Where the microbes aren't

Authors: Charles S Cockell

Abstract

Although a large fraction of Earth’s volume and most places beyond the planet lack life because physical and chemical conditions are too extreme, intriguing scientific questions are raised in many environments within or at the edges of life’s niche space in which active life is absent. This review explores the environments in which active microorganisms do not occur. Within the known niche space for life, uninhabited, but habitable physical spaces potentially offer opportunities for hypothesis testing, such as using them as negative control environments to investigate the influence of life on planetary processes. At the physico-chemical limits of life, questions such as whether spaces devoid of actively metabolizing or reproducing life constitute uninhabitable space or space containing vacant niches that could be occupied with appropriate adaptation are raised. We do not know the extent to which evolution has allowed life to occupy all niche space within its biochemical potential. The case of habitable extraterrestrial environments and the scientific and ethical questions that they raise is discussed.

Introduction

At the current time, Earth is the only planet in the observable universe that we know supports the phenomenon we call life (taken here to be soft matter capable of replicating and evolving, but see Mix 2015 for the substantial argument about the definition of life). This fact renders biology, the study of life, a science limited in spatial extent to a single planet, in contrast to physics and chemistry, whose reach extends across the observable universe. For example, telescopes can be used to study the chemical composition and structure of the earliest galaxies just hundreds of millions of years after the Big Bang (Curtis-Lake et al. 2023), and the Cosmic Microwave Background Radiation, which we can measure, is thought to be an energetic remnant of the Big Bang itself (Smoot and Scott 2000).

The discovery of life on other planets would expand biology to being a more universal science, but even with such discoveries, spatially, it would still be a limited material phenomenon in the universe compared to the rest of the matter we can study (although the study of life on Earth and elsewhere might well reveal general principles about life that are genuinely universal, just as the study of physics in terrestrial laboratories gives us information that can be applied universally).

Although most of the universe is likely uninhabitable to life, i.e. possessing physical and chemical conditions exceeding the capacities of life for active metabolism and reproduction, there are a variety of important and interesting questions that emerge as conditions approach those that can sustain life (where the limits of life are necessarily defined by the life that we know).

At the edges of the physical and chemical limits for life on Earth, are all possible physical spaces that can be colonized by life occupied? Since its origin, has evolution allowed life to exploit all possible niche space that its biochemistry is potentially able to use? Alternatively, are there physical spaces currently devoid of life, but which could, in theory, support it, and might in the future, with suitable evolutionary innovations? (A caveat is the possibility that pre-existing biochemical architectures may preclude the possibility of life exploiting certain niches, a question related to how flexible biochemistry is in adapting to all possibilities open to it.)

On account of the widespread distribution of life, especially microbial life, on the surface and within the crust of Earth, it is almost taken as paradigmatic that where there are conditions suitable for life to be metabolically active or to reproduce, life will generally be found in those locations. However, transient sterilization of substrates, such as by hydrothermal fluids or asteroid and comet impact events, may generate environments that are habitable, yet devoid of life that can exploit these conditions, breaking this link (Cockell 2011).

On other planets, locations with physical and chemical conditions suitable for life (even including physical and chemical combinations not found on Earth, but which in principle would allow for the growth of terrestrial organisms) may exist, but an absence of life (e.g. prior to its origin) would make these places habitable, but uninhabited environments.

In this review, I discuss places where active microorganisms cannot be found, focusing especially on those physical and chemical conditions close to, or within, the niche space currently known to support life. The importance of these environments and the hypotheses that can be tested with them are discussed.

Clarity on definitions

The discussion in this paper can best be approached by first considering some definitions. Although definitions can degrade into semantic gymnastics with no inherent scientific value, it is useful to define how certain terms are applied in this review (Fig. 1).

The first matter in which to achieve clarity is the concept of the niche and the habitat. Both terms have been the subject of intense discussion over decades, especially the niche, which has led to some classical concepts of what they are (Grinnell 1924, Hutchinson 1957, Elton 1958, Whittaker et al. 1973).

A niche is usually defined as a set of resources and physical and chemical conditions (including biotic interactions) required for life to achieve net population growth. This can be conceptualised as an n-dimensional hypervolume of an organism’s requirements (Hutchinson, 1957). Thus, a niche can also be viewed as the functional characteristics of an organism required to achieve net population growth (e.g. Whittaker et al. 1973). This is consistent with using metabolic capabilities and processes, e.g. in microorganisms, to establish niche space (Malard and Guisan, 2023).

A potential caveat to some traditional definitions of the niche that require net population growth is that microbes can persist for long periods of time, without reproducing, in a state of dormancy, such as through resting states (e.g. spores) or maintaining minimal metabolic functions to retain viability (Davey 2011, Dunchin 2018, Lloyd 2021, McDonald et al. 2024). For microorganisms, we could instead say that for a set of parameters to represent a niche, they must allow for active metabolism that enables reproduction or sustained cellular maintenance.

Temporary metabolic activity does not necessarily demonstrate niche space. A dormant organism may land in an environment in which it can never achieve sustained metabolic activity and yet still retain viability (even short-term metabolic activity) for some period. Analogously to a large multicellular organism temporarily surviving in an environment alien to it (e.g. a camel placed into a polar environment), this organism will eventually perish unless conditions change and become more favourable or unless it is transported elsewhere to more favourable conditions.

Having said all this, it seems that supplied with nutrients and other optimal conditions, ultimately a cell, even one capable of sustained metabolic activity, may be destined to perish (Rusconi et al. 2014, Taheri-Araghi et al. 2015). Our knowledge of the internal clocks and processes that establish the metabolic states of dormant cells and their longevity is still in its infancy, but the notion that a cell can simply persist indefinitely at some baseline level of metabolic activity seems too simplistic (Danchin 2018). Thus, in alignment with traditional concepts of the niche, perhaps for microorganisms as with multicellular organisms, net population growth is the most robust threshold to define niche space. Nevertheless, to encompass the enigmatic and grey area of dormancy, in this review, I will include sustained active metabolism (e.g. in dormancy) to be included in niche space and thus the definition of a physical space as being ‘habitable’.

In contrast, a habitat is often viewed as a physical space in which niche conditions are met for some type of life (McMahon et al. 1981, Hall et al. 1997). Using a simple analogy, Eugene Odum (1971) explained that a niche can be considered an organism’s profession, and a habitat its address. This is the concept of a habitat that I adopt in this review.

The definitions of niche and habitat used here are broadly similar to those described by Whittaker et al. (1973), who define the niche as ‘the role of the species within the community’ (implying the set of resources and conditions that allow it to perform this role) and a habitat is ‘the distributional response to intercommunity environmental factors’ implying the physical location in which these conditions are to be found.

