Chilling after Fermentation

[Steve, Cape Town]

There's a theory - supported by those in the industry (Fullers, for example) - that suggests that chilling, after fermentation, and prior to bottling, will improve the quality of the finished product.
I have tried to include this chilling process and have found that there is a noticeable extraction of yeast and other possible undesirables at the bottom of the fermenting barrel after about a week at +/-4C. (A darkish colour, not like the cleaner, caramel colour of the yeast that settles after a couple of days of fermenting).
My problem is that I have probably extracted too much yeast and so when I transfer to bottles the secondary fermentation fails to occur and the resulting beer is completely flat.
I have spoken to a micro brewer here in Cape Town, and he suggests that I need to add a small amount of yeast prior to bottling, which seems to make sense.
Does anyone have any experience in this area and can they suggest how much yeast I should add to a 40 pint barrel. And would the caramel-coloured yeast extracted soon after fermentation occurs (stored in a fridge until needed) be suitable?

[Aleman]

Funny that this has come up again

According to Jamil Zainasheff, dried yeast contain approx 20 billion cells per gramme. Most brewers bottle condition with between 1-3 million cells per ml of beer (Sierra Nevada use 2 million in their pale ale so we'll go with that).

So

20L of beer is 20,000ml and we need 2 million cells in each ml

= 40 billion cells in a 20L batch.

= 2g of rehydrated dried yeast assuming 100% viability.

I've achieved successful carbonation after 8 weeks lagering at -1 to -2C . . .. but when I add yeast to ensure carbonation I try an achieve a paint thin layer of yeast on the bottom of the bottle, so would use no more than 500,000 cells per ml