A Brief History of Comets II (1950 -1993)
The following text is adapted from a major review on Comets , prepared by Michel C. Festou (Observatoire Midi-Pyrenees, Toulouse, France), Hans Rickman (Astronomiska Observatoriet, Uppsala, Sweden) and Richard M. West (European Southern Observatory, Garching, Germany) and published in the review journal Astronomy & Astrophysics Reviews (A&AR) (Part I, Vol. 4 pp. 363-447, 1993)
This is the second part of this brief historical review, covering the period from 1950 to 1993, i.e. until just before the crucial years 1994-1997 that saw the impact of Comet Shoemaker-Levy 9 on Jupiter (1994), the apparition of Comet Hyakutake that passed only 15 million km from the Earth (1996), as well as the bright Comet Hale-Bopp that was discovered in 1995 and put on a marvellous display when it passed perihelion in early 1997. It includes some references to major papers in this period (by author of year of publication), but the original version of this review in Astronomy & Astrophysics Reviews must be consulted for the full details about these.
The history of cometary astronomy is naturally divided into five major periods, the transitions being marked by important new insights. Before 1600, comets were essentially considered to be heavenly omens and were not yet clearly established as celestial (astronomical), rather than meteorological phenomena in the terrestrial atmosphere. Then followed two centuries of mostly positional measurements with emphasis on the motions and the orbits, lasting until the early 19th century, when the era of cometary physics was inaugurated, in particular by the passage of P/Halley in 1835. The next major step forward occurred in 1950 with the sudden emergence of the modern picture of comets as being essentially very old solar system objects made of primordial ice and dust, generally in unstable orbits and intensively interacting with the solar electromagnetic and corpuscular radiation. Finally, the space missions to P/Giacobini-Zinner in 1985 and especially to P/Halley in 1986 provided the first in situ observations of comets and dramatically widened our scientific horizon, but also posed many new questions which are yet to be answered.
1952 - 1984: The modern era
Following the break-throughs in 1950-51, the entire concept about comets had to be revised. This process was a gradual one, as new observational facts were collected, and also because these observations were becoming increasingly quantitative, allowing a progressively more detailed verification of the new ideas. Although number density estimates for cometary comae had been derived since the time of Wurm's investigations in the 1930's, the figures obtained were rather uncertain and their reliability was limited by the lack of quantitative studies about the excitation mechanisms of the light. Thus it is not too surprising that, continuing the earlier investigations by Swings and McKellar, most spectroscopic studies between 1950 and 1970 were devoted to a never-ending attempt at discovering and identifying new emission lines and bands, as well as at unraveling the structure of the rotational and vibrational bands of the comet radicals and ions. A special reference must here be made to the numerous and important contributions from the Liege school , reviews of which are given by Swings (1956) and Arpigny (1965). During this epoch rather complete models were made of the fluorescence of the CN, CH, OH, and C 2 radicals. The advent of high resolution spectroscopy in the late 1950's allowed the identification of many unknown lines, most of which were due to C 2 and NH 2. It is worth mentioning here that this effort has never been carried through to completion and many observed cometary spectral lines have still not been assigned an emitter; the most likely are CO + , CO 2 + and C 3 in the near-UV, C 2 and NH 2 in the optical and NH 2 and H 2 O + in the IR.
The following Sections, divided according to the main areas of investigative thrust during this period, illustrate how cometary research over the most recent decades has vindicated the ideas put forward in 1950-51. The ultraviolet, infrared and radio windows were explored in the early 1970's, the emissions of H I, O I and OH were observed and the dissociation products of the main volatile constituent of the nucleus were finally detected observationally. The first radar detection of a comet in 1980 (P/Encke, Kamoun et al. 1982), and the first recording of an image of a comet nucleus in 1986 convinced the last sceptics that Whipple was correct. From refined studies of the orbital motion of comets it was demonstrated that Oort's distant reservoir was fully justified, even though some shortcomings of the theory are only now being overcome and have led to new and exciting developments. Above all, however, a wealth of quantitative data became available, making truly comparative studies of comets possible. In a not too distant future this should enable us to learn whether the differences we observe between individual comets are the results of evolutionary processes or rather reflect intrinsic diversity.
