M$ and $A|_{K^*}$ is cohomologous to $\nu(f)$ for some (or, equivalently, for any)
general position PL map $f:K\to\R^{2k}$;
(RV) for some (or, equivalently, for any) basis in $H_k(M)$ there is an $n\times\rk H_k(M)$-matrix $Y$ with $\Z_2$-entries such that for the matrix $I$ of $\cap_M$ in this basis and some general position PL map $f:K\to\R^{2k}$, we have $(Y^TIY)|_{K^*}=\nu(f)$ and $Y^TIY$ is even.
(RHV) is obtained from (RH) by adding `$Y^TIY$ is even'.
The property (RV) is the matrix form of (R) ($Y$ is the matrix formed by the columns $y_\sigma$ corresponding to $k$-faces of $K$) with `$Y^TIY$ is even' added.
So (CV$\Leftrightarrow$ RV) by Lemma \ref{l:main}, and ($RHV\Leftrightarrow RV$) by Lemma \ref{star}.
The implication $(EV\Rightarrow RHV)$ is proved analogously to the implication $(EH\Rightarrow RH)$ above.
%{\bf For the additivity.}
%We have $h_*\widehat\sigma=y_\sigma$ (after modification of $h$!).
Since $M$ is $(k-1)$-connected, by making a homotopy we may additionally assume that $h'(T)$ is contained in a small neighborhood of $\partial B$.
The obtained matrix $Y^TIY$ is even because for each $k$-face $\sigma\subset K-T$ we have
$$y_\sigma^TIy_\sigma \overset{(1)}= [f'\sigma\cup h'\sigma]^2_M \overset{(2)}= (h'_*\widehat\sigma)^2_M \overset{(3)}= 0.$$
Here
$\bullet$ (1) is proved in the proof of the implication $(EH\Rightarrow RH)$,
$\bullet$ (2) holds by Addendum, and
$\bullet$ (3) holds because $h$ is even, so $h'$ is even.
%%% Local Variables:
%%% mode: latex
%%% TeX-master: t
%%% End:
% SKOPENKOV:
% TO CHECK:
% TO ANSWER:
% Algorithms -> Remark 1.1?
% banded (2.3.5 CM_2) - Dima Trushin? Konyaev?
% r_2 for even/odd/closed
% 2.1.b -- unimodular???? (1.3.2c)
% TO EDIT:
% r:critsg f write (???)
% r:inte e prove?
% Fulek remarks
%! Kunneth Formula + 2.5.3
%! (R'g, RTg) -- conjecture or write sketch of proof if obvious
% \begin{proof}
% Let us prove that (a) follows from (b). Let $Z := \{(\sigma, \tau)\mid\{\sigma, \tau\}\in C\}$.
% By (b) there are $k$-cycles $\alpha_j, \beta_j, j = 1, \ldots, q$ such that
% $\displaystyle Z = \sum_{j = 1}^q(\alpha_j\times\beta_j + \beta_j\times\alpha_j)$.
% Then $C = \sum_{j = 1}^q[\alpha_j\times\beta_j]$.
% (b) Let $T$ be a maximal forest of $K$.
% Denote $\overline T := T\times K \cup K\times T$.
% In the following two paragraphs we prove that $\overline T$ is a maximal forest of $K\times K$.
% Let us prove that $H_{2k}(\overline T) = 0$.
% Let $C\subset \overline T$ be any $2k$-cycle.
% For any $k$-face $\sigma$ of $K$
% the set $C_\sigma := \{\tau\mid(\tau, \sigma)\in C\}$ is a $k$-cycle in $K$.
% For each $\sigma\in K - T$ we have $C_\sigma\subset T$ and thus we have $C_\sigma = \emptyset$.
% Hence $C\subseteq T\times T$. By K\"unneth formula we have $H_{2k}(T\times T) = H_k(T)\otimes H_k(T) = 0$.
% Thus $H_{2k}(C) = 0$ as well.
% % The subcomplex $\overline T$ is maximal because
% For any pair $\sigma\times\tau\in K\times K$
% we have a nontrivial $2k$-cycle
% $\widehat\sigma\times\widehat\tau\subseteq \overline T\cup\{\sigma\times\tau\}$.
% Thus, $\overline T$ is a maximal forest.
% Obviously,
% $\widehat{\sigma\times\tau} = \widehat\sigma\times\widehat\tau$.
% Then the following equality
% \begin{equation*}
% Z = \sum_{\sigma\times\tau\in Z}\widehat{\sigma\times\tau}
% = \sum_{\sigma\times\tau\in Z}\widehat\sigma\times\widehat\tau
% \end{equation*}
% gives the required representation of $Z$.
% \end{proof}
% \begin{proof}
% % Denote by $C^K$ the $\Z_2$-vector space of $k$-cocycles of $K$.
% Denote by $C_K$ the $\Z_2$-vector space of $k$-cycles of $K$.
% Let $z\colon C_K\times C_K\to\Z_2$ be the bilinear form
% defined by
% \begin{equation*}
% z(\alpha, \beta) := |\{(\sigma, \tau)\in Z\mid \sigma\in\alpha, \tau\in\beta\}|_2
% \end{equation*}
% Let us prove that $z$ is even. Indeed,
% \begin{equation*}
% z(\alpha, \alpha) \stackrel{(1)}= |\{(\sigma, \sigma)\in Z\mid \sigma\in\alpha\}|_2
% \stackrel{(2)}= 0,
% \end{equation*}
% where $(1)$ follows because $Z$ is symmetric, and
% $(2)$ follows because $(\sigma, \sigma)\notin\widetilde K\supseteq Z$.
% Hence by the??? classification theorem in some basis
% $(\alpha_1, \beta_1, \alpha_2, \beta_2, \ldots, \alpha_q, \beta_q, \gamma_1, \ldots, \gamma_{n - 2q})$
% of $C_K$
% the bilinear form $z$ has the matrix $
% \begin{pmatrix}
% H_q&0\\
% 0&0
% \end{pmatrix}$.
% Then??? $\displaystyle Z = \sum_{j = 1}^q (\alpha_j\times\beta_j + \beta_j\times\alpha_j)$
% as required.
% \end{proof}
Let $V$ be a $\Z_2$-vector space.
The complex $K$ is called {\bf realizable modulo 2
by a bilinear form $I\colon V\times V\to \Z_2$
and a collection of vectors $y_\sigma$ indexed by $k$-faces of $K$}
if there is a general position PL map $f:K\to\R^{2k}$
such that
$$(R_{f,I})\quad I(y_\sigma, y_\tau) = |f\sigma \cap f\tau|_2 \quad\text{for any non-adjacent $k$-faces $\sigma, \tau$ of $K$}.$$
\begin{lemma}\label{l:main}
Let $I\colon V\times V\to \Z_2$ be an even/odd symmetric bilinear form on a $\Z_2$-vector space $V$.
The complex $K$ is realizable modulo 2 by $I$ and some collection of vectors $y_\sigma$ if and only if $K$ is compatible modulo 2 to an even/odd matrix $A$ such that $\rk A\le \rk I$.
\end{lemma}
Lemma \ref{l:main} implicitly appeared in \cite[\S2]{Bi21}.
\talkno{
Lemma \ref{l:main} is easily deduced below from well-known Lemmas \ref{l:gram}, \ref{l:midrk},
known Lemma \ref{l:matrix_HgI} (see proof e.g. in \cite[Theorem 3]{Al} and \cite[Theorem 1]{MW69})
and simple Lemma \ref{l:alternate} which appeared in a discussion with A. Bikeev and which is proved in \cite[\S2]{Bi21}.
\begin{lemma}\label{l:gram}
Let $F$ be $\Z_2$ or $\Q$.
Let $I\colon V\times V\to F$ be a symmetric bilinear form on an $F$-vector space $V$.
% Then the rank of the Gramian matrix of any vectors $v_1,\ldots, v_n\in V$ does not exceed $\rk I$.
If $A$ is a Gramian matrix with respect to $I$, then $\rk A \leq \rk I$.
\end{lemma}
\begin{proof}[Proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main} for $I$ even]
Denote by $A$ the Gramian matrix of the vectors $y_\sigma$ with respect to $I$.
Then $A$ is even and $(C_{2,f})$ holds.
By Lemma \ref{l:gram} we have $\rk A \leq \rk I$.
\end{proof}
\begin{lemma}\label{l:alternate}
Let $I\colon V\times V\to \Z_2$ be an odd symmetric bilinear form on a $\Z_2$-vector space $V$.
If $A$ is an even Gramian matrix with respect to $I$, then $\rk A \leq \rk I - 1$.
\end{lemma}
\begin{proof}[Proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main} for an $I$ odd]
Denote by $A$ the Gramian matrix of the vectors $y_\sigma$ with respect to $I$.
If $A$ is odd, then by Lemma \ref{l:gram} $\rk A\le \rk I$ and $(C_{2,f})$ holds.
If $A$ is even, then by Lemma \ref{l:alternate} $\rk A \leq \rk I-1$.
Take the matrix $A'$ obtained from $A$ by replacing the entry $A_{1,1}$ by $1$.
Then $\rk A'\leq \rk I$, $A'$ is odd and $(C_{2,f})$ holds.
\end{proof}
\begin{lemma}\label{l:midrk}
Let $A, B$ be $n\times m$ and $m\times k$ matrices with entries in $\Z_2$, respectively.
Then $\rk AB \leq \rk B$. In particular, $\rk AB \leq m$.
\end{lemma}
Denote by $H_{2}$ the $2g\times2g$-matrix with $\Z_2$-entries formed by $g$ diagonal blocks
$\begin{pmatrix}0&1\\1&0\end{pmatrix}$ and zeros elsewhere.
Recall that the rank of an even symmetric matrix with entries in $\Z_2$ is even.
\begin{lemma}\label{l:matrix_HgI}
Let $A$ be a symmetric $a\times a$-matrix with $\Z_2$-entries.
Let $H$ be the matrix $H_{\frac{\rk A}{2}}$ if $A$ is even
and the $\rk A\times\rk A$ identity matrix otherwise.
There is a $\rk A\times a$-matrix $Y$ such that $A = Y^T HY$.
\end{lemma}
\begin{proof}[Proof of the implication $(\Leftarrow)$ of Lemma \ref{l:main}]
Denote by $I$ the matrix of the bilinear form $I$ in some basis $(e_i)$ of $V$.
Denote $m := \dim V$.
For $M = A, I$ let $H_M$ be the matrix $H_{\frac{\rk M}2}$ if $M$ is even
and the $\rk M\times \rk M$ identity matrix if $M$ is odd.
By Lemma \ref{l:matrix_HgI} for each $M = A, I$ there is a matrix $Y_M$ such that $M = Y_M^T H_MY_M$.
By Lemma \ref{l:midrk} we have $\rk Y_I \geq \rk H_IY_I \geq \rk I$.
Hence all the $\rk I$ rows of $Y_I$ are linearly independent.
Thus there is a nondegenerate $m\times m$-matrix $Y'_I$ whose first $\rk I$ rows are the rows of $Y_I$.
Denote by $H'_I$ the $m\times m$-matrix $\begin{pmatrix}H_I&0\\0&0\end{pmatrix}$. Then $I = (Y'_I)^T H'_I Y'_I$.
Denote by $Y$ the $m\times n$-matrix obtained from $Y_A$ by adding $m - \rk A$ zeroes below each column of $Y_A$.
Then $A = Y^TH'_IY = ((Y'_I)^{-1}Y)^T I ((Y'_I)^{-1}Y)$.
Denote by $y_\sigma\in V$ the vector represented in the basis $(e_i)$ by the corresponding column of the matrix $(Y'_I)^{-1}Y$.
Then $(R_{f,I})$ holds.
\end{proof}
}
% The proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main}.b is obtained from
% the proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main}.a for an even matrix $I$
% by replacing ``even'' with ``skew-symmetric'' and $\rk A \leq \rk I$ with $\rk_\Q A \leq 2g$.
% The proof of the implication $(\Leftarrow)$ of Lemma \ref{l:main}.b is analogous to
% the proof of the implication $(\Leftarrow)$ of Lemma \ref{l:main}.a.
Denote by $H_{g,\Z}$ the $2g\times2g$-matrix with $\Z$-entries formed by $g$ diagonal blocks
$\begin{pmatrix}0&1\\-1&0\end{pmatrix}$ and zeros elsewhere.
\begin{lemma}\label{l:Z_Hg} For every skew-symmetric $a\times a$-matrix $A$ with $\Z$-entries its rank $r$ is even
and there is an $r\times a$ matrix $Y$ with $\Z$-entries such that $A = Y^T H_{r/2,\Z} Y$.
\end{lemma}
Lemma \ref{l:Z_Hg} follows from \cite[Chapter IX, \S5, Theorem 1]{Bo81}.
% {\bf Comment.} Let $I\colon \Z^{2g}\times \Z^{2g}\to \Z$ be a bilinear form with the matrix $H_{g,\Z}$.
% The complex $K$ is realizable by $I$ and some collection of vectors $y_\sigma$
% if and only if $K$ is compatible to a matrix $A$ such that $\rk A \leq 2g$.
% Let us deduce this statement from Lemmas \ref{l:gram} and \ref{l:Z_Hg}.
% The implication $(\Rightarrow)$ is obtained from
% the proof of the implication $(\Rightarrow)$ of Lemma \ref{l:main}.a for $I$ even
% by replacing ``even'' with ``skew-symmetric'' and $\rk A \leq \rk I$ with $\rk A \leq 2g$.
% $(\Leftarrow)$
% By Lemma \ref{l:Z_Hg} the number $r:=\rk A$ is even and there is an $r\times n$ matrix $Y$ such that $A = Y^T H_{r/2,\Z} Y$.
% Denote by $Y_1$ the $2g\times n$ matrix obtained from $Y$ by adding $2g - r$ zeroes below each column of $Y$.
% Then $A = Y_1^TH_{g,\Z}Y_1$.
% Denote by $(e_i)$ the basis in which $I$ has the matrix $H_{g,\Z}$.
% Denote by $y_\sigma$ the vector whose coordinates in the basis $(e_i)$
% form the corresponding column of matrix $Y_1$.
% Then $(R_{f,I})$ holds.
\talkno{
\begin{lemma}\label{l:rank1} If a symmetric $n\times n$ matrix $A$ with $\Z$-entries has rank 1
and any diagonal element of $A$ is the square of an integer\arxiv{\footnote{The last requirement can be replaced with ``some nonzero diagonal element of $A$ is the square of an integer''.}}, then $A=b^T b$ for some vector $b\in\Z^n$\journal{ (see \cite[footnote 5]{KS21e})}.
\end{lemma}
\begin{proof}
Since $A$ is symmetric and has rank 1, there are a nonzero vector $u\in\Q^n$ and a number $q\in\Q$
such that $A = qu^Tu$. Then for each $i = 1, \ldots, n$ we have $qu_i^2 = A_{i,i}$.
