1. Introduction

From both theoretical and observational perspectives, the concept that the Earth has evolved in isolation, unaffected by its environment, has been in retreat for the past three decades. The appreciation that the Earth is a heavily bombarded planet came in the late 1970s following the lunar landing programs, which established that the craters of the moon had an impact origin; from the discovery of increasing numbers of terrestrial impact craters; and from the discovery of increasing numbers of small, Earth-crossing hazardous asteroids. It was soon realised that such bombardments were sufficiently frequent and devastating to have left geological and biological signatures, including mass extinctions (Hoyle and Wickramasinghe 1978, Napier 2009; Napier and Clube 1979). The first 'hard evidence' for such celestial input came with the seminal finding of Alvarez et al (1980) that an excess of the siderophile element iridium occurred at the Cretaceous-Tertiary boundary, where the dinosaurs became extinct. Later discoveries of exotic aminoacids (AIB and isolvaline) in the same sediments confirmed this result.

That the cosmos might bring life as well as destroy it is an idea which seems to have originated with Anaxagoras in the fifth century BC, who considered that invisible germs, "spermata", permeated the cosmos. Twenty two centuries later, in 1871, Kelvin proposed that life-bearing rocks could be transferred between planetary bodies as a consequence of collisions. Three years after that von Helmholtz suggested that the seeds of life might be carried from one planet to another. And in the beginning of the 20th century the Swedish chemist Arrhenius suggested that bacterial spores were transported between the stars by the radiation pressure of starlight. In the last quarter of the 20th century, Hoyle and Wickramasinghe (1982) revived and updated the panspermia concept, developing the view that comets are important as transmitters of life throughout the Galaxy. In this paper we describe the current status of this concept.

2. Transfer of Microorganisms within the Solar System

2.1 The inner planetary system

Of the 24,000 or so meteorites held in collections worldwide, about 130 are currently known to have come from the Moon and 35 from Mars. Many of these meteorites have low to moderate shock levels, never having been subjected to temperatures above about 100°C, and it has been claimed that three Martian meteorites -- Allen Hills, Nakhla and Yamato -- contain the fossilised remains of bacterial mats (Gibson et al 2001. These ejecta constitute hard evidence that material can be transferred between the inner planets, and in the process need not reach a temperature lethal to any microorganisms within it.

Numerical simulations of the delivery process, and cratering mechanics, have clarified the mechanisms involved. An impacting body large enough to punch its way through the atmosphere of one of the inner planets may strike the ground at some tens of kilometres a second, generating a crater and throwing boulders and soil at high velocities upwards. Shock cancellation in the target surface layers during a large impact may yield ejection into space without much heating, the rocks being thrown upwards at up to about half the speed of the impactor (Melosh 1984). A 20 km diameter impact, striking Mars at 15 km/s, would result in the ejection of something like 30 million boulders of mean diameter 6 m -- sufficient to protect microorganisms, buried deep inside, from the damaging effect of cosmic rays -- with ejecta temperatures less than 100°C.

Once ejected from a planet, the material will go into orbit around the Sun, possibly for millions of years, before either being expelled from the planetary system, falling into the Sun or landing on another planet. Mileikowsky et al (2000) found that, following a large impact on Mars, there is a steady infall of meteorites onto the Earth thereafter, ~6% of the ejecta reaching the Earth over the next 10 million years before a decline in flux sets in. Over that period, the impact yields about two million large, unshocked Martian boulders falling to Earth. If such impacts occur at say 100 million unit intervals, there is a clear potential for the transfer of viable microorganisms from Mars to Earth.

It has been argued that life is more likely to have originated on Mars than on Earth, since prebiotic processes might possibly have proceeded there for hundreds of millions of years before the Earth was cool enough to support them (Davies 1999). However, this argument has less force when one puts the solar system into a wider Galactic context (Sections 3 – 5).

Similar considerations apply in the opposite direction, with the fraction of terrestrial material ending up on Mars after 10 million years being about 0.16%. Even a gram of basalt 5 or granite may contain 10⁷-10⁸ microorganisms in cracks and fissures (Haldeman et al. 1994), and so a two-way fertilisation process between Earth and Mars may take place.

These processes depend on whether organisms can survive the shock of being thrown into space. Burchell et al. (2004) used a gas gun to fire pellets containing microorganisms at high speeds onto solid surfaces. At shock pressures of 30 GPa, survival fractions of 10^-4 – 10^-6 were obtained, declining as shock pressures increased.

