3 The catalogue

The catalogue contains:

• 494 new high-, medium- and low-resolution measurements described in Sect. 2.
• 373 nebulae from Beaulieu (1996), 73 from Kohoutek & Pauls (1995), 52 from Dopita & Hua (1997), 3 from Spyromilio (1995) and 5 from Kraan-Korteweg et al. (1996).
• 577 radial velocities published in the SECAT (including STPP83, MWF88 and KFL88 data).

3.1 Uncertainties and systematic deviations of literature data

The literature data originates mostly from a number of individual surveys. In order to test the accuracies of these surveys, we compare the lists with the high-resolution ZASW data, for those objects in common.

Figure 2 (right) displays the comparison of the literature data data with the high-resolution ZASW data for several surveys. Two individual objects of Dopita & Hua, He 2-118 and He 2-97, show a very large discrepancy with the ZASW values (see Fig. 2, right) and are excluded in the statistics of Table 1. Both have late-type [WC] central stars, and in such objects a velocity shift is sometimes seen between nebular and stellar lines (e.g. Menzies & Wolstencroft 1990; Le Bertre et al. 1989). Two large surveys included in STPP83 (Minkowski 1957 (right panel) and Mayall 1964 (left panel)) show significant scatter. Both are published only as private communications in the catalogue of Perek & Kohoutek (1967) and may originate from a variety of separate measurements.

  

Figure 2: Left: Comparison of the five principal sources of STPP83 with the high-dispersion ZASW data ($\Delta V$ stands for $(V_{\rm ZASW} - V_{i})$); the literature data are globally overestimated and present various uncertainties depending from their authors. Right: Comparison of recent data with the ZASW data ($\Delta V$ stands for $(V_{\rm ZASW} - V_{i})$); in New lit data: open squares are DH97, crosses are SB96 and circles are KP95 data

Table 1 presents the systematic deviations and the scatter for the principal sources. We calculate the significance of the systematic offsets between the samples by applying student's t statistics on the objects in common between ZASW and the individual samples. An uncertainty of 2 km$\,$s^-1 is assigned to ZASW (Zijlstra et al. 1997). The resulting probabilities are listed in Table 1. The offset is taken to be significant if the probability is less than 0.01. This is the case only for Minkowski (1957). STPP83 added 4 km$\,$s^-1 to all Webster and Kalnajs (1982) velocities based on a comparison with 10 previous measurements. This correction is not confirmed at the required significance level.

The data from Mayall (1964) shows a large scatter which is well approximated by a 1-$\sigma$ uncertainty of 40 km s^-1 (Table 1). This is much larger than the 25 km s^-1 assigned by STPP83. After removing the systematic offset for the Minkowski (1957) data, an uncertainty of 15 km s^-1 remains (as compared to 11 km s^-1 assigned in STPP83). The lists by MWF88 and KFL88 are not included in Table 1 as only a few of their objects are common with ZASW data. We assign an estimated uncertainty of 15 km s^-1 for all KFL88's velocities, based on the resolution of their spectra which is similar to Beaulieu's data. KFL suggest themselves an uncertainty of 8 km s^-1.

 

Table 1: Systematic deviations of the principal sources of available data compared with the ZASW data (n is the number of compared PNe) (offsets stand for $<V_{\rm ZASW} - V_{i}\gt$). The probability gives the likelihood that the offset is due to chance, using student's t statistics

 

3.2 The catalogue

Following the discussion above, we correct the sample of Minkowski (1957) for a systematic offset of -16.3 km$\,$s^-1. The offsets for other samples are not considered significant. For each PN the average value of its radial velocity is now obtained by weighting each existing velocity by the inverse of the square of its associated error, thus giving low weight to the poorest data:

$$\begin{eqnarray} \bar{V} = \bigg[\sum_{i}\bar{V}_{i}/Err_{i}^{2}\bigg] \bigg/ \bigg[\sum_{i} 1/Err_{i}^{2}\bigg].\end{eqnarray}$$ (1)

For the mean total error, we choose as the best estimate the larger of the two quantities:

$$\begin{eqnarray} Err = \bigg[\sum_{i}1/Err_{i}^{2}\bigg]^{-\frac{1}{2}}\end{eqnarray}$$ (2)

and

$$\begin{eqnarray} Err_{\rm rel}\!\!=\!\!\bigg[\sum_{i}(\bar{V}_{i}-\bar{V})\big/E... ...igg/\bigg[(P-1)\!\!\sum_{i}1\big/Err_{i}^{2}\bigg]^{\frac{1}{2}} .\end{eqnarray}$$ (3)

The catalogue is presented in Table 2 as follows:
Column 1: the standard PNG numbers as adopted in the SECAT.
Column 2: usual name of the PN.
Column 3: new high-resolution ZASW velocities.
Column 4: new medium-resolution ASC velocities.
Column 5: new low-resolution SA velocities.
Column 6: literature data published after 1992 with references.
Column 7: literature data from the SECAT (some are corrected from offsets seen in Table 1) with references.
Column 8: adopted radial velocities with their uncertainties.