\begin{table}%t3
\caption{\label{table3}Relative line fluxes.}
%\centerline
 {\begin{tabular}{lllllllll}
\hline\hline
\noalign{\smallskip}
 & \multicolumn{2}{c}{Area SI} & \multicolumn{2}{c}{Area SII} & \multicolumn{2}{c}{Area EI} & \multicolumn{2}{c}{Area EII}  \\ 
Line (\AA) & $F^{a}$ & $S/N^{b}$ & $F$ & $S/N$ & $F$ & $S/N$ & $F$ & $S/N$ \\
\hline
\hbeta\ 4861 & $<$6 & & $<$6 & & $<$4 & & $<$4 & \\
\oi\ 6300 &  43.5 & (23) & 31.3 & (16) & 31.7 & (33) & 67.0 & (18) \\
\nitrogen\ 6548 & 50.5 & (25) & 54.5 & (25) & 48.6 & (44) & 29.4 & (9)  \\
\ha\  6563 & 100 & (45) & 100 & (45) & 100 & (83) & 100 & (26)  \\
\nitrogen\ 6584 &  167.6 & (64) & 167.1 & (67) & 135.0 & (101) & 98.8 & (25)  \\
\sulfur\ 6716 & 97.1 & (42) & 93.8 & (41) & 88.1 & (70) & 109.8 & (27)  \\
\sulfur\ 6731 & 74.8 & (32) & 73.9 & (32) & 66.4 & (58) & 74.4 & (19)  \\
\hline
Absolute \ha\ flux$^{c}$ & \multicolumn{2}{c}{2.5} & \multicolumn{2}{c}{1.9} & \multicolumn{2}{c}{4.6} & \multicolumn{2}{c}{1.8}  \\
\sulfur/\ha\  & \multicolumn{2}{c}{1.72 $\pm$ 0.05} & \multicolumn{2}{c}{1.68 $\pm$ 0.05} & \multicolumn{2}{c}{1.54 $\pm$ 0.03} & \multicolumn{2}{c}{1.84 $\pm$ 0.09} \\
$F$(6716)/$F$(6731) & \multicolumn{2}{c}{1.30 $\pm$ 0.05} & \multicolumn{2}{c}{1.27 $\pm$ 0.05} & \multicolumn{2}{c}{1.33 $\pm$ 0.03} & \multicolumn{2}{c}{1.48 $\pm$ 0.09} \\
\nii/\ha\ & \multicolumn{2}{c}{2.18 $\pm$ 0.06} &
\multicolumn{2}{c}{2.22 $\pm$ 0.06} & \multicolumn{2}{c}{1.84 $\pm$
0.03}& \multicolumn{2}{c}{1.28 $\pm$ 0.07} \\ 
c(\hbeta)$^{d}$ & \multicolumn{2}{c}{$>$2.2} &
\multicolumn{2}{c}{$>$2.2} & \multicolumn{2}{c}{$>$2.7} & \multicolumn{2}{c}{$>$2.7}\\ 
\hline
\end{tabular}}
\par
\smallskip
${^a}$ Observed fluxes normalized to F(H$\alpha$)~=~100 and uncorrected  
for interstellar extinction. \\ 
${^b}$ Numbers in parentheses represent the signal to noise ratio of the
quoted fluxes.\\ 
$^{c}$ In units of \flux. \\ 
$^{d}$ The logarithmic extinction is derived by 
$c = 1/0.348$~$\times$~log((\ha/\hbeta)$_{\rm obs}$/2.85).\\
Listed fluxes are a signal to noise weighted average of two fluxes for
area~$S$.\\
The errors of the emission line ratios are calculated through standard
error propagation.
\par
\end{table}