The expression ‘tree of life’ is shorthand for four billion years of birth, death, reproduction, and relatedness. This extended family tree has been produced by four billion years of using energy from the environment to power biological systems. At present these systems, with which our planet is teeming, seem unique in the vastness of the cosmos. But they’re not. Their apparent uniqueness is an artefact produced by current limitations to human knowledge. One day we will have evidence of life on other planets, and that day may be close at hand. It’s not unreasonable to believe that our first evidence of extraterrestrial life will arrive in the next couple of decades.
In this book, our starting point for thinking about life in an interstellar context is the nature of life on Earth. Here on our home planet one particular tree of life has played out. This tree will continue to grow, though the directions in which its still-ungrown branches will extend are impossible to predict, so we cannot look with clarity into our evolutionary future. But we most certainly can examine our evolutionary past. And we can ask to what extent we would expect major features of that past to apply to trees of life that are playing out independently of ours – right now – on planets scattered across the Milky Way galaxy and beyond.
Notice that ‘tree of life’ is in the singular in the context of our own planet. Every living creature on Earth is related to every other. We humans are not just related to chimps, gorillas, and orang-utans. We are also related to the rest of the animal kingdom and, beyond that, to the trees we climbed as children, the yeast we use to make bread, and the bacteria that line our guts. The branches of the tree of life have no breaks in them. If we made a three-dimensional model tree of this kind, it would be possible to run a finger down from one terminal twig, such as humans, to a particular ancestor in the distant past, and then back up again to any other present-day twig, for example a maple tree.
But what shape should we choose when building our model? In other words, what shape characterizes the overall tree of life on Earth? It has been depicted in many ways since Darwin sketched an evolutionary tree diagram, in the form of lines gradually diverging from each other, in Chapter 4 of The Origin of Species. There are several caveats here, because the shape of the tree of life – or of parts of it – has been a source of heated argument among biologists over the years. So we need to tread carefully.
First, scale may be important. Let’s consider this in terms of the two-dimensional trees that have been drawn on pieces of paper ever since Darwin. The shape of one small branch and the shape of the overall tree may not be the same. Second, at any scale we choose to examine, the divergence of branches may be leisurely (picture a V) or rapid (picture a U with a flat base). The former corresponds to a ‘gradualist’ view of evolution, the latter to either a ‘punctuationist’ or ‘saltationist’ view depending on the scale. Third, the vertical axis can represent time in an exact way, so that it could be labelled with units such as millions of years; or it could just represent time in a more general way in that it shows only the order of branching events, not their relative distances apart. Fourth, the horizontal axis could represent ‘degree of difference’ or it could be there simply to allow us to picture divergences – something that can’t be done unless you have at least a two-dimensional diagram. The difference between these types of horizontal axis is that in the former case the distance apart of two twigs is a measure of their biological disparity, whereas in the latter it is not.
All the above four issues have been the focus of major debates at some stage in the history of evolutionary biology, and some of them continue to be debated. But the purpose of this book is not to examine such issues. We have a bigger picture to paint, so we’ll sweep these issues under the proverbial carpet and focus on something even more important – the question of whether a tree diagram of any kind is the right way to depict evolutionary relatedness in the first place.
Consider for a moment an actual tree, whether a maple, an ash, or an oak. If you inspect it carefully in winter when no foliage obscures its branches and twigs, what you’ll see are thousands of divergences but not a single convergence. Twigs grow apart from each other; they do not grow together and unite. But in the tree of life such growings-together do indeed happen to a degree. Two processes are responsible – interspecies hybridization and horizontal gene transfer. In the former process, two twigs, each representing a single species, hybridize and thus create a descendant species that is different from both of its parents. In theory this shouldn’t happen, because a species is defined by its inability to interbreed with others – but in practice it does happen, because definitions are rarely perfect in the biological realm. In the latter process, DNA (deoxyribonucleic acid) from one twig is transferred into another, often via a virus.