Water as the main constituent of comets
In 1958, high resolution spectroscopy allowed the separation of the terrestrial oxygen lines from the cometary ones and also led to the definitive confirmation of the presence of the isotopic lines of 13 C, long suspected to be present in comets. The detection of the [O I] red lines in comet Mrkos (1957 V) (Swings and Greenstein 1958) created a completely new problem: it was soon shown by Wurm (1963) that if fluorescence is at the origin of the emission, then very large amounts of oxygen are implied, much larger than those of for instance C 2. It seemed preferable to assume another emission mechanism and Wurm proposed corpuscular excitation. The idea that some coma species may be produced directly into an excited state can be traced back to McKellar (1943), but this suggestion was not explored in detail until 1964 (Biermann and Trefftz 1964). Their work led to the prediction that photodissociation of parent molecules is the main production mechanism and that not only oxygen but also hydrogen atoms must exist in large amounts in comets, with resulting production rates of e.g. log Q (mol s -1 ) ~ 30 for a bright comet, or much larger than those of the parents of CO + , CN or C 2. In some sense, the discovery in 1970 by the Orbiting Astronomical Observatory (OAO-2) and the Orbiting Geophysical Observatory (OGO-5) of huge Lyman-alpha haloes of neutral hydrogen (> 1.5 x 10 7 km) around comets Tago-Sato-Kosaka (1969 IX) and Bennett (1970 II) did not come as a complete surprise. However, the origin of these hydrogen atoms was not yet known with certainty.
While the OH emission band at 3090 AA was first identified in comet Cunningham (1941 I) by Swings, the first quantitative OH abundance measurements only date from the early 1970's (Code et al. 1972; Blamont and Festou 1974; Keller and Lillie 1974). The analysis of the Lyman-alpha isophotes of comet Bennett (1970 II) revealed that the velocity of the H-atoms was about 8 km s -1 (Bertaux et al. 1973). Following an investigation of the photolysis of water molecules by sunlight, this led these authors to speculate about the possibility that the majority of the observed H-atoms were coming from the dissociation of OH radicals. To prove this assertion, Blamont and Festou (1974) measured both the then unknown scalelength of OH and the production rate of that radical in comet Kohoutek (1973 XII). They proposed for the first time, on a quantitative basis, that water was the parent of most of the hydrogen atoms and the OH radicals. Horst Uwe Keller and co-workers reached similar conclusions in a series of independent papers: Keller (1971) discussed the possibility that the observed H-atoms in comet Bennett might arise from the direct dissociation of water and later (Keller 1973a, 1973b) developed these ideas further. His investigation though, as well as that of Bertaux et al. (1973), was limited by the fact that the parameters governing the water photolysis were not well known at that time. Keller and Lillie (1974) also measured the scalelength of OH (in comet Bennett) and found a value in complete agreement with that found for comet Kohoutek. The definitive clue that H 2 O was the main source of both the H-atoms and the OH radicals came when the velocity of the H-atoms was measured directly from COPERNICUS observations (Drake et al. 1976) and, indirectly, from the analysis of H I Lyman-alpha observations (cf. the review by Keller 1976), and was found to be compatible with the water photolysis scheme. This was confirmed by the direct observation of water in P/Halley in 1986.
After the discovery of the 18 cm maser emission of OH (Biraud et al. 1974; Turner 1974), OH radio observations became routine and the evidence for the ubiquitous presence of water in comets was overwhelming. The emission of an unknown ion was observed in comet Kohoutek (1973 XII) by Herbig (1973) and Benvenuti and Wurm (1974). Herzberg and Lew (1974) had just obtained the first laboratory spectra of the H 2 O + ion and tentatively identified this ion as the source of the new cometary emission. The same emission was later found in cometary spectra recorded as early as 1942 (Swings et al. 1943). Although the water ion might be an abundant species in comet tails, its presence there is not conspicuous: this is a clear indication that the ion is lost rapidly, unlike the other tail ions. Aikin (1974) showed that the main loss mechanism is a charge exchange reaction with water molecules leading to the formation of the H 3 O + ion, and this latter is likely to be destroyed in electron recombination reactions. H 3 O + was indeed found to be one of the main ions in the comae of P/Giacobini-Zinner and P/Halley.