These results have been confirmed by Horneck et al (2007) for a variety of microorganisms. The shock pressures applied are in the range of those inferred for Martian meteorites, and so these results confirm that a significant proportion of bacteria will survive ejection by a large impact, opening the door to the transmission of life from one planet to another.

Several of the assumptions which have gone into these explorations have been on the conservative side. For example, the ejection process as described by Mileikowsky et al (2000) has individual rocks punching their way through several times their own mass of air, whereas there is terrestrial evidence that ejected material may ride on the vapour jets, making it easier to attain escape speed. The assumption of an exponential decay of bacteria in the presence of cosmic rays may also be queried, both from observational and theoretical perspectives. There is also experimental evidence that freeze-dried bacteria and spores are resistant to flash heating up to at least 350°C for 30 seconds rather than 100°C adopted in these studies (Al-Mufti et al 1986). These issues are discussed by Wallis and Wickramasinghe (2004).

It may also be that the canonical impact velocity 15 km/s adopted in these studies is too low since the impact frequency from comets in Halley-type orbits appears to have been underestimated in the past (Emel'yanenko and Bailey 1998; Napier et al 2004). In the case of an impact on Mars from a comet in a Halley-type orbit, the impact velocity is ~37.5 km/s; for the Earth it is ~57 km/s. This latter figure, coupled with the mass distribution of comets, implies that 90,000,000 tons of boulders are ejected from the Earth at survivable temperatures every 10 million years or, in round figures, 10 tons/yr.

2.2 The outer planetary system

What about the transmission of life beyond the inner planets? The ejected boulders are subject to strong planetary perturbations which make the orbits chaotic on timescales typically 10³ --10⁵ years. These perturbations have been modelled as either a diffusion process due to distant encounters, or a Monte Carlo process due to a succession of close encounters. In reality, orbits tend to be influenced by a mix of the two, and the fate of the 6 boulders can only be determined by numerical integration. Median dynamical lifetimes of order a few million years are generally found (e.g. Dones et al 1996). There seems to be no impediment to the transmission of life to the outer planets or their satellites or beyond.

Beyond that, Wallis and Wickramasinghe (2004) have proposed that over the lifetime of the solar system an average of ~3 kg of unsterilised planetary ejecta may become embedded in the surface layers of Edgeworth-Kuiper Belt comets, a system of perhaps 10⁹ comets orbiting just beyond the fringes of the planetary system, with a probable mass less than that of the Earth. In reality, microorganisms would have to be preserved in large boulders for some millions of years before significant numbers of them could migrate to the Edgeworth-Kuiper (EK) region. Once safely embedded under the icy surface of an EK object, dormant microorganisms would presumably be preserved indefinitely. The Edgeworth-Kuiper belt may lose a 10 -- 100 km object every 1 -- 10 years, giving it a half-life ~3 Gyr.

Early claims that fossilised bacteria reside in carbonaceous meteorites were discounted because contamination by ragweed pollen was demonstrated in a few instances (Claus and Nagy, 1961). The first convincing evidence of microfossils in the Murchison meteorite was obtained by Pflug (1984) and Pflug and Heinz (1997). These authors made thin sections of the meteorite, placed them on membrane filters and used HF gas to leach out the mineral component, thus leaving indigenous carbonaceous structures intact. Laser ion probes further established that the structures had a biological provenance, and contamination was thus ruled out. Examples of such structures compared with a modern microorganism and a virus are shown in Fig.1. Further remarkable evidence for microfossils analogous to modern cyanobacteria in carbonaceous chondrites has been described by Hoover in a long series of papers (e.g. Hoover 2006, 2009).

Carbonaceous chondrites constitute a few percent of known chondrites. The most primitive of these may contain up to 20% water and various minerals altered in the presence of liquid water, as well as clay-like hydrous phyllosilicates, amino acids and aromatic hydrocarbons. They have never been heated above ~50°C. There is oxygen isotope evidence that the parent bodies of carbonaceous chondrites and Comet Wild 2 (the Stardust target) belong to a single family of objects (Aleon et al 2009). Microfossil identifications have still to be confirmed, but if they stand up, the implications for the origin, evolution and dissemination of life are clearly profound. Inter alia it would support the view that solar system life originated neither on Earth nor Mars, but came in from elsewhere.

