Observational Data for Galactic Globular Clusters

(Published in the Webb Society Quarterly Journal for January 1995, no. 99)
(Database (Table 1) and references updated 2 May, 1999 on this web site by the author)

Brian A. Skiff
Lowell Observatory
1400 West Mars Hill Road
Flagstaff AZ 86001-4499
e-mail: bas@lowell.edu


An up-to-date tabular summary of total magnitudes, brightest stars, and journal references is given for all 142 confirmed Galactic globular clusters. The text discusses how the information can be used to plan and analyze visual observations.


What you can observe depends a lot on what you know about. It is easy, for instance, to overlook a bright non-NGC galaxy in the same field as an NGC object merely because it is not plotted on the atlas you're using. To help improve this situation, over the years I have collected relevant data about all types of objects from the professional literature for use in analyzing my visual observations. Much of this material was condensed without attribution into Chris Luginbuhl's and my Observing Handbook and Catalogue of Deep-sky Objects. But a lot of data has come out since 1984, when work for the book was finished. Many of the most useful articles were mentioned in the 'Scanning the Literature' column that appeared in the back pages of each issue of late, lamented Deep Sky magazine. I continue to get inquiries from amateurs about specific objects, and it is usually possible to supply a citation to a relevant journal article dealing with practically any fairly bright object. This task has become much easier in the last several years thanks to the availability of on-line databases, such as the NASA Extragalactic Database (NED) and SIMBAD, maintained by the Centre de Données astronomiques in Strasbourg, France.

One file I've built up that has come in particularly handy is a list of 'observables' for globular clusters. The list of globulars is taken from the compilation by Djorgovski & Meylan (ref. 1), and lists all confirmed clusters deemed to belong to the Galaxy and classified as globulars. Comments about objects omitted and undetermined cases are dicussed below. Included in the file are total V magnitudes according to the 1992 Peterson compilation (ref. 2). In addition, I went through the literature and found essentially every photometric study ever published of a globular cluster (hundreds of papers!). From these publications, I determined the V magnitude and B-V colour of the brightest stars in each cluster, and the magnitude of the 'horizontal branch'. For each cluster I chose the best paper (sometimes more than one) dealing with the brighter stars. That is, I have listed in the bibliography the paper with the most useful finder chart for the field of each cluster, the one with magnitudes for bright field stars around the cluster---the best one for checking one's visual and photographic observations. This is not always the highest-quality or most recent study, just the one best for amateur purposes. This material is collected in Table 1.

Colour-Magnitude Diagrams

The purpose in making photometric measurements of globular cluster stars is to create a 'colour-magnitude diagram'. This is merely the observational version of the well-known Herzsprung-Russell diagram. In the original HR diagram, the luminosity of a star (or group of stars) is plotted against the temperature. We cannot observe these two things directly, but we can measure the apparent brightness of a star and its 'colour', that is, the difference in its brightness at two wavelengths. A hot star is brighter toward the blue end of the spectrum than toward the red end, and conversely for cool stars. In the commonly used UBV system, the temperature of a star can be estimated by the magnitude difference in the B (blue) and in the V (visual) band. The B-V ("B minus V") colour for a fairly hot star like Vega is close to zero in this system and is about 1.5 for ordinary cool giants such as Aldebaran. Some cool supergiant stars and carbon stars have much larger B-Vs, and appear quite red visually. The hottest ordinary stars, such as the stars in the Belt of Orion, have B-V about -0.2. Interstellar dust causes stars to appear redder than they would than otherwise, and so more distant stars are often 'reddened' by significant amounts.

Well, what happens when B and V magnitudes are measured for stars in a globular cluster? The figure shows an excellent example from a paper by Gonzalo Alcaino and William Liller (ref. 3). Notice first that the stars do not scatter everywhere in the plot of V versus B-V, but instead clump into narrowly-defined strands, called 'sequences' or 'branches'. The very brightest stars in any globular cluster are stars at the tip of the red-giant branch toward the upper right corner of the plot. These are giants and bright-giants of spectral class K and M. They have evolved from stars somewhat more massive than the Sun. Following the red-giant branch down, the more nearly vertical string of stars is called the subgiant branch, i.e., stars intermediate in brightness between main-sequence stars (like the Sun) and the brighter, cooler giants. Midway down this string, stars split off to the left (blueward) side of the diagram. These stars are called 'horizontal- branch' stars, and are even more evolved than the red giants, i.e., stars that are heading toward extinction. In the middle of the horizontal branch is a band of variables, the RR Lyraes, indicated by small x's in this plot. These are A- and F-type stars (somewhat hotter than the Sun) that pulsate with amplitudes of around a magnitude with periods of about half a day.

At the base of the red-giant/subgiant strip, the stars take a sharp turn redward as they get fainter. This is point where the most massive stars on the main-sequence are starting to evolve into giants. The stars at the turnoff in globular clusters are close to the Sun's temperature and mass. From here on down as far as anyone has been able to observe in any cluster, there are only unevolved main-sequence stars, becoming more and more numerous toward the faint end. The faintest stars measured in globulars so far are about absolute V magnitude 12, or eight magnitudes fainter than the Sun, corresponding to stars of only 0.2 of a solar mass. The most luminous stars at the tip of the red-giant branch, on the other hand, have absolute V magnitudes near -2.5, roughly half a million times brighter.

