1 Introduction

Photometry from the Hipparcos Satellite (European Space Agency (ESA) 1997) can be used to study stellar variability. This absolute photometry database contains about 100 observations for each measured star with most of the observations bunched into several intervals. I investigate the Chemically Peculiar stars of the Upper Main Sequence except for the $\lambda$ Boo stars. Single normal B, A, and F main sequence band stars for the most part are non-variable. Exceptions include the $\beta$ Cephei, 53 Persei, and slowly pulsating B stars in the B stars and the $\delta$ Scuti stars in the late A and early F stars.

Whether or not the non-magnetic CP stars: the Mercury-Manganese (HgMn) stars and the metallic-line (Am) stars (which are not also $\delta$ Scuti stars) are intrinsically constant still remains to be determined as the literature contains claims of low amplitude variability for some of these stars, many of which have been refuted for the better known stars. Still a few claims for variability might be valid. In looking for stars which might be investigated further to settle this question, I used both the standard error as given in the Hipparcos Catalogue and the "amplitude'', the difference between the 95 th and 5 th percentile magnitudes from its Photometry Annex. The Hipparcos Catalogue gives the scatter (one-half the difference between the 85 th and 15 th percentile magnitudes). Although the scatter and amplitude should correlate with one another and with the standard error, the amplitude might be more sensitive to potentially eclipsing systems. Tables 1, 2, and 3 give the Hipparcos photometry for the HgMn, Am, and magnetic CP stars, respectively. For each star, two identifiers are normally given, the HR or HD number and the Hipparcos Catalogue number, the mean Hipparcos magnitude, its standard error, its amplitude, and variability comments including the period if known except for the magnetic CP stars.

The standard error, the amplitude, and the scatter tend to increase with mean magnitude. These values are due to the intrinsic variability of the star as well as to errors in particular those due to photon statistics. Hence a noise model is needed in investigating variability. When one plots the standard error vs. the mean Hipparcos magnitude of the HgMn stars (Fig. 1), there is a lower bound at each magnitude which is consistent with that for the two other types of stars which are studied in this paper. The smallest standard errors occur among the HgMn stars at 0.0003 mag. If one looks instead at normal stars, then the smallest standard errors are often 0.0002 mag (Adelman 1998). The values of the lower bound increase with magnitude as would be expected due to photon statistics with the value for each magnitude increasing by 1.58 and the value at 5 th magnitude being 0.0002 mag. These values are given in Table 4.

  

Figure 1: The standard error vs. mean Hipparcos magnitude diagram for the HgMn stars. Closed diamonds indicate three definitely variable stars according to Hipparcos photometry. The other stars are indicated by plus signs

  

Figure 2: The amplitude vs. standard error diagram for HgMn stars

Usually for each type of star, the standard errors are proportional to the amplitude as seen in Fig. 2 for the HgMn stars. But there are some stars which appreciably deviate from the mean relation such as the eclipsing HgMn star binary AR Aurigae. As some CP stars are more variable in particular wavelength ranges than others, using Hipparcos magnitudes to search for photometric variability may not find all such stars.