diff --git a/src/tex/ms.tex b/src/tex/ms.tex
index 2668279..0bfd9d8 100644
--- a/src/tex/ms.tex
+++ b/src/tex/ms.tex
@@ -110,7 +110,7 @@ \section{Loop Simulations}
     \caption{Similar to Figure \ref{fig:L40}, showing the results for an 80 Mm loop heated impulsively for 20 s.  \label{fig:L80}}
 \end{figure*}
 
-We finally show two examples of nanoflare trains \citep{reep2013,cargill2014,barnes2016b}, where a series of nanoflare heating events occur in close succession before ceasing and allowing the loop to cool.  Figure \ref{fig:train} shows two cases for loops of 40 and 80 Mm, with 5 heating events of 0.01 erg s$^{-1}$ cm$^{-3}$, spaced 300 s apart.  The loops oscillate around temperatures of around 2.5 and 3.5 MK, respectively, during the heating period, before rapidly cooling.  In both cases, the evaporation is prolonged for more than 30 minutes, causing the abundance factor $f$ to fall successively with each heating event (and thus evaporation event), reaching a minimum of around $f=2$.  As with single heating events, the effect of time-variable abundance on temperature and density becomes most noticeable during the cooling phase.
+We finally show two examples of nanoflare trains \citep{reep2013,cargill2014,barnes2016b}, where a series of nanoflare heating events occur in close succession before ceasing and allowing the loop to cool.  Figure \ref{fig:train} shows two cases for loops of 40 and 80 Mm, with 5 heating events of 0.01 erg s$^{-1}$ cm$^{-3}$, spaced 300 s apart.  The loops oscillate around temperatures of around 2.5 and 3.5 MK, respectively, during the heating period, before rapidly cooling.  In both cases, the evaporation is prolonged for more than 30 minutes, causing the abundance factor $f$ to fall successively with each heating event (and thus evaporation event), reaching a minimum of around $f=2$.  As with single heating events, the effect of time-variable abundance on temperature and density becomes most noticeable during the cooling phase.  Of course, more realistic nanoflare trains would have stochastic waiting times, durations, and heating rates, but the abundances would similarly tend towards photospheric as each event causes evaporation.  Additionally, over long periods of time, it is likely that the Alfv\'en waves would fractionate the low FIP elements, causing a gradual return towards coronal abundances, which we have not considered here.  
 \begin{figure*}
     \script{render_figure3.py}
     \centering
@@ -133,7 +133,7 @@ \section{Discussion}
 
 In this work, we have not attempted to include the changes in abundances due to ponderomotive forces of Alfv\'en waves propagating from the corona to chromosphere, which is thought to be the cause of the FIP effect \citep{laming2015}, where low FIP elements like iron are generally enhanced in coronal loops.  Observations of such waves have been measured (\textit{e.g.} \citealt{murabito2024}), and shown to be consistent with producing the FIP effect.  However, it is not clear what timescales are required to produce the fractionation of elements, as the modeling work to date has been time-independent.  Assuming that the time-scales are short, then this is an important effect that should additionally be incorporated to properly model time-variable abundances.  One additional complexity is that for large enough Poynting fluxes, the resistive dissipation of Alfv\'en waves in the chromosphere would also cause significant amounts of evaporation \citep{reep2016}.  Estimates of Poynting fluxes suggest that waves might carry up to $\approx 10^{10}$ erg s$^{-1}$ cm$^{-2}$ in flares \citep{russell2023}, which is large enough to drive explosive evaporation and produce photospheric abundances in general, but this is an upper limit.   The net effect on elemental abundances is not clear, therefore, and more work is required to determine the interplay.  
 
-Finally, we have implemented this in the 0D \texttt{ebtel++} primarily as a test for its importance in loop evolution.  Higher dimensional models, however, could simulate not only the time variability but also spatial variability.  Spatial variations of abundances are observed regularly in ARs and flares \citep{doschek2018,baker2021,to2021,long2024}.  While the current model does not address this, it would be straightforward to extend this work to higher dimensions.  
+Finally, we have implemented this in the 0D \texttt{ebtel++} primarily as a test for its importance in loop evolution.  Higher dimensional models, however, could simulate not only the time variability but also spatial variability.  Spatial variations of abundances are observed regularly in ARs and flares \citep{doschek2018,baker2021,to2021,long2024}.  While the current model does not address this, it would be straightforward to extend this work to higher dimensions.  Since evaporation is a ubiquitous process in the solar corona, this effect is critical for (magneto)hydrodynamic simulations. 
 
 
 \begin{acknowledgements}