The importance of these processes varies according to position in the tree. In the animal kingdom as a whole, their role is minor compared to twig divergence – though that does not mean that they aren’t important. Some human genes appear to have originated by horizontal transfer from other species, including those as different from us as bacteria. Some of the best examples of interspecies hybridization come from the plant kingdom, while horizontal gene transfer is especially important in microbes.
How should we modify our picture of the tree of life to incorporate these two processes? Hybridization can be included simply by picturing twigs growing together – at least within some of the tree’s branches. Horizontal transfer is probably better pictured as a sort of thin wire connecting two twigs at the same level (i.e. the same point in time). Taking both of these modifications on board (Figure 1.1), we now have a tree of life that is still largely tree-like but with some additional forms of growth compared to a real tree. A 2018 book by John Archibald – The Tangled Tree – provides further discussion of this issue.
The final thing to say about our tree-of-life picture (or model) is that its top should be flat. It’s more like an African Acacia tree than a Norway spruce. This is because the present moment of time is the same for all the growing twigs, which collectively represent today’s biota – the animals, plants, and other life-forms that populate the Earth right now. Let’s take a look at this particular time-slice of the Earth’s biological history.
So, now we alter our angle of view of the tree from the side to the top. We hover over it as a kestrel might, to achieve the proverbial bird’s-eye view that we want. And we look at it as a photographer would when taking one of those shots where the foreground – in this case the present – is in sharp focus, and the background – in this case the past – is just a blur. We are then looking at a series of small circles, each one of them the tip of a growing twig. One is the human circle, another the bonobo circle, and so on. Species of cacti are represented by small circles far away from the ape ones. And mushrooms are represented by small circles far away from both of those other clusters.
Each circle is a species, though as we’ve already seen species can be badly behaved. The usual definition of a species is that while its members can breed among themselves none of them can breed with members of other species. And there is usually the proviso ‘in the wild’, so that we exclude information on what can happen in captivity, such as the production of ligers (lion–tiger hybrids). Of course, it would be naïve to expect all real organisms to conform to such a neat human concept. Some do, some don’t. But even those that don’t can be seen as fitting the definition in a probabilistic way – the density of reproductive interactions among members of a species is much higher than the density of such interactions between them and their sibling species.
Because there are at least a few million species on the Earth at present, and perhaps a few tens of millions, we need to have some way of structuring our knowledge of this vast biodiversity. And what better way than the method provided by the Swedish naturalist Carl Linnaeus in the mid-eighteenth century. Taking his approach, we group a bunch of neighbouring twigs together by drawing larger dotted circles around their small solid circles, thus representing groups of related species called genera (singular genus). For example, the orang-utan genus (Pongo) includes three twigs – those of the Bornean, Sumatran, and Tapanuli orangs (Figure 1.2). Our own genus (Homo) consists of only a single species in today’s fauna. In contrast, some genera – for example the insect genus Drosophila – have hundreds of species.
In this exercise of looking down from above on the growing tips of the tree of life’s twigs and drawing circles, we are doing something that can be described in terms of set theory. Our circles-within-circles picture is what a mathematician would describe as a large set that includes one or more smaller sets. But there’s something unique about our taxonomic sets: they are related to each other by their shared branches of the past. In a set of crockery types, where one subset is ‘cups’, another ‘plates’, and so on, there is no such underlying common ancestry – each item is made from scratch.
Taxonomists sometimes describe what they’re doing as discovering and describing ‘the pattern of natural classification’. The Linnaean approach draws bigger and bigger circles around progressively greater numbers of twigs, so that after species and genera we have families, orders, classes, and so on. Not only are these progressively more inclusive in terms of current biodiversity, but they are also progressively more deeply rooted in the tree of life. At the most inclusive end of the taxonomic hierarchy in Linnaeus’s scheme was the kingdom – still in use today but expanded in number. Linnaeus described just two kingdoms of life – plants and animals. Now we also recognize at least one more – the fungi – and almost all biologists would say that there are several others. For example, all the large conspicuous brown seaweeds that we observe around our coasts, including those that make up that wonderful marine habitat called the kelp forest, are outside of the plant, fungal, and animal kingdoms. Studies on their genes make this conclusion clear. Collectively they are brown algae, but beware the term ‘algae’ as it has many inconsistent usages. They are in a fourth kingdom, even though there is some debate over what its name should be.
