Our observational program was mainly oriented on monitoring the medium and long-term changes of Be stars. In many cases, only one series of observations of a particular star was obtained during the observing night. Later, however, some Be stars were also observed more intensively to study their rapid changes. Here, we report in detail only the cases of well-documented variations; otherwise, we only estimated possible timescales of variations on a statistical basis.

Information about the character of variability of frequently-observed stars is summarized in Table 4 (click here). The bibliographical remarks in this section are not aimed to provide complete reviews of known variability of individual stars. We only refer to important or principal reports of photometric variability of stars in question. Spectroscopic changes are reported only if they are clearly related to the photometric variations observed. In such cases, we shall often refer to the study of Hubert-Delplace & Hubert (1979), hereafter referred to as "Atlas''.

Those stars of Table 1 (click here) for which we cannot conclude anything about their variability, and for which we were unable to find any report of light variability in the astronomical literature, are simply omitted from the discussion below.

Percy & Welch (1983) found variations of 005 in b on a time scale of 3 - 8 days, but no significant variability on a time scale of hours. Rapid variations were reported, however, by Elst (1979) and by Badalia & Gurm (1982). Rather limited Hvar observations have the rms errors below the detection threshold for variables established from the observations of check stars.

This is the brighter component of the visual binary ADS 622. o Cas is a known long-term spectral variable, which changes between the B and Be phases cyclically over intervals of a few years. Photometric variations of o Cas were suspected by Haupt & Schroll (1974). Also Alvarez & Schuster (1981) and Schuster & Guichard (1984) listed it as a variable. Hvar observations of 1982-1984 have shown significant photometric activity of this star (Horn et al. 1985). The variations are shown in Fig. 1 (click here). Rapid flare-like brightenings were recorded during two consecutive seasons. Moreover, a secular brightening over the whole interval of observations can also be noted.

This is one of the most frequently observed Be stars, both in the optical and in other spectral regions. The long-term light variability was summarized by Doazan et al. (1983). Only a few observations were obtained during two seasons at Hvar; they fit into their light curve.

This star is the brighter component of the close visual binary ADS 940 composed of two B stars with an orbital period of 371.6 years (Abt & Cardona 1984). Only a weak and variable H

emission is reported in Atlas. Alvarez & Schuster (1981) have reported this star as "possibly variable'' while Bož ić et al.\ (1995) argued in favour of its secular constancy. No light or colour changes are indicated by the Hvar observations.

Figure 3: Color behaviour of

Tau during long-term light variations. Luminosity classes V, II and Ia represented by a solid, dotted and dashed line, respectively. Epoch of the first and last observation is indicated

This Be star is a peculiar double-line spectroscopic binary consisting of two emission-line objects. Its very detailed spectroscopic and photometric study was published by Bož ić et al. (1995) who critically summarized the existing knowledge and analyzed all published radial-velocity and photometric data along with new spectroscopic and photometric observations. Complicated low-amplitude light variations are a combination of the long-term ones (on a time scale of years), orbital variations with a period of 1266731 and rapid variations on a time scale of less than a day. The long-term variations show a positive correlation between the H

emission strength, brightness and colours. For this type of correlation, the apparent photometric spectral type is gradually changing from a main-sequence object to a bright supergiant as the brightness and emission increase and vice versa (Harmanec 1983b).

Table 4: The incidence of light variability among Be stars observed at the Hvar. Only well-observed stars are included here. The early- and late-Be stars are separately presented on the left and right side of the Table, respectively. Abbreviations L, R, O, and S denote (by black dots) the presence of variations on a time scale of years and months (L), days or fraction of day (R), orbital (O) and sudden light changes (S), respectively

Alvarez & Schuster (1981) list this star as a possible variable. Scarce Hvar observations show only marginal light variations in the V passband and apparently no variations above the detection threshold in B and U.

Alvarez & Schuster (1981) list this star as a variable. This supergiant is definitively variable according to the rms errors in all three passbands (cf. Table 1 (click here)).