Nevertheless, in some views, the niche is not mutually exclusive from a habitat. For example, Block and Brennan (1993) consider a niche to be all the biotic and abiotic factors that permit an animal to use part of an environment and a habitat to be specifically the physical factors within that set. In this sense, the habitat is part of the defined niche space and is inseparable from it. This follows Grinnell’s (1924) concept of the niche space whereby the environment in which an organism exists (what we might describe as its habitat) must also be part of the concatenation of conditions that define the niche. Similarly, Whittaker et al. (1973) suggest that although the niche and habitat can be defined separately, they can be combined to create the concept of the ecotope. Even in these concepts, however, the habitat as a physical space is viewed as a separable entity that is part of the niche.

Against all these definitions, it is incorrect to use niche as synonymous with habitat. This is a common problem, especially in the planetary sciences literature. A failure to distinguish these two concepts can lead to confusion.

One contentious issue relevant to this review is the concept of a vacant niche (Gibson-Reinemer 2015). Some people consider that a niche must be defined with respect to an organism using that niche. Thus, a vacant niche is a logical impossibility (see Lewontin 1978, Whittaker et al. 1973, and the seminal exchange between Woodley 2007, 2008, and Rhode 2008). However, if we consider a niche to be some theoretical parameter space in which some known type of life could exist, then we could talk about a vacant niche. This view has been widely adopted in the ecological literature (Kawata 2002, Rohde 2008, Lekevičius 2009, Dlugosch 2015), and this assumption is adopted here.

Given the difference between a niche and a habitat, this leads to one necessary point of clarity. A vacant niche is not synonymous with a physical space that is habitable, but uninhabited; it is a set of unused resources or functional capabilities in the environment that might happen to exist in a physical space completely devoid of actively metabolizing life. A vacant niche can (and often does) exist in a physical place that already contains life. Consider a rock that hosts a thriving microbial community and has within it sulfate and organics that are available to some known sulfate-reducing microorganisms absent from that rock. This biologically diverse rock can be said to have a vacant niche for these sulfate-reducing microorganisms. In this case, a macroscopic physical space containing life can contain vacant niches. The extent to which habitats become saturated with species or contain vacant niches is a matter of some contention in ecology (Lekevičius 2009), and the presence of unused niche space will vary in different environments.

What about the nomenclature for physical spaces that are habitable, but do not contain any actively metabolizing or reproducing life? One could call these vacant habitats, a physical space that is vacant for life. This was the approach adopted by Cockell (2011). However, this nomenclature has problems. As with vacant niches, a habitat could be vacant for a particular type of life, but contain other types of life. For example, at the macroscopic scale, we could talk about a grassy hillside being a vacant habitat for a species of butterfly. Indeed, a large part of conservation biology and animal reintroduction programmes is predicated on finding physical spaces at the macroscopic scale that are described as vacant habitats for a given species of animal (e.g. Errington 1940, Steele-Petrović et al. 1979, Sullivan 1979, Thomas et al. 1992, Johnson et al. 2010), but otherwise contain life (as animals usually cannot recolonize completely barren, lifeless environments, these vacant habitats usually already contain life). So, for historical reasons, and for consistency, this latter convention seems sensible to preserve to avoid confusion.

Another way to define an uninhabited, but habitable space, is to call it a sterile habitat. This term has been used to describe new volcanic terrains (Gudmundsson 1970). However, there are problems with this too. It is possible for a physical space to be habitable to some type of life, but to contain spores or other dormant states of life (e.g. microorganisms that have been atmospherically deposited) that are unable to achieve sustained metabolic activity or reproduction in that space. Thus, the environment does not contain organisms capable of using that space as habitat, but it is not sterile.

These problems led Cockell (2014a) to describe these physical spaces as uninhabited habitats, i.e. habitable spaces that are devoid of life that can be metabolically active or reproduce in that environment. In other words, the space is not being actively inhabited, but it is not necessarily devoid of any life at all (i.e. sterile). Some ecologists, in a manner similar to the concept of a vacant niche, could argue that a habitat is defined with respect to an organism within that physical space using the resources available, meaning that, like a vacant niche, the concept of an uninhabited habitat is a logical impossibility. However, given the acceptance of the vacant niche and the existing use of the term ‘vacant habitat’ by conservationists, if we allow for the fact that a physical space can be investigated sufficiently to show that it could, in principle, support the metabolic activity of at least one known form of life, then we could talk about a physical space empty of this life, but capable of supporting such life as an uninhabited habitat. This theoretical notion of habitat space would be consistent with Whittaker et al. (1973), who consider that a habitat hypervolume is ‘an abstract (my italics) formulation of this range (of environments or communities over which a species occurs) in terms of extensive environmental variables and the species’ limits in relation to them’.

A habitat that contains life capable of reproduction and/or sustained metabolic activity for maintenance we might refer to as an inhabited habitat, noting the point of principle above, which would make this definition non-tautologous if we accept the possibility of uninhabited habitats.

Finally, there exist physical and chemical conditions that exceed the metabolic capacities of all known life (Fig. 1). Physical spaces that contain these conditions we describe as uninhabitable. By default, there is no niche space for life in uninhabitable spaces. It would be incorrect to describe an uninhabitable location as a vacant niche.

These definitions are important because it is essential to distinguish uninhabitable spaces from habitable, but uninhabited spaces (despite the fact that they are both devoid of life capable of being maintained there). For example, Dragone et al. (2021) ‘and it is widely assumed that all environments on Earth should contain detectable microorganisms (Cockell et al. 2012). This assumption is likely incorrect’. However, Cockell et al. (2021) were not suggesting that all environments contain microbial life. They were pointing out that most habitable locations on Earth contain life, leading to an assumed coupling between habitable conditions and the presence of life. Even in this respect, however, they were elaborating on the different ways in which that coupling is broken and, more specifically, how it might be in places beyond Earth.

What we define as uninhabitable and habitable space depends on our knowledge of the limits to life. In general, the space we consider to be uninhabitable is likely to be an overestimate if we assume that our knowledge of the limits to life is incomplete and that greater knowledge will eventually lead to the limits of life being expanded. This problem will be discussed later.

In summary, the semantic issues are not necessarily resolved here; the nuance can become convoluted. The above definitions clarify those used in this review.

The distribution of environment types on Earth

Accepting the definitions considered above, how are such physical spaces distributed on Earth?