For Linnaeus, above kingdoms of life there was simply ‘life’. But now we insert an even higher level of taxon than kingdoms – domains. The American microbiologist Carl Woese refined the taxonomic scheme of the earlier Carl in a major paper published in Reference Woese, Kandler and Wheelis1990. He grouped life-forms on Earth into the domains Bacteria, Archaea, and Eukarya. The first of these is self-explanatory, the last contains animals, plants, fungi, brown algae, and all other life-forms that are built of complex (eukaryotic) cells. The middle one, Archaea, was new, and based on earlier work by Woese and his colleagues. Superficially, the organisms that comprise Archaea look like bacteria, and pre-Woese they’d been classified as such. But, as he showed, they have a different form of RNA (ribonucleic acid) from the other two domains; and they use different fats in their cell membranes too. These are very deep-seated differences, and reflect their early divergence from both bacteria and eukaryotes.
It’s important to realize that all the taxonomic categories above species – from genera up to domains – are arbitrary and have no clear definitions. They simply constitute a useful way of organizing information. The species is the only category that has biological meaning for the entities that comprise it, as opposed to for human observers; hence our ability to define, albeit imperfectly, what ‘species’ means. But there’s an even more fundamental definition that we now need to consider – that of life itself.
If you’d like a lengthy discussion of this issue I can recommend the 2012 book of the same title by the organic chemist Addy Pross. Here we’ll focus on just two approaches, which I’ll call evolutionary and metabolic. The first is tightly linked to Darwinian natural selection. The second is linked instead to the biochemical processes that go on within cells. It’s quite possible to be alive by one definition but not by the other; indeed, that’s the case with viruses.
The evolutionary definition of life is as follows. Entities that exhibit the three properties of variation, reproduction, and inheritance are alive; those that don’t are not. These are the very same three properties that are necessary for natural selection to occur. Consider a group of entities – we’ll not prejudge the issue by calling them organisms just yet – that are rather similar but not identical to each other. They reproduce, in at least one of an immense variety of ways (beautifully discussed by Italian biologists Giuseppe Fusco and Alessandro Minelli in their 2019 book The Biology of Reproduction), and the offspring resemble their parent(s) more than they resemble randomly chosen members of the group. Resemblance in this context is not just external, nor just structural; it is internal and behavioural too. In such a situation, whichever variants are best suited to the current environmental conditions will leave most offspring, and so the composite nature of the population will change over time. Such a situation does not ‘give rise to’ natural selection – rather, it is natural selection.
At first this definition seems clear. According to it, birds and ferns are alive, while rocks and clouds are not. But if we dig deeper we find problems. Mules seem just as alive to me as do the horses and donkeys that were their parents. But, as sterile hybrids, they generally cannot themselves reproduce. Surely we shouldn’t leap to the conclusion that they are inert entities, non-life-forms. And the converse problem of an entity that has the three requisite properties to be considered life but that we generally do not think of as life can also be encountered – for example computer viruses. These can exhibit variation, reproduction, and inheritance, but most of us would not consider them to be alive. And what about real (biological) viruses – are these alive? Many biologists see them as inhabiting a philosophical grey area between the living and the non-living. They can reproduce, but not on their own without hijacking another living system to help them. Then again, the same could be said of a tapeworm. So the evolutionary definition on its own is problematic.
The metabolic definition of life goes something like this. An entity is alive if it takes up energy and materials from its environment, uses these to maintain an internal state that is dynamic and yet buffered to some extent from environmental fluctuations, and ejects waste products from this process back into the environment from which the raw materials came. For the most part, this definition classifies entities as alive or not in the same way as does the evolutionary one: birds and ferns are alive, rocks and clouds aren’t.
