Annu. Rev. Astron. Astrophys. 1991. 29: 543-79
Copyright © 1991 by . All rights reserved

Next Contents Previous

2. SURVEYS AND DISCOVERIES

In a galaxy more distant than D ~ 5-10 Mpc, the images of typical globular clusters (with core radii rc ~ 1 pc and half-mass radii rh ~ 5 pc) appear indistinguishable from stars under even the best ground-based seeing conditions. In such cases, the galaxy's GCS appears primarily as a statistical excess of faint starlike objects with a central concentration toward the galaxy, and in general, the only way to identify clusters individually is by time-consuming radial velocity measurement. For nearby galaxies, identifying clusters one by one is virtually a necessity.

2.1 Giant Elliptical Galaxies

Large ellipticals are the most natural environments for GCS work: Globular clusters may number in the thousands per galaxy, and no problems arise with internal reddening or confusion with disk objects. Unfortunately, E giants are the one major type of galaxy not represented in the Local Group, and much effort is still needed to connect their GCS properties with the more familiar ones in the M31 and Milky Way spirals.

VIRGO: This all-important center of the Local Supercluster contains the nearest rich collection of gE-type galaxies. A major starting point for all modern GCS work remains the classic photometric study of Hanes (77, 78), who obtained spatial distributions and luminosity functions to Blim appeq 22.5 for several key Virgo members. During the next several years, additional photographic imaging studies were carried out at the CFHT, CTIO, and AAT observatories, which with limiting magnitudes Blim ltapprox 24 and one-degree fields of view were well suited to GCS detection in galaxies closer than D ~ 20 Mpc (cf 93). In Virgo, the supergiants M87 (NGC 4486) and M49 (NGC 4472) were once again the first targets of study (94, 107, 109). Strom et al. (192) and Forte et al. (64) carried out large-scale multicolor (UBR) surveys of four Virgo Es - a formidable task of plate measurement and data management during the era of photographic photometry and low-speed computing.

Major new progress took place with the advent of CCD detectors. Photometry reaching to, and beyond, the turnover of the cluster luminosity function (Section 3) was carried out by van den Bergh et al. (217), Grillmair et al. (76), Cohen (38), and Harris et al. (103), along with multicolor photometry for samples of the brighter clusters (38, 39, 40). Low-dispersion spectra obtained for a few dozen clusters in M87 and M49 have yielded direct metallicity estimates, radial velocities, and dynamical analyses (81, 119, 173, and especially 151, 152). In total, the Virgo systems are the best understood GCSs outside the Local Group, and stand as the baseline for all comparisons of global GCS properties.

FORNAX: At nearly the same distance as Virgo, the southern Fornax Cluster also contains several large ellipticals. Dawe & Dickens (47) discovered GCSs in three of these, and more complete starcount studies for five galaxies were carried out by Hanes & Harris (83) and Harris & Hanes (105). Several groups have obtained deep CCD photometry for the very rich cluster system around the central cD, NGC 1399 (18, 71, 144a, 219).

SMALL GROUPS: Several E/S0 members of smaller groups within the Local Supercluster region have been surveyed - in some cases the groups include interacting or peculiar galaxies (17, 82, 88, 100, 104, 111, 113, 147, 169). In general, these objects exhibit much greater variety than do the rather homogeneous Virgo and Fornax members, but on average have less populous GCSs (see Section 4). None have yet been studied to the same level of detail as in the two rich clusters described above.

NGC 5128: Investigation of the cluster system in this, the nearest large E galaxy, deserves special mention as an exemplary process of ingenuity and persistence. Only a decade ago, this galaxy - essentially a normal elliptical with accreted tidal debris - was thought to contain few, if any, globular clusters (201) and thus to stand as a strong counterexample to the Virgo Es. But the serendipitous discovery of a single bright cluster (74) proved to be only the tip of a rapidly explored iceberg. More clusters were soon found by their nonstellar visual appearance and radial velocities, while starcounts over a ~ 1° field showed that indeed NGC 5128 had a healthy GCS numbering many hundreds of objects (117, 215). The GCS had eluded notice for so long because, located at only ~ 4 Mpc distance and at low latitude, the GCS is very spread-out on the sky and thus almost invisibly dilute against the Galactic field star population, yet only a few of its biggest clusters stand out individually as nonstellar at appeq1" seeing. Additional photometry and multicolor starcounts (86, 89) refined the estimated size of the system and the cluster luminosity function. Frogel's (67) JHK photometry of a dozen of the brightest clusters demonstrated that they were indeed old globulars, with metallicities extending perhaps to above-solar levels. G. Harris et al. (85) presented a more extensive analysis of their metallicity distribution with CMT1T2 photometry. Velocities of many more objects have been obtained with various selection approaches (116, 191), and by now a remarkably complete understanding of the system has been reached. The present overview of the system is nicely summarized by H. Harris et al. (88).

