5 The Color-Magnitude Diagrams

In this section the V vs. (V-I) CMDs for the 13 GGCs are presented.

The same color and magnitude scales have been used in plotting the CMDs, so that differential measurements can be done directly using the plots. The adopted scale is the same used in previous Paper I. Two dot sizes have been used: the larger ones correspond to the better measured stars, normally selected on the basis of their photometric error ($\leq0.1$) and sharpness parameter. In some exceptional cases, a selection on the radial distance from the cluster center is also done, in order to make more evident the cluster CMD over the field stars, or to show differential reddening effects. The smaller size dots show all the measured stars with errors (as calculated by DAOPHOT) smaller than 0.15 mag.

The images of the fields are oriented with North at the top and East on the left side. Each field covers $5 \hbox{$.\mkern-4mu^\prime$ }6$ square. The same spatial scale has been used in all the cluster images.

In the next subsections, we briefly present the single CMDs and clusters, and give some references to the best existing CMDs. This does not pretend to be a complete bibliographical catalog: a large number of CMDs are available in the literature for many of the clusters of this survey; we will concentrate just on the best CCD photometric works. The tables with the position and photometry of the measured stars will be available on-line at the IAC (http://www.iac.es/proyect/poblestelares) and Padova (http://menhir.pd.astro.it/).

NGC 5053.

(Fig. 4)

NGC 5053 is a low concentration cluster, and, as all the sparse clusters, it has a small central velocity dispersion and central mass density. It is one of the clusters farthest from the Galactic center in our sample. NGC 5053 is one of the most metal-poor clusters in our Galaxy (Sarajedini & Milone 1995).

Figure 4: CMD and covered field for NGC 5053

There are several CMD studies for NGC 5053 in the literature. Nemec & Cohen (1989) presented the first CCD CMD in the Thuan & Gunn (1976) g and r filters, reaching $g\simeq23$. Their CMD is one magnitude deeper than ours, but the stellar distribution and number of stars above the $\rm 22^{\rm nd}$ magnitude is almost the same in both diagrams. The upper part of their RGB (above the HB) is saturated. Heasley & Christian (1991a) present B and V photometry extended to $V\simeq22$. Their CMD is poorly populated, presenting only a few stars in the HB region. Their upper RGB is also truncated at $V\sim15.5$. In the same year, Fahlman et al. (1991) present a study of the stellar content and structure of this cluster, including a B and V CMD that reaches V=24. They are mainly interested in the stellar content and structure, being most of the data obtained just for the V band, and only one field (field #2) in both, V and B bands. The corresponding CMD is deep but poorly populated, presenting $\sim 6$stars in their HB.

More recently, Sarajedini and Milone (1995) present B, V, and I photometry for the upper part of the CMD (above the cluster's TO), making a good sampling ($25\%$ larger than ours) of the evolved cluster stars.

We present a photometry ($\sim5300$ stars, seeing of $\sim1.1\hbox{$^{\prime\prime}$ }$) that covers the cluster from the brighter RGB stars down to the $\rm 22^{\rm nd}$ magnitude. All the CMD sequences are well defined, including a blue straggler sequence. NGC 5053 has a BHB with a few RR-Lyrae and also a few stars in the red side of the HB (evolved HB stars?).

NGC 5272 (M 3).

(Fig. 5)

The northern "standard couple'' of clusters affected by the second parameter effect is represented by M 3 and M 13. M 3 has a well populated HB, both on the red and blue sides, while the M 13 HB is populated only on the blue side. This is a typical example where the age cannot be advocated as the (unique) responsible of the differences in the HB morphology (Paper III, Johnson & Bolte 1998; Davidge & Courteau 1999).

Very recently, M 3 has been the subject of many studies. Laget et al. (1998) present UV, U, V and Iphotometry of the central part of M 3 from HST data. Their CMDs reach $V\sim20$, and is extremely well defined. Their stellar sample is smaller than our one, but clearly shows the mean regions of the CMD. Its overall structure is very similar to our "selected'' diagram. Kaluzny et al. (1998) look for variable stars in the B and Vbands. Their CMD has a limiting magnitude $V\sim21$, is less populated (probably a factor of two, based on the number of BHB stars), since the MS stars are very sparse. An excellent CMD in the V and Ifilters is presented by Johnson & Bolte (1998). Our field does not match that covered by them. However, the fiducial line representing their CMD well matches our diagram, with a small zero point difference (of the order of 0.02 magnitudes in V) in the RGB: a difference within the uncertainties of our absolute calibration. Recently, Davidge & Courteau (1999) published JHK data for four clusters, including M 3. Their CMD extends from the brightest stars down to the (sparse) MS. A large sample of M 3 stars ( $\sim37\ 000$) has been measured in the V and Ibands by Rood et al. (1999), combining groundbased and HST data extending from the cluster center to the outer radius. This CMD is very clean as the high HST resolution limits crowding effects in the cluster center. Notice that the RR-Lyrae stars were identified and removed by the authors.

