\begin{table}%t2
  \caption{\label{tab:correlationfits}Correlation properties between $\log$~\lx\ and $\log$~\lmir\ to various subsample populations.}
%\centerline
 {
\begin{tabular}{lcccccr}
    \hline\hline
    Subsample                 & $N$       & $\rho_{\cdot z}$              &    $a$      &     $b$    &  $\bar{r}$  &  $\sigma_r$\\
       (1)                     & (2) &             (3)                     &    (4)      &     (5)    &      (6)    &    (7)    \\
    \hline
    {\bf All}                  &  {\bf 42} & {\bf 0.81}                    &  {\bf 0.41$\;\pm\;$0.03} & {\bf 1.11$\;\pm\;$0.04} & {\bf 0.30} & {\bf 0.36}\\
    All, excluding CT          &  34       &      0.79                     &  0.43$\;\pm\;$0.04 & 1.11$\;\pm\;$0.04 &     0.35    &    0.36\\
    All Sy~1                   &  12       &      0.71                     &  0.34$\;\pm\;$0.09 & 1.18$\;\pm\;$0.14 &     0.34    &    0.31\\
    All Sy~2 + CT              &  27       &      0.77                     &  0.40$\;\pm\;$0.06 & 1.40$\;\pm\;$0.11 &     0.25    &    0.38\\
      & & & & \\  
    {\bf Well-resolved}        &  {\bf 22} & {\bf 0.89}                    &  {\bf 0.19$\;\pm\;$0.05} & {\bf 1.11$\;\pm\;$0.07} & {\bf 0.15} & {\bf 0.23}\\
    Well-resolved, excluding CT&  15       &      0.93                     &  0.18$\;\pm\;$0.05 & 1.08$\;\pm\;$0.07 &     0.17    &    0.18\\
    Less-resolved              &  20       &     0.76                      &  0.62$\;\pm\;$0.05 & 1.11$\;\pm\;$0.06 &     0.47    &    0.40\\
      & & & & \\[-1.5mm]  
      \multicolumn{7}{c}{\underline{Systematic variations}}\\  
    Well-resolved (Systematic 1$\uparrow$)  & 25 & 0.82                    & 0.25$\;\pm\;$0.05 & 1.21$\;\pm\;$0.08 &     0.17    &    0.30\\
    Well-resolved (Systematic 1$\downarrow$)& 16 & 0.90                    & 0.26$\;\pm\;$0.05 & 1.07$\;\pm\;$0.08 &     0.21    &    0.23\\
    Well-resolved (Systematic 2)            & 22 & 0.83                    & 0.22$\;\pm\;$0.05 & 1.09$\;\pm\;$0.07 &     0.16    &    0.27\\
    Well-resolved (Systematic 3)$^{\dag}$   & *  & *                       & 1.87$\;\pm\;$0.31 & 1.11$\;\pm\;$0.10 &       *     &      * \\
    Well-resolved (Systematic 4)            & *  & *                       & 0.15$\;\pm\;$0.05 & 1.02$\;\pm\;$0.07 &       *     &      * \\
    All (Systematic 5)                      & 42 & 0.74                    & 0.42$\;\pm\;$0.03 & 1.11$\;\pm\;$0.04 &     0.30    &    0.44\\
    Well-resolved (Systematic 5)            & 22 & 0.78                    & 0.21$\;\pm\;$0.05 & 1.09$\;\pm\;$0.07 &     0.13    &    0.31\\
    \hline
  \end{tabular}}
\par
\medskip
``CT'' refers to Compton-thick sources. The two solutions likely to be of most interest (see Sect.~\ref{sec:summary}) are written in bold. The final seven rows refer to tests for systematic variations in the correlation; see Sects.~\ref{sec:scale_systematics}--\ref{sec:filter_systematics}. $^\dag$ This test (Systematic 3) is carried out in linear (not log) luminosity space (see Sect.~\ref{sec:errors}). * For Systematics 3 and 4, the subsamples are identical to the main well-resolved sample of 22 sources, and so the parameters marked with a star are identical as well.
\end{table}