Документ взят из кэша поисковой машины. Адрес оригинального документа : http://star.arm.ac.uk/preprints/409.ps Дата изменения: Wed Oct 29 19:14:51 2003 Дата индексирования: Mon Oct 1 21:29:11 2012 Кодировка: Поисковые слова: star system

Mon. Not. R. Astron. Soc. 000, 000{000 (0000) Printed 1 October 2003 (MN L A T E X style le v2.2)
A Mechanism for Interstellar Panspermia
W.M. Napier ?
Armagh Observatory, College Hill, Armagh, BT61 9DG, Northern Ireland
Last update 2003 September 29; in original form 2003 March 20
ABSTRACT
Metre-sized boulders ejected from the Earth by large impacts are destroyed through
collisions and erosion by impacting zodiacal cloud dust particles. The timescale for such
disintegration in a dense zodiacal cloud may be <
 10 4 yr. Once reduced to a critical
size, the particles are rapidly ejected from the solar system by radiation pressure.
The critical size for ejection is of order a micron, large enough to protect groups of
micro-organisms within them from solar UV irradiation. Such life-bearing particles
are ejected at a mean rate of 10 20 per million years. During passages of the solar
system through or close to dense molecular clouds, a signicant proportion of the
particles may be incorporated into protoplanetary systems and protected from cosmic
rays within growing planetesimals. The specic number density of micro-organisms so
deposited is highest in small, dense molecular clouds. On the assumption that this
ejection mechanism is common in other planetary systems environmentally capable of
supporting life, a `chain reaction' may seed the disc of the Galaxy within a few billion
years. In that case it is unlikely that life originated on Earth.
Key words: astrobiology - interplanetary medium - molecular clouds - comets: gen-
eral - ISM: clouds
1 INTRODUCTION
In his presidential address of 1871 to the British Association,
Lord Kelvin suggested that impacts between planetary bod-
ies might scatter life-bearing meteoric stones through space
and on to the surfaces of other worlds, thereby propagat-
ing life between the stars (Thomson 1871). The basic con-
cept goes back at least to William Herschel, who suggested
that interstellar bodies may spread or sustain life, by ac-
creting on to the surfaces of stars like the Sun (see Schafer
1977). Interest in this hypothesis was revived following the
(disputed) claim that microfossils are present in the me-
teorite ALH 84001 of probable Martian origin (McKay et
al. 1996). Large land impacts may throw meteorite-sized
fragments of rock from the inner planets into interplane-
tary space (Melosh 1988, Gladman et al. 1996), and a sig-
nicant exchange of boulders between Earth, Mars and the
Moon has occurred throughout geological history. Boulders
more than 20 cm across, ejected from the topmost layers
of an impact site, are probably never heated to more than
100 ф C in their interiors during the few seconds' ight time
from ground to space, while bacteria seem able to survive
the accelerations involved (Mastrapa et al. 2001). Thus bac-
teria within such boulders will survive ejection into space.
Over 4 Gyr there may have been 40 land impacts on Earth
producing craters over 60 km in diameter, yielding in total
? Aфliated to Centre for Astrobiology, Cardi University
about 40 billion such boulders (loc. cit.). In the 10 million
years following an impact, there is a roughly constant deliv-
ery rate of meteorites between Earth and Mars. Conversely,
about 15 Martian meteorites may currently be falling on the
Earth each year (Gladman 1997). Lord Kelvin's conjecture
thus seems to have been vindicated by recent work: viable
micro-organisms may be exchanged between the planets of
the inner solar system.
Transfer of micro-organisms between stellar systems by
this mechanism is, however, a more formidable prospect.
Melosh (2003) has followed the fate of boulders ejected from
the Earth until they collide with a planet, fall into the Sun
or are ejected from the solar system. He found that about
15 metre-sized boulders originating from the inner planets
are ejected from the solar system every year. Their mean
residence time within the solar system is about 50 Myr, al-
though with a very wide dispersion. The mean ejection speed
is 53 km s 1 . With these gures, Melosh nds that only one
meteorite ejected from a planet within our solar system is
likely to have been captured into another stellar system in
1,000 Myr and that, having been so captured, there is only a
probability 10 4 that it will land on a terrestrial planet in
that system (if it has one) within another 4.5 Gyr. Further-
more, according to Mileikowsky et al. (2000), viable micro-
organisms decline in numbers exponentially due to destruc-
tion by galactic cosmic rays. Characteristically, a population
of 10 8 micro-organisms is reduced to  100 or less after a
million years. Thus the boulders will in any case be sterile
c
0000 RAS

2 W.M. Napier
by the time they are ejected from the solar system, even
before they have begun their interstellar voyage of several
billion years' duration.