Our photometry covers the cluster from the tip of the RGB to the $\rm 23^{\rm rd}$ magnitude ($\sim18750$ stars), under quite good seeing conditions. The HB bimodality is clearly visible. Stars spread above the MS turn-off and blue-ward from the RGB are at small distances from the cluster center. If these stars were not plotted, the CMD would become very well defined (larger dots). Moreover, a large number of RR-Lyrae stars are present in our diagram, and our dispersion in color (see the MS, for example) is smaller than that present on the Rood et al. (1999) diagram.

NGC 5466.

(Fig. 6)

NGC 5466 has one of the lowest central densities among the GGCs. Previous CMDs are not very recent; Nemec & Harris (1987) present photographic and CCD data, with the CCD CMD extending from $V\sim21$, to just below the HB (not present). We have not found more recent CMDs for this cluster.

We present a CMD from the RGB tip down to $V\sim22.5$ covering $\sim5000$ stars. NGC 5466 resembles in many aspects NGC 5053, including the metal content, and also their CMDs look very similar.

Figure 5: CMD and covered field for NGC 5272 (M 3)

Figure 6: CMD and covered field for NGC 5466

Figure 7: CMD and covered field for NGC 5904 (M 5)

Figure 8: CMD and covered field for NGC 6205 (M 13)

Figure 9: CMD and covered field for NGC 6218 (M 12)

Figure 10: CMD and covered field for NGC 6254 (M 10)

Figure 11: CMD and covered field for NGC 6341 (M 92)

Figure 12: CMD and covered field for NGC 6366

Figure 13: CMD and covered field for NGC 6535

Figure 14: CMD and covered field for NGC 6779 (M 56)

Figure 15: CMD and covered field for NGC 6838 (M 71)

Figure 16: CMD and covered field for NGC 7078 (M 15)

NGC 5904 (M 5).

(Fig. 7)

The globular cluster M 5 harbors one of the richest collection of RR-Lyrae stars in the Galaxy. It also hosts one of the only two known dwarf novae in GGCs.

The first CCD CMD for this cluster is published by Richer & Fahlman (1987), who present deep U, B, and V photometry. They give a well defined diagram, but poorly populated on the RGB and the HB. More recently, Sandquist et al. (1996) present B, V, and I photometry for more than 20 000 stars in M 5, and an excellent CMD extended down to $V\sim22$. The latest ground-based study is in Johnson & Bolte (1998), who presented very good V and I photometry for this cluster. They compare their photometric calibration with that of Sandquist et al. (1996), with an uncomfortable trend with magnitude and an offset that increases with decreasing brightness. Recently, HST data have been published by Drissen & Shara (1998), who studied the stellar population and the variable stars in the core of the cluster. Again, we compared our photometry with the fiducials from Johnson & Bolte (1998), finding a good agreement within the errors.

Also NGC 5904 has been observed during quite good seeing conditions. The CMD extends by $\ge4$ magnitudes below the TO, and includes $\sim18300$ stars. All the CMD branches are well defined. In particular, note the perfect distinction between the AGB and the RGB, quite rare even in recent CMDs also for other clusters, and the extended blue straggler sequence.

NGC 6205 (M 13).

(Fig. 8)

As already mentioned, M 13 and M 3 are a classical second parameter couple. M 13 has only a BHB, still, in several recent studies, it is found to be coeval with M 3 (e.g. Paper III). In our CMD it shows a sparse EBHB that arrives to the MS TO magnitude.

Previous studies include Richer & Fahlman (1986) who obtained U, B, and V CCD photometry for this cluster from $\sim2$ magnitudes above the TO to V $\sim23$. Their CMD is very well defined, but for a very small sample of stars. Moreover, just 6 RGB stars under the HB are present. More recently, CMDs are presented again by Johnson & Bolte (1998), with their (V,I) CMD and our one perfectly overlapping within the errors. Paltrinieri et al. (1998) present B,V photometry for $\sim5500$ stars from the RGB tip to about 2 magnitudes below the MS TO. Their photometry is more sparse in the SGB and specially in the MS. Davidge & Courteau (1999) have published a JHK photometry. They cover basically the more evolved branches of the diagram.

NGC 6218 (M 12).

(Fig. 9)

The only two CCD studies that we have found in the literature for this cluster are those of Sato et al. (1989), who presented UBVdata for the MS and SGB region in a poorly populated CMD, and Brocato et al. (1996), who present a sparse B and V CMD from the tip of the RGB to a few magnitudes below the cluster TO.

Our CMD is not very populated ($\sim7100$ stars measured), but the main lines of the RGB, BHB, SGB and MS are very well defined. It is noticeable that, despite its intermediate metallicity, this cluster shows only a BHB, resembling in some way M 13.

NGC 6254 (M 10).

(Fig. 10)
Hurley et al. (1989) present the first CCD CMD of this cluster in the B and V bands, covering from the RGB (poorly populated) to $V\sim21.5$. However, already in this diagram the remarkable EBHB is clearly visible. Recently, Piotto & Zoccali (1999) present HST data for the MS of this clusters and analyze the cluster luminosity function, using the present data for the evolved part of the CMD.