Interpreting colour-magnitude diagrams for clusters in terms of a star's evolutionary history remains a central problem for astrophysics even 75 years after the first diagrams were constructed and their significance shown. For further reading on this subject, see for instance refs. 4 & 5. But these details need not concern us for the immediate purpose visual observation.

What Are The Data Good For?

In visual observation of a globular cluster, obviously if you are going to see any resolution at all, your telescope will have to show stars at least as faint as the brightest stars in the cluster. Thus it is of interest to know what the magnitudes of the stars in the cluster are in order to predict whether a cluster will be resolved, or perhaps to get an estimate of how faint you can see based on whether a cluster is resolved or not. So the table includes the magnitudes and colours for the very brightest stars in the cluster [V(tip) and B-V(tip)]. I note the B-V colour because the eye will see red stars fainter than their V magnitudes would indicate, since the peak dark-adapted visual response is shifted toward the blue from the standard V passband. The shift amounts to roughly:

m(vis) = V + 0.2(B-V).

For example, a star with B-V = 1.5 will appear 0.3 magnitudes fainter visually than a star of B-V = 0.0 that has an identical V magnitude. As you can see by looking down the list, stars at the tip of the giant branch in most globulars have B-V between 1.5 and 2.0. So although these stars will appear slightly fainter than blue stars of the same V magnitude, comparisons among different globulars can be made without taking this effect into account.

In moderate-to-large apertures, most globulars are at least partially resolved, showing some dozens of stars. In the brightest objects, one often sees seemingly thousands of stars, and in a few such as 47 Tucanae and omega Centauri, it really is thousands. The cluster is 'well resolved'. The point at which this happens is when your telescope has a limiting magnitude at or below the level of the horizontal branch. The reason is simply that the number of stars in any given magnitude interval takes a sudden leap at the magnitude of the horizontal branch.

What about the Shapley-Sawyer concentration classes? These Roman numeral classes are often considered indicative of how easy a cluster is to resolve. The classes are actually correlated with central surface brightness, not the magnitudes of the stars, and my experience indicates they have little bearing on cluster resolution. In general, the highly-concentrated clusters are the most distant ones, and so are difficult to resolve simply because the stars are faint, not because they are close together. I have seen the brightest handful of stars in M15 and M2 using a 70mm refractor at 75x and 110x. They are both quite strongly concentrated objects, yet their brightest stars, between mag. 12.5 and 13.0, are just visible in this small aperture from a true-dark site. It may be that clusters with the same V(tip) but having different concentration classes will show a difference in resolution. Perhaps a reader can provide the necessary observations.

Cluster Rankings

Along with the main list, the supplementary lists (Tables 2,3,4) show how the clusters rank in terms of the three parameters: total magnitude, magnitude of the brightest stars, and by magnitude of the horizontal branch. The results are occasionally surprising compared to conventional wisdom. The first list shows the twenty-five clusters with total V of 7.0 or brighter. The two southern clusters omega Centauri and 47 Tucanae are roughly a magnitude-and-a-half brighter than the nearest competition. M5 and M13 rank seventh and eighth; they are nice clusters, but by comparison omega Cen and 47 Tuc will knock your socks off! A common debate among southern observers is which of the two is better. Some prefer the sheer richness of omega Cen, and others the star-density and remarkable (indeed unique) structure of 47 Tuc. Essentially all of the non-Messier clusters in the list are far-southern objects that certainly would have been catalogued by him had he worked at southern latitudes. NGC 3201 and NGC 6541 are accesible, however, to those at mid-northern latitudes.

The other two lists give the top ten clusters in terms of brightest stars and brightest horizontal branches---the most easily resolved clusters. Note that neither omega Centauri nor 47 Tucanae appears at the top of these lists. Both are luminous objects with large numbers of stars, as anyone who has observed them can confirm, and their combined light sends them to the top of the total magnitude list. But because of their distance, the brightest stars are not as bright as nearer clusters. Instead the winner is NGC 6397, a far-southern object that is known as the closest or second-closest globular (it and Messier 4 are both about two kiloparsecs distant). This object and the next few in the ranking are partially resolved in ordinary handheld binoculars. For northern observers, the most easily resolved clusters are M22, M4, and M55, all of which are nevertheless south of -20 Dec. The brightest stars in M13 are nearly two magnitudes fainter than those in NGC 6397. On the horizontal-branch list, NGC 6397 also rates as #1, again with stars two magnitudes brighter than those in M13, which falls off the list altogether.

Perhaps the most immediate lesson to be drawn from this is that to see the galactic globular system well at all, you need to head to the southern hemisphere!

The Database

The objects contained in Table 1 derive directly from Djorgovski & Meylan's list (ref. 1), which was presented at a 1992 conference. They included 143 objects, but one of them, known as the "Reticulum cluster", has since been shown to be an outlying member of the Large Magellanic Cloud. The cluster reported by Djorgovski as "Djorgovski 3" is identical with NGC 6540, which was previously considered to be an open cluster. This is the first NGC/IC object to be identified as a globular since the 1930s.

