Appendix A: Calibration of the Nançay data

Since at Nançay, all 9, 18 and 21-cm receivers are permanently installed in the carriage chariot, the noise diode measurements are not made in the laboratory, but are made through drift scan observations of standard radio-continuum sources. Full sets of noise diode temperatures are measured twice a year for many observation frequencies and polarizations, on several extragalactic radio-continuum sources.

Calibration measurements were made at three frequencies: 1370, 1400 and 1425 MHz. A general drift of the noise diode temperature as a function of time is seen. The jump occurring on July 24, 1994 (hereafter 940724) is due to a hardware change.

Our observations were performed between November 1992 and mid-1996, but most of the data were taken in 1993 and 1994. The main variations of the noise diode temperature are due to the changing observation frequencies, corresponding to radial velocities of 500- 6000 km s^-1. The following formulas were used: $T_{<940724} = 223.3 -0.14 \times F$ and $T_{\gt 940724} = 63.1 -0.032\times F$ where T is the estimated noise diode temperature in K and F the observation frequency in MHz. The overall calibration procedure accuracy is $\sim$ 10%.

A1. Comparison with data from other authors

  

Table 4: Comparison with other authors

Among the galaxies which have been detected with a high signal-to-noise ratio in the course of this survey and are not suffering any confusion, 12 have been observed by other authors and their HI parameters can be found in the literature. We have also observed 2 more Markarian galaxies (MK 303 and MK 307) to increase our comparison sample. The comparison sample is given in Table 4, with both our HI measurements and those of other authors.

  

Figure 6: Integrated HI flux densities from other authors (ordinate axis) versus our Nançay and Green Bank measurements (abscissa axis) for a sample of 15 common galaxies. The solid line represents equality

We have plotted in Fig. 6 our measurements against those of others along with our error bars. The solid line represents equality. The data points scatter evenly on either side of the equality line, demonstrating the correctness of our flux calibration.