Annu. Rev. Astron. Astrophys. 1988. 26: 245-294
Copyright © 1988 by . All rights reserved

Next Contents Previous


2.1. Results From Galaxy Reashift Surveys

Previous reviews of this topic include those by Chincarini (38, 39), Doroshkevich et al. (62), Geller et al. (79a), Oemler (126b), Schwarzschild (175), Shandarin & Zel'dovich (176), and Ze1'dovich et al. (216). The review by Geller et al. (79a) contains information in addition to that below.


The early Coma / A1367 redshift survey (81) covers 260 sq deg. The sample of galaxies is nearly complete to mp appeq 15 corresponding to an effective Doppler velocity limit Vlim ~ 8000 km s-1. The early Hercules/A2199 redshift survey (49) covers a 332 sq deg region. The sample of galaxies was drawn randomly in location and apparent magnitude from a population complete to mp appeq 15.5 (Vlim ~ 11,000 km s-1). Both surveys revealed a prominent void with a characteristic length > 40 Mpc. These voids were detectable because they are small compared with the effective depth of each survey (~ 160-220 Mpc).

Although redshifts have been measured for the galaxies in the all-sky Shapley-Ames survey (59b, 167a) complete to mp appeq 13 (i.e. Vlim ~ 3000 km s-1), the effective depth (~ 60 Mpc) and the location of our Local Group of galaxies within the Local Supercluster precluded the discovery from these data of any obvious void with characteristic length > 40 Mpc. Because of these same survey properties, however, the data on Shapley-Ames galaxies have contributed greatly toward increasing our knowledge of the structure of the Local Supercluster, and Tully (202, 203) [papers reviewed by Oort (128), pp. 380-384] finds that the great majority of galaxies in the supercluster are located in nine "clouds," with most of the space between empty.

Although the redshift surveys covering small solid angles of the sky led to pioneering discoveries of individual voids (e.g. Coma/A1367, Hercules/A2199), redshift surveys covering large solid angles of the sky are necessary to obtain statistical information on the locational and physical properties of voids in general, as well as to obtain more complete structural information about individual voids and their contiguous shells. The data from the CfA redshift survey of the 2400 galaxies complete to mp = 14.5 with b > 40°, delta > 0° and b < - 30°, delta > - 2.°5 [where b is Galactic latitude and delta is declination (epoch 1950)] have been published (93a), and initial results (59c) have been reviewed by Oort (128, pp. 385-94). Significant progress has been made toward the completion of a corresponding redshift survey of the > 12,000 galaxies in the same solid angle of the sky with mp leq 15.5 (J. Huchra, preprint, 1987); for example, initial results from part of this survey (discussed below) are beautifully illustrated by de Lapparent et al. (56) as reproduced in Figure 4. Da Costa et al. (51b) have carried out a complementary redshift survey of 1963 galaxies that is 84% complete (redshift measurements obtained for 1657 galaxies) to an angular diameter limit of 1 arcmin (very roughly mp ~ 15) in 1.75 ster of the southern sky (delta leq - 17.°5, b leq - 30°).

The early studies described in Section 1.3 established that the void in the foreground of the Coma/A1367 supercluster spans at least from the Coma cluster (A1656) (alpha = 12h57m , delta = 28°15', V = 6955 km s-1) to A1367 (alpha = 11h42m, delta = 20°7', V = 6446 km s-1)and that the void in the fore- ground of the Hercules/A2199 supercluster spans at least from the Hercules cluster (A2151) (alpha = 16h3m, delta = 17°53', V = 11, 122 km s-1) to A2199 (alpha = 16h27m, delta = 39°28', V = 9264 km s-1), where the equatorial coordinates are for epoch 1950 and the Doppler velocity is relative to the centroid of the Local Group. Neta Bahcall (private communication, 1986) pointed out that these two voids are in fact segments of the two largest voids in Figure 4a from de Lapparent et al. (56), which I call the Coma /A1367 void (or "Coma void," for short) and the Hercules/A2199 void (or "Hercules void," for short). Figure 4a is a plot of Doppler velocity vs. right ascension (over a range of 9 h) for a nearly complete sample of galaxies with mp leq 15.5 in a 6° strip of declination that contains the Coma cluster. Figure 4b is the same as Figure 4a but with mp leq 14.5. Figure 4(a,b,c) (which supersedes Figures 11 and 12 (pp. 390-91) of Oort (128)] provides a clear demonstration of the detailed new structural information on voids that is obtainable from the CfA redshift survey to an apparent magnitude limit of mp = 15.5 and demonstrates that the Coma and Hercules voids are much more well defined from data with a survey limit of mp = 15.5 (Figure 4a) than from data with a survey limit of mp = 14.5 (Figure 4b); this unequivocally demonstrates the physical reality of these voids with their contiguous shells - i.e. with improved statistics, these structures have risen significantly farther above the noise. (The reader may wish to assess quantitatively whether the statistical confidence level of this statement is actually tantamount to certainty; some considerations relevant to this problem are discussed in (57) and Section 3.1.]. De Lapparent et al. (56, 57) note that the Coma and Hercules voids each have a characteristic length ~ 50 Mpc, and their contiguous shells are sharply structured. Part of the Local Supercluster with its effective Doppler velocity ~ 1000 km s-1 that is evident in Figure 2 is also apparent in Figure 4a as part of the contiguous shell of the Coma void. Because the angular extent of the Coma and Hercules voids are each > 6°, de Lapparent et al., in order to obtain additional information on their structure, examined a map of the surface distribution of the galaxies with mp leq 15.5 in a strip with the same center and range of right ascension but with a width of 400 (Figure 4c). De Lapparent et a1. suggest that the cellular pattern evident in Figure 4a and the smoothness of Figure 4c are simply understood if the galaxies are distributed on the surfaces of shells tightly packed next to each other. (It is remarkable that 10 yr earlier Einasto and coworkers were able to reach very similar conclusions from examination of the heterogeneous data available at that time (cf. 99a), albeit with some theoretical guidance by Zel'dovich et al. (cf. 215b, 217).]