Thus two factors, (a) the extremely long residence and
travel times of boulders in interplanetary and interstellar
space, and (b) the extremely low probability that any ejected
boulder will ever impact on another terrestrial planet, seem
to ensure that `lithopanspermia' as currently envisaged is
a local mechanism. In the absence of other transfer mecha-
nisms, the Earth and any other life-bearing planets in the
Galaxy would seem to be biologically isolated from each
other.
In the present paper, however, I show that these life-
bearing boulders may be destroyed by erosion and fragmen-
tation within a few thousand years, especially when there is a
large, active comet in the inner planetary system to enhance
the mass of the zodiacal cloud. Their disintegration proceeds
until the boulder fragments become -meteoroids (for which
the repulsion due to sunlight exceeds the attraction due to
gravity) and are expelled from the solar system within a few
years or decades of their ejection from the surface of the
Earth, before solar UV or cosmic radiation can signicantly
damage the organisms within them. The mass so ejected is
an order of magnitude greater than that from the action
of gravity alone. However what matters is not so much the
mass expelled as the number of micro-organisms, essentially
codes for replication, expelled: a single micro-organism in
a suitable environment may, following Lord Kelvin, popu-
late a planet. Thus the expulsion of 10 14 particles/yr, each
with the potential to populate a planet, is a far more sig-
nicant event than the expulsion of the same mass in the
form of a few boulders/yr. The long travel times to other
star systems are obviated by the fact that, over its history,
the solar system has passed through about half a dozen gi-
ant molecular clouds and close to a few dozen dark cloud
complexes. During such passages the star factories within
these nebulae are seeded with micro-organisms which have
been protected from UV radiation and not yet destroyed
by galactic cosmic rays. For reasonable assumptions about
the frequency of Earth-like planetary systems, life may thus
spread throughout the Galactic disc in less than the age of
the Galaxy.
2 BOULDER EJECTION FROM THE EARTH
Boulders ejected from the Earth into heliocentric orbits are
subject to mean motion and secular resonances, and long-
term drift due to perturbations of the orbits by radiation
pressure (the Yarkovsky eect: Bottke et al. 2000). Monte
Carlo codes which neglect these eects yield median lifetimes
of tens of millions of years for near-Earth asteroids before
they strike a planet or fall into the Sun (Melosh 2003). In-
tegrations with these codes reveal that the Earth and Mars
are themselves incapable of hyperbolically ejecting boulders.
Melosh (2003) nds that about 15% of boulders ejected from
the Earth attain Jupiter-crossing orbits within 100 Myr, and
are thereafter ejected quite rapidly. The inclusion of the
above eects may change this gure by a factor of a few,
but still an enormous timespan remains during which boul-
ders would reside in the inner solar system, subject to solar
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
0.0035
0.004
0.0045
25 30 35 40 45 50 55 60 65 70 75
f
Impact speed (km/sec)
Figure 1. Mass me of ejecta thrown out from Earth (a fraction
f of the impactor mass m) as a function of impact speed v i .
are irradiation, erosion and fragmentation, before they are
expelled into interstellar space.
According to Armstrong et al. (1989), the mass mej of
material expelled by spalling of the target rocks on Earth, at
speeds greater than escape velocity, ve , is given as a fraction
f of the projectile mass m by
f = mej
m
= 0:75 Pm
cLv i
[(v i =2ve ) 5=3 1] (1)
Here Pm represents the maximum pressure to which the
debris (boulders) are subjected,  is their density and CL
is the speed of sound through the rock. v i is the impact
velocity of the projectile. Fig. 1 gives the survivable frac-
tion as a function of impactor velocity, employing the rock
data of Mileikowsky et al. (2000) (only material subject to
shock pressures to less than 10 10 dynes cm 2 is considered
as otherwise shock heating would destroy organisms within
the micro-pores of the rock). The equation has been derived
under a number of conservative assumptions (loc. cit.). In
particular it was assumed that ejecta must push their way
through an unperturbed atmosphere: in reality, small frag-
ments will be driven upwards through a cone evacuated by
a large incoming bolide (Wallis & Wickramasinghe, pers.
comm.). For material to escape at all from the Earth it is
assumed that the impactor must strike the ground at more
than twice the Earth's escape velocity.
