%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%Latex file, 15 pages: B.M. Zupnik. Geometry of solutions of
%N=2 SYM-theory in harmonic superspace
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\newcommand{\be}{\begin{equation}}\newcommand{\ee}{\end{equation}}
\newcommand{\bea}{\begin{eqnarray}}\newcommand{\eea}{\end{eqnarray}}
\newcommand{\nn}{\nonumber}\newcommand{\p}[1]{(\ref{#1})}
 \newcommand{\lb}[1]{\label{#1}}
\newcommand\bt{\bar{\theta}}\newcommand\q{\quad}\newcommand\qq{\qquad}
\newcommand\tp{\theta^+}
\newcommand\btp{\bar{\theta}^+}
\newcommand\tpm{\theta^{\mu+}}
\newcommand\tpb{\theta^{\beta+}}
\newcommand\tpa{\theta^{\alpha+}}
\newcommand\btpm{\bar{\theta}^{\dot{\mu}+}}
\newcommand\btpa{\bar{\theta}^{\dot{\alpha}+}}
\newcommand\btpb{\bar{\theta}^{\dot{\beta}+}}
\newcommand\btpn{\bar{\theta}^{\dot{\nu}+}}
\newcommand\tm{\theta^-}
\newcommand\btm{\bar{\theta}^-}
\newcommand\tma{\theta^{\alpha-}}
\newcommand\tmb{\theta^{\beta-}}
\newcommand\tmm{\theta^{\mu-}}
\newcommand\btma{\bar{\theta}^{\dot{\alpha}-}}
\newcommand\btmb{\bar{\theta}^{\dot{\beta}-}}
\newcommand\btmm{\bar{\theta}^{\dot{\mu}-}}
\newcommand\btmn{\bar{\theta}^{\dot{\nu}-}}
\newcommand\Tp{\Theta^+}
\newcommand\Tpb{\Theta^{\beta+}}
\newcommand\Tpa{\Theta^{\alpha+}}
\newcommand\hT{\hat\Theta}
\newcommand\hTp{\hat\Theta^+}
\newcommand\hTpa{\hat{\Theta}^{\dot{\alpha}+}}
\newcommand\hTpb{\hat{\Theta}^{\dot{\beta}+}}
\newcommand\hTpn{\hat{\Theta}^{\dot{\nu}+}}
\newcommand\Tm{\Theta^-}
\newcommand\hTm{\hat{\Theta}^-}
\newcommand\Tma{\Theta^{\alpha-}}
\newcommand\Tmb{\Theta^{\beta-}}
\newcommand\Tmm{\Theta^{\mu-}}
\newcommand\hTma{\hat{\Theta}^{\dot{\alpha}-}}
\newcommand\hTmb{\hat{\Theta}^{\dot{\beta}-}}
\newcommand\hTmm{\hat{\Theta}^{\dot{\mu}-}}
\newcommand\hTmn{\hat{\Theta}^{\dot{\nu}-}}
\newcommand\pada{\partial_{\alpha\dot{\alpha}}}
\newcommand\padb{\partial_{\alpha\dot{\beta}}}
\newcommand\pbda{\partial_{\beta\dot{\alpha}}}
\newcommand\pbdb{\partial_{\beta\dot{\beta}}}
\newcommand\pmdn{\partial_{\mu\dot{\nu}}}
\newcommand\ada{{\alpha\dot{\alpha}}}
\newcommand\adb{{\alpha\dot{\beta}}}
\newcommand\bdb{{\beta\dot{\beta}}}
\newcommand\bda{{\beta\dot{\alpha}}}
\newcommand\ab{{\alpha\beta}}
\newcommand\ba{{\beta\alpha}}
\newcommand\bg{{\beta\gamma}}
\newcommand\gb{{\gamma\beta}}
\newcommand\ag{{\alpha\gamma}}
\newcommand\ga{{\gamma\alpha}}
\newcommand\ar{{\alpha\rho}}
\newcommand\gr{{\gamma\rho}}
\newcommand\rg{{\rho\gamma}}
\newcommand\sr{{\sigma\rho}}
\newcommand\ra{{\rho\alpha}}
\newcommand\rb{{\rho\beta}}
\newcommand\br{{\beta\rho}}
\newcommand\mdn{{\mu\dot{\nu}}}
\newcommand\ap{{\alpha +}}
\newcommand\dap{{\dot{\alpha}+}}
\newcommand\am{{\alpha-}}
\newcommand\bm{{\beta-}}
\newcommand\dam{{\dot{\alpha}-}}
\newcommand\dbm{{\dot{\beta}-}}
\newcommand\da{{\dot{\alpha}}}
\newcommand\db{{\dot{\beta}}}
\newcommand\dg{{\dot{\gamma}}}
\newcommand\dr{{\dot{\rho}}}
\newcommand\cA{{\cal A}}
\newcommand\cB{{\cal B}}
\newcommand\cV{{\cal V}}
\newcommand\cD{{\cal D}}
\newcommand\bcD{\bar{\cal D}}
\newcommand\hcD{\hat{\cal D}}
\newcommand\hcA{\hat{\cal A}}
\newcommand\hcW{\hat{\cal W}}
\newcommand\cW{{\cal W}}
\newcommand\T{\mbox{Tr}}
\newcommand\s{\scriptscriptstyle}
\newcommand\A{{\s A}}
\newcommand\B{{\s B}}
\newcommand\C{{\s C}}
\newcommand\R{{\s R}}
\newcommand\M{{\s M}}
\newcommand\N{{\s N}}
\newcommand{\pp}{{\s ++}}
\newcommand{\m}{{\s --}}
\newcommand{\0}{{\s 0}}
\newcommand{\2}{{\s 2}}
\newcommand{\3}{\s 3}
\newcommand{\5}{\s 5}
\newcommand{\pT}{\s(+3)}
\newcommand{\Dp}{D^{\pp}}
\newcommand{\Dm}{D^{\m}}
\newcommand{\Vp}{V^{\pp}}
\newcommand{\Vm}{V^{\m}}
\newcommand{\Dep}{\Delta^{\pp}}
\newcommand{\Dem}{\Delta^{\m}}
\newcommand{\dpp}{\partial^{\pp}}
\newcommand{\dm}{\partial^{\m}}
\newcommand{\DP}{D^+}
\newcommand{\DM}{D^-}
\newcommand{\DPa}{D^+_\alpha}
\newcommand{\DPb}{D^+_\beta}
\newcommand{\DMa}{D^-_\alpha}
\newcommand{\DMb}{D^-_\beta}
\newcommand{\pM}{\partial^-}
\newcommand{\pPa}{\partial^+_\alpha}
\newcommand{\pPb}{\partial^+_\beta}
\newcommand{\pMa}{\partial^-_\alpha}
\newcommand{\pMb}{\partial^-_\beta}
\newcommand{\bDP}{\bar{D}^+}
\newcommand{\bDPa}{\bar{D}^+_{\dot{\alpha}}}
\newcommand{\bDPb}{\bar{D}^+_{\dot{\beta}}}
\newcommand{\bDM}{\bar{D}^-}
\newcommand{\bDMa}{\bar{D}^-_{\dot{\alpha}}}
\newcommand{\bDMb}{\bar{D}^-_{\dot{\beta}}}
\newcommand{\bpP}{\bar{\partial}^+}
\newcommand{\bpPa}{\bar{\partial}^+_{\dot{\alpha}}}
\newcommand{\bpPb}{\bar{\partial}^+_{\dot{\beta}}}
\newcommand{\bpM}{\bar{\partial}^-}
\newcommand{\bpMa}{\bar{\partial}^-_{\dot{\alpha}}}
\newcommand{\bpMb}{\bar{\partial}^-_{\dot{\beta}}}
\newcommand{\cDp}{\cD^+}
\newcommand{\cDm}{\cD^-}
\newcommand{\cDpa}{\cD^+_\alpha}
\newcommand{\cDpb}{\cD^+_\beta}
\newcommand{\cDma}{\cD^-_\alpha}
\newcommand{\cDmb}{\cD^-_\beta}
\newcommand{\hcDp}{\hcD^+}
\newcommand{\hcDm}{\hcD^-}
\newcommand{\hcDpa}{\hcD^+_\da}
\newcommand{\hcDpb}{\hcD^+_\db}
\newcommand{\hcDma}{\hcD^-_\da}
\newcommand{\hcDmb}{\hcD^-_\db}
\newcommand\tmA{\theta^{\A-}}
\newcommand\tmB{\theta^{\B-}}
\newcommand\tpA{\theta^{\A+}}
\newcommand\tpB{\theta^{\B+}}
\newcommand{\DPA}{D^+_\A}
\newcommand{\DPB}{D^+_\B}
\newcommand{\DMA}{D^-_\A}
\newcommand{\DMB}{D^-_\B}
\newcommand{\ppA}{\partial^+_\A}
\newcommand{\ppB}{\partial^+_\B}
\newcommand{\pmA}{\partial^-_\A}
\newcommand{\pmB}{\partial^-_\B}
\newcommand{\pAB}{\partial^+_{\A\B}}
\newcommand{\pMN}{\partial^+_{\M\N}}
\newcommand{\pBM}{\partial^+_{\B\M}}
\newcommand{\xAB}{x^{\A\B}}
\newcommand{\XAB}{X^{\A\B}}
\newcommand\dza{d\zeta^{\s(-4)}}


