Standard Stars - CCD Photometry, Transformations and Comparisons
Hwankyung Sung , Michael. S. Bessell, PASA, 17 (3), 244.

Next Section: Summary
Title/Abstract Page: Standard Stars - CCD
Previous Section: Atmospheric Extinction
Contents Page: Volume 17, Number 3

Standard System Transformations

Landolt Standards

After bias subtraction and flat-field division, we performed simple aperture photometry on the standard star frames. An aperture of 14'' diameter was used in the standard star photometry to simulate the photoelectric photometry done by Landolt (1992). In deriving the transformation coefficients, we used only stars with good signal-to-noise ratios ( $\epsilon \leq$ 0.01mag). In addition, the reliability of the standard star values was also very important. We used stars with $n_{pe} \geq 5$ only and assigned a weight for each star based on n_pe. Final values of the transformation coefficients were determined by averaging the individual determinations. The resulting transformation coefficients can be found in Sung et al. 1998) and a typical case is shown in Fig. 2. In the vertical axis labelling, the uppercase characters, lowercase characters with subscript 0, and $\zeta$ represent standard magnitudes, atmospheric extinction-corrected instrumental magnitudes and zero point, respectively.

There is no evidence for a change in slope in the V transformation up to V-I= 2. But for the SAAO standard stars, the slope changed at V-I = 1.5 (see Fig. 3). For the I transformation, there is a weak indication for a slope change at V-I = 0.7. Menzies (1993) outlined the non-linearities involved in the standardization of the current natural photoelectric UBVRI system at the SAAO. Linear transformations to V have a break at B-V = 1.5; to V-I have a break at V-I = 1.6; and to B-V have a break at 1.0 and again at 1.6. The neglect of such breaks has lead to some of the systematic differences between different versions of the UBVRI system.

In Fig. 2e, we see that a simple linear transformation is not suitable for U. A large scatter in U can be seen at U-B $\sim$ 0. This suggests that the scatter is related to the Balmer discontinuity and the confluence of the Balmer lines. The linear transformation coefficient for the U magnitude against U-B is the slope derived simply from two stars with extreme U-B colours in the T Phe region. The large value of the slope indicates that the effective wavelength of the U band is much too far to the red. After applying this linear U-B correction term, the residuals were plotted against (B-V) [(B-V)₀ for T Phe C] in Fig. 2f. The thin line represents the non-linear correction term f[(B-V)₀] in the U transformation. The size of the non-linear correction is at most about 0.13 mag.

Part of this correction arises from the rapid change in the UV sensitivity of the thinned SITe 2048 x 2048 CCD shifting the U passband redward so that the stellar fluxes above the Balmer discontinuity contribute too much to U and part of the correction also arises from the U-B of Landolt standard stars (three stars in SA 98 showed up to 0.06 mag difference relative to Menzies et al. (1991) measurements). We will discuss this further in section 3.3.

Several standard regions observed in the 1997 January, Feb/Mar observing runs (Ru 149, PG 0918 +029, & PG 1323 -086) are well described by the transformation relation found in Fig. 2(e & f). But stars in the SA 110 region observed in 1996 August do not show evidence of the non-linear term in the U transformation. This probably results from the fact that SA 110 is near the galactic plane and many stars may be strongly affected by interstellar reddening and follow a different correction curve.

**Figure 2:** A typical sample of transformation relations to the Landolt standard system (1996 Nov. 6)). The transformation relations can be found in Sung et al. (1998). The square, filled circle, and open circle represent, respectively, stars in the T Phe region, SA 98 region, and SA 98-185 (an Algol-type eclipsing binary Kim et al. 1997). The size of the symbols are proportional to the number of independent photoelectric observations in Landolt (1992). The extinction coefficients used for the extinction correction are marked in the figure.
$\begin{figure} \begin{center} \psfig{file=L961106.ps,height=20cm}\end{center}\end{figure}$

SAAO Standards

Transformation Coefficients

The problems with the U transformation to the Landolt standard U system encouraged us to observe the SAAO standard stars and also to shift the effective wavelength of the U band further to the UV and thus diminish the contribution of the light above the Balmer jump. To do this we added an additional 1mm of UG1 to the existing U filter. This unfortunately also had the effect of lowering the overall throughput of the U band requiring an increase in exposure times.