Mean impact velocities of the various classes of near-
Earth object have been computed by Jeers et al. (2001).
Their essential results are summarised in Table 1. It seems
that over a 10 Myr period one may expect about 50 impacts
of bodies (say) with masses in excess of 10 16 g. The Halley-
type comets peak strongly at 70km s 1 . At these highest
speeds, the contribution from dormant Halley-type comets
is signicant. Their existence is largely inferred rather than
observed, however (Bailey & Emel'yanenko 1998) { strong
selection eects operate against the discovery of dark objects
with 200 yr orbital periods (Jewitt & Fernandez (2001)
{ but it is also possible that long-period and Halley-type
comets disintegrate rather than become dormant (Levinson
et al. 2002). The intermediate disintegration products could
include many hazardous cometary meteoroids. The propor-
c
0000 RAS, MNRAS 000, 000{000

Panspermia mechanism 3
Table 1. Total impact probabilities and mean impact speeds 
v i
for various populations: near-Earth asteroids (NEA) >1km diam-
eter; short-period comets (SPC); Halley-type comets (HTC); and
long-period comets (LPC). After Jeers et al. (2001). The letter
(d) refers to dormant populations which are hard to detect: their
numbers are derived from considerations of population balance
and are very uncertain.
impactor impacts 
v i
type per 10 Myr km/s
NEA (> 1 km) 29 22
SPC 1 20
SPC (d) 14 20
HTC 0 57
HTC (d) 3 57
LPC 3 54
tion of comets which end up as dormant bodies is at present
an unresolved issue (Bailey 2002, Rickman et al. 2001). In
general there are still many uncertainties in this area and
the true impact rates of bolides in the various categories
may be higher or lower than indicated by factors of at least
two.
From Fig. 1 and the mean impact speeds of Table 1,
we adopt a mean f = me=m 210 3 . Two thirds of im-
pacts occur in the oceans and so an overall rate of one fast
land impact per million years is expected. Over any given
10 Myr period, for a population distributed as n(m) / m  ,
where the population index  1.7-2.0 down to a minimum
mass , one typically nds a mean impactor mass in the
range 
m 2.0-3.5. For  10 16 g, the total mass of boulders
ejected from the Earth at survivable temperatures is then

mf 400-700 million tons over a 10 Myr period. Of course
the longer the interval considered the greater the probability
of exceptionally large impact, and the interval 10 Myr has
been chosen to represent, conservatively, the characteristic
time which the solar system has spent in the environment
of star-forming nebulae (Section 4). With impactor speeds
of order 50 - 60 km s 1 rather than the 30 km s 1 adopted
by Mileikowsky et al. (2000), hyperbolic ejection speeds up
to 13 km s 1 may occur and so relatively few ejecta will
be recaptured on subsequent passes of the Earth (Gladman
et al. 1996).
3 FATE OF EJECTED BOULDERS
On leaving the Earth a boulder joins the zodiacal cloud
and is subject to erosion by impacting dust and fragmen-
tation by collisions with meteoroids. `Dust' conventionally
refers to particles with radii less than 0.1 mm (100 m), `me-
teoroids' to the larger bodies. About 95% of the zodiacal
light is due to dust particles. This size also corresponds to a
fairly sharp transition at 1 AU between radiation-dominated
and collision-dominated dynamics (Grun et al. 1983): below
100 m, the lifetimes of particles are limited by inspiralling
due to the Poynting-Robertson eect; above it, they are lim-
ited by collisions. For a 1 mm meteoroid close to the orbit of
the Earth (semi-major axis a=1 au, eccentricity e=0.1), the
timescale for disruption against collisions is, in the present
zodiacal cloud, c=40,000 yr (Steel & Elford 1986). Erosion
timescales e are at least an order of magnitude longer.
