The binarity of barium stars (McClure et al. 1980; McClure 1983; Jorissen & Mayor 1988; McClure & Woodsworth 1990) and the white dwarf (WD) nature of the companion (Böhm-Vitense 1980; Dominy & Lambert 1983; Böhm-Vitense et al. 1984) indicates that a transfer of heavy element rich matter from the AGB companion (now WD) towards the barium stars is a good explanation for the chemical peculiarities of the latter. Therefore, the binarity appears to be a necessary condition for producing barium stars (Jorissen & Boffin 1992). Mass transfer is possible either through Roche lobe overflow (Iben & Tutukov 1985) or wind accretion (Boffin & Jorissen 1988). Wind accretion within the binary system is very likely to be responsible for the contamination of the barium star, as suggested by the correlation between the orbital period and the level of chemical peculiarities (Zacs 1994; Boffin & Zacs 1994).
On the other hand, a lot of spectroscopic red giant binaries
with unseen companion (probably WD in some cases) whose orbits
have similar characteristics to those of the barium stars in
all probability have solar abundances of the heavy elements
(Jorissen & Boffin 1992; Boffin et al. 1993).
Furthermore they (Boffin & Jorissen 1992) identify three normal
giants with directly discovered WD companions (
, HD
21120, HD 81817). The existence of binary DR Dra
(K0III) which has a long orbital period (904 days) and
hot WD companion discovered by Fekel & Simon (1985)
and Fekel et al. (1993) lead to the conclusion that the
existence of WD companion in BaII star like system is
not sufficient to produce a barium star. Unfortunately,
the problem with chemical peculiarities is complicated,
because a part of the barium stars may be chemically
mildly peculiar and, therefore, an enhancement of
s-process elements is difficult to detect, especially,
on the base of a spectrum in a narrow spectral region.
For example, Pilachowski (1977) and Zacs (1994)
suggested that 
(HD 13611) has a mild s-process enhancement.
An important step towards a better understanding of the phenomenon of barium stars is a detailed abundance analysis of spectroscopic binaries with BaII stars like orbital elements, especially, binaries with directly observed WD companions. The purpose of this study is to obtain a well represented abundance pattern for a large sample of long-period spectroscopic binaries to compare these abundances with solar ones and to analyze possibly the influence of the unseen companion on the atmosphere of the primary star to understand better the sufficient conditions for the formation of abundance peculiarities. Since the observational material is very extensive, we give here only the analyses and results. The equivalent widths and individual abundances will be published in a companion paper, in the Supplement Series.