3. Transfer of life over interstellar distances

3.1 Ejection from the solar system

Melosh (2003) found that about 15 boulders a metre or more in diameter, originating from the inner planets, are ejected from the solar system every year. Their mean residence time within the solar system is about 50 million years, and their mean ejection speed is 5±3 km/s. With these figures, Melosh demonstrated that there was little prospect of any meteoroid from the solar system ever landing on an Earthlike planet in another system.The distances between stars are too vast, and the capture probabilities too low. These factors, and the powerful screening of Galactic cosmic rays, have been the prime objections to the idea of interstellar panspermia.

There is, however, a much more effective route for embedding microorganisms in an exoplanet.

These timescales are such that the boulders may be significantly affected by erosion and fragmentation. In the inner planetary system, boulders are effectively sandblasted and at risk of fragmenting if struck by small meteoroids. These small particles, in aggregate, constitute the zodiacal cloud. About 90% of the material in this cloud is dust generated by the breakup of Jupiter Family comets, with small admixtures of asteroid dust and particles from long-period comets.

Because of collisions and radiative drag, the lifetime of a zodiacal cloud particle is only ~20,000 years and the system must be replenished on a timescale of that order. Such replenishment comes largely from comets deriving from the Jupiter family, and is likely to be erratic, being dominated by the arrival of rare, large bodies into the system. The current mass of the zodiacal cloud has been estimated at ~5x10¹⁹ gm, with an appreciable uncertainty (Nesvorny et al. 2010), but a significant feature of the cloud is that on timescales of order 1 Myr its mass is subject to strong, random surges caused by the arrival of exceptionally large comets entering short-period orbits, followed by their disintegration into dust (Napier 2001).

Consider a boulder of density ρ_b struck by zodiacal cloud particles with speed V and space density ρ_z. Then its radius declines linearly with time due to erosion on a timescale τ

Consider for example a bacterial clump of radius 1 μm with a graphite coat of thickness 0.05 μm ejected from the Earth and orbiting one AU from the sun. This particle will have a ratio of radiation pressure to gravity of 1.5 and an asymptotic escape speed ~30 km/s; it would reach ~100 AU from the sun in about 15 years.

It can be estimated that 100 tons of boulders are ejected at survivable temperatures every 10 million years, corresponding to about 10 tons per annum. Taking a microorganism count in terrestrial soil to range from ~10⁷ --10⁹ per gram, then 10¹⁴ -- 10¹⁶ microorganisms per annum are ejected from Earth, in discrete episodes. In the presence of a large short period comet, soil or boulders will rapidly fragment to dust, the timescale being characteristically centuries. On collision, about half the mass of a 100 μm particle disintegrates into β-meteoroids and is ejected by radiation pressure. The terminal speed

Whatever the precise figure, it follows that the solar system is surrounded by an expanding biosphere of radius ≥5 parsecs containing typically 10¹⁹ -- 10²¹ microorganisms, depending inter alia on the fertility of the impacted ground; conceivably, the sphere could extend to over 20 parsecs.

A boulder is a relatively ineffective means of transmitting biological information. If we compare say ten metre-sized boulders ejected per annum with the same mass of material ground into β-meteoroids and ejected as ~μm-sized particles (10¹⁹ of them per annum), then for soil fertility of 10⁷ --10⁹ microorganisms per gram, one finds that the beta-meteoroids are 13 -- 15 powers of ten more effective in reaching targets than their same mass in rocks.

The largest impacts on Earth, capable of ejecting boulders, are probably due to comets, since asteroids more than about 2 km in diameter are too difficult to perturb out of the asteroid belt. Large, well-dated impact craters of the past 250 million years are found to have occurred, not at random, but in discrete episodes (Napier 2009). There is evidence that these bombardment episodes occur with a periodicity of ~35 Myr, and they can be understood in terms of Oort cloud disturbance caused by the Galactic tides which peak sharply whenever the Sun passes through the plane of the Galaxy. The comets need not impact directly from the cloud, but may pass through intermediary populations such as the Jupiter family of comets.

The Oort comet cloud is a major cometary reservoir, comprising perhaps 10¹¹ comets up to 50,000 astronomical units from the sun, and having a mass in the range 0.1 -- 250 Earth masses. The long period comets have orbital periods ~3 -- 6 million years. The system is only just gravitationally bound to the solar system and is unstable in the Galactic environment. In addition to regular modulation by the Galactic tides, sporadic disturbances by passing nebulae and stars result in the ejection of comets into interstellar space and give the outer regions of the Oort cloud a half-life ~1.9 Gyr. It is most probably replenished from an inner source.