The mass of the current zodiacal cloud, counting up
to particles of 0.1 mm radius, is given by Leinert et al.
(1983) as 10 17 g, broadly in agreement with the dust im-
pact rate measured with the Long Duration Exposure Fa-
cility (Love & Brownlee 1993). If the count is extended to
include meteoroids up to 100 g, however, the mass of the
zodiacal cloud has been variously estimated as 2:5  10 19 g
(Whipple 1967) and 3:0  1:0  10 20 g (Hughes 1996). How-
ever, over timescales of order 1 Myr, the zodiacal cloud is
subject to strong, random surges in mass caused by the ar-
rival of exceptionally large comets into short-period orbits
followed by their disintegration (Napier 2001). Simulations
allowing for both collisional disintegration and Poynting-
Robertson drift indicate that the cloud may, following the
entry of a Chiron-sized object into the inner planetary sys-
tem, reach a mass two or three powers of ten higher than its
current value for several millennia. When this happens the
Poynting-Robertson lifetime is unchanged but the lifetimes
against erosion and fragmentation are reduced by 2{3 or-
ders of magnitude: the system becomes collision-dominated
down to 10m. Over a 1 Myr period, therefore, there are
likely to be several epochs when ejected boulders are subject
to rapid destruction.
Consider a boulder of bulk density  b and radius r im-
mersed in zodiacal cloud particles whose density in space is
z and which strike the boulder at V km s 1 . In time
t the
boulder loses a mass
m, due to erosion, given by

m = r 2 zV
t (2)
and its radius decreases by
r obtained from

m = 4 b r 2
r
Here = me=mz where me is the mass of boulder excavated
by a colliding particle of mass mz . Thus

r = 1
4 (z= b )V
t (3)
and so
r = r0 kt (4)
The radius of the boulder thus decreases linearly until, in
the absence of a catastrophic disruption, it would disappear
in a time e = 1=k given by
e = 4r b =(V z ) (5)
The excavation factor { i.e. the mass excavated in units
of the projectile mass { is found experimentally to vary as
V 2 and has a value, for medium-strength rock with impact
speed 10 km s 1 , 510 4 (Grun et al. 1983). The ero-
sion rate V / V 3 . Consider a metre-sized boulder with
 b =2.5 g cm 3 (mass 1.3 tons) injected into the current zo-
diacal cloud. Then in the absence of fragmentation, eqn (5)
reveals that the rock would be destroyed by erosion within
e 19000 yr if the cloud is taken to have mass 3  10 20 g
(Hughes 1996), or 230000 yr if we adopt the lower zo-
diacal cloud mass of Whipple. With a zodiacal cloud en-
hanced in mass by a disintegrating large comet, however,
both number density and the mean encounter speed with an
expelled boulder become signicantly higher. The enhanced
encounter speed comes from the higher mean eccentricity
c
0000 RAS, MNRAS 000, 000{000

4 W.M. Napier
of particles derived from a short-period comet, before the
Poynting-Robertson eect has had time to circularise their
orbits. An enhancement by a factor 10 in cloud mass and 1.5
in particle velocity (Napier 2001, Fig. 5), reduces the erosion
time to 560 yr and 6,800 yr respectively for boulders of this
size.
When about half the mass of a metre-sized boulder has
disappeared through erosion, it becomes vulnerable to frag-
mentation with cm-sized cometary meteoroids, and it joins
the collisional regime of the zodiacal cloud as a whole. Napier
& Dodd (1974) found that, in an environment where imping-
ing particles have a power law distribution with population
index  1.8, the fragmentation timescale c for a basalt
target is an order of magnitude shorter than the erosion one
e (Napier & Dodd 1974). It follows that the survival time
of a boulder injected into an Earth-like orbit is very short,
much less than 10 4 yr, during a period of enhanced giant-
comet activity in the inner planetary system. A hierarchy of
fragmentations reduces the boulder to dust whose ultimate
fate (apart from planetary collisions which are a minor fac-
tor) is either infall to the Sun, or ejection from the solar
system, both processes being radiation-driven.
