# Rapid Environmental Quenching of Satellite Dwarf Galaxies in the Local Group

Abstract

In the Local Group, nearly all of the dwarf galaxies ($$M_{\rm star}\lesssim 10^{9}~{}\mbox{M}_{\odot}$$) that are satellites within $$300~{}\mbox{kpc}$$ (the virial radius) of the Milky Way (MW) and Andromeda (M31) have quiescent star formation and little-to-no cold gas. This contrasts strongly with comparatively isolated dwarf galaxies, which are almost all actively star-forming and gas-rich. This near dichotomy implies a rapid transformation after falling into the halos of the MW or M31. We combine the observed quiescent fractions for satellites of the MW and M31 with the infall times of satellites from the ELVIS suite of cosmological simulations to determine the typical timescales over which environmental processes within the MW/M31 halos remove gas and quench star formation in low-mass satellite galaxies. The quenching timescales for satellites with $$M_{\rm star}<10^{8}~{}\mbox{M}_{\odot}$$ are short, $$\lesssim 2~{}\mbox{Gyr}$$, and decrease at lower $$M_{\rm star}$$. These quenching timescales can be $$1-2~{}\mbox{Gyr}$$ longer if environmental preprocessing in lower-mass groups prior to MW/M31 infall is important. We compare with timescales for more massive satellites from previous works, exploring satellite quenching across the observable range of $$M_{\rm star}=10^{3-11}~{}\mbox{M}_{\odot}$$. The environmental quenching timescale increases rapidly with satellite $$M_{\rm star}$$, peaking at $$\approx 9.5~{}\mbox{Gyr}$$ for $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$, and rapidly decreases at higher $$M_{\rm star}$$ to less than $$5~{}\mbox{Gyr}$$ at $$M_{\rm star}>5\times 10^{9}~{}\mbox{M}_{\odot}$$. Thus, satellites with $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$, similar to the Magellanic Clouds, exhibit the longest environmental quenching timescales.

## Introduction

Galaxies in denser environments are more likely to have suppressed (quiescent) star formation and little-to-no cold gas than galaxies of similar stellar mass, $$M_{\rm star}$$, in less dense environments. The observed environmental effects within the Local Group (LG) on the satellite galaxies within the halos of the Milky Way (MW) and Andromeda (M31) are particularly strong (e.g., Einasto et al., 1974; Grcevich et al., 2009; McConnachie, 2012; Phillips et al., 2014; Slater et al., 2014), even compared to the already strong effects on (more massive) satellites within massive groups/clusters. Specifically, the dwarf galaxies around the MW/M31 show a strikingly sharp and nearly complete transition in their properties within $$\approx 300~{}\mbox{kpc}$$ (approximately the virial radius, $$R_{\rm vir}$$, of the MW or M31), from irregular to spheroidal morphologies, from having significant cold atomic gas to having little-to-no measured cold gas, and from actively star-forming to quiescent. This trend has just a few exceptions: 4 gas-rich, star-forming galaxies persist within the halos of the MW (the LMC and SMC) and M31 (LGS 3 and IC 10), and 4 - 5 quiescent, gas-poor galaxies reside well beyond $$R_{\rm vir}$$ of either the MW or M31: Cetus (Lewis et al., 2007), Tucana (Fraternali et al., 2009), KKR 25 (Makarov et al., 2012), KKs 3 (Karachentsev et al., 2015), and possibly Andromeda XVIII, though Cetus and Tucana may have orbited within the MW halo (Teyssier et al., 2012). This efficient satellite quenching is particularly striking because, other than KKR 25 and KKs 3, at $$M_{\rm star}<10^{9}~{}\mbox{M}_{\odot}$$ all known galaxies that are sufficiently isolated ($$>1500~{}\mbox{kpc}$$ from a more massive galaxy) are star-forming (Geha et al., 2012; Phillips et al., 2014). Thus, the MW and M31 halos show the strongest signal of environmental influence over their satellites of any known systems, and the LG is a compelling laboratory for studying environmental processes on galaxies.

Several such processes within a host halo can regulate the gas content, star formation, morphology, and eventual disruption of satellite galaxies, including gravitational tidal forces (e.g., Dekel et al., 2003), galaxy-galaxy tidal interactions (e.g., Farouki et al., 1981), galaxy-galaxy mergers (e.g., Deason et al., 2014), and ram-pressure stripping of extended gas (e.g., Larson et al., 1980; McCarthy et al., 2008) or cold inter-stellar medium (e.g., Gunn et al., 1972; Tonnesen et al., 2009). The key astrophysical challenge is understanding the relative importance of these processes, including which (if any) dominate, and how they vary across both satellite and host masses.

