Andre points us to SciImago — a Web site to mine science journals. Using their aggregates per country and some data from Wikipedia, I made up a table on number of science papers produced per country going back to 1996.
| Country | Science papers (1996-2006) | Population (current) | Papers per capita |
| US | 3,437,213 | 303,202,683 | 0.011 |
| Japan | 983,020 | 127,718,000 | 0.0077 |
| UK | 962,640 | 60,587,300 | 0.015 |
| Germany | 888,287 | 82,251,000 | 0.010 |
| China | 758,042 | 1,323,128,240 | 0.00057 |
| France | 640,163 | 64,102,140 | 0.010 |
| Canada | 473,763 | 33,148,682 | 0.014 |
| Italy | 461,292 | 59,206,382 | 0.0077 |
| Spain | 330,399 | 45,116,894 | 0.0073 |
| India | 286,109 | 1,131,043,000 | 0.00025 |
| Sweden | 194,921 | 9,174,082 | 0.021 |
| Switzerland | 188,134 | 7,508,700 | 0.025 |
| Israel | 120,257 | 7,222,222 | 0.0166 |
| Norway | 70,314 | 4,738,085 | 0.015 |
What is fascinating is that the picture changes dramatically if you just look at the most recent year (2006):
| Country | Science papers (2006) | Population (current) | Papers per capita |
| US | 340,268 | 303,202,683 | 0.0011 |
| China | 166,205 | 1,323,128,240 | 0.000125 |
| UK | 107,528 | 60,587,300 | 0.0018 |
| Japan | 97,073 | 127,718,000 | 0.00076 |
| Germany | 95,310 | 82,251,000 | 0.0012 |
| France | 67,652 | 64,102,140 | 0.0011 |
| Canada | 56,571 | 33,148,682 | 0.0017 |
| Italy | 54,298 | 59,206,382 | 0.0009 |
| Spain | 41,914 | 45,116,894 | 0.0009 |
| India | 38,140 | 1,131,043,000 | tiny |
| Switzerland | 22,966 | 7,508,700 | 0.003 |
| Sweden | 20,926 | 9,174,082 | 0.002 |
| Israel | 13,049 | 7,222,222 | 0.0018 |
| Norway | 8,670 | 4,738,085 | 0.0018 |
These numbers suggest some significant changes:
- The US is still leading in the number of papers produced, but it no longer dominates. And it may not lead for many more decades if China keeps this up.
- Canada, Switzerland, Norway, Spain and Italy are improving their per capita numbers.
- Switzerland has a surprisingly high number of papers per capita.
- Japan has a surprisingly low number of papers per capita.