$
\bf{<1 \times 10^{46}}
$

mixed

^{ 1} 

VLBI 
$
\bf{<1 \times 10^{35}}
$

single

^{ 2} 

CMB 
• • • We do not use the following data for averages, fits, limits, etc. • • • 
$
<1 \times 10^{32}
$

single

^{ 1} 

VLBI 
$
<3 \times 10^{33}
$

mixed

^{ 3} 

VLBI 
$
<4 \times 10^{31}
$

single

^{ 3} 

VLBI 
$<8.5 \times 10^{17}$ 
^{ 4} 


$
<3 \times 10^{28}
$

single

^{ 5} 

CMB 
$<5 \times 10^{30}$ 
^{ 6} 

TOF 
$<2 \times 10^{28}$ 
^{ 7} 


$<2 \times 10^{32}$ 


TOF 
^{1}
ALTSCHUL 2007B looks for AharonovBohm phase shift in addition to geometric phase shift in radio interference fringes (VSOP mission).

^{2}
CAPRINI 2005 uses isotropy of the cosmic microwave background to place stringent limits on possible charge asymmetry of the Universe. Charge limits are set on the photon, neutrino, and dark matter particles. Valid if charge asymmetries produced by different particles are not anticorrelated.

^{3}
KOBYCHEV 2005 considers a variety of observable effects of photon charge for extragalactic compact radio sources. Best limits if source observed through a foreground cluster of galaxies.

^{4}
SEMERTZIDIS 2003 reports the first laboratory limit on the photon charge in the last 30 years. Straightforward improvements in the apparatus could attain a sensitivity of $10^{20}~$e.

^{5}
SIVARAM 1995 requires that CMB photon charge density not overwhelm gravity. Result scales as $\Omega _{M}~$h${}^{2}$.

^{6}
RAFFELT 1994 notes that COCCONI 1988 neglects the fact that the time delay due to dispersion by free electrons in the interstellar medium has the same photon energy dependence as that due to bending of a charged photon in the magnetic field. His limit is based on the assumption that the entire observed dispersion is due to photon charge. It is a factor of 200 less stringent than the COCCONI 1988 limit.

^{7}
See COCCONI 1992 for less stringent limits in other frequency ranges. Also see RAFFELT 1994 note.