One strong constraint comes from determining the timescale over which environmental quenching occurs, as previous works have explored at higher masses (e.g., Balogh et al., 2000; Wetzel et al., 2013; Hirschmann et al., 2014; Wheeler et al., 2014). For the satellite dwarf galaxies in the LG, some works have shown that the environmental quenching efficiency is higher than than for more massive satellites in massive groups/clusters (Phillips et al., 2014; Slater et al., 2014). In this letter, we combine the observed quiescent fractions for satellites in the LG with the typical infall times of such satellites from cosmological simulations to infer the timescales over which environmental processes remove gas and quench star formation in the current satellite galaxies in the MW/M31 halos. Motivated by the results of Wetzel et al. (2015), we also consider the possible impact of group preprocessing on satellites before they fell into the MW/M31 halos. We also compare with the above works on more massive satellites, allowing us to examine satellite quenching timescales across the observable range of $$M_{\rm star}=10^{3-11}~{}\mbox{M}_{\odot}$$.

## Observations

To examine the observed properties of dwarf galaxies in the LG, we use the compilation from McConnachie (2012), which includes all galaxies known at that time within $$3~{}\mbox{Mpc}$$ of the Sun. We also include the more recent measurements or upper limits of cold atomic gas mass from Spekkens et al. (2014). We define “satellite” galaxies as those within $$300~{}\mbox{kpc}$$ of either the MW or M31, motivated by the observed sharp transition in star formation, gas, and morphological properties within this distance.

Observed dwarf galaxies show a tight correlation between their morphology, star formation, and cold gas content: all spheroidals have little-to-no detectable cold gas (e.g., Spekkens et al., 2014) or ongoing star formation (e.g., Weisz et al., 2014), and almost all irregulars have significant cold gas mass and ongoing star formation. Thus, we define “quiescent” galaxies as those with $$M_{\rm gas}/M_{\rm star}<0.1$$ or with colors and morphologies that resemble spheroidals if they have no cold gas constraints. By this definition, the only star-forming, gas-rich satellites are: LMC ($$M_{\rm star}=1.5\times 10^{9}~{}\mbox{M}_{\odot}$$, $$M_{\rm gas}/M_{\rm star}\approx 0.3$$) and SMC ($$M_{\rm star}=4.6\times 10^{8}~{}\mbox{M}_{\odot}$$, $$M_{\rm gas}/M_{\rm star}\sim 1$$) around the MW, LGS 3 ($$M_{\rm star}=9.6\times 10^{5}~{}\mbox{M}_{\odot}$$, $$M_{\rm gas}/M_{\rm star}\approx 0.4$$) and IC 10 ($$M_{\rm star}=9\times 10^{7}$$, $$M_{\rm gas}/M_{\rm star}\approx 0.6$$) around M31.

## Simulations

To measure infall times of satellites, we use ELVIS (Exploring the Local Volume in Simulations), a suite of cosmological zoom-in $$N$$-body simulations intended to model the LG (Garrison-Kimmel et al., 2014) in $$\Lambda$$CDM cosmology: $$\sigma_{8}=0.801$$, $$\Omega_{\rm matter}=0.266$$, $$\Omega_{\rm\Lambda}=0.734$$, $$n_{s}=0.963$$ and $$h=0.71$$. Within the zoom-in regions, the particle mass is $$1.9\times 10^{5}~{}\mbox{M}_{\odot}$$ and the Plummer-equivalent force softening is $$140~{}\mbox{pc}$$ physical.

ELVIS contains 48 dark-matter halos of masses similar to the MW or M31 ($$M_{\rm vir}=1.0-2.8\times 10^{12}~{}\mbox{M}_{\odot}$$), with a median $$R_{\rm vir}\approx 300~{}\mbox{kpc}$$, the distance where observed dwarf galaxies show a strong transition. Half of the halos are part of a pair that resemble the masses, distance, and relative velocity of the MW-M31 pair, while the other half are single isolated halos. Given the lack of systematic differences in satellite infall times for the paired versus isolated halos (Wetzel et al., 2015), we use all 48 to improve the statistics.

ELVIS identifies dark-matter (sub)halos using the six-dimensional halo finder rockstar (Behroozi et al., 2013). For each halo, we assign a virial mass, $$M_{\rm vir}$$, and radius, $$R_{\rm vir}$$, according to Bryan et al. (1998). A “subhalo” is a halo whose center is inside $$R_{\rm vir}$$ of a more massive host halo, and a subhalo experiences “first infall” and becomes a “satellite” when it first passes within $$R_{\rm vir}$$. For each subhalo, we compute the peak mass, $$M_{\rm peak}$$, that it ever reached, and we assign $$M_{\rm star}$$ to subhalos based on their $$M_{\rm peak}$$ using the relation from abundance matching in Garrison-Kimmel et al. (2014), which reproduces the observed mass function at $$M_{\rm star}<10^{9}~{}\mbox{M}_{\odot}$$ in the LG if one accounts for observational incompleteness (Tollerud et al., 2008; Hargis et al., 2014).

For more details on ELVIS and/or satellite infall times, see Garrison-Kimmel et al. (2014) and/or Wetzel et al. (2015).

