This is a story about a discovery and some of the developments which followed it. It is not a textbook. Although I hope it contains most of the relevant technical details I set out to show a little of how astronomy is actually done. Some of the characters spend their time looking through telescopes on the darkest of dark nights, others work in offices and laboratories far removed, both physically and psychologically, from mountaintop observatories. From time to time this diverse group of people come together, in small groups or en masse, to exchange ideas and dispute data. They do this in order to understand the origin and evolution of the solar system in which we live and work. A few names crop up frequently, for the community of solar system astronomers is a small one and our paths often meander across each other in unpredictable ways.

In the last few years a new, and dynamic, outer solar system has replaced the sterile border known to our predecessors. I still find it hard to believe how much our view of the solar system has changed in the last decade and even harder to credit that I have been a part of this adventure.

In July 1943 the Journal of the British Astronomical Association published a short article entitled ‘The Evolution of our Planetary System’. The paper had been submitted by a retired Irish soldier and part-time amateur theoretical astronomer, Lt-Col. Kenneth Edgeworth. Despite being greatly reduced in length due to wartime shortages of paper, the article contained a prophetic paragraph on the structure of the solar system. While discussing comets, Lt-Col. Edgeworth remarked, ‘It may be inferred that the outer region of the solar system, beyond the orbits of the planets, is occupied by a very large number of comparatively small bodies.’ Kenneth Edgeworth did not live to see his prediction confirmed, but almost 50 years later just such an object was discovered. This new body, initially called simply 1992 QB1, was the harbinger of a breakthrough in our understanding of the solar system. Within a few years hundreds of similar objects would be found in what, by an ironic twist, soon became known as the Kuiper, rather than Edgeworth, Belt.

During the first few hours after each impact, numerous phenomena were observed with telescopes on Earth, in orbit, and in space. The primary events in that time were: impacts themselves, rise and fall of large plumes of ejected material, and atmospheric waves; also of interest were the characteristic morphologies of fresh sites. Based on timing from Galileo instruments and ground-based observations, the Hubble Space Telescope (HST) recorded actual impact phenomena for fragments G and W, with the A and E impacts occurring just prior to the HST observation window. For these four events, plumes were directly imaged; plume development and collapse correlated with strong infrared emission near the jovian limb, supporting the interpretation that the IR brightness was created by the fall-back of plume material from high altitude (see chapter by Nicholson). For medium-to-large fresh impact sites imaged by HST within a few hours of impact, expanding rings were detected, caused by horizontal propagation of atmospheric waves (see chapters by Ingersoll and Zahnle). Initial site morphology at visible wavelengths was similar for all medium-to-large impacts: a dark streak surrounded by dark material, dominated by a large crescent-shaped ejecta to the southeast. Smaller impact sites typically only showed a dark patch (no ejecta) which dissipated quickly. This chapter summarizes the most recent measurements and interpretations of plumes and fresh impact sites as observed by HST.

The dark clouds that were easily seen in small telescopes after the comet impacts were caused by small particles which were deposited in Jupiter's stratosphere. Observations from the Hubble Space Telescope and from ground-based instruments at visible and infrared wavelengths indicate that the mean radius of the particles is in the range 0.1 to 0.3 μm, and the total volume of particles is approximately the same as that for a 1-km diameter sphere. In the dark core regions of freshly-formed impacts, the particles are distributed over a large vertical extent, between about 1 mb and 200 mb or deeper. The diffuse outlying haze is confined to the high-altitude end of the range. Such a distribution probably reflects different methods of emplacement of the debris as a function of distance from the impact. The color of the particles, and their volatility as required to make waves visible, suggest an organic material as the main constituent. These relatively volatile materials are thought to have condensed onto more refractory grains after the plume material cooled, some 30 minutes or more after impact. The most refractory materials expected to condense from an evolving fireball are Al2O3, magnesium and iron silicates, and soot, depending on the C/O ratio. A silicate spectral feature was observed, confirming that cometary material was incorporated into the grains, although silicate grains make up only 10–20% of the particle volume.

Galileo observations in the UV, visible, and infrared uniquely characterize the luminous phenomena associated primarily with the early stages of the impacts of SL9 fragments—the bolide and fireball phases—because of the spacecraft's direct view of the impact sites. The single luminous events, typically 1 min in duration at near-IR wavelengths, are interpreted as initial bolide flashes in the stratosphere followed immediately by development of a fireball above the ammonia clouds, which subsequently rises, expands, and cools from ∼ 8000 K to ∼ 1000 K over the first minute. The brightnesses of the bolide phases were remarkably similar for disparate events, including L and N, which were among the biggest and smallest of the impacts as classified by Earth-based phenomena. Subsequent fireball brightnesses differ much more, suggesting that the similar-sized fragments were near the threshold for creating fireballs and large dark features on Jupiter's face. Both bolides and fireballs were much dimmer than had been predicted before the impacts, implying that impactor masses were small (∼0.5 km diameter). Galileo data clarify the physical interpretation of the “first precursor,” as observed from Earth: it probably represents a massive meteor storm accompanying the main fragment, peaking ∼10s before the fragment penetrates to the tropopause; hints of behind-the-limb luminous phenomena, recorded from Earth immediately following the peak of the first precursor, may be due to reflection of the late bolide/early fireball stages from comet debris very high in Jupiter's atmosphere.

Measurements of thermal emission in spectral regions, ranging from the near-infrared to mm wavelengths provide information on the atmospheric thermal structure over impact sites from μbar levels in the upper stratosphere down to the upper troposphere. Systematic time series of observations relevant to this entire height range over individual spots do not exist. However, by piecing together information at different times from various spots, it is possible to obtain a provisional, semi-quantitative picture of the behavior of the thermal structure over a typical impact site. Immediately after fall-back of the ejecta plume, the upper stratosphere is heated to ∼ 600–1300 K above ambient temperature. The amplitude of the temperature perturbation diminishes with increasing depth in the atmosphere, but even in the upper troposphere a temperature increase of a few kelvins is observed. Initially, the upper stratosphere cools very rapidly with time scales of tens of minutes, presumably the result of strong radiative cooling associated with the high temperatures. After the initial cooling, all levels continue to cool at slower rates with time scales of a few days; however, this is still very rapid compared to radiative cooling of the ambient atmosphere. Enhancements in infrared opacity necessary to produce the cooling radiatively do not appear to be viable, suggesting that dynamical effects may play a dominant role. Possible mechanisms include horizontal mixing with the ambient atmosphere and adiabatic cooling produced by upward motion associated with an anticyclonic vortex.