required for ultraviolet survey work. The first, developed by Meinel and Tifft at the University of Arizona, uses three mirrors to provide a flat field 5 degrees in diameter with angular aberrations of 5 arc seconds or less. This optical system is illustrated in figure 3. The designers state that large aperture versions of this system can achieve image diameters in the range of 1 arc second or better.

telescope with a 1-meter aperture and an f/3 focal ratio should be ideal for a directimaging survey of the entire sky. If a suitable grating were to be ruled on the correcting mirror, this system would also serve as an extremely efficient two-mirror, objectivegrating spectrograph. Current grating technology makes it unlikely, however, that apertures in excess of 12 inches should be considered for such a system in the near future. Therefore, I would conclude that the problems of wide-angle ultraviolet optical systems are essentially solved.

Figure 3. Optical arrangement of the Arizona wide-field, all-reflecting telescope.

The second design is the all-reflecting Schmidt telescope, developed by Lewis C. Epstein of Chrysler Corporation. In this system, the conventional transmitting correcting plate of the classical Schmidt telescope is replaced by a reflecting element. Even though light must pass through the reflecting corrector element in an off-axis manner, the resulting aberrations are small. A 6-inch-aperture prototype has been constructed and yields image diameters less than 4 arc seconds in diameter over a 10-degree field (fig. 4). Ray tracing analysis indicates that the aberration-limited image diameter over this field is less than 1 arc second. Such a

Figure 4. The region of the Andromeda Galaxy photographed with the Epstein all-reflecting Schmidt telescope. Photo by Lewis Epstein, Chrysler Corporation.

In the little time left for discussing other problems, I would stress the development of ultraviolet longwave cutoff filters as one of the most important as well as one of the most difficult problems remaining. If such filters were available, the last obstacle in the way of extending the Palomar Atlas type of survey to ultraviolet wavelengths would be overcome. Perhaps the most practical current solution to this problem is the reflective broad-band interference coatings, which are being incorporated into telescopes proposed by Tifft. This is not yet an ideal solution, however, because the reflection bandwidths are broad and are affected by longer wavelength bands that are

difficult to eliminate when the desired wavelength band is in the 1000- to 1500-angstrom region. Another current solution is the simple aluminum-dielectric interference filter. Unfortunately, these filters are rather inefficient at wavelengths of less than 2000 angstroms, and there is a question of whether they could be produced in the sizes required to cover the photographic fields desirable for survey work. Such fields may be expected to have a diameter of at least 6 inches. Still a third approach is the alkali metal transmission filter. Although such filters appear promising in theory, the problems of producing and preserving them are severe and have not yet been solved. A small program for investigating these problems is being carried out at Northwestern University. The problem of achieving long wavelength cutoff has been efficiently solved for electronic detectors by photocathode technology. The cesium-iodide photocathode provides an excellent method of detecting wavelengths shorter than 1500 angstroms without interference from longer wavelengths.

Summary

In summary, I would like to emphasize two points. First, ultraviolet survey work is an important part of space astronomy and deserves greater attention than it is currently receiving. Not only is it necessary as a means of searching for that small but exciting percentage of stars that may be expected to show anomalies in their ultraviolet radiation, but also it is a means of making moderateaccuracy, spectral, and color data for hundreds of thousands of stars available to all astronomers for statistical studies or for studies of specific sets of stars. Second, if we grant that survey work should be a part of space astronomy, it is to be anticipated that, in the next 10 to 20 years, the most useful and significant of these surveys will probably employ the photographic emulsion as a detector. Figures 5 and 6 show the tremendous information-gathering capacity of the photographic emulsion. In these figures, we see a region of the southern Milky Way in

The analogy has been made that talking about the future of infrared astronomy at this time is like having to talk about visual astronomy if all the data that were available were observations which had been made in daylight. I think this comparison is a valid one; hence, most of this discussion today will be about the kinds of measurements that have been made in the past in order to give a frame of reference to what we are working on now and what this can lead to in the future. I will limit my discussion to the stellar type of objects.

For astronomy, the infrared region can be divided into two regions. One region, approximately from 1 to 20 microns, is accessible to ground observations. It is characterized by transmission windows in the earth's atmosphere at about 1, 2, 3, 5, 10, and 20 microns. Detectors used in this region are of the photoconductive type, which will work over this entire range. Examples of these are the lead sulfide and lead selenide detectors. At approximately 5 microns, the germanium bolometer of Frank Low becomes competitive with the photoconductive detectors. At 10 microns and 20 microns, the bolometer is probably a superior detector.

The second wavelength region in the infrared is from approximately 20 microns to 1 millimeter, where optical techniques start becoming radio techniques. Essentially, there are no windows in the earth's atmosphere in this region. Thus, in order to study astronomy at these wavelengths, we have to go outside the earth's atmosphere.

Observations made from the ground have been done with conventional telescopes. The idea is to use a conventional telescope and to

put a photometer at the focal plane. The only novel feature about this type of instrumentation is having a two-beam photometer so that one portion of the sky can be compared against an adjacent portion, thus eliminating the effects of the earth's atmosphere. The technique is fairly straight forward, and, in general, telescopes with diameters that range from 24 inches to 200 inches have been used in this way.

There has been one special telescope (fig. 1) built at the California Institute of Technology. It is a 62-inch-aperture telescope and is made of epoxy that has been spun in the earth's gravitational field to generate the parabolic primary. This fulfills the requirement for a lightweight telescope; it does not fulfill the requirement of a high-resolution surface figure. Its resolution is only about 1 minute of arc, which is good enough to meet the objective of a lightweight, easy to

use, inexpensive telescope for conducting a survey in the infrared.

To illustrate how the difficulty of making measurements increases with wavelength, it is interesting to note that on the survey, which was observing at 2.2 microns, we measured between 10,000 to 20,000 objects. At 10 microns, I estimate the number of objects that have been published is only 200; at 20 microns, it is well under 50.

Some measurements have already been made from above the earth's atmosphere. The first were done with balloons. This type of measurement is mainly the work of John Strong, who pioneered this field primarily in planetary work. There has been some stellar spectroscopy with the Stratoscope, and, more recently, there have been observations from balloons at 100 microns to look at the galactic center. It is interesting that this latter telescope had only a 1-inch aperture. Both the Naval Research Laboratory and Cornell have done some rocket work that has been designed mainly to look at the infrared background by using cooled telescopes; these have numbered in the tens of flights. There has also been an airplane program employing a Convair 990 and some work by Frank Low with a 12-inch-aperture telescope in a Lear jet. Most of the work has been done in connection with spectral measurements of the planets. Stellar observations are just beginning.

What types of objects do we plan to measure in the infrared? Ten years ago, if the question had been asked: "What are the infrared characteristics of the stars?", the answer would have been that most stars look just like black bodies with temperatures typically of the order of 3,000° Kelvin or higher. Johnson, at Arizona, measured a large fraction of the bright stars in the infrared and found that this was indeed the case.

Within the last five years, and especially in the last year, this picture has changed radically. This has been the result both of extending the wavelengths beyond 2 microns to the 5-, 10-, and 20-micron windows and of making an unbiased search for infrared objects. The types of infrared stars can be