Quantitative studies and comparative cometology
Many parameters for the OH radical were derived from radio observations at 18 cm. The detailed mechanism by which comets emit photons at that wavelength was investigated by Despois et al. (1981). The methodology for determining OH velocity profiles was worked out by Bockeee-Morvan and Gerard (1984). An overview of the production rate and velocity determinations is given by Bockelee-Morvan et al. (1990). Beginning with comet Bradfield (1979 X), a long series of high quality observations of the UV spectrum of comets was obtained by theInternational Ultraviolet Explorer (IUE) , from which a self-consistent set of water production rates was derived (e.g. Festou and Feldman 1987). The radio and UV determinations of these rates do not agree perfectly, because the models used in the interpretation of the data differ markedly. Schloerb (1988) and Gerard (1990) have discussed this problem.
The 1970's saw the development of quantitative observations of comet emissions, mainly by means of narrow-band photoelectric filter photometry through diaphragms encompassing a more or less large part of the coma. A review of the early observations and the observing techniques is given by A'Hearn (1983). One of the shortcomings of standard photoelectric photometry is the contamination by an underlying continuum and by gaseous emissions in the wings of the spectral transmission curve defined by the filter. It was therefore not surprizing that spectrophotometry developed rapidly in the early 1980's when linear detectors and image intensifier tubes became available; see the review by A'Hearn (1982). This method provided both a good separation of the band or line emissions and also spatially well-resolved information about the distribution of coma species. In parallel, numerous theoretical studies, aimed at calculating the fluorescence efficiencies of the coma radicals and ions, resulted in the establishment of reliable conversions of observed surface brightness into column densities of the different species. The last step in the data analysis process is then the derivation of gas production rates.
The data accumulated during the last 20 years or so by many dedicated observers, using both ground- and space-based instruments, have made possible the comparison of the relative abundances in comae of different comets. As direct sources of detailed information and for additional references on this subject as well as the radio OH measurements quoted above, we refer the reader in particular to the papers by A'Hearn and Millis (1980), Cochran (1987), Newburn and Spinrad (1989) and Osip et al. (1992). A list of all individual observations made with the IUE until late 1989 and a discussion of the resulting comparative cometology have been published by Festou (1990). The most striking observational fact is that, at first sight, all comets look alike (Cochran 1989). There are just a few well-known objects for which the chemical composition of the coma departs notably from that of an average comet, e.g. a few CO + rich comets or those that seem to contain only one or a few of the actual compounds of comet comae. For instance, P/Giacobini-Zinner is C 2 and C 3 depleted (Cochran 1989), while comet Yanaka (1988 XXIV) seems to be made almost exclusively of NH 2 and water (Fink 1992). As suggested by the direct inspection of optical (Swings 1948) and UV spectra (Festou 1990b), the main difference between individual comets is the continuum to gas emission ratio. Observations of P/Halley in 1986 added an interesting piece to the puzzle: CO and some other observed gases require an extended source in the coma. A key issue is now to determine the relationship between this source and the dust particles. The general picture beginning to emerge is that all comets basically have similar molecular abundances and that the observed differences might only reflect a variable dust to gas production ratio. It remains to be determined whether this ratio is an intrinsic property or the result of an evolutionary (i.e. ageing) process.
From the point of view of cometary dynamics, the modern era is first of all distinguished by the advent and development of efficient and powerful computers. This allowed, for the first time, extensive numerical simulations of the orbital evolution resulting from repeated close encounters with Jupiter and other planets. It also revolutionized the work on orbit determination and linkage of past apparitions for observed comets as well as the preparation of ephemerides for upcoming apparitions, even for long-lost comets.
