1995A&AS..110...59T

An analysis of 66 old and 651 recent electronic and photographic spectrograms of the archetype line-profile variable ε Per led to the following conclusions: 1) ε Per is the primary component of a binary system. There are radial-velocity variations with a full amplitude of ∼30km/s with a period 14.076d which can be traced back to old Allegheny observations from 1907-1912. The orbit is not circular but has a high eccentricity of 0.540±0.045. 2) The 14-d binary system probably moves in orbit with a distant, as yet unknown third star (which is not identical to the visual component ADS 2888B). A likely period is 4156d which then leads to an eccentric orbit with e=0.40±0.26 but several longer periods are also possible. The existence of the putative third component to ε Per could easily be checked by speckle-interferometric observations since its expected angular distance from the close pair in the sky does not decrease below ≃0.04" even at periastron (for the shortest of the possible periods). 3) Some basic physical elements of the component stars and system dimensions were estimated as follows: primary: mass (13.5±2)M_☉ , radius (6.9±0.2)R_☉, B0.5IV-III, T_eff=(27600±1000)K, log g =3.85±0.15[CGS], He/H=0.23, logε(O)=8.89, logε(Si)=7.51; secondary: mass between 1 and 2M_☉; tertiary: mass between 2.3 and 5.3M_☉. The semimajor axis of the 14-d orbit is about 60R_☉, that for the wide orbit 2800R_☉ or more (depending on what is the correct value of the period). The data presented here make the membership of ε Per in the α Per cluster very improbable.