Stellar populations

The first major prong of traversability is the stellar population of a galaxy. The stellar population limits the number and types of destination within a given range. It is thus particularly important for setting the scale of internal and endpoint hazards, but it can also play some role in the journey itself. The stellar population traits that bear on traversability include:

• • Space density – The space density ${n_{\star }}$ of stars is directly proportional to the expected number of destinations within a given range at any time (Di Stefano and Ray Reference Di Stefano and Ray2016; Carroll-Nellenback et al. Reference Carroll-Nellenback, Frank, Wright and Scharf2019). The travel time to the nearest potential destination should, all other things being equal, likewise also decrease with travel time ($t \propto {n_{\star }}^{-1/3}$). Patient ETIs can also wait for stars to come near their current world to shorten travel time, and the rate of these encounters increases with stellar density ($\Gamma \propto {n_{\star }}^{1/2}$; Hansen and Zuckerman Reference Hansen and Zuckerman2021). The space density of stars can compound other traversability hazards – a shorter flight can mean less exposure to cosmic radiation, for example. Overall regions with high stellar density should be more traversable. However, high stellar density might be detrimental to habitability for analogous reasons: there are more energetic sterilizing events within a given volume and stellar encounters potentially could wreak havoc on planetary systems (Lineweaver et al. Reference Lineweaver, Fenner and Gibson2004; Jiménez-Torres et al. Reference Jiménez-Torres, Pichardo, Lake and Segura2013; but see Gowanlock et al. Reference Gowanlock, Patton and McConnell2011 on the limited effect of supernovae).

• • Stellar kinematics – All stellar systems are at least partly supported against gravity by the random velocities of stars, characterized by a velocity dispersion $\sigma _{\star }$.Footnote ¹ The random motions can aid or hinder stellar travel. First, high $\sigma _{\star }$ can aid travel by increasing the rate that stars come close to an inhabited system ($\Gamma \propto \sigma _{\star }^{1/2}$) for patient ETIs that wait for stars to come to them. Second, high $\sigma _{\star }$ can aid the spread of ETI by mixing stellar populations. Landis (Reference Landis1998) proposed a percolation model where only systems at the ‘edge’ of inhabited space will be in range of uninhabited systems, but some of these are permanently uninterested in migration (or perhaps are by chance too far from any suitable systems). Large regions of the Galaxy can then end up uninhabited, blocked by these ‘no-go’ systems. Stellar mixing breaks up these walls, bringing societies interested in spreading to the ‘surface’ of the settlement front (Wright et al. Reference Wright, Mullan, Sigurdsson and Povich2014). Third, $\sigma _{\star }$ sets a floor on the typical Δv between origin and destination. High random velocities may inhibit interstellar travel, because the ETIs need to slow to a stop at the destination. If $\sigma _{\star } \sim 100\, {\rm km\ s}^{-1}$, for example, patient ETIs using chemical propulsion will not be able to rendezvous with most destinations. On the other hand, ETIs already using high-Δv propulsion could save propellant by going to systems that are already approaching at high speed and simply decelerating relative to the destination the whole way.

  A speculative fourth effect is that ETIs could use stellar flybys to perform gravitational slingshots and gain Δv for free. Significant boosts require very close approaches to large, compact masses like stellar remnants, since Δv is limited to be around the escape velocity at closest approach. Ancient neutron stars do in fact have a very high velocity dispersion (hundreds of km s⁻¹) resulting from the supernovae that birth them (Faucher-Giguère and Kaspi Reference Faucher-Giguère and Kaspi2006). ETIs might use them to accelerate craft up to ~ 1000 km s⁻¹ and cross galactic distances over a few million years, if they can withstand the tides and radiation environment a few hundred thousand kilometres from the remnant. Neutron stars have a lower space density than stars, however – in the Solar neighbourhood, the nearest is expected to be ~ 10 pc away (Ofek Reference Ofek2009) – so the ETIs need to be able to cross the larger distances first. The gravitational potential of a galaxy limits the $\sigma _{\star }$ before stellar objects escape: many neutron stars escape their galaxies, particularly small ones.

• • Star-formation history – Some astrophysical phenomena are associated with young, massive stars, and their abundance is limited by the star formation history (SFH) of the stellar population. These include O and B dwarf stars, supergiants, supernovae, bright pulsars, high-mass X-ray binaries and gamma-ray bursts. On the whole these phenomena may be considered dangerous to both inhabited worlds and interstellar travel because of the radiation they produce, particularly CRs (see next section). It is possible, however, that these same phenomena may draw ETIs interested in harnessing their power. For example, they might be used to accelerate light sails to high speeds (Heller et al. Reference Heller, Hippke and Kervella2017; Lingam and Loeb Reference Lingam and Loeb2020). The prevalence of these phenomena relative to old stars is encapsulated by the specific star formation rate ${\rm sSFR} \equiv {\rm SFR}/{M_{\star }}$, where SFR is the galaxy's star-formation rate and ${M_{\star }}$ is its stellar mass. The stellar populations of galaxies with high sSFR have more young stars. Some energetic phenomena, like low mass X-ray binaries, are present even in older stellar populations.

• • Other properties – Other characteristics of the stellar population may affect traversability. Stellar metallicity regulates the number and size of planets around stars, and perhaps the number of destinations (although terrestrial planets are not as affected as giant planets; Buchhave et al. Reference Buchhave, Bizzarro, Latham, Sasselov, Cochran, Endl, Isaacson, Juncher and Marcy2014). It also probably limits the mass in asteroids and other minor bodies (Gáspár et al. Reference Gáspár, Rieke and Ballering2016) that might be mined and used to build more craft. The initial mass function (IMF) sets the ratio of low mass stars like the Sun to high mass stars, and is usually similar to the local Milky Way's (Bastian et al. Reference Bastian, Covey and Meyer2010). Locations with top-heavy IMFs, which may include the inner parsec of our Galaxy (Lu et al. Reference Lu, Do, Ghez, Morris, Yelda and Matthews2013), have an overabundance of massive stars and all their associated phenomena. Bottom-heavy IMFs seem to exist within the inner regions of massive elliptical galaxies (Smith Reference Smith2020), with more red dwarfs and possibly substellar brown dwarfs and sub-brown dwarfs that do not contribute much to the galaxy's luminosity. Although habitable planets may be scarce around these objects, they may be convenient stopover points to restock and refuel, or have resources for building new structures.

The interstellar medium

The ISM fills the space a starship must travel through, and has largely been viewed as an external hazard to interstellar travel. Small, fast vessels like laser sails are much more vulnerable to these hazards than massive ‘world ships’, although CRs are sufficiently penetrating to be a hazard to smaller crewed vessels without heavy shielding.