Most of Earth’s volume is uninhabitable space. If we take the current upper temperature limit to life as 122°C (Takai et al. 2008) (this upper limit was achieved at a specific pressure of 20 MPa, and it is unlikely that it represents the true upper limit of life), then assuming a geothermal gradient of ∼25°C/km (Kolawole and Evenick 2023), the depth of the biosphere is ∼5 km, although geothermal gradients vary. These limits suggest that ∼99% of the planet’s physical rocky volume is uninhabitable space (Cockell 2021). Uninhabitable conditions are found in the atmosphere above the altitude at which microbial life can propagate in cloud droplets. Assuming the lower temperature limit of life to be around –20°C (Collins and Buick 1989, Clarke et al. 2013), then the upper altitude limit of life would be ∼5 km, rendering most of the atmospheric height uninhabitable. Other factors such as desiccation, UV irradiation, and nutrient limitation also impose severe stresses on life in the atmosphere (Bryan et al. 2019, Tignat-Perrier et al. 2020, Šantl-Temkiv et al. 2022). Even if these limits were to be expanded, they would not qualitatively alter the conclusion that most of the volume of Earth is uninhabitable. These simple estimates are consistent with the large volumes of uninhabitable space suggested by Jones and Lineweaver (2010) who calculate that 88% of the volume where liquid water is stable on and within Earth is likely uninhabitable because of extreme temperature and pressure conditions.

On the crust and within the first few kilometres of the subsurface, observed at macroscopic scales, spaces that are habitable to known life appear to be mostly inhabited, leading to a strong coupling between habitable conditions and the presence of life mentioned earlier (Cockell 2011). Following the definitions applied above, we might refer to these locations as inhabited habitats. Then there exist habitable environments that contain no life that can use the space to achieve sustained metabolic activity, uninhabited habitats. These environments appear to be rare at the macroscopic scale, but may be common at the micron scale. Examples of how these can come about are discussed in the next section.

The three types of environments are potentially interchangeable (Fig. 2). Where uninhabitable conditions lie far outside the limits of life, small deviations in physical and chemical conditions will not render them habitable. But closer to the limits of life, a change in conditions may transform an uninhabitable location to one that is habitable. An example would be a cooling lava flow. If this transition is contemporaneous with inoculation by a suitable organism, then the environment may become nearly instantaneously inhabited, but if the inoculation is delayed, the environment may retain the characteristics of an uninhabited habitat. Experimental analysis of this transition from uninhabited habitat to inhabited habitat has demonstrated examples of these conditions persisting for months (Cockell 2020).

In reverse, a deterioration of conditions can cause the loss of habitable conditions leading to uninhabitable space, whether the physical space was occupied or not.

Conceptually, we might consider these three types of environments as shown in Fig. 3. The figure illustrates, in three dimensions, the hypervolume of physical and chemical conditions in the universe (in reality it is an n-dimensional object of all physical and chemical conditions in the universe, simplified to three dimensions for representation, which include pH, temperature, pressure, etc.). The red area exceeds the physical and chemical conditions associated with life’s niche space and is uninhabitable according to our current knowledge.

Some space has physical and chemical conditions permissive to life and is occupied by life. This is represented by the blue spheroid. This diagram refers to known life (on Earth), but if life exists elsewhere, this niche space could also include that life as well. If life of an entirely different biochemistry exists somewhere beyond Earth, then it could define a separate habitable object in what we currently consider to be the uninhabitable hypervolume.

Regions that are uninhabitable may be occupied by life with the expenditure of energy to maintain internal cell or organismal conditions out of equilibrium with the environment (Hoehler 2007). For example, microorganisms are largely in thermal equilibrium with the environment so that the temperature conditions within the environment itself determine whether a microorganism can be metabolically active. However, organisms can maintain a near neutral cytosolic pH in highly acidic conditions by active proton pumping to the exterior, allowing them to grow optimally in exterior conditions that would otherwise be uninhabitable to them (Baker-Austin and Dopson 2007). Other, more complex examples of this phenomenon include endothermic multicellular organisms in the polar regions (e.g. Welch et al. 2014). A more extreme example would be the construction of a self-sustaining human settlement on another planetary body, such as the Moon (Hoehler 2007, Pittman et al. 2016). In these cases, with respect to Fig. 3, energy expenditure to maintain conditions in the organism sufficiently out of equilibrium with the local environment can allow inhabited regions (blue) to extend further into the uninhabitable (red) regions to become established locally in isolation in the red region beyond the blue spheroid.

Within this diagram, there are yellow regions (a ‘rind’) at the edge of the spheroid. These are physical and chemical extremes that may, in theory, be accessible to life using biochemical adaptations, but are not yet occupied by it (and we have no good biophysical reason to think that it could not be inhabited). These potentially habitable, but uninhabited regions at the edges of life will be discussed later.

There are regions within the spheroid (grey objects) that depict physical spaces with physical and chemical conditions well within those limits known to support life that can reproduce and/or sustain metabolic activity, but which are empty of such life. In practice, these combinations of physical or chemical extremes are not necessarily devoid of such life in all spaces in which they occur, so the grey objects depict local physical occurrences.

The n-dimensional object that contains life is a miniscule part of the physical and chemical parameter space in the universe, especially with respect to temperature and pressure. Considering temperature and pressure, Fig. 4 shows the location of Fig. 3 in the range of some representative values of pressure and temperature conditions in the universe that have been measured under laboratory conditions. Assuming life was to occupy the complete temperature and pressure range shown, it fills 3.8 × 10^−12^% of the area of the graph. Qualitatively, any alteration to the end-member values shown on the axes in Fig. 4 matters little to the conclusion that the replicating evolving matter we call life is remarkably parochial with respect to the conditions under which all matter can exist. Furthermore, Fig. 4 shows the parameter space, but not the physical prevalence of such conditions in the universe. Considered in terms of physical space, the uninhabitable space makes the life-supporting region shown in Fig. 4 even more diminutive.

Having considered briefly the distribution of habitat types on Earth, let us consider the specific example of habitable locations that are devoid of life that can reproduce and/or sustain metabolic activity and how they might come about.

Uninhabited habitats on Earth at macroscopic and microscopic scales

Macroscopic scales

This section considers environments capable of supporting life that can reproduce and/or sustain metabolic activity, but are categorically devoid of such life. Figure 5 shows a range of possible ways in which habitable but uninhabited spaces can exist on a planet such as Earth, which otherwise hosts life (the case of lifeless planets will be considered later). We can broadly consider them to exist at two scales—the macroscopic scale (on the order of millimetres to kilometres) and microscopic scale (microbial scales).

Macroscopic instances of physical spaces that are habitable but uninhabited can come into existence in a variety of ways (Table 1).

