\begin{table}%t1
\caption{\label{line_parameter}Parameters of the hyperfine lines $F' - F''$ in the observed $1_{11}{-}0_{00}, J=3/2{-}1/2$ ortho \hoplus{} transition, including predicted frequencies, Einstein-$A$ and optical depth at low temperatures.}
\vspace{-1mm}
\par
%\centerline
{\small
\begin{tabular}{lrrrrrr}
\hline\hline
$F' - F''$ & $\nu\sub{M\ddot{u}rtz}^a$ & $\nu\sub{Strahan}$\tablefootmark{a} & $\nu\sub{OH-based}$\tablefootmark{b} & $A$ &
$\int \tau {\rm d}v$ \\
& $[$MHz$]$ & $[$MHz$]$ &  $[$MHz$]$ & $[$s$^{-1}]$ & \tablefootmark{c} \\
\hline
$5/2{-}3/2$ &1~115~204.1 &1~115~175.8& 1~115~161& 0.031&	23.51 \\
$3/2{-}1/2$ &1~115~150.5 &1~115~122.0& 1~115~107& 0.017&	 8.67 \\
$3/2{-}3/2$ &1~115~263.2 &1~115~235.6& 1~115~221& 0.014&	 7.00 \\
$1/2{-}1/2$ &1~115~186.2 &1~115~158.0& 1~115~143& 0.027&	 6.96 \\
$1/2{-}3/2$ &1~115~298.9 &1~115~271.6& 1~115~257& 0.0035&	 0.88 \\
\hline
\end{tabular}}
\vspace{-1mm}
\tablefoot{\tablefoottext{a}{Predictions based on \citet{H2O+_LMR_1986} and
\citet{H2O+_LMR_1998}. Nominal uncertainties are $\approx 2$~MHz 
but this is inconsistent with the discrepancy between the two 
predictions so that the actual uncertainty is unknown;}
\tablefoottext{b}{from the matching DR21 OH pattern by \cite{Guilloteau};}
\tablefoottext{c}{$\int \tau\sub{low-\it T}/N\sub{H_2O^+} {\rm d}v$ in $10^{-14}$~km~s$^{-1}$~cm$^2$.}}\vspace*{2mm}
\vspace{-4mm}
\end{table}