\documentclass[12pt]{article}
\def\theequation{\arabic{section}.\arabic{equation}}
\def\thesection {\arabic{section}.}
\topmargin=-0.5cm \textheight=235mm\textwidth=16cm
\oddsidemargin=-0.2cm\evensidemargin=-0.2cm
%\def\baselinestretch{1.1}
\begin{document}
%\large
\renewcommand{\thefootnote}{\fnsymbol{footnote}}
\begin{flushright}
{ hep-th/0107012 }
\end{flushright}
\vspace{2cm}

\begin{center}
{\large\bf   GEOMETRY OF SOLUTIONS OF
N=2 SYM-THEORY IN HARMONIC SUPERSPACE}
\vspace{1cm} \\

{\bf B.M. Zupnik}\\
\vspace{0.5cm}

{\it Joint Institute for Nuclear Research, Bogoliubov Laboratory
of Theoretical Physics, Dubna, Moscow Region,
 141980, Russia. E-mail: zupnik@thsun1.jinr.ru }

\end{center}

\begin{abstract}
In harmonic superspace, the classical equations of motion of $D=4, N=2$
supersymmetric Yang-Mills theory for Minkowski and Euclidean spaces are
analyzed. We study dual superfield representations of equations and
subsidiary conditions corresponding to classical SYM-solutions with
different symmetries. In particular, alternative superfield constructions
of self-dual  and static  solutions are described in the framework
of the harmonic approach.
\end{abstract}

\setcounter{equation}0
\section{Introduction}
The off-shell superfield constraints of the $N=2 $ super-Yang-Mills theory
in the Minkowski $D=(3,1)$ space has been solved in the framework of
harmonic superspace (HSS) using the auxiliary  coordinates of the coset
space $S_2\sim SU_\A(2)/U_\A(1)$ where $SU_\A(2)$  is the corresponding
automorphism group \cite{GIK1}-\cite{Zu2}. An analogous harmonic formalism
can be applied to reformulate the off-shell $N=2$ supersymmetric models in
the Euclidean $D=4$ space. We shall use the notation with the subscripts
$HSS_M$ or $HSS_E$ for harmonic-superspace structures over Minkowski or
Euclidean  spaces, respectively. Auxiliary harmonic coordinates are used
in  covariant conditions of the Grassmann (G-) analyticity. The
unconstrained superfields of the $N=2$ theories live in the G-analytic
superspaces with the reduced odd dimension $4$.

Harmonic variables connected with coset $SU_\R(2)/U_\R(1)$ of the
subgroup of the Euclidean rotation group have been used to study self-dual
solutions of the YM- and  SYM-theories \cite{GIOS2,OO,DO}. These harmonic
superspaces have the reduced values of even and odd dimensions, and the
corresponding analytic superfields parametrize moduli spaces of self-dual
solutions.

The  superfield equations of motion in the most general harmonic
formalism can  be effectively used for the analysis of the geomery of
classical $N=2$ SYM-solutions. In particular, one can consider  dual
transformations of the $N=2$ SYM-superfield variables in the harmonic
superspace which allow us to formulate unusual representations of the
equations of motion and simple gauge conditions \cite{Zu3,Zu4}. It is
important to stress that the dual change of variables transforms the part
of SYM-constraints and equations in the harmonic formalism  to linear
restrictions on HSS-connection  $\Vm$, then the zero-curvature relation
between the harmonic connections can be interpreted as a dynamical
equation. Dual representations of the HSS-superfield equations allow us to
describe symmetry properties of different classes of partial solutions.

It is interesting to compare the alternative HSS-descriptions of $4D$
self-dual SYM-solutions or solutions of $D<4$ SYM-equations. We analyze
the  $(t=0)$ reduction of the $HSS_M$-formalism and formulate the static
3-dimensional superfield BPS-conditions. It is shown that the
supersymmetry of the static self-dual SYM-equations is equivalent to the
corresponding  3-dimensional subgroup of the Euclidean supersymmetry
with 8 supercharges.

The $HSS_M$-equations of the $N=2$ SYM-theory are considered in Sect. 2.
We discuss the off-shell formalism with independent harmonic connections
and dually-equivalent constructions of the superfield equations of motion.
One analyzes the conditions of the G-analyticity for connections in this
formalism. The superfield equations of motion can be transformed to
relations between the G-analytic functions arising in decompositions of
the non-analytic connection $\Vm$.

Sect. 3 is devoted to the discussion of the static $3D$ reduction of the
$N=2$ superfield SYM-equations. The time component of the $4D$ superfield
connection $A_t$ becomes the new $3D$-scalar superfield in this limit. We
consider the $3D$-superfield BPS-type relation between $A_t$ and the
superfield strengths $W$ and $\bar{W}$ which is equivalent to the 2nd
order differential constraint for  connection $\Vm$. It is shown that this
2nd order constraint generates the standard 4th order constraint for the
same connection  and all solutions of the superfield BPS-equation satisfy
the SYM-equation.

In Sect. 4 we study the $4D$ Euclidean version of the harmonic $HSS_E$
formalism. The chiral superfield self-duality condition in this approach
can also be interpreted as the 2nd order constraint on $\Vm$. The
alternative bridge representation and the nilpotent gauge condition
for the bridge superfield are used  to analyze the $HSS_E$ self-dual
solutions.

The comparison of these harmonic constructions with the analogous
self-dual $N=2$ solutions in the alternative $SU_\R(2)/U_\R(1)$ formalism
is discussed in Sect. 5. This version of the harmonic approach transforms
the super-self-duality condition to the specific Grassmann-bosonic
analyticity of the on-shell harmonic connection. We consider the
identification of the harmonic variables for different $SU(2)$ subgroups
and the simple Ansatz for the self-dual harmonic connection in the gauge
group $SU(2)$ which yield the explicit construction of solutions in
quadratures \cite{Zu}.