An inhabited space can be buried, exposed to crustal temperatures that sterilize it, and then it can be uplifted back into habitable conditions, but remain disconnected from colonists or separated from them by a time lag (Fig. 5A1). An example could be paleosterilized oil deposits, which are reported to contain low- or non-biodegraded oils on account of this process of ‘paleosterilization’ (Wilhems et al. 2001, Adams et al. 2006). A caveat to these observations is that the unaltered state of the oil shows a lack of oil-degrading organisms, but this does not rule out the presence of active organisms using other redox couples, so their status as uninhabited habitats remains open.

Paleosterilization may be caused by a transient thermal pulse induced by processes other than burial, such as by spatially fluctuating hydrothermal fluid flows (Fig. 5A2). An example of such an environment may be the Nankai Trough, Japan, in which sections within the subsurface are reported to be habitable but devoid of vegetative cells. These zones are attributed to transient sterilization by ∼145°C–220°C fluids (Heuer et al. 2020). The lack of observable microorganisms in other deep sediments near hydrothermal systems, such as the 27 samples apparently devoid of life collected from site 858 of the Middle Valley, Juan de Fuca Ridge, Pacific Ocean (Cragg and Parkes 1994), might be caused by paleosterilization. If the fluid flow is constant enough, however, these latter subsurface sites may lie permanently above the temperature limit for life (i.e. they are uninhabitable).

Another way in which transient sterilization can occur is by asteroid or comet impact, which can deliver a thermal pulse into the subsurface derived from the kinetic energy of the impacting object, sterilizing substrates. As the thermal anomaly dissipates over time by conduction and radiation, so the sterilized section will drop below the upper temperature limit for life, bringing the material back into available niche space. This habitable section may remain permanently disconnected from habitable regions or be separated by a time lag (Fig. 5A3). Undetectable microbial abundances in some sections of the Chesapeake Bay and Chicxulub impact structures (Cockell et al. 2009b, 2021) could represent instances of this type of uninhabited habitat. A caveat with these observations is that minimum cell numbers measured were ∼10^4^ cells/cm^3^, thus low biomass in these sections cannot be ruled out in the present day. Mineral evidence of thermal conditions well above the upper temperature limit for life immediately after impact shows that these sections (in Chesapeake, Malinconico et al. 2009, in Chicxulub, Kring et al. 2020) were sterilized during impact. Even if they are colonized by a low-biomass biota in the present day, unless they were contemporaneously colonized as soon as the temperature dropped below the upper temperature limit for life, these sections could have hosted uninhabited habitats during some interval in the aftermath of impact. Analogously to an impact event, a thermal pulse might be delivered by subsurface tectonic activity leading to a transient rise in local temperatures.

Yet another way to form an uninhabited habitat is for a new habitable substrate to be formed de novo under conditions too extreme for life, for the substrate to transition into habitable conditions, and for the substrate to remain disconnected from life, or transiently empty of suitable colonists (Fig. 5A4). An example of such an environment could be the interior of a lava flow (Cockell 2014a). Upon extrusion, its temperature can be above the upper temperature limit for life. Once cooled to below this limit, it becomes habitable. Although the surface may be colonized rapidly by the aerial biota, the interior permeable vesicular space may remain devoid of life, as observed in the Eyjafjallajökull lava flows, Iceland, 4 months after the eruption (Kelley et al. 2014).

In theory, yet another example of transiency could be an environment that is habitable only for a short period of time, too short for colonization to occur, before transitioning back into an uninhabitable state. A conceptual example of such an environment is a lava flow as described above which, after becoming habitable, but before colonization, is covered by a new lava flow or high-temperature geothermal fluids that render it uninhabitable.

In principle, any process that causes transient sterilization could produce an uninhabited habitat. Other mechanisms might include, for instance, a transient radiation pulse or a movement of a microbially lethal fluid through an environment that is subsequently washed out by more benign fluids permissive to microbial activity. In natural environments, thermal pulses would seem to be the most common potential mechanism, as illustrated here.

At these macroscopic scales, the evidence for a lack of organisms within the physical space is confounded by methodological problems, mainly the inability to rule out a low-biomass population or inactive cells. Nevertheless, the above examples illustrate several ways in which uninhabited habitats may come into existence, breaking the coupling between habitability and the presence of life. These spaces are illustrated in Fig. 3 by the grey objects within the blue habitable spheroid.

Not shown in Fig. 5 are artificial substrates and fluids that are uninhabited habitats. Examples of these abound in microbiological laboratories in which the deliberate formulation of microbial media, which is sterilized before inoculation as either liquid or solid material, is made. These media represent newly formed environments suitable for life that can reproduce and/or sustain metabolic activity in its presence but are devoid of such life until the media is purposefully inoculated or becomes contaminated. It would not be wrong to say that microbiologists’ efforts to undertake culture-dependent microbiological growth are essentially the science of making uninhabited habitats.

Types of uninhabited habitats that exist well within the niche space for life are distinguished in Fig. 3 from environments poised at the very edges of life (the yellow outer rind in Fig. 3). The nature of environments poised at the edges of life’s limits is discussed in the next section.

Ratliff et al. (2023) assembled a variety of reports of macroscopic environments devoid of life. In Table 1, some of those citations provided by Ratliff et al. (2023) are included in a compilation of environments reported to be devoid of life. These authors did not explicitly distinguish between lifeless samples thought to be uninhabitable or samples that are habitable but devoid of life. In Table 1, an attempt to segregate these environments has been made. The list of uninhabitable environments could be vastly expanded by including more samples with physical and chemical conditions far outside the limits of life, e.g. from the high atmosphere, from the deep interior of Earth, and especially extraterrestrial samples (discussed in the following section). The samples shown in Table 1 are those in Earth environments close to the limits of life or within the niche space of life. They attract attention because their status with respect to life that can reproduce and/or sustain metabolic activity is often equivocal.

Do the physical and chemical conditions within apparently lifeless environments show any commonality? Ratliff et al. (2023) collected a number of papers in which no life had been reported and carried out an NMDS and ANOSIM analysis to investigate the physical and chemical features of these environments. They found no trend. There might be two possible reasons for these results. First, consider uninhabited habitats. Many of the mechanisms for their formation illustrated in Fig. 5, such as burial or impact heating, leading to sterilization, are not associated with any particular chemical or physical substrate. Any substrate with any set of geochemical conditions can potentially be subjected to sterilization and cooling processes that render it an uninhabited habitat. We would not expect uninhabited habitats formed in these ways to show any particular physical or chemical similarities in the nature of the substrate.