A dynamic and complex medium, the ISM has many components that can bear on traversability:

• • Gas – The great majority of the ISM's mass is in the form of gas.Footnote ² Gas impacting a craft could harm it, with the hazards growing with speed. Gas heats a craft; beyond a certain speed, it may cause important structures to melt and damage the surface materials (Hoang et al. Reference Hoang, Lazarian, Burkhart and Loeb2017). At relativistic speeds, ISM particles are essentially high-energy CRs in the craft frame, inducing radiation damage (Semyonov Reference Semyonov2009). The resulting ionization may result in a charge gain, to the detriment of small craft (Hoang and Loeb Reference Hoang and Loeb2017). Potentially, it could also induce drag, although gas probably does not couple efficiently with thin light sails and the drag is too weak to slow large ‘world ships’ (Hoang Reference Hoang2017). Gas could also be an opportunity for ETIs, however, as material to be harvested for use in propulsion (Bussard Reference Bussard1960). The most plausible concept in the local ISM is the ram-augmented interstellar rocket (RAIR), which uses the ISM for propellant (Crawford Reference Crawford1990). Ionized gas might also be used as a ‘wind’ to brake small craft with a magnetic or electric sail (Zubrin and Andrews Reference Zubrin and Andrews1991; Perakis and Hein Reference Perakis and Hein2016). Finally, a dense, highly pressurized ISM can crush the stellar wind astrospheres around stars, while the astrospheres should be larger in galaxies with rarefied ISMs. This could affect some magnetic or electric sail designs.

• • Dust – Dust grains represent $\lesssim 1\percnt$ of the ISM mass (Draine et al. Reference Draine, Dale, Bendo, Gordon, Smith, Armus, Engelbracht, Helou, Kennicutt, Li, Roussel, Walter, Calzetti, Moustakas, Murphy, Rieke, Bot, Hollenbach, Sheth and Teplitz2007), but they could be a major hazard to interstellar travel. Each dust grain is effectively a hypervelocity micrometeoroid. When they hit, the kinetic energy is concentrated in one spot, carving out small craters in the ship surface and causing erosion (Hoang et al. Reference Hoang, Lazarian, Burkhart and Loeb2017). At relativistic speeds, dust grains act like clusters of CRs, and their impacts trigger enormous particle cascades.

  The prevalence of dust varies between and within galaxies. Dust grains require the presence of heavy elements, and dust is more prevalent in high metallicity galaxies (Rémy-Ruyer et al. Reference Rémy-Ruyer, Madden, Galliano, Galametz, Takeuchi, Asano, Zhukovska, Lebouteiller, Cormier, Jones, Bocchio, Baes, Bendo, Boquien, Boselli, DeLooze, Doublier-Pritchard, Hughes, Karczewski and Spinoglio2014). Generally, the density of dust is expected to track the gas density (Boulanger et al. Reference Boulanger, Abergel, Bernard, Burton, Desert, Hartmann, Lagache and Puget1996). On small scales (a few parsecs or less), though, the dust/gas ratio has been observed to fluctuate without correlation to the gas density itself (Hopkins and Lee Reference Hopkins and Lee2016). Hot ionized plasma sputters away small dust grains over millions of years, reducing their threat (Draine Reference Draine2011), although very big dust grains could survive.

  Finally, the size distribution of dust grains affects the nature of the threat. Small grains provide a steady background of tiny impacts, whereas rarer large grains concentrate their energy into potentially catastrophic impacts. Within the Milky Way, more mass is contained in the larger grains, up to a size of ~ 0.2 μm, but few grains are $\gtrsim 1\, \mu {\rm m}$ wide (Draine Reference Draine2011). Empirical and theoretical modelling suggests the size distribution of grains varies, with smaller grains more prevalent between older stellar populations and the larger sub-micron grains more prevalent in dense ISM (e.g. Cardelli et al. Reference Cardelli, Clayton and Mathis1989; Asano et al. Reference Asano, Takeuchi, Hirashita and Nozawa2013). A worrying prospect is that there is an undetectable ‘tail’ in the size distribution of much larger dust grains that does not contribute to interstellar extinction or infrared emission, our main ways of constraining dust (Hoang et al. Reference Hoang, Lazarian, Burkhart and Loeb2017): measurements by space probes demonstrate micron-size grains exist and radar and optical searches for still larger grains have been carried out (Musci et al. Reference Musci, Weryk, Brown, Campbell-Brown and Wiegert2012; Krüger et al. Reference Krüger, Strub, Grün and Sterken2015).

• • Cosmic rays – CRs are energetic particles, mostly protons and helium nuclei, that pervade the entire ISM. The bulk of the CR energy density is in protons with kinetic energies of order a few GeV. There may also be regions of the ISM with significant MeV components to the CR flux, although these would be localized (Indriolo et al. Reference Indriolo, Fields and McCall2009; Cummings et al. Reference Cummings, Stone, Heikkila, Lal, Webber, Jóhannesson, Moskalenko, Orlando and Porter2016). CRs are dangerous, especially to organic beings, because of the ionizing damage they inflict (Semyonov Reference Semyonov2009). They may also damage electronic equipment. Low-energy CRs lose their energy quickly in matter by ionizing surrounding atoms, essentially searing the skin of the craft. Higher energy CRs penetrate deeper, generating secondary particles like muons, electrons, positrons and gamma-rays as they go (Longair Reference Longair1992). In the Solar neighbourhood, the equivalent of several metres of water are necessary to reduce the ionizing flux from GeV CRs to tolerable levels for humans. Alternatively, the inhabitants and mechanisms of an interstellar craft must be able to self-heal from or tolerate this radiation damage. They may be a minor issue for thin light sails, which are mostly transparent to these particles (at the cost of being more vulnerable to gas and dust damage, as in Hoang et al. Reference Hoang, Lazarian, Burkhart and Loeb2017), and vast world ships where the surface shielding is a small fraction of the mass compared to the large inner volume.

• • Magnetic fields – Magnetic fields thread the gases between the stars. Their traversability effects mainly apply to lightweight fast vehicles like laser sails, which become charged as they pass through the ISM. Lorentz forces may deflect a Starshot-like sail off course by about 0.1 AU pc⁻¹ in the Solar neighbourhood and cause it to wobble (Hoang and Loeb Reference Hoang and Loeb2017). Very large deflections could make it difficult to deliver a probe or inscribed message into the habitable zone of fainter stars without onboard guidance. Galactic magnetic fields can include an ordered field with large-scale patterns and a fluctuating turbulent field that varies with space and time. Magnetic fields can be constrained on large scales through radio polarization and Faraday rotation mapping (e.g. Beck Reference Beck2015), but measuring the time-variable field along a particular route may require either regular traversals or setting up a radio beacon at each terminus to provide for Faraday rotation measurements.

• • Interstellar radiation fields – The interstellar radiation field (ISRF) is dominated by near-infrared, optical, and ultraviolet light from the entire stellar population; mid- and far-infrared emission from dust in star-forming regions; and the cosmic microwave background (e.g. Draine Reference Draine2011). Both the density and energy of photons encountered by a vessel increases with speed due to relativistic effects, and highly relativistic craft may be subject to ionization damage, radiative drag and even nuclear reactions (Hoang et al. Reference Hoang, Lazarian and Schlickeiser2015; Yurtsever and Wilkinson Reference Yurtsever and Wilkinson2018). However, these effects will generally only be a problem at extreme Lorentz factors ($\gamma \gtrsim 10$) well beyond those generally expected for interstellar craft. Otherwise, the effects are minimal because the ISRF is very dilute: of order 10⁻⁷ of the energy density of sunlight at Earth through most of the Milky Way. In the most extreme starbursts, the ISRF could reach as much as $\sim 1\percnt$ of solar insolation on Earth, but virtually all of it is far-infrared emission which becomes ionizing only at $\gamma \gtrsim 100$. Likewise, a spherical craft heats to temperatures $\gtrsim 1000\, {\rm K}$ only if $\gamma \gtrsim 100$ even in extreme starbursts.Footnote ³ A more indirect effect might be the inability of extreme starburst environments to form planets to go to (Thompson Reference Thompson2013), although pre-existing planets would be unaffected. It is unclear how the changes to planet formation affect the prevalence of minor bodies.