The static self-dual solutions are discussed in Sect. 6. We discuss the
relations between the conjugated spinor coordinates in the static limit of
the $4D$  superspace $HSS_M $and the corresponding pseudoreal spinor
coordinates of the $3D$ reduction of the Euclidean  superspace $HSS_E$.

The basic formulae of the harmonic-superspace approach are reviewed
in Appendix. We describe conjugation rules for different versions of
harmonic superspaces which are very important in the  SYM-theory.


\setcounter{equation}0
\section{\lb{B}Harmonic-superspace representations  of N=2 SYM-equations}

The harmonic $SU_\A(2)/U_\A(1)$ transform has been used first to solve
the off-shell  constraints for  superfield connections $A^k_a(z)$ of
the $N=2, D=(3,1)$ SYM-theory \cite{GIK1,GIOS}
\be
u^+_k(D^k_a+A^k_a)\equiv\nabla^+_a= e^{-v}D^+_a e^v
\ee
where $a=(\alpha,\da)$ are the $SL(2,C)$ indices, $k$ is the 2-component
index of automorphism group $SU(2)_\A$, $u^+_k$ are the auxiliary
harmonic variables (see Appendix) and $v(z,u)$ is the {\it bridge} matrix.
This transform connects formally different off-shell representations of
the gauge group in the central basis (CB) and the analytic basis (AB).

The basic off-shell harmonic superfield of the  SYM-theory is the
connection $\Vp=e^v\Dp e^{-v}$ for the harmonic covariant derivative of
the analytic basis
\be
\nabla^\pp=\Dp+\Vp~,\q \delta \Vp=-\Dp\lambda-[\Vp,\lambda]~.\lb{B1}
\ee
The prepotential $\Vp$ and the AB-gauge parameters satisfy
the G-analyticity conditions
\be
(\DPa, \bDPa)(\Vp, \lambda)=0~.\lb{B2}
\ee

The G-analytic superfields are described by the unconstrained functions
of the analytic coordinates $\zeta$ \p{F5} and harmonics $u^\pm_i$.
The second harmonic connection $\Vm$ satisfies the harmonic zero-curvature
 equation
\be
\Dp \Vm-\Dm\Vp+[\Vp,\Vm]=0~,\lb{B3}
\ee
which can be solved explicitly in terms of $\Vp$-serias.

In the analytic basis, the spinor and vector connections $A_\M$ as well as
the superfield strengths $W$ and $\bar{W}$ can be written directly via
spinor derivatives of connection $\Vm$ \cite{Zu2}
\bea
&&A^+_\alpha=\bar{A}^+_\da=0~,\q A^-_\alpha=-\DPa\Vm~,\q
\bar{A}^-_\da=-\bDPa\Vm~,\\
&&A_\adb\equiv i\DPa\bar{A}^-_\db\equiv-i\bDPb A^-_\alpha=-i\DPa\bDPb\Vm~,
\lb{B5b}\\
&& \bar{W}={1\over2}D^{\alpha+}A^-_\alpha=-(\DP)^2\Vm~,\q
W=-{1\over2}\bDP_\da\bar{A}^{\da-}=(\bDP)^2\Vm~.\lb{B5}
\eea

The nonlinear harmonic-superfield equation of motion for the prepotential
$\Vp$ can be reformulated as a linear differential condition on the
composed connection $\Vm(\Vp)$
\be
(\DP)^2(\bDP)^2\Vm=0~.\lb{B6}
\ee
 Eq.\p{B3} is treated as  a kinematic solvable equation in the standard
SYM-formalism with basic superfield $\Vp$. The alternative $HSS_M$
formalism with the independent off-shell superfield $\Vm$ has been
considered in Refs.\cite{Zu3}. This approach is analogous to the
first-order formalism for the ordinary Yang-Mills theory, since the
independent superfield $\Vm$ contains off-shell an infinite number of
auxiliary fields including the fields of dimension $-2:~ F, F_{mn}$ and
$F_m$. (Note that the changes of basic HSS variables $\Vp\rightarrow \Vm$
or $\Vp\rightarrow v$ can be interpreted as dual transformations in the
SYM-theory. We hope that the rich  structure of the dual
transformations in HSS will be useful for the explicit constructions
of different classical solutions.)

The G-analyticity of $\Vp$ is equivalent to the  nonlinear
analyticity  equation for $\Vm$
\be
[\nabla^-_a,\nabla^\m]=D^-_a\Vm+\Dm D^+_a\Vm+[\Vm,D^+_a\Vm]=0~.\lb{B18}
\ee

The linear constraint \p{B6} for the independent superfield $\Vm$ can be
readily solved  \cite{Zu3}
\be
\Vm=\DPa B^{\alpha(-3)}-\bDPa\bar{B}^{\da(-3)}\lb{B31b}.
\ee

In the picture with independent SYM superfield variables
$B^{\alpha(-3)}$ and $\Vp$, the basic dynamical equation is
\be
\nabla^\pp(\DPa B^{\alpha(-3)}-\bDPa\bar{B}^{\da(-3)})=\Dm\Vp\lb{B32b}~.
\lb{Beq}
\ee

Acting by  operators $\DPa ,\bDPa$  on both sides of this
equation one can obtain the following relations:
\bea
&&\nabla^\pp(\bDP)^2\DPa B^{\alpha(-3)}=0~,
 \lb{B33b}\\
&&\nabla^\pp
[(\bDP)^2\DPa\bar{B}_\da^{(-3)}-(\DP)^2\bDPa B_\alpha^{(-3)}]=i\pada\Vp~.
\eea

Decomposition of Eq.\p{B32b} in terms of all spinor coordinates
gives  the component SYM-equations; however, it is also useful  to analyze
the partial decomposition of this superfield equation in terms of $\tma$
and $\btma$. Let us define the G-analytic components of  superfield $\Vm$
\bea
&\bar\phi=-(\DP)^2\Vm |~,\q \phi=(\bDP)^2\Vm |~,\q B_\adb=\bDPb\DPa\Vm |
~,&\lb{B7}\\
&\lambda_\alpha^+=-\DPa(\bDP)^2\Vm |~,\q\bar\lambda_\da^+=
-\bDPa(\DP)^2\Vm |~,\q F^\pp=(\DP)^2(\bDP)^2\Vm |~,&\nn
\eea
where the symbol $|$ means $\theta^-=\bar\theta^-=0$.

The full supersymmetry transformation $\delta_\epsilon \Vm=0$ yields
the corresponding transformations of the analytic components
\bea
&&\delta_\epsilon \phi=u^-_k\epsilon^{\alpha k}\lambda_\alpha^+~,\q
\delta_\epsilon B_\adb=u^-_k\epsilon_\alpha^k\bar\lambda_\db^+
+u^-_k\bar\epsilon_\db^k\lambda_\alpha^+~,\lb{B9}\\
&&\delta_\epsilon\lambda_\alpha=-u^-_k\epsilon_\alpha^k F^\pp~,\q
\delta_\epsilon F^\pp=0~.
\eea

Using the gauge transformation $\delta \Vm=-\Dm\lambda-[\Vm,\lambda]$,
one can obtain the nilpotent gauge condition for the harmonic
connection
\bea
&&\cV^\m=(\tm)^2\bar\phi-(\btm)^2\phi+\tma\btmb B_\adb+
(\tm)^2\btma\bar\lambda^+_\da+(\btm)^2\tma\lambda^+_\alpha\nn\\
&& +
(\tm)^2(\btm)^2F^\pp~.\lb{B9b}
\eea