Evidence for the influence of the Galaxy on the Oort cloud is found not only in the occurrence of bombardment episodes but also in the aphelion distribution over the sky of long-period comets. The effect of Galactic tides is to cause their orbits to precess, resulting in a tendency to avoid the north and south galactic poles, along with a deficiency of aphelia around the equator. It takes about 100 million years for an initially isotropic distribution to develop this pattern. For comparison with observations, a set of high-precision orbital elements for 386 comets was obtained from Dr Piotr Dybczynski (Poznan Observatory). These were the orbital elements of the comets before they entered the planetary system, and were derived by backward numerical integrations of the orbits, taking account of planetary perturbations. The elements had typically six or seven significant digit accuracy. Of these, ~100 had semi-major axes a>15,000 AU and were subject to significant precession and nutation due to the Galactic tide. The celestial distribution of their perihelia is shown in Fig. 2 along with their galactic latitude distribution (Fig. 3). The latitudinal dependence has long been known (e.g. Lüst 1984) and continues to hold with these high-precision data. These results confirm the sensitivity of the Oort cloud to Galactic disturbances.

A consequence of this is that, when the solar system has a close encounter with a giant molecular cloud, there are sufficient Oort cloud comets to yield a strong bombardment episode (factor of 10-20 increase above background) with a corresponding surge of ejection of life-bearing boulders with essentially instantaneous erosion into dust and expulsion from the solar system directly into the passing nebula. A giant molecular cloud may have mass 5x10⁵ that of the Sun and radius 20 parsecs. They are major sites of star formation: an OB association within such a cloud may contain 10³ -- 10⁴ young stars. Such encounters are quite frequent in geological terms (Fig. 4).

On average once every 400 million years, the solar system will pass through such a cloud, the passage time taking ~3 million years. In the course of such a penetration, if the impact rate on Earth is unchanged from its current value, about 3x10²¹ microorganisms will be deposited into the giant molecular cloud. During the time which life has existed on Earth, there may have been a total interval of 25 – 30 Myr when the solar system was passing through such a nebula, with a similar interval during which close encounters with smaller nebulae took place. There is an expectation over this interval of about 10 impacts yielding craters ≥ 50 km in diameter, and one ≥ 100 kilometres across.

However, in the course of an encounter with a massive nebula, the Oort cloud is gravitationally disturbed, and this leads to a temporarily enhanced impact rate of large impactors. Fig. 5 shows the outcome of a numerical simulation involving 30,000 comets during the course of a grazing encounter with a giant molecular cloud. There is a strong bombardment episode (factor of 10 -- 20 increase above background) with a corresponding surge of ejection of life-bearing boulders with essentially instantaneous erosion into dust and expulsion from the solar system directly into the passing nebula. For reasonable numbers (Napier 2007) one finds that this corresponds to one microbe per 10¹⁶ grams of dust within the giant molecular cloud: roughly one per incipient comet, reduced by whatever sterilisation processes are at work and enhanced by whatever replication takes place in cometary interiors. Typically, a GMC may have ~50,000 stars passing through it at any given time, as well as several thousand protoplanetary systems. There is a clear potential for propagating life throughout the Galaxy, with the nebulae gathering up, amplifying and disseminating microorganisms as they go.

Less massive nebulae are more common and are also sites of low mass star formation. The solar system passes within 5 parsecs of such clouds every ~100 million years (Fig. 4) and the deposition of microorganisms per gram of cloud is an order of magnitude higher than for giant molecular clouds.

3.11 Dispersal of primordial microorganisms

The discussion so far has been confined to exchanges, transfer and amplification of microbial genes that became established on planetary surfaces via the mechanism of comet impacts. If life originated in processes taking place over a galactic scale (Napier et al, 2008; Joseph 2000; Joseph and Schild, 2010a) or even on a cosmological scale (Gibson and Wickramasinghe, 2010; Gibson et al, 2010; Joseph and Schild, 2010b) a population of primordia bacterial cells would be expected to be present within interstellar clouds (Joseph 2000, Joseph & Schild 2010a). Although the vast majority of these cells may suffer inactivation and eventual degradation into particles resembling anthracite, a minute fraction of primordial organisms would always remain viable. These may represent cells within clumps or those captured into planetary discs on short transit hops from an earlier replication site (Joseph and Schild, 2010b). The feed-back loop shown in Fig. 6 would then continue to amplify an initial population of primordial bacteria. With a large fraction of interstellar C,N,O and metals being cycled in this way the interpretation of the bulk of the observed population of interstellar organics would then represent the detritus of biology (Joseph and Schild, 2010b). Furthermore, the trickle of 'evolved' life recycled in the same loop would have access to ample amounts of high grade organic nutrients to serve as feedstock for efficient replication.