This object was announced to be a double-line spectroscopic binary with a cool secondary (Polidan & Peters 1976) but this finding has never been confirmed. A gradual increase of the emission strength between 1953 and 1976 is reported in Atlas but no light variations have been detected. Over almost 4 months of intensive monitoring at Hvar during the 1980/81 season, no light and colour variations exceeding 002 were detected (Pavlovski & Bož ić 1982). The constancy of the star was confirmed by continuing Hvar observations.

RX Cas is a strongly interacting binary in which the hotter and more massive component is completely hidden in an accretion disk which mimics a stellar photosphere of a spectral type later than the central star (Andersen et al. 1989). Emission in the Balmer lines is seen during the whole orbital cycle and - if not seen equator-on - this object would certainly be classified as a typical Be star. It was observed at Hvar from 1975 to 1977 (cf. K°ı ž et al. 1980) and in 1982. A pronounced long-term modulation of the light curve is present. Kalv (1979), who observed the star systematically, concluded that its light curve varied with a period of 516 d. Cycle-to-cycle variations of the orbital light curves, possibly similar to that of RX Cas, were also found for some other Be binaries (CX Dra, KX And,

Lyr, V360 Lac).

This star is a member of the

Per cluster. A shell spectrum was once observed by Kraft (1967). Hvar observations scatter within 004 in all three passbands.

This star is also a member of the

Per cluster. Slettebak (1982) found no obvious emission-line variations between 1949 and 1980. Neither light nor colour variations were detected at Hvar.

The long-term light variations of this star are well documented especially thanks to systematic UBV observations by Sharov and Lyutyj (see Sharov & Lyutyj 1992, 1994 and references therein). A few Hvar observations fit in their light curve.

This star shows only slight variations in the emission line profiles and brightness. Light variations for about 002 on a time-scale of hours were reported by Percy & Lane (1977). Limited Hvar data provide no compelling evidence of light variations. However, the rms error in the V passband is 0019, hence, some slight variations are probably present.

We are not aware of any report of photometric variability. A very slight change of the H

emission is reported in Atlas. No photometric variations larger than the detection threshold were found at Hvar.

In spite of many spectroscopic studies, only a few photometric measurements were published. Percy et al. (1981) observed probable variations of about 002 within a few hours. Alvarez at al.\ (1987) reported small quasi-periodic fluctuations on a time scale of days (or possibly years). Limited Hvar data suggest light variations in V but not in B or U.

Variations up to 01 in V were detected by Pavlovski & Bož ić (1982). Bossi et al. (1989) confirmed the variations on a short time-scale, but not their large amplitude. They interpreted them as multiperiodic nonradial pulsations.

This is a single-line spectroscopic binary with a period of 1329735 (Harmanec 1984c). A light decrease by 02 during 1968 - 1969 was reported by Haupt & Schroll (1974). Results of systematic Hvar photometry were published by Bož ić & Pavlovski (1988). Three types of variations are combined in a complicated way:

I) Secular light variations with an amplitude of a few tenths of the magnitude on a time-scale of years;

II) peculiar phase-locked orbital changes: a light minimum centered on the phase of expected mid-eclipse of the Be star was observed during some cycles but not on others; and

The latter finding is especially interesting in the light of the spectroscopic result of Yang et al. (1990). Using high-S/N spectra, they found line-profile variations with periods of 0683 and 0110, and also with the orbital period. However, a recent analysis of a series of He I

4921 Å line profiles by Hahula & Gies (1994) revealed single-periodic variations with a period of

d. Complete Hvar photometry of

Tau is shown in Fig. 2 (click here) while the accompanying colour changes are presented in Fig. 3 (click here). Recently, new photometric observations were obtained by Percy et al. (1994).

