Planetary radar astronomy was a problem-solving activity, an algorithm in search of a problem. Its fundamental driving force was the dynamic interaction between radar techniques and the kinds of problems radar astronomy solved. Improvements in radar hardware and innovative radar techniques, such as range-Doppler mapping, allowed radar astronomy to solve scientific problems of interest to astronomers and geologists. Conversely, problem-solving could bring attention to radar techniques and properties previously neglected or little used, such as the polarization of echoes.
The institutional and financial linking of radar astronomy to NASA at Arecibo and JPL gave the field a mission-oriented cast. The justification for funding was the field's utility to NASA space missions, and access to Goldstone antenna time required specific mission support. Beginning with Viking, participation in NASA missions also brought ground-based radar astronomers into closer collaboration with the radar scientists at the Stanford Center for Radar Astronomy. The distinction made in the 1960s between ground-based planetary radar astronomy and Stanford's "space exploration" held less and less meaning.
Planetary radar astronomy after about 1975 also remained above all else a science driven by technology, namely, access to radars with the transmitter power and antenna and receiver sensitivity to explore the planets. Without those radars, radar astronomy could not exist. The decline of radar astronomy at JPL followed directly from the deteriorating state of the Goldstone radar. Improvements in radar hardware, on the other hand, drove planetary radar forward.
Additional transmitter power and receiver sensitivity meant access to previously unexplored targets. The orbit of Mars defined the outer reaches of planetary radar astronomy until 1975, when both the Arecibo and Goldstone radars underwent upgrades that significantly enhanced their value as research tools, as discussed in Chapter Four. For the first time, the Galilean satellites of Jupiter, the rings of Saturn, cometary nuclei, and a number of both Earth-approaching and main belt asteroids came within reach of those planetary radars. Those targets represent considerable radar distances; the round-trip radar time to the moons of Jupiter is about 1 hour and 12 minutes and to Saturn's rings around 2 hours and 15 minutes.
Meanwhile, the planetary radar astronomy community remained small, and Arecibo and Goldstone were the only active research facilities. Arecibo was still a major NSF-funded center for radio astronomy and ionospheric research. On the other hand, funded by NASA, not the NSF, and associated with exploration of the solar system, radar astronomy there occupied a small, peculiar niche, a niche that, nonetheless, furnished a research facility for both Cornell and MIT graduate students to be trained as future radar astronomers.
In contrast, Goldstone did not train graduate students. The radar astronomers at JPL did not hold the kind of appointment at Caltech that permitted them to train graduate students as future radar astronomers, and no Caltech professor was interested in training radar astronomers. A similar situation had existed at Lincoln Laboratory during the 1960s until Pettengill's appointments at Arecibo and his subsequent teaching position at MIT changed that situation and provided the institutional matrix for the training of graduate  students as future radar astronomers. In short, the teacher-disciple pattern that prevailed at Arecibo was lacking at JPL, where radar astronomers propagated through job hiring. Planetary radar astronomy at JPL remained unofficial and invisible. Between 1978 and 1986, furthermore, essentially no radar astronomy work took place at Goldstone, because investigators lacked a reliable research instrument.
Among the new radar targets brought into range by the Goldstone X-band and Arecibo S-band upgrades were Ganymede, Europa, Callisto, and Io, named the Galilean moons of Jupiter after their discoverer, Galileo Galilei. The radar exploration of those moons illustrated the interactions between radar astronomers and geologists, as well as the increasing collaboration with Stanford researchers that came to typify ground-based planetary radar. Those moons also puzzled radar astronomers. Never before had they encountered such peculiar radar characteristics among the terrestrial planets. An explanation for the bizarre radar readings came from Earth and from leading edge research in the physics of light.
The first, though unsuccessful, attempt at the Galilean moons took place in 1970. Dick Goldstein (at Goldstone) and Dick Ingalls and Irwin Shapiro (at Haystack) tried to detect echoes from Callisto using the bistatic Goldstack radar, in which the Haystack 300-kilowatt telescope transmitted and Goldstone received.1 The experiment did not work, however, because of a misunderstanding over polarization.
After unsuccessfully attempting Venus with the Goldstack radar, Ingalls and Goldstein pointed the radar at the Moon and received "the weakest of signals." Goldstein, trained as an electrical engineer, realized what was wrong. Bistatic radars require investigators to agree on the polarization of the wave. Physicists, like Shapiro, use one definition for left-handed polarization, defining handedness from the view of a person looking in the direction that the wave is travelling, while electrical engineers use the opposite convention, defining handedness from the view of the receiving antenna, so left and right are reversed. Goldstack eventually searched for Ganymede and Callisto in late May and early June 1970.2 The polarization of radar echoes was about to become a key radar technique for studying the Galilean moons and other solar system bodies.
Dick Goldstein and George A. Morris succeeded in detecting Ganymede with the 400 kilowatts of Goldstone S-band radar power on six nights in late August 1974. Those echoes set a record for the longest time of flight to a radar target, one hour and seven minutes. The echoes, though, were very weak, well below the noise level. From those weak echoes, Goldstein and Morris drew conclusions about the surface of Ganymede.
From the total signal power returned and the width of the spectrum, they concluded that Ganymede "must have a considerable degree of roughness." Their data did not agree with accepted theory, derived from infrared spectra and polarization studies, that Ganymede's surface consisted mostly of ice.3 Goldstein and Morris ventured that the most  likely possibility was for the surface to consist of rocky or metallic material from meteoric bombardment embedded in a matrix of ice.4
Soon after the Arecibo S-band upgrade reached completion, Don Campbell (NAIC Research Associate) and Gordon Pettengill (MIT) made the first radar detections of Callisto and Europa on 28 September and 5 October 1975, respectively, and detected Ganymede on 30 September. Pettengill and Campbell noticed that the satellites had an unusual radar signature. The three moons were almost uniformly radar bright; they lacked the bright specular return from the subradar point, the area on the target closest to the Earth, that all terrestrial planets exhibit. The uniformity of brightness suggested that the satellite surfaces were probably extremely rough on scales comparable to or larger than the wavelength of 12 cm.
Io remained an elusive radar target. The innermost of the Galilean moons, Io is inside Jupiter's magnetosphere, which may have interfered with the radar waves aimed at Io. Campbell and Pettengill unsuccessfully attempted the satellite twice in 1975, and their attempt to detect Io in January 1976 yielded only a weak echo that indicated an error in the ephemeris large enough to explain the previous failed attempt. Not until 1987, when improved hardware was available, did radar astronomers begin to receive good echoes from Io.
After reducing their January 1976 data on the four Galilean moons, Campbell and Pettengill found surprisingly large radar cross sections for Europa and Ganymede, approximately 1.5 and 0.9 times the geometric cross section, respectively, while those for Callisto and Io were around 0.4 and 0.2, respectively. The radar cross section is a measure of target brightness. Although the values for Callisto and Io were low and typical of the terrestrial planets, the radar cross sections for Europa and Ganymede were abnormally high.5
When Pettengill and Campbell resumed their observations of Jupiter's moons in October 1976, the Arecibo radar had a dual polarized circular feed paid for with NASA S-band operations funds. The feed increased total system sensitivity over that available in 1975 and displayed the peculiar radar polarization properties of the Galilean satellites.
Previously, all observations of the Galilean moons had been made with linear feeds in both orthogonal linear polarizations. The transmitter sent out signals with one sense of polarization, and the antenna received both the same linear and orthogonal linear polarizations. The same linear echoes are much stronger than the orthogonal linear echoes for all targets detected by radar. Although the switch from linear to circular polarization did not alter the general character of the spectra for Callisto, Ganymede, and Europa, the circular polarization ratios of the echoes were totally unanticipated.
When radar astronomers transmit a right-handed circularly polarized signal, they expect the echo to return mostly left-handed circularly polarized, the opposite handedness. This type of polarization return is called variously the "expected," "polarized," or "opposite circular" (OC). The echo power returned right-handed circularly polarized is said to have "unexpected," "depolarized," or "same circular" (SC) polarization. The SC-to-OC ratio is known as the circular polarization ratio.
The terminology "expected" and "unexpected" is out of place today. The "unexpected" polarization returns from the Galilean moons and other icy targets are no longer considered unusual or "unexpected." The terms, however, reflected the surprise of radar astronomers in the past, as they discovered polarization returns that differed markedly from those of the terrestrial planets. For the sake of preserving that historical flavor of discovery, and to avoid using terms likely unfamiliar and perhaps confusing to the reader (such as "polarized" and "depolarized"), the terminology "expected" and "unexpected," or OC and SC, will be used throughout.
