No effect that requires more than 10 percent accuracy in measurement is worth investigating. See image Walther Nernst
