Well, without showing off about it, my background is in biological science, and I'm extremely sceptical about this. Let's see what's changed over the last couple of centuries. We don't take temperatures under armpits any more, a very unreliable method (in fact, I used to get older kids to experiment on this and we found wild inaccuracies, though I did find a lad in my class with pneumonia once...) Temperatures taken axially, orally, aurally or anally all give different answers and all suffer from a degree of unreliability via very human inconsistencies of approach by both measure and recipient. They are a great tool for doctors to confirm their diagnoses but they are rarely intended to be a firmly scientific measure. Second, in the good old days all was done in Fahrenheit. As the distance from one Fahrenheit degree to the next is much smaller than the distance from one Celsius degree to the next (a factor of 5:9), the temptation to round up or down to the nearest whole degree F was justified in consideration of human body temperature. A bit of a degree Fahrenheit one way or the other was so small as to be not worth medically bothering about, so rounding would have been rife. Third, thermometers have increased in accuracy and reliability. This issue has been addressed in the matter of global warming, in which much agonising and corrective techniques have been applied to older readings, still much argued about. We are talking about less than half a degree Celsius here. The study is interesting but that's as far as it goes.