Red and blue: galaxy evolution and trends in traversability

Most galaxies fall into two broad categories, the blue star-forming galaxies (SFGs) and the red quiescent galaxies (Strateva et al. Reference Strateva, Ivezić, Knapp, Narayanan, Strauss, Gunn, Lupton, Schlegel, Bahcall, Brinkmann, Brunner, Budavári, Csabai, Castander, Doi, Fukugita, Györy, Hamabe, Hennessy, Ichikawa, Kunszt, Lamb, McKay, Okamura, Racusin, Sekiguchi, Schneider, Shimasaku and York2001). At z ~ 0, the quiescent galaxies include most elliptical and lenticular (S0) galaxies and are biased towards dense galaxy clusters, while SFGs include most late-type spiral and irregular galaxies and are typically found in the field (Skibba et al. Reference Skibba, Bamford, Nichol, Lintott, Andreescu, Edmondson, Murray, Raddick, Schawinski, Slosar, Szalay, Thomas and Vandenberg2009). These groups differ fundamentally in their sSFR. The sSFR of blue galaxies with a given stellar mass ${M_{\star }}$ are similar at any given cosmic epoch (Brinchmann et al. Reference Brinchmann, Charlot, White, Tremonti, Kauffmann, Heckman and Brinkmann2004), evoking a SFR–${M_{\star }}$ relation sometimes called the ‘main sequence’ (Noeske et al. Reference Noeske, Weiner, Faber, Papovich, Koo, Somerville, Bundy, Conselice, Newman, Schiminovich, Le Floc'h, Coil, Rieke, Lotz, Primack, Barmby, Cooper, Davis, Ellis, Fazio, Guhathakurta, Huang, Kassin, Martin, Phillips, Rich, Small, Willmer and Wilson2007; Renzini and Peng Reference Renzini and Peng2015).Footnote ⁴ This characteristic sSFR has been decreasing for the past ~ 10 Gyr (Noeske et al. Reference Noeske, Weiner, Faber, Papovich, Koo, Somerville, Bundy, Conselice, Newman, Schiminovich, Le Floc'h, Coil, Rieke, Lotz, Primack, Barmby, Cooper, Davis, Ellis, Fazio, Guhathakurta, Huang, Kassin, Martin, Phillips, Rich, Small, Willmer and Wilson2007; Speagle et al. Reference Speagle, Steinhardt, Capak and Silverman2014). Quiescent galaxies, by contrast, have a very small sSFR and have formed few stars over the past gigayears. Thus, SFGs have a mixture of young and old stars, while quiescents predominantly have old, lower mass stars. Between the red and blue galaxies lie a transitional group of ‘green valley’ galaxies, which the Milky Way is entering (Licquia et al. Reference Licquia, Newman and Brinchmann2015), while starburst galaxies have abnormally high sSFR for their epoch (Elbaz et al. Reference Elbaz, Dickinson, Hwang, Díaz-Santos, Magdis, Magnelli, Le Borgne, Galliano, Pannella, Chanial, Armus, Charmandaris, Daddi, Aussel, Popesso, Kartaltepe, Altieri, Valtchanov, Coia, Dannerbauer, Dasyra, Leiton, Mazzarella, Alexander, Buat, Burgarella, Chary, Gilli, Ivison, Juneau, Le Floc'h, Lutz, Morrison, Mullaney, Murphy, Pope, Scott, Brodwin, Calzetti, Cesarsky, Charlot, Dole, Eisenhardt, Ferguson, Förster Schreiber, Frayer, Giavalisco, Huynh, Koekemoer, Papovich, Reddy, Surace, Teplitz, Yun and Wilson2011). A small fraction of galaxies at any given time have active galactic nuclei, with the central black holes emitting prodigious electromagnetic and particle radiation.

The dichotomy in sSFR, and its cosmic history, arises from the evolution of the gas supply in galaxies, a major factor in traversability. Star formation requires the presence of sufficient amounts of cold (generally molecular) gas that can collapse, with a larger star-formation density in regions of dense gas (Kennicutt Reference Kennicutt1998). Galaxies are fed gas from the surrounding intergalactic medium, and convert some of that gas into stars at an equilibrium rate set by feedback from the young stars, while some is ejected in a wind (Bouché et al. Reference Bouché, Dekel, Genzel, Genel, Cresci, Förster Schreiber, Shapiro, Davies and Tacconi2010). SFGs notionally climb up the main sequence, from dwarf to giant, but the main sequence as a whole has falling sSFR with time, and blue galaxies had much higher gas fractions (${f_{\rm gas}} \equiv {M_{\rm gas}} / ( {M_{\rm gas}} + {M_{\star }}$)) and star-formation rates in the early Universe (Santini et al. Reference Santini, Maiolino, Magnelli, Lutz, Lamastra, Li Causi, Eales, Andreani, Berta, Buat, Cooray, Cresci, Daddi, Farrah, Fontana, Franceschini, Genzel, Granato, Grazian, Le Floc'h, Magdis, Magliocchetti, Mannucci, Menci, Nordon, Oliver, Popesso, Pozzi, Riguccini, Rodighiero, Rosario, Salvato, Scott, Silva, Tacconi, Viero, Wang, Wuyts and Xu2014; Tacconi et al. Reference Tacconi, Genzel, Saintonge, Combes, García-Burillo, Neri, Bolatto, Contini, Förster Schreiber, Lilly, Lutz, Wuyts, Accurso, Boissier, Boone, Bouché, Bournaud, Burkert, Carollo, Cooper, Cox, Feruglio, Freundlich, Herrera-Camus, Juneau, Lippa, Naab, Renzini, Salome, Sternberg, Tadaki, Übler, Walter, Weiner and Weiss2018). A large main sequence SFG with ${M_{\star }} \sim 10^{11}\, {\rm M}_{\odot }$ in the present day can be expected to have a gas mass ${M_{\rm gas}} \sim 10^{10}\, {\rm M}_{\odot }$ and ${\rm SFR} \lesssim 10\, {\rm M}_{\odot }\, {\rm yr}^{-1}$, but a similarly sized galaxy 10 billion years ago would have had ${M_{\rm gas}} \sim 10^{11}\, {\rm M}_{\odot }$ and ${\rm SFR} \gtrsim 100\, {\rm M}_{\odot }\, {\rm yr}^{-1}$ (Daddi et al. Reference Daddi, Bournaud, Walter, Dannerbauer, Carilli, Dickinson, Elbaz, Morrison, Riechers, Onodera, Salmi, Krips and Stern2010; Tacconi et al. Reference Tacconi, Neri, Genzel, Combes, Bolatto, Cooper, Wuyts, Bournaud, Burkert, Comerford, Cox, Davis, Förster Schreiber, García-Burillo, Gracia-Carpio, Lutz, Naab, Newman, Omont, Saintonge, Shapiro Griffin, Shapley, Sternberg and Weiner2013). As both the star-formation rate and the gas fraction fall with time, the interstellar media of SFGs are becoming more traversable. The generations of massive stars born and dying during this growth results in metals accumulating in the ISM, and high-${M_{\star }}$ galaxies have higher metallicity even if they formed all their stars rapidly in the early Universe (Mannucci et al. Reference Mannucci, Cresci, Maiolino, Marconi and Gnerucci2010). A higher metallicity leads to more planets, but also more dust in ISM, favouring habitability at the expense of SFG traversability.