One difficulty in standardising CCD photometry is that there are only a few standard stars on a CCD image. The lack of stars having extreme colours is another difficulty in determining transformation coefficients accurately. It is clear that efforts are still required to provide more and better CCD standard fields. We included the equatorial stars measured by Menzies et al. (1991) and blue and red standard stars by Kilkenny et al. (1998). We averaged the individual determinations to derive the final transformation coefficients. The final transformation relations are,

$\begin{displaymath}V = v - k_{1V} X + 0.072 (B-V) + \zeta_V \hspace{4.0cm} \end{displaymath}$

$\begin{displaymath}V = v - k_{1V} X + 0.074 (V-I) + \zeta_V',~ \hspace{1.90cm}(V-I \leq 1.5) \end{displaymath}$

$\begin{displaymath}V = v - k_{1V} X + \zeta_V'',~ \hspace{3.9cm}(V-I > 1.5) \end{displaymath}$

$\begin{displaymath}B = b - [k_{1B} - 0.031 (B-V)] X - 0.101 (B-V) + \zeta_B \hspace{1.6cm} \end{displaymath}$

$\begin{displaymath}U = u - [k_{1U} - 0.013 (U-B)] X + 0.125 (U-B) + \zeta_U, ~ (U-B \leq 0.0) \end{displaymath}$

$\begin{displaymath}U = u - [k_{1U} - 0.013 (U-B)] X + 0.006 (U-B) + \zeta_U, ~ (U-B > 0.0) \end{displaymath}$

$\begin{displaymath}I = i - k_{1I} X + 0.028 (V-I) + \zeta_I. \hspace{4.0cm} \end{displaymath}$

In the V transformation with respect to the V-I colour, evidently there is no dependence on V-I for red stars (V-I > 1.5). This is probably related to the fact that V photometric standards were originally setup using B-V as the colour term. The B-V colour of late-K and M stars shows little variation while V-I continues to increase to later spectral types. As mentioned above, Landolt standard stars do not show evidence of a slope change up to V-I = 2.0. The transformation of U is also very different from that found previously (see section 3.1). Due to the use of an additional 1mm UG1 which suppresses the contribution of the Balmer discontinuity, nearly no U-B dependence for late type stars was found. Somewhat large scatters in the figure are caused mainly by the mixed use of data obtained on different nights.

**Figure 3:** Transformation relations to the SAAO standard system. The transformation relations are described in the text. The filled circles, open squares and triangles represent, respectively, E-region standard stars from Menzies et al. (1989), the Landolt equatorial standard stars from Menzies et al. (1991) and blue and red standard stars from Kilkenny et al. (1998). The uppercase character, lowercase character with subscript 0, and $\zeta$ denote standard magnitudes, atmospheric extinction corrected instrumental magnitudes, and zero point, respectively. The thin line represents the adopted transformation relation for each band.
$\begin{figure} \begin{center} \psfig{file=SAAO_tr.ps}\end{center}\end{figure}$

Residuals

We plot in Fig. 4 the residuals relative to the SAAO system after transforming to the standard system using the relations above. The blue and red stars in Kilkenny et al. (1998) show slightly larger scatter. This is probably not caused by the photometric errors, but by errors in the extinction correction for the stars having extreme colour. For V, there is no evidence of a systematic difference relative to the standard system. But for I, B, and U, there exist small but systematic differences. Evidently the systematic differences in U and B are related to the Balmer discontinuity. The maximum difference at (B-V) = 0.5 is about 0.02 mag. The difference seems to increase for red stars, but lack of very red stars (due to a lack of highly reddened stars among the standard stars) makes it difficult to confirm this proposition.