Second, if we include uninhabitable spaces within the set of data, this lack of correlation is even more likely to be the case, since uninhabitable conditions can be found in any combination of physical and chemical parameters that lie outside the limits of life. In its broadest universal sense (Fig. 4), uninhabitable conditions exist under a vast range of physical and chemical combinations and any set of points selected within uninhabitable space would not necessarily be expected to show commonality. Even at the edges of the limits to life (Fig. 3), the presence of spaces devoid of life, e.g. caused by a lack of appropriate evolutionary innovations, could in principle include many physical and chemical combinations.

Microscopic scales

The prevalence of uninhabited habitats at the micron scale, especially in locations containing life, is more enigmatic (Cockell 2021b). At macroscopic scales, we can easily appreciate that the three environments depicted in Figs 1 and 2 can be patchily distributed. However, are uninhabited, but habitable, spaces common at microscopic scales, even in habitats apparently inhabited at macroscopic scales?

From a crude volumetric point of view, it is apparent that large proportions of space, for example in rocks, are not occupied. Consider a hypothetical cell abundance of ∼10^5^ cells/cm^3^ in the deep subsurface in a rock with 5% porosity. Assuming that the pore space is available and accessible, and assuming that each microbe has a volume of 1 μm^3^, then we can see that only 2 × 10^−5^% of the volume is occupied. Even in more biologically colonized surface environments, often the total surface area colonized is a small fraction of that available. To provide an empirical example, Cockell et al. (2009a) measured microbial numbers in weathered basaltic glass in Iceland as ∼ 1 × 10^7^ cells/g, yet the surface area measured was 90 m^2^/g, suggesting that (assuming a typical cell covers 1 μm^2^) 1 × 10^−5^% of the available surface area is occupied. These numbers will vary greatly depending on cell size and the accessible (permeable) surface area available to microorganisms. However, they are qualitatively adequate to illustrate that large percentages of the surface area in rocky environments are unoccupied. A similar set of simple volumetric calculations might be applied to the oceans and other water bodies, and they would result in similarly large percentages of water unoccupied by microorganisms. What is the status of the unoccupied space?

Figure 5B depicts the status of spaces that lie between inhabited spaces at microscopic scales using a rock surface for illustration. Between two microorganisms, space could be uninhabitable on account of localized physical or chemical extremes that preclude active life (Fig. 5B1). This could be because of local conditions caused by the physical and chemical properties of that space (such as the presence of a toxic mineral) or because cells on either side of the space have extracted available nutrients from it, making it uninhabitable. Habitable conditions could be transient or fluctuating (Fig. 5B2), disallowing active microbes to persist in that location. The formation of habitable conditions, although stable, could be recent, and not enough time has elapsed for microbes to move into the space (Fig. 5B3), or some physical or chemical biogeographical dispersal filter might exist between the habitable space and the nearest colonized space (Fig. 5B4). As microbial colonists become detached from surface growth by fluid flow and other physical and chemical processes, it seems likely that at micron scales uninhabited habitats come and go regularly in natural environments.

These observations leave open the possibility that in some environments, despite the space at macroscopic scales appearing to be inhabited, microscopic scales may be dominated by, or at least harbour, large areas of uninhabited habitat. This would imply that habitable, but uninhabited spaces at small scales are more the norm and that the link between habitability and the presence of life is more tenuous than it might appear when life is observed at macroscopic scales.

What is happening at the limits of life?

Returning to environments near or at the limits of life. At the extremes of life’s capabilities, are there physical spaces that could, in theory, support life that can reproduce and/or sustain metabolic activity with appropriate evolutionary adaptations, but currently remain devoid of it? In Fig. 3, this potential region of habitable, but uninhabited space at the limits of life is shown as a yellow rind at the edges of the blue inhabited niche space.

If life has not yet evolved the capabilities to use the physical spaces of the rind, then, strictly speaking, these spaces we should say are uninhabitable with respect to known life (there is a seamless transition from inhabited (blue) to uninhabitable (red) space at the edge of the object). However, the diagram illustrates the possibility that there may be physical regions containing extremes just beyond the edges of life’s current occupancy that are open to its eventual occupancy. Put another way, we could Has life expanded to achieve its maximum biochemical potential?

This question seems reasonable because presumably, just before life first arose on Earth, the blue space in Fig. 3 was entirely yellow and represented the physical and chemical conditions theoretically accessible for the propagation of life with the biochemistry possible on this planet. When the origin of life occurred, one blue spot emerged in that potentially habitable volume, which expanded to fill the habitats we recognize as containing life today. How quickly does life tap into its total biochemical potential to fill space, and is this still an evolutionary work in progress? To what extent is life’s physical and chemical limit constrained by past evolutionary ‘decisions’, or is biochemistry sufficiently flexible to adapt to certain extremes regardless of prior evolutionary path?

To illustrate the uncertainties, consider the empirical data on the growth of Halomonas hydrothermalis at its limits (Fig. 6; adapted from Dickinson et al. 2021). Point X, a set of combined temperature and salinity extremes just beyond the limits of its growth (according to the criteria and period of observation adopted in that work) is shown. Yet each of the individual extremes of temperature and salinity imposed on their own at these same levels can allow for growth. Does this combination represent some fundamental biophysical limit, or, if the organism was gradually exposed to combined temperature and salinity extremes in a natural environment, could it evolve to grow within them? In other words, with respect to this organism, would a hypothetical environment with the conditions marked by X be uninhabitable or an uninhabited habitat (i.e. the physical and chemical conditions within that physical space represent a vacant niche to the organism, which could be occupied with appropriate adaptation/evolution)?

These difficulties relate to the idea of using natural environments that are devoid of life to help us establish the limits to life. Ratliff et al. (2023) suggested that by cataloguing space devoid of life, we can approach defining the limits of life from the ‘outside’ in, rather than the more conventional approach, which is to move ‘inside’ out by exploring the limits of biology itself.

Consider again Fig. 6 and the hypothetical environment X. Without the data presented here (e.g. assume that we only have laboratory data on the organism’s growth limits to temperature or salinity as separately studied parameters), we might conclude that the empty environment is a vacant niche for that organism since both extremes on their own, at the levels measured in that environment, allow for growth. This hypothesis would be incorrect since we would not know that the combined extremes render the environment uninhabitable to that organism. Even with the data that we present here, it remains an open question as to whether point X is truly uninhabitable or a vacant niche that could be filled by appropriate biochemical adaptation. We could only distinguish between an uninhabitable space and a vacant niche by performing laboratory experiments (e.g. directed evolution experiments) to see if the organism could adapt to grow or be metabolically active at point X from point Y, for instance. In principle, for all we know, no amount of evolutionary adaptation would allow this organism to inhabit point X because biophysical and/or energetic reasons make it a forbidden niche.