**Table 5:** Spectroscopic binaries
HD	Name	Sp. class	Period	Re
10516	Per	B2Vep	1266731	1
-	RX Cas	(B5)+K1III	3233	2
37202	Tau	B1IIIpe	1329735	3
41335	HR 2142	B2Ven	80860	4
109387	Dra	B6IIIpe	6155	5
142926	V839 Her	B7Ve	46194	6
162732	V744 Her	B6Ve	86722	7
174237	CX Dra	B2.5Ve+F5III	669603	8
183656	V923 Aql	B6e	21475	9
187399	V1507 Cyg	B8IIIe	279705	10
200120	V832 Cyg	B1ne	2914	11
200310	V1931 Cyg	B1Ve	248?	12
216200	V360 Lac	B3IVe	100854	13
217675	o And	B6IIIpe	330847	14
218393	KX And	B2IIIe+KOII	38918	15
224559	LQ And	B4Ven	7413	16

Sources: 1) Bož ić et al. (1995); 2) K°ı ž et al. (1980), Andersen et al.\ (1989); 3) Harmanec (1984c); 4) Peters (1983); 5) Juza et al. (1991); 6) Harmanec et al. (1976); 7) Doazan et al. (1982); 8) Koubský (1978), Horn et al. (1992), Šimon (1996); 9) Koubský et al. (1989); 10) Pavlovski et al. (1979); 11) Tarasov & Tuominen (1987); 12) Harmanec et al. (1986); 13) Hill et al. (1996); 14) Harmanec et al. (1987a); Hill et al.\ (1989); 15) Štefl et al. (1991); Floquet et al.\ (1995); 16) Matthews et al. (1991).

Small variations with a scatter of 004 in season 1981/82 were reported by Pavlovski & Bož ić (1982) from Hvar observations.

This Be star is a spectroscopic binary with a period of 80860 which regularly exhibits phase-locked shell phases lasting about 3 days prior, and 1 day after the lower conjunction of the Be component (Peters 1983). Light variations of a small amplitude were reported by Lynds (1959a). No measurable light variations corresponding to recurrent shell phases were found by Harmanec et al.\ (1983) and Dorren et al.\ (1984), who also excluded the presence of a 013 light decrease in January 1983, reported by Sterken (1983). However, long-term variations with an amplitude as large as 01 in u were found by Mennickent et al. (1994). Periodic light variations with a period of either 369 or 079 and an amplitude growing towards ultraviolet wavelengths were reported by Barrera et al. (1991). Rapid variations were also reported by Adelman (1992).

Peters (1976) classified the star as a long-period binary, but no report of light variations has been published as yet. The star is not a light variable according to limited Hvar observations.

Alvarez & Schuster (1981) list this star as possibly variable. Hvar observations have the rms errors just above the detection threshold.

Although Figer (1981) announced rapid light variations based on visual estimates in 1979, 1980 and the beginning of 1981, two later studies (Bož ić et al. 1982; Poretti 1982), based on photoelectric photometry, did not confirm any variations above the detection limit. Very pronounced emission-line variations of this star over two decades were reported in Atlas. This prompted us to include this star among regularly observed objects during the season 1981/82. No convincing evidence of rapid changes was found. However, the Hvar data indicate a secular brightness decrease over four consecutive seasons. Pronounced long-term emission-line changes were observed by Hubert-Delplace et al. (1982).

This bright Be star was classified as strongly variable by Schuster & Guichard (1984). Juza et al. (1994 and references therein) confirmed the long-term photometric variations, known for a long time, and derived the period of intensity variations of the Balmer emission lines and continuum polarimetry of about 23 years. The brightness attains maximum during the rise of the emission while the brightness minimum coincides with the maximum strength of the Balmer emission.

This is a relatively little-studied bright Be star. Roark (1971) detected a large decrease of its brightness (07) with superimposed rapid variations. Harmanec (1983c) speculated that

CrB could be an eclipsing binary with a period of about 511 days. This speculation was based on the coincidence between the time of primary minimum derived from radial velocity data and the occurence of a steep light decrease observed by Roark (1971). However, the subsequent spectroscopic and photometric data (Hubert et al. 1990; Percy et al. 1988b; Papouš ek 1986 and Fernandes et al.\ 1985) safely excluded Harmanec's suggestion. Significant photometric variations on a time scale of days occurred in 1986 and 1987 (Percy et al. 1988b). Percy & Attard (1992) did not detect any significant photometric variations of this star in 1990 or 1991. Fernandes et al. (1985) also found the star constant within a few hundreds of a magnitude in May, June and July 1984. Guerrero et al. (1992) reported rapid light variations with a full amplitude of about 003 which they interpreted as a combination of a slow variation (possible period of about 8 d) and a rapid variation with either 0869 or 0459 period. According to Hvar data, this star is a low-amplitude light variable, just above the detection limit.

