deletions | additions       diff --git a/figures/qu.time_v_m.star_sat.first.t/caption.tex b/figures/qu.time_v_m.star_sat.first.t/caption.tex  index df28e49..0f3c29b 100644  --- a/figures/qu.time_v_m.star_sat.first.t/caption.tex  +++ b/figures/qu.time_v_m.star_sat.first.t/caption.tex 
...
Top axis shows the satellite subhalo's $\mpeak$ based on abundance matching.  Blue circles show times for satellites of the Milky Way (MW) and M31, obtained by matching the observed quiescent fractions in Figure~\ref{fig:quiescent_fraction} to the virial-infall times of satellites from the ELVIS simulations \citep{Wetzel2015} in bins of $\mstar$.  Left panel shows time since first infall into the current MW/M31 halos, while right panel shows time since first infall into \emph{any} host halo, that is, including the (potential) effects of group preprocessing.  Satellites at $\star $\mstar  < 10 ^ 5 \msun$ (light blue) may have been quenched prior to virial infall via reionization, so their timescale should be interpreted with care. Gray triangle shows lower limit for the Large and Small Magellanic Clouds using their virial-infall time from their measured proper motion \citep{Kallivayalil2013}.  Red squares show the timescale inferred for satellites with $\mstar = 10 ^ {8.5 - 9.5} \msun$ around MW-like hosts in SDSS \citep{Wheeler2014}, and green curve show timescales inferred for more massive satellites in groups of $\mvir = 10 ^ {12 - 13}$ from SDSS \citep{Wetzel2013}.  The satellites in the Local Group quenched much more rapidly after falling into the MW/M31 halo than more massive satellites around other hosts.