1 Introduction

GRS 1915+105 (Castro-Tirado et al. 1994) is the prototypical microquasar, a galactic X-ray binary ejecting plasma clouds at $v\approx0.92$c (Mirabel & Rodriguez 1994). It exhibits unique X-ray variability patterns (Greiner et al. 1996) which have been interpreted as accretion disk instabilities leading to an infall of parts of the inner accretion disk (Belloni et al. 1997). Based on its X-ray properties GRS 1915+105 is suspected to be the most massive stellar black hole candidate in the Galaxy (Morgan et al. 1997). It is one of only two galactic sources which are thought to contain a maximally spinning black hole (Zhang et al. 1997). It is therefore of great importance to know some details about the system components in order to understand the conditions which lead to the unique X-ray, radio, and IR characteristics.

Previous infrared spectroscopy at Palomar (Eikenberry et al. 1998), UKIRT (Mirabel et al. 1997; Harlaftis et al. 2001) and the VLT (Martí et al. 2000) has shown that the He II emission line is variable, most probably depending on the X-ray activity. In particular, Eikenberry et al. (1998) found variations in line flux of 5 on 5-10 min timescales, suggesting that these IR lines are radiatively pumped by (presumably) jet ejection events rather than high X-ray luminosity. Based on the IR spectral variability and on its position on the IR H-R diagram, Castro-Tirado et al. (1996) suggested that GRS 1915+105 is a LMXB. On the contrary, based on the detection of He I, Mirabel et al. (1997) and Martí et al. (2000) suggested that the donor in GRS 1915+105 is a high-mass O or B star, and that accretion occurs predominantly from the wind of the donor.