But again problems emerge when we start digging. The tiny invertebrate animals called tardigrades (or water bears) are famous for being able to withstand extreme conditions. They can survive extended periods of temperatures close to absolute zero, which would freeze-kill most other animals very quickly. They go into a state of suspended animation, from which they wake up when the ambient temperature is increased again. They use a similar technique to survive the vacuum of space; some of the tardigrades that have been taken into space – on the outside of a spacecraft rather than in the relative comfort of its interior – have survived and reanimated themselves on return to Earth, as reported by the Swedish scientist Ingemar Jönnson and his colleagues in Reference Jönnson, Rabbow, Schill, Harms-Ringdahl and Rettberg2008. Is a ‘cryptobiotic’ tardigrade alive? Personally I’d say yes, and that it’s just a rather extreme form of a hibernating hedgehog; but not everyone will agree with this view.
And what does being ‘buffered to some extent’ mean? We mammals can maintain an internal body temperature buffered into a narrow range around 37 degrees Celsius (98 Fahrenheit). Crocodiles can’t do that. Their internal temperature is much more variable over time – though it’s still buffered ‘a bit’ from the prevailing temperature of the environment. This more modest buffering is partly metabolic and partly behavioural in the sense of a crocodile’s choice of microhabitat.
What about the converse problem to that presented by the deep-sleeping tardigrade, in other words an entity that could be called ‘alive’ by the metabolic definition but which common sense would suggest is not? A fridge takes up energy from outside itself, uses this to maintain a regulated internal state, and ejects heat back into the environment as a sort of waste product. In this case we perhaps escape from definitional problems in that the fridge doesn’t take up materials from its environment. But even then a qualification is needed, because it does take up materials (cartons of milk, bottles of beer) – but not without human help, and it doesn’t use them to produce its internal homeostasis.
The metabolic definition could be modified by adding a stipulation that the ‘inside’ and ‘outside’ of the entity we’re looking at should be separated by one or more membranes – otherwise we conclude that the entity is inert. Again this works to distinguish birds and ferns from rocks and clouds. But is it too restrictive? Its use would remove viruses from their grey area and classify them as non-living, which to me seems too certain a conclusion for such hard-to-classify entities. And if we want to have a definition of life that we can apply to entities that we find on other planets, then it might be unwise to insist that we definitely won’t call them life-forms if they don’t have membranes, especially in our current state of ignorance as to what form life takes elsewhere than on our own planet.
So, how to proceed? I suggest that we adopt the following policy. If an entity is metabolically alive and membrane-bound, and groups of individual entities of this kind are characterized by variation, reproduction, and inheritance, then we describe the situation as ‘life’. If none of these criteria are met, the type of entity we’re examining is inert – i.e. not alive. Situations in between these two, such as viruses, we treat on a case-by-case basis. And regarding extraterrestrial life we should try to keep as open a mind as possible – we’ll revisit this issue in Chapter 3.
The Omnipresent Force
A common distinction made by biologists about evolution is between pattern and process. Evolutionary pattern is what is produced after long-term operation of the process. The tree of life is a pattern, but although its existence is good evidence for some kind of evolutionary process having occurred, it tells us little about the nature of the evolutionary process or the mechanisms underlying it. What forces have driven it? More than one, for sure, but what I’m calling here ‘the omnipresent force’ is Darwinian natural selection. In some sense, this is the most important force, though that seemingly simple statement is trickier than it initially appears.
It’s easiest to observe selection in action when the process happens quickly. And for this the best combination is a major threat to life (i.e. strong selection) and a short generation time. Hence one of the best examples of the power of selection is the evolution of antibiotic resistance in bacteria, now a major public health concern. In this situation, bacteria are facing a novel chemical designed to exterminate them. In a particular population of bacteria exposed for the first time to a new antibiotic, the mortality level may well be in excess of 99%. But, due to natural variation among individuals, a few may have inbuilt resistance to the threat, just because of some tiny difference in metabolism compared to all their relatives. In such a situation, the progeny of the resistant individuals will prevail. If many generations can unfold in hours or days rather than years or decades, evolution of the population to the point where it is 99% resistant as opposed to 99% susceptible will be very rapid.
