The Role of Images in Astronomical Discovery
Page 35
isophotal map of the smoothed frame with about six isophotes at intervals of 1 magnitude
per square arcsecond; and the surface-brightness profiles along the major and the minor
axes. This was an impressive work that brought atlases of galaxies into the world of “big
data.” It was not until the Sloan Digital Sky Survey (SDSS), with its five-colour set of
filters and greater sensitivity (due to the use of 2.5-m telescope and CCDs), that the scope
and breadth of the Japanese atlas was surpassed.
1994. The Carnegie Atlas of Galaxies , the Ultimate Book of Galaxies
In 1994, Allan Sandage and John Bedke were back on the scene, and this time with a spec-
tacular product. They published the epitome of all galaxy atlases, a grandiose two-volume
publication. Initiated in the mid 1960s, The Carnegie Atlas of Galaxies represented over 25
years of effort.45 Its original concept as the definitive scholarly atlas of galaxies had actu-
ally driven the construction of new telescopes, optimized for wide-field photography in the
southern hemisphere. The project had also been enabled by the setting up of an important
and unique photographic laboratory at the Space Telescope Science Institute in Baltimore,
MD, the NASA Photolab. The Carnegie Atlas of Galaxies comprised images of 1,168 galax-
ies, compiling and illustrating most of the Shapley–Ames galaxies. The presentation was
very much like the successful The Hubble Atlas of Galaxies of 1961.
According to Sandage and Bedke, the new atlas extended and refined The Hubble Atlas
of Galaxies of 1961 and served as the definitive companion to the RSA. Its purposes were
stated as being (1) a textbook of the RSA system of galaxy classification and (2) an illustrated
compendium to aid in planning observing programs of bright galaxies, with both purposes
fulfilling the criteria for an ideal scientific “working object.”
With this ultimate atlas, the reader was to become familiar with the long-term program
of galaxy imaging that Sandage and his colleagues had brought to completion. It had been
a long adventure. The “Sandage project” had actually started in 1910 with the pioneering
work of George Ritchey (Chapter 3). Over a period of 80 years, many photographic plates
of galaxies had been obtained with the large optical telescopes in California and Chile. In
the later decades of the twentieth century, new telescopes designed with wider-field capa-
bilities (in particular the Chile-based Irénée du Pont 2.5-m telescope with its remarkable
2.2-square-degree field of view) helped Sandage and his collaborators to achieve the ambi-
tious and stunning goal of imaging all the bright galaxies in the Shapley–Ames catalogue
(Fig. 10.10). For Sandage, the atlas was the ultimate tool for “decoding cosmic evolution.”
The introductory part of the atlas included some pedagogical musing on classification.
“Can any classification be made independent of the classifier?” asked Sandage and Bedke.
A quote from Francis Bacon (1620) is particularly interesting: “The human understanding,
45 A. R. Sandage and J. Bedke, The Carnegie Atlas of Galaxies, Washington: Carnegie Institution of Washington Publications, 1994.
14:11:09, subject to the Cambridge Core
.012
228
Part III – Organizing the World of Galaxies
from its peculiar nature, easily supposes a greater degree of order and equality in things than
it really finds.” A warning that reminds us of Arthur Worthington’s work and his setback
on splashing liquid drops (Chapter 3).
For example, the idea of continuity or transition in the Hubble classification had been a
fundamental question (and still is). “A purpose of this Atlas is to present evidence for the
continuity. The ultimate purpose of the classification is to understand galaxy formation and
evolution.” Sandage and Bedke concluded, optimistically, “that the modern classification
indeed describes a true order among the galaxies, an order not imposed by the classifier.”
The new atlas aimed to show that there was a stronger case for continuity along the sequence
than could be made in 1961; continuity, but not evolution.
As will be seen in the next chapter, The Carnegie Atlas of Galaxies was extremely well
received. It is an exemplary working object, meeting the goals initially set by Sandage
in The Hubble Atlas of Galaxies 35 years earlier. The atlas is now available on-line, and
it is widely used, due in great part to the large sample of galaxies of all classes. A later
comment by Allan Sandage was almost an understatement: “With this atlas – an all-sky
Carnegie project covering both celestial hemispheres – the vast project of photographing
the galaxies, initiated by Pease and Ritchey in 1910, and continued by Hubble, Baade and
Humason, into the 1950s, was brought to a close.”46
2007. The de Vaucouleurs Atlas of Galaxies , a Masterly Response
Working somewhat independently of Sandage, French astronomer Gérard de Vaucouleurs
had published a new galaxy classification system in the Handbuch der Physik in 1957 and
1959 (Chapter 9). The new system followed the broad categories of Hubble/Sandage but
expanded them with finer details and subtle categories. Gérard de Vaucouleurs, assisted
by his wife Antoinette and other colleagues, had classified several thousands of galaxies
and published the results in a series of galaxy catalogues, e.g. Third Reference Catalogue
of Bright Galaxies ( RC3). In this revised and expanded edition, de Vaucouleurs and col-
laborators presented 23,024 galaxies and about 18,000 had classifications. The reference
catalogues were to some extent competitors to the RSA, but in several ways the works were
complementary to each other.