Growth in the main sequence mode does not continue forever, but is halted when the SFG runs out of gas. This can be a sudden process, triggered by a galaxy merger funnelling gas into the core to be consumed in a brief spectacular starburst, but more often is the result of a slower exhaustion or ‘strangulation’ of the external gas supply that results in the galaxy crossing the green valley (Schawinski et al. Reference Schawinski, Urry, Simmons, Fortson, Kaviraj, Keel, Lintott, Masters, Nichol, Sarzi, Skibba, Treister, Willett, Wong and Yi2014; Peng et al. Reference Peng, Maiolino and Cochrane2015; Bremer et al. Reference Bremer, Phillipps, Kelvin, De Propris, Kennedy, Moffett, Bamford, Davies, Driver, Häußler, Holwerda, Hopkins, James, Liske, Percival and Taylor2018). The building of a bulge may also be a factor in this quenching process: many green valley galaxies are early-type spirals and many quiescents appear to have stellar disks within them (Bremer et al. Reference Bremer, Phillipps, Kelvin, De Propris, Kennedy, Moffett, Bamford, Davies, Driver, Häußler, Holwerda, Hopkins, James, Liske, Percival and Taylor2018). Within a couple of billion years, the quenched former SFGs are red and quiescent. They are broadly expected to have little gas – and thus be favourable in terms of traversability. However, they can retain a residue of gas from their past lives as SFGs, accrete it from passing smaller galaxies with their tides, and replenish it as old stars eject their envelopes, so their ISMs are not entirely empty (Davis et al. Reference Davis, Alatalo, Sarzi, Bureau, Young, Blitz, Serra, Crocker, Krajnović, McDermid, Bois, Bournaud, Cappellari, Davies, Duc, de Zeeuw, Emsellem, Khochfar, Kuntschner, Lablanche, Morganti, Naab, Oosterloo, Scott and Weijmans2011; Young et al. Reference Young, Scott, Serra, Alatalo, Bayet, Blitz, Bois, Bournaud, Bureau, Crocker, Cappellari, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, McDermid, Morganti, Naab, Oosterloo, Sarzi and Weijmans2014).

Star-forming galaxies: a cold, shifting labyrinth

The ‘space climate’ of SFGs is shaped by the chaotic and violent interplay between gas and star formation. The space weather within them is thus very chaotic, inhomogeneous and constantly changing. On a large scale, the ISM is complicated by its ‘multiphase’ nature, where each phase has a different temperature and pressure. They all have roughly the same pressure, so hot gas is less dense (Draine Reference Draine2011). Classically, there are three phases to the ISM, although this is a great simplification (Cox Reference Cox2005). The cool/cold phase (~ 10 − 100 K) has dense atomic neutral (HI) and molecular (H₂) gas, the warm phase (~ 10 000 K) of neutral (HI) and ionized (HII) gas, and a third hot phase (~ 10⁶ K) is carved out by supernovae and young stellar clusters. In the Milky Way, the hot phase fills about half the volume, a patchwork of interstellar space covering much of the Galactic disk where gas and dust pose little obstacle to starfarers. Most of the remaining volume is filled with warm gas, and only small clouds and filaments with dense cold gas.Footnote ⁵ The different phases of gas also have different height distributions, with cold gas concentrated close to the Galactic plane, while warm ionized gas rises several hundred parsecs above the plane, punctured by chimneys of hot gas over star clusters (Norman and Ikeuchi Reference Norman and Ikeuchi1989; Cox Reference Cox2005). The standard three phase structure is expected to apply to most contemporary spiral galaxies, but it is not universal. Intensely SFGs, which includes both starbursts in galactic cores and most large SFGs in the early Universe, are dominated by vast amounts of cold gas (Daddi et al. Reference Daddi, Bournaud, Walter, Dannerbauer, Carilli, Dickinson, Elbaz, Morrison, Riechers, Onodera, Salmi, Krips and Stern2010; Tacconi et al. Reference Tacconi, Genzel, Saintonge, Combes, García-Burillo, Neri, Bolatto, Contini, Förster Schreiber, Lilly, Lutz, Wuyts, Accurso, Boissier, Boone, Bouché, Bournaud, Burkert, Carollo, Cooper, Cox, Feruglio, Freundlich, Herrera-Camus, Juneau, Lippa, Naab, Renzini, Salome, Sternberg, Tadaki, Übler, Walter, Weiner and Weiss2018). A high-volume phase of hot ISM may also exist in weaker starbursts, excavated by the high rate density of supernovae, but this is not agreed upon (Strickland and Heckman Reference Strickland and Heckman2007; Krumholz et al. Reference Krumholz, Crocker, Xu, Lazarian, Rosevear and Bedwell-Wilson2020).

These phases are not simple clouds or smooth media. Turbulence is found in at least the cold and warm phases of SFGs, with σ_gas ~ 10 km s⁻¹ in contemporary main sequence SFGs and up to ~ 100 km s⁻¹ in starbursts and high-redshift SFGs (e.g. Downes and Solomon Reference Downes and Solomon1998; Dib et al. Reference Dib, Bell and Burkert2006; Law et al. Reference Law, Steidel, Erb, Larkin, Pettini, Shapley and Wright2009). Turbulence creates random density fluctuations at a variety of length scales, the largest at the outer scale ℓ_outer ~ 0.1 −1 kpc (e.g. Draine Reference Draine2011). The amplitude of these fluctuations grows according to the Mach number of the turbulence (ratio of σ_gas to gas sound speed), surpassing the mean density when the turbulence becomes supersonic. In ISM with supersonic turbulence, most of the matter is concentrated into dense transient clumps while most of the volume is rarefied, with $\sim 10\percnt$ at any given time having $\rho _{\rm gas} \ge {\langle {\rho _{\rm gas}} \rangle }$ (Padoan et al. Reference Padoan, Nordlund and Jones1997; Ostriker et al. Reference Ostriker, Stone and Gammie2001; Federrath Reference Federrath2013). The fluctuations evolve rapidly, growing and dissipating within one flow crossing time ℓ_outer/σ_gas (Stone et al. Reference Stone, Ostriker and Gammie1998; Mac Low Reference Mac Low1999), resulting in constantly shifting space weather in ISM with supersonic turbulence. The Mach number of the turbulence is much higher in cold gas due to the lower sound speed, and thus the cold, denser gas has large density contrasts, with most of the mass concentrated into transient clumps, with the densest forming molecular clouds. Thus, even in starbursts with a mean ISM density hundreds of times higher than the Milky Way's, at any given location the gas and dust density may be relatively low. Yet although these relatively traversable conditions are the median, they are interrupted by intervals where traversability is low. Thus a map of the traversability of an SFG would constantly shift over millions of years.

If gas and dust truly are major impediments to interstellar travel, SFGs are essentially mazes where travel can occur freely along some paths (particularly in the hot ISM) but are walled off by dense matter along other paths. The free paths are not necessarily connected, which can lead to a mosaic of disconnected regions in a galaxy, between which travel is impractical. This labyrinthine patchwork does suggest that percolation barriers to interstellar migration could apply on short enough timescales. Ultimately, however, the dynamic nature of the ISM thwarts these blockages on long enough timescales: the walls of the labyrinth are constantly moving, opening up new paths and connecting some regions even as they close off and split others.