This star is an interacting binary (Table 5 (click here)). The long-term spectral changes (transition Be

B) were summarized by Koubský et al. (1994). Hill et al. (1976), Landis et al. (1977), Papouš ek (1979), Dapergolas et al. (1981), and Alvarez & Schuster (1981) did not detect any phase-locked light variations and concluded that the star was apparently constant in light and colour for quite a long time. Percy & Attard (1992) reported fading by 009 in B and V and 004 in U in 1991. Our systematic monitoring of the star for 19 years clearly reveals the presence of the long-term light and colour changes with a positive correlation between brightness and emission-line strength (Fig. 4 (click here)).

This star is a single-line spectroscopic binary with a well determined orbit (Table 5 (click here)). Long-term light variations were first announced by Harmanec et al. (1978) and further investigated by Doazan et al. (1982) and by Barylak & Doazan (1986). The star has shown only slight variations on time scales from days to months (see, e.g., Baldinelli et al. 1981). The long-term light and spectral changes have an inverse correlation between the brightness and emission-line strength (Doazan et al. 1982). Here, we show (Fig.\ 5 (click here)) a smoothed light curve defined by the Hvar data and also the colour behaviour of 88 Her in the U-B vs. B-V diagram (Fig. 6 (click here)).

Cousins (1952) observed small and irregular brightness variations within 007. Page & Page (1970) reported large flare-like events detected on patrol photographs on two different nights in 1969. The star was listed as variable by Haupt & Schroll (1974) and as a suspected variable by Hill et al.\ (1976). The latter authors abandoned their observations after the season 1971 since they did not succeed to confirm any variations in excess of 002. Percy et al. (1988b) detected small variations on a time scale as short as a few hours in each season between 1981 and 1987. A monotonic decrease of 01 in b during 60 days was found by Adelman (1992). The star is reported as a spectroscopic binary (Hoffleit 1982) and RV variable (Crampton 1968). Conspicuous variations of emission line-profiles and other kinds of spectral variability were described by Reynolds & Slettebak (1980), Andrillat & Fehrenbach (1982), Grady et al. (1987), Peters (1988) and Hanuschik et al.\ (1995), among others. At Hvar it was observed from May to July 1982. The measurements definitely indicate variations (Fig. 7 (click here)), but their true period is still unclear. Both, a period close to 1 day or periods of 22-25 days fit the Hvar data.

Alvarez & Schuster (1981) listed this shell star as a possible variable. We observed it in four seasons beginning 1982 and found some indication of variability (cf. the rms errors in Table 1 (click here)).

This star was classified as a Be star by Irvine (1975) who observed H

emission in its spectrum. Hvar observations indicate that the star varies with a period of about 55 d.

Alvarez & Schuster (1981) listed this star as possibly variable. Percy et al. (1988b) classified NW Ser as an "active" star which showed short-term variations up to 01 on time scales from hours to years. In a subsequent paper, Percy & Attard (1992) confirmed the short-term variability. The star was observed at Hvar since 1980. It is definitely a short-term variable, with an amplitude of about 015 in all three passbands (Fig. 8 (click here)). Moreover, it seems that the amplitude of the light changes varies from season to season.

This is an interacting Be binary (Table 5 (click here)). The orbit was determined for both components (Horn et al.\ 1992; Šimon 1996). The long-term fluctuations of its brightness were found by Koubský et al.\ (1980a). Moreover, low-amplitude orbital residual light variations were detected. The character of the orbital light variations is reminiscent of an ellipsoidal variable. However, large cycle-to-cycle variations in the depth of the primary minimum were observed (cf., e.g., Harmanec 1987b). The long-term variations were confirmed by Percy et al.\ (1988b) and by Percy & Attard (1992). The binary was regularly monitored at Hvar in almost all seasons since 1974.

