Yet Another Personally Untested Hypothesis
You ever get that thing, with the sink?
You know, the thing?
No? Seriously?
Come on, I bet you do.
Hrm.
Maybe I should explain: You know how sometimes, when you turn on a water faucet just a little bit, the water comes out in a thin, steady stream?
But by the time that steady stream of water reaches the bowl of the sink, it has broken up into discrete droplets?
Have you ever wondered why that happens?
I mean, we've all played surface tension games with water. Everywhere else in life, water likes to join up with more water, not split up from the water around it. In this one situation, though, water does the opposite of what I'd expect.
Here's (what I think is the reason) why:
Acceleration due to gravity.
As the water falls from the faucet to the bowl of the sink, it accelerates. This means that the water near the bottom of the drop, just above the drain hole, is moving faster than the water near the top of the drop, at the mouth of the faucet. Because it's moving faster, it will pull away from the slower water above it. If it weren't for surface tension (and a bunch of other factors too complex to work with here), the separation would happen at the molecular level.
Anyway, I've never bothered to research or test this hypothesis. I'm content to assume that it's an accurate and complete explanation of the phenomenon.