## Observed Quiescent Fractions for Satellites

Figure \ref{fig:quiescent_fraction} shows, for all satellite galaxies at $$M_{\rm star}\lesssim 10^{9}~{}\mbox{M}_{\odot}$$ within $$300~{}\mbox{kpc}$$ of the MW or M31, the fraction that are quiescent in 1-dex bins of $$M_{\rm star}$$ (see also Phillips et al., 2014; Slater et al., 2014). We do not correct for any observational completeness versus $$M_{\rm star}$$, because we measure the relative fraction in each bin, which is likely unbiased at distances $$\lesssim 300~{}\mbox{kpc}$$. We show fractions for all satellites (blue circles) and separately for those in the MW (violet squares) and M31 (green triangles) halos. Error bars show 68% uncertainty for the binomial counts using a beta distribution (Cameron, 2011). Of the 56 satellites, only 4 (7%) are star-forming/gas-rich: LMC and SMC of the MW, LGS 3 and IC 10 of M31. Moreover, at $$M_{\rm star}<8\times 10^{7}~{}\mbox{M}_{\odot}$$, only 1 (LGS 3) of the 51 satellites is star-forming, and at $$M_{\rm star}<9\times 10^{5}~{}\mbox{M}_{\odot}$$ all 40 satellites are quiescent.

These near-unity quiescent fractions for satellites of the MW/M31 contrast strongly with the nearly zero quiescent fraction observed for isolated (non-satellite) galaxies at $$M_{\rm star}<10^{9}~{}\mbox{M}_{\odot}$$ (Geha et al., 2012; Phillips et al., 2014). The only clear exceptions are the quiescent galaxies KKR 25 ($$M_{\rm star}=1.4\times 10^{6}~{}\mbox{M}_{\odot}$$) and KKs 3 ($$M_{\rm star}=2.3\times 10^{7}~{}\mbox{M}_{\odot}$$) at $$\approx 2~{}\mbox{Mpc}$$ from the MW/M31, although the limited completeness at $$M_{\rm star}\lesssim 10^{6}~{}\mbox{M}_{\odot}$$ beyond $$300~{}\mbox{kpc}$$ (Tollerud et al., 2008; Hargis et al., 2014) leaves open the possibility for more such galaxies.

\label{fig:quiescent_fraction}For all satellites galaxies with $$M_{\rm star}\lesssim 10^{9}~{}\mbox{M}_{\odot}$$ within $$300~{}\mbox{kpc}$$ of the Milky Way (MW) or Andromeda (M31), the fraction that are quiescent ($$M_{\rm gas}/M_{\rm star}<0.1$$) versus stellar mass, $$M_{\rm star}$$. Blue circles show all satellites, violet squares (green triangles) show those of just the MW (M31). Of these 56 satellites, only 4 are star-forming/gas-rich: LMC ($$M_{\rm star}=1.5\times 10^{9}~{}\mbox{M}_{\odot}$$) and SMC ($$M_{\rm star}=4.6\times 10^{8}~{}\mbox{M}_{\odot}$$) around the MW, LGS 3 ($$M_{\rm star}=9.6\times 10^{5}~{}\mbox{M}_{\odot}$$) and IC 10 ($$M_{\rm star}=9\times 10^{7}~{}\mbox{M}_{\odot}$$) around M31. At $$M_{\rm star}<8\times 10^{7}~{}\mbox{M}_{\odot}$$, 50 of 51 satellites are quiescent, and at $$M_{\rm star}<9\times 10^{5}~{}\mbox{M}_{\odot}$$ all are quiescent. Error bars show 68% uncertainty based on the observed counts.

## Inferred Environmental Quenching Timescales for Satellites

We now translate the quiescent fractions in Figure \ref{fig:quiescent_fraction} into the typical timescales over which environmental processes quench satellites after they fall into a host halo, following the methodology of Wetzel et al. (2013).

First, motivated by the dearth of isolated galaxies with $$M_{\rm star}<10^{9}~{}\mbox{M}_{\odot}$$ that are quiescent at $$z\approx 0$$ (see Introduction), our model assumes that all satellites with $$M_{\rm star}(z=0)<10^{9}~{}\mbox{M}_{\odot}$$ were actively star-forming prior to first infall. However, because most galaxies with $$M_{\rm star}(z=0)<10^{4}~{}\mbox{M}_{\odot}$$ may have been quenched at high redshift by cosmic reionization (e.g., Weisz et al., 2014; Brown et al., 2014), we do not model those masses. At $$M_{\rm star}(z=0)=10^{4-5}~{}\mbox{M}_{\odot}$$, satellites’ star-formation histories show a mix of complete quenching by $$z\gtrsim 3$$ (e.g., Bootes I, Leo IV) and signs of star formation at $$z\lesssim 1$$ (e.g., And XI, And XII, And XVI) (Weisz et al., 2014; Weisz et al., 2014a; Brown et al., 2014), so quenching at these masses may come from a mix of reionization and the host-halo environment. That said, the 100% quiescent fraction for satellites at this $$M_{\rm star}$$ means that if both processes are responsible, both are highly efficient. Furthermore, if the satellites that were quenched by reionization have a similar infall-time distribution to those that were quenched by the host-halo environment, our modeling approach remains valid. Thus, we include this $$M_{\rm star}$$ in our results, but we label it distinctly to emphasize caution in interpretation.