Regular light variations of this early B star with period of about 0.61 d and a range of 0.06

were found by Lynds (1959a) in his photometric survey. The variability has been confirmed by Percy et al. (1988b) and is also obvious from the Hvar data. From a recent line-profile study, Hahula & Gies (1994) reported changes with a period of about 1.27 d. The Hvar observations indicate low-amplitude variations with a period of either 06 or 12.

The star is a single-line spectroscopic binary (Table 5 (click here)). It is also one of the first stars for which periodic rapid light variations were found (a period of 0.8518 day and a variable amplitude up to 01, Lynds 1960). V923 Aql exhibits variations in radial velocity and V/R ratio of the double emission peaks on a time scale of about six years (Koubský et al. 1989). Small photometric variations on several time scales were reported by Percy et al. (1988b). Mennickent et al. (1994) indicated quasi-periodic long-term variations with a period of about 7 years and an amplitude up to 025 in u and short-term variations up to 01 within a few days. We observed V923 Aql at Hvar during 9 seasons. Significant light and colour variations are found (Fig. 9 (click here)) on both, long-term and short-term time scales.

This star was used as a secondary photometric standard for the Johnson UBV system. Its variability was first discovered by Lynds (1959a). History of its light variations was summarized by Horn et al. (1982b). Photoelectric observations were also accumulated by Percy et al. (1981, 1988b) and Percy & Attard (1992). Correlated long-term photometric and spectroscopic variations were investigated by Horn et al. (1982b), Horn et al. (1983) and Alvarez et al. (1987). Mennickent et al. (1994) confirmed the cyclic light variability of this star with a period of about 5.9 years. They reported also occasional short-term variations up to a few 001 in 10 days. The Hvar observations cover 9 seasons, starting 1975 (Fig. 10 (click here)).

13-colour photometry of the star was published by Schuster (1982). Rather scarce Hvar observations did not indicate any light variations.

Detailed spectroscopic study by Hutchings & Redman (1973) of this long-period binary system revealed its complex nature. They suggested that the primary, i.e. the star dominating in the optical spectrum, loses mass while the more massive component is observationally undetectable due to its huge circumstellar envelope. In this respect, V1507 Cyg is similar to

Lyr, RX Cas and other strongly interacting binaries of the W Serpentis type. The light variations are complex and without an a priori knowledge of the orbital period, photometric observations can hardly be properly phased. Hvar observations span 3 years (1976 - 1978) and consist of 38 UBV measurements. The orbital light curve is peculiar but stable (Pavlovski et al. 1979). Even, when compared to previous photometric masurements by Hill et al.\ (1976), no pronounced changes in the shape of the light curve were noticed as could be expected for such an "active'' binary.

Photometric observations of this star were described by Percy & Lane (1977), Percy et al. (1981) and Percy et al. (1988b). Comparing these studies one can conclude that if the star were variable at all, it would have a very small amplitude (001) and would vary on a time scale of hours. The star was observed at Hvar since 1985 in three seasons. Neither rapid nor slow variations above the detection threshold were found, although some suspicion of variations with a very small amplitude remains.