The same cannot be said if the main hazard is CRs. Plasma waves in the ISM trap CRs and prevent them from simply flying out of a galaxy. While GeV CRs are thought to be mainly accelerated in supernova remnants and bubbles surrounding young star clusters, they diffuse many hundreds of parsecs from these sites, when they are not destroyed by interactions with the ISM (Murphy et al. Reference Murphy, Helou, Kenney, Armus and Braun2008). Nor are they confined to the galactic plane: isotopic evidence in the Milky Way implies they can wander up and down by at least a kiloparsec off the plane until they start to escape (Lukasiak et al. Reference Lukasiak, Ferrando, McDonald and Webber1994; Strong and Moskalenko Reference Strong and Moskalenko1998). As such, the bulk CR density is not expected to vary too much within the stellar disks of most SFGs, although in the Milky Way it slightly decreases with distance from the centre(Acero et al. Reference Acero, Ackermann, Ajello, Albert, Baldini, Ballet, Barbiellini, Bastieri, Bellazzini, Bissaldi, Bloom, Bonino, Bottacini, Brandt, Bregeon, Bruel, Buehler, Buson, Caliandro, Cameron, Caragiulo, Caraveo, Casandjian, Cavazzuti, Cecchi, Charles, Chekhtman, Chiang, Chiaro, Ciprini, Claus, Cohen-Tanugi, Conrad, Cuoco, Cutini, D'Ammando, de Angelis, de Palma, Desiante, Digel, Di Venere, Drell, Favuzzi, Fegan, Ferrara, Focke, Franckowiak, Funk, Fusco, Gargano, Gasparrini, Giglietto, Giordano, Giroletti, Glanzman, Godfrey, Grenier, Guiriec, Hadasch, Harding, Hayashi, Hays, Hewitt, Hill, Horan, Hou, Jogler, Jóhannesson, Kamae, Kuss, Landriu, Larsson, Latronico, Li, Li, Longo, Loparco, Lovellette, Lubrano, Maldera, Malyshev, Manfreda, Martin, Mayer, Mazziotta, McEnery, Michelson, Mirabal, Mizuno, Monzani, Morselli, Nuss, Ohsugi, Omodei, Orienti, Orlando, Ormes, Paneque, Pesce-Rollins, Piron, Pivato, Rainò, Rando, Razzano, Razzaque, Reimer, Reimer, Remy, Renault, Sánchez-Conde, Schaal, Schulz, Sgrò, Siskind, Spada, Spandre, Spinelli, Strong, Suson, Tajima, Takahashi, Thayer, Thompson, Tibaldo, Tinivella, Torres, Tosti, Troja, Vianello, Werner, Wood, Wood, Zaharijas and Zimmer2016). There is no escape from the CR hazard unless one moves far into the galactic halo, potentially hindering travel between most of the stars in an SFG. The mean background level of CRs varies dramatically from SFG to SFG, however. Gamma-ray and radio observations indicate that the CR energy density increases rapidly with star formation rate: it is significantly lower in dwarf irregular galaxies (Abdo et al. Reference Abdo, Ackermann, Ajello, Baldini, Ballet, Barbiellini, Bastieri, Bechtol, Bellazzini, Berenji, Blandford, Bloom, Bonamente, Borgland, Bouvier, Brandt, Bregeon, Brez, Brigida, Bruel, Buehler, Buson, Caliandro, Cameron, Caraveo, Carrigan, Casandjian, Cecchi, Çelik, Charles, Chekhtman, Cheung, Chiang, Ciprini, Claus, Cohen-Tanugi, Conrad, Dermer, de Palma, Digel, Silva, Drell, Dubois, Dumora, Favuzzi, Fegan, Fukazawa, Funk, Fusco, Gargano, Gasparrini, Gehrels, Germani, Giglietto, Giordano, Giroletti, Glanzman, Godfrey, Grenier, Grondin, Grove, Guiriec, Hadasch, Harding, Hayashida, Hays, Horan, Hughes, Jean, Jóhannesson, Johnson, Johnson, Kamae, Katagiri, Kataoka, Kerr, Knödlseder, Kuss, Lande, Latronico, Lee, Lemoine-Goumard, Llena Garde, Longo, Loparco, Lovellette, Lubrano, Makeev, Martin, Mazziotta, McEnery, Michelson, Mitthumsiri, Mizuno, Monte, Monzani, Morselli, Moskalenko, Murgia, Nakamori, Naumann-Godo, Nolan, Norris, Nuss, Ohsugi, Okumura, Omodei, Orlando, Ormes, Panetta, Parent, Pelassa, Pepe, Pesce-Rollins, Piron, Porter, Rainò, Rando, Razzano, Reimer, Reimer, Reposeur, Ripken, Ritz, Romani, Sadrozinski, Sander, Saz Parkinson, Scargle, Sgrò, Siskind, Smith, Smith, Spandre, Spinelli, Strickman, Strong, Suson, Takahashi, Takahashi, Tanaka, Thayer, Thayer, Thompson, Tibaldo, Torres, Tosti, Tramacere, Uchiyama, Usher, Vandenbroucke, Vasileiou, Vilchez, Vitale, Waite, Wang, Winer, Wood, Yang, Ylinen and Ziegler2010a), but hundreds of times higher in starburst regions and probably large SFGs at high redshift (e.g. VERITAS Collaboration Reference Acciari, Aliu, Arlen, Aune, Bautista, Beilicke, Benbow, Boltuch, Bradbury, Buckley, Bugaev, Byrum, Cannon, Celik, Cesarini, Chow, Ciupik, Cogan, Colin, Cui, Dickherber, Duke, Fegan, Finley, Finnegan, Fortin, Fortson, Furniss, Galante, Gall, Gibbs, Gillanders, Godambe, Grube, Guenette, Gyuk, Hanna, Holder, Horan, Hui, Humensky, Imran, Kaaret, Karlsson, Kertzman, Kieda, Kildea, Konopelko, Krawczynski, Krennrich, Lang, Lebohec, Maier, McArthur, McCann, McCutcheon, Millis, Moriarty, Mukherjee, Nagai, Ong, Otte, Pandel, Perkins, Pizlo, Pohl, Quinn, Ragan, Reyes, Reynolds, Roache, Rose, Schroedter, Sembroski, Smith, Steele, Swordy, Theiling, Thibadeau, Varlotta, Vassiliev, Vincent, Wagner, Wakely, Ward, Weekes, Weinstein, Weisgarber, Williams, Wissel, Wood and Zitzer2009; Papadopoulos Reference Papadopoulos2010).

Magnetic fields are generally observed to have a nearly equal energy density as CRs in local star-forming spiral galaxies, although the fields likely become dominant in the most intense starburst regions (e.g. Thompson et al. Reference Thompson, Quataert, Waxman, Murray and Martin2006; Yoast-Hull et al. Reference Yoast-Hull, Gallagher and Zweibel2016). Their ubiquitous presence is attested by the radio synchrotron emission observed from SFGs throughout the Universe, and they may be maintained by a dynamo driven by ISM turbulence even in the early Universe (Arshakian et al. Reference Arshakian, Beck, Krause and Sokoloff2009).