Even though the scatter is large, we can be confident that systematic differences exist in I. Observations from individual nights show clearly the change of slope at V-I $\approx$ 0.7. The large scatter is probably caused by the difficulty in the determination of accurate zero points due to the lack of stars of extreme colour and by variations in atmospheric water vapour content that affects fluxes in the I band more than in other bands. The systematic differences in I are almost certainly caused by the fact that the standard I passband is defined on the red side by the abrupt cutoff in the GaAs response at about $9000\AA$ whereas the red edge of our CCD I passband is defined by the gradual cutoff of the CCD to longer wavelengths. As a result the contribution of the Paschen discontinuity (

$\lambda\sim 8200\AA$ ) to the I magnitude will be different for the two passbands.

**Figure 4:** The average of residuals relative to the SAAO system plotted against standard colour. The $\Delta$ represents the difference between the standard and observed magnitudes. The size of symbols is proportional to the number of observations. All the symbols are the same as in Fig. 3. The dotted lines represent the mean systematic difference found from the data.
$\begin{figure} \begin{center} \psfig{file=SAAO_fr.ps }\end{center}\end{figure}$

Table 3a. Photometric Data - Standard Stars

star	V	V-I	B-V	U-B	$\epsilon_V$	$\epsilon_{V-I}$	$\epsilon_{B-V}$	$\epsilon_{U-B}$	n_obs
E1-R	9.474	$\cdots$	1.424	1.662	0.006	$\cdots$	0.006	0.003	2		2	1
E1-O	9.764	0.652	0.598	0.106	0.007	0.001	0.005	0.001	2	2	2	1
E1-Q	9.861	0.802	0.751	0.271	0.005	0.002	0.005	0.001	2	2	2	1
E525	9.372	0.867	0.809	0.401	0.005	0.009	0.010	0.011	4	2	4	4
E544	9.968	0.360	0.175	0.038	0.007	0.007	0.004	0.012	6	6	6	6
E566	9.873	1.462	1.441	1.507	0.012	0.013	0.010	0.011	6	2	6	6
E567	10.200	1.138	1.116	0.925	0.014	0.011	0.012	0.006	6	5	6	6
E5-R	9.331	$\cdots$	1.696	2.096	0.015	$\cdots$	0.008	0.010	6		6	8
E5-O	9.996	0.399	0.356	0.164	0.005	0.007	0.008	0.007	12	9	10	10
E5-T	10.825	0.687	0.653	0.199	0.007	0.007	0.009	0.009	12	10	10	11
E661	10.199	1.233	1.254	1.243	0.003	0.002	0.004	0.011	4	2	4	4