Since $u\neq 0$, $q$ is the square of a rational number.
Hence for each $i=1, \ldots, n$ the number $\sqrt qu_i$ is an integer.
Then let $b = \sqrt qu$.
\end{proof}
}
% (or, equivalently, either $A=0$ or some non-zero diagonal element is a square),
% {\bf Comment.}
% From Lemmas \ref{l:main} and \ref{l:matrix_HgI} it is clear
% that if a complex $K$ is realizable by an even / odd nondegenerate $m\times m$ matrix $I$,
% then $K$ is realizable by $H_{2,m/2}$ / the $m\times m$ identity matrix.
% This makes parts (a) and (b) of Lemma \ref{l:main} analogous.
%%% Local Variables:
%%% mode: latex
%%% TeX-master: "invasurf"
%%% End:
%uncomment for arxiv
\UseRawInputEncoding
\newcommand{\abc}{\bibitem[ABC+]{ABC+} * \emph{M. Atiyah, A. Borel, G. J. Chaitin, D. Friedan, J. Glimm, J. J. Gray, M. W. Hirsch, S. MacLane, B. B. Mandelbrot, D. Ruelle, A. Schwarz, K. Uhlenbeck, R. Thom, E. Witten, C. Zeeman.} Responses to ``Theoretical Mathematics: Toward a cultural synthesis of mathematics and theoretical physics'', by A. Jaffe and F. Quinn. Bull. Am. Math. Soc. 30 (1994) 178--207. arXiv:math/9404229.}
\newcommand{\agles}{\bibitem[AGL]{AGL86} Mathematical Economics, ed. by A. Ambrosetti, F. Gori, R. Lucchetti,
Lect. Notes Math. 1330, Springer, 1986.}
%Lectures given at the 2nd 1986 Session of the Centro ...
\newcommand{\akzz}{\bibitem[Ak00]{Ak00} * \emph{Ï. Ì. Àõìåòüåâ.} Âëîæåíèÿ êîìïàêòîâ, ñòàáèëüíûå
ãîìîòîïè÷åñêèå ãðóïïû ñôåð è òåîðèÿ îñîáåííîñòåé, Óñïåõè Ìàò. Íàóê. 2000. 55:3. C.~3-62.}
\newcommand{\akoe}{\bibitem[AK19]{AK19} \emph{S. Avvakumov, R. Karasev.} Envy-free division using mapping degree. arXiv:1907.11183.}
\newcommand{\akto}{\bibitem[AK21]{AK21} \emph{G. Arone and V. Krushkal.}
Embedding obstructions in $\R^d$ from the Goodwillie-Weiss calculus and Whitney disks. arXiv:2101.10995. }
\newcommand{\akm}{\bibitem[AKM]{AKM} \emph{M. Abrahamsen, L. Kleist and T. Miltzow.}
Geometric Embeddability of Complexes is $\exists\mathbb R$-complete, arXiv:2108.02585.}
\newcommand{\aksoe}{\bibitem[AKS]{AKS} \emph{S. Avvakumov, R. Karasev and A. Skopenkov.} Stronger counterexamples to the topological Tverberg conjecture, submitted. arXiv:1908.08731.}
\newcommand{\akuoe}{\bibitem[AKu19]{AKu19} \emph{S. Avvakumov, S. Kudrya.}
Vanishing of all equivariant obstructions and the mapping degree. arXiv:1910.12628.}
\newcommand{\alto}{\bibitem[Al21]{Al21} \emph{E. Alkin,}
Hardness of almost embedding simplicial complexes in $\R^d$, II.}
\newcommand{\amsw}{\bibitem[AMS+]{AMSW} \emph{S. Avvakumov, I. Mabillard, A. Skopenkov and U. Wagner.}
Eliminating Higher-Multiplicity Intersections, III. Codimension 2, Israel J. Math. (2021). arxiv:1511.03501.}
%\url{http://arxiv.org/abs/1511.03501}
%DOI: https://doi.org/10.1007/s11856-021-2216-z.
\newcommand{\anzt}{\bibitem[An03]{An03} * \emph{Ä. Â. Àíîñîâ.} Îòîáðàæåíèÿ îêðóæíîñòè, âåêòîðíûå ïîëÿ è èõ ïðèìåíåíèÿ. Ì: ÌÖÍÌÎ, 2003.}
\newcommand{\arnf}{\bibitem[Ar95]{Ar95} \emph{V. I. Arnold,} Topological invariants of plane curves and caustics, University Lecture Series, Vol. 5, Amer. Math. Soc., Providence, RI, 1995.}
\newcommand{\arszo}{\bibitem[ARS01]{ARS01} \emph{P. Akhmetiev, D. Repov\v s and A. Skopenkov},
Embedding products of low-dimensional manifolds in $\R^m$, Topol. Appl. 113 (2001), 7--12.}
\newcommand{\arszt}{\bibitem[ARS02]{ARS02} \emph{P. Akhmetiev, D. Repovs and A. Skopenkov.} Obstructions to approximating maps of $n$-manifolds into $R^{2n}$ by embeddings, Topol. Appl., 123 (2002), 3--14.}
\newcommand{\asoed}{\bibitem[As]{As} \emph{A. Asanau,} \lowercase{A SIMPLE PROOF THAT CONNECTED SUM OF ORDERED
ORIENTED LINKS IS NOT WELL-DEFINED,} Math. Notes, to appear.}
\newcommand{\asoe}{\bibitem[As]{As} \emph{A. Asanau,} On the \lowercase{TRIPLE SELF-INTERSECTION NUMBER FOR GRAPHS IN THE PLANE,} unpublished, 2018.}
\newcommand{\bant}{\bibitem[Ba93]{Ba93} * \emph{T. Bartsch.} Topological methods for variational problems with
symmetries, Lecture Notes in Mathematics, 1560, Springer-Verlag, Berlin, 1993.}
\newcommand{\bbsn}{\bibitem[BB79]{BB} \emph{E.~G. Bajm{{\'o}}czy and I.~B{{\'a}}r{{\'a}}ny,}
\newblock On a common generalization of {B}orsuk's and {R}adon's theorem,
\newblock Acta Math.\ Acad.\ Sci.\ Hungar.\ 34:3 (1979), 347-350.}
\newcommand{\bbzos}{\bibitem[BBZ]{BBZ} * \emph{I.~B{{\'a}}r{{\'a}}ny, P.~V.~M. Blagojevi{{\'c}} and G.~M. Ziegler.} Tverberg's Theorem at 50: Extensions and Counterexamples, Notices of the Amer. Math. Soc., 63:7 (2016), 732--739.}
%\url{http://www.ams.org/journals/notices/201607}}
\newcommand{\bcm}{\bibitem[BCM]{BCM} * 13th Hilbert Problem on superpositions of functions, presented by A. Belov, A. Chilikov, I. Mitrofanov, S. Shaposhnikov and A. Skopenkov,
%\linebreak
\url{http://www.turgor.ru/lktg/2016/5/index.htm}.}
\newcommand{\beet}{\bibitem[BE82]{BE82} * \emph{V.G. Boltyansky and V.A. Efremovich.} Intuitive Combinatorial Topology. Springer.}
\newcommand{\beetr}{\bibitem[BE82]{BE82} * \emph{Â. Ã. Áîëòÿíñêèé è Â. À. Åôðåìîâè÷.} Íàãëÿäíàÿ òîïîëîãèÿ. Ì.: Íàóêà, 1982.}
%\url{http://ilib.mccme.ru/djvu/geometry/boltiansky-nagl-topo.htm}.}
\newcommand{\bfzof}{\bibitem[BFZ14]{BFZ14} \emph{P. V. M. Blagojevi{\'c}, F. Frick, and G. M. Ziegler,}
Tverberg plus constraints, Bull. Lond. Math. Soc. 46:5 (2014), 953-967, arXiv:1401.0690.}
%\url{http://arxiv.org/abs/1401.0690}
\newcommand{\bfzos}{\bibitem[BFZ]{BFZ} \emph{P. V. M. Blagojevi{\'c}, F. Frick and G. M. Ziegler,}
Barycenters of Polytope Skeleta and Counterexamples to the Topological Tverberg Conjecture, via Constraints,
J. Eur. Math. Soc., 21:7 (2019) 2107-2116. arXiv:1510.07984.}
%MR3959859
%\linebreak\url{http://arxiv.org/abs/1510.07984}
\newcommand{\bgos}{\bibitem[BG16]{BG16} \emph{A. Bj\"orner and A. Goodarzi}, On Codimension one Embedding of Simplicial Complexes, in book: A Journey Through Discrete Mathematics, arXiv:1605.01240.}
\newcommand{\biet}{\bibitem[Bi83]{Bi83} * \emph{R. H. Bing.} The Geometric Topology of 3-Manifolds. Providence, R.~I. 1983. (Amer. Math. Soc. Colloq. Publ., 40).}
\newcommand{\bitz}{\bibitem[Bi20]{Bi20} \emph{A. Bikeev.} Realizability of discs with ribbons on the M\"obius strip. arXiv:2010.15833.}
%\emph{À. Áèêååâ.} Ðåàëèçóåìîñòü äèñêîâ ñ ëåíòî÷êàìè íà ëåíòå Ìåáèóñà.
%Ïðåïðèíò. Ñð. \url{https://conf62.mipt.ru/conference/6923}
\newcommand{\bito}{\bibitem[Bi21]{Bi21} {\it A. I. Bikeev,}
Criteria for integer and modulo 2 embeddability of graphs to surfaces, arXiv:2012.12070v2.}
%Towards a short proof of the Fulek--Kyn\v cl criterion for modulo 2 embeddability of graphs to surfaces,
\newcommand{\bagos}{\bibitem[BG17]{BG17} \emph{S. Basu and S. Ghosh.} Equivariant maps related to the topological Tverberg conjecture, Homology, Homotopy and Applications 19:1 (2017) 155--170.}
\newcommand{\bkkmzof}{\bibitem[BKK]{BKK} \emph{M. Bestvina, M. Kapovich and B. Kleiner,}
Van Kampen's embedding obstruction for discrete groups, Invent. Math. 150 (2002) 219--235. arXiv:math/0010141.}
\newcommand{\bmzf}{\bibitem[BM04]{BM04} \emph{Boyer, J. M. and Myrvold, W. J.} On the cutting edge: simplified $O(n)$ planarity by edge addition, Journal of Graph Algorithms and Applications, 8:3 (2004) 241--273.}
\newcommand{\bm}{\bibitem[BM15]{BM15} \emph{I. Bogdanov and A. Matushkin.} Algebraic proofs of linear versions of the Conway--Gordon--Sachs theorem and the van Kampen--Flores theorem, arXiv:1508.03185.}
%\emph{È. Áîãäàíîâ è À. Ìàòóøêèí,} Àëãåáðàè÷åñêîå äîêàçàòåëüñòâî ëèíåéíûõ òåîðåì Êîíâåÿ-Ãîðäîíà-Çàêñà
%è âàí Êàìïåíà-Ôëîðåñà.
%\linebreak \url{http://arxiv.org/abs/1508.03185}
\newcommand{\bmzzn}{\bibitem[BMZ09]{BMZ09} \emph{P. V. M. Blagojevi{\'c}, B. Matschke, G. M. Ziegler,}
Optimal bounds for a colorful Tverberg-Vre\'cica type problem, Advances in Math., 226 (2011), 5198-5215, arXiv:0911.2692.}
\newcommand{\bmzof}{\bibitem[BMZ15]{BMZ15} \emph{P. V. M. Blagojevi{\'c}, B. Matschke, G. M. Ziegler,}
Optimal bounds for the colored Tverberg problem, J. Eur. Math. Soc., 17:4 (2015) 739--754,
arXiv:0910.4987.}
\newcommand{\bpns}{\bibitem[BP97]{BP97} * \emph{R. Benedetti and C. Petronio.} Branched standard spines of 3-manifolds, Lecture Notes in Math. 1653, Springer-Verlag, Berlin-Heidelberg-New York, 1997.}
\newcommand{\brst}{\bibitem[Br72]{Br72} \emph{J. L. Bryant.} Approximating embeddings of polyhedra in codimension 3, Trans. Amer. Math. Soc., 170 (1972) 85--95.}
\newcommand{\brts}{\bibitem[Br26]{Br26} \emph{P. Bruegel.} The Magpie on the Gallows, 1526, \url{https://en.wikipedia.org/wiki/The_Magpie_on_the_Gallows}.}
\newcommand{\bren}{\bibitem[Br82]{brown1982} * \emph{K.~S. Brown.} \newblock Cohomology of Groups. \newblock Springer-Verlag New York, 1982.}
%\bibitem[Bre]{Bre}G. Bredon \newblock {\em Cohomology Theory}
\newcommand{\bssos}{\bibitem[BS17]{BS17} * \emph{I.~B\'{a}r\'{a}ny and P. Sober\'{o}n,} Tverberg's theorem is 50 years old: a survey, Bull. Amer. Math. Soc. (N.S.) 55:4 (2018), 459--492. arXiv:1712.06119.}
\newcommand{\bsto}{\bibitem[BS21]{BS21} * \emph{A. Buchaev and A. Skopenkov,} Simple proofs of estimations of Ramsey numbers and of discrepancy, Mat. Prosveschenie, to appear, arXiv:2107.13831.}
\newcommand{\brsnn}{\bibitem[BRS99]{BRS99} \emph{D. Repov\v s, N. Brodsky and A. B. Skopenkov.}
A classification of 3-thickenings of 2-polyhedra, Topol. Appl. 1999. 94. P.~307-314.}
\newcommand{\bsseo}{\bibitem[BSS]{BSS} \emph{I.~B\'{a}r\'{a}ny, S.~B. Shlosman, and A.~Sz{\H{u}}cs,}
\newblock On a topological generalization of a theorem of {T}verberg,
\newblock J.\ London Math.\ Soc.\ (II. Ser.) 23 (1981), 158--164.}
\newcommand{\btzs}{\bibitem[BT07]{BT07} \emph{A. Bj\"orner, M. Tancer}, Combinatorial Alexander Duality --- a Short and Elementary Proof, Discr. and Comp. Geom., 42 (2009) 586. arXiv:0710.1172.}
\newcommand{\buse}{\bibitem[Bu68]{Bu68} \emph{A. R. Butz,} Space filling curves and mathematical programming, Information and Control, 12:4 (1968) 314--330.}
%\url{http://www.sciencedirect.com/science/article/pii/S0019995868903677}
\newcommand{\bz}{\bibitem[BZ16]{BZ16} * \emph{P. V. M. Blagojevi\'c and G. M. Ziegler,} Beyond the Borsuk-Ulam theorem: The topological Tverberg story, in: A Journey Through Discrete Mathematics, Eds. M. Loebl,
J. Ne\v set\v ril, R. Thomas, Springer, 2017, 273--341. arXiv:1605.07321v3.}
%\url{http://arxiv.org/abs/1605.07321}
\newcommand{\carm}{\bibitem[Ca]{Ca} \emph{J. Carmesin.}
Embedding simply connected 2-complexes in 3-space, I-V, arXiv:1709.04642, arXiv:1709.04643, arXiv:1709.04645, arXiv:1709.04652, arXiv:1709.04659.}
\newcommand{\cfsz}{\bibitem[CF60]{CF60} \emph{P. E. Conner and E. E. Floyd}, Fixed points free involutions and equivariant maps, Bull. Amer. Math. Soc., 66 (1960) 416--441.}
\newcommand{\cget}{\bibitem[CG83]{CG83} \emph{J. H. Conway and C. M. A. Gordon},
Knots and links in spatial graphs, J. Graph Theory 7 (1983), 445--453.}
\newcommand{\cten}{\bibitem[Ch]{Ch} \emph{Chuang Tzu,} translated by H. A. Giles, Bernard Quaritch, London, 1889.}
\newcommand{\ctruku}{\bibitem[Ch]{Ch} \emph{Chuang Tzu,} translated to Russian by S. Kuchera, in: Ancient Chinese Philosophy, v. I, Mysl, Moscow, 1972.}
%Äðåâíåêèòàéñêàÿ ôèëîñîôèÿ. Ñîáðàíèå òåêñòîâ â äâóõ òîìàõ. Ò. I. Ì., <<Ìûñëü>>, 1972.