Evolution of resistance to pesticides among insects is a similar phenomenon except that the typical insect generation time is much longer than the typical bacterial one, so the process takes a significantly extended period. Evolution of grasses and other plants to growing on the metal-contaminated spoil heaps surrounding old mine workings is similar too; as is the familiar textbook example of the evolution of certain insect species in response to smoke pollution caused by the Industrial Revolution. In the best-studied instance of this – the peppered moth – evolution from 99% peppered (camouflaged against the old lichen-covered clean tree trunks and branches) to 99% quasi-black (camouflaged against the novel sooty tree trunks and branches) took about a century.
Whether such examples constitute ‘natural’ selection is a moot point. It depends on the extent to which human-created agents of mortality are considered to be natural. But there is no clean line separating these anthropogenic environmental threats from others. Sometimes rapid environmental change can cause very strong selection without any human interference. It seems likely that this was the case when the asteroid that killed off the dinosaurs (and more than 75% of species of everything else too) crashed onto the piece of land that is now called Mexico. Although for most creatures mortality was complete, for the lucky few it was merely ‘very high’. The border between extinction and rapid evolution is a narrow one.
Most of the time, over most of the world, selection probably works quite slowly. But that doesn’t make it any less important – just harder to catch happening. When there are many geographical boundaries between small areas, in each of which selection favours different forms, we have an arena in which selection may be studied on what we might call a middling timescale – neither months nor millennia. Archipelagos constitute just such arenas, with the evolution of Darwin’s finches on the Galapagos archipelago being the best-known example. An early classic book on this system by British ornithologist David Lack has been supplemented by more recent books written by Princeton-based biologists Peter and Rosemary Grant.
Demonstrating natural selection in action as a driving force of evolution is one thing; demonstrating Darwin’s ‘theory of natural selection’ is quite another. In fact, we now need to ask: what exactly is this theory? To me, it is best summed up by Darwin’s own words at the end of the Introduction to The Origin of Species: ‘I am convinced that natural selection has been the main but not exclusive means of modification.’ How do we demonstrate not just that natural selection is one of the forces driving evolution but that it is the main one? Clearly, we must specify other such forces and somehow assess their relative importance. And we must acknowledge that what has been included in the ‘other’ category has changed since Darwin’s time. Now, in the early twenty-first century, the other forces of which we need to take account are: random genetic drift, gene mutation, and developmental channelling. Of these three, only genetic drift is a direct challenger to natural selection; the others are complementary to it, as we’ll see below.
Genetic drift and natural selection are both things that influence populations of organisms as opposed to individual organisms. In that respect, they’re similar. Where they differ is that natural selection is a systematic process producing a quasi-predictable result, at least in the short term, while genetic drift is a random process whose influence is indeterminate. But describing drift as a random process is not sufficient to give a clear picture of the nature of the beast. So let’s zoom in on it more closely.
A digression into coin-tossing is helpful at this point. If I toss a coin, the odds of it ending up heads are 50%. If I toss it five times, the odds of getting five heads are less than 10%. So usually a person tossing a coin five times doesn’t get this result. But if you have a large lecture theatre full of students and you ask them all to do this experiment simultaneously, some of them will get the magical five heads.
The same applies to the fate of variant versions of genes in populations of organisms. Suppose there is a gene called g, and in a particular population of animals there are two versions of it – g1 and g2. If one of them – say g2 – conveys enhanced fitness, it will tend to spread through the population by natural selection. However, if g1 and g2 don’t affect fitness, their relative frequency in a population will be determined by genetic drift. This is like coin-tossing. The probability of the frequency of g2 going up rather than down in one generation is 50% (assuming that its chances of remaining exactly the same are negligible). The probability of it going up in five successive generations is less than 10%. Yet if we have enough populations of the species concerned scattered across the globe, some of them will show exactly this outcome. And in a few of them g1 may be lost entirely, with g2 thus becoming ‘fixed’ in the population, despite the lack of natural selection.