 In radar observations of the terrestrial planets and the Moon, more power normally returns in the expected than in the unexpected mode. The circular polarization ratio for these targets is about 0.1; for Venus and the Moon, it is only about 0.05. In the case of Jupiter's moons, however, more power returned in the unexpected mode, a phenomenon called circular polarization inversion. For Europa, Ganymede, and Callisto, the average circular polarization ratios were 1.61 ± 0.20, 1.48 ± 0.27, and 1.24 ± 0.19, respectively. They were the first solar system objects for which circular polarization inversion was observed.6
The dominance of unexpected polarization from the Galilean satellites was enigmatic and even unbelievable. "That was a bit of a puzzle," Don Campbell recalled. "There was a lot of skepticism, frankly, about the results....That was a really significant puzzle to everybody."7 The phenomenon was also a puzzle to Steve Ostro, then a graduate student at MIT working under Gordon Pettengill. Ostro was looking for a dissertation topic. He joined Pettengill and Campbell in observing the Galilean satellites at Arecibo in late 1976. "The anticipation," Ostro explained, "was that working on those observations, as well as on the data reduction and interpretation, would evolve into a good thesis topic."8
When the bizarre circular polarization inversion first appeared during the 26 October through 7 December 1976 observations, Ostro recalled, "We tested to the point of grasping at straws. Maybe we had crossed the cables. Or maybe somebody had screwed up in the data acquisition program. We checked everything. We couldn't believe it, just couldn't believe it." A test on Venus returned normal echoes. Then they pointed the telescope at Europa, and the circular polarization ratio was about one and a half. At that point, Ostro remembers watching Pettengill reflecting then saying, "Well, now I have to believe it." Then he turned to Ostro and said, "If you can explain this, it would be a good thesis topic."9
In order to investigate systematically the unusual radar cross sections and polarization ratios of the Galilean moons, Ostro, Campbell, and Pettengill undertook a new series of 20 observation sessions in November and early December 1977 and obtained results similar to those found the previous year.10
Ostro, Pettengill, and Campbell continued their campaign on the Galilean satellites in February 1979 and March-April 1980, when the satellites were in different phases. Also, in order to determine whether the strange polarization ratios were a function of frequency, Don Campbell undertook a separate series of observations with the old 430-MHz (70-cm) radar and obtained a weak detection of Europa, but not of Ganymede.11 Jupiter then left the declination window of the Arecibo Observatory until 1987.
In order to account for the unusual radar signatures of Europa, Ganymede, and Callisto, Steve Ostro developed a model, published in 1978. The model postulated a thick surface layer of ice saturated with nearly hemispherical surface craters. Hemispherical craters would favor double reflection of radar waves at a 45° angle at each reflection, so that most of the signal would return with the same handedness of polarization. The same craters could be made to explain the high radar cross sections, as well.12
 Dick Goldstein and Richard R. Green at JPL proposed a different model based on their own observations of the Galilean satellites. After the pioneering observations of 1974 at S-band, Goldstein took additional data on Ganymede during six nights in December 1977 with the Goldstone X-band radar and received alternately right-handed and left-handed circular polarization, in order to compare the expected and unexpected echo strengths. Despite the high transmitter power (343 kilowatts) and low system noise temperature (23 K), the Ganymede echoes were noisy. Nonetheless, the Goldstone data confirmed the Arecibo results, which had been the subject of great incredulity. As Don Campbell recalled, "That confirmation started a significant discussion about the phenomenon. Why were we getting these odd reflections?"13
From the spectral data, Goldstein and Green measured the radar cross section and polarization ratios and posited a model of Ganymede's surface. They assumed that the upper few meters of its surface consisted of ice "crazed and fissured and covered by jagged ice boulders." The critical part of the model was a large number of interfaces between ice and vacuum where, depending on the angle of incidence above or below a certain limit (called the critical angle), the sense of polarization was largely preserved and most of the power remained in the original polarization sense. In a 1982 review article, Steve Ostro concluded that "many questions remain about interpretation of the radar results, but we seem to be pointed in a sensible direction."14
Voyager 1 had begun sending back pictures of the Jupiter system in early 1979. Geologic activity on Ganymede appeared varied, while Callisto's entire surface was densely cratered. Europa probably was covered completely by ice.15 More information than ever was available about the surfaces of the Galilean satellites, yet none of it resolved the questions raised by planetary radar astronomers, who, in the meantime, attempted to explain the strange radar characteristics of the Galilean satellites based on reflection geometries and radar scattering rules, not the geology of those worlds as revealed by Voyager imagery.
Among those offering explanations for the high cross section and circular polarization inversion was Tor Hagfors. He proposed that the satellites' unusual radar signatures echoes were due not to reflections at the interfaces of ice and vacuum, as Goldstein and Green had suggested, but rather to the bending of the incident wave around continuous gradients in refractive index.16 Von Eshleman developed an argument around refraction scattering from imperfect spheroidal lenses. Then he modified his argument and incorporated Ostro's notion of hemispheroidal impact craters, as well as elements from the Goldstein-Green model.17
The Ostro, Goldstein-Green, Hagfors, and Eshleman models all rested on radar geometries and scattering mechanisms. Not a single model linked surface or subsurface structure realistically to the radar signatures, nor did the models explain the origins of those structures. Positing the existence of hemispherical craters was one thing; finding geologic evidence for them was another. Not surprisingly, Voyager revealed no hemispherical craters on any of the Galilean satellites. Ostro now sought an explanation for the radar signatures of the Galilean moons in collaboration with USGS planetary geologist Eugene Shoemaker.
 Shoemaker had a rather simple and elegant geologic solution to the problem. In developing his solution, Shoemaker drew upon his knowledge of the lunar regolith and Voyager data. He assumed that the surfaces of the Galilean moons were exactly like that of the Moon. From statistics of craters observed in Voyager images of Ganymede and Callisto, Shoemaker inferred that the surfaces of those moons had a history of meteor bombardment similar to that of the Moon. He concluded that they were probably blanketed with fragmental debris produced by prolonged meteoroid bombardment. The only difference, then, between the Moon and Jupiter's moons was that the rocks on the Galilean satellites were made of ice, and the ice, given the extremely low ambient temperatures, would behave like a silicate rock. Ice is highly transparent to radar waves, so the icy surfaces of the Galilean moons would permit radar waves to penetrate those surfaces to a far greater extent than if they were made of silicate rock. The combination of the greater penetrating depth and the greater number of scattering events could provide an explanation for the peculiar radar signatures of the Galilean satellites.18
The primary contribution of the Ostro-Shoemaker model was its geological perspective. Nonetheless, the model only partially explained the radar results; a satisfactory understanding of the detailed scattering mechanism that gave rise to the odd radar signatures still remained beyond reach. Meanwhile, Steve Ostro and Don Campbell had begun a new series of radar observations of the Galilean satellites at Arecibo in 1987. Unlike the previous campaign, Stanford researchers under the leadership of Von Eshleman participated. Dick Simpson took data at Arecibo, while a graduate student, Eric Gurrola, was charged with the analysis. Tor Hagfors, who also was interested in experimenting on the Galilean satellites for reasons similar to those of the Stanford group, joined their group.
This new series of S-band observations were to provide thorough phase coverage for all three icy satellites (Ganymede, Callisto, and Europa). Started in November 1987, the campaign continued into 1988, then November-December 1989, January 1990, and February-March 1991, when Ostro observed the satellites at rotational/orbit phases chosen to fill in gaps in the 1987-1990 phase coverage.19 Then Jupiter left the Arecibo declination window.
At the same time, Arecibo obtained the first good echoes from Io. Its radar properties were unlike those of the other Galilean satellites. Data collected in 1976 already had shown that Io's surface was significantly rougher on average than the terrestrial planets, but much smoother than the other Galilean moons. Its radar cross section and polarization ratio were more typical of the inner planets, however, and argued strongly against the presence of significant quantities of surface ice.20
In parallel with the 2,380-MHz (12.6-cm) observations, Don Campbell studied the Galilean moons with the 430-MHz (70-cm) radar beginning in November 1988, the first time in 25 years that the UHF radar had been used in the continuous-wave mode. He detected Ganymede and Callisto, then in November-December 1989, made the first UHF detection of Europa. The purpose of the experiment was to compare the polarization properties of the Galilean satellites at both S-band and UHF. Campbell discovered that the echoes from Ganymede at UHF were reminiscent of those at S-band. Additional UHF measurements made in January 1990 apparently confirmed that the peculiar polarization ratios of the Galilean moons were independent of frequency.21
 Steve Ostro, who now had a position at JPL, also observed the Galilean satellites with the Goldstone X-band radar between 1987 and 1991 and measured polarization ratios and radar cross sections. The combined X-band, S-band, and UHF radar data taken over a long period of time documented the degree to which the satellites' radar properties depended on target, rotational phase, and frequency.22 They provided a considerable base upon which to explain the bizarre radar signatures of the Galilean moons, and a reasonable explanation soon was in hand.
