1 Introduction

The exciting discoveries of giant planets orbiting solar-type stars by precise Doppler searches have caused a shift in our paradigm of the structure and formation of planetary systems. Although we now know that planets have also formed around stars other than the Sun, their orbital characteristics turned out to be quite exotic (for an overview see e.g. Marcy et al. 2000).

Extrasolar giant planets were detected in very close-by orbits around their host stars with periods of the order of a few days, while orbital eccentricities at longer periods appear to be distributed quite uniformly. To date no Solar System analogue has been detected which is primarily due to the insufficient time baseline and long-term RV precision of present Doppler surveys. However, the detection of Jovian-mass companions with P>10 yrs will become possible in the near future.

One of these long-term RV surveys is the planet search program at the Coudé Echelle Spectrometer (CES) at ESO La Silla, which was begun in Nov. 1992 using the 1.4 m CAT telescope. The highlight of this program so far was the discovery of an extrasolar giant planet in an Earth-like orbit around the young (ZAMS) and modestly active G0V star $\iota$ Horologii (Kürster et al. 2000).

It is important to set such discoveries into the context of the complete results obtained by planet search programs. The pioneering study by Walker et al. (1995) first presented long-term (12 years) RV results for a sample of 21 stars and discussed the implications of their non-detections on the occurrence of Jovian-type planets around solar-type stars. In an even earlier work, Murdoch et al. (1993) presented an analysis of their RV measurements for 29 stars over 2.5 years, finding no brown dwarf companions within 10 AU in their sample. Since then RV measurement precision (e.g. Butler et al. 1996) and the size of target samples has increased dramatically. Extrasolar giant planets, which can be detected by present Doppler searches, exist around $\approx$3-5% of the observed solar-type stars. Another study of the long-term RV behaviour of a sample of stars was presented by Cumming et al. (1999). These authors examined 11 years of RV data collected by the Lick survey for 76 F-, G-, and K-type stars and derived companion limits for these stars.

With this work we present all RV measurements of the CES survey and a complete analysis thereof over the time period of November 1992 to April 1998. During that time observations were performed with the same telescope and spectrograph configuration and thus form a homogenous data set. After April 1998 the CES instrument underwent major modifications and the results based on data collected after that point of time will be presented in an upcoming paper.

The structure of this paper is the following: Sect. 2 gives an overview of the CES planet search program, Sect. 3 presents the complete RV results of the CES targets (Appendix A displays the RV measurements graphically for each star), Sect. 4 is a statistical examination of the CES RV data where we perform tests to identify variable stars, the presence of linear and non-linear trends and periodic signals, in Sect. 5 we set quantitative upper mass-limits for orbiting planets based on the RV results (Appendix B shows the derived limits for each survey star) and finally Sects. 6 and 7 contain the discussion and summary.