Whereas Oort had been working on a small sample of comets to build his theory, Marsden et al. (1978) improved the earlier statistics by using 200 well-determined long-period orbits. They found a concentration of inverse semimajor axes corresponding to an average aphelion distance of about 45,000 AU, only about half as remote as Oort's original distance. A major problem remained the apparent overabundance of Jupiter family comets. Edgar Everhart (1972) found a possible route of direct transfer from the Oort cloud via Jovian perturbations at repeated encounters with the planet, beginning with a special type of initial orbits with perihelia near Jupiter's orbit and low inclinations. However, some authors questioned the efficiency of this transfer or the fit of the orbital distribution of the captured comets. An alternative scenario came from orbital integrations of the observed comets by Elena I. Kazimirchak-Polonskaya (1972): the comets might not be captured by Jupiter alone, but rather by a stepwise process involving all the giant planets.
The ideas about the long-term dynamics of the Oort cloud evolved considerably. While passages of individual stars were mostly considered in earlier investigations, the tidal effects of the Galaxy as a whole, preliminarily modelled by Chebotarev (1965), have become recognized in recent years as the prime mechanism to provide new comets from the outer cloud. The dramatic effects that might follow upon close encounters with massive perturbers, such as giant molecular clouds (Biermann and Luest 1978), also received a great deal of attention. In particular, the question of the stability of the outer, classical regions of the Oort cloud over the age of the solar system has been debated.
A major step forward taken during this period dealt with the modelling of nongravitational effects in cometary motions. Based on Whipple's concepts, Brian Marsden (1969) introduced a nongravitational force into the Newtonian equations of motion with simple expressions for the radial and transverse components in the orbital plane. These involved a function of the heliocentric distance r expressing a standard `force law', multiplied by a coefficient whose value was determined along with the osculating orbital elements by minimizing the residuals of the fit to positional observations. The radial coefficient was called A 1 and the transverse A 2 . It was realized that the model might not be physically realistic and that more meaningful parameters might be derived from a generalized formalism, but attempts in this direction were not successful (Marsden 1970). The final update of the model was made in 1973 (Marsden et al. 1973), stimulated by calculations of the H 2 O sublimation rate as a function of r (Delsemme and Miller 1971). This was taken as the model force law, expressed as an algebraic function g(r) whose parameters were chosen to fit Delsemme and Miller's results. Eventually more realistic models were constructed for the jet force as resulting from asymmetric H 2 O outgassing, including the heat flow in the surface layers of the nucleus (Rickman and Froeschle 1983). As a result it was found that the true force law might be very different from the g(r) formula, and hence there should be room for an improved model.
The long-term variations of the nongravitational forces were found to involve a wide range of behaviour. Thus the well-determined A 2 -values found over different periods of time for the same comet usually vary in a more or less regular fashion, often including changes of sign. This was generally interpreted in terms of spin axis precession, which in turn may be caused by the torque associated with the jet force of outgassing. An early suggestion of such a scenario was made for P/Kopff (Yeomans 1974). Quantitative models were first derived by Whipple and Zdenek Sekanina (1979) to fit the secular decrease of the nongravitational perihelion shift of P/Encke. These models, and similar ones developed later on for a number of other comets (Sekanina 1984-85; Sekanina and Yeomans 1985), led to predictions of some physical parameters of the nuclei - in particular, the orientations of the spin axes. They treated the jet force in a physically more realistic way than the g(r) formula. However, the results were still dependent on model assumptions and thus questionable (cf. Sekanina 1988).
The introduction of the basic concepts of the Oort cloud and the icy conglomerate nucleus have naturally influenced modern ideas about the origin of comets. Oort (1950) already paid attention to the problem of formation of the cloud and hypothesized that it could have originated as a result of Jovian perturbations after the explosion of a planet-sized body in the asteroid belt. Thereby the asteroids and comets would have a common origin, the former being devolatilized variants of the latter. However, this revival of Olbers' old idea did not gain wide acceptance, partly due to the growing evidence that meteorites, obviously part of the same complex of minor bodies, have nearly solar elemental abundances and can not originate from a planet-sized parent body.
