In an effort to determine statistical properties of visual pre-main se
quence (PMS) binaries, we have carried out an extensive systematic CCD
imaging survey of young low-mass stars in nearby southern dark clouds
, all at about 150 pc distance. The observations were made with the ES
O New Technology Telescope at a wavelength of 0.9 mum under sub-arcsec
ond seeing conditions. Among the 238 young stars systematically observ
ed, we detect 37 binaries (22 new discoveries) and 1 triple system (Sz
30, also new) in the range of separations between 1'' and 12''. Furth
er binaries were identified outside these limits, including 6 sub-arcs
ec binaries. The lower limit of 1'' was imposed to ensure completeness
of our sample, which we have tested through numerical simulations. We
show that the upper limit is small enough that we have essentially no
contamination by fore- and background stars. Thus, we find a frequenc
y of 16% for PMS binaries with projected separations between about 150
AU and 1800 AU. The distribution of separations is a steeply rising f
unction towards smaller separations. We compare these observations wit
h the known main sequence (MS) distributions of G and K dwarfs in the
same interval of separations, and find an apparent excess of PMS binar
ies. Under the assumption that total PMS and MS binary frequencies sho
uld be the same, the excess may imply that binary orbits undergo secul
ar evolution towards the main sequence. The histogram of flux-ratios a
t 0.9 mum shows a gradual increase towards unequal components. We find
very few binaries with components of nearly equal brightness. We disc
uss the difficulties involved in translating this flux-ratio distribut
ion for PMS binaries into a distribution of mass-ratios. There is no s
ignificant dependence of flux-ratio on the separation of the component
s. We briefly discuss implications of our results for the star formati
on efficiency and the initial mass function. Our data suggest that in
T associations binary star formation appears to be the rule, while the
formation of single stars is probably the exception. Our distance-lim
ited systematic survey sample is actually a sub-sample of our total su
rvey sample which comprised most known PMS objects accessible from the
southern hemisphere except Orion objects. We have in the same way sea
rched for multiplicity in this extended sample. As a result, we have a
dditionally discovered another 18 new PMS binaries and have reobserved
another 32 already known PMS binaries. Therefore, in total we here pr
ovide information (component separations, brightness-ratios, and posit
ion angles) on 87 PMS binaries and 1 triple, approximately half of whi
ch are newly discovered.