Within each 1-dex bin of $$M_{\rm star}$$, we use ELVIS to compute the distribution of infall times for satellites at $$z=0$$. Assuming that environmental quenching correlates with time since infall, we designate those that fell in earliest as having quenched, and we adjust the time-since-infall threshold for quenching until we match the observed quiescent fraction in each bin.

Several works have shown that this model successfully describes the dependence of satellite quiescent fractions on host-centric distance (e.g., Wetzel et al., 2013; Wetzel et al., 2014; Wheeler et al., 2014) because infall time correlates with host-centric distance (e.g., Wetzel et al., 2015). However, this correlation means that we must account for observed satellites’ distances in computing their infall times. Thus, in ELVIS we only select satellites out to the maximum host-centric distance that they are observed in each $$M_{\rm star}$$ bin. In fact, this matters most at the highest $$M_{\rm star}$$ bin, where all observed satellites (M32, NGC 205, LMC/SMC) reside $$<61~{}\mbox{kpc}$$ from the MW or M31.

Figure \ref{fig:quench_times} shows the inferred environmental quenching timescales (the time duration from first infall to being fully quiescent/gas-poor) versus $$M_{\rm star}$$ (top axis shows corresponding subhalo $$M_{\rm peak}$$). Blue circles show the satellites in the MW and M31, and we shade the lowest $$M_{\rm star}$$ bin to highlight caution in interpretation because of reionization. We derive error bars from the 68% uncertainty in the observed quiescent fractions in Figure \ref{fig:quiescent_fraction}.

As explored in Wetzel et al. (2015), many satellites first fell into a another host halo (group), typically of $$M_{\rm vir}=10^{10-12}~{}\mbox{M}_{\odot}$$, before falling into the MW/M31 halos. Because the importance of this environmental preprocessing in lower-mass groups remains unclear, we present quenching timescales both neglecting (left panel) and including (right panel) such group preprocessing. The latter results in longer quenching timescales, though it primarily shifts the upper 16% of the distribution.

Both panels show shorter median quenching timescales for less massive satellites: $$\sim 5~{}\mbox{Gyr}$$ at $$M_{\rm star}=10^{8-9}~{}\mbox{M}_{\odot}$$, $$2-3~{}\mbox{Gyr}$$ at $$M_{\rm star}=10^{7-8}~{}\mbox{M}_{\odot}$$, and $$<1.5~{}\mbox{Gyr}$$ at $$M_{\rm star}<10^{7}~{}\mbox{M}_{\odot}$$, depending on the inclusion of group preprocessing. Moreover, the median timescale for two of the lowest $$M_{\rm star}$$ bins is $$0~{}\mbox{Gyr}$$ because 100% of those satellites are quiescent, which implies extremely rapid quenching after infall.

Figure \ref{fig:quench_times} also shows infall/quenching timescales directly measured for satellites of the MW. The 3-D orbital velocity measured for the LMC/SMC strongly suggests that they are on their first infall and passed inside $$R_{\rm vir}$$ of the MW $$\approx 2~{}\mbox{Gyr}$$ ago (Kallivayalil et al., 2013). Given that both remain star-forming, this places a lower limit to their quenching timescale (gray triangle), consistent with our statistical timescales. Similarly, measurements of the 3-D orbital velocity and star-formation history for Leo I ($$M_{\rm star}=5.5\times 10^{6}~{}\mbox{M}_{\odot}$$) indicate that it fell into the MW halo $$\approx 2.3~{}\mbox{Gyr}$$ ago and quenched $$\approx 1~{}\mbox{Gyr}$$ ago (near its $$\approx 90~{}\mbox{kpc}$$ pericentric passage), implying a quenching timescale of $$\approx 1.3~{}\mbox{Gyr}$$ (Sohn et al., 2013, gray pentagon), again consistent with our results.

The mass trend in Figure \ref{fig:quench_times} is broadly consistent the star-formation-history-based results of Weisz et al. (2015) that more massive dwarf galaxies in the LG quenched more recently. Also, the overall timescale is broadly consistent with the related analysis of Slater et al. (2014), who inferred a typical quenching time since first pericenter of $$1-2~{}\mbox{Gyr}$$), which implies a quenching time since infall of $$\sim 3~{}\mbox{Gyr}$$, though they did not examine mass dependence.

