The gloriously mild weather in the south-east over the past couple of days has increased my sense of urgency to get on with a little experiment I have been planning for some time. The objective is to come up with a formula for estimating the temperature, based on a survey of the outer clothing worn by a sample of people observed going to work in the morning.

When it was first decided, about a quarter of a century ago, to phase out Fahreneheit temperatures and replace them with Celsius (which we then called Centigrade), I heard a good instant guide to the new system: at 10C a man takes his coat off, at 20C he sheds his jacket, and at 30C shirts are removed. More recently, however, I have learnt that the system may be extended for colder conditions: when the temperature drops to 5C, scarves are added, and when it hits zero, hats and gloves appear too.

What I plan to do over the next few weeks - if the cold weather returns - is to keep records of the numbers of hats, coats, scarves and gloves being worn by fellow bus-passengers each morning, then correlate the percentages wearing each garment with the temperature.

Initially at least, the sample will have to be limited to male commuters because of the hat factor. I understand that some females don millinery more as a fashion statement than as protection against the weather, and such behaviour would produce an unwanted bias in the figures. Ultimately, we should be able to able to produce a formula of the form: T = K - (hH+gG+sS+cC) which will enable the temperature T to be calculated from the percentages H, C, G and S of men wearing hats, coats, gloves and scarves respectively. All that remains is to work out the value of the constant K and the weightings, h, c, g, and s to be given to each garment. Ultimately, we may refine the figures to include such measures as Hf (the percentage wearing furry hats) and Cr and Cw (raincoats and winter coats).

If this works, we should be able to dispense with the rather cumbersome formulae currently being used to assess wind-chill factor, which is the cooling effect caused by the cold wind constantly blowing away the comforting duvet of warm air that our bodies exude around them to protect from the external chilliness.

I do not know what clothing people shed if the temperature reaches 38.9C, but it would bring a particularly warm feeling to one person in Hayle in Cornwall. For bookmakers William Hill report that one such man has placed a pounds 2,500 bet at odds of 20 to 1 that the temperature in England, Scotland or Wales will, at some time this year, reach a record 100F. In view of this large wager, the odds have already dropped to 14 to 1.