Backwards water drops on third rail

June 4, 2007 9:15 AM

In the rain at my outdoor subway station in Brooklyn this morning, I observed once again a bizarre phenomenon: water drops appear to drip upwards from the third rail to the wooden cover above it. My wife saw this too. What's going on?
posted by rustcellar to science & nature (21 comments total)

Could it actually be that many drops are falling downward, creating an illusion like when car wheels appear to spin backwards at certain speeds?

This is easy to demonstrate with a variable-speed fan and a strobe light, and makes sense to me in that context (the strobe lights up the fan when it has gone just less than one full revolution since the previous strobe,) but I've never understood how it can happen in (presumably constant) sunlight unless the speed of the rotation (or in this case the distance between the falling drops) is somehow related to the "refresh rate" of the human visual system.

Anyway, maybe what you're seeing is unrelated to the charge on the third rail and is actually just an optical illusion?
posted by contraption at 9:35 AM on June 4, 2007

There might be a large electric potential difference (voltage) between the wet third rail and the cover. This polarizes both the cover and the water. I don't know which surface is at the highest voltage, but if, for example, the rail is, then the electrons in the cover become attracted to the rail while the positive charges become repelled. Even if no charges actually move, the atoms stretch so that there is a layer of negative charge on the wooden cover surface closest to the rail. The water on the rail is made of molecules with a permanent dipole moment -- one side of each molecule is positive and the other side is negative. These molecules will rotate so that the positive side is closest to the cover, since this side is attracted to the surface charge on the cover and the negative side is repelled. Now there is a net force on the molecule, since the attraction is a bit greater than the repulsion (since the distance is smaller, and the electric force falls off with distance). (This is the same thing that happens with this demonstration.) With enough voltage, the upward force on the water is greater than the combination of weight and surface tension keeping the water on the rail.
posted by Killick at 9:35 AM on June 4, 2007

could it actually be that many drops are falling downward, creating an illusion like when car wheels appear to spin backward at certain speeds?

bingo. don't overthink this.
posted by bruce at 9:43 AM on June 4, 2007

I thought car wheels only seem to spin backwards on TV or film, due to the framerate being just a bit off the period of the spin. I'll second the illusion theory, though.
posted by monkeymadness at 9:54 AM on June 4, 2007

Is it possible that the drops you see 'moving upward' are actually illuminated in some way by a fluorescent light, and you're actually seeing a pseudo-strobe effect?
posted by pjern at 9:59 AM on June 4, 2007

But that subway station is likely lit with fluorescent lights, which flicker and can cause that illusion.
posted by wzcx at 10:01 AM on June 4, 2007

Maybe the strobing effect of the old crappy flourescent lights is producing the 'framerate' effect?
posted by dirtdirt at 10:01 AM on June 4, 2007

could it actually be that many drops are falling downward, creating an illusion like when car wheels appear to spin backward at certain speeds?

I am pretty sure that only happens on camera and has to do with the frame rate of the camera. There's the off chance that florescent lights in the station could cause the same phenomenon but it's damn unlikely because the drip rate and the falling rate of the drops would have to be insanely fast.
posted by 517 at 10:02 AM on June 4, 2007

Are these drips moving a few inches or a few feet?
posted by 517 at 10:04 AM on June 4, 2007

If you're seeing the reflection of drops in a puddle or some other reflective surface, that could create this illusion.
posted by aubilenon at 10:10 AM on June 4, 2007

Not only fluorescent lights but light coming through a fan or other rapidly repeating interruption can cause a stroboscopic effect like the one described. So can CRT tubes. It is also called temporal aliasing .
posted by TedW at 10:11 AM on June 4, 2007

The drops are moving (I'd say) about 3 inches. I doubt that it's just an illusion from many drops falling down because there simply aren't that many. There are old but not particularly flickery fluorescent bulbs, so I suppose that's a possibility, although it happens so regularly (although at unpredictable points on the track) that I'm really drawn to Killick's explanation.

If the rain keeps up I might try to capture some video, but probably the scale will keep that from being useful.

Simply in visual terms, the drops might (maybe possibly but not really) resemble the "steps" of a Jacob's Ladder of my science-museum childhood. But since there is a clear motion from the bottom to the top, that doesn't seem to be what's happening (plus they don't seem to be luminescent).
posted by rustcellar at 10:16 AM on June 4, 2007

I am pretty sure that only happens on camera and has to do with the frame rate of the camera.

Nope. I've seen it outdoors on many a sunny day, through no equipment but my own eyes. And no, it's not always those "spinner" hubcaps.

Wait a... Oh fuck, am I a replicant?!?!
posted by contraption at 10:23 AM on June 4, 2007

I've seen fluorescent lights causing hubcaps to appear to be moving backward for cars moving at freeway speeds. But it doesn't seem plausible to me here, because I think the dripping would have to be very regular and fairly quick. Also, -- of all the situations where water is dripping, the one where this effect was observed involved a third rail -- there is a statistical argument that can be made here in favor of an electrical explanation.

rustcellar, if it is an electrical effect, then the points where it happens might be somewhat predictable, since the field will be stronger where there are sharp points like splinters in the wood, or where the wood and the rail are slightly closer together.
posted by Killick at 10:29 AM on June 4, 2007

Check out The Time Fountain for an example of the strobe effect.
posted by monkeymadness at 10:30 AM on June 4, 2007 [1 favorite]

Wikipedia on the "wagon-wheel effect" under continuous illumination.
posted by contraption at 10:41 AM on June 4, 2007

If it happens at specific points on the track, make sure there are no hidden ventilation or cooling fans which could be blowing the drops upward.

If the rail were to be hot enough at those points to cause the leidenfrost effect, which seems highly doubtful, they could crawl up the rail on a cushion of their own steam, as discussed previously on MetaFilter, and demonstrated here.
posted by jamjam at 11:05 AM on June 4, 2007

The third rail, according to Wikipedia at least, is electrified with DC; I seem to recall it's about 400 volts in NYC, although I don't know where I'm getting that figure.

That suggests that the water droplets would acquire a considerable charge by touching the rail. My guess is that they then go to the nearest ground; "electromotive force" or EMF is the physicist's name for the physical force derived from electric fields which can accelerate charged masses.
posted by ikkyu2 at 11:39 AM on June 4, 2007

Yeah, if it's daytime I doubt the fluorescent lights are causing that effect. I feel like I've seen this before and concluded it was a reflection causing it. I'm seconding aubilenon.
posted by miniape at 2:24 PM on June 4, 2007

I've seen this a dozen times on the DC metro third rail, in natural daylight, with no vents to push the water up.

Always assumed it was an electrically-generated phenomenon--thanks for the explanation Killick.
posted by garfy3 at 2:37 PM on June 4, 2007

If enough salt is dissolved in the water droplets to make them a little bit conductive, current flowing through the rail could exert force on the droplets magnetohydrodynamically.

Current passed through electrodes protruding into a thin tube from the sides, for example, will pump mercury along that tube-- with a scream like a banshee from cavitation, by all reports.
posted by jamjam at 7:21 PM on June 4, 2007