DISTANT CLUSTERS: In galaxies progressively further beyond the Local Supercluster region, only the most luminous globular clusters remain visible even under excellent ground-based imaging conditions and very deep limiting magnitudes. Harris (93) argued that by careful attention to image classification and selective identification of the faint nonstarlike images (that make up most of the contaminating background field population at high latitude), one could straightforwardly detect GCSs in Virgo-like giant ellipticals as distant as gtapprox 100 Mpc. The most distant system studied with photographic methods was NGC 3311, the central cD in Hydra I (A1060) at a redshift V0 appeq 3500 km/s (110). Deeper exploratory CCD imaging has revealed GCSs in the Coma Cluster supergiants NGC 4874 and 4889 (V0 appeq 7000 km/s) (95, 194), NGC 3842 in A1367 (V0 appeq 6000 km/s) (26), and the classic giant cD NGC 6166 in A2199 (V0 appeq 9000 km/s) (168). Several other distant galaxies are currently under study, including the central giant ellipticals in Hydra I, Centaurus, Pegasus I, A262, and other clusters in the appeq 2000-5000 km/s redshift range.

2.2 Giant Disk Galaxies

In large Sa/Sb galaxies the distinction between open (disk) and globular (halo) clusters is still relatively clear, but the GCS can readily be found only if the target galaxy is seen nearly edge-on to minimize confusion from the disk and spiral arm populations. (This requirement is stringent: even for M31, at i ~ 78° and at a distance of only 0.7 Mpc distant, roughly half its globulars are seen projected against the disk, with consequent severe reddening, crowding, and contamination effects.) In brief, the handful of such galaxies that have been surveyed to date amounts to a rather homogeneous group, appearing more or less as M31 or the Milky Way would be expected to look if placed at the same distance. Positive GCS detections include NGC 2683 (90); NGC 4565 (214); NGC 4594 (106); NGC 5170 (62); NGC 7814 (87); and the nearby M81 (21). Null or marginal detections have been reported for NGC 891 (214), 3717, 5866, and 5907 (87) - all of which are relatively isolated large galaxies or members of sparse groups. Hanes (77, 78) found slight excess counts around the Virgo spirals NGC 4216 and 4569, but these are barely significant within the statistical uncertainties and should probably also be regarded as marginal detections.

For M31 itself, several comprehensive searches for globular clusters have now been carried out with a variety of spectroscopic and imaging techniques (8, 41, 183, 224; see 69 for a summary); these now provide a nearly complete census of globular clusters to V(lim)appeq 18 (well past the peak of the cluster luminosity function) and out to almost a 2° (25-kiloparsec) radius. Several current observational programs are now leading a much-improved understanding of the clusters' distributions over metallicity, radial velocity, and luminosity, and even to measurement of color-magnitude diagrams for selected clusters (e.g. 20, 34a, 112, 122, 175; and 171a).

2.3 Sc/Irr Systems

Detecting and measuring globular clusters in late-type galaxies is a truly difficult task. In these, clusters of all ages are found throughout the body of the host galaxy in a rather uniform mixture, and even in the biggest Sc spirals the globular (old-halo) clusters are likely to number only in the dozens. Experience has shown that the globulars in late-type galaxies must be isolated one by one through an exhaustive process of wide-field multicolor photometry and image analysis, thorough removal of field objects, and (if possible) spectra and velocities (e.g. 11, 34). These requirements in turn limit the possible targets to systems in or not far beyond the Local Group. Even so, only in the Magellanic Clouds have researchers been possible to select the old clusters definitively (through the use of color-magnitude diagrams and RR Lyrae variables.) Nevertheless, these systems help establish just how active the earliest epoch of star and cluster formation was along the full continuum of the Hubble sequence.