The geometric superfields have the following form in this gauge:
\bea
&&W=\phi-\tma\lambda^+_\alpha-(\tm)^2F^\pp~,\lb{gauge1}\\
&&\bar{W}=\bar\phi+\btma\bar\lambda^+_\da+(\btm)^2F^\pp~,\lb{gauge2}\\
&&A_\adb=-iB_\adb+i\tm_\alpha\bar\lambda^+_\db+i\btm_\db\lambda^+_\alpha
-i\tm_\alpha\btm_\db F^\pp\lb{gauge3}~.
\eea

The constraint  $F^\pp=0$ gives the following  analytic equations
of motion equivalent to Eq.\p{B32b}
\bea
&&\nabla^\pp\phi=\tpa\lambda^+_\alpha~,\q(\nabla^\pp)^2\phi=0~,\lb{B17b}
\\
&&\nabla^\pp B_\adb=-i\padb\cV^\pp+\theta^+_\alpha\bar\lambda^+_\db+
\bar\theta^+_\db\lambda^+_\alpha~,\lb{B17c}\\
&&\nabla^\pp\lambda^+_\alpha=0
~.\lb{B30b}
\eea

The component solutions for the on-shell $N=2$ analytic superfields are
\bea
&&\phi=\varphi(x)+\tpa u^-_k\psi^k_\alpha(x)~,\\
&&B_\adb=a_\adb(x)+\theta^+_\alpha u^-_k\bar\psi_\db^k(x)+
\bar\theta^+_\db u^-_k\psi^k_\alpha(x)~,\\
&&\cV^\pp=(\tp)^2\bar\varphi(x)-(\btp)^2\varphi(x)+\tpa\btpb a_\adb(x)-
\tpa(\btp)^2 u^-_k\psi^k_\alpha(x)\nn\\
&&-\btpa(\tp)^2 u^-_k\bar\psi^k_\da(x)~.\lb{D1}
\eea

The  $N=2$ equations of motion fot these components follow from
G-analytic superfield equations. Of course, the main purpose of the
harmonic approach is to find nonstandard superfield methods of solving the
SYM-equations which cannot be completely reduced to the analysis of
component equations. We hope that a comparison of dual superfield
representations of  SYM-solutions will be useful for the search of unusual
symmetry properties which cannot be seen directly in the component
representation.

\setcounter{equation}0
\section{\lb{C}Static superfield equations}

Three-dimensional monopole and dyon solutions play an important role in
modern non-perturbative methods of the quantum $N=2$ SYM-theory
\cite{SW,Ga}. The harmonic analysis of the monopole Yang-Mills solutions
has been considered in Ref.\cite{OO}. We shall analyze the alternative
harmonic-superfield constructions of the static $N=2$ SYM-solutions and
their relations with the $4D$ self-dual solutions.

Let us consider the nonrelativistic representation of the coordinates in
the $D=(3,1),N=2$ harmonic superspace based on the static group $SO(3)$
\bea
&&x^\adb_\A\rightarrow iy^\alpha_\beta+\delta^\alpha_\beta
 t_\A~,\q y^\alpha_\alpha=0~,\q y^\ab=\varepsilon^\ar y^\beta_\rho
~,\lb{C1}\\
&&\padb~\rightarrow~ -i\partial^\alpha_\beta+\delta_\alpha^\beta
\partial_t~,\q\partial_\gamma^\sigma y^\alpha_\beta=2\delta^\sigma_\beta
\delta^\alpha_\gamma-\delta^\alpha_\beta\delta^\sigma_\gamma~,\\
&&\bar\theta^{\da\pm}~\rightarrow~\bar\theta^\pm_\alpha~,\q
\bar\theta^{\alpha\pm}=\varepsilon^\ab\bar\theta^\pm_\beta~,\\
&&\bar\partial^\pm_\da~\rightarrow~\bar\partial^{\alpha\pm}~,\q
\bar\partial^{\alpha\pm}\bar\theta_\beta^\mp=\delta^\alpha_\beta~.
\eea

We shall use the following conjugation rules:
\bea
&&(y^\alpha_\beta)^\dagger=-y_\alpha^\beta~,\q
(y^\ab)^\dagger=y_\ab~,\q(t_\A)^\dagger=t_\A~,\\
&&(\theta^{\alpha\pm})^\dagger= \bar\theta_\alpha^\pm~,\q
(\theta_\alpha^\pm)^\dagger= -\bar\theta^{\alpha\pm}~,\q
(\varepsilon^\ab)^\dagger=-\varepsilon_\ab~,\\
&&(\bar\theta^{\alpha\pm})^\dagger= \theta_\alpha^\pm~,\q
(\bar\theta_\alpha^\pm)^\dagger= -\theta^{\alpha\pm}~,\lb{C2}\\
&&(D_\alpha^\pm)^\dagger= \bar{D}^{\alpha\pm}~,\q
(\bar{D}^{\alpha\pm})^\dagger=-D_\alpha^\pm~,\q
(\bar{D}_\alpha^\pm)^\dagger=D^{\alpha\pm}~.
\eea
Note that changes of the position of dotted $SL(2,C)$ indices after the
static reduction is connected with a convention of the conjugation of the
$SU(2)$-spinors which transforms upper indices to lower ones. The time
reduction $(t=0)$ transforms 4-dimensional harmonic superfields to the
3-dimensional Euclidean  superfields which are covariant with respect to
the reduced supersymmetry with 8 supercharges.

The corresponding representation of the  4-vector connection is
\be
A_\adb~\Rightarrow~A^\beta_\alpha+\delta^\beta_\alpha A_t~,\q
(A^\beta_\alpha)^\dagger=-A_\beta^\alpha~,\q(A_t)^\dagger=-A_t~,
\lb{C3b}
\ee
where the time component of the connection  becomes covariant with
respect to residual gauge static transformations
\bea
&&A^\beta_\alpha=i(\bar{D}^{\beta+}\DPa-{1\over2}\delta^\beta_\alpha
(\bar{D}^{+}\DP) )\Vm~,\\
&&A_t={i\over2}(\bar{D}^{+}\DP) \Vm~,
\q\delta_\lambda A_t=[\lambda, A_t]~.\lb{C4}
\eea

It seems useful to construct the manifestly supersymmetric
generalization of the Bogomolnyi equation for the $N=2$ gauge theory
in order to analyze the symmetry properties of the corresponding
monopole solutions. In addition to the $t=0$ reduction of the HSS-equation
\p{B3}, we propose to consider the following superfield BPS-type relation:
\be
S\equiv-2irA_t+p(W+\bar{W})=0~,\lb{C6}
\ee
where $r$ and $p$ are some real parameters. Note that this condition
breaks  the $R$-symmetry $W\rightarrow e^{i\rho}W$.

We shall interpret \p{C6} as a linear constraint on $\Vm$
\be
S= [r(\bDP \DP)+p(\bDP)^2-p(\DP)^2]\Vm =0~,
\ee
which gives also the spinor condition
\be
\left(-r(\DP)^2\bar{D}^+_\alpha+p(\bDP)^2\DPa\right)\Vm=0~.\lb{C7}
\ee

For the case $r\neq p$ , the last spinor relation is not self-adjoint,
so one can obtain the strong restrictions
\be
\DPa(\bDP)^2\Vm=\DPa W=0~\Rightarrow~\nabla^\m\DPa W=\nabla^-_\alpha W=0
~.\lb{gen}
\ee
It is evident that the corresponding  covariantly constant solution
$W=const$ preserves all 8 supercharges \cite{IKZ}.  The case $r=p$
corresponds to the self-adjoint  relation for 3D spinors
\be
\lambda^\alpha_k=\bar\lambda^\alpha_k=\varepsilon^\ab\varepsilon_{kl}
(\lambda^\beta_l)^\dagger~.
\ee
The corresponding self-dual static solutions  will be discussed in
Sect. 6.