Light variability was first detected by Gies & Percy (1977). Since then, the star attracted attention of several photometrists. Rapid variations were confirmed by Percy & Lane (1977), Spear et al. (1981) who reported a periodicity of about 07, Percy et al.\ (1981), Schuster & Guichard (1984) and by Adelman (1992). Peters & Penrod (1988) obtained simultaneous IUE and ground-based spectral observations. They confirmed conspicuous mass-loss activity and moving sub-features in the optical line profiles. They interpreted the observed effects as NRP with a low-order mode l = 2 and a period of 068, and a high-order mode l = 10 with a period of 013. Pavlovski & Ruž ić (1990) analyzed Hvar measurements obtained during the same season as the Peters & Penrod (1988) data. Light variations with an amplitude of 008 in V could be reconciled with either a 068 period (sinusoidal light curve) or a 136 period (a double-wave light curve). However, the authors concluded that the scatter around the mean curve was much larger than the observational errors. Using a Fourier decomposition, they were able to model the observed variations by a sum of four sinusoids with different periods. Ruž ić et al. (1994) attempted to analyze all available photometric data on this star. They tentatively concluded that the variations are either controlled by two principal periods, 0690 and 111 or that regular variations are being disturbed by some stochastic process. The light amplitude varies from season to season. Bossi et al. (1993) found several different periods in the line-profile variations of HeI

6678 Å and H

, including those reported by Peters & Penrod, and by Pavlovski & Ruž ić. Recently, Hahula & Gies (1994) analyzed series of observations of He I

4921 Å line and found a single-periodic variation with a period of 064. They attributed the variation to an m = -2 NRP mode. Early Hvar observations of this star suffered from the fact that (at that time unknown) variable, V1644 Cyg = 29 Cyg (HD 192640) was originally used as the comparison. During the final reduction, differential magnitudes were derived relative to many different, accidentally observed comparison stars and are not, therefore, particularly suitable for analyses of low-amplitude changes. All reliable photometric observations obtained at Hvar are shown in Fig. 11 (click here)s.

Figure 12: Photometry for 20 Vul, 25 Vul, QR Vul, and their check star 18 Vul, respectively

Historical observations of this luminous blue variable star can be found in Lamers & de Groot (1992). The 1985 and 1986 Hvar observations were incorporated into the study by Percy et al. (1988a). They found irregular variations up to 02 on time scales ranging from a few days to a few months.

No report of light variability was found. Hvar observations show this star to be non-variable (cf. Table 1 (click here) and Fig. 12 (click here), respectively).

This relatively bright Be star has been almost neglected by photometrists. Our data which cover 3 consecutive observing seasons clearly revealed light variations on a short-term scale (Fig.\ 13 (click here)).

No report of variability was found and no changes in the emission strength between 1953 and 1975 are reported in Atlas. This star was observed at Hvar since 1980 and seems to be constant at our detection treshold.

Percy et al. (1988b) report small photometric variations on a time scale of weeks to months. Apparently there were no short-term variations. On the other hand, Percy & Welch (1983) present indications of shorter changes on a time scale of 5 to 10 d. Rufener & Bartholdi (1982) list this star among microvariables. Hvar observations (cf. Fig. 14 (click here)) show significant changes on a time scale of days, with a possible pseudoperiod of 45284.

Figure 16: Phase diagram of V360 Lac in the V passband folded with a period of 100854 for the Hvar observations

Summary of observations of this star in 1986 and 1987 was given by Percy et al. (1988b). Slight variations on a time scale of days to months can be deduced from their data.

This bright and frequently observed Be star is a member of the Cyg OB7 association and the brightest component of a multiple system of stars. It is characterized by variable emission lines and a temporary shell of hydrogen, helium and ionized metals and by variable mass-loss (studied intensively in UV superionized lines, e.g.\ Doazan et al. 1989 and references therein). For a long time, it has been known as a light variable. Lynds (1959b) reported a steady decline of brightness by 01 during five months in 1958. Large changes in V (

04) and in the near IR was observed in 1977-1979 by Schuster & Alvarez (1983). Rapid variations on a time scale of days were reported by Percy et al. (1988b) and Hadrava et al. (1989).

Light variations were found byHarmanec et al. (1986). They found a possible period of 248, which was consistent with the published radial velocity variations, and suggested that the star was a spectroscopic binary. Percy et al.\ (1988b) confirmed that the star varies on a time-scale of days, but their data did not support the 248 period. Rapid line-profile variations were detected by Koubský et al. (1995).

A summary of observations of this star in 1986 and 1987 was given by Percy et al. (1988b). Slight variations on a time scale of days to months can be derived from their monitoring. Alvarez & Schuster (1981) list this star as possibly variable.