Once the particles have become `dust' (< 100 m in ra-
dius) then 50% of the mass excavated by collisions is im-
mediately expelled as -meteoroids, which are accelerated
out of the solar system (Leinert et al 1983), although even
before reaching this size, an increasing proportion of parti-
cles produced during erosive collisions may be expelled by
radiation pressure.
Putting  = PR=g (PR the solar radiation pressure
on a grain and g the gravitational acceleration), a silicate
sphere of mass 10 10 g (4 m diameter) has  0.1, increas-
ing to  0.5 at 10 13 g or 0.4 m diameter, before declin-
ing to  0.1 at 10 15 g (Ishimoto et al. 1993). Irregularly
shaped grains in this size range may attain  values in excess
of unity. In this size range the Poynting-Robertson timescale
is 1000 yr and probably comparable with or shorter than
the timescale for collisions.
Unshielded micro-organisms will not survive expo-
sure to solar ultraviolet radiation, but clusters of micro-
organisms, imbedded in micropores of rock, may be ade-
quately shielded: a layer of graphite 0.024 m thick has
optical depth  3 at 2200 A, enough to protect the or-
ganisms within (Wickramasinghe 1967). Wickramasinghe &
Wickramasinghe (pers. comm.) have pointed out that bac-
teria in clusters will, if exposed to ultraviolet light, develop
a thin carbonised outer skin (0.03 m thick) adequate to
shield interior organisms. Thus groups of micro-organisms
within -meteoroids may be self-shielded in their interiors
from ultraviolet light.
In summary, during an episode of enhanced zodiacal
cloud mass, by a factor of ten to a hundred over the present,
the boulders ejected from Earth are destroyed in situ, be-
coming eroded or broken down to dust in a few thousand
years. At 1 AU, the timescale c for destruction by collisions
is less than that for infall to the Sun PR at all sizes down
to about 10 m, so that the dust particles remain in situ, at
least half their mass then being expelled as -meteoroids in
the course of fragmenting collisions. The true proportion of
mass so expelled is probably higher as losses occur during
every eroding or fragmenting impact. It is likely that a large
proportion of the expelled microbes are well protected from
damaging ultraviolet radiation.
Moreno (1988) has suggested that sub-micron debris
containing micro-organisms, and ejected from the Earth by
impacts, may be driven to Mars by solar radiation pressure.
We consider here the fate of small particles thrown into in-
terstellar realms.
4 PASSAGE THROUGH A DENSE
INTERSTELLAR CLOUD
A mean annual supply of 10 tons of life-bearing boulders,
collisionally ground to -meteoroids of (say) 1 m radius,
yields 10 18 such particles/yr. For comparison the extremely
common microbe staphylococcus has radius 0.125 m, the T1
bacteriophage 0.03 m (Secker et al. 1994). A gram of rich
soil contains typically 10 9 micro-organisms, rock presumably
at least one or two orders of magnitude less, depending on
porosity, location and so on. Most of the lightly-shocked
boulders ejected from Earth come from within a few metres
of the surface due to interference and cancellation of shock
waves near a free surface (Melosh 1988). We assume 10 8
micro-organisms per gram of ejecta, but clearly this gure
will vary by orders of magnitude from one impact to another
depending on accident of location and epoch.
The -meteoroids ejected from the solar system will be
concentrated in a number of expanding shells corresponding
to discrete past impacts. The terminal velocities of the shells
are given by
v t =
p
(2 0 =r) (6)
where r 1AU is the distance at which the -meteor was
formed and  0 =r 2 is the net repulsive acceleration of the me-
teor due to the excess of solar radiation pressure over gravity.
Thus for -meteoroids with radiation force 10% in excess of
the gravitational one, v t =13km s 1 and the shell has moved
1.9 pc from the solar system in 140,000 yr; for meteoroids
with radiation force 40% over gravity, v t =27 km s 1 and the
shell has moved 1.9 pc from Earth in 70,000 yr. A meteoroid
with solar repulsion four times greater than gravity will at-
tain a terminal speed v t 85km s 1 out of the solar system
and travel 6 pc in this time. A 1 m grain with a graphite
coat 0.02 m thick is an example (Wickramasinghe & Wick-
ramasinghe, pers. comm.).