Figure 4

Figure 4. (a) Map of the Doppler velocity vs. right ascension in the decl 26.°5 leq delta leq 32.°5. Data are plotted For 1061 galaxies with mp leq 15.5 and V leq 15,000 km s-1. (b) Same as Figure 4a for mp leq 14.5 and V leq 10,000 km s-1. Data are plotted for 182 galaxies. (c) Projected map of the 7031 objects with mp leq 15.5 listed by Zwicky et al. (220) in the region bounded by 8h leq alpha leq 17h and 8.°5 leq delta leq 50.°.5. Reproduced from (56) by permission of Margaret J. Geller.

The de Lapparent et al. study (56, 57) underlines the potential value for future work of (a) the entire CfA redshift survey complete to mp = 15.5 over a large fraction of the sky and (b) the completed subset of data covering the 6° strip that was analyzed in (56, 57). [This data is being prepared for publication and will also be made available on magnetic tape for distribution by the Astronomical Data Center at the Goddard Space Flight Center (M. Geller, private communication, 1987)]. With (a), one should be able to establish with certainty whether the galaxy distribution is in fact cellular, filamentary, some combination of both, or something else. With (b), one could, e.g., subtract out the distorting effect of virial motions of galaxies, especially from the Coma cluster, to establish whether the Coma and Hercules voids are independent entities, or possibly, e.g., two parts of one void or the network of voids in the sponge-like model of Gott and coworkers (cf. 80). And one could examine the declination of each galaxy with the same Doppler velocity and right ascension as a point in the interior of the Coma or Hercules void to determine whether the galaxy is nevertheless outside or actually within the void.

In 1978, Kirshner et al. (103) set out on a pioneering study of the deep-space distribution of galaxies. They obtained redshifts and direct photometric magnitudes for most of the galaxies (164 of 184) to a completion limit of mp appeq 15.5 (Vlim ~ 11,000 km s-1) in eight representative square fields appeq 4° on a side, four in the northern and four in the southern Galactic hemisphere. They planned to redetermine the galaxy luminosity function and, stimulated by the recent progress in our understanding of the distribution of galaxies in space derived from studies of the two-point correlation function and the early redshift surveys [e.g. (81) and other work described in Section 1.3], to redetermine this space distribution. Their data immediately suggested that the galaxy distribution might be significantly smoother in the south Galactic polar cap (b leq - 40°) than in the north Galactic polar cap (b 40°), and the correlation function that they derived was inconsistent with results from previous studies of the surface distribution of large samples of galaxies (104). This was sufficient to motivate Kirshner et al. (105) to obtain similar data for six representative fields appeq 1.4° on a side, three in the northern and three in the southern Galactic hemisphere, located near six of the fields in the previous survey but penetrating to the pioneering apparent magnitude limit mp ~ 17.5 (corresponding to Vlim ~ 30,000 km s-1 appeq 0.1c). The three northern fields are located approximately 350 equidistant apart on the sky near the periphery of the constellation Boötes, as illustrated in Figure 5.

Figure 5

Figure 5. Location on the sky of (a) the original Boötes survey fields of Kirshner et al. (105). (b) the 283 small square survey fields (represented by small circles) in the recent, more detailed survey of Boötes (106), and (c) the Corona Borealis survey field of Postman et al. (148a). Reproduced from (148a) by permission of M. Postman.