\setcounter{equation}0
\section{\lb{D}Euclidean $N=2$ SYM-equations}

We shall analyze the superfield constraints of the $N=2$ SYM-theory in the
Euclidean superspace
\bea
&&\{\nabla^k_\alpha,\nabla^l_\beta\}=\varepsilon_\ab\varepsilon^{kl}
\cW~,\lb{E1}\\
&&\{\hat\nabla^k_\da,\hat\nabla^l_\db\}=\varepsilon_{\da\db}
\varepsilon^{kl}\hcW~,\lb{E2}\\
&&\{\nabla^k_\alpha,\hat\nabla^l_\db\}=\varepsilon^{kl}\nabla_\adb~,
\lb{E3}
\eea
where  $\cW$ and $\hcW$ are independent real superfield strengths and
indices $\alpha, \beta; \da, \db$ and $k, l$ describe, respectively, the
2-spinor representations of the Euclidean group $SU_{\s L}(2)\times
SU_\R(2)$ and the automorphism group $SU_\A(2)$.

Superspace analogues of the Euclidean self-duality equations have been
considered in Refs.\cite{Se}-\cite{Si}. We shall analyze the $N=2$
self-dual solutions in the framework of the Euclidean version of the
harmonic superspace $HSS_E$ (see Appendix). Underline that conjugation
rules of the  $N=2$ superspace coordinates and derivatives are
essentially different for the cases of the Euclidean group $SO(4)$ and the
Lorentz group $SO(3,1)$.

The $SU_\A(2)/U_\A(1)$ harmonic-superspace description of the Euclidean
$N=2$ SYM-theory is  similar to the corresponding formalism in $HSS_M$,
in particular, one can  use the same notation $\Vp, \Vm$ for the Euclidean
harmonic connections. The Euclidean spinor and vector connections of the
analytic basic have the following form:
\bea
&&\cA^+_\alpha=\hcA^+_\da=0~,\q \cA^-_\alpha=-\cDpa\Vm~,\q
\hcA^-_\da=-\hcDpa\Vm~,\lb{E4}\\
&&\cA_\adb\equiv \cDpa\hcA^-_\db\equiv\cDpa\hcDpb\Vm~.\lb{E5}
\eea

The reality of the  Euclidean superfield strengths is evident in $HSS_E$
\be
\cW=(\cD^+)^2\Vm~,\q\hcW=(\hcD^+)^2\Vm~.\lb{E6}
\ee

The system of self-duality equations has the
following form in the harmonic-superspace:
\be
\cD^\pp\Vm+[\Vp,\Vm]=\cD^\m\Vp~,\q(\cD^+)^2\Vm=0~.\lb{E7}
\ee
The manifest solution of the 2nd constraint
\be
\Vm=\cDpa \cB^{\alpha(-3)}~,\lb{E8}
\ee
yields the simple relation between $\cB^{\alpha(-3)}$ and $\Vp$
\be
\cDpa\nabla^\pp\cB^{\alpha(-3)}=\cD^\m\Vp~.
\ee

The self-duality equation is equivalent to the relation
\be
\cDpa\cA^-_\beta=0~.\lb{E9}
\ee

Using the equation $[\nabla^\m,\nabla^-_\alpha]=0$ one can derive the
chirality relations
\be
\{\nabla^\pm_\alpha,\nabla^\pm_\beta\}=0~.\lb{E10}
\ee

Let us reformulate now the self-duality condition in the bridge
representation of the harmonic connection \cite{GIK1}
\be
\Vm\equiv e^v\cD^\m e^{-v}~.\lb{E10b}
\ee
The  condition \p{E9} for  $\Vm$ corresponds to the partial analyticity
condition for the self-dual bridge matrix
\be
\cD^+_\alpha v=0~\Rightarrow~-\cD^+_\alpha \left(e^v\cD^\m e^{-v}\right)
=e^v\cD^-_\alpha e^{-v}~.
\ee

Comparing representations \p{E8} and \p{E10b} one can obtain the relation
\be
\cB^{\alpha(-3)}={1\over2}\Theta^{\alpha-} e^v\cD^\m e^{-v}+
{1\over2}(\Theta^-)^2 e^v\cD^{\alpha-} e^{-v}~.
\ee

Harmonic projections of the central-basis connections
have the following form in this representation:
\bea
&&A_\alpha^\pm=0~,\q \hat{A}_\da^+(v)=e^{-v}\hcD_\da^+e^v~,\lb{vrep}\\
&&\hat{A}_\da^-=\cD^\m\hat{A}_\da^+(v)~,\q A_\ada(v)=\cD^-_\alpha
\hat{A}_\da^+(v)~.\nn
\eea

In the bridge  representation, the system of self-duality equations \p{E7}
is equivalent to the single equation
\be
\hcD^+_\da \left( e^v\Dp e^{-v}\right)\equiv\hcD^+_\da \Vp(v)=0~.
\ee

Let us introduce the nilpotent gauge condition for the bridge matrix
\be
v=\hT^{\da-}b^+_\da+(\hTm)^2b^\pp~,\q v^2=-(\hTm)^2b^{\da+}b^+_\da~,
\lb{sdgauge}
\ee
where $b^+_\da$ and $b^\pp$ are G-analytic matrix coefficients.
Note that the analogous nilpotent bridge representation has been used
recently in the $N=3$ SYM-theory \cite{NZ}.

Let us compose  harmonic connection $\Vp(v)$
\bea
&&e^v\Dp e^{-v}=-\hTpa b^+_\da+\hTma\left(-\Dp b^+_\da-\hT^+_\da b^\pp+
{1\over2}\hTpb\{b^+_\da,b^+_\db\}\right)\\
&&+(\hTm)^2\left(\Dp b^\pp+{1\over2}\{b^{\da+},\Dp b^+_\da\}+
\hTpa[b^+_\da,b^\pp]+{1\over2}\hTpa\{b^+_\da,b^{\db+}b^+_\db\}\right),\nn
\eea
where we have used the identity
\be
\{b^+_\da,b^{\db+}b^+_\db\}={1\over3}[b^{\db+},\{b^+_\da,b^+_\db\}]~.
\ee

The dynamical analyticity equation
\be
 \Vp(v)=-\hTpa b^+_\da
\ee
yields the following analytic relations:
\bea
&&\Dp b^+_\da+\hT^+_\da b^\pp-
{1\over2}\hTpb\{b^+_\da,b^+_\db\}=0~,\lb{hsd2}\\
&&\Dp b^\pp+{1\over2}\{b^{\da+},\Dp b^+_\da\}+
\hTpa[b^+_\da,b^\pp]+{1\over6}\hTpa[b^{\db+},\{b^+_\da,b^+_\db\}]=0~.
\nn
\eea

 Let us use the intermediate decomposition
of (2,2)-analytic superfields in $\hTpa$
\bea
&&b^+_\da=\beta^+_\da+\hTpb C_{\da\db}+(\hTp)^2\eta^-_\da~,\\
&&b^\pp=B^\pp+\hTpa\gamma^+_\da+(\hTp)^2B~,
\eea
where all coefficients are polynomials in $\Tpa$. The (2,2)-equations
for $b^+_\da$ and $b^\pp$ yield (2,0)-equations for these coefficients,
for instance,
\bea
&&\dpp \beta^+_\da=0~,\q\dpp B^\pp=0~,\lb{hom}\\
&&\dpp C_{\da\db}+\Tpb\pbdb\beta^+_\da+\varepsilon_{\da\db}B^\pp
-{1\over2}\{\beta^+_\da,\beta^+_\db\}
=0~.
\eea

The linear harmonic equations for $\beta^+_\da$ and $B^\pp$ can be solved
explicitly. The inhomogeneous linear harmonic equation for $C_{\da\db}$
contains the composed source calculated at the previous stage, so it is
also solvable, in principle. The inhomogeneous harmonic equations for
other (2,0)-coefficients $\eta^-_\da, \gamma^+_\da$ and $B$ can be derived
and solved analogously. It should be stressed that self-dual equations
\p{E7} and \p{hsd2} are convenient for the search of dimensionally reduced
solutions.