We also compare these timescales with previous studies of more massive satellites of other hosts. The red squares in Figure \ref{fig:quench_times} show the timescales from Wheeler et al. (2014), who used nearly identical methodology, combining the the galaxy catalog from Geha et al. (2012) with satellite infall times (including group preprocessing) from simulation. They examined satellites with $$M_{\rm star}\approx 10^{8.5}$$ and $$10^{9.5}~{}\mbox{M}_{\odot}$$ around hosts with $$M_{\rm star}>2.5\times 10^{10}~{}\mbox{M}_{\odot}$$, which they found likely spans $$M_{\rm vir}\approx 10^{12.5-14}~{}\mbox{M}_{\odot}$$, much higher than the MW/M31. Similarly, the green curves in Figure \ref{fig:quench_times} show the quenching timescales for more massive satellites in groups with $$M_{\rm vir}=10^{12-13}~{}\mbox{M}_{\odot}$$ from Wetzel et al. (2013), who also used identical methodology, combining a galaxy group catalog from SDSS (Tinker et al., 2011; Wetzel et al., 2012) with satellite infall times (including group preprocessing) measured in their cosmological simulation.

Altogether, Figure \ref{fig:quench_times} indicates a complex dependence of the environmental quenching timescale on satellite $$M_{\rm star}$$. The typical timescale for the low-mass satellites in the MW/M31 halos increases with $$M_{\rm star}$$, from $$\lesssim 1~{}\mbox{Gyr}$$ at $$M_{\rm star}<10^{7}~{}\mbox{M}_{\odot}$$ to $$\sim 5~{}\mbox{Gyr}$$ at $$M_{\rm star}\approx 10^{8.5}~{}\mbox{M}_{\odot}$$. Wheeler et al. (2014) indicate that this mass dependence continues, though with a rapid increase ($$\sim 2\times$$) to $$\approx 9.5~{}\mbox{Gyr}$$, and no change from $$M_{\rm star}\approx 10^{8.5}$$ to $$10^{9.5}~{}\mbox{M}_{\odot}$$. This rapid increase implies some tension with our results based on the two quiescent satellites of M31, NGC 205 and M32 ($$M_{\rm star}\approx 10^{8.5}~{}\mbox{M}_{\odot}$$), unless both experienced unusually early infall $$>9.5~{}\mbox{Gyr}$$ ago or M31 quenched its satellites much more rapidly than the (more massive) hosts in Wheeler et al. (2014). At higher $$M_{\rm star}$$, Wetzel et al. (2013) indicate that the quenching timescale rapidly decreases by $$5\times 10^{9}~{}\mbox{M}_{\odot}$$, and it continues to decline with increasing $$M_{\rm star}$$.

Overall, the typical environmental quenching timescales are shortest for the lowest-mass satellites and are longest for satellites with $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$, similar to the Magellanic Clouds.

left panel

\label{fig:quench_times}right panel

Environmental quenching timescales for satellite galaxies across the observable range of stellar mass, $$M_{\rm star}$$ (top axis shows subhalo $$M_{\rm peak}$$ from abundance matching). Blue circles show satellites of the MW and M31, obtained by matching the observed quiescent fractions in Figure \ref{fig:quiescent_fraction} to rank-ordered infall times of satellites from the ELVIS simulations (Wetzel et al., 2015) in 1-dex bins of $$M_{\rm star}$$. At $$M_{\rm star}=10^{4-5}~{}\mbox{M}_{\odot}$$ (light blue), reionization may have quenched some satellites prior to infall. Error bars come from the 68% uncertainty in observed quiescent fractions in Figure \ref{fig:quiescent_fraction}. Left panel uses time since first infall into the current MW/M31-like halo, while right panel uses time since first infall into any host halo, thereby including possible effects of group preprocessing. Gray triangle shows lower limit for the LMC/SMC system from its measured orbit (Kallivayalil et al., 2013), and gray pentagon shows the quenching timescale for Leo I from its measured orbit and star-formation history (Sohn et al., 2013). Red squares show times inferred for satellites with $$M_{\rm star}=10^{8.5}$$, $$10^{9.5}~{}\mbox{M}_{\odot}$$ around hosts with $$M_{\rm star}>2.5\times 10^{10}~{}\mbox{M}_{\odot}$$ in SDSS (Wheeler et al., 2014), and green curve shows the same for more massive satellites in groups of $$M_{\rm vir}=10^{12-13}~{}\mbox{M}_{\odot}$$ in SDSS (Wetzel et al., 2013). The satellites in the MW/M31 halos quenched more rapidly after infall than more massive satellites (around other hosts). Overall, the quenching timescale increases with $$M_{\rm star}$$, is longest at $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$ (near Magellanic-Cloud masses), then decreases with further increasing $$M_{\rm star}$$.

## Discussion

We conclude by briefly discussing the dependence of satellite quenching timescales on $$M_{\rm star}$$ from Figure \ref{fig:quench_times} in the context of the underlying physics.