Toward the end of the Arecibo and Goldstone campaign on the Galilean satellites, Bruce Hapke, an optical astronomer and scattering expert, drew attention to a growing body of literature on laboratory and theoretical investigations of a phenomenon called alternatively "coherent-backscatter effect" or "weak localization." The effect has potential application in a new class of semiconductors in which photons, rather than electrons, perform circuitry functions. Weak localization of light takes place at the microscopic level and arises from a combination of coherent multiple scattering and interference. Backscattered intensity is enhanced, and the forward diffusion through the low-loss medium reduced, by constructive interference between fields propagating along identical but time-reversed paths.23
At the suggestion of Steve Ostro, Kenneth J. Peters of Caltech did calculations that demonstrated that coherent backscattering from forward scatterers could explain the high reflectivity and polarization ratios of the Galilean satellites.24 Coherent backscattering now appeared to explain adequately the high radar cross sections and circular polarization ratios of the icy satellites, and it was consistent with the geologic picture of those moons painted by Gene Shoemaker. The scattering might arise less from individual pieces of ejecta, but more likely from uncoordinated changes in porosity (and hence refractive index) that occur randomly throughout "smoothly heterogeneous" regoliths, argued Ostro and Shoemaker.25
Additional data on the radar properties of icy surfaces came from observations of the Earth. In June 1991, the NASA/JPL airborne synthetic aperture radar (AIR-SAR) flew over a vast portion of the Greenland ice sheet called the percolation zone, where summer melting generates water that percolates down through the cold, porous dry snow then refreezes in place to form massive layers and pipes of solid ice. The AIR-SAR radar observed the Greenland ice sheet at several wavelengths (5.6-, 24-, and 68-cm) and obtained values for the circular polarization ratio greater than one.26
The riddle of the strange radar signatures of the Galilean satellites focused radar astronomers' attention on epistemological questions, the fundamental need to understand and interpret radar echoes and their relationship to the target. Such questions, though, were of interest only to radar astronomers; their solutions contributed to an  understanding of the radar characteristics of planetary surfaces, but not to the more general scientific questions posed by non-radar planetary astronomers. However, if radar astronomers were going to contribute to our knowledge of the Jupiter and Saturn systems, they first had to resolve such basic epistemological issues relating to the radar properties of those planetary systems.
Although the central focus of radar research on the Galilean satellites had been the solution of the satellites' strange radar signatures, the data also has served to correct their ephemerides as part of the Planetary Ephemeris Program of Irwin Shapiro and John Chandler of the Harvard-Smithsonian Center for Astrophysics. The radar data uncovered errors in the ephemerides as early as 1976. A round of Callisto observations carried out beginning in 1987, though, were intended mainly for orbital ephemeris refinement in support of the Galileo mission.27
Sensitized to the needs of planetary geologists, Ostro also attempted to relate radar data collected at Arecibo and Goldstone between 1987 and 1991 to surface features on the Galilean moons. The most prominent features tentatively identified in the echo spectra were Ganymede's Galileo Regio and Callisto's Valhalla Basin.28 Using a new radar coding technique, John Harmon and Steve Ostro observed Ganymede and Callisto at Arecibo from February to March 1992 and obtained the first range-Doppler images of the moons. These observations also constituted the first successful ranging measurements to the Galilean satellites and the farthest radar distance measurements ever reported.29
The exploration of the Galilean moons of Jupiter illustrated the increasing complexity of the planetary radar paradigm. Hardware improvements, coding techniques, and even discoveries made in optics laboratories shaped the science done by radar astronomers. Moreover, despite the shift toward geology, planetary radar remained oriented toward astronomical questions and NASA missions, such as Galileo.
The rings of Saturn, like the Galilean moons of Jupiter, presented radar astronomers with a target very different from the terrestrial planets. The rings of Saturn were believed to be icy and until the 1970s, were thought to consist of tiny, micron-sized particles. Radar astronomy upset that conception of the rings. In doing so, radar astronomy also set a distance record: the round-trip light time to the rings was about 2 hours and 15 minutes.
After an unsuccessful try in 1967, Haystack researchers successfully bounced X-band radar waves off the rings in 1973.30 Earlier, however, in December 1972 and January 1973, Richard Goldstein and George A. Morris, Jr., at JPL detected the rings with the S-band Goldstone Mars Station. Making the observation was not easy. The orientation of the rings is optimum for radar observations only twice during each 29-year orbit of Saturn, when the rings are most tilted to the line of sight and present the largest projected area. At the same time, the Doppler spreading and consequent dilution of the signals in the noise is the least.
 The echoes Goldstein and Morris found were unexpectedly strong. The rings were inclined at an angle about 26° with respect to the line of sight, and the amount of power returned from the rings was about 10 times that for Mercury and five times that for Venus. Moreover, wrote Goldstein and Morris: "Particles of any material that are much smaller than our wavelength [12.6 cm] are ruled out by our data....Large (compared to the wavelength), irregular, rough particles could produce the observed echoes."31
Shortly thereafter, on 31 July and 1 August 1973, JPL organized a workshop on Saturn's rings at the request of S. Ichtiaque Rasool of the Planetary Programs Office, NASA Headquarters. Gordon Pettengill organized the scientific program. The workshop responded to an upsurge in interest in the Saturn system, and the outer systems in general, in anticipation of the 1977 Mariner Jupiter/Saturn mission, later known as Voyager.
The interpretation of the JPL radar experiment on Saturn's rings surprised astronomers32 and caused rethinking about the ring particles and models published by radio astronomers. The amazingly large particle size also raised questions about the safety of a spacecraft near the rings and gave rise to NASA and JPL interest in the radar results, which George Morris discussed at the workshop. Excited by the Goldstone radar findings, astronomers during the general discussion expressed an interest in obtaining more radar data on the rings.33
The JPL results also surprised radar astronomers. For example, Gordon Pettengill (MIT) and Tor Hagfors (then at the Department of Electrical Engineering of the Norges Tekniske Hogskole, Trondheim, Norway), based on their own radar experience with the terrestrial planets and the asteroids Icarus and Toro, felt that the radar cross section observed by Goldstein and Morris, 0.62 ± 0.15, was unreasonably high. "Even by assuming the particulate matter in the rings to have linear dimensions comparable to or larger than the radar wavelength," they wrote, "we are left with the need to explain a radar scattering mechanism more efficient by a factor of about 10 than that of the inner planets, unless we wish to postulate an unreasonable ring particle density or composition."34
Astonished, too, were radio astronomers. The high radar return had to be reconciled with the rings's low radio emission, as well as with optical and infrared results.35 As the enigma of Saturn's rings continued to puzzle astronomers, the Arecibo S-band upgrade reached completion. It seemed only natural, as Don Campbell explained, that the first radar experiment with the upgraded telescope should be an attempt to detect echoes from the rings of Saturn: "When Arecibo first came on line in 1974, the very first thing we did to test the transmitting system, apart from trying to communicate with a star system 25,000 light years away, was to run a bistatic radar measurement on the rings of Saturn  with Goldstone. At that time, we had transmitting capability, but we had not yet installed the receivers. The dedication of the upgraded telescope had been in November 1974, and this was in December, when we were trying to get the transmitter really working properly."36
Despite equipment difficulties at Arecibo, Goldstone received echoes from Arecibo by way of Saturn.37 In addition to the bistatic Arecibo-Goldstone radar test on Saturn's rings in December 1974, Arecibo and Goldstone performed dual-polarization experiments on two nights in January 1975. These bistatic linear polarization experiments established that echoes from the rings of Saturn were highly depolarized, that is, more power appeared in the unexpected than in the expected polarization.
Goldstein also conducted monostatic dual-polarization observations with the Goldstone X-band radar on five nights in December 1974 and January 1975 and measured a high circular polarization ratio. Goldstone and Arecibo investigators now knew that Saturn's rings exhibited high linear and circular polarization ratios and that the phenomenon was independent of frequency. Moreover, they confirmed at both X-band and S-band that the rings had high radar cross sections.38
The high radar cross sections and polarization ratios of Saturn's rings were puzzling. Campbell and Goldstein considered several possible explanations for those radar properties. Two models appeared plausible. One model hypothesized a thick cloud of irregular water-ice chunks a few centimeters or larger in radius. The other posited a monolayer of multimeter-sized water-frost-coated metallic chunks. Voyager data later rejected the metallic composition of the rings.39 In summing up the state of knowledge on Saturn's rings in 1975, Allan F. Cook and Fred A. Franklin of the Smithsonian Astrophysical Observatory speculated that the ring particles consisted of water ice, clathrated hydrates of methane, and ammonia hydrates,40 in agreement with one of the radar models.
Meanwhile, James Pollack and other astronomers proposed that the ring system was diffuse and many particles thick. In order to determine whether the rings of Saturn consisted of one or several layers, and in general to test various models of the thickness and composition of the rings, Gordon Pettengill, Don Campbell, and Steve Ostro undertook further radar observations in 1977, 1978, and 1979 on a total of 13 nights. Like those on the Galilean satellites of Jupiter, the observations became part of Ostro's thesis.41 In March 1977, also, Gordon Pettengill and Dick Goldstein resumed bistatic observations of Saturn's rings with the Arecibo and Goldstone S-band radars.42
The key to the radar observations made in 1977, 1978, and 1979 was the differing tilt angles of the rings during the 13 total nights of observations. The tilt angle of the rings relative to the line of sight declined over those three years from 18.2° to 11.7°, then to 5.6°. The astronomers also received in both senses of circular polarization in order to measure the polarization ratio as a function of tilt angle. Their results, when combined  with earlier radar data and the theoretical calculations of Jeffrey N. Cuzzi and James Pollack,43 provided significant constraints on ring structure.
