Mike T wrote an article on the 'Long Mynd Valley' thread, suggesting that below 20degrees one need not drink at all during a marathon. This is it:

"Running a marathon, at 100 calories a mile, takes approximately 2600 calories. About 25% of this results in work/movement but 75% just results in heat which has to be dissipated to the environment or we would quickly become hyperthermic and die. 75% of 2600 is 1850 calories of heat to lose, and as long as the ambient temperature is less than 20C it is estimated that sweating only has to dissipate half of this ( or less as it gets colder ) as the rest is lost through radiation and convection. 1ml of sweat loses about 0.6 of a calorie so 925 calories needs about 1.5 litres of sweat. But glycogen breakdown liberates water at 2.7mls/gm - if half of the energy of the marathon is derived from glycogen at 4 calories/gm (and the rest from fat ) then about 800 mls of water is liberated and about another 400mls is generated as a by product of energy use ( carbohydrate plus O2 being converted to CO2,H2O and energy ). So the body has come up with 1.2litres of the 1.5 needed. But during exercise our temperature rises, so we don't need to lose all the heat we generate. A typical temperature rise would be from 36C to 39C, and as each degree of temperature increase is about 60 calories, or 100mls of sweat, we can subtract 300 from 1,500 leaving 1,200 - so there is no need to drink at all!"

So to those who understand the science in this article how many will now stop drinking during their marathons?