Exposed to Galactic cosmic rays, the viable microbe
population declines exponentially, at a rate depending on the
specic organism. From the examples given by Mileikowsky
et al. (2000), half-lives of 50,000{100,000 yr seem character-
istic. This kill rate includes the eect of irreparable dam-
age to the DNA. Adopting a half-life of 75,000 yr, then in
say 100,000 yr almost half the ejected organisms will still be
alive, albeit dormant. For an ejection of n0 microbes each
year, the solar system is surrounded by an equilibrium `bio-
sphere' of n0 t 1=2 810 20 living micro-organisms, extending
out to 5 pc. The half-life adopted may, however, be highly
conservative (Wickramasinghe, pers. comm.)
In the situation where the Sun is passing through a
dense nebula, the distances travelled by the -meteoroids
allow them to penetrate small molecular clouds, or clumpy
structure within giant molecular clouds, in timescales com-
c
0000 RAS, MNRAS 000, 000{000

Panspermia mechanism 5
fortably less than the half-lives for destruction of unshielded
micro-organisms by galactic cosmic rays.
The sun has penetrated giant molecular clouds (mean
mass M3{1010 5 M , mean radius R20 pc) about
5 times in the last 4 Gyr. At an encounter speed
V 20km s 1 , the mean passage time through a GMC is
4=3(R=V ) 3 Myr. Thus if the impact rate on Earth is
unchanged from its current value the GMC is infected with
terrestrial micro-organisms at a rate of 10 7 g yr 1 , 10 8 micro-
organisms g 1 , for 3 Myr, yielding a deposition per passage
of 310 21 micro-organisms.
A key feature of an encounter with a GMC is that the
Oort cloud becomes gravitationally disturbed, generating a
comet shower. Infall times of 2 - 3 Myr correspond to comets
with semi-major axes in the range 25,000 - 33,000 AU so that
the shower may peak while the Sun is still immersed in the
GMC. Such encounters will take place when the Sun passes
through the spiral arms of the Galaxy, where GMCs are
concentrated. Detailed modelling of such a prolonged dis-
turbance of the Oort cloud has not been carried out at the
time of writing, but it seems inevitable that there will be
at least an order of magnitude increase in the ux of long-
period comets during the penetration. In that situation (see
Table 1) terrestrial bombardment may become dominated
by high-speed impactors, if Oort cloud comets are indeed
signicant contributors (directly or indirectly through re-
plenishing the Halley comet reservoir).
In addition to the GMC penetrations, the Sun may pass
close to a larger number of dark cloud complexes (DCCs) of
smaller dimensions. Low-mass stars are formed in both types
of cloud, but GMCs are the primary creation sites for stars
of high mass (e.g. Mundy 1994). GMCs have a clumpy in-
ternal structure whose mass distribution can be represented
as a power law with population index  = 1:60.2 (Mundy
1994). The DCC population may likewise be so tted: the
mass spectrum of over 1500 clouds in the Perseus Arm has
index 1.75 (Heyer & Terebey 1998). This t extends down to
nebulae with masses as low as 100 M and suggests that
both DCCs and GMCs may be part of the same population
(Blitz & Williams 1998). The solar system passes within d pc
of such nebulae at intervals
tMyr given by

t  800(M=5  10 5 ) 0:75 (d=20) 2 (7)
assuming the mass of molecular gas in the Galactic disc is
unchanged over the lifetime of the solar system.
From (7) the solar system passes within 5 pc of a 5,000
M nebula every 400 Myr or so, and a 1,000 M nebula ev-
ery 120 Myr. Biologically active material may therefore be
deposited in passing star-forming nebulae at relatively fre-
quent intervals in geological terms. The characteristic radii
of nebulae in this mass range is one or two parsec.
Particles scattered into a dark cloud are preferentially
absorbed into dense regions where the collision rate with
nebular material is higher. Mantle accretion and grain
growth occur in dense molecular cores (e.g. Pollack et al.
