Fig.s 4 and 5 are quite interesting by their own rights 
since they concisely show the main results of the paper. 
However, one thing that really bugs me here is the 
estimated total dust mass with Spitzer and Herschel 
data. Why would it be "less" than the one which was 
estimated 'only' from the Spitzer data? I was expecting 
the ratio shown at the bottom panels of these figures 
would either 1 (very good prediction from Driane & Li 
model for the galaxy with only Spizter data) or greater 
1 (Driane & Li model was approximate for some galaxies 
with only Spitzer data but with Herschel+Spitzer data 
it is better).           

I know you have given a brief explanation but a little 
more elaboration along this line would do better for 
a person like myself who has no expertise in this kind 
of complex issues, although, I do understand that it 
may not the be main focus of this paper.

*** The Herschel+Spitzer dust mass could be lower than the Spitzer dust mass if, for example, the SPIRE data show less dust emission than that predicted by extrapolating from just the Spitzer data alone.  So in my view, it's fine to have the numbers bounce around unity in these plots; it's not a problem to have some above unity and some below unity.