• Number of papers
  The numbers of papers published are growing rapidly, as are the number of science journals. A study a few years ago in Issues of Science and Technology (Caroline Wagner, 2011) shows a basic dataset that exemplifies how this has been changing, comparing countries:(the data splits multiple authors fractionally between different countries).
  Most clearly apparent is the rise of China, while other mature nations still increase strongly their outputs.

The number of papers published in the journals Nature and Science though increases only slowly over time, with the largest increase in the 1960s, followed by a reduction back to around 1000/year to maintain the highest competition.

[data counts number of articles in each journal per year]

• That this is exponential growth is clear from looking over a longer timeline on a log scale:
  (from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2909426/ which has a good discussion about the problems of counting properly).

• I like the motivation behind the idea of giving each new academic scientist a lifetime quota of some number of published printed journal pages. They can either use these pages or sell/trade them to more productive researchers (who because they are more successful will have the funds to pay for them). A market in delivering outputs and impacts could force science to consider this as a limited resource, then competing for attention of others.

• The amount of time scientists spend reading each journal paper has decreased over time:
  (from Science 2009, http://people.ischool.illinois.edu/~renear/norobots/StrategicReadingSCI09.pdf).

• There is some evidence that interdisciplinarity makes impacts slower and smaller, though it is unclear if this is a problem of visibility or of quality.
  http://arxiv.org/ftp/arxiv/papers/0908/0908.1776.pdf: as the science output becomes more interdisclinary, the paper is less likely to have impact.
  http://www.cwts.nl/TvR/documents/AvR-Citation-delay-Scientometrics.pdf: citations to interdisciplinary papers take longer to rise.