LOCAL GROUP IRREGULARS: In the Magellanic Clouds, dozens of recent studies have reduced the list of old globular clusters to just 9 in both the LMC and SMC combined [for literature compilations, see (178, 213, 220); I add the Reticulum cluster (75) to the objects listed by Westerlund (220) and van den Bergh (213)]. The dwarf irregulars NGC 6822 (118, 216) and WLM (1, 182) each contain one cluster of the right luminosity and color range to be a possible globular, but both could as well be intermediate-age objects. IC 1613 contains no known globular cluster candidates (cf 118).

M33: Christian & Schommer and their collaborators (35, 36, 37, 184a) have extensively studied the clusters in this nearest Sc spiral. They find a rather uniform age distribution from young clusters to very old, with a sample of about 20 probable old globulars now identified. The difficulty of separating the oldest clusters unambiguously from intermediate-age objects here is particularly severe; as Christian (34) discusses, integrated color indices or spectra are not sensitive discriminants in the ~ 5-15 Gy age range. For a sample of the old clusters, Brodie & Huchra (21) find a range of metallicities similar to those in the Milky Way halo clusters.

NGC 3109: Demers et al. (49) and Blecha (12) have reported finding a small number (~ 10-20) of globular-like clusters in the halo of this Magellanic-type galaxy, which lies marginally beyond the Local Group.

SCULPTOR GROUP: In NGC 55, Da Costa & Graham (44) found two clusters projected near the disk whose spectra suggest they are old, metal-poor globulars. A more systematic halo search conducted by Liller & Alcaino (141) yielded no significant result. The bigger and somewhat more distant spiral NGC 253 has a more noticeable GCS, with perhaps ~ 20 halo clusters brighter than B appeq 21 (11, 142). See the latter paper particularly for a careful discussion of methodology.

M81 GROUP: For the Sc spiral NGC 2403, Battistini et al. (9) find 8 halo globular cluster candidates. More work is reportedly in progress (145).

2.4 Dwarf Ellipticals

Four dwarf elliptical members of the Local Group (NGC 147, 185, 205, and Fornax) all have handfuls of metal-poor globulars (45, 68, 108, 118). For the Fornax clusters, color-magnitude diagrams have also been published which confirm their generic similarity to Milky Way halo clusters (22, 48). The lack of any known clusters in M32 (the high-surface-brightness companion to M31) remains a minor puzzle: If it had a normal specific frequency it would be expected to carry ~ 15-20 globulars, and no one has quantitatively demonstrated that any original population of clusters could have been totally removed by a combination of tidal stripping from M31 and dynamical friction and tidal shocking from the dense nucleus of M32 itself.

Half a dozen GCSs have been more-or-less accidentally discovered in dwarf ellipticals beyond the Local Group (102): One is a companion to NGC 3115 (82); four are in the Virgo Cluster (38, 181); and one in the Fornax Cluster (60). Notably, all six of these GCS dwarfs are nucleated dEs. This connection may support the evidence of Ferguson & Sandage (61) that the dE, N galaxies are spatially and generically related to the true big ellipticals and not the spirals or irregulars. More systematic deep CCD imaging of these interesting small systems could well be rewarding, as they may resemble the fundamental units from which the galaxies are hypothesized to form (cf 126, 127).

In Table 1, the properties of 60 galaxies with GCSs are cataloged.