At $$M_{\rm star}\gtrsim 10^{9}~{}\mbox{M}_{\odot}$$, the long timescales suggests that satellite quenching is caused by gas depletion in the absence of cosmic accretion, via the stripping of extended gas around the satellite, after infall (“strangulation”). This scenario also can explain the decline of the quenching timescale with increasing $$M_{\rm star}$$, because higher-$$M_{\rm star}$$ (non-satellite) galaxies generally have lower $$M_{\rm gas}/M_{\rm star}$$ (in either cold atomic or molecular gas, e.g., Schiminovich et al., 2010; Huang et al., 2012; Boselli et al., 2014, Bradford et al., submitted) and thus shorter gas depletion timescales in the absence of accretion. Conversely, (non-satellite) galaxies at $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$ have $$M_{\rm gas}\approx M_{\rm star}$$, with gas depletion timescales comparable to a Hubble time. Thus, satellite quenching timescales at $$M_{\rm star}\gtrsim 10^{9}~{}\mbox{M}_{\odot}$$ do not necessarily require strong environmental processes beyond truncated gas accretion (see also discussions in Wetzel et al., 2013; Wheeler et al., 2014; McGee et al., 2014).

However, strangulation cannot explain the rollover in satellite quenching times at $$M_{\rm star}\lesssim 10^{9}~{}\mbox{M}_{\odot}$$, because the gas-rich dwarf galaxies of the LG also have $$M_{\rm gas}\gtrsim M_{\rm star}$$ (Grcevich et al., 2009) and thus contain enough cold gas to fuel star formation for a Hubble time, even absent accretion. Thus, the rapid decline at lower $$M_{\rm star}$$ requires an additional process(es) to remove gas from satellites after infall. This likely arises from the increased efficiency of ram-pressure stripping in removing cold gas from such low-mas galaxies with shallower potential wells. Moreover, for dwarf galaxies, the same internal stellar feedback that regulates their low star-formation efficiency likely heats/drives significant cold gas to large radii on short timescales (e.g., Muratov et al., 2015), which would assist any such environmental stripping, making it even more efficient. Thus, the rapid environmental quenching timescales for dwarf galaxies may arise from the non-linear interplay of both internal feedback and external stripping (e.g., Nichols et al., 2011; Bahé et al., 2015).

Overall, satellites with $$M_{\rm star}\sim 10^{9}~{}\mbox{M}_{\odot}$$ (similar to the Magellanic Clouds) represent the transition between gas consumption and gas stripping, and no quenching mechanism (either internal or external) appears to operate efficiently near this mass (see also Weisz et al., 2015).

Finally, we note that the above scenario may explain the curious, though qualitative, similarity of Figure \ref{fig:quench_times} with the mass dependence of the underlying galaxy-halo $$M_{\rm star}/M_{\rm vir}$$ relation, which also is low at both high and low $$M_{\rm star}$$ and peaks at $$M_{\rm star}\sim 10^{10}~{}\mbox{M}_{\odot}$$ (e.g., Behroozi et al., 2013). In particular, at high $$M_{\rm star}$$, the same physical process(es) that lowers $$M_{\rm star}/M_{\rm vir}$$ also lowers a galaxy’s cold gas fraction, which in turn causes more massive satellites to quench more rapidly, absent accretion. At low $$M_{\rm star}$$, the same shallower potential wells that cause stellar feedback to lower $$M_{\rm star}/M_{\rm vir}$$ also allows external stripping to occur more easily and thus quenching to occur more rapidly.

While preparing this letter, we became aware of Fillingham et al. 2015 (submitted), who also used ELVIS to constrain the quenching timescales of satellites of the MW/M31.

We thank the Aspen Center for Physics and the Kavli Institute for Theoretical Physics, both supported by the National Science Foundation, for stimulating environments during the preparation of this letter. ARW gratefully acknowledges support from the Moore Center for Theoretical Cosmology and Physics at Caltech. Support for DRW is provided by NASA through Hubble Fellowship grants HST-HF-51331.01.

### References

1. Y. M. Bahé, I. G. McCarthy. Star formation quenching in simulated group and cluster galaxies: when, how, and why?. 447, 973-996 (2015). Link

2. M. L. Balogh, J. F. Navarro, S. L. Morris. The Origin of Star Formation Gradients in Rich Galaxy Clusters. 540, 113-121 (2000). Link

3. P. S. Behroozi, R. H. Wechsler, H.-Y. Wu. The ROCKSTAR Phase-space Temporal Halo Finder and the Velocity Offsets of Cluster Cores. 762, 109 (2013). Link

4. P. S. Behroozi, R. H. Wechsler, C. Conroy. The Average Star Formation Histories of Galaxies in Dark Matter Halos from z = 0-8. 770, 57 (2013). Link

5. A. Boselli, L. Cortese, M. Boquien, S. Boissier, B. Catinella, C. Lagos, A. Saintonge. Cold gas properties of the Herschel Reference Survey. II. Molecular and total gas scaling relations. 564, A66 (2014). Link