\newcommand{\chnn}{\bibitem[Ch99]{Ch99} * \emph{À. Â. ×åðíàâñêèé,} Òåîðåìà Æîðäàíà. Ìàò. Ïðîñâåùåíèå, 3 (1999), 142--157.}
\newcommand{\hcon}{\bibitem[HC19]{HC19} * \emph{C. Herbert Clemens.} Two-Dimensional Geometries. A Problem-Solving Approach, Amer. Math. Soc., 2019.}
\newcommand{\ckmoo}{\bibitem[CKMS]{CKMS} \emph{M. \v Cadek, M. Kr\v c\'al. J. Matou\v sek, F. Sergeraert,
L. Vok\v r\'inek, U. Wagner.} Computing all maps into a sphere, J. of the ACM, 61:3 (2014). arXiv:1105.6257.}
%http://arxiv.org/abs/1105.6257.
\newcommand{\ckmvwot}{\bibitem[CKM12+]{CKM12+} \emph{M. \v Cadek, M. Kr\v c\'al. J. Matou\v sek, L. Vok\v r\'inek, U. Wagner.} Polynomial-time computation of homotopy groups and Postnikov systems in fixed dimension, SIAM J. Comput., 43:5 (2014), 1728--1780. arXiv:1211.3093.}
\newcommand{\ckmvw}{\bibitem[CKM+]{CKM+} \emph{M. \v Cadek, M. Kr\v c\'al. J. Matou\v sek, L. Vok\v r\'inek, U. Wagner.} Extendability of continuous maps is undecidable, Discr. and Comp. Geom. 51 (2014) 24--66.
arXiv:1302.2370.}
\newcommand{\ckppt}{\bibitem[CKP+]{CKP+} \emph{E. Colin de Verdi\'ere, V. Kalu\v za, P. Pat\'ak, Z. Pat\'akov\'a and M. Tancer.} A direct proof of the strong Hanani-Tutte theorem on the projective plane. Journal of Graph Algorithms and Applications, 21:5 (2017) 939--981.}
\newcommand{\cksof}{\bibitem[CKS+]{CKS+} * New ways of weaving baskets, presented by G. Chelnokov, Yu. Kudryashov, A.Skopenkov and A. Sossinsky, \url{http://www.turgor.ru/lktg/2004/lines.en/index.htm}.}
\newcommand{\ckv}{\bibitem[CKV]{CKV} \emph{M.~{\v{C}}adek, M.~Kr\v{c}\'{a}l, and L.~Vok\v{r}\'{\i}nek.}
Algorithmic solvability of the lifting-extension problem, Discr. Comp. Geom. 57 (2017), 915--965. arXiv:1307.6444.}
%DOI 10.1007/s00454-016-9855-6
\newcommand{\clr}{\bibitem[CLR]{CLR} * \emph{Ò. Êîðìåí, ×. Ëåéçåðñîí, Ð. Ðèâåñò.} Àëãîðèòìû:
ïîñòðîåíèå è àíàëèç, ÌÖÍÌÎ, Ìîñêâà, 1999.}
\newcommand{\clreng}{\bibitem[CLR]{CLR} * \emph{T. H. Cormen, C. E.Leiserson, R. L.Rivest, C. Stein.} Introduction to Algorithms, MIT Press, 2009.}
\newcommand{\crzfru}{\bibitem[CR]{CR} * \emph{Ð. Êóðàíò, Äæ. Ðîááèíñ,} ×òî òàêîå ìàòåìàòèêà. Ì.: ÌÖÍÌÎ, 2004.}
\newcommand{\crzfen}{\bibitem[CR]{CR} * \emph{R. Courant and H. Robbins,} What is Mathematics, Oxford Univ. Press.}
\newcommand{\crsne}{\bibitem[CRS98]{CRS98} * \emph{A. Cavicchioli, D. Repov\v s and A. B. Skopenkov.}
Open problems on graphs, arising from geometric topology, Topol. Appl. 1998. 84. P.~207-226.}
\newcommand{\crsot}{\bibitem[CRS]{CRS} \emph{M. Cencelj, D. Repov\v s and M. Skopenkov,}
Classification of knotted tori in the 2-metastable dimension, Mat. Sbornik, 203:11 (2012), 1654--1681.
arxiv:math/0811.2745.}
\newcommand{\csoo}{\bibitem[CS08]{CS08} \emph{D. Crowley and A. Skopenkov.} A classification of smooth embeddings of 4-manifolds in 7-space, II, Intern. J. Math., 22:6 (2011) 731-757, arxiv:math/0808.1795.}
\newcommand{\csos}{\bibitem[CS16]{CS16} \emph{D. Crowley and A. Skopenkov,} Embeddings of non-simply-connected 4-manifolds in 7-space. I. Classification modulo knots, Moscow Math. J., 21 (2021), 43--98. arXiv:1611.04738.}
\newcommand{\csoso}{\bibitem[CS16o]{CS16o} \emph{D. Crowley and A. Skopenkov,} Embeddings of non-simply-connected 4-manifolds in 7-space. II. On the smooth classification, Proc. A of the Royal Soc. of Edinburgh 52:1 (2022), 163--181. arXiv:1612.04776.}
%https://doi.org/10.1017/prm.2020.103
\newcommand{\crsk}{\bibitem[CS]{CS} \emph{D. Crowley and A. Skopenkov,} Embeddings of non-simply-connected 4-manifolds in 7-space. III. Piecewise-linear classification. draft.}
\newcommand{\cutz}{\bibitem[Cu20]{Cu20} \emph{C. Culter,} Cantor sets are not tangent homogeneous,
Topol. Appl. 271 (2020) 1--9.}
\newcommand{\dies}{\bibitem[Di87]{Di} * \emph{T. tom Dieck,} Transformation groups, Studies in Mathematics, vol. 8, Walter de Gruyter, Berlin, 1987.}
\newcommand{\dent}{\bibitem[De93]{De93} \emph{T.K. Dey.} On counting triangulations in $d$-dimensions. Comput. Geom. 3:6 (1993) 315--325.}
\newcommand{\denf}{\bibitem[DE94]{DE94} \emph{T.K. Dey and H. Edelsbrunner.} Counting triangle crossings and halving planes, Discrete Comput. Geom. 12 (1994), 281--289.}
\newcommand{\embo}{\bibitem[Eb]{Eb} * \url{http://www.map.mpim-bonn.mpg.de/Embeddings_of_manifolds_with_boundary:_classification}}
\newcommand{\embe}{\bibitem[Em]{Em} * \url{http://www.map.mpim-bonn.mpg.de/Embedding_(simple_definition)}}
\newcommand{\ers}{\bibitem[ERS]{ERS} * Invariants of graph drawings in the plane, presented by A. Enne, A. Ryabichev, A. Skopenkov and T. Zaitsev, \url{http://www.turgor.ru/lktg/2017/6/index.htm}}
\newcommand{\feto}{\bibitem[Fe21]{Fe21} \emph{M. Fedorov.} A description of values of Seifert form for punctured $n$-manifolds in $(2n-1)$-space, arXiv:2107.02541.}
\newcommand{\ffen}{\bibitem[FF89]{FF89} * \emph{À. Ò. Ôîìåíêî è Ä. Á. Ôóêñ.} Êóðñ ãîìîòîïè÷åñêîé òîïîëîãèè. Ì.: Íàóêà, 1989.}
\newcommand{\ffene}{\bibitem[FF89]{FF89} * \emph{A.T. Fomenko and D.B. Fuchs.} Homotopical Topology, Springer, 2016.}
%A course in homotopy theory (in Russian). Moscow: Nauka, 1989.
%English translation of a preliminary edition: Budapest, AK, 1986.}
\newcommand{\fkosc}{\bibitem[FK17]{FK17} \emph{R. Fulek, J. Kyn{\v{c}}l,} Counterexample to an Extension of the Hanani-Tutte Theorem on the Surface of Genus 4, Combinatorica, 39 (2019) 1267--1279, arXiv:1709.00508.}
\newcommand{\fkos}{\bibitem[FK17]{FK17} \emph{R. Fulek, J. Kyn{\v{c}}l,} Hanani-Tutte for approximating maps of graphs, arXiv:1705.05243.}
\newcommand{\fkon}{\bibitem[FK19]{FK19} \emph{R. Fulek, J. Kyn{\v{c}}l,}
$\Z_2$-genus of graphs and minimum rank of partial symmetric matrices,
35th Intern. Symp. on Comp. Geom. (SoCG 2019), Article No. 39; pp. 39:139:16,
\url{https://drops.dagstuhl.de/opus/volltexte/2019/10443/pdf/LIPIcs-SoCG-2019-39.pdf}.
We refer to numbering in arXiv version: arXiv:1903.08637.}
\newcommand{\fktnf}{\bibitem[FKT]{FKT} \emph{M. H. Freedman, V. S. Krushkal and P. Teichner.} Van Kampen's
embedding obstruction is incomplete for 2-complexes in~$\R^4$, Math. Res. Letters. 1994. 1. P.~167-176.}
\newcommand{\fltf}{\bibitem[Fl34]{Fl34} \emph{A. Flores}, \"Uber $n$-dimensionale Komplexe die im $E^{2n+1}$ absolut selbstverschlungen sind, Ergeb. Math. Koll. 6 (1934) 4--7.}
\newcommand{\fo}{\bibitem[Fo]{Fo} * \emph{L. Fortnow.} Time for Computer Science to Grow Up, \url{https://people.cs.uchicago.edu/~fortnow/papers/growup.pdf}.}
\newcommand{\fozf}{\bibitem[Fo04]{Fo04} * \emph{R. Fokkink.} A forgotten mathematician, Eur. Math. Soc. Newsletter 52 (2004) 9--14.}
%\url{http://www.ems-ph.org/journals/newsletter/pdf/2004-06-52.pdf}.}
\newcommand{\fpstz}{\bibitem[FPS]{FPS} \emph{R. Fulek, M.J. Pelsmajer and M. Schaefer.}
Strong Hanani-Tutte for the Torus, arXiv:2009.01683.}
\newcommand{\frse}{\bibitem[Fr78]{Fr78} \emph{M. Freedman,} Quadruple points of 3-manifolds in $S^4$, Comment. Math. Helv. 53 (1978), 385-394.}
\newcommand{\fres}{\bibitem[FR86]{FR86} \emph{R. Fenn, D. Rolfsen.}
Spheres may link homotopically in 4-space, J. London Math. Soc. 34 (1986) 177-184.}
\newcommand{\frof}{\bibitem[Fr15]{Fr15} \emph{F. Frick}, Counterexamples to the topological Tverberg conjecture,
Oberwolfach reports, 12:1 (2015), 318--321. arXiv:1502.00947.}
%http://arxiv.org/abs/1502.00947
\newcommand{\fros}{\bibitem[Fr17]{Fr17} \emph{F. Frick}, O\lowercase{N AFFINE TVERBERG-TYPE RESULTS WITHOUT CONTINUOUS GENERALIZATION}, arXiv:1702.05466}
%https://arxiv.org/pdf/1702.05466.pdf,
\newcommand{\fstz}{\bibitem[FS20]{FS20} \emph{F. Frick and P. Sober\'on}, The topological Tverberg problem beyond prime powers, arXiv:2005.05251.}
\newcommand{\fvto}{\bibitem[FV21]{FV21} \emph{M. Filakovsk\'y, L. Vok\v r\'inek.} Computing homotopy classes for diagrams, arXiv:2104.10152.}
\newcommand{\fwz}{\bibitem[FWZ]{FWZ} \emph{M. Filakovsk\'y, U. Wagner, S. Zhechev.} Embeddability of simplicial complexes is undecidable. Oberwolfach reports, to appear.}
\newcommand{\fwztz}{\bibitem[FWZ]{FWZ} \emph{M. Filakovsk\'y, U. Wagner, S. Zhechev.} Embeddability of simplicial complexes is undecidable.
Proceedings of the 2020 ACM-SIAM Symposium on Discrete Algorithms, \url{https://epubs.siam.org/doi/pdf/10.1137/1.9781611975994.47}}
\newcommand{\ga}{\bibitem[GA]{GA} * \url{https://en.wikipedia.org/wiki/Galactic_algorithm}}
\newcommand{\gdikrse}{\bibitem[GDI]{GDI} * {\it A. Chernov, A. Daynyak, A. Glibichuk, M. Ilyinskiy, A. Kupavskiy, A. Raigorodskiy and A. Skopenkov,} Elements of Discrete Mathematics As a Sequence of Problems (in Russian),
MCCME, Moscow, 2016. Update: \url{http://www.mccme.ru/circles/oim/discrbook.pdf} .}
\newcommand{\gdikrs}{\bibitem[GDI]{GDI} * {\it À.À. Ãëèáè÷óê, À.Á. Äàéíÿê, Ä.Ã. Èëüèíñêèé, À.Á. Êóïàâñêèé, À.Ì. Ðàéãîðîäñêèé, À.Á. Ñêîïåíêîâ, À.À. ×åðíîâ,} Ýëåìåíòû äèñêðåòíîé ìàòåìàòèêè â çàäà÷àõ, Ì, ÌÖÍÌÎ, 2016.
