4 Sources of new data

Most of the new solutions given in the following sections were made possible because of the availability of several pieces of information that were not used during the data processing for lack of time, or simply because the information did not exist at that time or was overlooked. Essentially, the new data consists of updated values for the relative astrometry and/or photometry of double stars.

As mentioned above the grid step error could be avoided if a good a priori separation and position angle could be secured to select the most likely solutions among the two or three possible. It is worth emphasising that for the vast majority of the double stars solved from the Hipparcos data, the processing was self-sufficient and did not need to rely on a good starting point. Only when the double star signal was too weak or the observation equations poorly conditioned was the starting value more crucial. This concerned only a small fraction of the Hipparcos double stars, but a large fraction of the subset reinvestigated in this paper.

Several sources were used to reprocess the Hipparcos observations of the selected doubtful solutions:

• The Tycho Catalogue includes components of wide binaries ($\rho \ge 3$ arcsec) observed as two independent entries provides they are brighter than V = 10.5. From the individual astrometric solutions it was easy to derive the separation and position angle of the two components.
• For each of the suspected binaries not solved from the Hipparcos data, we have systematically extracted a field from the Digitized Sky Survey, either from the first or second epoch. When the separation between the two components is larger than about 5 arcsec this was easily detected as an elongated image or two well-defined stellar images. A rough measurement was sufficient to provide a reliable input for $\rho$ and $\theta$ which was then used for the starting value in the Hipparcos double star software. Figures 1-2 show two examples taken respectively from the the first and second epoch of the Digitized Sky Survey.
• In some cases the Input Catalogue provided valuable information not fully exploited during the automatic data processing essentially because of the complexity of sorting out the components of the multiple systems, a practical problem which did not fit easily within a mass processing with stringent time constraints.
• Finally recent CCD observations of visual double stars yielded updated or new astrometric parameters for several systems (Cuypers et al. 1998; Lampens et al. 1998; Oblak 1998).

  

Figure 1: Example of a binary system (HIP 34707) measured over a field of $120 \times 120$ arcsec from the STScI Digitized Sky Survey (1st epoch) and showing well detached components. A crude measurement in the image gave $\rho = 12\hbox{$.\!\!^{\prime\prime}$}0$ and $\theta = 180$ deg. From the Hipparcos observations one gets respectively $12\hbox{$.\!\!^{\prime\prime}$}14$ and 183.8 deg

  

Figure 2: Example of a binary system (HIP 21000) measured over a field of $120 \times 120$ arcsec from the STScI Digitized Sky Survey (2nd epoch) near the lower limit of resolution. A crude measurement in the image gave $\rho = 4\hbox{$.\!\!^{\prime\prime}$}3$ and $\theta = 290$ deg. From the Hipparcos observations one gets respectively $4\hbox{$.\!\!^{\prime\prime}$} 37$ and 289 deg