E6-U	9.956	0.533	0.449	0.041	0.003	0.010	0.003	0.003	4	4	4	4
E6-W	10.538	0.280	0.256	0.138	0.002	0.001	0.001	0.004	4	4	4	4
E6-Y	10.150	0.885	0.821	0.395	0.004	0.005	0.004	0.003	4	3	4	4
E737	9.783	1.114	1.125	0.971	0.008	0.008	0.006	0.015	10	5	10	10
E7-W	10.549	0.253	0.168	0.034	0.004	0.008	0.005	0.015	13	12	11	11
E7-X	10.773	0.038	0.020	-0.393	0.009	0.011	0.005	0.010	13	12	11	11
E7-c	10.405	1.216	1.122	0.838	0.004	0.005	0.005	0.013	13	8	11	11
SA93-317	11.558	0.596	0.509	-0.025	0.004	0.003	0.008	0.008	7	6	6	7
SA93-326	9.573	0.517	0.456	-0.012	0.006	0.009	0.004	0.003	5	4	5	5
SA93-332	9.791	0.604	0.527	0.002	0.005	0.009	0.008	0.009	5	5	5	5
SA93-333	12.032	0.887	0.851	0.493	0.003	0.010	0.011	0.018	7	6	6	7
SA93-424	11.640	1.057	1.082	0.947	0.007	0.006	0.009	0.018	6	5	5	6
SA93-417	11.942	0.816	0.755	0.280	0.003	0.007	0.010	0.013	7	6	6	7
SA93-407	11.972	0.889	0.884	0.618	0.004	0.006	0.015	0.018	7	6	6	7
SA93-405	12.212	0.598	0.509	-0.012	0.004	0.007	0.012	0.016	7	6	6	7
SA93-312	12.035	0.667	0.598	0.054	0.008	0.009	0.016	0.006	6	5	5	5
SA93-422	12.143	0.745	0.608	0.063	0.005	0.006	0.011	0.021	4	3	3	4
SA99-408	9.802	0.492	0.430	0.046	0.009	0.010	0.004	0.007	4	3	4	3
SA99-418	9.452	-0.032	-0.031	-0.154	0.012	0.011	0.004	0.005	3	3	3	3
SA99-438	9.390	-0.158	-0.154	-0.711	0.007	0.013	0.002	0.005	3	3	3	3
SA99-447	9.418	-0.082	-0.070	-0.205	0.005	0.008	0.004	0.005	3	3	3	3
SA114-750	11.915	-0.007	-0.047	-0.359	0.011	0.015	0.004	0.009	12	12	11	9
SA114-755	10.918	0.635	0.580	0.000	0.005	0.012	0.005	0.005	12	12	11	9
SA114-670	11.117	1.215	1.213	1.212	0.007	0.013	0.006	0.011	12	12	11	9
SA114-548	11.609	1.389	1.362	1.547	0.005	0.015	0.006	0.016	12	12	11	9
SA114-654	11.853	0.706	0.660	0.205	0.002	0.017	0.002	0.007	6	6	5	5
SA114-651	10.272	0.686	0.612	0.070	0.001	0.001	0.001	0.001	1	1	1	1
CD-38 222	10.462	-0.184	-0.219	-0.928	0.001	0.001	0.001	0.002	1	1	1	1
GL27.1	11.406	2.044	1.480	1.198	0.001	0.001	0.001	0.006	1	1	1	1
GL747.4	11.320	1.958	1.438	1.094	0.001	0.001	0.002	0.005	1	1	1	1
HD188112	10.196	-0.260	-0.177	-0.795	0.018	0.003	0.007	0.005	2	2	2	2
CD-31 4800	10.524	-0.309	-0.310	-1.218	0.002	0.013	0.007	0.001	2	2	2	2