The observations confirmed that the radar reflectivity of the rings was quite high and that depolarization was also high. The polarization ratio for the Galilean satellites, a mystery not yet solved, however, was higher. The data ruled out all large-particle monolayer models. On the other hand, the polarization and radar cross section results favored ring models of several layers. The radar data also appeared to support particle composition of ice or metal, but not silicate rock.44
Ostro, Pettengill, and Campbell also concluded that the A and B rings (the outermost rings) were responsible for most, if not all, of the S-band radar echoes, and that the radar reflectivity of the A-ring was nearly as great as the B-ring radar reflectivity. The radar reflectivity of the C ring was notably less than that of the B ring. Also, they found no evidence for radar echoes from beyond the A ring or from the planet itself.45
The case of Saturn's rings resulted in radar astronomers contributing to planetary science, in contrast to their studies of the Galilean moons. Those studies for a long time had been limited to epistemological issues, namely, what caused the Galilean moons' strange radar signatures? Radar contributed to Saturn science, on the other hand, by focusing less on such questions of radar technique and more on scientific questions, such as the size of the ring particles and the number and thickness of the ring layers. Although the solution of technical problems was a prerequisite for any radar astronomy problem solving, the lack of obvious relevance to planetary science was a serious matter; the ability to solve scientific problems, especially those relating to NASA space missions, was the basis on which scientists judged the value of radar astronomy and on which funding decisions were made.
The nuclei of comets provided radar astronomers additional icy research subjects. Comets are believed to represent samples of the most primitive material of the solar nebula and to hold clues to the origin of the solar system.46 They make challenging radar targets, because close approaches are rare. The relatively small size of comets dictates that they be studied by radar only when they approach Earth at distances of a fraction of an astronomical unit. Also, ephemerides derived from optical data lack the accuracy demanded for radar observations. Only the S-band and X-band upgrades of the Arecibo and Goldstone antennas made radar studies of comets possible.
 Early attempts all ended in failure. For example, after an attempt in January 1971 on Comet Kohoutek stymied by rain and snow, the Haystack telescope again failed to detect that comet in January 1974. Although Irwin Shapiro had prepared an accurate ephemeris in advance, neither the bandwidth nor the center frequency of the radar echo was known precisely, so they had to search for the echo.47
It took the S-band upgrade of the Arecibo Observatory to make the first comet detections possible. Paul G. D. Kamoun, a French student of Gordon Pettengill at MIT, built his dissertation research around those detections. The main objective of his dissertation was to use cometary radar data to discriminate between two different models of cometary nuclei.48 One model was that proposed by Fred Whipple, who served on Kamoun's dissertation committee, and supported by Zdenek Sekanina, an established expert on comets.
In the Whipple model, the cometary nucleus was like a rotating "dirty snowball," an icy matrix of water ammonia, methane, carbon dioxide, or carbon monoxide, combined with rock, dust and other meteoric debris. A popular model for the nucleus in the early 20th century predicated a "dust swarm" or swarm of solid particles of unknown sizes, each particle carrying with it an envelope of gas, mostly hydrocarbons. However, that model had a number of difficulties, and by the 1970s Whipple's "dirty snowball" model prevailed.49 Consequently, Kamoun's dissertation did not contribute meaningfully to the comet debate.
Kamoun's research on comets turned around the unsuccessful cometary research begun at Arecibo by Gordon Pettengill, Brian Marsden (Harvard-Smithsonian Astrophysical Observatory), and Irwin Shapiro (who prepared the ephemerides). In late July 1976, they attempted to detect echoes from Comets d'Arrest and Grigg-Skjellerup during three observing sessions. Both attempts failed, although Comet d'Arrest came within 0.15 astronomical units of Earth.50
The first comet detected by radar was Comet Encke. As Don Campbell explained, "It was a historic first. We had never actually seen a comet before."51 French and German astronomers had observed Encke earlier; its name came from the German mathematician and physicist Johann Encke, who initially suggested an elliptical orbit with a period of 12.2 years, then correctly recalculated an elliptical orbit of 3.3 years, the shortest period of any known comet.52 Comet Encke was due back in November-December 1980. Although Encke had a relatively stable and therefore predictable orbit, optical observations were neither sufficiently numerous nor sufficiently accurate to formulate a satisfactory ephemeris for the radar. Irwin Shapiro and Antonia Forni (Lincoln Laboratory) based the radar ephemerides on optical data from both past appearances and new observations associated with the 1980 appearance supplied by Brian Marsden. The ephemeris difficulties resolved, Kamoun, Campbell, and Ostro observed Encke for 12 hours on seven consecutive days, 2-8 November 1980, about 30 days before the comet reached perihelion and at a distance of slightly more than 0.3 astronomical units from Earth. They found distinct, but very weak, echoes during each observing session.53
 Next, Kamoun attempted radar observations of the Comet Grigg-Skjellerup, which was discovered in 1902 by Grigg in New Zealand, then re-discovered as a new comet in 1922 by Skjellerup in South Africa. Grigg-Skjellerup has an orbital period of 5.1 years, making it the second shortest periodic comet after Encke. The time of perihelion passage was 15 May 1982, at a perihelion distance of nearly one astronomical unit (0.989).
Compared to other cometary experiments, Kamoun spent an unprecedented and never repeated 49 hours observing the comet between 20 May and 2 June 1982, about a week after it passed perihelion, while the comet was about 0.33 astronomical units from Earth. He received echoes in both senses of circular polarization, but technical problems prevented the acquisition of data on five days. An interesting feature was the very narrow (less than one Hz) Doppler bandwidth of the echo, which indicated either a very specular echo, a slow rotation rate, or collinearity of the polar axis with the line-of-sight.54
Comet Austin came next. Unlike Encke and Grigg-Skjellerup, Comet Austin had only been discovered on the morning of 19 June 1982 by Rodney Austin in New Zealand. Alan Gilmore, of Mount John University Observatory, New Zealand, confirmed the discovery. The comet was first reported on 21 June 1982 in IAU circular 3705 of the Central Bureau for Astronomical Telegrams by Brian Marsden, who also computed and made public a set of orbital elements showing that the comet was moving on a parabolic orbit. From the Marsden ephemeris, it appeared that Comet Austin would pass close enough to Earth to detect it with the Arecibo radar.
Following receipt of IAU circular 3706 containing the improved elements of the comet's orbit, Kamoun undertook the task of obtaining telescope time. He attempted to observe the comet on the mornings of 8-12 August 1982. Despite equipment problems that plagued observations on 8 and 9 August, the last three days yielded normal performance. On the last day, 12 August, the analyzing bandwidth was doubled from 380 to 760 hz, with a corresponding increase in the frequency resolution, in order to widen the search window. They computed an ephemeris after the experiment, using all the astrometric observations available for Comet Austin between June 1982 and November 1982. That ephemeris turned out to be substantially different from the ephemeris used during the actual radar observations. Despite correcting for this, and despite the distance from Earth being very similar to that of Comets Encke and Grigg-Skjellerup, five days of observations in August 1982 did not result in a successful detection.55
Radar detections of comets were obviously fairly difficult to make, even with the best radar telescope then available. Another opportunity to attempt a newly-discovered comet came later that year. Comet Churyumov-Gerasimenko was discovered on a photograph taken on 11 September 1969 at the Alma-Ata observatory in the Soviet Union by K. I. Churyumov and S. I. Gerasimenko. At the time of Kamoun's radar observations in November 1982, Comet Churyumov-Gerasimenko was 0.39 astronomical units from Earth. It ought to have been detectable by the Arecibo radar. Kamoun attempted Comet Churyumov-Gerasimenko for 33 hours between 7 and 16 November 1982. Serious technical problems on 7 and 16 November prevented acquisition of data. Further difficulties on 8 and 11 November caused loss of some data. In the end, the attempt on Comet Churyumov-Gerasimenko was not successful.56
From his successful and unsuccessful observations of comets, Kamoun estimated the radii of their nuclei, which were 0.4-3.6 km for Encke, 0.4-2.2 km for Grigg-Skjellerup,  less then 1.5 km for Austin, and less than 2 km for Churyumov-Gerasimenko. He also placed upper limits on the number of millimeter and centimeter-sized particles in the coma of the four comets (Table 7).57
Composition Magnetite Iron Sulfide
Grigg-Skjellerup and Austin
In setting forth a program of future cometary radar studies, Kamoun noted that the comets attempted in his dissertation could not be observed again during the next 10 years. Despite the scheduled reappearances of Encke in 1984 and 1987, of Grigg-Skjellerup in 1987, and of Churyumov-Gerasimenko in 1989, none of the comets would approach close enough for radar observation. On the other hand, he calculated, even if no improvement in radar sensitivity occurred, other comets would be accessible, particularly Comets Haneda-Campos (1984), Giacobini-Zinner (1985), Borelly and Denning-Fujikawa (1987), and Brorsen-Metcalf and Dubiago (1989).58 None of those comets, however, was ever observed by radar.