We are not aware of any report of light variations. No convincing evidence of light variations was found from one season of monitoring at Hvar.

This is a relatively little-studied Be star. Photometric observations were presented by Alvarez & Schuster (1981) and Schuster & Guichard (1984). They found no evidence of light variability. No evidence of light variations was found during 7 seasons of observations at Hvar.

Walker (1952) discovered variability of 14 Lac with an amplitude of a few 001. Hill et al. (1976) confirmed the light variability and concluded that 14 Lac is an ellipsoidal variable with a period near 20 days. Mantegazza (1980), clearly unaware of these previous results, found evidence of Balmer emission and reported light variations with a possible (pseudo)-periodicity of about 5 days. This fact prompted Morris (1985) to reject the star from his list of ellipsoidal variables. Hvar observations extend over 8 seasons. Period analysis of combined Hvar and MtKobau observations yield a period of P = 100854, similar to that found by Hill from radial velocities (see Richardson 1977). Phase diagram of the light and colour variations is shown in Fig.\ 16 (click here). A detailed analysis of the combined photometric and spectroscopic data has been carried out by Hill et al. (1997).

This often studied Be star has exhibited a strong shell spectrum since 1940. Its light variations were discovered by Walker (1953). He was able to reconcile "periods'' in the range of 07 - 08 in some series of observations, but no apparent periodicity was present in other epochs. Lester (1975) confirmed the variations with a characteristic cycle of

07 from his uvby observations. Harmanec et al. (1979) demonstrated the presence of long-term light variations between 1972 and 1978. Percy (1981a) found periods of either 0725 or 0418 from his 1980 observations but expressed doubts about the presence of a real periodicity. Similar periods were found by several authors: 072 (Pavlovski 1983, 1987), 075 (Iliev et al. 1984), 0716 (Stagg et al. 1988). Pavlovski (1983) pointed out the double-wave character of the light curve. Later, he derived a multifrequency fit with 3 periods of 0364, 0784 and 0619 (Pavlovski 1987).

The star shows conspicuous spectroscopic variations on the same time scale. Multifrequency analysis of He I 6678 Å line profiles and equivalent width (Floquet et al. 1992) and of rapid V/R variations of the H

line (Pavlovski & Schneider 1990) led to periods close to those derived by Pavlovski (also, Pavlovski et al. 1993a) from the light variations and a longer one, 222. Results of the above mentioned multifrequency analysis for the shorter periods have been interpreted as nonradial pulsations while the longer period was suggested to be due to inhomogeneities slightly above the photosphere. Hvar photometric measurements of EW Lac are displayed in Fig. 17 (click here). Both long-term and rapid variations are obvious. Full UBV amplitudes of the long-term changes exceed the amplitudes of the rapid light variations, and amount to 035, 035, and 065 in V, B, and U passbands, respectively. Seasonal changes of the amplitudes of the short-term variations, first observed for o And (see below), are clearly seen in Fig. 17 (click here). Colour changes are also significant (Fig.\ 18 (click here)) and the star becomes redder when brighter in both

and

indices. However, changes in the

are not linear and their maximum does not coincide with the brightness extremum. This effect was also found for several other Be stars by Dachs et al. (1988) who interpreted it as due to selective contributions of circumstellar envelope and underlying stellar photosphere in the course of significant changes of the envelope structure.

Schuster & Alvarez (1981) listed HD 217543 as a variable star. Jerzykiewicz (1993) confirmed its variability with an amplitude of 0018 without giving details. Our observations clearly indicate variability - see the rms errors in Table 1 (click here) and Fig. 19 (click here), respectively.