Of course, these conclusions depend on how much of a hindrance each part of the ISM is. It is possible that ETIs use RAIR propulsion and favour cold, dense ISM to acquire propellant – then SFGs may instead have fleeting ‘islands’ of ETI travel separated by empty gulfs instead of a maze. Starbursts and high-redshift SFGs may then actually be favoured for interstellar travel because of their dense gas reservoirs, if ETIs can cope with the extreme levels of CRs in them. Or the reality may be more nuanced, with a mosaic of different interstellar travel methods, each adapted to presiding local conditions.

Quiescent galaxies: placid until they're not

The space climate of quiescent galaxies is very different. While SFGs are filled with stormy and chaotic ‘weather’, the ‘atmosphere’ of these galaxies is generally thinner and more placid. In some galaxies this seeming quiet can be disturbed by the wildness of star formation – or long galaxy-wide disasters when the central black hole turns on.

Like SFGs, quiescent galaxies come in a vast ranges of mass and size, from ultrafaint dwarf spheroidal galaxies to supergiant ellipticals in the hearts of galaxy clusters. Generally, the larger the galaxy, the larger the stellar velocity dispersion: in dwarf spheroidals, $\sigma _{\star }$ can be as small as 1 km s⁻¹ (Simon Reference Simon2019); in the hearts of the largest ellipticals, over 300 km s⁻¹ (Cappellari et al. Reference Cappellari, Scott, Alatalo, Blitz, Bois, Bournaud, Bureau, Crocker, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, McDermid, Morganti, Naab, Oosterloo, Sarzi, Serra, Weijmans and Young2013; Norris et al. Reference Norris, Kannappan, Forbes, Romanowsky, Brodie, Faifer, Huxor, Maraston, Moffett, Penny, Pota, Smith-Castelli, Strader, Bradley, Eckert, Fohring, McBride, Stark and Vaduvescu2014). Moderate to large ETGs, which contain more stellar mass, have been discovered to fall into two broad categories, the fast rotators and the slow rotators (Emsellem et al. Reference Emsellem, Cappellari, Krajnović, van de Ven, Bacon, Bureau, Davies, de Zeeuw, Falcón-Barroso, Kuntschner, McDermid, Peletier and Sarzi2007; Cappellari Reference Cappellari2016). Fast rotators, which include the lenticulars and the majority of ellipticals, have stellar disks like their progenitors, the spirals, although their bulge components are usually dominant (Krajnović et al. Reference Krajnović, Alatalo, Blitz, Bois, Bournaud, Bureau, Cappellari, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Kuntschner, McDermid, Morganti, Naab, Oosterloo, Sarzi, Scott, Serra, Weijmans and Young2013). Slow rotators are a diverse class, with some being disks or containing two counterrotating populations. The largest slow rotators are giant elliptical galaxies with rounder ‘boxy’ shapes (Kormendy and Bender Reference Kormendy and Bender1996), and thus I will refer to them as boxy ETGs. The dichotomy extends to many galaxy properties (Kormendy et al. Reference Kormendy, Fisher, Cornell and Bender2009) and appears to have a profound effect on the ISM.

The classical expectation of quiescent galaxy ISMs is that they do not have any, but observations have shown that this is not universally true. A significant fraction of ETGs (particularly fast rotators outside of galaxy clusters) have cool atomic and molecular gas. Estimates of the ETG fraction with HI range from 10 to $70\percnt$ depending on environment and the sensitivity to gas (Morganti et al. Reference Morganti, de Zeeuw, Oosterloo, McDermid, Krajnović, Cappellari, Kenn, Weijmans and Sarzi2006; Serra et al. Reference Serra, Oosterloo, Morganti, Alatalo, Blitz, Bois, Bournaud, Bureau, Cappellari, Crocker, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, Lablanche, McDermid, Naab, Sarzi, Scott, Trager, Weijmans and Young2012), and $\sim 20\percnt$ have molecular gas (Young et al. Reference Young, Bureau, Davis, Combes, McDermid, Alatalo, Blitz, Bois, Bournaud, Cappellari, Davies, de Zeeuw, Emsellem, Khochfar, Krajnović, Kuntschner, Lablanche, Morganti, Naab, Oosterloo, Sarzi, Scott, Serra and Weijmans2011). In some cases, the amount of cool gas can reach levels comparable to the Milky Way, but in most cases the gas mass is much lower ($\sim 10^{6} - 10^{9}\, {\rm M}_{\odot }$). Dust is also commonly present in ETGs, if usually at much lower abundances than in normal star-forming spirals (Goudfrooij and de Jong Reference Goudfrooij and de Jong1995; Ferrari et al. Reference Ferrari, Pastoriza, Macchetto and Caon1999). Smith et al. (Reference Smith, Gomez, Eales, Ciesla, Boselli, Cortese, Bendo, Baes, Bianchi, Clemens, Clements, Cooray, Davies, De Looze, di Serego Alighieri, Fritz, Gavazzi, Gear, Madden, Mentuch, Panuzzo, Pohlen, Spinoglio, Verstappen, Vlahakis, Wilson and Xilouris2012) detect dust in 62% of lenticular galaxies and 24% of ellipticals. In those galaxies with detections, the mass ratio of dust to stars is ~ 10⁻⁴, compared to ~ 10⁻³ in the Milky Way (Smith et al. Reference Smith, Gomez, Eales, Ciesla, Boselli, Cortese, Bendo, Baes, Bianchi, Clemens, Clements, Cooray, Davies, De Looze, di Serego Alighieri, Fritz, Gavazzi, Gear, Madden, Mentuch, Panuzzo, Pohlen, Spinoglio, Verstappen, Vlahakis, Wilson and Xilouris2012). The dust is often concentrated in a small disk around the galactic nucleus or in filaments, so ETIs may simply be able to dodge what dust is present (Tran et al. Reference Tran, Tsvetanov, Ford, Davies, Jaffe, van den Bosch and Rest2001). This is not always the case, however (Goudfrooij and de Jong Reference Goudfrooij and de Jong1995; Temi et al. Reference Temi, Brighenti and Mathews2007).