Instead, opportunities, in fact far better opportunities, came from comets never before seen. In early May 1983, as Paul Kamoun was writing his dissertation, preparations were underway at Arecibo to observe Comet IRAS-Araki-Alcock. On 25 April 1983, the Infrared Astronomical Satellite (IRAS) discovered Comet IRAS-Araki-Alcock. Initially, scientists believed it was an asteroid. In either case, it was sure to approach near the Earth. Astronomers calculated that the object would pass Earth at a distance of only 0.03 astronomical units (450,000 km), that is, about 10 times closer than any other comet that Kamoun had observed for his dissertation. In fact, such a close approach for a comet had not been known to have occurred in more than two hundred years. Although Kamoun had pioneered cometary radar, he would miss the most spectacular cometary opportunity. After writing up his thesis, he returned to France and took a position with a French aerospace firm.59
But observing Comet IRAS-Araki-Alcock was not going to be easy. Its orbit was highly inclined relative to the Earth's equator, and to make observation at Arecibo that much harder, as Don Campbell explained, "It was moving in declination so rapidly, that it actually went through the entire sky coverage of Arecibo in one day. We had a two-and-a-half-hour observing window, and that was it!"60
 The ability to get good data on Comet IRAS-Araki-Alcock depended heavily on having an accurate ephemeris. That was the job of Brian Marsden and Irwin Shapiro, who had just become Director of the Harvard-Smithsonian Astrophysical Observatory in January 1983, four months before the comet's discovery. "Taking over this place was an all-consuming job," he recalled. "I worked day and night. But for a few days, I dropped this job like a ton of bricks, literally, to develop the ephemeris needed to observe IRAS-Araki-Alcock at Arecibo and Goldstone."61
Working closely with Brian Marsden, Shapiro generated an ephemeris for the comet. "It was a big mess," Shapiro explained. "I was up until 2:30 in the morning every night. The difficulty was due to there being very few comet observations, mostly bad. We had to try numerous combinations to sort the good from the bad." Then Shapiro turned to the task of preparing an ephemeris for the radar. "The radar ephemeris was prepared at Lincoln Laboratory; the radar observations were to be made at Arecibo. It was a logistical nightmare, because of the incredible time pressure," Shapiro explained. "As the time of close approach of the comet to Arecibo neared, we sent the ephemeris electronically. It arrived an hour before the comet was to make its one and only pass over head. It worked brilliantly."62
Don Campbell took high quality data on IRAS-Araki-Alcock for about three hours during the single observation evening of 11 May when the comet was in the telescope's declination window. Campbell recalled: "We got extremely nice data. You could actually see the echo on the oscilloscope right there in the control room. It was all over the place. A nice sine wave popping in and out. It was all very exciting. We measured only spectra and obtained a lot of very interesting data on IRAS-Araki-Alcock in just that two-hour period."63
More surprising than a powerful echo from a relatively large nucleus, the spectra showed a broad low-level skirt distinct from the nucleus echo. The skirt suggested the possible existence of a cloud of unexpectedly large, centimeter-sized ejected particles from the comet. The IRAS-Araki-Alcock skirt spectrum appeared to be consistent with a model in which large grains were ejected from the nucleus by the same gas-drag mechanism used to explain the ejection of the smaller particles making up the dust coma and tail.64 "This was the first time that such particles had ever been discovered," Campbell explained. "It made the whole experiment much more interesting."65
At the same time, Dick Goldstein and Ray Jurgens, in collaboration with JPL comet specialist Zdenek Sekanina, prepared to look at IRAS-Araki-Alcock with the Goldstone radar. Previously, they had made failed attempts at Comets d'Arrest (1976), Kohoutek (1974), and Bradfield (1974).66 IRAS-Araki-Alcock would be their first successful cometary detection. Their chief obstacle was the resuscitation the Goldstone radar. As Jurgens wrote: "As luck would have it, the JPL radar system had been shut down following  the unsuccessful tracks of asteroid 4 Vesta on 28 May 1982. Since the radar system has seen only sporadic usage over the past few years, the X-band transmitter, the 20 year old computer and the data acquisition equipment were unreliable. We were in the midst of a major rebuilding project that would not be put into operation until March 1985. Fortunately, we had not removed the old equipment."67
Jurgens and a team of JPL engineers refurbished the radar equipment, while Mike Keesey prepared a radar ephemeris based on orbital elements supplied by Brian Marsden and Irwin Shapiro, who also had supplied the Arecibo ephemeris. The Goldstone observations took place on 11 and 14 May 1982 at both S-band and X-band. On a few runs, echoes were received in the same circular polarization.68 Goldstein, Jurgens, and Sekanina concluded that the nucleus of Comet IRAS-Araki-Alcock was very rough on a scale larger than the radar wavelength. They did not believe that the predominant backscattering mechanism was similar to that observed from the icy surfaces of the Galilean satellites, but instead consisted of single reflections from very rough surfaces. They posited, furthermore, that the shape of the nucleus appeared to be irregular. Jurgens believed that the nucleus's shape could be represented fairly well by a triaxial ellipsoid having equatorial radii in a ratio of two to one. The JPL radar astronomers estimated its radius to be between three and six km (larger than any comet observed by Kamoun) and its rotational period to be from one to two days.
Because of Jurgens' interest in asteroids, he and his JPL colleagues compared the comet to known asteroids. "The observed spectral shapes are typical of those measured for small Earth-crossing asteroids except for the broadband skirt," they noted. "Due to distance and sensitivity limitations, such a skirt would not have been detected on any asteroid observed so far even if it existed."69 However, they did not carry out a detailed analysis of the skirt.
Within weeks after Comet IRAS-Araki-Alcock, another new comet, Sugano-Saigusa-Fujikawa, passed the Earth. The two comets coming so closely together created a "once in a lifetime" opportunity. Comet Sugano-Saigusa-Fujikawa came within 0.06 astronomical units of Earth in early June 1983. Don Campbell attempted Sugano-Saigusa-Fujikawa on the one day it was within the Arecibo telescope's declination window, while Jurgens and Goldstein tried during four full days of observations, which delayed the renovation of the Mars Station antenna for one month. "Night after night," Jurgens wrote, "we searched the sky in the area of the comet with no indication of an echo."70 Arecibo, on the other hand, did find echoes; however, Sugano-Saigusa-Fujikawa was about three times further away than IRAS-Araki-Alcock had been, and it was a smaller comet, so that it was a less interesting and "somewhat disappointing" target.71
Despite the many unsuccessful and disappointing attempts to detect comets, until the passing of Comet Halley, only the radar observations of IRAS-Araki-Alcock made at Arecibo and Goldstone contributed to the vast amount of data collected by comet scientists at optical, radio, infrared, and ultraviolet wavelengths.72 Comet Halley returns every 76 years. Its reappearance prompted a global effort, the International Halley Watch, to coordinate ground and space observations. Unlike previous comets, Halley was investigated from a number of spacecraft sent by Japan (Suisei and Sakigake), the Soviet Union (Vega 1 and 2), and the European Space Agency (Giotto). The radar results, however, did  not play a part in the international effort.73 Radar was still a marginal tool for cometary research.
Comet Halley was to make two close approaches to Earth during its appearance in 1985-1986. At its closest approach in November 1985, it was to be 0.61 astronomical units from Earth, and during its second approach, even closer, 0.41 astronomical units, to Earth in April 1986. At the November 1985 approach, Halley would be visible at both Arecibo and Goldstone, though far below likely detectability at the latter site. Moreover, Halley was not within the Arecibo telescope's limited declination coverage during its closer approach to Earth in April 1986.74 The chances for viewing Halley thus were small; the best chance was in November and December 1985, when Halley was to be 0.62 astronomical units distant from Earth, not a good distance for observing comets.
At Arecibo, John Harmon observed Halley on 24, 28, 29 November and 1 and 2 December 1985 during its inbound Earth approach and detected a weak echo from Halley at a distance of 0.62 to 0.64 astronomical units, the most distant comet yet detected with radar. With the exception of IRAS-Araki-Alcock, comets observed earlier generally had been about 0.3 astronomical units away. A broadband feature with a high radar cross section and a large Doppler bandwidth dominated the echo spectrum, properties that were inconsistent with an echo from the nucleus. Halley, then, became the second comet to yield a radar detection of grains larger than two cm in radius ejected from the nucleus. Comet Halley also was the first radar bright comet observed; it had the largest radar cross section to date of any comet detected by radar. "If our interpretation of the echoes is correct," Don Campbell explained, "Halley is the first comet to give a stronger echo from particles than from the nucleus itself."75
The Arecibo attempt on Halley in 1985 was the last successful radar detection of a comet. In 1990, John Harmon attempted Comet Austin in cooperation with Steve Ostro, who tried to obtain echoes with the Goldstone X-band radar. Harmon also attempted Comet Honda-Mrkos-Pajddusakova in 1990, but again without success.76 These failures only served to highlight the extreme difficulty of doing radar research on comets and, as a result, the lack of major radar contributions to cometary science.
Asteroids did not make easy radar targets, either. Their small size and distance from Earth placed them at the limits of planetary radar capabilities. Also, the known population of asteroids outside the mainbelt between Mars and Jupiter, that is, the known number of asteroids that might approach Earth close enough for radar study, was far smaller than the quantity we know today. After the detection of Icarus at Haystack and Goldstone in June 1968, only six more asteroids came under radar investigation between then and July 1980: five near-Earth asteroids (1566 Icarus, 1685 Toro, 433 Eros, 1580 Betulia, and Phocaea) and two mainbelt asteroids (1 Ceres and 4 Vesta).
