Of course not. I agree with with you totally and thought the point about the bubble missed the mark entirely. But it's hard to be a expert on an item that is almost 100 years old and has been copied endlessly w/o the benefit of the original designer being still around. I suspect it got into my feed because of the pressings in the sheet metal, which could be done with bead rolling which I am always searching up. I don't think they were there for expansion purposes, as they would actually make the metal less flexible, not more (think bomber seats). So I think he may be off base about that too. But I like the two part design, the handles, the spout, the fact that no wrench was required, the rubber seal, the recessed welded seam and a few other points. If you ever use an older Fluke multimeter, they use the exact principle in reverse—the outside protruding yellow rubber is high enough above the plastic controls and display that it protects it when it falls. I actually saw a guy throw one across a shop once and it survived perfectly. There was a video (can't find it now) about the engineering in that rubber case and how much math was crunched to get it the right so as to protect the face of the meter. (Sadly the newer ones seem to abandon this principle I think).
The cynic in me says the poster got the idea in his head about buoyancy from an old timer who passed it on. I can totally picture a Master Corp. or Sgt. telling a group of recruits, "Don't worry boys, if you get tossed overboard and you can grab a Jerrycan, that bubble there (LOOK WHERE I AM POINTING DUMB*SSES) will save your bacon." It's the exact kind of thing I was told in the reserves in many cases, about many things. It was total B.S., but calmed you down and made you feel safer. 80 years later everyone remembers those instructions, and thus history is falsely created. But no one questioned it at the time.
