AGE DISTRIBUTION OF LMC CLUSTERS FROM THEIR INTEGRATED UBV COLORS - HISTORY OF STAR-FORMATION

In this paper we revise the relationship between ages and metallicitie s of LMC star clusters and their integrated UBV colors. The study stan ds on the catalog of UBV colors of the Large Magellanic Cloud (LMC) cl usters by Bica et al: (1994; BCDSP) and the photometric models of sing le stellar populations (SSP) calculated by Bertelli et al. (1994). The se photometric models nicely describe the color distribution of LMC cl usters in the (U - B) vs. (B - V) plane together with the observed dis persion of the colors and the existence of a gap in a certain region o f this diagram. In the case of blue clusters, most of the dispersion i n the colors can be accounted for by the presence of stochastic effect s on the mass distribution of stars, whereas for the red ones addition al dispersion's of similar to 0.2 dex in metallicity and of similar to 0.05 mag in color excess are needed. From comparing the observed dist ribution of integrated colors in the (U - B) vs. (B - V) diagram with the theoretical models, it turns out that: 1) The data are consistent with the presence of a gap (period of quiescence) in the history of cl uster formation. If the age-metallicity relation (AMR) for the LMC obe ys the simple model of chemical evolution, the gap is well evident and corresponds to the age interval similar to 3 Gyr to (12 - 15) Gyr. On the contrary, if the chemical enrichment has been much slower than in the simple model, so that intermediate age clusters are less metal ri ch, the gap is expected to occur over a much narrower color range and to be hidden by effects of color dispersion. 2) The bimodal distributi on of B - V colors can be reproduced by a sequence of clusters almost evenly distributed in the logarithm of the age, whose metallicity is g overned by a normal AMR. No need is found of the so-called phase trans itions in the integrated colors of a cluster taking place at suitable ages (Renzini & Buzzoni 1986). 3) The gap noticed by BCDSP in the (U - B) vs. (B - V) plane can be explained by the particular direction alo ng which cluster colors are dispersed in that part of the (U - B) vs. (B - V) diagram. Also in this case, no sudden changes in the integrate d properties of clusters must be invoked. The results of this analysis are used to revise the empirical method proposed by Elson and Fall (1 985, EF85) to attribute ages to LMC clusters according to their integr ated U B V colors. We show that the EF85 method does not provide the c orrect relation between ages and colors for clusters of low metallicit y and hence its inability to date the old clusters. We propose two mod ifications to the definition of the parameter S of EF85 such that the age sequence of red clusters is suitably described, and the intrinsic errors on ages caused by the heavy presence of various effects dispers ing the colors are reduced to a minimum. The age sequence is calibrate d on 24 template clusters for which ages were independently derived fr om recent color-magnitude diagrams (CMD). Finally, we attribute ages t o all clusters present in BCDSP catalog, and derive the global age dis tribution function (ADF) for LMC clusters. The ADF presents new featur es that were not clear in previous analyses of UBV data, but were alre ady suggested by a number of independent observational studies. The fe atures in question are periods of enhanced cluster formation at simila r to 100 MYr and 1 - 2 Gyr, and a gap in the cluster formation history between similar to 3 and (12 - 15) Gyr. The peaks observed in the dis tribution of B - V colors are found to be sensitive to the presence of these periods of enhanced cluster formation and the lack of extremely red clusters caused by the age gap between intermediate-age and old c lusters.