6. T. M. Brown, J. Tumlinson, M. Geha, J. D. Simon, L. C. Vargas, D. A. VandenBerg, E. N. Kirby, J. S. Kalirai, R. J. Avila, M. Gennaro, H. C. Ferguson, R. R. Muñoz, P. Guhathakurta, A. Renzini. The Quenching of the Ultra-faint Dwarf Galaxies in the Reionization Era. 796, 91 (2014). Link

7. G. L. Bryan, M. L. Norman. Statistical Properties of X-Ray Clusters: Analytic and Numerical Comparisons. 495, 80-99 (1998). Link

8. E. Cameron. On the Estimation of Confidence Intervals for Binomial Population Proportions in Astronomy: The Simplicity and Superiority of the Bayesian Approach. 28, 128-139 (2011). Link

9. A. Deason, A. Wetzel, S. Garrison-Kimmel. Satellite Dwarf Galaxies in a Hierarchical Universe: The Prevalence of Dwarf-Dwarf Major Mergers. 794, 115 (2014). Link

10. A. Dekel, J. Devor, G. Hetzroni. Galactic halo cusp-core: tidal compression in mergers. 341, 326-342 (2003).

11. J. Einasto, E. Saar, A. Kaasik, A. D. Chernin. Missing mass around galaxies - Morphological evidence. 252, 111-113 (1974). Link

12. R. Farouki, S. L. Shapiro. Computer simulations of environmental influences on galaxy evolution in dense clusters. II - Rapid tidal encounters. 243, 32-41 (1981). Link

13. F. Fraternali, E. Tolstoy, M. J. Irwin, A. A. Cole. Life at the periphery of the Local Group: the kinematics of the Tucana dwarf galaxy. 499, 121-128 (2009). Link

14. S. Garrison-Kimmel, M. Boylan-Kolchin, J. S. Bullock, K. Lee. ELVIS: Exploring the Local Volume in Simulations. 438, 2578-2596 (2014). Link

15. M. Geha, M. R. Blanton, R. Yan, J. L. Tinker. A Stellar Mass Threshold for Quenching of Field Galaxies. 757, 85 (2012). Link

16. J. Grcevich, M. E. Putman. H I in Local Group Dwarf Galaxies and Stripping by the Galactic Halo. 696, 385-395 (2009). Link

17. J. E. Gunn, III Gott. On the Infall of Matter Into Clusters of Galaxies and Some Effects on Their Evolution. 176, 1-+ (1972). Link

18. J. R. Hargis, B. Willman, A. H. G. Peter. Too Many, Too Few, or Just Right? The Predicted Number and Distribution of Milky Way Dwarf Galaxies. 795, L13 (2014). Link

19. M. Hirschmann, G. De Lucia, D. Wilman, S. Weinmann, A. Iovino, O. Cucciati, S. Zibetti, Á. Villalobos. The influence of the environmental history on quenching star formation in a $$\Lambda$$ cold dark matter universe. 444, 2938-2959 (2014). Link

20. S. Huang, M. P. Haynes, R. Giovanelli, J. Brinchmann, S. Stierwalt, S. G. Neff. Gas, Stars, and Star Formation in ALFALFA Dwarf Galaxies. 143, 133 (2012). Link

21. N. Kallivayalil, R. P. van der Marel, G. Besla, J. Anderson, C. Alcock. Third-epoch Magellanic Cloud Proper Motions. I. Hubble Space Telescope/WFC3 Data and Orbit Implications. 764, 161 (2013). Link

22. I. D. Karachentsev, L. N. Makarova, D. I. Makarov, R. B. Tully, L. Rizzi. A new isolated dSph galaxy near the Local Group. 447, L85-L89 (2015). Link

23. R. B. Larson, B. M. Tinsley, C. N. Caldwell. The evolution of disk galaxies and the origin of S0 galaxies. 237, 692-707 (1980). Link

24. G. F. Lewis, R. A. Ibata, S. C. Chapman, A. McConnachie, M. J. Irwin, E. Tolstoy, N. R. Tanvir. Inside the whale: the structure and dynamics of the isolated Cetus dwarf spheroidal. 375, 1364-1370 (2007). Link

25. D. Makarov, L. Makarova, M. Sharina, R. Uklein, A. Tikhonov, P. Guhathakurta, E. Kirby, N. Terekhova. A unique isolated dwarf spheroidal galaxy at D = 1.9 Mpc. 425, 709-719 (2012). Link

26. I. G. McCarthy, C. S. Frenk, A. S. Font, C. G. Lacey, R. G. Bower, N. L. Mitchell, M. L. Balogh, T. Theuns. Ram pressure stripping the hot gaseous haloes of galaxies in groups and clusters. 383, 593-605 (2008). Link

27. A. W. McConnachie. The Observed Properties of Dwarf Galaxies in and around the Local Group. 144, 4 (2012). Link

28. S. L. McGee, R. G. Bower, M. L. Balogh. Overconsumption, outflows and the quenching of satellite galaxies. 442, L105-L109 (2014). Link

29. A. L. Muratov, D. Keres, C.-A. Faucher-Giguere, P. F. Hopkins, E. Quataert, N. Murray. Gusty, gaseous flows of FIRE: galactic winds in cosmological simulations with explicit stellar feedback. ArXiv e-prints (2015).