\url{http://www.mccme.ru/circles/oim/discrbook.pdf} .}
\newcommand{\giso}{\bibitem[Gi71]{Gi71} * {\it S. Gitler,} Immersion and Embedding of Manifolds,
Proc. Symp. Pure Math. 22, 87-96 (1971).}
\newcommand{\gkp}{\bibitem[GKP]{GKP} * {\it R. Graham, D. Knuth, and O. Patashnik,} Concrete Mathematics: A Foundation for Computer Science, AddisonWesley, first published in 1989, \url{https://www.csie.ntu.edu.tw/~r97002/temp/Concrete\%20Mathematics\%202e.pdf}.}
\newcommand{\gmpptw}{\bibitem[GMP+]{GMP+} \emph{X. Goaoc, I. Mabillard, P. Pat\'ak, Z. Pat\'akov\'a, M. Tancer, U. Wagner}, On Generalized Heawood Inequalities for Manifolds: a van Kampen--Flores-type Nonembeddability Result,
arXiv:1610.09063.}
\newcommand{\group}{\bibitem[Gr]{Gr} * \url{https://en.wikipedia.org/wiki/Groupthink}}
\newcommand{\grsz}{\bibitem[Gr69]{Gr69} \emph{B. Gr\"unbaum.} Imbeddings of simplicial complexes. Comment. Math. Helv., 44:1, 502--513, 1969.}
%https://link.springer.com/article/10.1007/BF02564551
\newcommand{\gres}{\bibitem[Gr86]{Gr86} * \emph{M. Gromov}, Partial Differential Relations,
Ergebnisse der Mathematik und ihrer Grenzgebiete (3), Springer Verlag, Berlin-New York, 1986.}
\newcommand{\groz}{\bibitem[Gr10]{Gr10} \emph{M. Gromov,}
\newblock Singularities, expanders and topology of maps. Part 2: From combinatorics to topology via algebraic isoperimetry, \newblock Geometric and Functional Analysis 20 (2010), no.~2, 416--526.}
\newcommand{\grsn}{\bibitem[GR79]{GR79} \emph{J. L. Gross and R. H. Rosen}, A linear time planarity algorithm for 2-complexes, Journal of the ACM, 26:4 (1979), 611--617.}
\newcommand{\gs}{\bibitem[GS]{GS} \emph{Ì. Ãîðòèíñêèé è Î. Ñêðÿáèí.} Êðèòåðèé âëîæèìîñòè ãðàôîâ â ïëîñêîñòü âäîëü ïðÿìîé, ïðåïðèíò.}
\newcommand{\gssn}{\bibitem[GS79]{GS} \emph{P.~M. Gruber and R.~Schneider,} Problems in geometric convexity. In {\em Contributions to geometry ({P}roc. {G}eom. {S}ympos., {S}iegen, 1978)}, 255--278. Birkh{\"a}user, Basel-Boston, Mass., 1979.}
\newcommand{\gsnn}{\bibitem[GS99]{GS99} \emph{R. Gompf and A. Stipsicz,}
4-manifolds and Kirby calculus, GSM20, AMS, Providence, RI, 1999.}
%Graduate Studies in Mathematics 20, American Mathematical Society, MR1707327 (2000h:57038)
\newcommand{\gszs}{\bibitem[GS06]{GS06} \emph{D. Goncalves and A. Skopenkov,} Embeddings of homology equivalent manifolds with boundary, Topol. Appl., 153:12 (2006) 2026-2034. arxiv:1207.1326.}
\newcommand{\gssoe}{\bibitem[GSS+]{GSS+} * Projections of skew lines, presented by A. Gaifullin, A. Shapovalov, A. Skopenkov and M. Skopenkov, \url{http://www.turgor.ru/lktg/2001/index.php}.}
\newcommand{\gtes}{\bibitem[GT87]{GT87} * \emph{J. L. Gross and T. W. Tucker.}
Topological graph theory. New York: Wiley-Interscience, 1987.}
\newcommand{\guzn}{\bibitem[Gu09]{Gu09} \emph{A. Gundert.} On the complexity of embeddable simplicial complexes. Diplomarbeit, Freie Universit\"at Berlin, 2009. arXiv:1812.08447.}
\newcommand{\ha}{\bibitem[Ha]{Ha} * \emph{F. Harary.} Graph theory.
Ðóñ. ïåð.: Ô. Õàðàðè. Òåîðèÿ ãðàôîâ. Ì., Ìèð, 1973.}
\newcommand{\hats}{\bibitem[Ha37]{Ha37} \emph{W. Hantzsche,} Einlagerung von Mannigfaltigkeiten in euklidische R\" aume, Math. Zeitschrift, 43:1 (1937) 38--58.}
\newcommand{\hastk}{\bibitem[Ha62k]{Ha62k} {\em A.~Haefliger,} Knotted $(4k-1)$-spheres in $6k$-space, Ann. of Math. 75 (1962) 452--466.}
\newcommand{\hastl}{\bibitem[Ha62l]{Ha62l} \emph{A. Haefliger,} Differentiable links, Topology, 1 (1962) 241--244.}
\newcommand{\hast}{\bibitem[Ha63]{Ha63} \emph{A.~Haefliger,} Plongements differentiables dans le domain stable, Comment. Math. Helv. 36 (1962-63) 155--176.}
\newcommand{\hassa}{\bibitem[Ha66A]{Ha66A} \textit{A. Haefliger}. Differential embeddings of~$S^n$ in $S^{n+q}$ for $q>2$. Ann. Math. (2), 83 (1966), 402--~436.}
\newcommand{\hass}{\bibitem[Ha66C]{Ha66C} \emph{A.~Haefliger,} Enlacements de spheres en codimension superiure a 2, Comment. Math. Helv. 41 (1966-67) 51--72.}
\newcommand{\hase}{\bibitem[Ha68]{Ha68} \emph{A. Haefliger,} Knotted Spheres and Related Geometric Topic,
in Proc. Int. Congr. Math., Moscow, 1966 (Mir, Moscow, 1968), 437--445.}
\newcommand{\hasn}{\bibitem[Ha69]{Ha69} \emph{L.~S.~Harris,} Intersections and embeddings of polyhedra, Topology 8 (1969) 1--26.}
\newcommand{\hasf}{\bibitem[Ha74]{Ha74} * \emph{P. Halmos,} How to talk mathematics. Notices of the Amer. Math. Soc., 21 (1974) 155--158.}
\newcommand{\haef}{\bibitem[Ha84]{Ha84} \emph{N. Habegger,} Obstruction to embedding disks II: a proof of a conjecture by Hudson, Topol. Appl. 17 (1984).}
\newcommand{\haes}{\bibitem[Ha86]{Ha86} \emph{N. Habegger,} Knots and links in codimension greater than 2, Topology, 25:3 (1986) 253--260.}
\newcommand{\hifn}{\bibitem[Hi59]{Hi59} \emph{M. W. Hirsch.} Immersions of manifolds, Trans. Amer. Math. Soc. 93 (1959) 242--276.}
\newcommand{\hjsf}{\bibitem[HJ64]{HJ64} \emph{R. Halin and H. A. Jung.}
Karakterisierung der Komplexe der Ebene und der 2-Sph\"are, Arch. Math. 1964. 15. P.~466-469.}
\newcommand{\hkne}{\bibitem[HK98]{HK98} \emph{N. Habegger and U. Kaiser,} Link homotopy in 2--metastable range, Topology 37:1 (1998) 75--94.}
\newcommand{\hmsnt}{\bibitem[HMS]{HMS93} * \emph{C. Hog-Angeloni, W. Metzler and A. J. Sieradski.}
Two-dimensional homotopy and combinatorial group theory. Cambridge: Cambridge Univ. Press, 1993. (London Math. Soc. Lecture Notes, 197).}
\newcommand{\ho}{\bibitem[Ho]{Ho} * The Hopf fibration, \url{https://www.youtube.com/watch?v=AKotMPGFJYk}}
\newcommand{\hozs}{\bibitem[Ho06]{Ho06} \emph{H. van der Holst,} Graphs and obstructions in four dimensions, J. Combin. Theory Ser. B 96:3 (2006), 388--404.}
%https://www.sciencedirect.com/science/article/pii/S0095895605001425?via%3Dihub
\newcommand{\hpzn}{\bibitem[HP09]{HP09} \emph{H. van der Holst and R. Pendavingh,} On a graph property generalizing planarity and flatness, Combinatorica, 29 (2009) 337--361.}
\newcommand{\htsf}{\bibitem[HT74]{HT74} \emph{J. Hopcroft and R. E. Tarjan,} Efficient planarity testing, J. of the Association for Computing Machinery, 21:4 (1974) 549--568.}
\newcommand{\hufn}{\bibitem[Hu59]{hu59} * \emph{S. T. Hu,} Homotopy Theory, Academic Press, New York, 1959.}
\newcommand{\husn}{\bibitem[Hu69]{Hu69} * \emph{J. F. P. Hudson.} Piecewise linear topology, W. A. Benjamin, Inc., New York-Amsterdam, 1969.}
\newcommand{\io}{\bibitem[Io]{Io} * \url{https://en.wikipedia.org/wiki/Category:Impossible_objects}}
\newcommand{\info}{\bibitem[IF]{IF} * \url{http://www.map.mpim-bonn.mpg.de/Intersection_form}}
\newcommand{\irsf}{\bibitem[Ir65]{Ir65} \emph{M.~C.~Irwin,} Embeddings of polyhedral manifolds, Ann. of Math. (2)
82 (1965) 1--14.}
\newcommand{\isot}{\bibitem[Is]{Is} * \url{http://www.map.mpim-bonn.mpg.de/Isotopy}}
\newcommand{\jqnt}{\bibitem[JQ93]{JQ93} * \emph{A. Jaffe, F. Quinn,} ``Theoretical mathematics'': Toward a cultural synthesis of mathematics and theoretical physics. Bull.Am.Math.Soc. 29 (1993) 1-13. arXiv:math/9307227.}
\newcommand{\jozt}{\bibitem[Jo02]{Jo02} \emph{C. M. Johnson.} An obstruction to embedding a simplicial $n$-complex into a $2n$-manifold, Topology Appl. 122:3 (2002) 581--591.}
\newcommand{\jvz}{\bibitem[JVZ]{JVZ} D. Joji\'c, S. T. Vre\'cica, R. T. \v Zivaljevi\' c,
Topology and combinatorics of 'unavoidable complexes', arXiv:1603.08472v1.}
%\url{http://arxiv.org/abs/1603.08472}
\newcommand{\kalai}{\bibitem[Ka]{Ka} G. Kalai, From Oberwolfach: The Topological Tverberg Conjecture is False, `Combinatorics and more' blog post, February 6, 2015, \url{gilkalai.wordpress.com}}
\newcommand{\kh}{\bibitem[Kh]{Kh} \emph{À.È. Õðàáðîâ.} Ðóêîâîäñòâî ïî ÷òåíèþ ëåêöèé
\url{http://vm.tstu.tver.ru/topics/pdf_tests/lection.pdf}}
\newcommand{\kho}{\bibitem[Kho]{Kho} \emph{N. Khoroshavkina.} A simple characterization of graphs of cutwidth 2, arXiv:1811.06716.}
\newcommand{\kkrot}{\bibitem[KKR]{KKR} \emph{K. Kawarabayashi, Y. Kobayashi and B. Reed.} The disjoint paths problem in quadratic time, J. of Comb. Theory, Ser. B, 102:2 (2012), 424--435.}
\newcommand{\kmsth}{\bibitem[KM63]{KM63} \emph{M. A. Kervaire and J. W. Milnor,} Groups of homotopy spheres. I, Ann. of Math. (2) 77 (1963), 504-537.}
\newcommand{\kozeru}{\bibitem[Ko18]{Ko18} * \emph{Å. Êîëïàêîâ.}
Äîêàçàòåëüñòâî òåîðåìû Ðàäîíà ïðè ïîìîùè ïîíèæåíèÿ ðàçìåðíîñòè, Ìàò. Ïðîñâåùåíèå, 23 (2018), arXiv:1903.11055.}
\newcommand{\koze}{\bibitem[Ko18]{Ko18} * \emph{E. Kolpakov.}
A proof of Radon Theorem via lowering of dimension, Mat. Prosveschenie, 23 (2018), arXiv:1903.11055.}
\newcommand{\ko}{\bibitem[Ko]{Ko} \emph{E. Kolpakov.} A `converse' to the Constraint Lemma, arXiv:1903.08910.}
\newcommand{\koon}{\bibitem[Ko19]{Ko19} \emph{E. Kogan.} Linking of three triangles in 3-space, arXiv:1908.03865.}
\newcommand{\koto}{\bibitem[Ko21]{Ko21} \emph{E. Kogan.} On the rank of $\Z_2$-matrices with free entries on the diagonal, arXiv:2104.10668.}
\newcommand{\koee}{\bibitem[Ko88]{Ko88} \emph{U. Koschorke.} Link maps and the geometry of their invariants,
Manuscripta Math. 61:4 (1988) 383--415.}
\newcommand{\kps}{\bibitem[KPS]{KPS} * \emph{A. Kaibkhanov, D. Permyakov and A. Skopenkov.}
Realization of graphs with rotation, \url{http://www.turgor.ru/lktg/2005/3/index.htm}.}
\newcommand{\krzz}{\bibitem[Kr00]{Kr00} \emph{V. S. Krushkal.} Embedding obstructions and 4-dimensional thickenings of 2-complexes, Proc. Amer. Math. Soc. 128:12 (2000) 3683--3691. arXiv:math/0004058. }
\newcommand{\ksnn}{\bibitem[KS99]{KS99} * \emph{Ï. Êîæåâíèêîâ è À. Ñêîïåíêîâ.} Óçêèå äåðåâüÿ íà ïëîñêîñòè, Ìàò. Îáðàçîâàíèå. 1999. 2-3. Ñ.~126-131.}
\newcommand{\kstz}{\bibitem[KS20]{KS20} \emph{R. Karasev and A. Skopenkov.}
Some `converses' to intrinsic linking theorems, arXiv:2008.02523.}
\newcommand{\ksto}{\bibitem[KS21]{KS21} \emph{E. Kogan and A. Skopenkov.} A short exposition of the Patak-Tancer theorem on non-embeddability of $k$-complexes in $2k$-manifolds, arXiv:2106.14010.}
\newcommand{\kstoe}{\bibitem[KS21e]{KS21e} \emph{E. Kogan and A. Skopenkov.}
Embeddings of $k$-complexes in $2k$-manifolds and minimum rank of partial symmetric matrices, arXiv:2112.06636.}
\newcommand{\kuse}{\bibitem[Ku68]{Ku68} * \emph{Ê. Êóðàòîâñêèé.} Òîïîëîãèÿ. Ò.~1,~2. Ì.: Ìèð, 1969.}
\newcommand{\lazz}{\bibitem[La00]{La00} \emph{F. Lasheras.} An obstruction to 3-dimensional thickening,
Proc. Amer. Math. Soc. 2000. 128. P.~893-902.}
\newcommand{\lfma}{\bibitem[LF]{LF} \url{http://www.map.mpim-bonn.mpg.de/Linking_form}}
\newcommand{\lloe}{\bibitem[LL18]{LL18} \emph{A.S. Levine and T. Lidman.} Simply connected, spineless 4-manifolds, Forum of Math., Sigma, 7 (2019) e14, 1--11, arxiv:1803.01765.}
\newcommand{\lo}{\bibitem[Lo]{Lo} M.~de~Longueville. Notes on the topological Tverberg theorem.