 Interest in asteroids was growing among astronomers during that 12-year period. Tom Gehrels, University of Arizona at Tucson, was the most vocal advocate of asteroid research. During the 1970s, he organized three asteroid conferences at Tucson which provided much of the impetus for the modern investigation of asteroids. He also initiated a program of asteroid detection called Spacewatch. Spacewatch, a survey telescope located on Kitt Peak to discover new asteroids, started operating in May 1963. Tom Gehrels also led an effort to use a modern CCD scanning camera on a specially designed telescope beginning in 1979. In its first two years, Spacewatch discovered 69 new asteroids. The rapid discovery rate of asteroids that started in the 1970s was due largely, however, to the Palomar Planet-Crossing Asteroid Survey (PCAS), begun in 1973 by Eleanor Helin and Eugene Shoemaker. The Survey initially used a 46-cm Schmidt camera to detect asteroids on the four to five nights each month around the new Moon. The exposed photographic plates were subjected to stereoscopic examination the same night they were taken, in case a new asteroid was recorded on the film. If an object were discovered, positional data was relayed by telephone to Brian Marsden at the Harvard-Smithsonian Astrophysical Center, where he headed a center for data on minor planets starting in 1978. Marsden then computed the orbit and ephemerides for further observations.
As a result of the Spacewatch and PCAS programs, the asteroid literature, as measured by citations of asteroid papers, underwent the kind of swift growth that is typical of Big Science.77 Although radar astronomers at first simply attempted to detect asteroids, both Arecibo and Goldstone investigators initiated systematic programs of asteroid detection and research in the mid-seventies. The focus was on measuring radii, surface roughness, and composition, and on improving orbits. In addition, Ray Jurgens pioneered the modeling of asteroid shapes.
Dick Goldstein, using the Goldstone Mars Station, obtained echoes from 1685 Toro, the first asteroid detected after Icarus, in 1972. After he combined the radar and optical data, Goldstein inferred that the asteroid had an irregular rocky surface slightly smoothed by a mantle of loose material.78 The following asteroid opportunity, 433 Eros, arrived in January 1975. The experiment carried out on Eros at Goldstone was, in the words of Steve Ostro, "The most important asteroid experiment before 1980," because data was taken at two frequencies (X-band and S-band) and in both senses of circular polarization. "As a result," according to Ostro, "they achieved the best characterization of an asteroid's centimeter-to-decimeter scale surface properties until the late 1980s. By then, all work was dual polarization. Jurgens and Goldstein were well ahead of their time."79
The data Goldstein and Jurgens collected indicated that the surface of Eros was much rougher than the Moon or any of the terrestrial planets. They described a surface completely covered with sharp edges, pits, subsurface holes, or embedded chunks. They also estimated the asteroid to have equatorial dimensions of 18.6 and 7.9 km.80 In order to better describe the shape of Eros and other asteroids, Ray Jurgens developed a triaxial ellipsoid model. His work represented an important first step toward modeling asteroids with radar data. Optical observations often provide the spin rate and pole, prerequisite parameters for determining the shape of an asteroid from radar data.81
Continuing his pursuit of radar asteroid research, Ray Jurgens outlined an ambitious 10-year program of asteroid opportunities in 1977. The program laid out the kinds of measurements that ground-based radars (both Goldstone and Arecibo) could make with currently available transmitter power and receiver sensitivity. Jurgens estimated that the number of detectable asteroids available for study over the following 10 years was 60. It took a few years longer to reach that number, however, for a variety of reasons. Jurgens also pointed out that astronomers could use the radar data in many cases to calculate the radius, average surface roughness, rotational rate, and polar axis direction, and in some cases the radar albedos and orbital parameters, of asteroids.82
In a memorandum to NASA Headquarters, Jurgens described the kinds of asteroid opportunities that would become available upon the upgrading of the Goldstone radar and argued for the scientific value of determining object size, rotation period, shape, and surface properties from range and Doppler measurements, in the hopes of funding asteroid radar research at JPL.83 Jurgens had foreseen and mapped out the kind of radar asteroid research program that only a few years later would materialize, but at Arecibo. Jurgens' asteroid research program did not take root at JPL; the Goldstone radar was shut down after some unsuccessful tracks on Vesta on 28 May 1982.84 In contrast, radar asteroid studies at Arecibo were far more energetic. There, Brian Marsden worked with Irwin Shapiro, the guru of the Planetary Ephemeris Program, to undertake a systematic study  of asteroid astrometry and composition.85 The first asteroid observations that formed part of that program were of Eros.
In 1975, Don Campbell and Gordon Pettengill observed Eros with the old UHF (430-MHz; 70-cm) transmitter. That was the first asteroid detected by the Arecibo telescope radar. Campbell and Pettengill measured the radar cross section of the asteroid and estimated its radius to be about 16 km. They also found the surface of Eros to be rough compared to the surfaces of the terrestrial planets and the Moon. When Pettengill and Campbell attempted to determine the composition of the surface, they could only conclude that it could not be a highly conductive metal.86
After unsuccessful attempts at asteroids Ceres and Metis, Pettengill and Marsden observed 1580 Betulia in 1976, the first asteroid target of the new S-band radar. Steve Ostro, then a graduate student at MIT, did the analysis. He measured the asteroid's average radar cross section and set a lower limit to the asteroid's radius of 2.9 ± 0.2 km.87
Pettengill and Ostro next turned their attention to the mainbelt asteroid Ceres. Already, in December 1975, Pettengill and Marsden, in collaboration with Goldstein (JPL) and Tom Gehrels and Benjamin Zellner (University of Arizona) had failed to obtain echoes from both Ceres and Metis, another mainbelt asteroid. The lack of an echo from Metis was not surprising, but Ceres should have been easy to detect; they interpreted the absence of an echo as indicating a smaller cross section than they had expected.88
Ostro, Pettengill, and Campbell finally detected Ceres with the Arecibo S-band radar in March and April 1977. This was the first mainbelt asteroid detected by radar. The greater sensitivity of the S-band instrument made it possible for the radar to reach into the mainbelt of asteroids and detect such a small body. The opportunity of March 1977 was slightly more favorable than that of 1975, thanks to the installation of a more sensitive line feed in 1976. Ceres was found to have a low radar cross section, less than that for the Moon, the terrestrial planets, and even Eros. On the other hand, the asteroid appeared to have a very rough surface at some scale comparable to, or larger than, the 12.6-cm wavelength of the radar, that is, rougher than the Moon and terrestrial planets, but smoother than the Galilean satellites of Jupiter.89
Noisy data taken on mainbelt asteroid Vesta during three nights of observations in November 1979 returned only a weak detection.90 Each asteroid detection seemed to bring a new revelation; no pattern emerged. Unlike the terrestrial planets, asteroids presented not a few bodies to study but an entire population, a population, moreover, that the growing discovery rate kept increasing. Although the systems of Jupiter and Saturn defined the outer limits of planetary radar astronomy after 1975, the asteroids defined its future. They were on their way to deposing Venus from its position as the favored target of radar astronomers.
1. Referred to in Campbell, Chandler, Pettengill, and Shapiro, "Galilean Satellites of Jupiter: 12.6-Centimeter Radar Observations," Science 196 (1977): 650.
2. Shapiro 1/10/93; "Funding Proposal, 'Plan for NEROC Operation of the Haystack Research Facility as a National Radio/Radar Observatory,' NSF, 7/1/71-6/30/73," 26/2/AC 135, and Sebring to Hurlburt, 27 March 1970, 18/2/AC 135, MITA; NEROC, Proposal to the National Science Foundation for Programs in Radio and Radar Astronomy at the Haystack Observatory, 8 May 1970, pp. III.8-III.10, LLLA; JPL 1970 Annual Report, p. 14, JPLA.
3. See Joseph Veverka, "Polarization Measurements of the Galilean Satellites of Jupiter," Icarus 14 (1971): 355-359; John S. Lewis, "Low Temperature Condensation from the Solar Nebula," Icarus 16 (1972): 241-252. Although Io, Ganymede, and Europa were believed covered with frost, Callisto was believed to be different, more like Moon, though with some frost possibly present.
4. Goldstein and Morris, "Ganymede: Observations by Radar," Science 188 (1975): 1211-1212.
5. Campbell 8/12/93; NAIC QR Q3/1975, 4-5; NAIC QR Q4/1975, 5; NAIC QR Q1/1976, 6.
6. Campbell, Chandler, Pettengill, and Shapiro, "Galilean Satellites of Jupiter: 12.6-Centimeter Radar Observations," Science 196 (1977): 650-653; Campbell, Chandler, Steven J. Ostro, Pettengill, and Shapiro, "Galilean Satellites: 1976 Radar Results," Icarus 34 (1978): 254-267; NAIC QR Q1/1976, 17; Ostro, "Radar Properties of Europa, Ganymede, and Callisto," in David Morrison, ed., Satellites of Jupiter (Tucson: University of Arizona Press, 1982), p. 213.
7. Campbell 8/12/93.
8. Ostro 18/5/94; NAIC QR Q1/1976, 6.
9. Ostro 18/5/94.
10. Campbell, Chandler, Ostro, Pettengill, and Shapiro, "Galilean Satellites: 1976 Radar Results," Icarus 34 (1978): 254-267; Ostro, "The Structure of Saturn's Rings and the Surfaces of the Galilean Satellites as Inferred from Radar Observations," Ph.D. dissertation, MIT, 1978; NAIC QR Q4/1977, 5-6; NAIC QR Q1/1978, 6.
11. Ostro, Campbell, Pettengill, and Shapiro, "Radar Observations of Europa, Ganymede, and Callisto," Icarus 44 (1980): 431-440; NAIC QR Q1/1979, 10; NAIC QR Q2/1980, 11.