1994) and it is to be expected that the injected particles
will take part in these processes. There is direct evidence
that many comets have formed at temperatures which put
them in the trans-Neptunian zone of the classical solar neb-
ula, or even in a molecular cloud environment (Mumma
1996; see also the review by Napier & Clube 1997 for refer-
ences). Large, fragile interplanetary dust particles of prob-
able cometary origin have D/H isotope ratios approach-
ing those of molecular clouds, indicating that molecular
cloud material has been incorporated intact into comets
(Messenger 2000). Thus micro-organisms may be incorpo-
rated directly into cold, growing cometary masses. Once so
incorporated, they are protected from further destruction
by Galactic cosmic rays. The growth of comets may take
place on timescales as little as 1000 yr (Hills 1982, Napier &
Humphries 1986).
Distributing 310 21 micro-organisms amongst 10 39 g of
molecular material in a GMC yields one micro-organism
per 10 16 g of nebular dust, assuming a dust-to-gas ratio
of 3%. Consider the eect of this bio-material on a hypo-
thetical planetary system identical to our own. At present
40,000 tons of dust fall on Earth annually. The smaller dust
particles at least are braked gently by the current atmo-
sphere and micro-organisms within them would survive ash
heating during the fall (e.g. Hoyle & Wickramasinghe 2000,
Coulson 2003). This ux has probably been constant over
the past 3 Gyr, and if the current value is an `average' then
1.210 20 g of cosmic material has fallen on Earth over that
period, permitting 10000 dormant microbes to reach the
ground, assuming the infalling material is relatively unpro-
cessed material. However, the lunar cratering rate shows the
mean impact rate to have been at least two or three powers
of ten higher over the rst 500 Myr of the record (Neukum &
Ivanov 1994) or 210 21 {210 22 g This would imply the fall
of some 210 5 {210 6 microbes on to the primordial Earth.
The conditions would possibly be too harsh for life to take
a hold, but as the cratering rate declined, there would be a
transition to a time where incoming microbes could survive
and replicate.
The mass M of molecular clouds  varies with cloud
radius R as M / R 2 over at least eight decades of mass
(Elmegreen & Falgarone 1996), whence small clouds are sys-
tematically more dense. As a result, for a dark cloud of
mass 1000 M passing through the biosphere and sweeping
up micro-organisms, the resulting microbe concentration is
readily found to be an order of magnitude higher than that
in a GMC. Infall on to hypothetical early Earths forming
within such a cloud is correspondingly higher. Since DCCs
are the sites of low-mass stars and close encounters with
them are an order of magnitude more frequent than with
GMCs, it seems that these dark cloud complexes may be
prime sites for the replication of life throughout the Galaxy.
5 DISCUSSION AND CONCLUSIONS
5.1 Earth to Galaxy
The mechanism described herein bridges the lithopansper-
mia of Lord Kelvin and the cometary panspermia of Hoyle
& Wickramasinghe (2000), and indeed one may follow from
the other. The largest uncertainty resides in the assumption
that solar systems environmentally capable of supporting
and disseminating life are common throughout the Galactic
disk. However Europa and Mars have both been considered
as possible sites for archaean or current life, and the exis-
tence of at least two other candidates for life in the solar
system suggests that biofriendly planets may be common
throughout the Galaxy. To populate 10 10 suitable planets
c
0000 RAS, MNRAS 000, 000{000

6 W.M. Napier
within the lifetime of the Galaxy would require about 33
generations with a doubling time of about 300 Myr, and,
neglecting the contribution from DCCs, this would require
inoculation of something like a dozen planetary systems dur-
ing each GMC encounter. Since an OB association may typ-
ically contain 10 3 {10 4 T Tauri stars, this requirement seems
reasonable. Within the molecular ring of the Galaxy (3{
8 kpc from the nucleus, peaking at 5.5 kpc), the number
density of clouds is 5 or 6 times higher than the local den-
sity (Blitz & Williams 1998). In that case a microbe-losing
star system would penetrate DCCs every few million years,
and GMCs every 50 Myr, so providing an environment in
which the reaction could propagate in no more than a few
Galactic rotations.
5.2 Galaxy to Earth
On this hypothesis the Earth is part of a chain reaction
whereby life has spread throughout the Galaxy, taking root
and replicating wherever suitable environments occur: if the
Galaxy were initially sterile, it could have been inoculated
by transmission of life from Earth. However the fact that
terrestrial life is part of this chain reaction does not mean
that it initiated the reaction, and simple probability argues
against it. Thus on the present hypothesis the origin of ter-
restrial life should be sought elsewhere, in the Galaxy or
beyond.