Discrete Math. 247 (2002), no.~1--3, 271--297.
(The paper first appeared in
%a volume of selected papers in honor of Helge Tverberg,
Discrete Math. 241 (2001) 207--233, but the original version suffered from serious publisher's typesetting errors.)}
\newcommand{\loot}{\bibitem[Lo13]{Lo13} \emph{M. de Longueville.} A course in topological combinatorics. Universitext. Springer, New York (2013).}
\newcommand{\lssn}{\bibitem[LS69]{LS69} \emph{W. B. R. Lickorish and L. C. Siebenmann.}
Regular neighborhoods and the stable range, Trans. Amer. Math. Soc.. 1969. 139. P.~207-230.}
\newcommand{\lsne}{\bibitem[LS98]{LS98} \emph{L. Lovasz and A. Schrijver,}
A Borsuk theorem for antipodal links and a spectral characterization of linklessly embeddable graphs, Proc. Amer. Math. Soc. 126:5 (1998), 1275-1285.}
\newcommand{\ltof}{\bibitem[LT14]{LT14} \emph{E. Lindenstrauss and M. Tsukamoto,} Mean dimension and an embedding problem: an example, Israel J. Math. 199 (2014).}
%http://www.math.huji.ac.il/~elon/Publications/mdim-example.pdf
\newcommand{\lyzf}{\bibitem[LY04]{LY04} * \emph{Y. Lin and A. Yang,} On 3-cutwidth critical graphs, Discrete Mathematics, 275 (2004), 339--346.}
\newcommand{\lz}{\bibitem[LZ]{LZ} * \emph{S. Lando and A. Zvonkin.} Embedded Graphs. Springer.}
%\newcommand{\ma}{\bibitem[MA]{MA} * %\url{http://www.map.mpim-bonn.mpg.de/High_codimension_links#The_linking_coefficient}
%Manifold Atlas Project}
\newcommand{\mast}{\bibitem[Ma73]{Ma73} \emph{Ñ. Â. Ìàòâååâ.} Ñïåöèàëüíûå îñòîâû êóñî÷íî-ëèíåéíûõ ìíîãîîáðàçèé, Ìàò. Ñáîðíèê. 1973. 92. Ñ.~282-293.}
\newcommand{\maste}{\bibitem[Ma73]{Ma73} \emph{S. V. Matveev.} Special skeletons of PL manifolds (in Russian), Mat. Sbornik. 1973. 92. P.~282-293.}
\newcommand{\mans}{\bibitem[Ma97]{Ma97} \emph{Yu. Makarychev.} A short proof of Kuratowski's graph planarity criterion, J. of Graph Theory, 25 (1997), 129--131.}
\newcommand{\mazt}{\bibitem[Ma03]{Ma03} * \emph{J.~Matou{\v{s}}ek.} Using the {B}orsuk-{U}lam theorem:
Lectures on topological methods in combinatorics and geometry. Springer Verlag, 2008.}
%http://books.google.ru/books/about/
%Using\_the\_Borsuk\_Ulam\_Theorem.html?id=tru-TWQFDAIC\&redir\_esc=y
\newcommand{\mazf}{\bibitem[Ma05]{Ma05} \emph{V. Manturov.} A proof of the Vasiliev conjecture on the planarity of singular links, Izv. RAN 2005.}
\newcommand{\metn}{\bibitem[Me29]{Me29} \emph{K. Menger.} \"Uber pl\"attbare Dreiergraphen und Potenzen nicht pl\"attbarer Graphen, Ergebnisse Math. Kolloq., 2 (1929) 30--31.}
\newcommand{\mezf}{\bibitem[Me04]{Me04} \emph{S. Melikhov.} Sphere eversions and realization of mappings, Trudy MIAN 247 (2004) 159-181 (in Russian) arXiv:math.GT/0305158.}
\newcommand{\mezs}{\bibitem[Me06]{Me06} \emph{S. A. Melikhov}, The van Kampen obstruction and its relatives,
Proc. Steklov Inst. Math 266 (2009), 142-176 (= Trudy MIAN 266 (2009), 149-183), arXiv:math/0612082.}
%https://arxiv.org/abs/math/0612082
\newcommand{\meoo}{\bibitem[Me11]{Me11} \emph{S. A. Melikhov}, Combinatorics of embeddings, arXiv:1103.5457.}
%\url{https://arxiv.org/abs/1103.5457}
\newcommand{\meos}{\bibitem[Me17]{Me17} \emph{S. Melikhov,} Gauss type formulas for link map invariants, arXiv:1711.03530.}
%unpublished, 2007.
\newcommand{\meoe}{\bibitem[Me18]{Me18} \emph{S. A. Melikhov,} A triple-point Whitney trick, J. Topol. Anal., 2018, 1--6.}
%https://doi.org/10.1142/S1793525319500730
\newcommand{\miso}{\bibitem[Mi61]{Mi61} \emph{J. Milnor,} A procedure for killing homotopy groups of differentiable manifolds, Proc. Sympos. Pure Math, Vol. III (1961), 39--55.}
\newcommand{\mins}{\bibitem[Mi97]{Mi97} \emph{P. Minc.} Embedding simplicial arcs into the plane, Topol. Proc. 1997. 22. 305--340.}
%\bibitem[MF90]{MF90} \emph{*Ñ. Â. Ìàòâååâ è À. Ò. Ôîìåíêî.}
%Àëãîðèòìè÷åñêèå è êîìïüþòåðíûå ìåòîäû â òðåõìåðíîé òîïîëîãèè. Ì.: Èçä-âî ÌÃÓ. Èëè: Ì.: Íàóêà. 1990.
\newcommand{\moss}{\bibitem[Mo77]{Mo77} * \emph{E. E. Moise.} Geometric Topology in Dimensions 2 and 3 (GTM), Springer-Verlag, 1977.}
\newcommand{\moen}{\bibitem[Mo89]{Mo89} \textit{B. Mohar}. An obstruction to embedding graphs in
surfaces. Discrete Math. 78 (1989) 135--142.}
\newcommand{\mrst}{\bibitem[MRS+]{MRS+} \emph{A. de Mesmay, Y. Rieck, E. Sedgwick, M. Tancer,}
Embeddability in $\R^3$ is NP-hard. arXiv:1708.07734.}
\newcommand{\mesczs}{\bibitem[MS06]{MS06} \emph{S.A. Melikhov, E.V. Shchepin,} The telescope approach to embeddability of compacta. arXiv:math.GT/0612085.}
\newcommand{\mstwof}{\bibitem[MST+]{MST+} \emph{J. Matou\v sek, E. Sedgwick, M. Tancer, U. Wagner}, Embeddability in the 3-sphere is decidable, Journal of the ACM 65:1 (2018) 1--49, arXiv:1402.0815.}
%DOI10.1145/3078632 Proc. 30th Ann. ACM Symp. on Comput. Geom. (SoCG), 2014, 78--84.
%\url{http://arxiv.org/abs/1402.0815}
\newcommand{\mtzo}{\bibitem[MT01]{MT01} * \emph{B. Mohar and C. Thomassen.} Graphs on Surfaces.
The John Hopkins University Press, 2001.}
\newcommand{\mtwoz}{\bibitem[MTW10]{MTW10} \emph{J. Matou\v sek, M. Tancer, U. Wagner.} A geometric proof of
the colored Tverberg theorem, Discr. and Comp. Geometry, 47:2 (2012), 245--265. arXiv:1008.5275.}
%\url{http://arxiv.org/abs/1008.5275}
\newcommand{\mtwoo}{\bibitem[MTW]{MTW} \emph{J. Matou\v sek, M. Tancer, U. Wagner.}
Hardness of embedding simplicial complexes in $\R^d$, J. Eur. Math. Soc. 13:2 (2011), 259--295. arXiv:0807.0336.}
%\url{http://arxiv.org/abs/0807.0336}
\newcommand{\mwofo}{\bibitem[MW14]{MW14} \emph{I. Mabillard and U. Wagner.} Eliminating Tverberg Points, I. An Analogue of the Whitney Trick, Proc. of the 30th Annual Symp. on Comp. Geom. (SoCG'14), ACM, New York, 2014, pp. 171--180.}
\newcommand{\mwof}{\bibitem[MW15]{MW15} \emph{I. Mabillard and U. Wagner.}
Eliminating Higher-Multiplicity Intersections, I. A Whitney Trick for Tverberg-Type Problems. arXiv:1508.02349.}
%\linebreak \url{http://arxiv.org/pdf/1508.02349v1.pdf}
\newcommand{\mwos}{\bibitem[MW16]{MW16} \emph{I. Mabillard and U. Wagner.} Eliminating Higher-Multiplicity Intersections, II. The Deleted Product Criterion in the $r$-Metastable Range. arXiv:1601.00876v2.}
\newcommand{\mwosd}{\bibitem[MW16']{MW16'} \emph{I. Mabillard and U. Wagner.} Eliminating Higher-Multiplicity Intersections, II. The Deleted Product Criterion in the r-Metastable Range,
Proceedings of the 32nd Annual Symposium on Computational Geometry (SoCG'16).}
\newcommand{\neno}{\bibitem[Ne91]{Ne91} \emph{S. Negami.} Ramsey theorems for knots, links and spatial graphs,
Trans. Amer. Math. Soc., 324 (1991), 527--541.}
%book on giving math talks
%https://sites.google.com/site/tutamnguyenphan/book.pdf
%My teaching statement
%https://sites.google.com/site/tutamnguyenphan/TeachingStatement.pdf
\newcommand{\nkon}{\bibitem[NKS]{NKS} * \emph{L. T. Nguyen, J. Kim, B. Shim.}
Low-Rank Matrix Completion: A Contemporary Survey. arXiv:1907.11705.}
\newcommand{\noss}{\bibitem[No76]{No76} * \emph{Ñ. Ï. Íîâèêîâ.} Òîïîëîãèÿ-1. Ì.: Íàóêà, 1976. (Èòîãè íàóêè è òåõíèêè. ÂÈÍÈÒÈ. Ñîâðåìåííûå ïðîáëåìû ìàòåìàòèêè. Îñíîâíûå íàïðàâëåíèÿ, 12).}
\newcommand{\nwns}{\bibitem[NW97]{NW97} \emph{A. Nabutovsky, S. Weinberger}. Algorithmic aspects of homeomorphism problems. arXiv:math/9707232.}
\newcommand{\omoe}{\bibitem[Om18]{Om18} * \emph{À. Îìåëü÷åíêî,} Òåîðèÿ ãðàôîâ. Ì.: ÌÖÍÌÎ, 2018.}
\newcommand{\ossf}{\bibitem[OS74]{OS74} \emph{R. P. Osborne and R. S. Stevens.} Group presentations
corresponding to spines of 3-manifolds, I, Amer. J.~Math. 1974. 96. P.~454-471; II, Amer. J.~Math. 1977. 234.
P.~213-243; III, Amer. J.~Math. 1977. 234 P.~245-251.}
% ??? óòî÷íèòü íîìåð æóðíàëà --- ñòðàííî, ÷òî â 1974 ã. íîìåð 96, à â 1977 óæå 234.
\newcommand{\oz}{\bibitem[Oz]{Oz} \emph{M. \"Ozaydin,} Equivariant maps for the symmetric group, unpublished,
\url{http://minds.wisconsin.edu/handle/1793/63829}.}
\newcommand{\panof}{\bibitem[Pan15]{Pan15} \emph{K. Panagiotis.} A note on the topology of irreducible $SO(3)$-manifolds, arXiv:1508.06150.}
\newcommand{\paof}{\bibitem[Pa15]{Pa15} \emph{S. Parsa,} On links of vertices in simplicial $d$-complexes embeddable in the euclidean $2d$-space, Discrete Comput. Geom. 59:3 (2018), 663--679.
This is arXiv:1512.05164v4 up to numbering of sections, theorems etc; we refer to numbering in arxiv version.}
\newcommand{\paoe}{\bibitem[Pa18]{Pa18} \emph{S. Parsa,} On links of vertices in simplicial $d$-complexes
embeddable in the euclidean $2d$-space, arXiv:1512.05164v6.}
\newcommand{\patz}{\bibitem[Pa20]{Pa20} \emph{S. Parsa,} On links of vertices in simplicial $d$-complexes
embeddable in the euclidean $2d$-space, arXiv:1512.05164v8.}
%still no reference to Sk
\newcommand{\patzl}{\bibitem[Pa20]{Pa20} \emph{S. Parsa,}
Correction to: On the Links of Vertices in Simplicial $d$-Complexes Embeddable in the Euclidean $2d$-Space,
Discrete Comput. Geom. (2020).}
\newcommand{\patza}{\bibitem[Pa20a]{Pa20a} \emph{S. Parsa,} On the Smith classes, the van Kampen obstruction and embeddability of $[3]*K$, arXiv:2001.06478.}
\newcommand{\patzb}{\bibitem[Pa20b]{Pa20b} \emph{S. Parsa,} On the embeddability of $[3]*K$, arXiv:2001.06506.}
\newcommand{\pak}{\bibitem[Pa]{Pa} * \emph{I. Pak}, Lectures on Discrete and Polyhedral Geometry, \url{http://www.math.ucla.edu/~pak/geompol8.pdf}.}
\newcommand{\peze}{\bibitem[Pe08]{Pe08} \emph{Ä. Ïåðìÿêîâ.} Êëàññèôèêàöèÿ ïîãðóæåíèé ãðàôîâ â ïëîñêîñòü,
Âåñòíèê ÌÃÓ, ñåð.1, 2008, N5, 55-56.}
\newcommand{\peos}{\bibitem[Pe16]{Pe16} \emph{Ä. Ïåðìÿêîâ.} Ìàòåì. ñá., 207:6 (2016), 93--112.}
\newcommand{\pest}{\bibitem[Pe72]{Pe72} * \emph{B. B. Peterson.} The Geometry of Radon's Theorem, Amer. Math. Monthly 79 (1972), 949-963.}
%\url{http://hipatia.dma.ulpgc.es/profesores/personal/aph/ficheros/resolver/ficheros/crp/radon.pdf}.}
\newcommand{\prnf}{\bibitem[Pr95]{Pr95} * \emph{V. V. Prasolov.} Intuitive topology. Amer. Math. Soc., Providence, R.I., 1995.}
\newcommand{\prnfr}{\bibitem[Pr95]{Pr95} * \emph{Â. Â. Ïðàñîëîâ.} Íàãëÿäíàÿ òîïîëîãèÿ. Ì.: ÌÖÍÌÎ, 1995.}
%http://carlossicoli.free.fr/P/Prasolov\_V.V.-Intuitive\_topology\_4-Universities\_Press\_India\_1998.pdf
%http://prasolov.loegria.net/nagltop-3ed.pdf
\newcommand{\przs}{\bibitem[Pr06]{Pr06} * \emph{V. V. Prasolov.}
Elements of Combinatorial and Differential Topology, 2006, GSM 74, Amer. Math. Soc., Providence, RI.}
\newcommand{\przsru}{\bibitem[Pr04]{Pr04} * \emph{Â. Â. Ïðàñîëîâ.}
Ýëåìåíòû êîìáèíàòîðíîé è äèôôåðåíöèàëüíîé òîïîëîãèè. Ì.: ÌÖÍÌÎ, 2004. \url{http://www.mccme.ru/prasolov}.}
\newcommand{\przse}{\bibitem[Pr07]{Pr07} * \emph{V. V. Prasolov.} Elements of homology theory. 2007, GSM 74, Amer. Math. Soc., Providence, RI.}
%Russian version: Moscow, MCCME, 2006,
\newcommand{\przseru}{\bibitem[Pr06]{Pr06} * \emph{Â. Â. Ïðàñîëîâ.} Ýëåìåíòû òåîðèè ãîìîëîãèé. Ì.: ÌÖÍÌÎ, 2006.}
% \url{http://www.mccme.ru/prasolov}.}
\newcommand{\psns}{\bibitem[PS96]{PS96} * \emph{V. V. Prasolov, A. B. Sossinsky } Knots, Links, Braids, and 3-manifolds. Amer. Math. Soc. Publ., Providence, R.I., 1996.}
%Russian version: \url{http://www.mccme.ru/prasolov}.}
\newcommand{\pszf}{\bibitem[PS05]{PS05} * \emph{Â. Â. Ïðàñîëîâ è Ì. Á. Ñêîïåíêîâ.}
Ðàìñååâñêàÿ òåîðèÿ çàöåïëåíèé, Ìàò. Ïðîñâåùåíèå. 2005. 9. Ñ.~108--115.}
\newcommand{\pszfen}{\bibitem[PS05]{PS05} * \emph{V. V. Prasolov and M.B. Skopenkov.}
Ramsey link theory, Mat, Prosvescheniye, 9 (2005), 108--115.}
\newcommand{\psoo}{\bibitem[PS11]{PS11} \emph{Y. Ponty and C. Saule.} A combinatorial framework for designing (pseudoknotted) RNA algorithms, Proc. of the 11th Intern. Workshop on Algorithms in Bioinformatics, WABI'11, 250--269.}
%\url{http://dl.acm.org/citation.cfm?id=2039967}
\newcommand{\pstz}{\bibitem[PS20]{PS20} \emph{S. Parsa and A. Skopenkov.} On embeddability of joins and their `factors', arXiv:2003.12285.}
%\bibitem[Pa20]{Pa20} S. Parsa, On the embeddability of $[3]*K$. arXiv:2001.06506.
