When activated, in the figure at the left, the PARTICLE SOURCE emits a countable stream of photons; particles of light; about 20 per second. They must pass the slitted blue barricade to impact the dark screen. Think of the screen as a large array of small (pixel sized) photo detectors. Any one of them, when struck by a photon, lights up.

1. Cover the right slit so that the rain of particles can only go through the left one. To do so hold-down-click and drag the purple barrier to the right until it covers whatever slit you wish to cover. Then click on the 'activate source' button to release the particle stream.

2. Next, without clearing the screen, send them through the right slit (cover the left one). For another 10 seconds.

The resulting pattern on the screen is for particles that passed either through the left slit or the right one. You may 'clear the screen' later but now

3. scroll down to the second motion graphic. It's a copy of the first.

4. Leaving both slits open, activate the particle source.
Allow 20 seconds of accumulation.

The screen pattern that emerges is not what one gets from particles passing either through one slit or the other. 'The 'both slits open' pattern is not the same as the 'one slit open plus other slit open' pattern.

You see it clearly by comparing the results in the two motion graphics.

This is the key finding in quantum mechanics. Its significance is this: The experimental results defeat elementary logic.

Elementary logic: any particle falling on the screen must have gone through one or the other of the slits. The left-slit-only and the right-slit-only pattern together represent particle paths through one or the other of the alternatives - left or right slit. It follows that the only-left-slit-open pattern plus the only-right-slit-open pattern should equal the both-slits-open pattern.

But they do not! The both-slits-open pattern does not match this sum. Thus, the experimental result contradicts elementary logic. Instead of the expected sum the both-slits-open experiment exhibits interference; nothing detected at some screen positions.

It is upon this finding that quantum mechanics is built. Quantum mechanics is the logic that embraces the actual physical world observations; results typified by this experiment. The important thing is its match to experiment - even though quantum logic does not match our traditional common sense.

Quantum logic acknowledges that, in this experiment, it is not knowable through which of the two slits a particle goes. Since, with both paths available one cannot say which path was taken, quantum mechanics posits that the particle takes both! Each particle takes all paths available simultaneously. Not just one. This takes-all-paths notion contradicts our common sense. But experimental tests over many years validate it. If the path taken is not knowable then the particle takes all paths.

When both slits are open something called a particle's amplitude to go via one slit interferes destructively with its amplitude to go via the other. That's why we see interference: screen regions where no particles fall. In those places the particle's amplitude to arrive at the screen via one path interferes with its amplitude to arrive via the other.

The indisputably best place to acquire understanding about this is Richard Feyman's little book called "QED" (1985).

Here is an experiment to mark and, therefore, know which path is taken; a 'which path taken' experiment.

The 45° polarizer behind the 'photon source' insures that only 45° linearly polarized photons meet the slits. These are polarized in the direction symbolized by the lines drawn on the polarizer; at 45° off the vertical. (For the nature of polarization see http://chesters.org/poles/poles.swf)

The polarizers behind the slits are path markers. Photons passing through the left slit are tagged as horizontally polarized; those passing through the right are tagged as vertically polarized. So each photon, when it arrives at the screen, carries the mark of its path. Which path it took is knowable; its polarization tags the path.

(3). Activate the photon source with the 'activate the source' button. Accumulate screen detections for about 10 seconds.

The resulting screen pattern shows no interference. The pattern is the common sense expectation; the sum of left-slit-only plus right-slit-only photons. Since they're tagged, traditional logic prevails. A photon arriving at the screen is known to have gone through one or the other of the slits because it carries a marker. There is no interference because we can know whether the particle went through the left slit or the right one. It's polarization - horizontal or vertical - tells us.

That we choose not to know - not to measure the polarization of particles arriving at the screen - is of no consequence. It is not a matter of what an observer decides to measure. It is a matter of identifiability; whether the thing is measurable in principle. What destroys interference is that the polarization of the arriving particle is identifiable, even if we choose not to identify it. The consciousness of the observer has nothing to do with it.

This slit-path knowledge can be erased.

For convenience continue the experiment in this new motion graphic.

(5) Hold-down-click on the high floating 45° polarizer and drag it down behind the two slit-path-tag polarizers - behind the horizontal and vertical pair. Now, regardless of the polarization of the entering photon, every photon exiting this polarizer must be 45° polarized.

(6) Click on the 'activate source' button.

Passing photons through this arrangement restores the interference.

The which-slit-path information is erased. Reason: the photons falling on the screen are all polarized at 45°. No way exists to know their history; to know whether a photon was formerly horizontally or formerly vertically polarized. Hence no photon falling on the screen has a slit-path tag. It's not knowable through which slit it went. The two paths interfere and we see an interference pattern.

Remarkably, the which-way information can be changed continuously from unknown to known! Altering the polarization direction of the lowered polarizer does it. Setting it at angle θ instead of 45° it is found that the screen pattern changes continuously from interference to non-interference and then back again as θ is varied between -45° and +45°.