5 Comparison with other observations

After our Nançay survey program had begun, H I observations for a number of the galaxies in our sample were published by Giovanelli et al. (1997) and by Theureau et al. (1998). The data of Giovanelli et al. were obtained with the Arecibo telescope, while the data of Theureau et al. were obtained at Nançay, hence we have samples available to provide both external and internal checks on our absolute flux calibration.

We have 15 galaxies in common with the sample of Theureau et al. (1998). Figure 3 shows a comparison of our respective heliocentric radial velocities, integrated line fluxes (after correction for beam attenuation), and linewidths ( $W_{\rm 20,c}$ and $W_{\rm 50,c}$) after correction for instrumental resolution.

Our radial velocities show excellent agreement with those of Theureau et al.; velocity differences are smaller than our quoted errors in most cases. Agreement between the integrated fluxes is less good; Fig. 3 reveals that on average, our derived fluxes appear systematically higher than those reported by Theureau et al. Since both of our groups corrected for beam attenuation in a similar manner, the discrepancy must stem from either differences in calibration or spectral quality. We return to this below. Finally we see that agreement between our $W_{\rm 50,c}$ values is quite good, and well within estimated errors. As expected, noise and scatter increases somewhat among the $W_{\rm 20,c}$ values, which are more sensitive to spectral noise and baseline uncertainties.

Figure 3: A comparison of H I parameters measured from the present survey (along the y-axis) and those derived by Theureau et al. (1998; along the x-axis) for 15 galaxies in common between the samples. Top left: heliocentric radial velocities, in km s-1. Top right: integrated line fluxes, in Jy km s-1, after correction for beam attenuation. Lower left: line widths measured at 20% peak maximum, in km s-1, after correction for spectral resolution. Lower right: line widths measured at 50% peak maximum, after correction for spectral resolution. Error bars are overplotted for our data, except on the $V_{\rm h}$ panel, where uncertainties are smaller than the data points. To guide the eye, the solid line in all four panels delineates y=x; these lines are not fits to the data

Sixty of the galaxies observed by Giovanelli et al. (1997) are in common with our present sample. Thirty-seven of these galaxies are mutual detections, and 6 are mutual non-detections. Giovanelli et al. also detected 11 of our undetected targets at velocities outside our search ranges, while we detected 3 objects outside the velocity ranges that they searched. Finally, there are 2 cases for which we were unable to confirm the Giovanelli et al. detections in our spectra obtained over the corresponding velocity range (noted in Appendix C), and one case where our measured velocities show substantial disagreement ( $\Delta V\sim1000$ km s^-1; see Appendix B).

Figure 4 shows a comparison of the respective $V_{\rm h}$, $W_{\rm 50,c}$, and $S_{\rm c}$ values for the 37 mutual detections measured by us and by Giovanelli et al. (Giovanelli et al. did not publish $W_{\rm 20,c}$ estimates). Once again, we see that agreement between our radial velocity measurements is generally excellent, and in most cases, differences are smaller than our estimated uncertainties.

Scatter between our linewidth measurements is somewhat larger than that seen in our comparison with Theureau et al. data (Fig. 3). The three most discrepant galaxies include UGC 11188 (which fell near the edge of our observed bandpass) and two objects for which our spectra are considerably noisier than those of Giovanelli et al. The remaining scatter may partially stem from the different techniques our respective groups used to measure the linewidths.

Figure 4: A comparison of H I parameters measured from the present survey (along the y-axis) and those derived by Giovanelli et al. (1997; along the x-axis) for 37 galaxies in common between the samples. Top left: heliocentric radial velocities, in km s-1. Top right: integrated line fluxes, in Jy km s-1, after correction for beam attenuation. Bottom: line widths measured at 50% peak maximum, after correction for spectral resolution. Error bars are overplotted for our data, except on the $V_{\rm h}$ panel, where uncertainties are smaller than the data points. To guide the eye, the solid line in all three panels delineates y=x; these lines are not fits to the data

For sources with $S\le$ 4 Jy km s^-1, we find agreement between our beam-corrected flux integrals is in most instances consistent with calibration uncertainties, although a few cases disagree by more than 50%. For higher values of $S_{\rm c}$, we see a systematic increase in the integrated fluxes reported by Giovanelli et al. compared with our values. Since in most cases, corrections for beam attenuation in our sample are small, this suggests that the beam correction factors applied by Giovanelli et al. may tend to overestimate the true line flux of larger galaxies. An additional implication is that the offset of several of our fluxes compared with Theureau et al. (Fig. 3) is likely to result from a systematic underestimation of integrated fluxes by those workers, rather than an overestimation by us. This may stem from calibration differences, differences in baseline fits, or differences in spectral quality. Unfortunately Theureau et al. did not publish their individual spectra, hence we are unable to resolve this issue through a direct comparison of the data.

 

Table 4: Detection rates for Hubble types observed in the present survey

Type	$\char93$ Observed	$\char93$ Detected	% Detected
Sab	1	0	0
Sb	1	1	100
Sbc	2	0	0
Sc	13	1	8
Scd	36	12	33
Sd	303	148	49
Sdm	99	60	61
Sm	13	7	54
Irr	4	3	75