7 Discussion

We find differences in CMD (as well as in CPD) positions for HB stars and lower giants (at the level of HB stars and below it) in two parts of NGC 1978, in several colors (all of them including B magnitudes). However, the upper giants show no obvious differences in the CMD positions for the NWH and SEH of the cluster. Do these features reflect real properties of the star population in the cluster field? We cannot exclude the photometric errors to have influenced the result. But, in this case, photometric errors should be rather specific because they influence only B magnitudes and have almost abrupt change just near the position of the cluster's minor axis. If the stars of NGC 1978 really have nonuniform parameters (chemical abundance, age), what range of them could explain the differences discussed? To answer this question, we considered isochrones by Bertelli et al. (1994) and Yale isochrones by Demarque et al. (1996) with values of parameters (age and [Fe/H]) close to those of NGC 1978. In spite of a number of sensible differences between isochrones (with similar or close values of age, [Fe/H], and Y) of the two sets, it is possible to conclude the following (cf. Fig. 7).

 

Figure 7: Comparison of isochrones in the $(M_{v}, \ B-V)$ plane [panel a)] and in the $(M_{v}, \ V-I)$ plane [panel b)]. The pair of isochrones of the set by Bertelli et al. (1994) with an age of 2 Gyr, Y=0.24, z=0.004 ([Fe/H] = -0.7) and with an age of 2 Gyr, Y=0.25, z=0.008 ([Fe/H] = -0.4) is shown by the solid line. A Yale isochrone from the set by Demarque et al. (1996) with an age of 2 Gyr, Y=0.23, z=0.008 ([Fe/H] = -0.7) is shown by the dashed line to demonstrate the difference between the two sets

Because there is no obvious variation, if any, of the turnoff magnitude in the two parts of the cluster, an age spread cannot be significant and, consequently, it cannot cause the differences under consideration. A metallicity spread might play an important role, especially taking into account that isochrones of both sets show that metallicity variations lead to a shift of the lower giants in B-V greater than that in V-I at least by a factor of $\sim1.5-2.0$. But it is not the case for the upper parts of the GB, especially for the GB tip where a shift in V-I may be even greater than in B-V.

The isochrones by Bertelli et al. (1994), reproducing the position of the clump of the HB stars, indicate that, similar to the lower giants, the clumps of the HB stars with different metallicity are closer to each other in V-I than in B-V color.

To explain the observed variations, the difference in metallicity between stars in the NWH and SEH of NGC 1978 should be as large as $\Delta[{\rm Fe/H}]\sim0.1-0.2$, the stars of the cluster's NWH being more metal-rich. Also, stars of the cluster's NWH still may be slightly younger (by a few tenths of a Gyr) than those of the cluster's SEH. The permitted age difference between the two subpopulations (so that their turnoff magnitudes remain indistinguishable in the CMD) depends on the possible difference of the helium content.

Thus, we cannot exclude the existence of a real inhomogeneity of the stellar content in NGC 1978, presumably a result of merging. But it has to be noted that studies of dynamics and of surface brightness isophotes do not provide evidence for merging in NGC 1978 (Fischer et al. 1992). To verify our results and to make more definite conclusions, an additional multicolor photometric study of NGC 1978 with high accuracy in a large magnitude range is needed.