... 108

Based on observations obtained at the European Southern Observatory using the Very Large Telescope (VLT) (programme 077.C-0660(A)), on Cerro Paranal, Chile.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Table 2 is only available in electronic form at http://www.aanda.org

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... CSR-012A

We use the nomenclature CSR-nnn to refer to the stars listed in Table 2 of CSR05. Stars detected in the present work and not listed in CSR05 are named CS-nnn. Members of multiple systems are referred to by adding A, B,... to their numbers. In this way, CSR-012A is the brightest member of the system collectively identified as Star #12 in CSR05, which the current observations show to be composed of five members.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... IRAF

IRAF is distributed by NOAO, which is operated by the Association of Universities for Research in Astronomy, Inc., under contract to the National Science Foundation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...$K_{\rm S}$bands

L' measurements are also available for many stars near the center of the cluster, and the procedure described here can in principle be easily extended towards that band as well. Nevertheless, since L' strongly dominates over the photosphere for many objects in our sample, the fit would become greatly sensitive to the approximation used to represent the excess in that band, which is necessarily rough. We have thus preferred to base our analysis on the ${\it JHK}_{\rm S}$ bands alone.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... KLF

More rigorously we should instead refer to the $K_{\rm S}$ luminosity function here. However, the difference between the magnitude of any of the stars in our sample in the Kand $K_{\rm S}$ bands is virtually irrelevant for our purposes, as it is much smaller than the size of the bins used to build the KLF and the size of the errors introduced by our rather schematic corrections for extinction and infrared excess. We thus use the name K luminosity function as is normally done in the literature.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... of 0.43

We define the multiplicity fraction as

$${{\sum_{n=2}^\infty n S_n} \over {\sum_{n=1}^\infty n S_n}},$$

where S_n is the number of systems containing n stars; this is, the number of stars in multiple systems divided by the total number of stars. An alternative definition used by other authors (e.g. Duchêne et al. 2001) is ${\sum_{n=2}^\infty S_n} / {\sum_{n=1}^\infty S_n}$, i.e., the number of multiple systems over the total number of systems, including single stars. Using this second definition we obtain a multiplicity fraction of 0.21, which is close to the value 0.18 found by Duchêne for NGC 6611 also using adaptive optics observations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.