Another problem is that to know whether a place devoid of life helps define the outer limit of life or whether it is an uninhabitable space with conditions far from the limits of life, we need to have some idea about where the limits for life are in the first place.

Contemplate Fig. 6 again and two hypothetical saline environments in which we might search for H. hydrothermalis. One of the locations has a temperature of 44 and the other 70°C. Both would be devoid of H. hydrothermalis. If I knew nothing about the growth of the organism, I would be none the wiser as to which of these helped define the limit of life, since in the former I would only know it was less than 44°C, and the latter less than 70°C. To know that the former did in fact define a temperature just outside the limit (see Fig. 6), I would need the growth data in the habitable region. Thus, an environment that lacks life can only be useful when it is employed to confirm a limit that is first obtained by investigating life, i.e. moving ‘outwards’ towards the edge using biological investigations.

In the natural environment, where we are seeking the limits of whole ecosystems, similar arguments apply. In Fig. 7, the discovery of a natural environment devoid of life (e.g. X1 or X2) can only become useful for establishing the limit of life when we have information on the left side of red line, which gives us the upper growth limits of all organisms. This conclusion is the case because, as shown in Fig. 6, subtle changes in the combinations of extremes (Harrison et al. 2013) generate ambiguity about the difference between uninhabitable spaces and vacant niches. Other extremes that have been shown to act in non-additive ways include water activity tolerance and temperature (Nichols et al. 1999), high pressure and temperature (Takai et al. 2008), high pressure and salinity (Kish et al. 2012), iron starvation, microaerophilic conditions, and temperature (Harrison et al. 2015), or water activity, UV radiation, and low temperatures (Dragone et al. 2021). At the physical and chemical limits of life, these interactions between extremes can be complex, involving many parameters. This forces us back into working from the inside out using studies of microbial growth under multiple extremes, either in laboratory or natural settings, to establish the limits to life under multiple extremes.

Ambiguities about the characterization of environments as uninhabitable are reflected in results from natural environments. Putatively lifeless samples can leave us confounded as to whether they are categorically uninhabitable, lifeless but potentially habitable (uninhabited habitats), or inhabited with extremely low biomass. Examples of locations where these difficulties exist include deep marine and subsurface continental low-water activity and high chaotropicity brines (e.g. Hallsworth et al. 2007, Payler et al. 2019). Their status with respect to the presence of life has been controversial (e.g. Hallsworth et al. 2007, Yakimov et al. 2015, Steinle et al. 2018) on account of the presence of cells or biomolecules such as RNA whose origin from in situ metabolically active cells is unclear. Analyses from hypersaline near-surface low water activity pools, such as Don Juan Pond, Antarctica (Meyer et al. 1962, Cameron et al. 1972, Peters et al. 2014), and ponds in the Dallol Rift, Ethiopia, where low water activity and high chaotropicity (high Mg-brines) brines in geothermal pools can be found (Belilla et al. 2019), have resulted in similarly equivocal microbial activity.

In Antarctica, dry permafrost putatively at, or beyond, the limits of life (Goordial et al. 2016, 2017) has later been shown to support some active populations (Wood et al. 2024). Ice-free, arid, salty soils in Antarctica have similarly generated uncertainty in their biological status (Dragone et al. 2021). Studies in the Atacama Desert have resulted in a series of papers that have reported conflicting accounts of microbial activity in extremely desiccated soils (e.g. Navarro-González et al. 2003, Drees et al. 2006, Connon et al. 2007, Crits-Christoph et al. 2013). The discovery of cells in conditions above the currently known upper temperature limits for life in proximity to subsurface hydrothermal systems has raised questions about the temporally changing conditions in these environments and whether they remain permanently outside the conditions for life or whether they transiently cool into conditions permissive for growth or the transport of active cells (Cragg and Parkes 1994, Cragg et al. 2000). Challenging questions remain about the height at which atmospheric conditions become uninhabitable (e.g. Bryan et al. 2019).

At the limits of life, an increasing number of microbial metabolisms can become energetically unsustainable, and the number of different types of organisms that can subsist is reduced (Murray et al. 2012). This is conceptually illustrated in Fig. 7 as a reduction in biomass of specific species (and metabolic types) and a reduction in total species richness as one moves closer to the limit of life. At these controversial limits, we are often compelled to return to research that examines life to verify or refute a putative lack of active metabolisms. Table 1 shows some extremes and combinations of extremes in natural environments apparently at the edges of life’s niche space in which there has been some debate about the lack of actively growing organisms.

Nevertheless, in agreement with Ratliff et al. (2023), collections of samples that contain conditions close to those known to support life (i.e. samples representing rind regions in Fig. 3 and to the right of the limit to life line in Fig. 7) might lead us to carry out investigations on life near those extremes. They can alert us to combinations of extremes that are especially important in natural environments in setting limits to life, and thus they can direct us towards further work.

Uninhabited habitats beyond Earth

Beyond Earth, habitable spaces that contain life are not yet known. This may alter in the future if we find life either in our Solar System or on an exoplanet. Some extraterrestrial samples have been proposed to have been once inhabited, such as a Martian meteorite (McKay et al. 1996), although the evidence for life is disputed.

Inhabited environments aside, what about the other two environmental types depicted in Fig. 2? Figure 4 illustrates that a vast proportion of the known universe is uninhabitable (to known life). Samples that are uninhabitable and contain no evidence of life from extraterrestrial environments abound, and we could list many papers related to meteoritic materials or samples returned from the Moon, Mars, asteroids, and comets that show no evidence of life (e.g. LSPET 1969, Bridges et al. 2001, Goesmann et al. 2015). These are not included in Table 1. Some of these samples that were uninhabitable when they reached Earth may have become contaminated by life, even colonized by actively growing organisms (Toporski and Steele 2007, Tait et al. 2017), but this does not represent their original state.

Under certain circumstances on planetary bodies that possess physical conditions that overlap with, even transiently, conditions supportive of known life, we could expect the appearance of habitable spaces devoid of life. An example could be transient aqueous environments in asteroids (Suttle et al. 2021). Such conditions might even have occurred on the Moon (Schulze-Makuch and Crawford 2018).

