Thomas J. Katz signs a small letter in Science warning us about the propagation of errors in review articles. The author gives a scary example of an incorrect citation propagated through 9 reviews (if I counted correctly). The cited paper does not contain the experiment that all the reviews mention and as it seems it was actually never published anywhere. Very scary.
As more science moves online with more individual voices, will this propagation of errors be accentuated or reduced?