Since they are relatively easy to pick up visually in moderate apertures, I have also included the five globular clusters in the Fornax system. The Clouds appear to contain only another eight or ten true globulars, and along with the "Fornax five", appear to be the only globulars outside the three major spirals in the Local Group (M31, M33, and Milky Way). Yes, I know Uranometria and the RNGC indicate scores of globulars in the LMC, but those clusters are not similar to galactic globulars. (Let's talk about this another time.)

According to preliminary studies, Palomar 1 exhibits no horizontal- branch stars, and thus is likely to be an old open cluster. I've left it in the list for now since the work is not formally published.

It is also worth noting that E 3, AM-2, and AM-4 are evidently clusters of intermediate age, i.e. between the ages of the oldest ordinary open clusters (6-8 billion years) and those of globulars (13-15 billion years). Although not uncommon in the Magellanic Clouds, these appear to be essentially absent from the Milky Way. One can suspect that, over the course of the dynamical evolution of the Local Group of galaxies, the Milky Way has glommed onto a couple of stray clusters from other galaxies. Indeed this may help explain the bi-modal character of the Galactic globular cluster system.

If you are one of those cluster-fetishists familiar with obscure objects on older lists of globulars, you will notice a number of objects missing and others that are new. Most of the nebulous objects reported as possible globulars by A. Terzan and his collaborators turn out to be galaxies or planetary nebulae upon spectroscopic investigation. Still, for several of these no definitive conclusion has yet been reached. Djorgovski & Meylan indicate that the objects TJ 5, TJ 23, and TBJ 3 could be either distant, obscured globulars, chance groupings of stars, or background galaxies with Milky Way stars superposed. Other "suspect" objects are:

  • ESO 166-SC11 = UKS 2, probably an open cluster;
  • ESO 93-SC?08, probably an open cluster;
  • ESO 224-SC08 = vdB-H 176, a sparse globular or an open cluster.
More data are required for all these six objects.

Among objects confirmed to be "not globular" are the Terzan objects TBJ 1 = TJ 17, TBJ 2 = TJ 16, TJ 15, and others, all of which are galaxies. Grindlay 1 and Kodaira 1 are both nonexistent, there being no cluster at the published coordinates of either.

I maintain this list on a computer and keep it up to date as new journal articles are published. Contact me if you would like a copy of the latest version. Ordinary mail and electronic mail are the best ways to get my attention. If you have an e-mail connection (or know someone who does), it is easy for me to simply send you the current file.


1.  Djorgovski, S. G., and Meylan, G.  1993, "The Galactic Globular Cluster
     System:  A List of the Known Globular Clusters and Their Positions"; in
     Structure and Dynamics of Globular Clusters, A.S.P. Conference 
     Series vol. 50, S. G. Djorgovski and G. Meylan, eds., pg. 325.
2.  Peterson C. J.  1993, "Integrated Photometric Properties of Globular
     Clusters"; in Structure and Dynamics of Globular Clusters, A.S.P. 
     Conference Series vol. 50, S. G. Djorgovski and G. Meylan, eds., pg. 337.

3.  Alcaino, G., and Liller, W.  1984, "BVRI Main-Sequence Photometry of the
     Globular Cluster M4", Astrophys. J. Suppl., 56, 13.

4.  Bok B. J., and Bok, P. F.  The Milky Way, Harvard Univeristy Press,
     Cambridge MA, 1981.

5.  Payne-Gaposchkin, C.  Stars and Clusters, Harvard University Press,
     Cambridge MA, 1979.

Table 2.

Globular clusters brighter than V = 7.0

Name               V
N5139=omega Cen   3.9   
N 104=47 Tuc      4.0   
N6656=M22         5.2   
N6397             5.3   
N6752             5.3   
N6121=M4          5.4   
N5904=M5          5.7   
N6205=M13         5.8   
N6218=M12         6.1   
N2808             6.2   
N6809=M55         6.3   
N6541             6.3   
N5272=M3          6.3   
N7078=M15         6.3   
N6266=M62         6.4   
N6341=M92         6.5   
N6254=M10         6.6   
N7089=M2          6.6   
N 362             6.8   
N6723             6.8   
N6388             6.8   
N6273=M19         6.8   
N7099=M30         6.9   
N3201             6.9   
N6626=M28         6.9   

Table 3.

Ranking by brightest stars

Name                V
N6397             10.0 
N6752             10.5 
N6656=M22         10.7 
N6121=M4          10.8 
N6809=M55         11.2 
N5139=omega Cen   11.5 
N3201             11.7 
N 104=47 Tuc      11.7 
N6205=M13         11.9 
N6254=M10         12.0 

Table 4.

Ranking by horizontal-branch magnitude

Name                V
N6397             12.9   
N6121=M4          13.4   
N6752             13.8   
N 104=47 Tuc      14.1   
N6656=M22         14.2   
N6809=M55         14.4   
N6838=M71         14.4   
N5139=omega Cen   14.5   
N6254=M10         14.7   
N3201             14.8   

The Database - Table 1

Download This Paper As A Text File