Table 1. Galaxies with globular systems a

NGC Group Type V0 (m-M)V MTV Nobs Blim Nel SN Ref

Galaxy Local Sbc -21.3 140   160 ± 20 0.5 ± 0.1 172,
219a
LMC Local Sm 12 18.64 -18.6 8 ltapprox 15 0.5: 213,
220
SMC Local Im -30 18.92 -16.9 1 2 ± 1 0.4 ± 0.2 213,
220
Fornax Local dE0 -51 20.70 -12.3 5 6 ± 1 73 ± 12 22,
118
147 Local dE5 89 24.45 -15.0 4 19 (V) 4 ± 1 4.0 ± 1.0 108,
118
185 Local dE3 39 24.40 -15.2 6 19 (V) 8 ± 2 6.5 ± 1.6 108,
118
205 Local dE5 -1 24.49 -16.5 8 19 (V) 9 ± 1 2.3 ± 0.3 108,
118
221 Local E2 35 24.54 -16.3 0 18.5 (V) ltapprox 3 ltapprox 0.8 108
M31 Local Sb -59 24.54 -21.7 300 ± 50 18.5 (V) 350 ± 100 0.7 ± 0.2 8,
190
M33 Local Scd 3 24.75 -19.2 19 19.5 (V) 30: 0.6: 21,
37
55 Sculptor Sm 106 25.75 -19.0 2 44
253 Sculptor Sc 260 26.8 -20.4 20: 21 (V) 22: 0.2 ± 0.1 11
524 CfA13 S0 2600 32.7 -22.1 700 ± 150 24.9 2830 ± 880 4.0 ± 1.2 104
1052 Cetus E4 1400 31.3: -20.8 220 ± 25 24.4 460 ± 90 2.3 ± 0.4 104
1374 Fornax E0 1390 31.1 -19.8 104 ± 47 23.6 330 ± 160 3.9 ± 1.9 83
1379 Fornax E0 1390 31.1 -20.0 108 ± 29 23.8 290 ± 100 2.8 ± 0.9 83
1387 Fornax S0 1390 31.1 -20.3 150 ± 30 24.2 320 ± 90 2.5 ± 0.7 83
1399 Fornax E1/cD 1390 31.1 -21.2 1690 ± 380 24.2 3600 ± 1100 12 ± 4 83
1404 Fornax E1 1390 31.1 -20.8 97 ± 28 24.4 190 ± 80 0.9 ± 0.4 83
1549 Doradus E0 1010 30.7: -20.8 110 ± 40 24.9 140 ± 60 0.7 ± 0.3 17
1553 Doradus S0 1010 30.7: -21.2 450 ± 100 25.2 540 ± 160 1.8 ± 0.6 17
2403 M81 Scd 260 27.64 -19.5 8: 19 (V) 9
3031 M81 Sab 260 27.75 -21.2 8 21
2683 Sb 370 29.75 -20.8 300 ± 93 25.0 310 ± 100 1.7 ± 0.5 90
3109 Im 130 26.35 -17.3 20: 20 (V) 12,
49
3115 S0 460 30.2 -21.1 520 ± 120 24.7 630 ± 150 2.3 ± 0.5 82
3115B dE1, N 460 30.2 -16.7 25 ± 10 24.7 30 ± 15 6.3 ± 3.1 82
3226 CfA58 E2 1210 31.4: -20.0 115 ± 40 23.6 470 ± 200 4.5 ± 2.0 111
3311 A1060 E0/cD 3420 33.5 -22.4 414 ± 31 24.7 17000 ± 6000 18 ± 6 94,
110
3377 Leo E5 630 30.00 -19.8 140 ± 30 23.7 210 ± 50 2.6 ± 0.6 111
3379 Leo E1 630 30.00 -20.7 170 ± 90 23.7 260 ± 140 1.3 ± 0.7 111
3384 Leo S0 630 30.00 -20.0 75 ± 35 23.7 110 ± 60 1.1 ± 0.5 111
3557 E3 2560 33.0 -22.6 130 ± 90 25.5 400 ± 300 0.4 ± 0.3 147
3607 CfA77 S0 1090 31.3 -21.3 50 ± 35 22.3 800 ± 600 2.5 ± 1.8 111
4278 Coma I E1 910 31.0 -20.8 470 ± 55 23.7 1300 ± 300 6.1 ± 1.5 111
4565 Coma I Sb 910 31.0 -22.2 90 ± 20 23.8 240 ± 70 0.3 ± 0.1 214
4216 Virgo Sb 1080 31.3 -22.0 20 ± 10 22.6 700 ± 380 1.1 ± 0.6 77
4340 Virgo SB0 1080 31.3 -20.3 25 ± 10 22.6 880 ± 380 6.8 ± 2.9 77
4365 Virgo E2 1080 31.3 -21.6 310 ± 30 26.2 3500 ± 1200 7.7 ± 2.7 103
4374 Virgo E1 1080 31.3 -22.0 98 ± 13 22.6 3400 ± 800 5.6 ± 1.3 77
4406 Virgo E3 1080 31.3 -22.1 108 ± 13 22.6 3800 ± 900 5.4 ± 1.3 77
4472 Virgo E2 1080 31.3 -22.9 2300 ± 400 23.3 (V) 7400 ± 2100 5.0 ± 1.4 94
4486 Virgo E0 1080 31.3 -22.7 5700 ± 600 24.2 16000 ± 4000 14 ± 3 94
4526 Virgo S0 1080 31.3 -21.7 87 ± 13 22.6 3000 ± 800 6.6 ± 1.6 77
4564 Virgo E6 1080 31.3 -20.4 34 ± 10 22.6 1200 ± 400 8.5 ± 3.0 77
4569 Virgo Sab 1080 31.3 -22.1 30 ± 10 22.6 1000 ± 400 1.5 ± 0.6 77
4596 Virgo SB0 1080 31.3 -20.8 82:: 22.6 2800:: 13:: 77
4621 Virgo E5 1080 31.3 -21.5 62 ± 12 22.6 2200 ± 440 5.4 ± 1.1 77
4636 Virgo E0 1080 31.3 -21.7 150 ± 20 22.6 4900 ± 1200 9.9 ± 2.4 77
4649 Virgo E2 1080 31.3 -22.5 165 ± 15 22.6 5800 ± 1300 5.9 ± 1.3 77
4697 Virgo E6 1080 31.3 -22.3 80 ± 20 22.6 2800 ± 900 3.5 ± 1.1 77
4594 Virgo SE Sa 840 30.5 -22.9 1250 ± 120 23.1 (V) 2500 ± 600 1.7 ± 0.4 106
4874 A1656 E0 6950 35.0 -23.1 121 ± 14 25.5 21000: 12: 95
4889 A1656 E4 6950 35.0 -23.6 60 ± 12 25.5 11000: 4: 95
5128 Centaurus E0 320 28.25 -22.0 1400 ± 300 22.2 (V) 1700 ± 400 2.6 ± 0.6 85
5170 Virgo SE Sb 1350 31.7 -21.2 310 ± 50 25.0 (V) 390 ± 140 1.2 ± 0.4 62
5813 CfA150 E1 1810 32.3 -21.6 450 ± 100 24.2 2400 ± 700 5.2 ± 1.5 111
5846 CfA150 E0 1810 32.3 -22.1 135 ± 80 23.5 1800 ± 1300 2.6 ± 1.9 111
6166 A2199 E2/cD 9080 35.6 -23.6 156 ± 35 26.8 10000 ± 5000 4 ± 2 168
7814 Sab 1110 30.5 -20.1 197 ± 28 24.5 (V) 230 ± 80 2.2 ± 0.7 88

