\begin{table}%t2
\caption{\label{t2}Coefficients for the l-samples.}
%\centerline
{
\begin{tabular}{cc cc cc c}   
\hline
\hline
{}&\multicolumn{2}{c} {Linear Fits} &{}&\multicolumn{3}{c} {Polynomial Fits}\\
l-sample&$b_0$&$b_1$&$\sigma$&$a_0$&$a_1$&$a_2$\\\hline
$a$&9.139&0.575&0.097&9.117&0.329& --0.447\\\
$b$&9.130&0.564&0.091&9.112& 0.408&--0.226\\\
$c$&9.137&0.599&0.114&9.114&0.438& --0.183\\\hline
\end{tabular}}
\medskip
 For linear fits we assume $y=b_0+b_1x$, and for polynomial fits $y=a_0+a_1x+a_2x^2$, with $y$ = 12~+~log(O/H) and $x$ = log([{N~\textsc{ii}}] $\lambda$6583/[{O~\textsc{ii}}] $\lambda$3727).
\end{table}