The medical literature is growing at an alarming rate. Research assessment exercises, research quality frameworks, league tables and the like have attempted to quantify the volume, quality and impact of research. Yet the established measures (such as citation rates) are being challenged by the sheer number of journals, variability in the "gold standard" of peer-review and the emergence of open-source or web-based journals. In the last few years, we have seen a growth in downloads to individual journal articles that now easily exceeds formal journal subscriptions. We have recorded the 10 top cited articles over a 12-month period and compared them to the 10 most popular articles being downloaded over the same time period. The citation-based listing included basic and applied, observational and interventional original research reports. For downloaded articles, which have shown a dramatic increase for the International Journal of Cardiology from 48,000 in 2002 to 120,000 in 2003 to 200,000 in 2004, the most popular articles over the same period are very different and are dominated by up-to-date reviews of either cutting-edge topics (such as the potential of stem cells) or of the management of rare or unusual conditions. There is no overlap between the two lists despite covering exactly the same 12-month period and using measures of peer esteem. Perhaps the time has come to look at the usage of articles rather than, or in addition to, their referencing.