\setcounter{equation}0
\section{\lb{E}Harmonic constructions of self-dual solutions}

The alternative harmonic formalism for the  self-dual SYM-solutions has
been considered in Ref.\cite{DO}. This formalism harmonizes one of the
space  groups: $SU_\R(2)$ acting on the indices $\da$.

In order to compare alternative harmonic approaches to the $N=2$
self-dual equations we shall consider the non-covariant procedure
of identification $SU_R(2)=SU_A(2)$ for solutions of these equations.
Let us identify also the corresponding spinor indices $\da\equiv k,
\db\equiv l\ldots$ The $N=2$ superspace coordinates have the following
form in this notation:
\be
z^M=(y^{\alpha k}, \Theta^\alpha_k, \hT^k_l)~.
\ee

Using the single set of the $SU(2)/U(1)$ harmonics
\be
u^A_i\equiv \varepsilon^{AB}u_{Bi}~,\q A, B=\pm
\ee
one can consider the harmonic projections of central, analytic and chiral
coordinates, correspondingly:
\bea
&&y^{\alpha \A}=u^\A_ky^{\alpha k}~,\q Y^{\alpha \A}=y^{\alpha \A}-
\Theta^{+\alpha}\hT^{-\A}-\Theta^{-\alpha}\hT^{+\A}\\
&&y^{\alpha \A}_r=y^{\alpha \A}+
\Theta^{+\alpha}\hT^{-\A}-\Theta^{-\alpha}\hT^{+\A}\\
&&\Theta^\alpha_\A\equiv\varepsilon_{\A\B}\Theta^{\B\alpha}=u^k_\A
\Theta^\alpha_k~,\\
&&\hT^\B_\A\equiv\varepsilon_{\A\C}\hT^{\B\C}=u^l_\A u^\B_k\hT^k_l~,
\eea

The corresponding projections of the flat derivatives are
\bea
&&\partial_{\alpha\A}~,\q \cD_\alpha^\A~,\q\hcD^\A_\B~,\\
&&\{\cD_\alpha^\A,\hcD^\C_\B\}=-\varepsilon^{\A\C}\partial_{\alpha\B}.
\eea

It the central basis, one can obtain the following representations of the
$N=2$ covariant self-dual derivatives
\be
\nabla_\alpha^\pm=\cD^\pm_\alpha~,\q\hat\nabla_-^\pm=h^{-1}\hcD_-^\pm h~,
\q h^{-1}\partial_{\alpha-} h~,\lb{altrep}
\ee
where $h(z,u)$ is the chiral self-dual bridge
\be
\cD^\pm_\alpha h=0~.
\ee

The harmonic connection for the self-dual bridge is the basic
chiral-analytic prepotential of this formalism
\bea
&&h\cD^\pp h^{-1}=v^\pp~,\\
&&(\partial_{\alpha-}, \cD^\pm_\alpha, \hcD^\pm_-)v^\pp=0~.\lb{sdhs}
\eea
Stress that self-dual solutions possess the combined  analyticity. The
self-dual prepotential $v^\pp$ can be treated as unconstrained matrix
function of coordinates $y^{\alpha+}, u^\pm_i$ and $\hT^+_\pm$ which
parametrizes the general $N=2$ self-dual solution.

In the gauge group $SU(2)$, the simple solvable Ansatz for the self-dual
prepotential can be choosen
\be
(v^\pp)_i^k=u^{+k}u^+_i b^0 +(u^{+k}u^-_i+u^{-k}u^+_i)b^\pp~,
\ee
where $b^0$ and $b^\pp$ are real chiral-analytic functions. The bridge
matrix for this Ansatz can be calculated via harmonic quadratures
\cite{Zu}. In the simple example of this parametrization with $b^0=0$ and
\be
b^\pp=Y^{\alpha+}Y^{\beta+}\rho_\ab+\hT^+_+Y^{\alpha+}u^+_l
\lambda_\alpha^l
\ee
the solution depends on constant tensor and spinor coefficients.

It is not difficult to relate self-dual representations for different
methods of harmonization of superfield equations
\be
e^{-v}\hcD^+_-e^v=h^{-1}\hcD^+_-h~,\q \cD^-_\alpha(e^{-v}\hcD^+_-e^v)=
h^{-1}\partial_{\alpha-}h~.
\ee

\setcounter{equation}0
\section{\lb{G}Self-dual static $N=2$ solutions}

Consider now the subsidiary conditions in BPS-equations \p{C6} $r= p$ then
these equations can be transformed  to the following simple constraint:
\bea
&&-2i A_t+W+\bar{W}\equiv \left((\bDP\DP)+(\bDP)^2
-(\DP)^2\right)\Vm\nn\\&&
=-{1\over2}(D^{\alpha+}-\bar{D}^{\alpha+})(D_\alpha^+ -\bar{D}_\alpha^+)
\Vm=0~.\lb{bps1}
\eea
One can consider also the equivalent 2nd order constraints on $\Vm$ using
the transformation $\DPa\rightarrow e^{i\rho/2}\DPa$.

It is convenient to introduce the  new pseudoreal spinor coordinates
of the static harmonic superspace
\bea
&\Theta^{\alpha\pm}\equiv{1\over\sqrt{2}}(\theta^{\alpha\pm}+
\bar\theta^{\alpha\pm})~,\q\hT^{\alpha\pm}\equiv{1\over\sqrt{2}}
(\theta^{\alpha\pm}-\bar\theta^{\alpha\pm})~,&
\lb{rot3}\\
&(\Theta^{\alpha\pm})^\dagger =\Theta_\alpha^\pm~,\q
(\hT^{\alpha\pm})^\dagger =-\hT_\alpha^\pm~.&
\eea

The corresponding transformed  spinor derivatives
\bea
&&\cD^\pm_\alpha\equiv {1\over\sqrt{2}}(D^\pm_\alpha-\bar{D}^\pm_\alpha
)~,\q(\cD^\pm_\alpha)^\dagger=-\cD^{\alpha\pm}~,\lb{rot1}\\
&&\hcD^\pm_\alpha=-{1\over\sqrt{2}}(D^\pm_\alpha+\bar{D}^\pm_\alpha)~,\q
(\hcD^\pm_\alpha)^\dagger=\hcD^{\alpha\pm}~.
\eea
 have the following algebra:
\bea
&&\{\cD^+_\alpha,\hcD^-_\beta\}=-\{\hcD^+_\alpha,\cD^-_\beta\}=
-\partial_\ab~,\lb{stalg}\\
&&\{\cD^+_\alpha,\cD^-_\beta\}=
\{\hcD^+_\alpha,\hcD^-_\beta\}=0~.
\eea

Consider the $3D$-covariant representation of the Euclidean
superspace coordinates  \p{eucl1}
\be
y^\adb~\Rightarrow~y^\ab+\varepsilon^\ab y_4~.
\ee

Stress that the  algebra of spinor derivatives \p{stalg} arises also in
the dimensional reduction $y_4=0$ of the Euclidean $N=2$ harmonic
superspace (see Appendix). It is clear that the transformation \p{rot3}
connects the equivalent $3D$ subspaces of the Minkowski and Euclidean
types of harmonic superspaces. Thus, the (3,1)-equation \p{bps1} is
equivalent to the $3D$ limit of the Euclidean $N=2$ self-duality equation
\p{E7}
\be
(\cD^+)^2\Vm=0~.
\ee