Around 1950, the Kant-Laplace nebular hypothesis for the origin of the solar system was also reconsidered in the light of the chemical compositions of the planets and their variation with heliocentric distance. Edgeworth (1949) and Gerald P. Kuiper (1949, 1951) argued that it is unlikely for the solar nebula to have ended abruptly at the position of Neptune's orbit, and thus a large population of planet precursors with a generally icy composition would have existed outside the giant planets. Kuiper (1951) claimed that such bodies could be identified with Whipple's cometary nuclei and suggested that Pluto's gravitational action (its mass was then thought to be in the 0.1 - 1 Earth-mass range) might have scattered the objects into Neptune's zone of influence, whereupon ejection into the Oort cloud would ensue. In particular, outside Pluto's orbit, the population might still remain intact.
During the following decades, Lyttleton (1952, 1974) challenged both basic concepts (the solid nucleus and the Oort cloud), arguing instead for cometary formation by aggregation of dust during the Sun's passages through interstellar clouds. Cometary origin thus would not be coupled to the origin of the solar system but to capture events throughout its lifetime. This scenario never received as much support as the one due to Kuiper, since it faced obvious difficulties, e.g. in explaining the cometary 1/a orig distribution and nongravitational effects. As both Oort's and Whipple's concepts have been consolidated in recent years, the basis for Lyttleton's picture has now virtually disappeared.
However, the idea of interstellar comets embraces many different scenarios that are subject to continued investigations. Aspects that have attracted particular attention are the distribution of aphelion directions of long-period comets and the possible signature of the solar apex, the mechanisms for formation of cometary nuclei under interstellar cloud conditions, the role of comets in galactic chemical evolution, and the significance of the fact that no hyperbolic comets have as yet been observed.
The standard concept of the solar nebula was criticized by Alfven and Arrhenius (1970, 1976), who argued for the importance of electromagnetic forces in the collapsing cloud, leading to a different picture of the radial arrangement of orbiting material and a different scenario for the accretion of larger bodies. In particular, the formation of comets was considered to occur by longitudinal focussing produced by self-gravitation and inelastic collisions in narrow streams of particles, so-called jet streams . This idea has not gained general acceptance, however. An eruptive origin of comets continued to attract attention as well. Van Flandern (1977, 1978) proposed, based on the distribution of orbits of long-period comets, that comets and asteroids originate from the break-up of a 90 Earth-mass planet in the asteroid belt only 5.5 x 10 6 years ago. This suggestion did not gain support, mainly on physical grounds (see the discussion following Van Flandern 1977). It was in stark disagreement with the picture building up during the 1970's and 80's, according to which minor bodies in general, and comets in particular, represent undifferentiated, pristine solar system material (Delsemme 1977). Sergej K. Vsekhsvyatskij (1972, 1977) continued to favour a variant of the Lagrange ejection hypothesis, involving the satellites of the giant planets, but he remained quite isolated in a community dominated by the view of comets as primordial bodies probing the solar nebula. The idea is fraught with many problems of different nature - let us mention here only that of explaining the abundance of long-period comets.
1985 - 1986: Encounters with P/Giacobini-Zinner and P/Halley
Following the enormous increase of interest for comets in the late 1970's, another giant leap in our understanding of comet phenomena occurred in March 1986 when six spacecraft (henceforth `S/C') observed P/Halley in situ, and future cometary scientists will undoubtedly speak about the pre- and post-Halley eras, much as historians describe the transition from the dark ages to the renaissance period. However, the first cometary encounter took place already six months earlier on September 11, 1985, when the ISEE-3 spacecraft, released from its earlier task of monitoring the Earth's radiation belts and, renamed as the International Cometary Explorer (ICE) , passed through the tail of P/Giacobini-Zinner , about 8000 km from the nucleus. The main results were the confirmation of the plasma tail model, indications about the ion composition and the detection of a neutral current sheet at the center of the tail. ICE flew on to register the effects of P/Halley on the interplanetary medium from a distance of 28 x 10 6 km sunward.
The detailed results from the extraordinary P/Halley campaign during the 1985-86 apparition fill many volumes - it will here suffice to give a very condensed overview of the main results.
