Rate coeffs for use with Shull and Van Steenburg's formula.  The third
floating point column contains a value I call bcol which should
replace the value 0.1 present in the ionization formula these guys
used.  These should be far more accurate than the original Shull ones.

el ion  acol tcol  bcol  arad  xrad  adi   bdi  t0  t1  vari  varr

vari= variance in the fit to the logarithm of ionization rates
varr= variance in the fit to the logarithm of recombination rates

The rates are given, in cm+3/s (multipley by Ne to get the rate per atom/ion) as

ifit=acol * sqrt(temp) * exp( -tcol / temp) $
    / (1.0 + bcol * temp / tcol) 


rfit=arad*(temp/1.e4)^(-xrad)  +  $
         adi /temp/sqrt(temp) * exp(-t0/temp)* $ 
         (1.e0+bdi * (exp(- t1/temp)))