The journal Nature published a study of the most 100 most cited scientific papers that provides food for thought to the working stiffs (I used to be one). First of all, I find it quite natural for any researcher to keep an eye on how often his or her work is cited. Other than a personal reaction from colleagues and friends, there is no better way to find out how well one's paper is accepted beyond the confines of one's own lab. And who would not want to know? So, the topic of frequency of citations touches most of us.
The paper says that by far the most cited papers deal with techniques. The Lowry method (1951) for determining protein concentration tops the list with over 300,000 citations and is followed by a raft of others that include methods of biochemistry, bioinformatics, and phylogenomics. Neither Einstein's paper on the special theory of relativity nor Watson & Crick's discovery make the list of the top 100. The reason for this seems clear. Once an important discovery becomes part of the scientific canon of the time, there is no need to quote who did it. Woe if we did — every time we mention water, we would have to add: Lavoisier, A. 1783. Observations sur la Physique, 23. 452 − 455.
But let me be a bit more nuanced. A factor that comes to mind is not just the number of citations, but their timing. This brings up what I will call the Magasanik Paradox, named after that wise bacterial physiologist, who is quoted as saying — more or less verbatim — that the more important a publication, the shorter the time it will be cited. This metric was not included in this report, although the data certainly exits and can be extracted for any individual paper. So, there is a bit of homeostasis-for-the-ego here: if your paper isn't all that important, at least it stands a chance of being occasionally cited for a long time!