Key points are not available for this paper at this time.
We study the statistics of citations from all Physical Review journals for the 110-year period 1893 until 2003. In addition to characterizing the citation distribution and identifying publications with the highest citation impact, we investigate how citations evolve with time. There is a positive correlation between the number of citations to a paper and the average age of citations. Citations from a publication have an exponentially decaying age distribution; that is, old papers tend to not get cited. In contrast, the citations to a publication are consistent with a power-law age distribution, with an exponent close to -1 over a time range of 2 -- 20 years. We also identify a number of strongly-correlated citation bursts and other dramatic features in the time history of citations to individual publications.
S. Redner (2004) studied this question.