3.2 Planetary systems in the Galaxy

A key requirement of interstellar panspermia is that there exist habitable locations in the Galaxy between which life can migrate. It would only take a handful of viable microorganisms -- perhaps even one -- to populate a receptive planet within a few years. This leads to the question of whether planetary systems capable of supporting life do actually exist, and this in turn leads to the concept of a circumstellar habitable zone. At its simplest, such a zone may be modelled by requiring that a planet within the zone should be capable of maintaining liquid water. In our planetary system, for example, the inner edge of the zone has often taken to be 95% of the Earth's current distance -- any closer in and the oceans would boil -- and the outer edge has often been taken to be 1or 2% further out, as beyond this the Earth would be permanently locked into a worldwide glaciation. More realistic models taking account of albedo extend this extremely narrow range to about 0.8 -- 1.5 AU.

The limitations of this line of argument are evidenced by the presence of internal heat sources and possible sub-surface oceans on Jupiter's moon Europa, and on the Saturnian satellites Enceladus and Titan, all which orbit too far from the sun to have liquid water by solar radiation alone, and yet are regarded as prime candidates for life in the solar system. McKay et al. (2008), for example, have proposed that the biological activity in a subsurface water environment on Enceladus may be possible for the energy source which drives jets of icy particles and water vapour from its south pole. The hydrocarbon mix in the jets has been taken to indicate such activity.

In general, tidal stresses in a moon orbiting a giant planet may generate sufficient internal heat for billions of years, while radiogenic heating is also likely to be significant.

At the time of writing about 500 exoplanetary systems have been discovered, about an order of magnitude up on the number known a decade ago. At least a quarter of the cool main sequence stars within about 50 parsecs are known to have planets in orbit around them. The masses range from a few Earth masses up to about five Jupiter masses, with a preponderance of Jupiter masses which may simply be an observational selection effect (Doppler spectroscopy being the prime mode of discovery: our own solar system, seen from say 25 parsecs, would be detected as having a single planet after at least 12 years of observation). About a quarter of the exoplanets discovered so far belong to multiple systems.

Binary stars will inhibit the range within which dynamically stable planetary orbits are possible but this seems not to be a major restriction. On the basis of extensive numerical trials, David et al (2003) estimate that about 50% of binaries allow an earthlike planet to remain stable for something like the age of the solar system. According to Lada (2006), the binary star fraction decreases with declining mass, with about 70% of late-type stars being single.

A white dwarf is essentially a helium-rich degenerate core (the endpoint of stellar evolution for over 90% of stars in the Galaxy), and heavy elements added to its atmosphere rapidly sink into the interior because of the strong surface gravity. Thus the presence of such elements in the atmospheres of some white dwarfs implies recent pollution, certainly within the last ~Myr. In the past this has been ascribed to interstellar accretion, but a study of the Galactic position and kinematics of 146 cool white dwarfs in the Sloan Digital Sky Survey by Farihi et al. (2010) revealed that this is not generally an adequate explanation. At least 3.5% of them had atmospheres polluted by heavy elements and hydrogen which seem to be the visible remains of water-rich, rocky planetary systems. The progenitor stars, of spectral types A and F, are little more massive than the Sun. This is likely to be a severe lower limit to the proportion of rocky exoplanets around such stars.

The upshot is that, although hard evidence is still lacking, it seems that a major proportion of stars in the Galaxy contain planetary systems, but the percentage of those which contain earthlike planets in stable orbits in habitable zones is still uncertain. If say the proportion is 1%, then one expects a billion such planets in the Galactic disc with mean separation ~10 light years. If 0.1%, then 100 million such systems exist in this galaxy alone, with mean separation ~20 pc.