12. Ostro and Pettengill, "Icy Craters on the Galilean Satellites?" Icarus 34 (1978): 268-279.
13. Campbell 8/12/93.
14. Goldstein and R. Green, "Ganymede: Radar Surface Characteristics," Science 207 (1980): 179-180; Ostro, "Radar Properties of Europa, Ganymede, and Callisto," in Morrison, Satellites of Jupiter, pp. 225-233, and quote p. 235.
15. Morrison and Jane Samz, Voyage to Jupiter, NASA SP-439 (Washington: NASA, 1980), pp. 58, 60 & 142.
16. Hagfors, Gold, and M. Ierkic, "Refraction Scattering as Origins of the Anomalous Radar Returns of Jupiter's Satellites," Nature 315 (1985): 637-640.
17. Eshleman, "Mode Decoupling during Retrorefraction as an Explanation for Bizarre Radar Echoes from Icy Moons," Nature 319 (1986): 755-757; Eshleman, "Radar Glory from Buried Craters on Icy Moons," Science 234 (1986): 587-590.
18. Shoemaker 30/6/94; Ostro and Eugene M. Shoemaker, "The Extraordinary Radar Echoes from Europa, Ganymede, and Callisto: A Geological Perspective," Icarus 85 (1990): 335-345.
19. E-mail, Simpson to author, 9 November 1994; NAIC QR Q2/1987, 7; Q3/1987, 8-9; Q4/1987, 9; Q2/1988, 9; Q4/1988, 8; Q4/1989, 7; Q1/1990, 7; Q1/1991, 7; Q1/1992, 8.
20. Campbell, Chandler, Ostro, Pettengill, and Shapiro, "Galilean Satellites: 1976 Radar Results," Icarus 34 (1978): 254-267; NAIC QR Q1/1976, 6; Q4/1977, 5-6; Q2/1987, 7; Q3/1987, 8-9; Q4/1987, 9.
21. NAIC QR Q4/1989, 7; Q1/1990, 7.
22. Ostro, Campbell, Simpson, R. Scott Hudson, Chandler, Keith D. Rosema, Shapiro, Standish, R. Winkler, Donald K. Yeoman, Ray Vélez, and Goldstein, "Europa, Ganymede, and Callisto: New Radar Results from Arecibo and Goldstone," Journal of Geophysical Research 97 (1992): 18,227-18,244. The Goldstone observations were made 10-11, 13, 15-16, 22, 26, & 29-30 November 1988; 5 & 8 December 1988; 13, 14, 15, 18, 19, 20, 22, 24, 27, & 29 December 1989; 13, 18, 22 & 27 December 1990.
23. Ostro 18/5/94; Bruce Hapke, "Coherent Backscatter and the Radar Characteristics of Outer Planet Satellites," Icarus 88 (1990): 407-417; Hapke and David Blewett, "Coherent Backscatter Model for the Unusual Radar Reflectivity of Icy Satellites," Nature 352 (1991) 46-47; Sajeev John, "Localization of Light," Physics Today 44 (May 1991): 32-40.
24. Kenneth J. Peters, "Coherent-Backscatter Effect: A Vector Formulation Accounting for Polarization and Absorption Effects and Small or Large Scatterers," Physical Review B 46 (1992): 801-812; John, "Localization of Light," Physics Today 44 (May 1991): 32-40; Ostro 18/5/1994.
25. Ostro 18/5/94; Ostro and Shoemaker, "The Extraordinary Radar Echoes from Europa, Ganymede, and Callisto: A Geological Perspective," Icarus 85 (1990): 335-345.
26. Eric J. Rignot, Ostro, Jakob J. Van Zyl, and K. C. Jezek, "Unusual Radar Echoes from the Greenland Ice Sheet," Science 261 (24 September 1993): 1710-1711.
27. NAIC QR Q1/1976, 7; Q4/1977, 5-6; Q3/1987, 8-9; Q2/1988, 9; Q1/1992, 8; Campbell, Chandler, Pettengill, and Shapiro, "Galilean Satellites of Jupiter: 12.6-Centimeter Radar Observations," Science 196 (1977): 651; Ostro, Campbell, Simpson, Hudson, Chandler, Rosema, Shapiro, Standish, Winkler, Yeoman, Vélez, and Goldstein, "Europa, Ganymede, and Callisto: New Radar Results from Arecibo and Goldstone," Journal of Geophysical Research 97 (1992): 18,227-18,244.
28. Ostro, Campbell, Simpson, Hudson, Chandler, Rosema, Shapiro, Standish, Winkler, Yeoman, Vélez, and Goldstein, "Europa, Ganymede, and Callisto: New Radar Results from Arecibo and Goldstone," Journal of Geophysical Research 97 (1992): 18,227-18,244; NAIC QR Q1/1991, 7.
29. Ostro, Pettengill, Campbell, Goldstein, Icarus 49 (1982): 367.
30. NEROC, Final Progress Report Radar Studies of the Planets, 29 August 1974, pp. 1, 3, 6 & 8-9; Log Book, Haystack Planetary Radar, HR-73-1, 27 June 1973 to 26 November 1973, SEBRING; and Goldstein, R. Green, Pettengill, and Campbell, "The Rings of Saturn: Two-Frequency Radar Observations," Icarus 30 (1977): 105.
31. Goldstein and Morris, "Radar Observations of the Rings of Saturn," Icarus 20 (1973): 260-262; Morris, "Distribution and Size of Elements of Saturn's Rings as Inferred from 12-cm Radar Observations," in Frank Don Palluconi and Pettengill, eds., The Rings of Saturn, SP-343 (Washington: NASA, 1974), p. 73.
32. Campbell 8/12/93. See, for example, Allan F. Cook, Fred A. Franklin, and F. D. Palluconi, "Saturn's Rings: A Survey," Icarus 19 (1973): 317-337 and Pollack, "The Rings of Saturn," American Scientist 66 (1978): 30-37.
33. Rasool, "Foreword," in Palluconi and Pettengill, pp. v-vi; ibid., pp. 192-195; and Morris, "Distribution and Size of Elements of Saturn's Rings as Inferred from 12-cm Radar Observations," pp. 73-82. Interestingly, when a subsequent workshop on Saturn's rings was held at the Reston International Conference Center, Reston, Virginia, 9-11 February 1978, and sponsored by the NASA Office of Space Science, no radar presentations were made. The purpose of the workshop was more tightly defined than the 1973 workshop; the 1978 workshop strictly prepared for the Voyager mission.
34. Pettengill and Hagfors, "Comment on Radar Scattering from Saturn's Rings," Icarus 21 (1974): 188-190, esp. 188.
35. Jeffrey N. Cuzzi and David Van Blerkom, "Microwave Brightness of Saturn's Rings," Icarus 22 (1974): 149-158; Pollack, A. L. Summers, and B. Baldwin, "Estimates of the Size of the Particles in the Rings of Saturn and their Cosmogonic Implications," Icarus 20 (1973): 263-279; Morrison and D. P. Cruikshank, "Physical Properties of the Natural Satellites," Space Science Review 15 (1974): 722-732; Pollack, "The Rings of Saturn," Space Science Review 18 (1975): 3-97.
36. Campbell 7/12/93.
37. NAIC QR Q1/1975, 4.
38. Goldstein, R. Green, Pettengill, and Campbell, "The Rings of Saturn: Two-Frequency Radar Observations," Icarus 30 (1977): 104-110.
39. L. W. Esposito, Cuzzi, J. B. Holberg, E. A. Marouf, Tyler, and C. C. Porco, "Saturn's Rings: Structure, Dynamics, and Particle Properties," in Tom Gehrels and Mildred Shapley Matthews, eds., Saturn (Tucson: University of Arizona Press, 1984), p. 466.
40. Allan F. Cook and Fred A. Franklin, "Saturn's Rings: A New Survey," in Joseph A. Burns, ed., Planetary Satellites (Tucson: University of Arizona Press, 1977), pp. 412-419. See also Cuzzi and Pollack, "Saturn's Rings: Particle Composition and Size Distribution as Constrained by Microwave Observations." Icarus 33 (1978): 233-262.
41. Campbell 8/12/93; Esposito, Cuzzi, Holberg, Marouf, Tyler, and Porco, "Saturn's Rings: Structure, Dynamics, and Particle Properties," in Gehrels and Matthews, Saturn, p. 467; NAIC QR Q1/1978, 7; NAIC QR Q1/1979, 9; Ostro, "The Structure of Saturn's Rings," pp. 105-157.
42. NAIC QR Q1/1977, 7.
43. Cuzzi and Pollack, "Saturn's Rings: Particle Composition and Size Distribution as Constrained by Microwave Observations." Icarus 33 (1978): 233-262.
44. Ostro, Pettengill, and Campbell, "Radar Observations of Saturn's Rings at Intermediate Tilt Angles," Icarus 41 (1980): 381-388.
45. Ostro, Pettengill, Campbell, and Goldstein, "Delay-Doppler Radar Observations of Saturn's Rings," Icarus 49 (1982): 367-381. See also Ostro and Pettengill, "A Review of Radar Observations of Saturn's Rings," in A. Brahic, ed., Planetary Rings 1982 (Toulouse: CEPADUES Éditions, 1982), pp. 49-55.
