\documentstyle[paasms4,pptwocol]{preprint}

\begin{document}

\title{RELATIVISTIC POSITRONS IN NON-THERMAL RADIO SOURCES}

\author{A.\ S.\ Wilson\/\thanks{Also Astronomy Department, University of Maryland,
College Park, MD 20742}\\
\\
Space Telescope Science Institute\\
\\
3700 San Martin Drive\\
\\
Baltimore, MD 21218\\
\and
K.\ W.\ Weiler\\
\\
Naval Research Laboratory\\
\\
Remote Sensing Division, Code 7214\\
\\
Washington, DC 20375-5320}

\tobe{1 February 1997}{The Astrophysical Journal}

\recacc{10 June 1996}{20 August 1996}

\maketitle

\abstract{
We describe a procedure for measuring the contribution of relativistic
positrons to radio synchrotron radiation. The method relies on the fact that
synchrotron radiation from particles of one sign (\eg electrons) is
circularly polarized by a small but measurable amount. If, on the other hand,
there are equal numbers of relativistic positrons and electrons,
the net circular polarization is zero. The method is illustrated through 
high accuracy mapping of the circular polarization of the Crab Nebula at
610~MHz. No significant 
circular polarization was detected: a very conservative limit is
$<0.05$\%, and a more realistic one $<0.03$\%. We calculate the degree of
circular polarization expected if only electrons are present, allowing for
the reduction in polarization resulting from non-uniformities in the magnetic
field along the line of sight and across the telescope beam. This reduction
due to field non-uniformity is estimated from measurements of the degree of
{\it linear\/} polarization at optical and high radio frequencies with similar
angular resolution to the circular polarization measurements. We find that the
observed upper limit on the degree of circular polarization is comparable
to or below
that expected if only electrons radiate. Various explanations of this result
are discussed, including (a)~a weaker than assumed magnetic field, (b)~a field
preferentially nearly perpendicular to the line of sight, (c)~a field
structure of such a type that non-uniformities reduce the degree of circular
polarization by more than they reduce the degree of linear
polarization, and (d)~relativistic positrons are present. Although~(a) is
implausible, possibilities~(b) and~(c) cannot be excluded. If future
observations establish that the degree of circular polarization at 610~MHz
is $<0.01$\%, a contribution from positrons would be strongly favored.}