3.3 Liquid water in comets

Liquid water may also form briefly in the interiors of comets, due to the heat generated by radioactive elements. The heat available from the decay of ²⁶Al and ⁶⁰Fe in cometary material is more than two orders of magnitude higher than the heat of fusion of water ice. The half-lives of these elements are respectively 0.74 and 1.5 Myr, whence comet formation would have to take place on timescales of this order after these radionuclides had been created by a supernova and injected into the primordial solar system. Typically half the volume of a 30 km comet might become liquid within half a million years of its formation, declining to 20% after 1.5 million years (Wickramasinghe et al 2010). While it must frequently happen that comet formation takes place in these circumstances throughout the Galaxy, long-lived radioactive elements such as ²³²Th and ²³⁸U may likewise yield melting, but only in the case of comets 300 km or more in diameter. Whether comets of this size exist is not known: Comet Sarabat, which in 1729 approached the Sun to within 4 AU on a parabolic orbit, is a possible. Even without radiogenic melting, however, the extremely low albedos of cometary surfaces (typically 0.02 -- 0.04) may yield subsurface liquid water less than a metre down when a comet approaches to less than about 1 AU of the Sun. The significance of this brief melting is that, should viable anaerobic microorganisms be present in the initial nutrient-rich material of the comet, they would grow exponentially in numbers.

The discovery of the amino acid glycine, a precursor of proteins, in a sample of Comet Wild 2 brought back by the Stardust spacecraft (Elsila et al. 2010) strengthens the view that amino acids were widespread in the early solar system and the interstellar cloud from which it collapsed.

3.4 The Galactic habitable zone

The idea of a Galactic habitable zone is one in which supernovae, gammaray bursters, explosive events in Galactic nuclei and the like limit the spread of life.

The Galactic habitable zone is, on its inner edge (7 kiloparsecs), supposedly set by radiation sources such as supernova and gammaray bursters. The outer edge, at 9 kpc, is set by the lack of heavy elements necessary for the development of carbon-based life. Thus the zone is currently only 2 kpc wide. The width of this zone is supposed to be increasing slowly with time, the zone being composed primarily of stars that formed between four and 8 billion years ago. About three quarters of the stars in this zone are older than the Sun (Lineweaver et al 2004; Gonzalez et al 1991).

This concept has been taken further by Franck et al (2002), who introduced the additional consideration that a biosphere on a planet will itself modify the conditions in which life can continue to exist, in particular through amplifying weathering processes through an amplified global carbon cycle. They used an Earth system model to compute the number of Earthlike planets in the Galaxy which may have developed primitive and complex life, based on the carbon cycle mediated by life and driven by stellar evolution and plate tectonics. Franck et al deduced that the number of planets which could bear complex (multicellular) life is about two orders of magnitude less than that capable of supporting primitive (unicellular) life and that the number of stars with Earthlike planets is a few million.

The whole concept of a Galactic habitable zone has been criticised by Prantzos (2006) on the grounds that the lethality of supernovae and the role of metallicity in planet formation are not sufficiently well understood for strong statements to be made. For example, the Galactic disc should probably not be considered as a closed box, but rather something which evolves with a prolonged input of gaseous infall which, with time, will flatten out the metal abundance gradient. And even if a nearby supernova destroyed all land life, marine life would largely survive and repopulate continents within a few percent of the lifetime of a G-type star. Prantzos considers that the Galactic habitable zone, within which complex life may exist, could extend throughout the whole Milky Way.

4. Beyond the Milky Way

Could life have spread widely in the early universe, when matter density was much higher and much more gas was present? Any expansion of life must be limited by the various cosmological horizons, a past horizon set by the consideration that there may be realms whose light has not had time to reach us, and a future horizon -- should the universal expansion be accelerating -- set by the fact that distant objects may recede faster than information sent out from any given locality could catch up. Such horizons, however, do not even remotely apply to the mechanisms discussed here. At the diffusion speeds considered here, an upper limit of about 10^-4 of the present-day visible horizon, or about 1 million light years, is the most that could be contemplated.

It seems the only way to avoid a horizon-limited spread of life would be to adopt a non-conformist cosmological model such as that of Gibson and Schild (2007), which involves the creation of Earth-mass clouds at the recombination stage 300,000 years after the Big Bang. In this model, these fragments condense into frozen planets which are then polluted with heavy elements from the first generation of supernovae. The creation of life at one locality in this extremely dense, collision-dominated system would lead to the fertilisation of other habitats, and the spread of life across the entire primordial universe within a million years can then be envisaged.

The conventional view, however, is that the first generation of stars formed about 200 million years after the Big Bang, before the first galaxies appeared. Such stars contained no elements heavier than carbon, and theoretical modelling indicates that most of them were bright and hot, with masses typically several hundred times that of the Sun. These supermassive objects formed black holes. For stars to collapse out of gas, angular momentum had to be removed, but in the absence of heavy elements it is not clear how disc fragmentation would proceed and what form planetary systems, if any were formed, could take.