McAlister et al. (1983) have confirmed that o And is a speckle interferometric triple star. Later Harmanec et al. (1987a) and Hill et al. (1988, 1989) have shown that o And is a quadruple system. The Be variable is the brightest component o And A. At a distance of about 0

3, o And B apparently revolves in a common orbit with A - its estimated orbital period being well over 100 years. Component B itself is a double-line spectroscopic binary with a period of 3308 composed of two B6-B8 stars. There is a faint fourth component a at a distance of some 0

05. If this component is also gravitationally bound to A, the expected orbital period is about 10 years. Since the beginning of this century several transitions B-Be-Be shell and vice versa of o And A were reported. Since 1918, o And has been known to show rapid photometric variability (Guthnick 1918). The photometric history of the star has been summarized by Harmanec (1983b). Harmanec (1984b) suggested that the available photometric observations could be explained as superposition of a long-term variation with an amplitude of 01 in B and V and a short-term periodic variation of 1571272. The 1.571-day curve has a double-wave character and is secularly changing both in amplitude and shape. These variations are connected to the long-term photometric and spectral changes. Harmanec (1984b) also suggested that the long-term cycle (photometry and reappearance of shell lines) is 8.5 years long and might correspond to the period of the system o And Aa. The presence of the 1.571-day modulation in the light of o And was confirmed by Harmanec et al. (1987b) and during two international campaigns - Stagg et al. (1988) and Sareyan et al. (1992). Figures 20 (click here) and 21 (click here) show the light and colour changes, respectively, observed at Hvar between 1973 and 1990. Long-term variations with peak-to-peak amplitudes of 023 in V, and up to 033 in U are well seen. Also, significant colour changes occured over the same cycle of about 8-9 years, in good agreement with Harmanec's (1984b) result.

A comprehensive review of the observational history of this star was given by Štefl et al. (1991). Light and colour variations of the star were discovered by Harmanec et al.\ (1977b). Štefl et al. (1991) showed that the light and colour variations are consistent with the period of 38919 derived from the RV variations. The variations are the most conspicuous in the U magnitude and in the U-B but their amplitude varies significantly from cycle to cycle. Low-amplitude light variations on a time scale of about half a day were reported by Stagg et al. (1988) and further studied by Pavlovski & Ružić (1989). The conclusion of Štefl et al.\ (1991) that the 389-period reflects the orbital motion in an interacting binary system was confirmed by Floquet & Hubert (1991) and Floquet et al. (1995). They discovered a K-type secondary in the near-IR region and derived half of its RV curve. The orbital light variations are not caused by stellar eclipses. The rapid light variations may be a result of processes in the accretion stream or in the region of its interaction with the disk.

Earlier photometric and spectral studies of KY And were summarized by Pavlovski & Ruž ić (1988). They analyzed rapid light variations observed in 1988 at Hvar but failed to find any strict periodicity. They suggested that variations of the star may be multiperiodic, with a dominant period of 075. Also Stagg et al.\ (1988) were unable to reconcile the rapid changes of KY And, observed during the multi-site observing campaign in 1983 with a single periodic process. Recently, Hahula & Gies (1994) analyzed a series of observations of He I

4921 Å line-profile and found periodic variations with a single period of 079 which they attributed to low-order NRP mode with m = -2. Hvar photometric data are displayed in Fig.\ 22 (click here). Light variations are rapid and have a large amplitude.

The light variability of HD 224559 was discovered by Provin (1953). Percy (1983) derived a sinusoidal light curve of LQ And with a period of 0307 and an amplitude of 003 in ubvy. Harmanec (1984a) argued that Percy's observations could be better represented by a double-wave light curve, with a twice as long period of 062. The ambiguity in the light-curve shape and period was not settled with new observations obtained during three intercontinental campaigns (Stagg et al. 1988; Sareyan et al. 1988; Harmanec et al. 1991). Only Matthews et al. (1991) resolved it finally in favour of Harmanec's longer period. Moreover, they concluded that LQ And might also be a spectroscopic binary with a period over 7 d.

Hill et al. (1976) list this star as a possible irregular variable. Guinan et al. (1982) confirmed that the star varied on a time scale of several days with a full amplitude of about 01. They could not exclude that the variations were in fact periodic, with a possible period of about 2 days or shorter. Our limited data seem to indicate some variability, too.

3. Variability of individual Be stars