Table 3b. Photometric Data - Additional Stars

star	V	V-I	B-V	U-B	$\epsilon_V$	$\epsilon_{V-I}$	$\epsilon_{B-V}$	$\epsilon_{U-B}$	n_obs
E5-Y	12.875	0.037	0.041	0.032	0.012	0.013	0.010	0.009	18	16	15	14
E5-c	13.426	0.956	0.900	0.445	0.019	0.010	0.006	0.004	18	16	16	15
E5-h	14.280	1.308	1.239	1.111	0.014	0.014	0.014	0.199	18	16	16	15
E5-b	13.573	0.717	0.668	0.181	0.009	0.010	0.006	0.006	12	11	11	11
E5-X^a	12.027	0.812	0.738	0.283	0.005	0.009	0.001	0.009	14	11	12	11
E5-W	12.053	0.488	0.412	0.083	0.005	0.008	0.004	0.012	17	13	14	13
E5-Z	12.367	0.833	0.778	0.413	0.003	0.003	0.003	0.007	2	2	2	1
E5-a	13.104	0.689	0.590	0.065	0.009	0.005	0.012	0.006	18	16	16	13
E5-d	13.635	0.770	0.693	0.178	0.012	0.007	0.012	0.013	14	12	12	11
E5-e	13.892	0.705	0.677	0.169	0.017	0.023	0.016	0.006	12	11	11	11
E5-f	14.039	0.891	0.816	0.378	0.014	0.009	0.007	0.019	18	16	16	14
E5-g	14.656	0.750	0.578	-0.074	0.007	0.009	0.006	0.029	10	9	9	8
E6-S	10.153	0.173	0.041	-0.563	0.003	0.001	0.003	0.000	2	2	2	2
E6-a	10.989	0.574	0.500	0.069	0.003	0.005	0.004	0.007	4	4	4	4
E6-c	10.672	1.658	1.523	1.800	0.003	0.002	0.003	0.022	4	2	4	4
E6-d	11.756	0.625	0.541	0.140	0.007	0.002	0.004	0.009	4	4	4	4
E6-e	11.996	0.664	0.574	0.037	0.006	0.008	0.008	0.009	4	4	4	4
E6-f	11.590	1.169	1.106	0.790	0.005	0.008	0.003	0.011	4	4	4	4
E6-n	12.868	1.141	1.046	0.751	0.008	0.014	0.007	0.028	4	3	4	4
E6-p	13.615	0.765	0.689	0.229	0.024	0.024	0.040	0.034	4	4	4	4
E6-r	14.374	1.290	1.185	1.034	0.006	0.003	0.004	0.083	2	2	2	2
E6-s	14.618	0.813	0.792	0.162	0.041	0.049	0.047	0.132	4	4	4	4
E6-t	14.701	0.822	0.757	0.416	0.013	0.029	0.008	0.044	2	2	2	2
E721	10.137	0.453	0.408	0.176	0.013	0.007	0.010	0.006	9	8	5	5
E7-Y	10.826	0.201	0.131	-0.003	0.005	0.010	0.005	0.011	13	12	11	11
E7-a	10.961	0.560	0.485	0.029	0.010	0.007	0.005	0.008	13	12	11	11
E7-b	10.963	0.861	0.646	0.403	0.004	0.010	0.005	0.013	13	12	11	11
E7-d	11.159	0.567	0.461	0.311	0.005	0.008	0.006	0.012	12	11	10	10
E7-e	11.984	0.187	0.113	-0.087	0.012	0.010	0.011	0.015	9	8	7	7
E7-f	12.047	0.281	0.228	0.192	0.011	0.021	0.009	0.013	13	12	11	11
E7-h	11.913	1.034	0.817	0.532	0.016	0.021	0.011	0.015	13	12	10	8
E7-i	12.604	0.574	0.445	0.284	0.013	0.013	0.016	0.011	13	12	11	11
E7-k	12.936	0.389	0.331	0.247	0.023	0.017	0.011	0.019	10	9	8	8
E7-l	12.500	1.232	1.220	1.159	0.016	0.010	0.011	0.041	13	12	11	11
E7-m	12.537	1.295	1.306	1.385	0.016	0.009	0.009	0.051	8	8	7	7
E7-n	13.026	1.261	1.168	0.889	0.020	0.015	0.016	0.046	12	11	10	10
E7-o	12.844	1.546	1.536	1.642	0.015	0.007	0.023	0.104	7	6	6	6
E7-p	13.226	1.292	1.107	0.652	0.017	0.022	0.019	0.058	12	11	10	10
E7-q	13.053	1.851	1.662	1.950	0.010	0.014	0.032	0.206	13	12	11	11
E7-g	12.069	0.566	0.478	0.058	0.005	0.006	0.005	0.012	13	12	11	11
E7-r	13.406	1.697	1.623	1.805	0.014	0.010	0.023	0.198	13	12	11	11
E7-s	14.281	0.753	0.686	0.203	0.032	0.049	0.039	0.071	13	12	11	11

^aOptical double

Final Results

The final results after making systematic difference correction to U, B, and I are listed in Table 3a,b. U-B is little affected by the systematic corrections to U and B because they mainly cancel each other out. We also list the photometric data for several stars brighter than V = 15 in the observed fields. We give in Table 4 the mean differences between our Table 3a values and the original SAAO data and plot differences for the individual stars against appropriate colours in Fig. 5.