\newcommand{\psszn}{\bibitem[PSS]{PSS} \emph{M. J. Pelsmajer, M. Schaefer and D. Stasi.} Strong Hanani-Tutte on the projective plane. SIAM J. Discrete Math., 23:3 (2009) 1317--1323.}
\newcommand{\pton}{\bibitem[PT19]{PT19} \emph{P. Pat\'ak and M. Tancer.} Embeddings of $k$-complexes into $2k$-manifolds. arXiv:1904.02404.}
\newcommand{\pw}{\bibitem[PW]{PW} \emph{I. Pak, S. Wilson}, G\lowercase{EOMETRIC REALIZATIONS OF POLYHEDRAL COMPLEXES}, \url{http://www.math.ucla.edu/~pak/papers/Fary-full31.pdf}.}
\newcommand{\razf}{\bibitem[RA05]{RA05} * \emph{J. L. Ram\'irez Alfons\'in.} Knots and links in spatial graphs: a survey. Discrete Math., 302 (2005), 225--242.}
\newcommand{\rep}{\bibitem[Rep]{Rep} Referee's report on the paper ``Some `converses' to intrinsic linking theorems', \url{https://www.mccme.ru/circles/oim/materials/ksreport.pdf}}
\newcommand{\rnoo}{\bibitem[RN11]{RN11} * \emph{R. L. Ricca, B. Nipoti.} Gauss' linking number revisited.
J. of Knot Theory and Its Ramif. 20:10 (2011) 1325--1343. \url{https://www.maths.ed.ac.uk/~v1ranick/papers/ricca.pdf} .}
\newcommand{\rrstz}{\bibitem[RRS]{RRS} * \emph{V. Retinskiy, A. Ryabichev and A. Skopenkov.}
Motivated exposition of the proof of the Tverberg Theorem (in Russian).
Mat. Prosveschenie, 27 (2021), 166--169. arXiv:2008.08361.}
\newcommand{\rssec}{\bibitem[RS68]{RS68} \emph{C. P. Rourke and B. J. Sanderson,} Block bundles II, Ann. of Math. (2), 87 (1968) 431--483.}
\newcommand{\rsst}{\bibitem[RS72]{RS72} * \emph{C. P. Rourke and B. J. Sanderson,}
\newblock Introduction to Piecewise-Linear Topology,
\newblock \emph{Ergebn.\ der Math.} 69, Springer-Verlag, Berlin, 1972.}
\newcommand{\rsstr}{\bibitem[RS72]{RS72} * \emph{Ê. Ï. Ðóðê è Á. Äæ. Ñàíäåðñîí.} Ââåäåíèå â êóñî÷íî-ëèíåéíóþ òîïîëîãèþ, Ìîñêâà. Ìèð. 1974.}
\newcommand{\rsns}{\bibitem[RS96]{RS96} * \emph{D. Repov\v s and A. B. Skopenkov.}
Embeddability and isotopy of polyhedra in Euclidean spaces,
%Òðóäû ÌÈÐÀÍ. 1996. 212;
Proc. of the Steklov Inst. Math. 1996. 212. P.~173-188.}
\newcommand{\rsne}{\bibitem[RS98]{RS98} \emph{D. Repov\v s and A. B. Skopenkov.}
A deleted product criterion for approximability of a map by embeddings, Topol. Appl. 1998. 87 P.~1-19.}
\newcommand{\rsnn}{\bibitem[RS99]{RS99} * \emph{D. Repov\v s and A. B. Skopenkov.} New results on embeddings of polyhedra and manifolds into Euclidean spaces,
%(in Russian), Uspekhi Mat. Nauk, 54:6 (1999) 61--109. English transl.:
Russ. Math. Surv. 54:6 (1999), 1149--1196.}
%\emph{Ä. Ðåïîâø è À. Ñêîïåíêîâ.}
%Íîâûå ðåçóëüòàòû î âëîæåíèÿõ ïîëèýäðîâ è ìíîãîîáðàçèé â åâêëèäîâû ïðîñòðàíñòâà, ÓÌÍ. 1999. 54:6. Ñ.~61-109.
\newcommand{\rsnnd}{\bibitem[RS99']{RS99'} * \emph{Ä. Ðåïîâø è À. Ñêîïåíêîâ.}
Êîëüöà Áîððîìåî è ïðåïÿòñòâèÿ ê âëîæèìîñòè, Òðóäû ÌÈÐÀÍ. 1999. 225. Ñ.~331-338.}
\newcommand{\rszz}{\bibitem[RS00]{RS00} \emph{D. Repov\v s and A. Skopenkov.} Cell-like resolutions of polyhedra by special ones, Colloq. Math. 2000. 86:2. P. 231--237.}
\newcommand{\rszzd}{\bibitem[RS00']{RS00'} * \emph{Ä. Ðåïîâø è À. Ñêîïåíêîâ.} Õàðàêòåðèñòè÷åñêèå êëàññû äëÿ íà÷èíàþùèõ, Ìàò. Ïðîñâåùåíèå. 2000. 4. Ñ.~151-176.}
\newcommand{\rszo}{\bibitem[RS01]{RS01} \emph{D. Repovs and A. Skopenkov.} On contractible $n$-dimensional compacta, non-embeddable into $\R^{2n}$, Proc. Amer. Math. Soc. 129 (2001) 627--628.}
\newcommand{\rszt}{\bibitem[RS02]{RS02} * \emph{Ä. Ðåïîâø è À. Ñêîïåíêîâ.} Òåîðèÿ ïðåïÿòñòâèé äëÿ íà÷èíàþùèõ,
Ìàò. Ïðîñâåùåíèå. 2002. 6. C.~60-77.}
\newcommand{\rszf}{\bibitem[RS04]{RS04} \emph{N. Robertson and P. Seymour.} Graph Minors. XX. Wagner's conjecture, J. of Comb. Theory, B, 92:2 (2004) 325--357.}
\newcommand{\rssnf}{\bibitem[RSS]{RSS95} \emph{D. Repov\v s, A. B. Skopenkov and E. V. \v S\v cepin.}
On uncountable collections of continua and their span, Colloq. Math. 1995. 69:2. P.~289-296.}
\newcommand{\rssnfd}{\bibitem[RSS']{RSS95'} \emph{D. Repov\v s, A. B. Skopenkov and E. V \v S\v cepin.}
On embeddability of $X\times I$ into Euclidean space, Houston J.~Math. 1995. 21. P.~199-204.}
\newcommand{\rssz}{\bibitem[RSS+]{RSSZ} * \emph{A. Rukhovich, A. Skopenkov, M. Skopenkov, A. Zimin},
Realizability of hypergraphs, \url{http://www.turgor.ru/lktg/2013/1/index.htm}.}
\newcommand{\rstnt}{\bibitem[RST']{RST93} \emph{N. Robertson, P. Seymour and R. Thomas}, Linkless embeddings of graphs in 3-space, Bull. of the Amer. Math. Soc., 21 (1993) 84--89.}
\newcommand{\rstno}{\bibitem[RST]{RST91} \emph{N. Robertson, P. Seymour and R. Thomas}, A survey of
linkless embeddings, Graph Structure Theory (Seattle, WA, 1991), Contemp. Math. 147, (1993) 125--136.}
%Amer. Math. Soc. Providence, RI,
%\bibitem[Ru]{Ru} \emph{Yu. B. Rudyak.} Piecewise linear structures on topological manifolds.
%\url{http://arxiv.org/abs/math/0105047}
\newcommand{\rwzl}{\bibitem[RWZ+]{RWZ+} \emph{Y. Ren, C. Wen, S. Zhen, N. Lei, F. Luo, D.X. Gu},
Characteristic class of isotopy for surfaces, J. Syst. Sci. Complex. 33 (2020) 2139--2156.}
\newcommand{\saeo}{\bibitem[Sa81]{Sa81} \emph{H. Sachs.} On spatial representation of finite graphs,
in: Finite and infinite sets, Colloq. Math. Soc. Janos Bolyai, North Holland, Amsterdam (37) 1981.}
\newcommand{\sano}{\bibitem[Sa91]{Sa91} \emph{K. S. Sarkaria.}
A one-dimensional Whitney trick and Kuratowski's graph planarity criterion, Israel J.~Math. 73 (1991), 79--89.
\url{http://kssarkaria.org/docs/One-dimensional.pdf}.}
\newcommand{\sanov}{\bibitem[Sa91g]{Sa91g} \emph{K. S. Sarkaria.} A generalized Van Kampen-Flores theorem, Proc. Amer. Math. Soc. 111 (1991), 559--565.}
\newcommand{\sant}{\bibitem[Sa92]{Sa92} \emph{K. S. Sarkaria.} Tverbergs theorem via number fields. Israel J. Math., 79:317320, 1992.}
\newcommand{\sann}{\bibitem[Sa99]{Sa99} O. Saeki {\em On punctured 3-manifolds in 5-sphere}, Hiroshima Math. J. 29 (1999) 255--272.}
%\moreref MR1704247 (2000h:57045)
\newcommand{\sazz}{\bibitem[Sa00]{Sa00} \emph{K. S. Sarkaria.} Tverberg partitions and Borsuk-Ulam theorems. Pacific J. Math., 196:1 (2000) 231--241.}
\newcommand{\sczf}{\bibitem[Sc04]{Sc04} \emph{T. Sch\"oneborn.} On the Topological Tverberg Theorem, arXiv:math/0405393.}
%http://arxiv.org/abs/math/0405393
\newcommand{\scot}{\bibitem[Sc13]{Sc13} * \emph{M. Schaefer.} Hanani-Tutte and related results. In Geometry --- intuitive, discrete, and convex, Bolyai Soc. Math. Stud., 24 (2013), 259--299.
\url{http://ovid.cs.depaul.edu/documents/htsurvey.pdf} }
%J\'anos Bolyai Math. Soc., Budapest, 2013.
\newcommand{\sctz}{\bibitem[Sc20]{Sc20} \emph{M. Schaefer.} The Graph Crossing Number and
its Variants: A Survey. The Electr. J. of Comb. (2020), DS21, \url{https://www.combinatorics.org/files/Surveys/ds21/ds21v5-2020.pdf}}
\newcommand{\scef}{\bibitem[Sc84]{Sc84} \emph{E.~V.~\v S\v cepin.} Soft mappings of manifolds, Russian Math. Surveys, 39:5 (1984).}
%, 209--224 (in Russian).
\newcommand{\shfs}{\bibitem[Sh57]{Sh57} \emph{A. Shapiro,} Obstructions to the embedding of a complex in a Euclidean space, I, The first obstruction, Ann. Math. 66 (1957), 256--269.}
%\url{http://www.maths.ed.ac.uk/~aar/papers/shapiro2.pdf}.}
\newcommand{\shen}{\bibitem[Sh89]{Sh89} * \emph{Þ. À. Øàøêèí,} Íåïîäâèæíûå òî÷êè, Ì., Íàóêà, 1989.}
\newcommand{\shoe}{\bibitem[Sh18]{Sh18} * \emph{S. Shlosman}, Topological Tverberg Theorem: the proofs and the counterexamples, Russian Math. Surveys, 73:2 (2018), 175182. arXiv:1804.03120.}
\newcommand{\sisn}{\bibitem[Si69]{Si69} \emph{K. Sieklucki.} Realization of mappings, Fund. Math. 1969. 65. P.~325-343.}
\newcommand{\sios}{\bibitem[Si16]{Si16} \emph{S. Simon,} Average-Value Tverberg Partitions via Finite Fourier Analysis, Israel J. Math., 216 (2016), 891-904, arXiv:1501.04612.}
%\url{http://arxiv.org/abs/1501.04612}
\newcommand{\sknf}{\bibitem[Sk94]{Sk94} \emph{À. Ñêîïåíêîâ.} Ãåîìåòðè÷åñêîå äîêàçàòåëüñòâî òåîðåìû
Íîéâèðòà îá óòîëùàåìîñòè 2-ìåðíûõ ïîëèýäðîâ, Math. Notes. 1995. 58:5. P.~1244-1247.}
% Ìàò. çàìåòêè. 1994. 56:2. Ñ.~94-98 English transl.
