Since long time ago, many authors advocated for more presence of physics history in physics teaching and learning in order to give students a better vision of the "nature of science", in other words, to let them learn not only established physics knowledge but also the ways of how physicists managed to get that knowledge. Generally, historical episodes are either not treated in physics textbooks (and consequently in teaching) or their treatment is reduced to a minimum: scientists' names and years. When an episode is treated with more details then some of those details are erroneous. In this article we analyze how physics textbooks authors treat one of the most famous experiments in physics history: Cavendish's determination of the density of Earth. Authors of all revised textbooks continue to repeat erroneous information, claiming that Cavendish measured the gravitational constant. As the erroneous nature of that claim for the Cavendish experiment was demonstrated many times in pedagogical and other journals, it is normal to ask: why do the authors continue to repeat it? Our hypothesis is that the "culture of teaching" is different from the "culture of research" regarding the appearance and correction of errors. We believe that only way to fight against errors in textbooks is to establish better mechanisms of veracity control, similar to those in research journals.