No systematic differences are now evident. In particular, there are no systematic differences between the three sets, although the stars from Kilkenny et al. (1998) show a somewhat larger scatter. That is as mentioned above, solely due to these stars having extreme colours. The relatively large scatter in V-I is caused by the difficulty in determining the zero point with too few stars of extreme colour and by variations in absorption from atmospheric water vapour. To minimize the difference relative to the SAAO system we should use a multi-line transformation equation for I.

A few stars (E1-R, GL747.4, & E661) showed large differences in U-B. For two of them, (E1-R, GL 747.4) the SAAO data are probably at fault, as indicated by the quoted standard error in U-B (for example, GL 747.4 U = 13.794, $\sigma_{U-B}$ = 0.031). For E661, the magnitude as well as the colours showed large differences ( $\Delta V$ = -0.049, $\Delta (V-I)$ = -0.012, $\Delta (B-V)$ = -0.016, and $\Delta (U-B)$ = -0.077). E661 may be a variable or spectroscopically peculiar star.

Our results show that CCD photometry can achieve better than 0.015 mag accuracy. To improve the accuracy, several well-observed standard regions with many standard stars covering a wide colour and magnitude range are necessary.

**Figure 5:** Difference between Table 3a values and the original SAAO photometry. The symbols used are the same as in Fig. 2. The zero point difference and its scatter (s.d.) are marked in each panel. Three deviant stars are shown. The sense of $\Delta$ is photoelectric magnitude minus CCD magnitude.
$\begin{figure} \begin{center} \psfig{file=SAAO_comp.ps}\end{center}\end{figure}$

Table 4. Comparison with Photoelectric Photometry

source	$\Delta V^*$	$\Delta (V-I)^*$	$\Delta (B-V)^*$	$\Delta (U-B)^*$
Menzies et al. (1989)	-0.44 $\pm$ 5.51 (16^a)	-1.53 $\pm$ 12.00 (15)	-1.41 $\pm$ 8.12 (17)	-3.87 $\pm$ 13.73 (15^b)
Menzies et al. (1991)	+2.00 $\pm$ 14.43 (13)	+0.31 $\pm$ 11.38 (13)	-1.77 $\pm$ 7.12 (13)	+0.15 $\pm$ 15.28 (13)
Kilkenny et al. (1998)	+2.20 $\pm$ 17.04 (5)	+8.20 $\pm$ 24.00 (5)	-1.80 $\pm$ 10.59 (5)	-4.75 $\pm$ 11.93 (4^c)
SAAO (total)	+0.88 $\pm$ 11.24 (34^a)	+0.67 $\pm$ 13.96 (33)	-1.09 $\pm$ 7.98 (35)	-2.34 $\pm$ 13.92 (32^d)
Landolt (1973)	-10.00 $\pm$ 16.58 (20)	$\cdots$	-6.05 $\pm$ 8.54 (20)	-12.45 $\pm$ 15.46 (20)
Landolt (1983)	+6.31 $\pm$ 13.97 (13)	+5.00 $\pm$ 9.81 (13)	-7.31 $\pm$ 9.90 (13)	-14.92 $\pm$ 25.47 (13)
Landolt (1992)	-10.60 $\pm$ 13.10 (10)	+6.00 $\pm$ 9.42 (10)	-7.30 $\pm$ 12.51 (10)	-16.40 $\pm$ 26.95 (10)
Landolt (total)	-9.02 $\pm$ 14.84 (43)	+5.43 $\pm$ 9.44 (43)	-6.77 $\pm$ 9.81 (43)	-14.12 $\pm$ 21.26 (43)

^* in mmag ^a E661 excluded; ^b E1-R & E661 excluded; ^c GL 747.4 excluded; ^d E1-R, E661, & GL 747.4 excluded

