For competeness, might as well write up the other one.

Howard also looked at recent gullies that have been imaged in stereo by HiRISE. He's an expert of morphology of fluvial systems, debris flows, etc.

Based on what we know from Earth, at least, he concludes the following:

1) The gully systems are far more consistent with short-duration, high-energy events than with long-running trickles
2) They are not consisent with a liquid or even a wet debris flow (lack of any runout channels from the base deposits is a key diagnostic)
3) But, they do show viscosity reduced from dry dust, so he think it's likely there is some volatile involved, be it little bit of water or some CO2.

He seemed perfectly happy with CO2, and noted that someone really needs to do modelling/experimenting with sublimating-CO2 lubricated debris flows so we know what they look like. He also allowed that we may not understand the Martian fines and other conditions there well enough to use our models and so he wouldn't rule out that them being dry. But definitely NOT, for half a dozen reasons, burbling springs running down the sides of craters.