We would expect uninhabited habitats to exist on many young rocky planets with liquid water and physical and chemical conditions within the limits of known life just prior to, or after, an origin of life. Presumably, in the period just before the emergence and radiation of life on Earth, our planet possessed large areas that were potentially habitable to life (such as the oceans and early volcanic materials), but devoid of it. Some Martian meteorites that have evidence for aqueous alteration (such as clay formation), but no evidence of life (Bridges et al. 2001, Changela and Bridges 2010), may be examples of uninhabited habitats, as might environments detected on Mars itself that have been suggested to have been habitable (e.g. Stoker et al. 2010, Grotzinger et al. 2014, Hurowitz et al. 2017).

The prevalence of uninhabited habitats in the universe largely depends on the probability of the origin of life on habitable planets and the rate of radiation of that life into available niche space. If the required chemical reactions leading to an origin of life are statistically unlikely and/or the environmental conditions allowing for it are rare (e.g. if an origin of life requires conditions for the formation of certain prebiotic molecules that are different and less abundant than those subsequently required by life; Chopra and Lineweaver 2016), or if the conditions for an origin of life are too transient, then habitable but uninhabited spaces may be common. Yet another theoretical scenario for the formation of such spaces would be for an origin of life to occur, for life to be extinguished by some cataclysm (e.g. a large planetary-scale impact), and for habitable conditions to subsequently exist without a second origin of life.

Taking these considerations together with what has been described for Earth, we can recognize the presence of environments habitable to life but devoid of active life on both inhabited and uninhabited planets (Fig. 8).

How might transient habitable, but uninhabited conditions be locally created on other planets? Figure 9 shows a potential process for the formation of an uninhabited habitat and its preservation on Mars. A collision of an asteroid or comet with the Martian surface causes melting and sterilization of near-surface ice, generating a short-lived water body (which eventually boils, sublimates, and/or freezes in the thin Martian atmosphere). Even if Mars was inhabited (e.g. in its more clement deep subsurface), the disconnection of inhabited regions from this transient water body by the cryosphere, a lack of liquid water fluid paths, and an inclement UV-irradiated atmosphere could ensure that the habitat remains devoid of active life for its duration.

Uninhabited habitats such as that shown in Fig. 9 have practical implications for missions to Mars since they suggest that we could explore a transient body of water of this kind, conclude that it was lifeless, and then conclude that Mars itself was lifeless, or, even if we had shown that Mars hosted life, we might conclude that life failed to evolve to live in the conditions represented in this specific instance of habitable, but uninhabited conditions. If we find evidence for ancient sterile liquid water on a planetary body that we think may have had conditions suitable for life, we should ask whether it reflects the biological state of wider regions or whether it might be an instance of an uninhabited habitat on an otherwise inhabited planetary body.

As Fig. 9 shows, we might also produce uninhabited habitats artificially, either deliberately or inadvertently by a spacecraft collision. In the former case, this might be done to test scientific hypotheses (discussed in the next section). In the latter case, the spacecraft could melt local permafrost either from the thermal pulse generated by its kinetic energy or from heat generated by an on-board energy source such as a radioisotope thermoelectric generator.

At a more universal scale, we could eventually find exoplanets that are habitable but apparently lacking evidence of life, just as Earth was in the period prior to the origin of life. In addition to young planets about to originate life, there may be a class of worlds that remain habitable, but uninhabited, indefinitely if an origin of life does not occur (Cockell 2014b). The difficulty with detecting and verifying the existence of such a class of planets is that habitable planets apparently devoid of life may be inhabited planets with low biomass or biomass producing weak atmospheric signatures below the detection limits of our telescopes.

Viewed at the universal scale, we could ask several questions about Fig. 2 that are depicted in Fig. 10. First, we could ask what proportion of the universe is habitable space, at least with respect to the known limits of life. This proportion is already known to be very small given the range of extremes that exist in the universe that exceed life’s limits. Of the uninhabited, but habitable regions in the universe (uninhabited habitats), we could ask what proportion of those environments are to be found on inhabited planets and what proportion on uninhabited planets. If the origin of life is rare, then we would presume that most habitable but uninhabited spaces are to be found on the latter. If it is a common process, then we should expect that the proportion of uninhabited habitats on inhabited planets is much higher. Finally, more specifically, we could ask what proportion of habitable space on planets hosting life is inhabited, the answer to which would reflect the extent and rate at which life, once it emerges on a planet, radiates to fill all possible physically manifested niche space and the rate at which it evolutionarily adapts to make use of its full biochemical potential to colonize spaces in which it can persist.

At least with respect to characterizing the distribution of habitable conditions in the universe and the prevalence of life, Fig. 10 summarizes the relationship between habitability and life and the key fractions that we want to quantify.

Use of lifeless environments in hypothesis testing

Beyond a fundamental characterization of habitable conditions and the extent to which they are used by life, what value do environments devoid of active life have in hypothesis testing? Earlier, the potential for using them to establish the limits to life was discussed (Ratliff et al. 2023).

In addition to helping define the limits to life, environments that are theoretically capable of sustaining life but are devoid of it can be used to test the hypothesis that life has expanded to occupy all possible niches that its biochemistry is capable of occupying. A comprehensive study of physical environments at the limits to life will presumably allow us to map the edges of the niche space illustrated in Fig. 3 and determine whether the yellow rind of uninhabited habitat space exists at all or whether life has already expanded to fill all possible niche spaces on Earth such that there is a categorical transition from inhabited space to uninhabitable space. We could imagine a much richer research programme than at the current time in the fields of extreme microbiology and synthetic biology to explore how far the limits to life can be artificially expanded in the laboratory. This research can be coupled to more in-depth field observations in many extreme environments on Earth to investigate the degree to which native microorganisms have reached the limits of their capabilities and whether over the next billion years or so of Earth’s habitable window we could expect new innovations in microbial evolution and a much expanded envelope of life’s physical and chemical domain before the whole of Earth becomes uninhabitable.

One use of uninhabited habitats is that they can provide negative control environments for understanding the effects of life on geological, geochemical, and geomorphological processes (Cockell 2011). Although we could use uninhabitable locations as negative controls, the fact that they have conditions too extreme for life can mean that their associated geological processes are poorer negative controls compared to environments that have conditions permissive for life, but are uninhabited. For example, sterile environments deep in the Earth could be said to be non-biological controls for understanding biosphere–geosphere interactions at the surface, but deep in the Earth, high temperatures and pressures produce mineral and chemical states rarely seen in life-bearing regions (e.g. Haggerty 1995).

If we could find uninhabited habitats, then we could get a better description and quantification of background non-biological processes in places devoid of active life that are otherwise similar or identical to inhabited places. We could then quantify the influence of biology. These measurements could include isotopic fractionation patterns, weathering rates, mineral formation and dissolution rates, water flow and erosion rates, etc. At large scales, we could address questions about the extent to which a biota influences landscape-changing geomorphological processes such as erosional rates and their associated laws (Dietrich and Taylor Perron 2006) or the formation of the macroscale sedimentary record, such as carbonate deposition (Davies et al. 2020).

