The second one mentions Jacques Benveniste who is the guy I was on about.
For the sake of fun I'll describe the experiment and lets see if anyone can spot the flaw before I get to it

I'm going from memory but I think I've got the basics down.
He was studying allergic reactions. The chemical that caused the allergic reaction was made into a solution in water. When added to the blood cells and cell that reacted would turn bluish. A researcher would then stare at a slide under a microscope and manually count the cells that turned blue.
What the researchers found was that when they diluted the solution down to a level where it should have no effect and repeated the experiment with new cells they found that some of the cells still turned blue. They continued to dilute the solution and repeat the experiment, eventually getting it down to a level where the odds were good that there wasn't even a single molecule of the chemical in the solution but still the cells would always be blue.
Jacques thought that this might be proof that water was somehow "remembering" the chemical that had been present in the solution and designed an experiment to prove it.
One set of cells were given water while another set of cells were given a highly diluted solution that was so diluted that it should have no more of an effect than plain water. The researcher again sat down at the microscope and counted the number of stained cells from the water sample and then from the diluted sample.
The result was that the highly diluted sample always had a much higher number of blue cells than the water sample even though there was very little chance of there actually being any of the allergen in it.
A paper on the subject was written up and submitted to Nature who hated it. They said that since this was an amazing claim they would only publish it if they were allowed to send their own experts there to verify the experiment. The paper was published and then the experts went in (James Randi a famous debunker was also sent along).
This time the experts insisted that the researcher who counted the cells didn't know which was the dilute sample and which was the water sample. At this point the bottom fell out of the entire experiment and Jacques dream of a nobel prize failed completely.
What had of course been happening was that if a cell looked a little blue the researchers were counting it for the dilute sample and discarding it for the water sample.

As soon as the person doing the counting didn't know which sample was which the reults were more accurate.
What amazed me is that a man who was considered on the track to a nobel prize didn't even consider doing a double (or even single) blind test on something this important.

EDIT : Here's a link to the
BBC Horizon programme on the subject.
It mentions that there is a 1 Million pound prize for anyone who can prove that homeopathy (i.e memory of water) works under lab conditions

2nd Edit : Hopefully some more clarity added