In the early days of population genetics, this result of genetic drift was thought by most biologists to apply only to small populations. But the Japanese geneticist Motoo Kimura showed that drift could be important in large populations too. He devised ‘the neutral theory of molecular evolution’ (and published a book of the same name in 1983), according to which, at the level of macromolecules (DNA, RNA, and proteins), genetic drift may be responsible for most of the changes that we observe in natural populations. If this is true, Darwin’s theory of natural selection, at least in the form I’ve given it here, is wrong. Interestingly, although Kimura promoted this view, he also argued that Darwin’s theory was correct at the level of the whole organism. For example, he agreed with Darwin that giraffes evolved long necks because of natural selection, not because of genetic drift.
It seems paradoxical that drift could be the main agent of change at the molecular level, while at the same time natural selection is the main agent of change at the level of the whole animal (or plant or microbe). But actually it’s fine. If the vast majority of single amino-acid changes in a protein that is made up of 500 of them don’t affect its function, a predominance of drift at this level might be expected. But by definition none of those changes will affect the animal’s form. When we focus on morphology these neutral changes are invisible. The only molecular changes that are relevant to neck length in giraffes, or to body size and shape in animals more generally, are: (a) those that occur in genes that have an effect on body size/shape; and (b) those that affect the functioning of such genes and their products. These changes will all be subject to natural selection.
In my view it remains to be seen whether Kimura was right about the predominance of drift at the molecular level of organization. But regardless of this, Darwin’s belief that natural selection is the ‘main’ agent responsible for evolutionary change at the organismic level has been borne out by more than 150 years of observational and experimental evidence. Natural selection really is omnipresent in biological systems – providing that there is some variation on which it can act.
We tend to take biological variation for granted. It’s all around us. And the role that genes play in its generation can be seen from the natural experiment of identical twins. Although such twins are never truly identical, their differences in physical form are amazingly slight compared with those between two randomly chosen humans whose genetic make-ups are different. In humans and most other animals, a major player in reassorting genes and hence in influencing the nature of variation is sexual reproduction. But this is indeed just an influence – albeit a very large one. It only works when there are variant genes present in the first place. Furthermore, the very earliest life-forms probably didn’t have sex – at least as we know it. This is still the case with some microbes (and indeed some larger organisms) today. Ultimately, different versions of genes come from the type of molecular accident that we call mutation.
Mutation is a change in the sequence of the DNA that’s found in the genomes of all Earthly life-forms (excluding the RNA viruses, if we deem viruses to be alive). A typical gene has more than 1000 DNA bases arranged in a particular order. Mutation involves an accidental change in at least one of them – an accident that usually occurs when the DNA is being replicated. The fact that such changes occur at a background level in all biological systems is hardly a surprise. Rather, the surprise is that the background level is so low – think of the level of errors when someone transcribes a substantial body of text by hand. The rate of mutations can be radically increased by many factors, including exposure to carcinogenic chemicals or ionizing radiation such as x-rays. But such high rates aren’t necessary for natural selection to work – with only the background level it works just fine.
However important mutation is, we have to remember that it’s just a change in a gene. The step from there to changing the organism as a whole is complex – just how complex depends on the identity of the organism concerned. The step is longer for a mammal than it is for a bacterium. And in any multicellular creature, whether mammal, butterfly, or birch tree, it involves the developmental process. Exactly how this process channels mutational changes, and hence interacts with natural selection, is a fascinating, but still largely open, question – it was the subject of my 2004 book Biased Embryos and Evolution.
So, inherited variation is ultimately produced by mutation. The form that it takes is influenced (in different ways) by sexual reproduction and by the developmental system – in the case of organisms that have these things. Natural selection acts on the array of variants at its disposal and moves the average form of the population in a particular direction – by definition the one of higher fitness under the prevailing environmental conditions. Although there are many complexities glossed over here, this is the essence of the main force driving evolution. And it has been in the driver’s seat ever since life began. But it cannot in itself explain how life originated. So now we ask: how did that happen?
The steps from carbon atoms to life-forms are as follows: first, synthesis of very simple organic molecules such as methane; second, synthesis of more complex ones involving two or more carbon atoms such as alcohols; third, synthesis of the complex organic molecules that are found as repeating units in the macromolecules of life, including amino acids (found in proteins) and sugars (found in nucleic acids and carbohydrates); fourth, formation of the macromolecules themselves; fifth, formation of aggregations and interactions of such molecules together with smaller ones, thus giving rise to a sort of proto-metabolism; sixth, the development of quasi-autonomy from the environment through becoming membrane-bound; finally, the origin of a simple form of reproduction, perhaps involving the budding of smaller membrane-bound units from a larger parental one.