However, de Vaucouleurs’ classification scheme had never been well illustrated, cer-
tainly not to the level achieved for Hubble’s scheme by Sandage in 1961, or the Hubble–
Sandage revisions and updates of the mammoth atlases by Sandage and Bedke of 1988 and
1994. Although arriving late, Ronald Buta and his colleagues filled the gap most expertly.
In 2007, the researchers came up with a wonderful galaxy atlas: The de Vaucouleurs Atlas
of Galaxies, published in 2007, is the latest of the great atlases of galaxies.47
It is a masterpiece. The authors followed a hint from the astronomer Ivan King who
suggested that the quality of the illustrations was the likely reason why the classification
46 A. R. Sandage, Centennial History of the Carnegie Institution of Washington, Volume 1: The Mount Wilson Observatory, Cambridge: Cambridge University Press, 2004, pp. 490–491.
47 R. J. Buta, H. G. Corwin Jr. and S. C. Odewahn, The de Vaucouleurs Atlas of Galaxies, Cambridge: Cambridge University Press, 2007.
14:11:09, subject to the Cambridge Core
.012
10. Atlases of Galaxies, Picturing “Island-Universes”
229
scheme of the RSA has been so popular. So Ronald Buta, Harold Corwin Jr. and Stephen
Odewahn recreated for the de Vaucouleurs classification scheme what Sandage et al. did for
the Hubble classification scheme: they illustrated the types and the meanings of the class
notation by using large-scale images obtained with large ground-based telescopes, all of
which were carefully annotated.
Paraphrasing de Vaucouleurs’ perspective on galaxy shapes, the authors stated that one
>
should avoid seeing sharp boundaries between the morphological classes of the galaxies.
This is the same idea of “continuity” emphasized by Sandage and Bedke. Buta et al. showed
that the structures of galaxies are wide-ranging, with a continuum of forms, requiring a
classification system not only of precision but also of great flexibility.48 For example “the
intermediate bar classification SAB is one of the hallmarks of the de Vaucouleurs system,
and is used to recognize galaxies having characteristics intermediate between barred and
non barred galaxies.”49
In a noticeable technical difference and innovation, Buta and his collaborators used
modern CCD electronic detectors to image the galaxies, instead of the traditional photo-
graphic plates. It was the first major and systematic classification of galaxies based on digi-
tal images. An obvious advantage of CCDs is the greater dynamic range, allowing features
of highly different contrast to be seen and measured on the same image. They also com-
pleted their illustrative work with the benefit of a solid historical perspective and mature
understanding of galaxies that had been acquired in the 50 years since de Vaucouleurs had
proposed his morphological types.50
The authors also described carefully how the de Vaucouleurs scheme differed from the
Hubble–Sandage revised types. Instead of using the RSA as a sample base, Buta et al. tapped
the general database of images from the astronomical community. Of the 523 images, very
few were intentionally taken for the atlas, in sharp contrast to all other atlases presented
in this chapter. As a working object, the atlas demonstrated the reliability and advantages
of digital images for galaxy classification and magnificently illustrated the de Vaucouleurs
revised classification scheme. It also contains one of the most complete and finest introduc-
tions to galaxy classification. The work is accompanied by a galaxy morphology website,
with the images of the atlas plus additional ones covering as wide a redshift as possible.
Citing examples from the space-based Spitzer Space Observatory, which had provided
images at infrared wavelengths from 3.6 to 8.0 microns, and the satellite GALEX in the
ultraviolet, Buta et al. wrote “there is no question at this time that galaxy morphology is
a vibrant subject.” The authors re-emphasized that the field of galaxy classification had
finally emerged from its pre-1990s second-class status, when it was viewed with disdain,
probably because it was too qualitative and somewhat arbitrary.
48 See for example: J. Kormendy, Observations of Galaxy Structure and Dynamics, in Morphology and Dynamics of Galaxies, Proceedings of the 12th Advanced Course, Saas-Fee, Switzerland: Observatoire de Genève, 1983, pp. 113–288; and R. Buta, Galaxy Morphology, in Planets, Stars, and Stellar Systems, Volume 6, W. C. Keel (editor), Dordrecht: Springer, 2011.
49 R. Buta, et al., Mid-infrared Galaxy Morphology from the Spitzer Survey of Stellar Structure in Galaxies (s4G): The imprint of the de Vaucouleurs Revised Hubble–Sandage Classification System at 3.6 micron, Astrophysical Journal Supplement Series, 2010, Vol. 190, pp. 147–165.
50 G. de Vaucouleurs, Classification and Morphology of External Galaxies, Handbuch der Physik, 1959, Vol. 53, p. 275.
14:11:09, subject to the Cambridge Core
.012
230
Part III – Organizing the World of Galaxies
The new atlas was also didactic, using galaxy morphology to explain their formation
and evolution. The Hubble sequence “did not always exist but was built up over time via
mergers or secular evolution or both.” When did it all fall into place? The authors indeed
proposed a clear answer: between redshift 0.5 < z < 2, i.e. when the universe was between
one third and two thirds of its present age.51
This atlas brings to a conclusion my review of the main galaxy atlases. Nearly half a
dozen more specialized galaxy atlases were published during the same period, which are
discussed in the appendix.
