From the detected stars, 681 have sufficient number of consistent observations to be useful as UBVRI secondary standard stars. The selection of these stars was done applying the following criteria.

Table 4: Secondary standard stars: first 50 entries of Table 4 (full table is available only in electronic form)

$\includegraphics[width=18cm,clip]{H2151T4.PS}$

Notes: a 1 - 22 AC 858005, TYC 226200156701.

b 2 - 2 AC 8677, HIP 8815, TYC 3000024401, SA 93 101.

$\star$ Star belonging to set #2.

The field of view of the detectors used in each observational run is different and, thus, not all the stars in the neighbourhood of our Landolt stars have been observed in all the observation periods. The final sample has been split into two different sets: #1 stars having al least 3 useful measurements in at least 2 different observation periods and #2 stars having at least 4 useful measurements belonging to the same period.

$\begin{figure} \par\includegraphics[width=7cm,clip]{H2151F1.PS}\end{figure}$

Figure 1: Histogram of the sample of selected stars as a function of V magnitude

$\begin{figure} \par\includegraphics[width=7cm,clip]{H2151F2.PS}\end{figure}$

Figure 2: Histogram of the sample of selected stars as a function of B-V colour index

$\begin{figure} \par\includegraphics[width=7cm,clip]{H2151F3.PS}\end{figure}$

Figure 3: Histogram of the sample of selected stars as a function of U-B colour index

$\begin{figure} \par\includegraphics[width=7cm,clip]{H2151F4.PS}\end{figure}$

Figure 4: Histogram of the sample of selected stars as a function of V-R colour index

Among this sample, we considered as candidates to be secondary standard stars only those whose photometric errors were smaller than 0.06 mag in all bands (except from U-B, where the limit was set at 0.10 mag). Stars rejected were mainly faint ones.

$\begin{figure} \par\includegraphics[width=7cm,clip]{H2151F5.PS}\end{figure}$

Figure 5: Histogram of the sample of selected stars as a function of V-I colour index

$\begin{figure} \par\includegraphics[width=8cm,clip]{H2151F8.PS}\end{figure}$

Figure 8: Identification chart of field #1 around Landolt stars SA 44 28 (labelled "A") and SA 44 113 (labelled "B"). Labels are placed right from the stars, except for #32, placed left

$\begin{figure} \par\includegraphics[width=8cm,clip]{H2151F9.PS}\end{figure}$

Figure 9: Identification chart of field #2 around Landolt star SA 93 103 (labelled "A")

$\begin{figure} \par\includegraphics[width=8cm,clip]{H2151F10.PS}\end{figure}$

Figure 10: Identification chart of field #3 around Landolt stars SA 95 16 (labelled "A"), SA 95 96 (labelled "B") and SA 95 15 (labelled "C")

$\begin{figure} \par\includegraphics[width=8cm,clip]{H2151F11.PS}\end{figure}$

Figure 11: Identification chart of field #4 around Landolt stars SA 96 405 (labelled "A"), SA 96 406 (labelled "B") and SA 96 393 (labelled "C")

In order to avoid possible variables, Landolt (1983, 1992) computed the average of the standard deviation in V for his complete sample and omitted those stars whose standard deviation were larger than twice this average. Landolt's standard deviations were similar along all the magnitude range covered by his stars, but we have very different values from the bright to the faint end. For this reason, we applied the same clipping procedure not to the whole sample, but by intervals of two magnitudes. Inside each interval, errors increase with magnitude in a continuous manner, but this bias is very slight and only affects the rejection/acceptance of stars with larger standard deviation. Reducing the interval would lead to small number statistics in several bins.

In average, the 431 stars in set #1 were observed in 2.1 periods and have 11.7 measurements in filter V. The 250 stars in set #2 have, in average, 6.6 measurements in filter V and were observed in 1.2 periods.

Notes:	a	1 - 22	AC 858005, TYC 226200156701.
	b	2 - 2	AC 8677, HIP 8815, TYC 3000024401, SA 93 101.
	$\star$	Star belonging to set #2.

3 Selection criteria