deletions | additions       diff --git a/section_Science_2_Mass_metallicity__.tex b/section_Science_2_Mass_metallicity__.tex  index 9efb6e4..74bdc4c 100644  --- a/section_Science_2_Mass_metallicity__.tex  +++ b/section_Science_2_Mass_metallicity__.tex 
...
\section{Science 2: Mass-metallicity}  The formation and evolution of galaxies is controlled by the accretion and outflow of  baryons (Bouch\'e et al.\ 2010; Dav\'e et al. 2011). \citep{Bouche_2010,Dave_2011}.  The gas-phase metallicities of galaxies are sensitive to this cycle. On one hand, pristine gas inflows dilute metallicity and drive star-formation which creates metals. On the other hand, feedback from star-formation driven outflows removes gas (which may be preferentially enriched) and halts further star and metal production. Consequently, the correlation between galaxies' stellar mass and metallicity, observed in the SDSS by \citet{Tremonti_2004}, is one of the key scaling relations that can constrain cosmological models for galaxy formation \cite{Peeples_2011,Zahid_2011,Henry_2013a,Henry_2013b,Lilly_2013,Wuyts_2014,Sanders_2015,Guo_2016,Ma_2016}. Recent studies have shown that the scatter in the mass-metallicity relation contains even more information about the ways that galaxies grow. For example, it is now established (for low-redshifts) that-- for fixed stellar mass-- galaxies with higher SFRs have lower metallicities, and vise vice  versa (XXX I think this is too informal XXX?;;; Mannucci et al. 2010; 2011, Andrews \& Martini 2013; Yates \& Kauffman 2014; Salim et al 2014). \citep{Mannucci_2010,Mannucci_2011,Andrews_2013,Yates_2014,Salim_2014}.  The simple interpretation of this result is that stochastic variations in gas accretion alter both metallicity and SFRs. However, another source of scatter has been detected, which is not so easily explained in this context: galaxy size. Using the same SDSS data, Ellison et al. (2008) \citet{Ellison_2008}  show a  strong anti-correlation between the  half-light radius and metallicity (at fixed mass) for SDSS galaxies. Nevertheless, this result relied on SDSS imaging (XXX insert some awesome GZ stuff here?). A correlation between galaxy morphology and size could be an indication that outflows must also be considered. Indeed, we expect the surface density of mass or star-formation to impact the efficiency of outflows. For example, Hopkins et al. (2012) \citet{Hopkins_2012}  predict greater mass-loss through outflows (relative to SFR) when gas surface density is low. Because the efficiency of galaxy outflows sets the slope of the mass-metallicity relation (e.g. Henry et al. 2013a,b), \citep[eg,][]{Henry_2013a,Henry_2013b},  we may expect galaxies with lower gas surface density (and therefore lower SFR density) to show a steeper mass-metallicity relation. At the same time, mergers may drive gas into galaxies, implying that morphologically disturbed systems should have lower metallicities (Peeples et al. 2009). \citep{Peeples_2009}.  (Wrap up with quantitative sentences about how this proposal will do the job.)