1 Introduction

Theories of cometary origin fall into two groups: primordial theories in which comets formed concurrently with the Sun and planets and have been stored in the Oort cloud over the age of the Solar System, and episodic theories in which comets either formed in distant interstellar clouds and were subsequently captured by the Solar System or comets formed as a result of catastrophic, recent and possibly repeated events in the planetary region (Mumma et al. 1993; Irvine et al. 2000). Current observational findings favour primordial theories; in this scenario, comets scattered outwards gave rise to the Oort cloud and comets scattered inwards contributed to the heavy bombardment in the inner Solar System and may have brought the ocean water and pre-biotic organic material to Earth.

A widely accepted model of the formation of comets in the Solar System assumes that most long-period comets originated in the Uranus-Neptune zone or just beyond with subsequent dynamical ejection by the growing protoplanets to distant orbits to form the Oort cloud (see, e.g., Duncan et al. 1987 and Stern & Weissman 2001 for details). However, if comets, and possibly some other minor bodies, were formed soon after the first planet-sized objects (likely proto-Jupiter and proto-Saturn), they should appear in a region not too close to the growing protoplanets as to allow temporarily steady dynamical evolution of the rocky material. Mumma et al. (2001a) have found that the measured compositions of four well studied comets (Halley, Hyakutake, Hale-Bopp and Lee) are consistent with formation from interstellar ices in the nebular region beyond Uranus. Nevertheless, the unusual organic composition of comet C/1999 S4 LINEAR suggests that it was formed in the inner part of the outer disk of the Solar System, most likely in the Jupiter-Saturn region (Mumma et al. 2001b) as it was depleted in carbon-chain molecules relative to most other comets. In addition, the chemistry of the building blocks of the nucleus was very uniform (Bockelée-Morvan et al. 2001) and this may thus also apply to their formation region between Jupiter and Saturn in the protoplanetary disk. Recently, Kawakita et al. (2001) have collected high-resolution spectra from the Subaru telescope of this object to determine the ortho-to-para ratio of NH₂. By assuming that the NH₂ is derived from the photodissociation of ammonia by the solar radiation, they derived a spin temperature of the ammonia of about 38 K. If this ammonia is primordial, then the spin temperature indicates that the comet formed between 8 and 15 AU in the solar nebula. Peculiar objects with unusual organic composition, tentatively identified as formed in the Jupiter-Saturn nebular region, have been named Jovian-class comets (Mumma et al. 2001b), although they should not be confused with the Jupiter-family comets, a dynamical grouping with Tisserand invariant in the span 2-3 and orbital periods <20 yr that formed over a much wider range of heliocentric distance and only later entered their present orbits (Fernández et al. 1999). In a very recent paper, Nurmi et al. (2002) conclude that the Jupiter-family comets cannot originate in the spherically distributed Oort Cloud and that the existence of this dynamical class can be explained only if they are captured from the extended disk of comets with semimajor axes a < 5 000 AU. Typical members of the Jupiter-family are the comets 9P/Tempel 1 and 19P/Borelly (Soderblom et al. 2002).

C/1999 S4 LINEAR was a unique object in many different aspects (Mäkinen et al. 2001; Mumma et al. 2001b; Weaver et al. 2001; Kawakita et al. 2001). This peculiar comet was discovered on 27 September 1999 by the Lincoln Near Earth Asteroid Research (LINEAR) program, while still at 4.2 AU from the Sun. The comet became a favorable target for observations during July 2000. It passed its closest point to Earth, 0.364 AU, on 23 July 2000, 3 days before it reached perihelion. Surprisingly, the comet completely disintegrated within a few days of its closest approach to Earth. Photometric measurements (Farnham et al. 2001) suggest a lower limit of 0.44 km for the radius of the cometary nucleus before breakup. Photometric analysis (Weaver et al. 2001) of the fragments indicates that the largest ones have effective spherical diameters of about 100 m. The total mass detectable in the dust tail after the breakup was $3 \times 10^8$ kg suggesting that most of the comet's original mass after breakup was in the form of remnants between 1 mm and 50 m in diameter. Both its orbit and activity evolution during its approach to perihelion suggest that C/1999 S4 LINEAR was a dynamically new comet from the Oort cloud, possibly on one of its first close visits to the Sun (Farnham et al. 2001).

In this paper we suggest a plausible mechanism for the formation of Jovian-class comets: outer resonance trapping resulting from gas drag (Weidenschilling & Davis 1985; Patterson 1987; Beaugé & Ferraz-Mello 1993; Kary et al. 1993; Scholl et al. 1993; Beaugé et al. 1994). We model the evolution of solid particles subject to nebular gas drag and perturbation from growing protoplanets (Jupiter and Saturn). Particles of various masses are distributed on almost circular coplanar orbits external to the protoplanets. As their orbits decay sun-ward, solid particles become trapped in external gas-induced mean-motion resonances with the protoplanets. While in theory this trapping may be permanent, trapping lifetime is limited by several mechanisms, such as growth of the protoplanets, formation of Uranus and Neptune and dispersal of the protoplanetary nebula. We begin in Sect. 2 by introducing the basic concepts considered in this paper. We analyze the results of our calculations in Sect. 3. In a concluding section a discussion of the implications of these results is presented.