CP-2156 Life In The Universe


Identifiability of Suitable Stars



[335] In view of the uncertainties, a sine qua non for a targeted search is a catalog of the properties of all stars near the Sun. Constructing such a catalog would be useful for a variety of astrophysical problems in addition to the search for intelligent life elsewhere in the Universe.


We do not yet know what kinds of stars are most likely to be the homes of intelligent beings, or even what kinds of stars are most likely to have planets. It would seem reasonable to assume that very young stars (all those of spectral types O and B, for example), variables, some close binaries, red giants, or flare stars are rather inhospitable places, and the anthropomorphic point of view, of course, encourages us to consider most seriously the stars from spectral types F5 to KO. However, such assumptions could easily be wrong and, since we have no idea how common civilizations are, it may be necessary to search widely before we find any. In view of these considerable uncertainties, an important resource for any search program would be a catalog of the properties of all stars within some substantial distance of the Sun. In any type of search program, such a catalog would be useful, and for a targeted search it would be essential.

In this review, I want to describe the current state of our knowledge about the nearby stars, the methods used in learning about them, and the possibilities for a large-scale extension of present catalogs. Such a catalog would be useful for a variety of astrophysical problems in addition to a SETI program.




We must first identify the stellar characteristics that we can hope to measure. Although the fundamental properties listed in the left-hand column of table 1 provide an essentially complete description of a star, several of these properties are not directly observable but must be inferred from a set of observable properties, such as those listed in the right-hand column. By combining observables with a theory of stellar evolution, one should be able to derive the quantities on the right-hand side. Not all of the quantities are of equal importance for either astrophysical purposes or for exploring the possible development of life, and in doing theoretical models most astronomers invoke the so-called Vogt-Russell theorem, which states that the physical properties of a star are determined entirely by its mass, age, and initial chemical composition. The equivalent observable parameters are temperature, luminosity, and surface chemical composition. The Vogt-Russell theorem is not entirely correct in that angular momentum, magnetic fields, and the presence of stellar companions can affect the life of a star and might be important for the evolution of life.

In theory, several (more than two) measurements of the precise position and spectral energy distribution would lead to a complete determination of the character of a star, but in practice a great deal of work is required to deduce even the approximate characteristics of a star. The complete set of properties listed in table 1 are known for only a small number of stars. The position, apparent magnitude, and proper motion of a star are the easiest to determine since they require only two photographs taken many years apart,



Fundamental properties




Apparent magnitude

Angular momentum




Initial composition

Proper motion


Radial velocity

Magnetic fields

Right ascension and declination




Spectral peculiarities

Surface composition

Surface gravity (luminosity)

(spectral type)




[337] ....and its temperature can also be estimated relatively easily from either photographic or photoelectric measurements of the star's color or from low-resolution spectroscopy (spectral classification). A star's surface gravity (which is closely related to its luminosity) and its composition require higher spectral-resolution spectroscopy or narrow-band multicolor photometry and are therefore substantially more tedious. Parallaxes are even more time-consuming, requiring many photographs over a period of several years to obtain even an approximate value, although the distance to a star can be inferred roughly from its apparent magnitude and spectral class. Radial velocity and rotation require relatively high-dispersion spectroscopy, but variability, stellar companions, and some peculiarities are usually found in the course of observing the other properties. Finally, interstellar absorption should be mentioned, though it is not, strictly speaking, a stellar property, but is the dimming and reddening of starlight by dust between the star and the Solar System. This becomes increasingly important for stars more distant than 100 parsecs, but if the physical properties of a star are measured by a combination of photometry and spectroscopy, then the amount of interstellar reddening can usually be separated from the other properties.

For our present purpose, the important properties of a star are its spectral class (or temperature and surface gravity), position, composition, and distance. All of these can be determined approximately for large numbers of stars. In some cases, the presence of a close companion may profoundly affect the character of a star, and I have hidden a considerable variety of phenomena under the heading "spectral peculiarities"; but as a rule, when these properties are important, they are also easy to detect.

Finally, a star's age is a property of obvious importance for SETI. Although in a formal sense a star's age is a function of the other observable quantities, as a practical matter it is not generally possible to measure a star's properties with sufficient precision to permit a useful estimate of its age. It is possible to say something about the ages of certain spectral types (such as O stars or T-tauri stars) or stars in clusters, but generally one would have to say that at present the age of a star is an indeterminable quantity.