Our CMD is well defined and extends for more than 4 magnitudes below the TO, covering a total of $\sim13\ 000$ stars. The cluster has only a BHB, and we confirm that it is extended. In many respects, the CMD of M 10 resembles that of M 13.

NGC 6341 (M 92).

(Fig. 11)

M 92 is one of the most metal-poor and one of the best studied globular clusters in the Galaxy. It was (together with M 3) the first GGC to be studied down to the TO (Arp et al. 1952, 1953). Since then, many CMDs have been built for NGC 6341. The first CCD photometry is presented by Heasley & Christian (1986). They obtain a CMD down to V=22 in the B and V filters, but poorly populated. Another exhaustive work on M 92 is presented by Stetson & Harris, who present a deep B and V CMD with a very well defined MS, but still poorly populated in the evolved part of the diagram. More recently, Johnson & Bolte (1998) present an excellent V and I diagram of M 92 where the principal sequences are very well defined, but there is still a small number of stars in the evolved regions. JHK photometry of this cluster is presented by Davidge & Courteau (1999). Piotto et al. (1997) present a deep CMD from HST/WFPC2 extended down to $\sim 0.15~M_\odot$.

Our diagram, with $\sim13900$ stars measured, is well defined and extends from the RGB tip to about 4 magnitudes below the MS TO.

NGC 6366.

(Fig. 12)

This cluster is somehow peculiar. While located in the disk, and being a metal-rich cluster (as M 71 or 47 Tucanae), it has a kinematics typical of a halo cluster. It is also highly reddened, and its CMD is affected by some differential reddening.

Alonso et al. (1997) present the only other CCD B and VCMD existent for this cluster. NGC 6366, together with NGC 5053, are the only two clusters that were observed under not exceptionally good seeing conditions. Still, all the sequences in the CMD can be identified (apart from the upper RGB), including what seems to be a well populated blue straggler sequence. The HB is very red, as expected on the basis of the metallicity, and tilted. We measured a total of $\sim5500$ stars for this cluster.

NGC 6535.

(Fig. 13)
To our knowledge, Sarajedini (1994) has published the only previous CCD study of this cluster: a B and V CMD down to $V\sim21$. His stellar population is slightly smaller than ours for this range of magnitudes (we reach $V \sim23$). NGC 6535 is the least luminous object of our northern sample, and probably the one with the smallest number of stars. We measured $\sim7800$ stars for this cluster. Its RGB is identifiable, but not clearly defined, due also to field star contamination. Its CMD somehow resembles the CMD of NGC 6717 (Paper I).

NGC 6779 (M 56).

(Fig. 14)

We have not found any previous CCD study on this cluster. Our CMD is well defined, though it is slightly contaminated by foreground/background stars. The broadening of the SGB-RGB might suggest the existence of some differential reddening. The distribution of the stars along the BHB seems to be not homogeneous, with the possible presence of a gap. The total of measured stars was of $\sim11300$.

NGC 6838 (M 71).

(Fig. 15)

As suggested also by its CMD, M 71 is a metal rich cluster, similar to 47 Tuc (Paper I). Our CMD is well defined and extends for more than 4 magnitudes below the TO, covering a total of $\sim12500$ stars. The cluster has only a RHB, and the upper part of the RGB is not very well defined. This cluster is located close to the Galactic plane, and this explains the contamination by disk stars clearly visible in the CMD. It is very bright and relatively nearby.

Despite this, there is no CMD in the literature after Hodder et al. (1992). They present a good B and Vdiagram, less populated than ours, reaching V=22.

Previous CCD studies are in Richer & Fahlman (1988), who present U, B, V photometry for the main sequence, down to V=22(U=25). No evolved stars are present in this work.

NGC 7078 (M 15).

(Fig. 16)

This cluster has been extensively studied in the past, both with groundbased facilities and a large number of HST observations.

HST studies include Stetson (1994) and Yanny et al. (1994), were a CMD of the cluster center is presented. The CMD does not arrive to the MS TO, and is quite disperse. Conversely, Sosin & King (1997) and Piotto et al. (1997) present and extraordinarily well defined MS, but no evolved stars are present.

The most recent ground-based study is the composite CMD of Durrell & Harris (1993) based on CCD data from two telescopes. This is the kind of problem that we try to avoid with the present catalog.

Our diagram is well populated ( $\sim27\ 000$ stars) from the RGB tip down to V=22.5. The CMD features are better identifiable when a radial selection, avoiding the clusters center, is done. The CMD in Fig. 16 gives the visual impression that there are three distinct groups of stars in the HB. The third possible group, on the red side of the RR Lyrae gap is surely a statistical fluctuation in the distribution of the RR Lyrae magnitudes and colors at random phase. It is present neither in the CMDs of M 15 in the above quoted works nor in our CMD of a larger stellar sample, with more accurate photometry from our HST data base.

Acknowledgements

This paper has been partially supported by the Ministero della Ricerca Scientifica e Tecnologica under the program "Treatment of large format astronomical images'' and by CNAA.