2 Observations and results

Our observations of LS I +61$^{\circ }$303 were carried out during 11 hours on 1994 June 7, at 5 GHz, using four antennas of the EVN, namely the Effelsberg 100 m, Medicina 32 m, Noto 32 m and Onsala 25 m telescopes. The data were recorded, in left hand circular polarization, using the Mark III mode A recording system, corresponding to a total bandwidth of 56 MHz. The correlation of the data was done on the Mark III correlator operated by the Max-Planck-Institut für Radioastronomie in Bonn.

Our EVN observations were made within a time interval during which a multiwavelength monitoring campaign was conducted that included simultaneous radio, optical, infrared, and hard X-ray/$\gamma$-ray observations and covered almost three orbital cycles of LS I +61$^{\circ }$303 (Strickman et al. 1998). As can be seen from the radio light curve taken with the Green Bank Interferometer (GBI) as part of that campaign (Fig. 1),

Figure 1: Radio light curve of LS I +61$^{\circ }$303 obtained with the GBI at 8.4 GHz (Strickman et al. 1998). The shaded area indicates the time interval during which our VLBI data were taken.

our observations took place about 10 days after the radio outburst maximum (i.e. almost 13 days after the onset of the outburst). Total flux density observations of LS I +61$^{\circ }$303, carried out by the 100 m telescope of Effelsberg at the beginning of each VLBI scan, are shown in Fig. 2.

Figure 2: Effelsberg flux density measurements of LS I +61$^{\circ }$303 at 5 GHz obtained during our VLBI observations. LCP and RCP flux density measurements are represented by filled and open circles respectively. The solid line corresponds to the total flux density.

The variations of left and right hand circular polarization flux are within their error bars. This lack of significant circular polarization agrees with previous results of Peracaula et al. (1997). The total flux density is nearly constant and with an average value of $34\pm5$ mJy.

In our EVN observations, LS I +61$^{\circ }$303 was clearly detected on all baselines. The data were reduced using standard procedures within the AIPS and Difmap software packages. The EVN uniform weighted map is shown at the top of Fig. 3.

It exhibits an extended structure elongated to the southeast. The elongation is not artificially created by the beam, in fact the elongation is at ${\rm PA} \simeq120^{\circ}$ while the beam, of size 5.9  ${\rm mas} \times 3.8$ mas, has ${\rm PA}=74.2^{\circ}$. The flux density recovered in the cleaning process amounts to $\sim$35 mJy.

In Table 1 we show the parameters obtained after fitting several models to the (u,v) data.

 

 

Table 1: Parameters of different models fitted to the data.

Model	$\chi^2_{\rm r}$	Components	Flux Density	r	$\theta$	Maj. axis	Min. axis	PA
			(mJy)	(mas)	( $\hbox{$^\circ$ }$)	(mas)	(mas)	( $\hbox{$^\circ$ }$)
1	0.798	Elliptical Gaussian	33.0	0.04	3	3.7	2.5	123
2	0.786	Elliptical Gaussian	21.6	0.5	135	5.5	3.6	123
		Point source	13.1	0.5	-43

The best fit using a single component is obtained with an elliptical Gaussian (Model 1 in Table 1). A circular Gaussian or a point source gives higher values of $\chi^2_r$. The elliptical Gaussian has a PA consistent with the observable elongation in the map. However, there is still residual emission not accounted for by Model 1, in the direction of the elongation.

In previous VLBI observations (Paredes et al. 1998) the best fits to the (u,v) data were consistent with an unresolved core plus a Gaussian halo. Hence, we have added a point source to Model 1, and we have obtained a slightly improved fit to the data (Model 2 in Table 1). It is important to note that, in this model, the elliptical Gaussian conserves its PA and moves towards the direction of the elongation, while the point source moves in the opposite direction. We have tried several different initial conditions and always the fit converged to the same solution. Finally, for a distance of 2.0 kpc (Frail & Hjellming 1991), the brightness temperature of the components of the models listed in Table 1 are in the range of $T_{\rm B}=10^{7}$-10⁸ K, which are characteristic of non-thermal emission.