Of these seven steps, the first three are relatively ‘easy’ and occur all over the universe. They do not require a finely tuned planetary environment. Methane, alcohols, sugars, and amino acids have all been found in space. They arrive on Earth via meteorites, most of which have come from the asteroid belt. The making of these molecules is chemistry, not biology.
The last four steps still involve chemistry, of course, as do all processes involving matter of any kind – with the proviso that we can as yet only speculate about the nature of dark matter. But they also involve biology. Neither DNA nor proteins have ever been found beyond the Earth. Finding an amino-acid molecule in a meteorite that arrives on Earth tomorrow would not be news – except to the tabloid pseudo-press and its electronic equivalent. But finding a meteorite-transported protein molecule whose tangled form was made up of 100+ amino acids joined end-to-end would be mind-boggling.
The Panspermia hypothesis proposes that terrestrial life began with dormant spores arriving on the proto-Earth from space. Their dormancy was broken by their arrival here and they proceeded to produce the tree of life that was our focus of attention in the first section of this chapter. This hypothesis has been championed by various scientists, including the Swedish chemist Svante Arrhenius and the British astronomers Fred Hoyle and Chandra Wickramasinghe. My personal view is that it is wrong, and that the alternative hypothesis of steps 4 to 7 having occurred on Earth – let’s call it Terraspermia – is correct.
My belief in the correctness of the Terraspermia hypothesis is not based on us having a good understanding of the origin of life on Earth – we don’t. Rather, it is based on three things: the likelihood that there is no life on Mars, the moons of Jupiter and Saturn, or anywhere else in our solar system; the improbability of the survival of any form of life, dormant or otherwise, through the nightmare environment of interstellar space; and the desirability of using Occam’s razor. Let’s quickly look at these three bases for terraspermia.
Life beyond the Earth but within our solar system can’t yet be ruled out, but we have no evidence for it, and at least in some places – such as parts of the surface of Mars – we’ve looked quite thoroughly. To date, we haven’t drilled through the icy surface of the moons Europa (orbiting Jupiter) or Enceladus (orbiting Saturn), so there might yet be life in their sub-surface oceans. But I doubt it. Perhaps some terrestrial extremophiles could survive in such an environment, at least for a while. But could life actually originate and evolve there? That would be much harder.
If there’s no extraterrestrial life in our solar system, then an incoming spore would have to come from the planet now known to orbit the nearest star to the Sun – Proxima Centauri – or from even further afield. This means that it would have to survive distances vastly greater than those separating the planets within a single system. The probability of such survival is, in my view, zero. Even those ultimate survival machines the tardigrades, which we met earlier, mostly died in Earth orbit – a fact that is less often emphasized than the survival of a few of them. The difference in journey-time between a few Earth orbits and making the trip from the Proxima Centauri planetary system to our own is immense: the latter would take about a billion times as long, travelling at a similar speed.
It might be better to say that the probability of a dormant spore surviving such a journey would be vanishingly small rather than zero. So there is still a glimmer of hope for panspermia. But this is where Occam’s razor comes in. Why propose a hypothesis involving two improbable events instead of just one? Science prefers simpler solutions whenever possible. Amino acids, sugars, and other small molecules becoming proto-cells containing macromolecules of DNA and proteins is a tall order anywhere; it doesn’t become any easier if we transplant it to a planet orbiting Proxima Centauri. But if we do so transplant it, we must invoke a second tall order – that of surviving a journey of more than 4 light years. So terraspermia is preferable to panspermia on the basis that it requires fewer improbable causes.
However improbable is an origin of life from routinely found organic molecules on any particular initially barren planet in any particular millennium of its history, the probability of life originating rises massively when we consider vast collections of planets over vast spans of time. And that takes us to where we’re going next – the entire Milky Way galaxy.