Cool gas is fuel for star formation, suggesting a CR threat. Star formation is indeed observed, usually in fast rotator galaxies, as is the radio synchrotron associated with CRs (Shapiro et al. Reference Shapiro, Falcón-Barroso, van de Ven, de Zeeuw, Sarzi, Bacon, Bolatto, Cappellari, Croton, Davies, Emsellem, Fakhouri, Krajnović, Kuntschner, McDermid, Peletier, van den Bosch and van der Wolk2010; Young et al. Reference Young, Scott, Serra, Alatalo, Bayet, Blitz, Bois, Bournaud, Bureau, Crocker, Cappellari, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, McDermid, Morganti, Naab, Oosterloo, Sarzi and Weijmans2014; Kokusho et al. Reference Kokusho, Kaneda, Bureau, Suzuki, Murata, Kondo and Yamagishi2017; Nyland et al. Reference Nyland, Young, Wrobel, Davis, Bureau, Alatalo, Morganti, Duc, de Zeeuw, McDermid, Crocker and Oosterloo2017). According to Shapiro et al. (Reference Shapiro, Falcón-Barroso, van de Ven, de Zeeuw, Sarzi, Bacon, Bolatto, Cappellari, Croton, Davies, Emsellem, Fakhouri, Krajnović, Kuntschner, McDermid, Peletier, van den Bosch and van der Wolk2010), the star-formation appears to have two ‘modes’, either spread throughout the galaxy or concentrated near the core. The former only occurs after the ETG receives a windfall of cool gas from an encounter. Shapiro et al. (Reference Shapiro, Falcón-Barroso, van de Ven, de Zeeuw, Sarzi, Bacon, Bolatto, Cappellari, Croton, Davies, Emsellem, Fakhouri, Krajnović, Kuntschner, McDermid, Peletier, van den Bosch and van der Wolk2010) speculates that nuclear star formation occurs in occasional spurts within all fast rotators, supplied by gas from the old stellar population. Either way, however, star formation is a brief episode in a largely quiescent history, limiting the CR threat temporally. Nuclear star formation also likely results in a CR threat only in the inner regions of the host galaxy. Because there usually is less cool gas than in proper SFGs, the star-formation rate should also be lower. There is also evidence that the star-formation efficiency – the rate at which gas is converted into stars – is lower in elliptical galaxies, which would suppress the SFR in these galaxies even further (Martig et al. Reference Martig, Crocker, Bournaud, Emsellem, Gabor, Alatalo, Blitz, Bois, Bureau, Cappellari, Davies, Davis, Dekel, de Zeeuw, Duc, Falcón-Barroso, Khochfar, Krajnović, Kuntschner, Morganti, McDermid, Naab, Oosterloo, Sarzi, Scott, Serra, Griffin, Teyssier, Weijmans and Young2013; Kokusho et al. Reference Kokusho, Kaneda, Bureau, Suzuki, Murata, Kondo and Yamagishi2017; Nyland et al. Reference Nyland, Young, Wrobel, Davis, Bureau, Alatalo, Morganti, Duc, de Zeeuw, McDermid, Crocker and Oosterloo2017 dispute this, however). Finally, the radio-faintness of ETGs with star formation could suggest that CRs escape easily in these galaxies (Nyland et al. Reference Nyland, Young, Wrobel, Davis, Bureau, Alatalo, Morganti, Duc, de Zeeuw, McDermid, Crocker and Oosterloo2017). All of these factors would limit the CR energy density in intensity, space and time.

Quiescent galaxies also have hot gas supplied by dying old stars. This gas gets stirred up by the stellar population, thermalizing it to 10⁶ −10⁷ K where the sound speed is roughly the stellar velocity dispersion and further heated by the Type Ia supernovae of old white dwarfs (Ciotti et al. Reference Ciotti, D'Ercole, Pellegrini and Renzini1991; Mathews and Brighenti Reference Mathews and Brighenti2003). The fate of this gas depends on the depth of the galaxy's gravitational potential. In the majority of ETGs, the hot gas is heated until it escapes, as borne out by a lack of X-ray emission (Mulchaey and Jeltema Reference Mulchaey and Jeltema2010; Sarzi et al. Reference Sarzi, Alatalo, Blitz, Bois, Bournaud, Bureau, Cappellari, Crocker, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, Lablanche, McDermid, Morganti, Naab, Oosterloo, Scott, Serra, Young and Weijmans2013). The most massive ETGs (including the boxy ETGs) hold onto their hot gas, which builds over time. Because the hot gas is glowing so brightly X-rays, it is losing energy, cooling down and sinking towards the galactic centre. Some of this gas may form stars, but the real trouble is that it provides a fuel source for the central black hole, which in turns powers relativistic jets seen in radio. Indeed, all massive ETGs appear to be radio galaxies on some level (Brown et al. Reference Brown, Jannuzi, Floyd and Mould2011; Sabater et al. Reference Sabater, Best, Hardcastle, Shimwell, Tasse, Williams, Brüggen, Cochrane, Croston, de Gasperin, Duncan, Gürkan, Mechev, Morabito, Prandoni, Röttgering, Smith, Harwood, Mingo, Mooney and Saxena2019), though most of the time the activity appears to be confined to a small core (Slee et al. Reference Slee, Sadler, Reynolds and Ekers1994; Nyland et al. Reference Nyland, Young, Wrobel, Sarzi, Morganti, Alatalo, Blitz, Bournaud, Bureau, Cappellari, Crocker, Davies, Davis, de Zeeuw, Duc, Emsellem, Khochfar, Krajnović, Kuntschner, McDermid, Naab, Oosterloo, Scott, Serra and Weijmans2016). From time to time, high accretion onto the central black hole fuels extraordinary episodes of activity in which jets blast their way through the galaxy and inflate lobes of CRs. The CR lobes can fill much of the galaxy's volume with dangerous levels of radiation, which can be ten to hundreds of times the Galactic background (McNamara and Nulsen Reference McNamara and Nulsen2007; de Gasperin et al. Reference de Gasperin, Orrú, Murgia, Merloni, Falcke, Beck, Beswick, Bîrzan, Bonafede, Brüggen, Brunetti, Chyży, Conway, Croston, Enßlin, Ferrari, Heald, Heidenreich, Jackson, Macario, McKean, Miley, Morganti, Offringa, Pizzo, Rafferty, Röttgering, Shulevski, Steinmetz, Tasse, van der Tol, van Driel, van Weeren, van Zwieten, Alexov, Anderson, Asgekar, Avruch, Bell, Bell, Bentum, Bernardi, Best, Breitling, Broderick, Butcher, Ciardi, Dettmar, Eisloeffel, Frieswijk, Gankema, Garrett, Gerbers, Griessmeier, Gunst, Hassall, Hessels, Hoeft, Horneffer, Karastergiou, Köhler, Koopman, Kuniyoshi, Kuper, Maat, Mann, Mevius, Mulcahy, Munk, Nijboer, Noordam, Paas, Pandey, Pandey, Polatidis, Reich, Schoenmakers, Sluman, Smirnov, Sobey, Stappers, Swinbank, Tagger, Tang, van Bemmel, van Cappellen, van Duin, van Haarlem, van Leeuwen, Vermeulen, Vocks, White, Wise, Wucknitz and Zarka2012; McNamara and Nulsen Reference McNamara and Nulsen2012; Pfrommer Reference Pfrommer2013). Magnetic field strengths of tens of microgauss are estimated in these regions (de Gasperin et al. Reference de Gasperin, Orrú, Murgia, Merloni, Falcke, Beck, Beswick, Bîrzan, Bonafede, Brüggen, Brunetti, Chyży, Conway, Croston, Enßlin, Ferrari, Heald, Heidenreich, Jackson, Macario, McKean, Miley, Morganti, Offringa, Pizzo, Rafferty, Röttgering, Shulevski, Steinmetz, Tasse, van der Tol, van Driel, van Weeren, van Zwieten, Alexov, Anderson, Asgekar, Avruch, Bell, Bell, Bentum, Bernardi, Best, Breitling, Broderick, Butcher, Ciardi, Dettmar, Eisloeffel, Frieswijk, Gankema, Garrett, Gerbers, Griessmeier, Gunst, Hassall, Hessels, Hoeft, Horneffer, Karastergiou, Köhler, Koopman, Kuniyoshi, Kuper, Maat, Mann, Mevius, Mulcahy, Munk, Nijboer, Noordam, Paas, Pandey, Pandey, Polatidis, Reich, Schoenmakers, Sluman, Smirnov, Sobey, Stappers, Swinbank, Tagger, Tang, van Bemmel, van Cappellen, van Duin, van Haarlem, van Leeuwen, Vermeulen, Vocks, White, Wise, Wucknitz and Zarka2012). Thus these larger ETGs may have long time periods where interstellar travel is relatively safe from CRs punctuated by catastrophes lasting tens of millions of years wherein radio jets and lobes make interstellar space extremely hazardous.

