An expert in semi-classical acoustics and quantum chaos has found mathematical proof for a theorem that many women knew instinctively to be true, writes Caroline Davis.
Matthew Wright of Southampton University's Institute of Sound and Vibration Research has shown that the formula widely used to calculate a woman's bra size bears almost no resemblance to the size that she will end up buying.
"Consumers will usually find their preferred size by trial and error," Dr Wright said, "but if you look on websites for lingerie, they all have this section on how to calculate your size."
The research, "Graphical analysis of bra size calculation procedures", published in the International Journal of Clothing , came about after a friend of Dr Wright's mentioned she had lost weight but found that her cup size had increased.
His interest in the problem, as an engineer, was to look at the effects of changing a single parameter in a calculation.
Dr Wright said he did not use empirical research methods. He sat down with at his computer for an evening, translating the measurement procedure into algebra.
The standard procedure is to measure around the ribcage in inches, adding four to an even result, five to an odd to give the band size. Cup size is calculated by subtracting this result from the fullest bust measurement.
But because one rounded number is subtracted from another - or as Dr Wright describes it "the difference between two rounded quantities" - errors of up to 3in can creep in. On a graph of bra size domains, this procedure led to bra sizes such as 34C and 36A sitting adjacent to one another.
Dr Wright suggested a modified method for calculating size, where the cup size subtraction is done before the band measurement is rounded up, which would produce more accurate results.
The research could have important consequences. For example, statistics on incidence of cancer to breast size may be based on data that detail cup size.