Perhaps there's a way to visualize the actual operation of the Crookes radiometer, I've a proposal below but I'd welcome a better one.<p>As with many of us, I've been fascinated with the radiometer's seeming 'magic' operation since I was a young kid and first saw one in a shop window. Back then and when I was doing high school science we were led to believe that radiation pressure from incoming light alone was its <i>modus operandi</i> and none of us questioned otherwise. When I learned later of the now accepted Reynolds explanation I was somewhat surprised that we weren't taught it at school as it was already known (seems too our science teachers were unaware of it, and for some reason they never questioned the anomalous direction of rotation, nor did we think to ask).<p>I admit I've always been surprised that radiation pressure was strong enough to cause rotation because it just seemed that overcoming inertia alone would be a problem which would mean the vanes would have taken a much longer time to spin up than they do in practice. Furthermore the frictional losses in the glass/pivot bearing would have had to have been incredibly small. BTW, I'm not taking credit for smart thinking here as others I'd discussed it with thought similarly. When we eventually learned of the Reynolds explanation it made sense.<p>it's just occurred to me—and I haven't thought this through with any finesse—but perhaps there's a way to visualize the radiometer's operation by making the movement of the residual low pressure air visible by some means, but that seems quite a challenge.<p>Coloring the air wouldn't work but there might be another way. Suppose the air was swopped with say radon gas or perhaps even uranium hexafluoride (although I'd suspect it would be much less efficient) and we cover the inside of the radiometer's bulb with numerous radiation detectors but not enough to block sufficient incoming light to stop it functioning. (The detectors would have to be on the inside of the bulb as its glass would block the alpha particles emitted from the radon).<p>We could then track density changes in the gas and build up a moving image of its movement versus incoming radiation levels. Little doubt it'd be a difficult to do and likely considered too dangerous and frivolous an experiment to make it worthwhile. Nevertheless it'd be fascinating to see visual images of the gas's movement, also they'd put any remaining arguments about the radiometer's operation to rest.<p>Anyone got any suggestions for better ways of making the gas molecules visible?