\newcommand{\skne}{\bibitem[Sk98]{Sk98} \emph{A. B. Skopenkov.} On the deleted product criterion for embeddability in $\R^m$, Proc. Amer. Math. Soc. 1998. 126:8. P.~2467-2476.}
\newcommand{\skzz}{\bibitem[Sk00]{Sk00} \emph{A. Skopenkov,} On the generalized Massey--Rolfsen invariant for link maps, Fund. Math. 165 (2000), 1--15.}
\newcommand{\skzt}{\bibitem[Sk02]{Sk02} \emph{A. Skopenkov,} On the Haefliger-Hirsch-Wu invariants for embeddings and immersions, Comment. Math. Helv. 77 (2002), 78--124.}
\newcommand{\skzth}{\bibitem[Sk03]{Sk03} \emph{M. Skopenkov,} Embedding products of graphs into Euclidean spaces,
Fund. Math. 179 (2003),~191--198, arXiv:0808.1199.}
\newcommand{\skzthd}{\bibitem[Sk03']{Sk03'} \emph{M. Skopenkov,} On approximability by embeddings of cycles in the plane, Topol. Appl. 134 (2003),~1--22, arXiv:0808.1187.}
\newcommand{\skzf}{\bibitem[Sk05]{Sk05} * \emph{A. Skopenkov,}
On the Kuratowski graph planarity criterion, Mat. Prosveschenie, 9 (2005), 116-128. arXiv:0802.3820.}
%À. Ñêîïåíêîâ Âîêðóã êðèòåðèÿ Êóðàòîâñêîãî ïëàíàðíîñòè ãðàôîâ, Ìàò. Ïðîñâåùåíèå. 2005. 9. Ñ.~116-128.
\newcommand{\skzs}{\bibitem[Sk05i]{Sk05i} \emph{A. Skopenkov,} A new invariant and parametric connected sum of embeddings, Fund. Math. 197 (2007) 253--269. arxiv:math/0509621.}
%Sk08
\newcommand{\skzei}{\bibitem[Sk05]{Sk05} \emph{A. Skopenkov,} A classification of smooth embeddings of
4-manifolds in 7-space, I, Topol. Appl., 157 (2010) 2094--2110. arXiv:math/0512594.}
\newcommand{\skze}{\bibitem[Sk06]{Sk06} * \emph{A. Skopenkov,} Embedding and knotting of manifolds in Euclidean spaces, London Math. Soc. Lect. Notes, 347 (2008) 248--342. arXiv:math/0604045.}
\newcommand{\skzsi}{\bibitem[Sk06']{Sk06'} \emph{A. Skopenkov,} A classification of smooth embeddings of 3-manifolds in 6-space, Math. Zeitschrift, 260:3 (2008) 647--672. arxiv:math/0603429.}
%Sk08p
\newcommand{\skozp}{\bibitem[Sk08]{Sk08} \emph{A. Skopenkov,} Embeddings of $k$-connected $n$-manifolds into
$\R^{2n-k-1}$. arxiv:math/0812.0263; earlier version published in Proc. Amer. Math. Soc., 138 (2010) 3377--3389.}
\newcommand{\skoz}{\bibitem[Sk10]{Sk10} * \emph{À. Ñêîïåíêîâ,} Âëîæåíèÿ â ïëîñêîñòü ãðàôîâ ñ âåðøèíàìè ñòåïåíè 4,
Ìàò. Ïðîñâåùåíèå, 21 (2017), arXiv:1008.4940.}
\newcommand{\skoo}{\bibitem[Sk11]{Sk11} \emph{M. Skopenkov,} When is the set of embeddings finite up to isotopy? Intern. J. Math. 26:7 (2015), 28 pp. arXiv:1106.1878.}
\newcommand{\sks}{\bibitem[Sk14]{Sk14} * \emph{A. Skopenkov,} Realizability of hypergraphs and Ramsey link theory, arXiv:1402.0658.}
%\url{http://arxiv.org/abs/1402.0658}
\newcommand{\skof}{\bibitem[Sk15]{Sk15} * \emph{À. Ñêîïåíêîâ,} Àëãåáðàè÷åñêàÿ òîïîëîãèÿ ñ ãåîìåòðè÷åñêîé òî÷êè çðåíèÿ, Ìîñêâà, ÌÖÍÌÎ, 2015 (1å èçäàíèå).}
\newcommand{\skofe}{\bibitem[Sk15]{Sk15} * \emph{A. Skopenkov,} Algebraic Topology From Geometric Viewpoint (in Russian), MCCME, Moscow, 2015 (1st edition). }
\newcommand{\skofel}{\bibitem[Sk15e]{Sk15e} * \emph{À. Ñêîïåíêîâ,} Àëãåáðàè÷åñêàÿ òîïîëîãèÿ
ñ ãåîìåòðè÷åñêîé òî÷êè çðåíèÿ, ýë. âåðñèÿ, \url{http://www.mccme.ru/circles/oim/home/combtop13.htm\#photo}}
%http://arxiv.org/abs/0808.1395,
\newcommand{\skotzr}{\bibitem[Sk20]{Sk20} * \emph{À. Ñêîïåíêîâ,} Àëãåáðàè÷åñêàÿ òîïîëîãèÿ ñ ãåîìåòðè÷åñêîé òî÷êè çðåíèÿ, Ìîñêâà, ÌÖÍÌÎ, 2020 (2å èçäàíèå).
×àñòü êíèãè: \url{http://www.mccme.ru/circles/oim/obstruct.pdf}}
\newcommand{\skotz}{\bibitem[Sk20]{Sk20} * \emph{A. Skopenkov,} Algebraic Topology From Geometric Viewpoint (in Russian), MCCME, Moscow, 2020 (2nd edition).
%Electronic version: \url{http://www.mccme.ru/circles/oim/home/combtop13.htm\#photo}
Part of the book: \url{http://www.mccme.ru/circles/oim/obstruct.pdf} .
Part of the English translation: \url{https://www.mccme.ru/circles/oim/obstructeng.pdf}. }
%Accepted for the English translation by Springer. Preprint of a part:
\newcommand{\skofp}{\bibitem[Sk15]{Sk15} \emph{A. Skopenkov,} Classification of knotted tori,
Proc. A of the Royal Soc. of Edinburgh, 150:2 (2020), 549-567. Full version: arXiv:1502.04470.}
%\url{https://doi.org/10.1017/prm.2018.141},
\newcommand{\skos}{\bibitem[Sk16]{Sk16} * \emph{A. Skopenkov,} A user's guide to the topological Tverberg Conjecture, arXiv:1605.05141v4. Abridged publishe version: Russian Math. Surveys, 73:2 (2018), 323--353.}
%Section 4 of the published version is available as {\it A. Skopenkov,}
%On van Kampen-Flores, Conway-Gordon-Sachs and Radon theorems, arXiv:1704.00300.
%https://doi.org/10.1070/rm9774
\newcommand{\skosd}{\bibitem[Sk16']{Sk16'} * \emph{A. Skopenkov,} Stability of intersections of graphs in the plane and the van Kampen obstruction, Topol. Appl. 240(2018) 259--269, arXiv:1609.03727.}
%https://doi.org/10.1016/j.topol.2018.02.029
\newcommand{\skosc}{\bibitem[Sk16c]{Sk16c} * \emph{A. Skopenkov,} Embeddings in Euclidean space: an introduction to their classification, to appear in Boll. Man. Atl. http://www.map.mpim-bonn.mpg.de/Embeddings\_in\_Euclidean\_space:\_an\_introduction\_to\_their\_classification}
\newcommand{\skosie}{\bibitem[Sk16e]{Sk16e} * \emph{A. Skopenkov,} Embeddings just below the stable range: classification, to appear in Boll. Man. Atl.
http://www.map.mpim-bonn.mpg.de/Embeddings\_just\_below\_the\_stable\_range:\_classification}
\newcommand{\skost}{\bibitem[Sk16t]{Sk16t} * \emph{A. Skopenkov,} 3-manifolds in 6-space, to appear in Boll. Man. Atl.
http://www.map.mpim-bonn.mpg.de/3-manifolds\_in\_6-space}
\newcommand{\skosf}{\bibitem[Sk16f]{Sk16f} * \emph{A. Skopenkov,} 4-manifolds in 7-space, to appear in Boll. Man. Atl. http://www.map.mpim-bonn.mpg.de/4-manifolds\_in\_7-space}
\newcommand{\skosh}{\bibitem[Sk16h]{Sk16h} * \emph{A. Skopenkov,} High codimension links, to appear in Boll. Man. Atl. \url{http://www.map.mpim-bonn.mpg.de/High_codimension_links}.}
\newcommand{\skosi}{\bibitem[Sk16i]{Sk16i} * \emph{A. Skopenkov,} Isotopy, submitted to Boll. Man. Atl.
\url{http://www.map.mpim-bonn.mpg.de/Isotopy}.}
\newcommand{\skose}{\bibitem[Sk17]{Sk17} \emph{A. Skopenkov,}
Eliminating higher-multiplicity intersections in the metastable dimension range. arXiv:1704.00143.}
\newcommand{\skosed}{\bibitem[Sk17v]{Sk17v} * \emph{A. Skopenkov,}
On van Kampen-Flores, Conway-Gordon-Sachs and Radon theorems, arXiv:1704.00300.}
\newcommand{\sk}{\bibitem[Sk17o]{Sk17o} \emph{A. Skopenkov,} On the metastable Mabillard-Wagner conjecture. arXiv:1702.04259.}
\newcommand{\skmos}{\bibitem[Sk17d]{Sk17d} \emph{M. Skopenkov}. Discrete field theory: symmetries and conservation laws, arXiv:1709.04788.}
\newcommand{\skoe}{\bibitem[Sk18]{Sk18} * \emph{A. Skopenkov.} Invariants of graph drawings in the plane.
Arnold Math. J., 6 (2020) 21--55; full version: arXiv:1805.10237.}
%DOI 10.1007/s40598-019-00128-5
%https://link.springer.com/article/10.1007/s40598-019-00128-5
\newcommand{\skoeo}{\bibitem[Sk18o]{Sk18o} * \emph{A. Skopenkov.} A short exposition of S. Parsa's theorems on intrinsic linking and non-realizability. Discr. Comp. Geom. 65:2 (2021), 584--585; full version: arXiv:1808.08363.}
%http://link.springer.com/10.1007/s00454-019-00158-y
%https://link.springer.com/epdf/10.1007/s00454-019-00158-y?sharing_token=
%XGdS8VlnK-pDzbEwtf3mBfe4RwlQNchNByi7wbcMAY6PJ4b9nP990kJl5DGh-51rxqVRnymDKW334OTc8lvTb8dMBR5PKi5EgDDh7R4xQBwahNi
%GoRrM0MA4cWoY55UZZGfGzKyrd9k-v04HAYod1Bqt4TGtvvU_Gv8j3-aCwKA%3D
\newcommand{\skona}{\bibitem[Sk19]{Sk19} * \emph{A. Skopenkov,} A short exposition of the Levine-Lidman example of spineless 4-manifolds, arXiv:1911.07330.}
\newcommand{\sktze}{\bibitem[Sk21m]{Sk21m} * \emph{A. Skopenkov.} Mathematics via Problems. Part 1: Algebra. Amer. Math. Soc., Providence, 2021. Preliminary version: \url{https://www.mccme.ru/circles/oim/algebra_eng.pdf}}
\newcommand{\sktz}{\bibitem[Sk20u]{Sk20u} * \emph{A. Skopenkov.} A user's guide to basic knot and link theory,
in Topology, Geometry, and Dynamics, Contemporary Mathematics, vol. 772, Amer. Math. Soc., Providence, RI, 2021, pp. 281--309.
%https://doi.org/10.1090/conm/772.
Russian version: Mat. Prosveschenie 27 (2021), 128--165. arXiv:2001.01472.}
\newcommand{\sktzo}{\bibitem[Sk20o]{Sk20o} \emph{A. Skopenkov.} On some results of S. Abramyan and T. Panov, arXiv:2005.11152.}
\newcommand{\sktzr}{\bibitem[Sk20e]{Sk20e} * \emph{A. Skopenkov.}
Extendability of simplicial maps is undecidable, arXiv:2008.00492.}
%Topological undecidability results for hypergraphs, Realizability of hypergraphs slides for talks,
%\url{https://www.mccme.ru/circles/oim/exalg_eng.pdf} .
\newcommand{\sktzd}{\bibitem[Sk21d]{Sk21d} * \emph{A. Skopenkov.}
On different reliability standards in current mathematical research, arXiv:2101.03745.
More often updated version: \url{https://www.mccme.ru/circles/oim/rese_inte.pdf}.}
\newcommand{\skd}{\bibitem[Sk]{Sk} * \emph{À. Ñêîïåíêîâ.} Àëãåáðàè÷åñêàÿ òîïîëîãèÿ ñ àëãîðèòìè÷åñêîé òî÷êè çðåíèÿ, %\linebreak
\url{http://www.mccme.ru/circles/oim/algor.pdf}.}
\newcommand{\skde}{\bibitem[Sk]{Sk} * \emph{A. Skopenkov.} Algebraic Topology From Algorithmic Standpoint, draft of a book, mostly in Russian,
%\linebreak
\url{http://www.mccme.ru/circles/oim/algor.pdf}.}
%LaTeX Error: Command \ss already defined Or name \end... illegal, see p.192 of the manual.
\newcommand{\skon}{\bibitem[Skw]{Skw} * \emph{A. Skopenkov.} Whitney trick for eliminating multiple intersections, slides for talks at St Petersburg, Brno, Kiev, Moscow, \url{https://www.mccme.ru/circles/oim/eliminat_talk.pdf}.}
\newcommand{\skl}{\bibitem[EEF]{EEF} * {\it Proposed by D. Eliseev, A. Enne, M. Fedorov, A. Glebov, N. Khoroshavkina, E. Morozov, A. Skopenkov, R. \v Zivaljevi\'c.}
A user's guide to knot and link theory, \url{https://www.turgor.ru/lktg/2019/3} .}
\newcommand{\skr}{\bibitem[Skr]{Skr} * \emph{A. Skopenkov.} Realizability of hypergraphs, slides for talks, \url{https://www.mccme.ru/circles/oim/algor1_beamer.pdf}.}
\newcommand{\skt}{\bibitem[Skt]{Skt} * \emph{A. Skopenkov.} Transparent anonymous peer review,
\linebreak
\url{https://www.mccme.ru/circles/oim/home/transp_peer_review.htm} .}
\newcommand{\rslktg}{\bibitem[KRR+]{RRSl} * Towards higher-dimensional combinatorial geometry, presented by
E. Kogan, V. Retinskiy, E. Riabov and A. Skopenkov, \url{https://www.turgor.ru/lktg/2020/3/index.html}.}
\newcommand{\sm}{\bibitem[Sm]{Sm} S. Smirnov.}
\newcommand{\sper}{\bibitem[Sp]{Sp} * Sperner's lemma defeats the rental harmony problem, \url{https://www.youtube.com/watch?v=7s-YM-kcKME}.}
\newcommand{\sset}{\bibitem[SS83]{SS83} \emph{Å. Â. Ùåïèí, Ì. À. Øòàíüêî.} Ñïåêòðàëüíûé êðèòåðèé âëîæèìîñòè êîìïàêòîâ â åâêëèäîâû ïðîñòðàíñòâà, Òðóäû Ëåíèíãðàäñêîé Ìåæäóíàðîäíîé Òîïîëîãè÷åñêîé êîíôåðåíöèè. Ë.: Íàóêà, 1983. Ñ.~135-142.}
\newcommand{\ssnt}{\bibitem[SS92]{SS92} \emph{J.~Segal and S.~Spie\.z.} Quasi embeddings and embeddings of polyhedra in $\R^m$, Topol. Appl., 45 (1992) 275--282.}
\newcommand{\sszt}{\bibitem[SS03]{SS03} \emph{F. W. Simmons and F. E. Su.}
Consensus-halving via theorems of Borsuk-Ulam and Tucker, Math. Social Sciences 45 (2003) 1525. \url{https://www.math.hmc.edu/~su/papers.dir/tucker.pdf}.}
\newcommand{\ssot}{\bibitem[SS13]{SS13} \emph{M. Schaefer and D. Stefankovi\v c.} Block additivity of $\Z_2$-embeddings. In Graph drawing, volume 8242 of Lecture Notes in Comput. Sci., 185--195.
