Last year I tried to explain away my added marathon minutes by suggesting that I may have had too many red blood cells. So I've been looking again in doctor daughter's medical books for confirmation of that and, yes, it's possible.

It's the "Ketchup Effect" : tomato sauce comes out of the bottle very much slower than red wine does. The graph in Sue's book shows blood thickening only slowly for anaemic people, but when their red cell counts creep into the "normal" range, the effect is more dramatic. Indeed, beyond the thicker end of normal, the rate of blood flow through the body's pipework goes down to only a third of what it was below the thin end. As the old saying goes, "blood is thicker than water".

The Ketchup Effect would be neutral if the oxygen-carrying part of the blood, the haemoglobin, increased in volume to offset the reduction in ketchup flow. Unhappily, it does not. The upshot of this is that the rate of energy production - and hence the speed we can race at - reduces twice as fast as the rate of haemoglobin increase.

That is not all. Blood will get thicker simply by losing water when we sweat it out in a race. From time to time, I check my fluid loss rate by weighing myself, stripped, before and then after a run. On an ordinary Spring day, I lose a pound for every 3 miles : that would be almost 9 pounds in 26 miles. Aaaargh !! Every pound of water short, if we're near the thick end of the blood's normal range, will slow us down by around 1%, say 2 seconds a mile at 3-hour marathon pace. Nine twos are 18 (seconds per mile), call it 8 minutes in a marathon. Aaaargh !! So that's why I pour water over myself at every "drinks" station.

I shall be taking extra precautions in London this year. I've bought two bottles of Isostar, a litre in total, which I shall drink just before the start. That's almost a quarter of what I shall need during the race and it'll see me through the first few aid stations which are usually so crowded. And I'll be carrying the water bottles with me for half a mile to pour over me then so that I don't waste as much as I usually do.

If this occasion's Experiment of One works out well, I may even run a negative-split race for the first time in 10 years.

And now for something completely different ...