a Column entries are as follows: 1. Galaxy name or NGC number. 2. The group or cluster in which the galaxy is found; CfA numbers are from Geller & Huchra (72). 3. Hubble type. 4. Mean radial velocity of the group V0, in km/s. (For the Local Group members, individual velocities are given.) 5. Apparent visual distance modulus. For the nearby galaxies [(m-M)V ltapprox 30.5], the values given are individually determined from Cepheids, RR Lyraes, planetary nebula luminosity functions, and other near-field standard candles. Data are taken from the review by van den Bergh (209) with additional material for individual galaxies from several other sources (46, 49, 65, 73, 121, 149, 150, 179, 180, 182a, 195). For more distant systems, moduli are calculated assuming a Hubble parameter H0 = 75 km s-1 Mpc-1 and a Local Group infall to Virgo of 250 km s-1. The Fornax cluster is taken to be 0.2 mag closer than Virgo (15, 18, 167). 6. Total luminosity of the galaxy MTV, in most cases from the BT magnitudes and (B - V) colors in the de Vaucouleurs RC2 catalog (50); other sources are Tully (199) for the Milky Way, NGC 1374, 1379, 1387, and 5170; Carignan (28) for NGC 3109; Hanes & Harris (82) for NGC 3115B; Burkhead (23) for NGC 4594; and Harris et al. (110) for NGC 3311. The values listed for spirals and irregulars are face-on luminosities corrected for internal absorption. 7., 8. Number of observed globular clusters, brighter than the magnitude limit Blim. Reference sources are listed in Column 11. If indicated, the limiting magnitude is in V. 9. Estimated total number of globular clusters in the galaxy, extrapolated from Nobs over all magnitudes and radii. For E galaxies, the cluster luminosity distribution (see Section 3 below) is assumed to be Gaussian with mean at MB = -6.84 and dispersion sigma = 1.4 mag following Harris et al. (103). For spirals, sigma = 1.2 mag is assumed, although Nt is insensitive to the choice of sigma as long as mlim is fairly close to the peak of the cluster luminosity function (94, 111). 10. Specific frequency SN (number of clusters per unit galaxy luminosity; see Section 4 below). Here, SN is calculated using the total light MTV of the galaxy, not just the spheroid light.

Next Contents Previous