Another potential source of CRs in quiescent galaxies are the Type Ia supernovae in old stellar populations (as in Lien and Fields Reference Lien and Fields2012), but it is not clear how many CRs these accelerate or how the radiation propagates. Likewise, magnetic fields in the quiescent ISM are very poorly constrained, although they probably are weaker than in the Milky Way (Seta et al. Reference Seta, Rodrigues, Federrath and Hales2021).

Role of size evolution of galaxies

Galaxies grew from the inside-out, with the denser cores forming at high redshift (e.g. Nelson et al. Reference Nelson, van Dokkum, Förster Schreiber, Franx, Brammer, Momcheva, Wuyts, Whitaker, Skelton, Fumagalli, Hayward, Kriek, Labbé, Leja, Rix, Tacconi, van der Wel, van den Bosch, Oesch, Dickey and Ulf Lange2016). This secular growth is supplemented by galaxy mergers. Today's galaxies were more compact in the distant past, which partly offsets the lower stellar mass when considering $\rho _{\star }$ and the typical duration of interstellar journeys. This size evolution is not solely due to galaxies with less stellar mass being smaller: in fact, galaxies of any fixed stellar mass were more compact in the early Universe. At z ~ 2, SFGs were about half as wide as present galaxies of the same mass, implying four times the stellar surface density (Trujillo et al. Reference Trujillo, Förster Schreiber, Rudnick, Barden, Franx, Rix, Caldwell, McIntosh, Toft, Häussler, Zirm, van Dokkum, Labbé, Moorwood, Röttgering, van der Wel, van der Werf and van Starkenburg2006b; van der Wel et al. Reference van der Wel, Franx, van Dokkum, Skelton, Momcheva, Whitaker, Brammer, Bell, Rix, Wuyts, Ferguson, Holden, Barro, Koekemoer, Chang, McGrath, Häussler, Dekel, Behroozi, Fumagalli, Leja, Lundgren, Maseda, Nelson, Wake, Patel, Labbé, Faber, Grogin and Kocevski2014). Note that at fixed ${M_{\star }}$, we select different galaxies at different cosmic epochs. When following an individual SFG, $\rho _{\star }$ increases over time in all regions of the galaxy (van Dokkum et al. Reference van Dokkum, Leja, Nelson, Patel, Skelton, Momcheva, Brammer, Whitaker, Lundgren, Fumagalli, Conroy, Förster Schreiber, Franx, Kriek, Labbé, Marchesini, Rix, van der Wel and Wuyts2013). Still, the typical inhabited planet of the early Universe could very well have been located in a relatively compact and massive galaxy, simply because the sparser low-${M_{\star }}$ galaxies had not formed many stars yet.

Despite the smaller radii, the median population-weighted $\rho _{\star }$ probably was not much greater than at present. SFGs in the early Universe had higher levels of turbulence, which puffs out their gaseous disk and the stars formed from them (e.g. Genzel et al. Reference Genzel, Burkert, Bouché, Cresci, Förster Schreiber, Shapley, Shapiro, Tacconi, Buschkamp, Cimatti, Daddi, Davies, Eisenhauer, Erb, Genel, Gerhard, Hicks, Lutz, Naab, Ott, Rabien, Renzini, Steidel, Sternberg and Lilly2008). Wisnioski et al. (Reference Wisnioski, Förster Schreiber, Wuyts, Wuyts, Bandara, Wilman, Genzel, Bender, Davies, Fossati, Lang, Mendel, Beifiori, Brammer, Chan, Fabricius, Fudamoto, Kulkarni, Kurk, Lutz, Nelson, Momcheva, Rosario, Saglia, Seitz, Tacconi and van Dokkum2015) estimates the typical σ_gas of SFGs was ~ 50 km s⁻¹ at z ~ 2.3, whereas it is now ~ 10 km s⁻¹.Footnote ⁶ As the disk thickness-to-radius ratio varies proportionally to σ_gas, the more compact radial scales at high redshift were offset by greater vertical scales, with a relatively weak net effect on traversability.

Although quiescent galaxies were comparatively rare in the early Universe, they were even more compact, about one-quarter the radius of present-day quiescent galaxies with the same ${M_{\star }}$ (Trujillo et al. Reference Trujillo, Feulner, Goranova, Hopp, Longhetti, Saracco, Bender, Braito, Della Ceca, Drory, Mannucci and Severgnini2006a; van der Wel et al. Reference van der Wel, Franx, van Dokkum, Skelton, Momcheva, Whitaker, Brammer, Bell, Rix, Wuyts, Ferguson, Holden, Barro, Koekemoer, Chang, McGrath, Häussler, Dekel, Behroozi, Fumagalli, Leja, Lundgren, Maseda, Nelson, Wake, Patel, Labbé, Faber, Grogin and Kocevski2014). These ‘red nuggets’ had ${M_{\star }}$ of $\sim 10^{10} - 10^{11}\, {\rm M}_{\odot }$ concentrated into a disk of scale radius ~ 1 kpc, with $\rho _{\star }$ hundreds of times that found in the Milky Way (van Dokkum et al. Reference van Dokkum, Franx, Kriek, Holden, Illingworth, Magee, Bouwens, Marchesini, Quadri, Rudnick, Taylor and Toft2008). Most of these galaxies have seemingly disappeared. Nonetheless, environments of similar $\rho _{\star }$ – and presumably traversability – are found in the inner kiloparsecs of today's massive early-type galaxies, suggesting that some of them are red nuggets that accreted stars during galaxy mergers (Bezanson et al. Reference Bezanson, van Dokkum, Tal, Marchesini, Kriek, Franx and Coppi2009).

Role of intergalactic environment

Galaxies are not isolated, but are grouped hierarchically. In particular, dwarf galaxies are frequently satellites of large galaxies, and even large galaxies are subhalos when they are located in galaxy clusters. The parent system usually has an associated gaseous medium itself, resulting in drag on the smaller system's ISM as it moves with high speed through the surrounding gas. Tidal forces can also dislodge ISM and stars from the smaller galaxy. These effects deprive the galaxies of their ISM, shutting down star formation, and rendering them super-traversable.

Summary

Overall, and perhaps not too surprisingly, quiescent galaxies appear to have the best prospects for interstellar travel, significantly better than the Milky Way. Globular clusters and compact ellipticals likely are super-traversable: their ISM is minuscule and the density of stars is extremely high. Larger compact ellipticals like M32 have the added advantage of relatively high stellar velocity dispersions, mixing stellar populations thoroughly and bringing a stream of new potential destinations to each sun. But even the majority of ETGs are plausibly super-traversable most of the time, with relatively dilute ISMs and dense stellar populations near their cores. The centres of the largest dwarf spheroidals may also be super-traversable, but smaller ones could be sub-traversable because of the low $\rho _{\star }$. SFGs are expected to be relatively poor in comparison, although dwarf irregulars with sufficiently high $\rho _{\star }$ are somewhat favourable compared to the Milky Way. Starbursts and the SFGs of the distant past – including the ancient Milky Way – were subtraversable from an ISM standpoint, filled with hazards to interstellar travel. The dense stellar populations of nuclear starbursts, like those in the cores of other galaxies, may aid traversability by shortening interstellar trip.