Five spacecraft encountered P/Halley in early 1986: Vega 1 (closest approach on March 6 at 8890 km distance),Suisei (March 8; 150,000 km); Vega 2 (March 9; 8030 km), Sakigake (March 11; 7 x 10 6 km) and Giotto (March 14; 600 km). At the same time, an unequalled long-term Earth-based observational effort was coordinated by theInternational Halley Watch (IHW) (Newburn and Rahe 1990); the IHW Archive with more than 25 Gbytes of data was released in December 1992 (IHW 1992) and the associated Summary Volume (Sekanina and Fry 1991) contains detailed information about the data obtained within the various IHW Networks. The observations were carried out in all wavebands from the UV at 120 nm to the radioband at 18 cm, by professionals and amateurs. It has proven particularly fruitful to combine space- and Earth-based observations for calibration and long-term monitoring purposes. In general, the earlier developed cometary models were confirmed and could now be quantified by in situ measurements, leading to many new insights.
The nucleus was observed at close distance for the first time; it was found to be larger (equivalent radius about 5.5 km) and darker (albedo about 4 percent) than expected. Surface features (craters, ridges, mountains etc.) and the emitting vents were observed. The coma was found to be highly structured on all scales (jets, shells, ion streamers, etc.) and the gaseous component (parent molecules, radicals, ions and atomic species) was analyzed in situ; H 2 O was confirmed to be by far (about 85 percent by weight) the most abundant constituent in the gas phase. A cavity devoid of magnetic field was detected within about 5,000 km of the nucleus. The dust was analyzed by size and composition; there was an unexpectedly high fraction of very small grains, down to the sensitivity limit (about 10 -19 kg). In addition to those of possibly chondritic composition, carbonaceous `CHON' particles were seen for the first time; they may be a new source of gas. Atomic masses from 1 to 100 amu. were detected by mass spectroscopy, and the likely presence of large organic polymeric molecules was indicated. The maximum measured production rates were larger than 10 4 and about 3 x 10 4 kg/sec for dust and gas, respectively, i.e. a dust/gas ratio larger than 0.3. The integrated mass loss experienced by the nucleus at this passage, of the order of about 4 x 10 11 kg (but very uncertain) was about 0.5 percent of the total mass of the nucleus, estimated at 1 - 3 x 10 14 kg. The brightness of the central condensation appears to be varying with pseudo-periods of about 2 and 7 days, but it was not possible to determine unambiguously the rotational state of the nucleus. The various predicted plasma effects were confirmed, including the existence of a bow shock and the adjacent interplanetary medium was found to be kinematically and magnetically extremely turbulent. Several disconnection events in the ion tail were observed, also at the time of the encounters, and the suspected connection with magnetic field reversals was partly confirmed.
1986 - 1993: P/Halley follow-up
Much of the period after the Halley encounters has been spent reducing the enormous amount of data on this comet. Ground-based observations of a number of other bright comets, including Wilson (1987 VII), Austin (1990 V), P/Brorsen-Metcalf (1989 X) and Levy (1990 XX), have served for comparison and have also resulted in several important discoveries, for instance of some new parent molecules, e.g. H 2 CO, H 2 S and CH 3 OH. Thanks to improved instrumentation and reduction techniques, it has become possible to observe fainter and more distant comets than ever before. To some surprise it has been found that several comets continue to be active many years after perihelion passage, in some cases at heliocentric distances well beyond 10 AU; this has implications for the models of the nuclei.
Another space encounter with a comet took place on July 10, 1992, when the Giotto spacecraft flew through P/Grigg-Skjellerup during the Giotto Extended Mission (GEM), cf. Schwehm and Grensemann (1992). A preliminary overview of some of the early results was published by Boehnhardt et al. (1992). The foremost virtue of GEM has been to provide direct comparison between a very active and a supposedly less active comet and to search for the underlying causes. However, P/Grigg-Skjellerup was found to be at least as active as expected, and the first presence of cometary ions was detected at a distance of about 6 x 10 5 km, while a magnetic disturbance resembling a bow shock or wave was passed at about 1.5 x 10 4 km distance. A few dust impacts occurred just after the closest approach which took place at about 200 km from the nucleus.