Beyond Earth, examples of candidate environments could be areas on the surface of Mars where transient liquid water may activate geological processes such as aqueous weathering in the absence of life and thus potentially provide quantitative negative control measurements of such processes (Adcock and Hausrath 2015). On Mars, if habitable, but lifeless samples could be identified, then they might be used to understand the abiotic baseline of isotopic cycling from which to better interpret enigmatic isotopic signatures (Franz et al. 2017, 2020, House et al. 2022) that have been observed on the planet.

Habitable negative controls may be of value in assessing the earliest influences of life recorded in ancient terrestrial rocks. As habitable environments devoid of life may be places where chemical structures that have similarity to life can form (McMahon and Jordan 2022), then rocks formed in uninhabited habitats may be the best environments to study the abiotic baseline of life-like features (biomorphs) to better distinguish true biosignatures in inhabited ancient environments. Here again, environments in places like Mars, where ancient sedimentary rocks are more extensive and better preserved than on Earth, may be of use. Even rocks that have undergone aqueous weathering on asteroids and similar solar system bodies where transient habitable conditions may have existed could provide these negative controls for the study of biomorphs.

Uninhabited habitats offer the potential for hypothesis testing in ecology. The introduction of life into extant examples of these environments can be used to study the earliest stages of succession from an empty baseline. Many primary succession habitats, such as newly exposed glacial forelands, do not begin devoid of life (e.g. Tscherko et al. 2003, Nicol et al. 2005). However, those environments that begin lifeless, such as newly formed lava flows, might be useful for such studies. Processes of interest include the early establishment of microbial communities, priority effects on the sequence of colonization, long-term stability of newly established communities, biogeographical phenomena in the establishment of new biotic communities, the effects of newly emerging communities on geochemical cycling, including isotopic fractionation, etc. These studies encompass many of the methods associated with ecology, microbial ecology, and allied fields. Experiments have been done to study how quickly uninhabited habitats become colonized by aerial inoculants, and they provide some insight into how quickly new volcanic lava flows can become colonized (Cockell 2018). These researchers found that after 1 week, the interior of initially habitable but sterile stacks of silicate substrates with a total thickness of 1.2 cm remained uncolonized, but that after a month they were colonized. Colonization in these time scales is consistent with the observation that microbial communities were established just months after the formation of a new lava flow in Iceland (Kelly et al. 2014).

Intersection with ethics

Finally, the presence of extant uninhabited habitats raises potential ethical questions, especially on other planetary bodies. As they contain conditions that can support the activity of life but are vacant of it, can we use them to conduct experiments on biological colonization? In principle, we can imagine a range of scientific investigations in such environments (McKay and Marinova 2001, Cockell 2021a), which include those described above on the dynamics and ecology of primary succession.

On Earth, the deliberate inoculation of these environments might be cause for few ethical concerns where local biota is introduced since the environment may be poised to be colonized anyway (such as a fresh lava field). However, the introduction of non-local or non-indigenous biota to a site might cause environmental concerns.

Beyond Earth, the colonization of habitable, but uninhabited environments potentially has ethical implications, including those impinging on planetary protection protocols that militate against ‘harmful’ contamination of extraterrestrial environments (Coustenis et al. 2023). Planetary protection has been largely focused on protecting the scientific potential of extraterrestrial environments (including deleterious impacts of contamination on the search for life), but in recent years discussion has expanded to ethical concerns (Rummel et al. 2012). Planetary protection has largely focused on concerns about environments that might already host life, but as illustrated here, environments that are habitable, but uninhabited, or could become so with technological intervention, should be of interest.

In addition to the natural and artificial Martian examples already discussed, other examples of uninhabited habitats could include moons hosting subsurface oceans. Moons of the outer planets, such as Saturn’s moons Enceladus and Titan, or Jupiter’s moons Europa and Ganymede, are of potential interest. A case in point is Enceladus, from which plumes of erupting water have been shown to contain organic carbon as well as molecules containing atoms of hydrogen, nitrogen, oxygen, phosphorus, and sulfur (i.e. CHNOPS elements). The presence of H2 and CO2 demonstrates a potential redox couple for methanogenesis. Temperature, pH, and salinity conditions suggest clement conditions for life (McKay et al. 2014, Affholder et al. 2021, Davila and Eigenbrode 2024). All the ingredients for life and a suitable environment seem to be present.

If any of these bodies were determined to be habitable but lifeless, then potentially, they could be deliberately inoculated, which would offer the possibility of long-term ecological colonization experiments to study primary succession and ecosystem establishment on a planetary scale. This work would inform our understanding of how ecosystems become established on young planets and how microorganisms radiate into new metabolic niches and habitat spaces. Once life is introduced, then the process of biotic alteration is largely irreversible and could compromise investigations on the geochemistry or chemistry of extraterrestrial environments (such as the study of extraterrestrial organic molecules and their processing).

These ethical concerns would be magnified if we were to find indigenous life in other regions of planets in which these uninhabited environments have been detected or are proposed to be created. Artificially introduced life could compromise the indigenous biota if it spread beyond the inoculated environment, or the introduced life might preclude the use of natural uninhabited habitats by indigenous life.

It is not the purpose of this paper to resolve these issues, some of which, with respect to contamination of indigenous extraterrestrial life, remain speculation, but rather to draw attention to the many examples of environments now known to exist beyond Earth that are either habitable or poised close to habitable conditions (and in which technological intervention would bring them into the realms of habitable conditions). The way we use these environments and their considerable scientific value raises important ethical and policy questions that will need to be satisfactorily addressed.

Conclusion

Physical spaces without microbes capable of sustained metabolic activity or reproduction can be split into uninhabitable conditions, which are pervasive throughout the universe (including on Earth), and habitable, but uninhabited spaces (uninhabited habitats). They must be distinguished from vacant niches, which can exist in habitable environments devoid of life or in macroscopic environments containing life. Uninhabited habitats, produced at macroscopic and microscopic scales by a variety of mechanisms, can be used for hypothesis testing, including as negative control environments to understand the role of biota in altering geochemical processes and to study microbial and community succession dynamics. Their presence and abundance in environments at the physical and chemical limits of life remain enigmatic. Uninhabited habitats might eventually be located beyond Earth. These environments will have significance for understanding the distribution of habitable conditions and life throughout the universe. Their scientific use can raise important ethical questions, especially where the artificial introduction of life is considered.