Starforming galaxies have been detected at distances such that the universe was then only 500 million years old and, on currently fashionable models, were formed by the aggregation of smaller galaxies. When the cosmic microwave background cooled to ~20 degrees Kelvin (at cosmological redshift z~6), molecular hydrogen could form, according to conformist cosmological theory, leading to the growth of molecular clouds. This would happen about 1.3 Gyr after the Big Bang, when the scale factor a was about 15% of its present-day value. The molecular clouds would help to shield microorganisms from radiation and cosmic rays, and the spread of life generated at a particular locality might then proceed efficiently, provided the radiation environment was not too hostile (Joseph and Schild 2010). With the second generation of stars, metals were now present and it is likely that planet formation became an integral part of the star formation process, and biologically friendly solid-state surfaces -- whether on planets, or their satellites, or cometary interiors -- could develop.

But could life then have spread from a single locus within the framework of conventional cosmology? Planetary bombardment rates would have been high, facilitating the ejection of microorganisms and the passage of stars through molecular clouds would have been extremely common; both these factors would facilitate dispersal. Against this, radiation and cosmic ray densities were much higher than those in our present-day Galactic environment. For a fixed number ν of damaging radiation sources per unit proper volume (say highly luminous B stars or supernovae), the energy density of damaging radiations varies as a³, survival half-lives likewise, and travel times by a, where a is the scale factor of the expanding universe. At 350 million years the scale factor of the universe was about 5% of its present value, enhancing radiation densities by ~5 x 10⁸ for a given ν. It is hard to see how life could have dispersed significantly or even thrived in such a hostile environment.

Even at redshifts z~2.3 corresponding to a time ~3 gigayears after the Big Bang, stars in the early universe formed much more rapidly, due to the greater abundance of molecular gas (Tacconi et al. 2010), and the UV and cosmic ray energy densities were probably about an order of magnitude larger than at present. Given the exponential scaling of survival times, interstellar panspermia may still have been strongly inhibited.

According to Forbes and Bridges (2010), about a quarter of the globular star clusters in our galaxy are invaders from other galaxies. These appear to be the remnants of 6--8 dwarf galaxies which have been assimilated into the body of our Galaxy, but their associated globular clusters remaining as coherent entities. One satellite galaxy, a dwarf in Sagittarius, is in process of being eaten up, its stars being assimilated into the halo of the Galaxy.

On the basis of these arguments it seems that life might spread throughout a region about the size of a small cluster of galaxies, such as the Local Group, but it is not obvious that it could have travelled much further, at any rate within the framework of standard cosmology. On a cosmic scale, we can therefore think of panspermia as a local phenomenon, not extending beyond a few megaparsecs. Beyond that, any existing lifeforms probably have no ancestral connection with our own. Thus in the context of standard cosmology we should think of life in the universe as made up of individual 'empires' rather than a single, all-pervasive family. How lifeforms would develop in regions where empires intersect is a matter for speculation.

5. Some Implications of Interstellar Panspermia

5.1 The transfer mechanism in summary.

The fact that from time to time solar system passes through or close to starforming regions means that the forbidding distances usually associated with interstellar panspermia are reduced to a short hop. These short hops occur preferentially at times of Oort cloud disturbance, high bombardment rate and maximal creation of β-meteoroids from ejected boulders. A layer of ~0.03 μm of charred bacteria will effectively shield bacteria in the interior of a grain from ultraviolet photons. Cosmic rays may reduce the viable fraction of a microbial species contained within grains to say 10^-6 of its original population if the 'short hop' is of duration say 10⁴ –10⁵ years. However since a survival rate ~10^-21 or so is all that is required for a successful transfer of life into protoplanetary nebulae (Wickramasinghe 2004), many orders of magnitude are in hand to fertilise a starforming cloud as the Sun passes through it. In entering a planetary atmosphere like that of the Earth, a bacterium is subjected to flash heating. For 10 μm clumps the heating is sufficiently brief that survival is possible; grazing encounters will enhance the survival probability (Coulson and Wickramasinghe 2003).

Thus we may start with life taking root in a planetary system, or the interior of a comet, or even in some environment not yet contemplated. If we assume a doubling time of about 300 million years for fertilising a planet (about the encounter time of a disc star with starforming nebulae) and give it 33 generations, then 10¹⁰ planetary systems in the Galactic disc are populated within 10¹⁰ years. Only 1.12 habitable planets or their precursors need to be inoculated during an encounter with a molecular cloud for panspermia to go to completion within the age of the Galaxy. The local number density of clouds within 3 -- 5 kiloparsecs of the galactic centre is about five times that of our locality, implying that saturation could be reached in much less than the age of the Galaxy.