New Trends in Galaxy Imaging
If we count “A Survey of the External Galaxies Brighter Than the Thirteenth Magni-
tude” published by Harlow Shapley and Adelaide Ames in 1932, as the start of galaxy
atlases, there have been about 16 such atlases, including some more specialized atlases (see
appendix), over a period of about 80 years, corresponding to a rate of about one atlas per five
years. It is also remarkable that almost all these atlases used material from photographic,
emulsion-covered glass plates.
Only two research atlases, Hickson’s Atlas of Compact Groups of Galaxies (see
appendix) and Buta et al.’s The de Vaucouleurs Atlas of Galaxies, used modern CCD elec-
tronic cameras to produce images of galaxies for their atlases. This may appear somewhat
surprising since CCDs have been around for the last 25 years. In his review of The Carnegie
Atlas of Galaxies, Brian Skiff commented on the “mature technology of photographs,” and
rightly warned that Sandage and Bedke finished their atlas “just in time before the demise of
the photographic plate. . . . It will almost certainly be the last of major photographic atlases
of large-scale astronomical images to be published.” This statement appeared to imply that
CCD cameras (then offering relatively small fields) would not be able to produce large
images. It has taken some time for CCDs to be in formats large enough to image galaxies
in an effective way. But this situation has dramatically changed, with the current manufac-
turing of large-sized CCD chips and techniques that allow their mosaic assembly in CCD
cameras with a field of view now larger than even the largest photographic plates used by
Sandage and others.52
Nevertheless, compilers of galaxy atlases of the previous decades had good reason to
favor photographic plates as they were of sufficient size to capture the large fields of view of
the telescopes. It would actually be very interesting to see a replication of Allan Sandage’s
The Hubble Atlas of Galaxies or of James Wray’s Color Atlas of Galaxies with images
obtained with the modern electronic imagers on the HST and the large, ground-based fine-
imaging telescopes such as the Very Large Telescope (VLT), Gemini and Subaru.53
51 Redshift ( z) is a widely used unit of distance based on the expansion of the universe; z is measured by spectroscopy. It is defined as follows: 1 + z = λ
/ λ , where λ
is the observed wavelength of a cosmic source, and λ
is the laboratory wavelength
obs
lab
obs
lab
of the same spectral line. The larger z is, the further away is the source and the further in time we are looking.
52 The Ritchey–Chretien 2.5-m Irénée du Pont Telescope at Las Campanas Observatory in Chile had a photographic camera with 50 × 50-cm-sized plates, which fully exploited the 2.1-degree field of view.
53 The giant HyperSuprimeCam on the Subaru 8-m telescope feeds a 116-CCD array (60 cm in diameter) with 870 mega pixels covering a 1.5-degree field of view.
14:11:09, subject to the Cambridge Core
.012
10. Atlases of Galaxies, Picturing “Island-Universes”
231
Atlases of galaxies have also undergone a more subtle evolution. Martin J. S. Rudwick
has described the increasing importance of visual content vis-à-vis text in geological pub-
lications of the nineteenth century and showed this to be mainly the result of evolving
printing technologies.54 Looking at the successive atlases of galaxies, an interesting trend
can be noted: the first generation of atlases (those published before 1990) are dominated by
large-sized images accompanied by very concise, almost telegraphic, text; see, for example,
Sandage’s The Hubble Atlas of Galaxies, or Takase, Kodaira and Okamura’s An Atlas of
Selected Galaxies. In contrast, the images of The Carnegie Atlas of Galaxies (1996), while
still filling the main content of each page, are accompanied by a much longer and descrip-
tive text. The most recent The de Vaucouleurs Atlas of Galaxies (2007) has the highest ratio
of space for text relative to image per page. It is almost as if, as more and more complexities
and details of galaxy morphology are being revealed, authors feel they have to say more,
because the images are less self-evident for the non-expert.
Although their classification has been driven mainly by observations in the visible part
of the spectrum, those in the infrared have reinforced the traditional approach to sorting
galaxies. Galaxy classification is a dynamic area of research and is brimming with new
developments.55 The availability of large-format infrared imagers also opens new possi-
bilities. This has been well demonstrated by Ron Buta and his colleagues with their mid-
IR classification of galaxies using the Spitzer Space Observatory Infrared Array Camera
(IRAC) images of more than 200 galaxies at a wavelength of 3.6 microns. The Spitzer Space
Telescope is a small telescope (with mirror of 85 cm in diameter) flown in an Earth-trailing
orbit around the Sun. It was cooled to 5.5 K and had instruments that produced images and
obtained spectra at wavelengths between 5 and 40 microns (see Fig. 0.4). Having used up
its active cooling liquid helium, the telescope is still operational in the shorter wavelengths
although with limited capabilities. The observatory was named after Princeton University
astrophysicist Lyman Spitzer (1914–1997), who first proposed the idea of telescopes oper-
ating in space in 1946. Buta et al.’s key finding is “that 3.6 micron classifications are well
correlated with blue-light classifications, to the point where the essential features of many