Kirshner et al. found that the histogram of observed redshifts for 133 galaxies in these three northern fields (90% complete redshift survey) is nearly empty in the interval between 12,000 km s-1 and 18,000 km s-1 (Figure 6a); this surprising result led them to make the plausible inference that the region within Boötes may contain a very large void (characteristic length ~ 120 Mpc) (105). To test this inference, they decided to survey the entire Boötes region in a representative manner to an apparent magnitude limit mp ~ 17.5, measuring both redshifts and magnitudes for a large homogeneous sample of galaxies. This was accomplished by first selecting 283 small square survey areas 15 arcmin on a side, covering 2% of the sky area of the Boötes region, and distributed as shown in Figure 5. Then for each of these fields they visually identified galaxies and ranked them according to apparent luminosity. Apparent magnitudes of the galaxies were then estimated from these ranks (typically, for a given galaxy, the adopted rank is a logarithmic average of four values, one measurement per observer) calibrated by photoelectric magnitudes measured for 59 of the galaxies. Finally, redshifts were measured for 231 galaxies in 239 of the 283 survey fields (90% complete redshift survey to mp ~ 17.5 in the subset of 239 fields). The resulting histogram is shown in Figure 6b. The region between 12,000 km s-1 and 18,000 km s-1, while not empty, remains significantly deficient in galaxies relative to the number expected if the galaxies were homogeneously distributed in space. Moreover, the largest empty sphere that Kirshner et al. could place in the effective volume occupied by their sample has a radius appeq 60 Mpc and is centered at alpha = 14h50m, delta = 46°, and Doppler velocity V = 15,500 km s-1. Kirshner et al. point out that, curiously, this empty sphere does not extend as far as the three survey fields (NP 5, NP 7, or NP 8) on the basis of which the original discovery of the Boötes void was made. Kirshner et al. suggest that the most plausible explanation is that the void, although apparently sharply bounded in front and back, is surrounded in other directions by a larger region of low density that fields NP 5, NP 7, and NP 8 happen to penetrate at particularly empty spots (106). A direct test of this hypothesis requires an even more extensive redshift survey. [Additional review of work on the Boötes void is provided by Oort (128, p. 417) and Oemler (126b).]

Figure 6

Figure 6. (a) Histogram of Doppler velocities in 1000 km s-1 intervals for a sample of 133 galaxies ~ 90% complete to mp appeq 17.8 in the 1.4 sq deg fields NPS (dark), NP7 (light), and NP8 (hatched) depicted in Figure 5. The smooth curve is the distribution expected in a homogeneous universe. Reproduced from (105) by permission of R.P. Kirshner. (b) Corresponding histogram for a sample of 231 galaxies comprising (i) a subset ~ 97% complete to mp appeq 17.5 (dark shading) and (ii) other galaxies in 239 square fields 15' on a side (out of the 283 small fields depicted in Figure 5). The smooth curve is the distribution expected in a homogeneous universe. Reproduced by courtesy of A. Oemler.

The catalog by Zwicky et al. (220), constructed primarily by visual techniques, provides coordinates and magnitudes for the galaxies north of declination delta appeq - 3° complete to a limiting apparent magnitude mp appeq 15.5. This information is prerequisite for the CfA redshift surveys to (a) mp leq 14.5 and (b) mp leq 15.5. To construct deeper redshift surveys, and to study in more detail the structure of the Boötes void, it seems advisable to first create an analog of the Zwicky et al. (220) catalog, but to a limiting apparent magnitude of, say, mp = 17.5 and by vastly more automated procedures. This goal is now within practical reach through the application of an automated plate scanner. One that is extremely fast, accurate, and in refinement-developmental stages is the automated plate scanner (APS) at the University of Minnesota; astronomers there are planning to use this instrument to create useful primary data bases of this type. The APS was initially built under the direction of W.J. Luyten specifically to carry out a stellar proper motion survey, but it has been refurbished with new detection electronics and data-taking and processing equipment that make it a practical instrument to apply to a plate copy of the Palomar Observatory Sky Survey to construct an all-sky catalog of locations, magnitudes, and other properties of galaxies nearly complete to mp = 17.5 (61, 96). The APS can automatically identify and measure all objects with mp leq 17.5 registered on a 14-inch × 14-inch plate of the Palomar Observatory Sky Survey in a scan time of leq 2h45m plus an equal reduction time. Completely automatic discrimination between stars and galaxies is achieved for most of the objects, and a semiautomated procedure achieves excellent discrimination for nearly all of the remaining objects. It appears that the APS system now provides a state-of-the-art technique for constructing the next-generation galaxy catalog of coordinates and apparent magnitudes, data that constitute a crucial prerequisite for next-generation galaxy redshift surveys.

It should be recognized, however, that the advanced technology (e.g. laser-beam scanning, computer and supercomputer reduction methods) represented by APS cannot now make morphological classifications of galaxies (61), so visual inspection of galaxies on photographic plates is still needed to construct catalogs with extensive, detailed, and internally consistent morphological information on galaxies (59b, 123a, 167a, 188b).

Next Contents Previous