Good point, Stefan, and it is one I understand well as an engineer. That is one of the reasons I let the water "age" (i.e. - degas) so that it is able to come to a natural saturation level at the lower pressure and at ambient conditions (i.e. - room temperature & atmospheric pressure). Natural stream water is at equilibrium with the environment and its amount of dissolved gasses. Fresh tap water has not yet come to equilibrium with the amount of dissolved gasses (it is colder than ambient conditions so it will tend to hold more dissolved gasses, and it it was at elevated pressure which also raises the amount of dissolved gasses the water can hold). By letting the water come to "age", the dissolved gasses will come to equilibrium to better match the conditions in a stream. The materials of the fly still want to hold onto some air as the fly is submerged into the water.
I am sure that if I knocked the fly around hard enough I could dislodge the bubbles from the fly, but that wasn't the intent of my experiment. I was initially trying to prove to myself that I could take macro photos of a fly in simulated fishing conditions (i.e. - underwater). I've seen a number of discussions about "What does this fly looks like underwater, not just wet?", which was the source of my inspiration. It's fun to try to answer these questions with actual photos. I need to find time to do some more experimental photos to see flies in the surface film (right at the water/air meniscus).
Gary LaFontaine used to do a lot of underwater observation of fly patterns to try to imitate certain appearances (like his studies of antron yarn on sparkle caddis patterns). After I read about about Gary's observations, it made me wonder what I could try with my waterproof digital camera. Last summer I was crazy enough to jump in a swimming pool to photograph flies with CDC on them. Now with my little set-up, I don't need to jump in a pool to take the pictures.