Later radar data collected at Goldstone by Goldstein and Jurgens and at Arecibo by Ostro, Pettengill, and Campbell in 1981, when the rings were at a 6° tilt angle, confirmed that the ring particles were large, irregular, and jagged in shape and made of ice; the researchers finally abandoned the notion that they might be metallic. Moreover, they affirmed the conclusion that the A and B rings reflected most, if not all, of the radar echo from Saturn's rings. Goldstein and Jurgens, "Radar Observations of the Rings of Saturn," Journal of Geophysical Research submitted for publication; Ostro, Pettengill, Campbell, and Goldstein, "Delay-Doppler Radar Observations of Saturn's Rings," Icarus 49 (1982): 367-381; Ostro, Pettengill, and Campbell, "Radar Observations of Saturn's Rings at Intermediate Tilt Angles," Icarus 41 (1980): 381-388. This research is summarized in: Ostro and Pettengill, "A Review of Radar Observations of Saturn's Rings," pp. 49-55.
46. Whipple, "Comets," in J. A. M. McDonnell, ed., Cosmic Dust (New York: John Wiley & Sons, 1978), pp. 1-73.
47. Log book, Haystack Planetary Radar, HR-73-2, 9 December 1970 to 11 August 1971, SEBRING; Shapiro 1/10/93; Eric J. Chaisson, Ingalls, Rogers, and Shapiro, "Upper Limit on the Radar Cross Section of the Comet Kohoutek," Icarus 24 (1975): 188-189.
48. Paul Gaston David Kamoun, "Radar Observations of Cometary Nuclei," Ph.D. diss., MIT, May 1983.
49. Whipple, "A Comet Model. I. The Acceleration of Comet Encke," Astrophysical Journal 111 (1950): 375-394; Whipple, "A Comet Model. II. Physical Relations for Comets and Meteors," ibid., 113 (1951): 464-474.
50. Kamoun, p. 31; NAIC QR Q3/1976, 6-7.
51. Campbell 9/12/93.
52. John E. Bortle, "Comet Digest," Sky and Telescope 60 (1980): 290; Kamoun, pp. 37-38.
53. Kamoun, p. 51; Kamoun, Campbell, Ostro, Pettengill, and Shapiro, "Comet Encke: Radar Detection of Nucleus," Science 216 (1982): 293-295; NAIC QR Q4/1980, 8-9.
54. Kamoun, pp. 90 & 85; NAIC QR Q2/1982, 7-8.
55. Kamoun, pp. 108-110; NAIC QR Q3/1982, 8.
56. Kamoun, pp. 21 & 122; NAIC QR Q4/1982, 7-8; K. I. Churyumov and S. I. Gerasimenko, "Physical Observations of the Short-Period Comet 1969 IV," in G. A. Chebotarev, E. I. Kazimirchak-Polonskaya, and Brian G. Marsden, eds., The Motion, Evolution of Orbits, and Origin of Comets IAU Symposium 45 (New York: Springer-Verlag, 1972), pp. 27-34. Both Churyumov and Gerasimenko were in the Department of Astronomy, University of Kiev.
57. Kamoun, p. 230.
58. Kamoun, p. 237.
59. Campbell 9/12/93; Kamoun, p. 238; Jurgens, "Seeing Comet IRAS," p. 221; information supplied by Pettengill.
60. Campbell 9/12/93.
61. Shapiro 1/10/93.
62. Shapiro 1/10/93.
63. Campbell 9/12/93; Harmon, Campbell, Hine, Shapiro, and Marsden, "Radar Observations of Comet IRAS-Araki-Alcock 1983d," The Astrophysical Journal 338 (1989): 1071; Harmon, Campbell, Hine, Shapiro, and Marsden, Radar Observations of Comet IRAS-Araki-Alcock (1983d) Report 245 (Ithaca: NAIC, September 1988), Pettengill materials.
64. Campbell 9/12/93; NAIC QR Q2/1983, 7; Harmon, Campbell, Hine, Shapiro, and Marsden, "Radar Observations of Comet IRAS-Araki-Alcock 1983d," The Astrophysical Journal 338 (1989): 1071-1093; Campbell, Harmon, Hine, Shapiro, Marsden, and Pettengill, "Arecibo Radar observations of Comets IRAS-Araki-Alcock and Sugano-Saigusa-Fujikawa," Bulletin of the American Astronomical Society 15 (1983): 800; Goldstein, Jurgens, and Zdenek Sekanina, "A Radar Study of Comet IRAS-Araki-Alcock 1983d," The Astronomical Journal 89 (1984): 1745-1754; and Shapiro, Marsden, Whipple, Campbell, Harmon, and Hine, "Interpretations of Radar Observations of Comets," Bulletin of the American Astronomical Society 15 (1983): 800.
65. Campbell 9/12/93.
66. Jurgens, "Seeing Comet IRAS," p. 221; Goldstein, Jurgens, and Sekanina, pp. 1745-1754.
67. Jurgens, "Seeing Comet IRAS," p. 222.
68. Jurgens, "Seeing Comet IRAS," p. 222; Goldstein, Jurgens, and Sekanina, pp. 1745-1747.
69. Goldstein, Jurgens, and Sekanina, p. 1754.
70. Jurgens, "Seeing Comet IRAS," p. 224.
71. Campbell 9/12/93; NAIC QR Q2/1983, 7.
72. Sekanina, "Nucleus of Comet IRAS-Araki-Alcock (1983 VII)," The Astronomical Journal 95 (1988): 1876-1894.
73. E. Grün, ed., "Halley and Giacobini-Zinner," Advances in Space Research vol. 5, no. 12 (1985): 1-344; J. W. Mason, ed., Comet Halley: Investigations, Results, Interpretations, 2 vols. (New York: Ellis Horwood, 1990); R. Reinhard and B. Battrick, eds., The Giotto Mission: Its Scientific Investigations (Noordwijk: ESTEC, European Space Agency, 1986); M. Grewing, F. Praderie, and R. Reinhard, eds., Exploration of Halley's Comet (New York: Springer-Verlag, 1986).
74. Kamoun, pp. 239-240; Campbell, Harmon, and Shapiro, "Radar Observations of Comet Halley," The Astrophysical Journal 338 (1989): 1094-1105; Campbell, Harmon, and Shapiro, Radar Observations of Comet Halley Report 246 (Ithaca: NAIC, September 1988), Pettengill materials.
75. Campbell 9/12/93; Campbell, Harmon, and Shapiro, "Comet Halley," pp. 1094 & 1103; NAIC QR Q4/1985, 8.
76. NAIC QR Q2/1990, 6.
77. Clifford J. Cunningham, Introduction to Asteroids: The Next Frontier (Richmond: Willmann-Bell, 1988), pp. 2 & 97-101; Tom Gehrels, "The Asteroids: History, Surveys, Techniques, and Future Work," in Gehrels and Matthews, eds., Asteroids (Tucson: University of Arizona Press, 1979), pp. 4-5 & 13-14.
78. Goldstein, D. B. Holdridge, and J. H. Lieske, "Minor Planets and Related Objects: 12. Radar Observations of (1685) Toro," The Astronomical Journal 78 (1973): 508-509.
79. Ostro 25/5/94.
80. Goldstein and Jurgens, "Radar Observations at 3.5 and 12.6 cm Wavelength of Asteroid 433 Eros," Icarus 28 (1976): 1-15.
81. Jurgens, "Radar Backscattering from a Rough Rotating Triaxial Ellipsoid with Applications to the Geodesy of Small Asteroids," Icarus 49 (1982): 97-108.
82. Jurgens and D. F. Bender, "Radar Detectability of Asteroids: A Survey of Opportunities for 1977 through 1987," Icarus 31 (1977): 483-497. The asteroid research program grew out of a larger work Jurgens did while at JPL, namely, Jurgens, A Survey of Ground-based Radar Astronomical Capability Employing 64 and 128 Meter Diameter Antenna Systems at S and X Band, Report 890-44 (Pasadena: JPL, March 1977). See also Pettengill and Jurgens, "Radar Observations of Asteroids," in Gehrels and Matthews, Asteroids, pp. 206-211.
83. Jurgens to Geoffrey A. Briggs, 12 August 1982, Jurgens materials.
84. Jurgens, "Seeing Comet IRAS," p. 222.
85. NAIC QR Q2/1976, 6-7.
86. Campbell, Pettengill, and Shapiro, "70-cm Radar Observations of 433 Eros," Icarus 28 (1976): 17-20; NAIC QR Q1/1975, 4.
87. Pettengill, Ostro, Shapiro, Marsden, and Campbell, "Radar Observations of Asteroid 1580 Betulia," Icarus 40 (1979): 350-354.
88. NAIC QR Q4/1975, 5.
89. Ostro, Pettengill, Shapiro, Campbell, and R. Green, "Radar Observations of Asteroid 1 Ceres," Icarus 40 (1979): 355-358; NAIC QR Q1/1977, 6-7.
90. Ostro, Campbell, Pettengill, and Shapiro, "Radar Detection of Vesta," Icarus 43 (1980): 169-171; Ostro 25/5/94; NAIC QR Q4/1979, 7.