Comparison

Landolt

Cousins (1984b) showed the difference between the SAAO E-region standard system and Landolt's equatorial standard system for the first time. Later, Menzies et al. (1991) confirmed the difference from extensive photometry of Landolt's equatorial standard stars. Bessell (1995) derived the transformation relation between SAAO measurements and Landolt's measurements in a polynomial form. We observed a few standard regions along the celestial equator to extend the colour range. We also measured a few stars that were not measured by Menzies et al. (1991) but measured by Landolt. The difference in magnitude and colour for stars in common with Landolt (1973, 1983, 1992) were also listed in Table 4 and compared in Fig. 6. The differences with respect to Landolt's measurements are consistent with those found by Menzies et al. (1991) and Bessell (1995).

The U band mismatch between the our initial natural and the standard system and shown in Figs. 2(e,f) highlights the effects that such mismatches can have. It is likely that a bandpass mismatch exacerbated by a transformation methodology is reponsible for much of the differences between the Landolt and SAAO U-B systems. The practice of transforming natural u-b colours to standard U-B using a colour term in U-B (as done by Landolt) can result in a U-B system dependent on spectral-type, reddening and luminosity for some stars. This is because a colour term in U-B implies that all stars with a given U-B will have the same correction to the standard system which is not necessarily true. Stars with very different spectra, such as late B, A or F stars or reddened B and A stars, can have the same standard U-B value but because of a passband mismatch will have different natural u-b colours. These differences can be partially corrected for using a colour term in B-V but it is obvious that if this is not done, the resultant system will have within it systematic differences for late B, A and F stars. However, even were this done, reddened stars with different U-B versus B-V loci to normal stars will not follow the same regression against B-V and so will not follow the same correction curve. Landolt's natural GaAs u-b system is not very close to the standard U-B system and so it is likely that, because the transformations were made using a U-B colour term alone, systematic U-B differences exist for some kinds of stars namely, A-F stars, reddened B-A stars and giant A-F stars. The effect of reddening appears more pronounced in Landolt's stars because he usefully provided many fainter stars in his lists ensuring a higher proportion of reddened stars compared to other lists of brighter standard stars.

It is worth noting that the difference in U-B increased from Landolt's earlier work (Landolt 1973) through his more recent work (Landolt 1992). This can be understood were the early 1P21 passbands closer to the standard passbands.

**Figure 6:** Comparison with Landolt's photometry. Triangles, open squares, filled circles represent, respectively, photoelectric data from Landolt (1973), Landolt (1983) and Landolt (1992). The dotted lines in the figure denote the mean difference between SAAO measurements and Landolt's measurements from Menzies et al. (1991). The thick curved line represents the fitted polynomial curve by Bessell (1995).
$\begin{figure} \begin{center} \psfig{file=Landolt.ps}\end{center}\end{figure}$

Graham

Graham (1982) observed many E-region stars. The purpose of his observations was to extend the magnitude range of E-region stars. He included several faint E-region stars up to $V \approx$ 17. We measured many E-region stars in the vicinity of SAAO standard stars (Table 2b). Our results are compared with Graham's data in Fig. 7. The meaning of $\Delta$ is the same as before. For fainter stars (V > 12), Graham's V magnitudes are systematically brighter than ours. V-I and B-V, on the other hand, do not show any systematic differences relative to ours. However U-B shows a large scatter for fainter stars. This is probably caused by the poor S/N of these stars in the U passband. Our photometry as well as Graham's were affected by the low signal-to-noise ratio in U. One star E6-r shows a large difference in magnitude as well as colour.

**Figure 7:** Comparison with Graham's photometry. The size of a symbol is proportional to the number of measurements.
$\begin{figure} \begin{center} \psfig{file=Graham.ps}\end{center}\end{figure}$

Next Section: Summary
Title/Abstract Page: Standard Stars - CCD
Previous Section: Atmospheric Extinction
Contents Page: Volume 17, Number 3

ASKAP

Public