5.2 Implications for the origin of life

So long as potential habitats in the Galaxy were thought to be separated from each other by an impenetrable screen of cosmic rays and ultraviolet radiation, it was reasonable to look to a terrestrial factory for the creation of life. Historically, therefore, a distinctly geocentric perspective has guided the direction of abiogenesis, going back to Darwin, throughout the 20th century via Oparin and Haldane in the 1920s and Urey and Miller in the 1950s. However given that we now seem to have one or more mechanisms for transporting microlife between the star systems, and that the fundamental building blocks of life are already present in comets and nebulae, it is reasonable to look at the possibilities offered by a broad range of Galactic environments.

A further consideration rests on the sheer vastness of the Galaxy in relation to the Earth. The mass of water in comets belonging to the Oort cloud probably exceeds that in all the world's oceans by four or five powers of ten. Multiplying this by the number of stars in the Galaxy yields an excess factor ~10¹⁵. Most of this is water ice, but it is probable that a significant proportion of it has been liquid for period of order a million years (or gigayears in the case of rare, giant comets). And a true mass-for-mass comparison lies not with the Earth's oceans in general, which are more likely to dilute than concentrate organic molecules, but with small, specialised watery environments -- the warm pond, the ocean spray or the hydrothermal vent. If the conditions which are assumed to have given rise to life on Earth are found also in cometary interiors, then the advantage for abiogenesis is overwhelmingly with the comet.

These conditions seem to be satisfied if one models the origin of life through the RNA world or through clay crystals. Clay substrates in comets might provide the necessary conditions for the growth of early cell membranes; post-impact spectra of Comet Tempel I following the Deep Impact event show good agreement with the standard clay spectrum over 8 -- 13 μm, and clay structures could form when submicron silica particles from interstellar dust come into contact with liquid water (Wickramasinghe et al 2010).

Dworkin et al (2001) have reproduced the condensation of common interstellar gases (water, methane, ammonium, carbon monoxide) onto grains in cold interstellar conditions. The iced gases were irradiated with UV, and the substrate alternatively heated and cooled. On exposure to liquid water, the resulting organic mixture self-organised into 'membranous vesicles' – droplets -- similar to lipid membranes found in living cells. The existence of enclosing membranes is regarded as vital to the development of early life.

The hypothesis that late accretion infall seeded the inner planets with complex organic compounds has been strengthened considerably since it was first proposed by Oró in 1961. The mass of carbonaceous material acquired by the early Earth was probably ~10²² gm, about 10,000 times greater than the mass of the current biosphere. Given that the building blocks of life are distributed throughout the Galaxy, there seems to be no logical reason why they should have awaited arrival on Earth before self-assembly: if the deciding factors are mass, range of environments, timescale and ease of dissemination, the Galaxy is an overwhelmingly more likely candidate for abiogenesis.

The question of how life originated is still a matter of much speculation. The leap from an assembly of amino acids to proteins to prokaryotes is huge and not yet understood. Panspermia does not directly address this question, but it clearly has a major bearing on the issue.

5.3 Implications for the cosmic evolution of life

For significant evolution of life to take place on a planet (for example to the extent where intelligence might appear), rather than simply the instant amplification and dispersal of incident microorganisms, a planet needs to remain within a habitable zone for something like a billion years at least. Jones et al (2006) modelled the 152 exoplanetary systems known by April 2006 to examine whether Earth-mass planets could be present and survive in the habitable zone, or whether the proximity of giant planets would eject them or prevent their formation. They found that about half the systems examined could contain Earth-mass planets for this order of time provided, of course, that such planets could be created in the protoplanetary disc. Whatever impediments there may be to the development of multicellular or intelligent life, it seems the condition of prolonged habitability is not one of them. Probably, the Galaxy is a fantastic zoo (cf Joseph & Schild 2010b).

The idea that comets may spread genes within and between adjacent galaxies, placing them in stable, long-lived habitats, also impinges on many of the 'big questions' of astrobiology and cosmology such as the Goldilocks enigma (Davies 2007), anthropics and the lifetimes of civilizations (Cirković et al 2009), SETI and the Fermi paradox, and even Darwinian evolution on Earth (Hoyle & Wickramasinghe 2000; Joseph 2000, 2009; Gibson & Wickramasinghe 2010). Discussion of these issues lies beyond the scope of the present paper.