Astronomers have been compulsive catalogers, beginning in the 2nd century B.C. with Hipparchus, who compiled a catalog of the positions of 850 stars. The great age of astronomical catalogs was the last half of the 19th century and the beginning of the 20th. We have from this period the Harvard Revised Photometry of 9110 stars (which has been transformed into the Catalogue of Bright Stars), the Bonner Durchmusterung and the remarkable [338] Cape Photographic Durchmusterung, giving accurate positions and approximate magnitudes for 454,875 stars. The greatest of all these catalogs is, of course, the Henry Draper (HD) Catalogue (Harvard Observatory Annals, volumes 91-99, 1918-1924) published almost 60 years ago, which contains the spectral types of 225,300 stars.

All this work was done by hand, yet the HD catalog at least has not been surpassed, despite the enormous improvements of modern technology. More recently, the interest of astronomers has turned to other problems, and the catalogs of the last 50 or 60 years are somewhat specialized in nature (such as catalogs of proper motions or photometry) or are collections of data published separately by many astronomers. Table 2 is a partial listing of the major astronomical catalogs presently in use. Although some of the brightest stars are well studied, the overwhelming majority of the fainter stars appear in only one or two of the catalogs; that is to say, our knowledge of them is incomplete.

The lack of completeness and uniformity in catalogs of astrophysical data has been a continuing source of difficulty to astronomers, particularly since many of the catalogs in common use are simply compilations of work done in piecemeal fashion by several astronomers. This is one of the important aspects of the HD catalog: it is a uniform, systematic, and (nearly) complete tabulation of spectral types of stars brighter than magnitude 9.5 or so.

For our present purposes, ordinary catalogs of stellar data have another serious deficiency. They are generally magnitude-limited; that is, they are....




Number of stars

Type of data


Henry Draper


Position, apparent magnitude, spectral type



Position, proper motion

G.C. Trigonometric Parallaxes


Trigonometric parallaxes



Information on visual binaries

G.C. Variable Stars


Information on variable stars

Radial Velocity Catalog


Radial velocities

Michigan Spectral Catalog


Two-dimensional spectral classes

Bright Star Catalog


Properties of bright stars

Nearby Stars (Gliese,1969)


Properties of stars within 20 parsecs

Nearby Stars (Woolley et al. ,1970)


Properties of stars within 25 parsecs

C.S.I. (Strasbourg)


Cross-reference to all major catalogs

a Will eventually reach 225,300.



[339] ....limited to stars brighter than some limit. It would be preferable to deal with a volume-limited sample, but the enormous range in the intrinsic luminosities of stars has made this an impractical goal. Consider, as an example, a volume of space 100 parsecs in radius centered on the Sun. There are within this volume approximately 250,000 stars brighter than absolute magnitude 14 (spectral type M5), but only 15,000 or so of these stars are in the HD catalog (Allen, 1973). Even in a volume 20 parsecs in radius around the Sun (as in the Gliese catalog in table 2), as many as 2/3 of all the stars may not yet have been noted in any catalogs (Wielan, 1974). It is true that the missing stars are extremely low-luminosity objects (many of them white dwarfs) and may not be friendly places for life, but we do not really know if that is a reasonable assumption.

Since the HD catalog was completed in 1924, there have been few systematic, large-scale surveys of stellar properties, although the basic astronomical data base has been enormously enlarged through a large number of limited studies.

Several projects are now underway to survey various stellar properties, and I will mention two projects to redo the HD catalog to give more detailed spectral classifications of the stars. Houk and colleagues are using objective prism spectra from the Schmidt telescope located at Cerro Tololo Observatory in Chile to reclassify the spectral types of the HD stars in the Southern Hemisphere. A traditional approach is being used in this project in the sense that the stars are being classified by visual inspection of the spectra by one person (Houk and Bidelman, 1979). In the Northern Hemisphere, a group of astronomers have set up a new observatory near Monterey, California, and plan to reobserve the northern half of the HD catalog (Overbye, 1979). They are taking a more modern approach and will be imaging the spectrum of one star at a time onto a solid-state, diode array detector, producing a spectrum in digital form that will be fed directly into a computer. These data should permit rapid, semiautomatic determination of the properties of the stars.




Present catalogs provide a somewhat limited search list for SETI. Is it practical or useful to undertake a large-scale extension of the HD catalog using available technology? A survey that reached magnitude 14 (i.e., a 5 magnitude increase over the HD limit) would contain about 90,000 stars within 100 parsecs, including virtually all stars of spectral type M0 or hotter. However, to identify the nearby stars, it would be necessary to observe all 15 million or so stars brighter than magnitude 14, virtually all of them luminous stars located farther than 100 parsecs from the Sun (Allen, 1973).