The component static self-dual solutions con be obtained in the following
gauge:
\bea
&&\cV^\pp={1\over2}\hT^{\alpha+}\hT^+_\alpha c(x)+{i\over2}
\Theta^{\alpha+}\hT^+_\alpha a_t(x)+i\Theta^{\alpha+}\hT^{\beta+}a_\ab(x)
\nn\\
&&-\hT^{\alpha+}\hT^+_\alpha \Theta^{\beta+}u_k^-\Psi^k_\beta(x)~,
\eea
where all fields are Hermitian
\be
(c, a_t, a_\ab, \Psi^k_\alpha)^\dagger=(c, a_t, a^\ab, \Psi_k^\alpha)~.
\ee

One can derive the component $3D$ self-dual equations
\bea
&&F_\ab=-\nabla_\ab a_t~,\\
&&\nabla^\ab\nabla_\ab\,c=2[a_t,[a_t,c]]-\{\Psi^\alpha_k,\Psi^k_\alpha\}~,
\\
&&\nabla_\bg\Psi^{\gamma k}+{i\over2}[a_t,\Psi^k_\beta]=0~,
\eea
where
\be
\nabla_\ab\equiv\partial_\ab-i[a_\ab,]~,\q
F_\ab=\partial_\ar a^\rho_\beta+\partial_\br a^\rho_\alpha
-i[a_\ar,a_\beta^\rho].
\ee
Note that the static gauge field strength and field $a_t$ are connected
by the self-dual Bogomolnyi equation.

The superfield analysis of the static self-duality equations
can be made by analogy with the analysis of Euclidean $4D$ self-dual
equations in Sect. 4. The static limit of the bridge representation is
\be
A_\alpha^\pm=0~,\q\hat{A}_\alpha^+(v)=e^{-v}\hcD_\alpha^+e^v~,\lb{vrepst}
\ee
where $v$ is the static self-dual bridge $\cD^+_\alpha v=0$.

By analogy with the 4D self-duality equation one can use
the formal identification of all groups $SU(2)$ in the $3D$ Euclidean
harmonic superspace
\be
y^\ab\rightarrow y^{ik},\q\Theta^\alpha_l\rightarrow \Theta^k_l,\q
\hT^\alpha_l\rightarrow \hT^k_l~.
\ee

It is clear that one can use harmonic projections of the
superspace coordinates
\bea
&&y^{\A\B}_r=u^\A_ku^\B_ly^{kl}_r~,\\
&&\Theta^\B_\A=u^k_\A u_l^\B
\Theta^l_k~,\q\hT^\B_\A=u^l_\A u^\B_k\hT^k_l~.
\eea

The covariant derivative $\nabla^l_k$ are flat in the chiral
self-dual representation. The corresponding bridge representation
of the $3D$ self-dual covariant derivatives has the following form:
\be
h^{-1}\frac{\partial}{\partial y^{--}_r}h~,\qq
h^{-1}\frac{\partial}{\partial \hT^-_\A}h~.
\ee

The prepotential $v^\pp=h\Dp h^{-1}$ for these solutions depends on the
(1+2) analytic coordinates $y^\pp_r$ and $\hT^+_\A$ only. It is clear that
these solutions can be interpreted as the dimensionally reduced self-dual
$4D$ solutions \p{sdhs}.
\vspace{0.5cm}

{\bf Acknowledgement.}
Author is grateful to E. Ivanov and J. Niederle for discussions.
This work is supported in part by grants RFBR 99-02-18417,
RFBR-DFG-99-02-04022, INTAS-2000-254 and NATO PST.CLG 974874,
and by the  Votruba-Blokhintsev program in LTP JINR.

\appendix
 \def\theequation{A.\arabic{equation}}
\setcounter{equation}0
\section{\lb{F} Appendix}

{\bf Harmonic-superspace coordinates in $D=(3,1), N=2$ superspace}\\

The $SU(2)/U(1)$ harmonics  \cite{GIK1} parametrize the sphere $S^2$. They
form an $SU(2)$ matrix $u^\pm_i$ and are defined modulo $U(1)$.

The $SU(2)$-invariant harmonic derivatives act on the harmonics
\be
[\dpp ,\dm ]=\partial^0
~,\q [\partial^0 ,\partial^{\s\pm\pm}]=\pm 2\partial^{\s\pm\pm}~.
\ee

The special $SU(2)$-covariant conjugation of harmonics preserves the
$U(1)$-charges
\be
\widetilde{u^\pm_i}=u^{\pm i}\;,\q \widetilde{u^{\pm i}}=-u^\pm_i
~.\lb{conj1}
\ee
On the harmonic derivatives of an arbitrary harmonic function $f(u)$ this
conjugation acts as follows
\be
\widetilde{\partial^{\s\pm\pm} f}=\partial^{\s\pm\pm}\widetilde{f}
\;.\lb{conj2}
\ee

Let us consider the coordinates of the $N=2$ superspace $M(3,1|8)$
over the $D=(3,1)$ Minkowski space
\bea
&&z^M=(x^\adb~,\theta^\alpha_k~,\bar\theta^{\da k})~,\\
&&(x^\adb)^\dagger=x^\bda~,\q(\theta^\alpha_k)^\dagger=\bar\theta^{\da k}
~,
\eea
where $\alpha, \da$ are the $SL(2,C)$ indices.

One can define the Minkowski analytic harmonic superspace with 2 coset
harmonic dimensions $u^I_i$ and the following set of 4 even and (2+2) odd
coordinates:
\bea
&&\zeta =(x^\adb_\A , \tpa, \btpa)
\;,\nn\\
&&x^\adb_\A=x^\adb-i(u^-_ku^+_l+u^-_lu^+_k)\theta^{\alpha k}\bar
\theta^{\dot\beta l}\lb{F5}\\
&&
\theta^{\alpha\pm}=\theta^{\alpha k} u_k^\pm~,\q\bar\theta^{\dot\alpha\pm}
=\bar\theta^{\dot\alpha k}u_k^\pm~.\nn
\eea
This superspace is covariant with respect to the $N=2$ supersymmetry
transformations
\be
\delta x^\adb_\A=2i\tpa\bar\epsilon^{\dot\beta k}u_k^-
+2i\btpb\epsilon^{\alpha k}u_k^-~,\q \delta \theta^{\alpha\pm}=
\epsilon^{\alpha k}u_k^\pm~,\q\delta\bar\theta^{\dot\beta\pm}=
\bar\epsilon^{\dot\beta k}u_k^\pm~.
\ee

The conjugation of the odd analytic coordinates has the following
form:
\be
\theta^{\alpha\pm}~\rightarrow~\bar\theta^{\dot\alpha\pm}~,\q
\bar\theta^{\dot\alpha\pm}~\rightarrow~-\theta^{\alpha\pm}
\lb{conj3}
\ee
and coordinates $x^\adb_\A$  are real.