ACKNOWLEDGEMENTS
The author is indebted to Chandra Wickramasinghe and
Max Wallis for discussions on this topic, and to David
Hughes for suggesting many improvements to the paper.
REFERENCES
Armstrong, J.C., Wells, L.E. & Gonzalez, G., Icarus, 2002, 160,
183
Bailey, M.E., 2002, Science, 296, 2151
Bailey, M.E. & Emel'yanenko, V., MNRAS, 298, 212
Blitz, L. & Williams, J.P., 1998, in The Origin of Stars and Plan-
etary Systems, eds. C.J. Lada & N.D. Kylas, Kluwer, Dor-
drecht, p.3
Bottke, W.F., Rubincam, D.P. & Burns, J.A., 2000, Icarus, 40, 1
Coulson, S.G., 2003, Dynamical and radiative interactions be-
tween interplanetary objects and an atmospheric aerosol, the-
sis
Elmegreen, B.G. & Falgarone, E., 1996, ApJ, 471, 816
Gladman, B.J., Burns, J.A., Duncan, M. et al., 1996, Science,
271, 1387
Hills, J.G., 1982, AJ, 87, 906
Hoyle, F. & Wickramasinghe, N.C., 2000, Astronomical Origins
of Life, Kluwer, Dordrecht
Hughes, D.W., 1996, QJRAS 37, 593
Ishimoto, H., Mukai, T., Kimura, H. & Maruyama, K., 1993,
in Meteoroids and their Parent Bodies, eds. J. Stohl & I.P.
Williams, Astron. Inst., Slovak Acad. Sci., Bratislava, p.65
Jeers, S.V., Manley, S.P., Bailey, M.E. & Asher, D.J., 2001, MN-
RAS 327, 126
Jewitt, D. & Fernandez, Y., 2001, in Collisional Processes in the
Solar System (eds. M. Ya. Marov & H. Rickman), 143-161,
Kluwer, Dordrecht
Leinert, C., Roser, S. & Buitrago, J., 1983, AA, 118, 345
Levison, H.F., Morbidelli, A., Dones, L. et al., 2002, Science, 296,
2212
Mastrapa, R.M.E., Glanzberg, H., Head, J.N., Melosh, H.J. &
Nicholson, W.L., 2001, EPSL, 189, 1
McKay, D.S. et al., 1996, Science, 273, 924
Melosh, J., 1988, Nature, 332, 687
Melosh, J., 2003, Astrobiology, 3, 207
Messenger, S., 2000, Nature, 404, 968
Mileikowsky, C., Cucinotta, F.A., Wilson, J. et al., 2000, Icarus,
145, 391
Moreno, M.A., 1988, Nature, 336, 209
Mumma, M.J., 1996, Nature, 383, 581
Mundy, L.G., 1994, in Clouds, Cores and Low Mass Stars, eds.
D.P. Clemens & H. Barvainis, ASP series Vol. 65, p.35
Napier, W.M., 2001, MNRAS, 321, 463
Napier, W.M. & Clube, S.V.M., 1997, Rept. Prog. Phys. 60, 293
Neukum, G. & Ivanov, B.A., 1994, in Hazards due to Comets
and Asteroids, ed. T. Gehrels, University of Arizona, Tucson,
p.359
Pollack, J.B., Hollenback, D., Beckwith, S. et al., 1994, ApJ, 421,
615
Rickman, H., Fernandez, J.A., Tancredi, G. & Licandro, J., 2001,
in Collisional Processes in the Solar System (eds. M. Ya.
Marov & H. Rickman) 131-142, Kluwer, Dordrecht
Secker, J., Lepock, J. & Wesson, P., 1994, ASS, 1
Schafer, E.H., 1977, Pacing the Void, University of California
Thomson, W., 1871, British Association for the Advancement of
Science, Presidential Address, 1871 Edinburgh meeting, p.26
Whipple, F.L., 1967, in The Zodiacal Light and the Interplane-
tary Medium, ed. J.L. Weinberg, NASA SP-150, Washington,
p.409
Wickramasinghe, N.C., 1967, Interstellar Grains, Chapman &
Hall, London
c
0000 RAS, MNRAS 000, 000{000