[340] Although the more distant stars may not be relevant to SETI, such a massive catalog could be useful for a wide variety of astrophysical problems, particularly in helping to determine the structure and evolutionary history of the Galaxy.

Most of the catalogs of the past century have been based primarily on the measurement of photographs. It is possible, in fact, to photograph the entire sky in a relatively short time with a modest telescope. A photograph is also an incredibly efficient information-storage device (a single 35-mm picture can contain the equivalent of at least 107 bits of information), but the problem is one of obtaining rapid access to that information. So far, the best way to do it has been to make use of two other remarkable instruments, the human eye and brain, and this is why the HD catalog was so successful and why Houk is using the same approach 60 years later. Unfortunately, 15 million stars are too many to measure even for the most dedicated astronomer.

It is now possible to transmit data directly from the focus of a telescope to a computer for analysis, thereby saving the time and expense of intermediate storage in the form of a photograph. There are now several types of photoelectric array detectors, ranging from television systems to solid-state photodiode arrays, the most promising being the charge-coupled device (CCD), consisting of an array of small photodiodes on a single silicon chip. The quantum efficiency of the CCD is close to unity with spectral sensitivity from 4000 to 9000 Å. When extreme care is taken in their manufacture, CCDs can be made that will detect fewer than 10 photoelectrons per picture element (pixel). The chief limitation of the CCD is the relatively small number of pixels compared to a photograph. The largest experimental CCD arrays that have been manufactured are the 800 x 800 pixel arrays to be used in the space telescope.

Although perfect arrays of this size will always be difficult to produce, reasonably good, large arrays should become widely available in a few years, and many astronomers are already talking in terms of replacing photography with CCDs. While this may be a somewhat overoptimistic view, CCDs would appear to have tremendous potential.

A possible design for a survey system might consist of a meridian telescope (fixed to operate only on the celestial meridian) with a mosaic of CCD arrays at its focus. The telescope itself would be inexpensive; the system would acquire combined astrophysical and positional data at a rapid rate and be precise in measuring the positional data. By dividing the light into several beams to measure several bandpasses simultaneously, some spectral information could be acquired.

Such a system would operate by reading out the CCD arrays about 15 times per second and calculating the time and precise path of each star as it drifts across the field of view. If each pixel covered 0.5 arcsec, a star on the celestial equator would drift across an 800X800 array in a little under [341] 30 sec. Its right ascension and declination could be calculated with a precision of about ±0.01 arcsec and the brightness could be recorded with a precision of about 0.02-0.03 magnitude in each bandpass.

By separating the light into 1000-A bandpasses, it would be possible to record the magnitude of the star in five bands with sufficient precision to determine the basic astrophysical character of the star. A telescope with an aperture of 1.5 m or so would permit observations of stars as faint as magnitude 14. The possible characteristics of a meridian CCD telescope are given in table 3.

It is important to emphasize what this survey would not do. It would not yield highly accurate parallaxes or spectral types, nor would it detect planetary or even, in many cases, stellar companions. It would tell us the approximate temperature, distance, luminosity, and any unusual properties of 15 million or so stars. One could then abstract from this data set those stars of interest for a particular project. For SETI, one might want to separate out all nearby stars, which could then be studied in more detail using more conventional techniques.




1.5 m

Focal ratio



9 CCDs

Array size

7200 x 800 pixels

Field of view

1° x 0°.11

Pixel size

0.5 arcsec

Data rate

109 bits/sec

Average number of stars < mag 14 in field of view


Counts/readout/1000 Å (for magnitude 14 star)






- Allen, Clabon Walter: Astrophysical Quantities, 3rd ed. Athlone Press, London, 1973.

- Gliese, W.: Catalogue of Nearby Stars. Veroff. Astron. Rechen-Inst. Heidelberg, no. 22, 1969.

- Houk, N.; and Bidelman, W. P.: Reclassification of the HD Stars on the MK System: Current Status and Future Plans. Bull. Am. Astron. Soc., vol. 11. no. 2.1979. n. 395.

[342] - Overbye, D.: Making It in Monterey. Sky & Telescope, vol. 57, 1979, pp. 223-230.

- Wielan, R.: In: Highlights of Astronomy, G. Contopoulos, ed., D. Reidel, Dordrecht, vol. 3, 1974, p. 395.

- Woolley, R.; Epps, E. A.; Penston, M. J.; and Pocock, S. B.: Catalogue of Stars Within Twenty-Five Parsecs of the Sun. Roy. Obs. Ann., no 5, 1 970.