The corresponding CR-structure involves the derivatives
\be
\DPa,\;\bDPa,~\Dp
\lb{F7}
\ee
which have the following explicit form in these coordinates:
\bea
&& \DPa=\pPa~,\q\bDPa=\bpPa \;,\lb{F8}\\
&&D^{++}=\dpp
-i\tpa\btpb\padb+\tpa\pPa+\btpa\bpPa~,
\lb{F9}
\eea
where the partial derivatives satisfy the following relations:
\be
\pada x^\bdb=2\delta^\beta_\alpha\delta^\db_\da~,\q
\pPa\tmb=\delta^\beta_\alpha~,\q \bpPa\btmb=\delta^\db_\da~.
\ee

One can construct also all harmonic and Grassmann derivatives in these
coordinates
\bea
&&\Dm=\dm
-i\tma\btmb\padb+\tma\pMa+\btma\bpMa~,\\
&& \DMa =-\pMa +i\btmb\padb~,\q
\bDMa =-\bpMa -i\tma\pada~,\\
&&\pMa\tpb=\delta^\beta_\alpha~,\q \bpMa\btpb=\delta^\db_\da~.
\eea

It is useful to write down the properties of these derivatives
with respect to Hermitian conjugation
\bea
&&(\Dp, \Dm) \rightarrow -(\Dp, \Dm)~,\q D^\pm_\alpha \rightarrow
\bar{D}^\pm_\da~,\q\bar{D}^\pm_\da\rightarrow -D^\pm_\alpha ~.
\eea

In the text, we use the following conventions:
\bea
&&\varepsilon^{ik}\varepsilon_{kl}=\delta^i_l~,\q
\varepsilon^{\alpha\gamma}\varepsilon_{\gamma\beta}=\delta^\alpha_\beta~,
\\
&&(\theta^\pm)^2={1\over2}\theta^{\alpha\pm}\theta^\pm_\alpha~,\q
(\bar\theta^\pm)^2={1\over2}\bar\theta_\da^\pm\bar\theta^{\da\pm}~,\\
&&(D^\pm)^2={1\over2}D^{\alpha\pm}D^\pm_\alpha~,\q
(\bar{D}^\pm)^2={1\over2}\bar{D}_\da^\pm\bar{D}^{\da\pm}~.
\eea
\vspace{0.5cm}

{\bf Harmonic-superspace coordinates in Euclidean superspace}\\

The Euclidean $N=2$ superspace $E(4|8)$ has the coordinates
\be
z^M=(y^\adb,\Theta^\alpha_k,\hat{\Theta}^{\da k}) \lb{eucor}
\ee
 with the alternative conjugation rules
\be
(y^\ada)^\dagger=y_\ada~,\q(\Theta^\alpha_k)^\dagger=\Theta^k_\alpha~,\q
(\hat{\Theta}^{\da k})^\dagger=\hat{\Theta}_{\da k}~,\lb{eucl1}
\ee
where $\alpha, \da$ and $k$ are the indices of the direct product of
three $SU(2)$ groups.

The Euclidean $N=2$ analytic-superspace coordinates can be defined as
follows
\bea
&&\zeta_E =(y^\adb_\A , \Tpa, \hTpa)
\;,\nn\\
&&y^\adb_\A=y^\adb-(u^{k-}u^+_l+u_l^-u^{k+})\Theta^\alpha_k
\hat\Theta^{l\db}\\
&&
\Theta^{\alpha\pm}=\Theta^\alpha_k u^{k\pm}~,\q\hat\Theta^{\dot\alpha\pm}
=\hat\Theta^{k\da}u^\pm_k~.\nn
\eea

The Euclidean spinor and harmonic derivatives have the following form in
these coordinates:
\bea
&&\cDpa=\partial/\partial \Tma\equiv d^+_\alpha~,\q\hcDpa=
\partial/\partial \hTma\equiv \hat{d}^+_\da~,\\
&&\cD^{\pm\pm}=\partial^{\pm\pm}
-\Theta^{\alpha\pm}\hat\Theta^{\db\pm}\padb+\Theta^{\alpha\pm}
d^\pm_\alpha+\hat\Theta^{\da\pm}\hat{d}^\pm_\da~,\\
&& \cDma =-d^-_\alpha +\hat\Theta^{\db-}\padb~,\q
\hcD^-_\da =-\hat{d}^-_\da -\Theta^{\alpha-}\pada~.
\eea

We shall use the following rules of conjugation for
 the Euclidean harmonized odd coordinates and derivatives:
\bea
&&(\Theta^{\alpha\pm})^\dagger=\Theta_\alpha^\pm~,\q
(\cD^\pm_\alpha)^\dagger=-\cD^{\alpha\pm}~,\\
&&(\hat\Theta^{\dot\alpha\pm})^\dagger=-\hat\Theta_\da^\pm~,\q
(\hcD^\pm_\da)^\dagger=\hcD^{\pm\da}~.
\eea

Our conventions for the bilinear combinations of the Euclidean spinors
are
\bea
&&(\Theta^\pm)^2={1\over2}\Theta^{\alpha\pm}\Theta^\pm_\alpha~,\q
(\hT^\pm)^2={1\over2}\hT^{\da\pm}\hT^\pm_\da~,\nn\\
&&
(\cD^\pm)^2={1\over2}\cD^{\alpha\pm}\cD^\pm_\alpha~,\q
(\hcD^\pm)^2={1\over2}\hcD^{\da\pm}\hcD^\pm_\da~.
\eea

\begin{thebibliography}{99}

\bibitem{GIK1}A. Galperin, E. Ivanov, S. Kalitzin, V. Ogievetsky and
E. Sokatchev, Class. Quant. Grav. 1 (1984) 469.
\bibitem{GIOS}A. Galperin, E. Ivanov,  V. Ogievetsky and
E. Sokatchev, Class. Quant. Grav. 2 (1985) 601.
\bibitem{Zu2}B.M. Zupnik,  Phys. Lett. B 183 (1987) 175.
\bibitem{GIOS2}A. Galperin, E. Ivanov,  V. Ogievetsky and
E. Sokatchev, Ann. Phys. 185 (1988) 1.
\bibitem{OO}O. Ogievetsky,  Lect. Notes in Physics, v. 313,
p.548, Springer-Verlag, 1988.
\bibitem{DO}C. Devchand and V. Ogievetsky,
Nucl. Phys. B 414 (1994) 763; Erratum, Nucl. Phys. B 451 (1995) 768.
\bibitem{Zu3}
B.M. Zupnik, Phys. Lett. B 375 (1996) 170;
 Lecture Notes in Physics, v. 509, p.157-160. Springer-Verlag, 1998.
\bibitem{Zu4}B.M. Zupnik, Short harmonic superfields and light-cone gauge
in super-Yang-Mills equations, Proceedings of the International
 conference "Quantization, gauge theory and strings"
devoted to the memory of prof. E.S. Fradkin, eds. A. Semikhatov,
M. Vasiliev and V. Zaikin, p. 277, Scientific World, Moscow,  2001;
hep-th/0011012.
\bibitem{IKZ} E.A. Ivanov, S.V. Ketov and B.M. Zupnik, Nucl. Phys. B 509
(1998) 53.
\bibitem{Zu5} B.M. Zupnik, Nucl. Phys. B 554 (1999) 365.
\bibitem{Zu}B.M. Zupnik,
 Phys. Lett. B 209 (1988) 513.
\bibitem{NZ} J. Niederle and B. Zupnik, Nucl. Phys. B 598 (2001) 645;
 hep-th/0012114.
\bibitem{Se} A. Semikhatov, Phys. Lett. B 120 (1983) 171.
\bibitem{Vo} I.V. Volovich, Phys. Lett. 129 (1983) 429; Lett. Math. Phys.
7 (1983) 517.
\bibitem{Ma}Yu.I. Manin, Gauge fields and complex geometry, Nauka, Moscow,
 1984; English version, Springer-Verlag, Berlin, 1988.
\bibitem{Si} W. Siegel, Phys. Rev. D 52 (1995) 1042.
\bibitem{SW} N.Seiberg and E.Witten, Nucl.Phys., B426 (1994) 19
\bibitem{Ga} J. Gauntlett, Nucl. Phys. B 411 (1994) 443.

\end{thebibliography}
\end{document}

