A thesis is often a few papers bound together. Mine made it through the committee partly because it was made of 5 accepted articles and two successful HST proposals.
The thing is...often you end up with a chapter here or there or some follow-up idea that you'd really like to do...you know...sometime.
I had a few and as a result there were three more papers that were based in part on ideas I first mentioned in the thesis. I can't say that I had already forseen that it would be really interesting to compare the results from the number of background galaxies to the dust surface densities from spectral energy distribution model fits. We had to wait for Herschel to be launched first.
But a large part of the original science question was to compare the spiral disk opacity to the column density of neutral hydrogen gas. I finally did that. Not the radial profile, the actual surface density derived from a THINGS map. This kind of stuff is only possible if the HI survey is public. In this sense the THINGS survey has been the game-changer. No more, would a grad student have to beg the fits file off 20 different authors...or read in the radial profile from crappy gifs.
Data is online. Go play.
So the last idea using the thesis data points is done: http://arxiv.org/abs/1209.0306 We get a high dust-to-gas ratio. Most likely an upper limit if I am being honest but certainly a point worth making. SED models tend to underestimate the dust surface density in spirals unless it's the Draine et al model. I have made my point. It is the bookend for all the thesis work and we're done.
unless I want to count the number of galaxies seen through M101 and M81...
just for fun...