30. M. Nichols, J. Bland-Hawthorn. Gas Depletion in Local Group Dwarfs on ~250 kpc Scales: Ram Pressure Stripping Assisted by Internal Heating at Early Times. 732, 17 (2011). Link

31. J. I. Phillips, C. Wheeler, M. C. Cooper, M. Boylan-Kolchin, J. S. Bullock, E. Tollerud. The Mass Dependance of Satellite Quenching in Milky Way-like Halos. ArXiv e-prints (2014).

32. D. Schiminovich, B. Catinella, G. Kauffmann, others. The GALEX Arecibo SDSS Survey - II. The star formation efficiency of massive galaxies. 408, 919-934 (2010). Link

33. C. T. Slater, E. F. Bell. The Mass Dependence of Dwarf Satellite Galaxy Quenching. 792, 141 (2014). Link

34. S. T. Sohn, G. Besla, R. P. van der Marel, M. Boylan-Kolchin, S. R. Majewski, J. S. Bullock. The Space Motion of Leo I: Hubble Space Telescope Proper Motion and Implied Orbit. 768, 139 (2013). Link

35. K. Spekkens, N. Urbancic, B. S. Mason, B. Willman, J. E. Aguirre. The Dearth of Neutral Hydrogen in Galactic Dwarf Spheroidal Galaxies. 795, L5 (2014). Link

36. M. Teyssier, K. V. Johnston, M. Kuhlen. Identifying Local Group field galaxies that have interacted with the Milky Way. 426, 1808-1818 (2012). Link

37. J. Tinker, A. Wetzel, C. Conroy. Are Halo and Galaxy Formation Histories Correlated?. (2011).

38. E. J. Tollerud, J. S. Bullock, L. E. Strigari, B. Willman. Hundreds of Milky Way Satellites? Luminosity Bias in the Satellite Luminosity Function. 688, 277-289 (2008). Link

39. S. Tonnesen, G. L. Bryan. Gas Stripping in Simulated Galaxies with a Multiphase Interstellar Medium. 694, 789-804 (2009). Link

40. D. R. Weisz, A. E. Dolphin, E. D. Skillman, J. Holtzman, K. M. Gilbert, J. J. Dalcanton, B. F. Williams. The Star Formation Histories of Local Group Dwarf Galaxies. I. Hubble Space Telescope/Wide Field Planetary Camera 2 Observations. 789, 147 (2014). Link

41. D. R. Weisz, E. D. Skillman, S. L. Hidalgo, M. Monelli, A. E. Dolphin, A. McConnachie, E. J. Bernard, C. Gallart, A. Aparicio, M. Boylan-Kolchin, S. Cassisi, A. A. Cole, H. C. Ferguson, M. Irwin, N. F. Martin, L. Mayer, K. B. W. McQuinn, J. F. Navarro, P. B. Stetson. Comparing M31 and Milky Way Satellites: The Extended Star Formation Histories of Andromeda II and Andromeda XVI. 789, 24 (2014). Link

42. D. R. Weisz, A. E. Dolphin, E. D. Skillman, J. Holtzman, K. M. Gilbert, J. J. Dalcanton, B. F. Williams. The Star Formation Histories of Local Group Dwarf Galaxies III. Characterizing Quenching in Low-Mass Galaxies. ArXiv e-prints (2015).

43. A. R. Wetzel, J. L. Tinker, C. Conroy, F. C. van den Bosch. Galaxy evolution in groups and clusters: satellite star formation histories and quenching time-scales in a hierarchical Universe. 432, 336-358 (2013). Link

44. A. R. Wetzel, A. J. Deason, S. Garrison-Kimmel. Satellite Dwarf Galaxies in a Hierarchical Universe: Infall Histories, Group Preprocessing, and Reionization. (2015).

45. A. R. Wetzel, J. L. Tinker, C. Conroy, F. C. v. d. Bosch. Galaxy evolution near groups and clusters: ejected satellites and the spatial extent of environmental quenching. 439, 2687-2700 (2014). Link

46. A. R. Wetzel, J. L. Tinker, C. Conroy. Galaxy evolution in groups and clusters: star formation rates, red sequence fractions and the persistent bimodality. 424, 232-243 (2012). Link

47. C. Wheeler, J. I. Phillips, M. C. Cooper, M. Boylan-Kolchin, J. S. Bullock. The surprising inefficiency of dwarf satellite quenching. 442, 1396-1404 (2014). Link