Because of the enormous diffusion created by Sirius-A (mv=-1.46), the star field around Sirius is very difficult to reach, and remained long unexplored. A first observation was carried out in 1985 with the 1.5m telescope of E.S.O. by J.M. Bonnet-Bidaud of the CEA Service of Astrophysics and C. Gry of the Marseilles LAS. It revealed for the first time about fifteen stars in a 4 arcminute field around Sirius-A.
A second image was obtained, in January 1999, by J.M. Bonnet-Bidaud (CEA), F Colas (IMC) and J Lecacheux (OPM), thanks to a coronographic device adapted on the 1m telescope of Pic du Midi.

The image, obtained at Pic du Midi in 1999, of the field of (4.5x6) arcmin around Sirius. Sirius is hidden by a mask (22 arcsec) in the center of the image, of which the horizontal support is also visible. On this animation, trajectories of Sirius-A (white) and Sirius-B (red) are illustrated, the crosses showing the position of A in 1985 and 1999. Between these two dates, Sirius-A moved by 14 arcsec, whereas no star of the field changed position. The sinusoidal path of A and B is the result of the rectilinear proper motion of the center of gravity of the system and the exentric ellipse (e=0.59) of the orbital movement of A and B with a 50.09 years period. Current A-B distance is 6 arcsec.

In both cases an astrometrical study made it possible to measure the precise position of stars within a field of 3 arcmin around Sirius. In the 13 years interval between the two images, Sirius-A moved 14 seconds of arc and a companion in orbit would have to undergo the same displacement. The superposition of the two images shows that no star of the field moved by more than 0.5 arcsec. This method thus makes it possible to eliminate as probable companion all visible stars from the field.
These images in addition made it possible to reconstitute the trajectory of Sirius in the star field. They show that Sirius passed, in 1937, at an apparent distance of less than 7 arcsec of a star of magnitude 12. The presence of a bright companion, visually observed during the period 1920-1930, is thus most probably the result of this fortuitous conjunction with a background star.

The absence of proper motion of stars in the field made it possible for the authors to exclude the presence of a Sirius companion with a separation larger than 30 arcsec, corresponding to a distance of 80 astronomical units from the A-B system, down to a limit magnitude of 17.
This limit compared with the magnitudes and colors expected from a low mass star at Sirius distance, makes it possible to exclude a companion of mass higher than 0.08 Mo. Only a brown dwarf, similar to the weakest currently observed with a magnitude of 18-19, could be still undetected.

The most central area (< 30 arcsec) around Sirius remains for the moment unexplored.
The possible trajectory of a companion of Sirius is an orbit of long period (P>2000 yr) very eccentric (e>0.9) (Bonnet-Bidaud & Gry 1991). According to the orientation of this orbit on the sky, there is a possibility that a companion is currently at a very short projected distance of Sirius. A program is currently carried out with ESO to observe the area nearest to Sirius with high spatial resolution with adaptive optics. Let us note that Kuchner and Brown (2000, PASP 112, 827) also recently obtained constraints on the possible companions of Sirius, with the Space Telescope (NICMOS).

Among all the binary stars comprising a white dwarf, Sirius is a singular system. Sirius-A is the star of the youngest spectral type, and Sirius-B is among the most massive white dwarfs known.
These characteristics still largely unexplained could well be the result of the evolution of a more complex system comprising a third star still undetected.

Contacts
Jean-Marc Bonnet-Bidaud Service d'Astrophysique, DSM/DAPNIA/Sap, CEA Saclay
François Colas , Obs de Paris, IMCCE-BDL
Jean Lecacheux, Observatoire de Paris, DESPA