Springer, Cham, 2013. \url{http://ovid.cs.depaul.edu/documents/genus.pdf}}
\newcommand{\sssne}{\bibitem[SSS]{SSS} \emph{J. Segal, A. Skopenkov and S. Spie\. z.}
Embeddings of polyhedra in $\R^m$ and the deleted product obstruction, Topol. Appl. 1998. 85. P.~225-234.}
\newcommand{\sstnf}{\bibitem[SST95]{SST95} \emph{R. S. Simon, S. Spie\. z and H. Toru\'nczyk.}
T\lowercase{HE EXISTENCE OF EQUILIBRIA IN CERTAIN GAMES, SEPARATION FOR FAMILIES OF CONVEX FUNCTIONS
AND A THEOREM OF BORSUK-ULAM TYPE}, Israel J. Math 92 (1995) 1--21.}
\newcommand{\sstzt}{\bibitem[SST02]{SST02} \emph{R. S. Simon, S. Spie\. z and H. Toru\'nczyk.}
E\lowercase{QUILIBRIUM EXISTENCE AND TOPOLOGY IN SOME REPEATED GAMES WITH INCOMPLETE INFORMATION},
Trans. Amer. Math. Soc. 354:12 (2002) 5005-5026.}
\newcommand{\stez}{\bibitem[ST80]{ST80} * {\it H.~Seifert and W.~Threlfall.}
A textbook of topology, v~89 of {\em Pure and Applied Mathematics}.
Academic Press, New York-London, 1980.}
%Translated from the German edition of 1934 by Michael A. Goldman.
\newcommand{\stzs}{\bibitem[ST07]{ST07} * \emph{À. Ñêîïåíêîâ è À. Òåëèøåâ.}
È âíîâü î êðèòåðèè Êóðàòîâñêîãî ïëàíàðíîñòè ãðàôîâ, Ìàò. Ïðîñâåùåíèå, 11 (2007), 159--160.}
\newcommand{\stzse}{\bibitem[ST07]{ST07} * \emph{A. Skopenkov and A. Telishev}, Once again on the Kuratowski graph planarity criterion, Mat. Prosveschenie, 11 (2007), 159-160. arXiv:0802.3820.}
\newcommand{\stos}{\bibitem[ST17]{ST17} \emph{A. Skopenkov and M. Tancer,}
Hardness of almost embedding simplicial complexes in $\R^d$, Discr. Comp. Geom., 61:2 (2019), 452--463. arXiv:1703.06305.}
\newcommand{\stwh}{\bibitem[SW]{SW} * \url{http://www.map.mpim-bonn.mpg.de/Stiefel-Whitney_characteristic_classes}}
\newcommand{\sz}{\bibitem[SZ05]{SZ} \emph{T. Sch\"oneborn and G. Ziegler}, The Topological Tverberg Theorem and Winding Numbers, J. Comb. Theory, Ser. A, 112:1 (2005) 82--104, arXiv:math/0409081.}
\newcommand{\szno}{\bibitem[Sz91]{Sz91} \emph{A.~Sz\"ucs,} On the cobordism groups of immersions and embeddings,
Math. Proc. Camb. Phil. Soc., 109 (1991) 343--349.}
%\bibitem[SZ]{SZ} \emph{A. Skopenkov and A. Zimin},
%Realizability of hypergraphs in Euclidean spaces.
%\linebreak
%\url{http://www.mccme.ru/circles/oim/exalong.pdf}
\newcommand{\ta}{\bibitem[Ta]{Ta} * Handbook of Graph Drawing and Visualization. ed. by R. Tamassia, CRC Press, 2016.}
%\url{https://cs.brown.edu/~rt/gdhandbook/}.}
\newcommand{\tazz}{\bibitem[Ta00]{Ta00} \emph{K. Taniyama,} Higher dimensional links in a simplicial complex embedded in a sphere, Pacific Jour. of Math. 194:2 (2000), 465-467.}
\newcommand{\theo}{\bibitem[Th81]{Th81} * \emph{C.~Thomassen,} Kuratowski's theorem, J.~Graph. Theory 5 (1981), 225--242.}
\newcommand{\tooo}{\bibitem[To11]{To11} \emph{Tonkonog D.} Embedding 3-manifolds with boundary into closed 3-manifolds, Topol. Appl. 158 (2011), 1157-1162. arXiv:1003.3029.}
%http://arxiv.org/abs/1003.3029.
\newcommand{\tsbzf}{\bibitem[TSB]{TSB} \emph{D. M. Thilikos, M. Serna and H. L. Bodlaender},
Cutwidth I: A linear time fixed parameter algorithm, J. of Algorithms, 56:1 (2005), 1--24.}
%https://www.sciencedirect.com/science/article/pii/S0196677405000167
\newcommand{\tsbzfd}{\bibitem[TSB05']{TSB05'} \emph{D. M. Thilikos, M. Serna and H. L. Bodlaender},
Cutwidth II: , J. of Algorithms, 56:1 (2005), 25--49.}
%https://www.sciencedirect.com/science/article/pii/S0196677405000179
%c.tuffley@massey.ac.nz
%https://arxiv.org/pdf/1112.4558.pdf
\newcommand{\umse}{\bibitem[Um78]{Um78} \emph{B. Ummel.} The product of nonplanar complexes does not imbed in 4-space, Trans. Amer. Math. Soc., 242 (1978) 319--328.}
%\bibitem[Vd95]{Vd95} \emph{A. Vdovina.} Construction of orientable Wicks forms and estimation of their number, %Comm. Algebra 23 (1995), 3205-3233.
%\bibitem[Vd96]{Vd96} \emph{A. Vdovina.} On the number of nonorientable Wicks forms in a free group,
%Proc. Royal Soc. Edinb. Sect. A 126 (1996), 113-116.
%\bibitem[Vd97]{Vd97} \emph{A. Vdovina.} Product of commutators in free products,
%Internat. J. Algebra Comput. 7 (1997), 471-485.
\newcommand{\val}{\bibitem[Val]{Val} Valknut, \url{https://en.wikipedia.org/wiki/Valknut}}
%http://katlas.math.toronto.edu/wiki/File:Noeuds-3carres_en_translation.png
\newcommand{\vi}{\bibitem[Vi]{Vi} * \emph{O. Viro.}
Some integral calculus based on Euler characteristic, Lect. Notes in Math. 1346.}
\newcommand{\vizt}{\bibitem[Vi02]{Vi02} * \emph{Ý. Á. Âèíáåðã.} Êóðñ àëãåáðû. Ìîñêâà. Ôàêòîðèàë Ïðåññ. 2002.}
\newcommand{\vizteng}{\bibitem[Vi02]{Vi02} * \emph{E. B. Vinberg.} A Course in Algebra. Graduate Studies in Mathematics, vol. 56. 2003.}
\newcommand{\vinhzs}{\bibitem[VINH07]{VINH07} * \emph{Î. ß. Âèðî, Î. À. Èâàíîâ, Í. Þ. Íåöâåòàåâ è Â. Ì. Õàðëàìîâ.}
Ýëåìåíòàðíàÿ òîïîëîãèÿ, ÌÖÍÌÎ. 2007.}
\newcommand{\vktt}{\bibitem[vK32]{vK32} \emph{E.~R.~van~Kampen}, Komplexe in euklidischen R\"aumen, Abh. Math. Sem. Hamburg, 9 (1933) 72--78; Berichtigung dazu, 152--153.
English translation by Tu T$\hat a$m Ngu$\tilde{\hat e}$n-Phan:
\url{https://sites.google.com/site/tutamnguyenphan/van_Kampen.pdf}}
\newcommand{\kafo}{\bibitem[vK41]{vK41} \emph{E. R. van Kampen,} Remark on the address of S. S. Cairns,
in Lectures in Topology, 311--313, University of Michigan Press, Ann Arbor, MI, 1941.}
\newcommand{\vo}{\bibitem[Vo96]{vo96} \emph{A. Yu. Volovikov,} On a topological generalization of the Tverberg theorem. Math. Notes 59:3 (1996), 324--326.}
\newcommand{\vopns}{\bibitem[Vo96v]{Vo96v} \emph{A. Yu. Volovikov,} On the van Kampen-Flores Theorem.
Math. Notes 59:5 (1996), 477--481.}
\newcommand{\vznt}{\bibitem[VZ93]{VZ93} \emph{A. Vu\v ci\'c and R. T. \v Zivaljevi\'c}, Note on a conjecture of Sierksma, Discr. Comput. Geom. 9 (1993), 339-349.}
\newcommand{\vzzn}{\bibitem[VZ09]{VZ09} \emph{S. T. Vre\'cica and R. T. \v Zivaljevi\'c}, Chessboard complexes
indomitable, J. of Comb. Theory, Ser. A 118:7 (2011), 2157--2166. arXiv:0911.3512.}
\newcommand{\walst}{\bibitem[Wa62]{Wa62} \emph{C.~T.~C.~Wall}, Classification of $(n-1)$-connected $2n$-manifolds, Ann. of Math., 75 (1962) 163--189.}
%https://www.maths.ed.ac.uk/~v1ranick/papers/n-1con2n.pdf
\newcommand{\wallss}{\bibitem[Wa67]{Wa67} \emph{C.~T.~C.~Wall.} Classification problems in differential topology, IV, Thickenings, Topology 1966. 5. P. 73--94.}
\newcommand{\waldss}{\bibitem[Wa67m]{Wa67m} \emph{F. Waldhausen.} Eine Klasse von 3-dimensional Mannigfaltigkeiten, I. Invent. Math. 1967. 3. P.~308-333.}
\newcommand{\walsz}{\bibitem[Wa70]{Wa70} \emph{C. T. C. Wall,} Surgery on compact manifolds,
1970, Academic Press, London.}
\newcommand{\wess}{\bibitem[We67]{We67} \emph{C.~Weber.} Plongements de poly\`edres dans le domain metastable, Comment. Math. Helv. 42 (1967), 1--27.}
\newcommand{\whit}{\bibitem[Wl]{Wl} * \url{https://en.wikipedia.org/wiki/Whitehead_link}}
\newcommand{\winum}{\bibitem[Wn]{Wn} * \url{https://en.wikipedia.org/wiki/Winding_number}}
\newcommand{\wrss}{\bibitem[Wr77]{Wr77} \emph{P. Wright.} Covering 2-dimensional polyhedra by 3-manifolds spines.
Topology. 16 (1977), 435--439.}
\newcommand{\wufe}{\bibitem[Wu58]{Wu58} \emph{W. T. Wu.} On the realization of complexes in a euclidean space (in Chinese): I, Sci Sinica, 7 (1958) 251--297; II, Sci Sinica, 7 (1958) 365--387; III, Sci Sinica, 8 (1959) 133--150.}
\newcommand{\wufn}{\bibitem[Wu59]{Wu59} \emph{W.~T.~Wu.} On the isotopy of a finite complex in Euclidean space, I, II, Science Record, N.S. 3:8 (1959) 342--347, 348--351.}
\newcommand{\wusf}{\bibitem[Wu65]{Wu65} * \emph{W. T. Wu.} A Theory of Embedding, Immersion and Isotopy of Polytopes in an Euclidean Space. Peking: Science Press, 1965.}
% Peking - OK? Da. A. S.
\newcommand{\yann}{\bibitem[Ya99]{Ya99} \emph{Z. Yang.} Computing Equilibria and Fixed Points: The Solution of Nonlinear Inequalities, Kluwer, Springer Science + Business Media, 1990.}
\newcommand{\zesz}{\bibitem[Ze60]{Ze60} \emph{E. C. Zeeman}, Unknotting spheres in five dimensions, Bull. Amer. Math. Soc. 66 (1960) 198.
\linebreak
\url{https://www.ams.org/journals/bull/1960-66-03/S0002-9904-1960-10431-4/S0002-9904-1960-10431-4.pdf}}
\newcommand{\z}{\bibitem[Ze]{Z} * \emph{E. C. Zeeman}, A Brief History of Topology, UC Berkeley, October 27, 1993, On the occasion of Moe Hirsch's 60th birthday, \url{http://zakuski.utsa.edu/~gokhman/ecz/hirsch60.pdf}.}
\newcommand{\zioz}{\bibitem[Zi10]{Zi10} * \emph{D. \v Zivaljevi\'c}, Borromean and Brunnian Rings
\url{http://www.rade-zivaljevic.appspot.com/borromean.html}.}
\newcommand{\zioo}{\bibitem[Zi11]{Zi11} * \emph{G. M. Ziegler}, 3N Colored Points in a Plane, Notices of the Amer. Math. Soc., 58:4 (2011), 550-557.}
%\url{http://www.ams.org/notices/201104/rtx110400550p.pdf}.}
\newcommand{\zot}{\bibitem[Zi13]{Z13} \emph{A. Zimin.} Alternative proofs of the Conway-Gordon-Sachs Theorems, arXiv:1311.2882.}
%http://arxiv.org/abs/1311.2882
\newcommand{\zss}{\bibitem[ZSS]{ZSS} * Ýëåìåíòû ìàòåìàòèêè â çàäà÷àõ: ÷åðåç îëèìïèàäû è êðóæêè ê ïðîôåññèè
Ñáîðíèê ïîä ðåäàêöèåé À. Çàñëàâñêîãî, À. Ñêîïåíêîâà è Ì. Ñêîïåíêîâà. Èçä-âî ÌÖÍÌÎ, 2018.
\url{http://www.mccme.ru/circles/oim/materials/sturm.pdf}.}
\newcommand{\zu}{\bibitem[Zu]{Zu} \emph{J. Zung.} A non-general-position Parity Lemma,
\url{http://www.turgor.ru/lktg/2013/1/parity.pdf}.}
%. There is an improved unpublished version.