From the six-year-old's "my dad can beat your dad" to the war-monger's "our political system is freer than your political system", human beings have been comparing themselves and their artifacts since at least the time of Cain and Abel. Scientists, predictably, have evolved a set of quantitative methods for doing this sort of thing, called scientometrics ("the science of science"). It addresses issues like patterns of collaboration among institutions and countries, recognition of outstanding contributions and contributors (before they go to Stockholm), the processes by which underdogs join the mainstream, and evaluation of how much a given facility or research group is contributing to progress in its field.
Now that the streets of scientific research are no longer paved with gold (whether you think this happened in 1990, 1980, 1970, 1960 ... depends mostly on how old you are - it was always just before your time), agencies that support our salaries and our equipment are beginning to task for relatively objective ways of picking the right horses to back.
It was against this background that I recently completed a study of large optical telescopes throughout the world, asking about which ones have contributed most to astronomical knowledge in recent years. The method was a simple one: go through all the journals in which astronomical papers are published (there are only a couple dozen) over a representative period (18 months in 1990-91) and count the published papers and pages reporting data from every telescope mentioned. Thirty-nine telescopes with mirrors larger than 2 meters across, located in 10 different countries, turned up. Then look into the Science Citation Index and count how many times each of the 1163 papers was referred to by someone else in 1993-94.
The number of papers says something about how much new science had been produced, and the number of citations indicates how important the results were to other astronomers. Such data can never be completely up-to-date. Images you collect tonight (August 2nd) will be processed (by our graduate student?) in the fall, submitted for publication in spring, 1997, published in fall 1997, and be cited most often in about 1999. Thus neither the Hubble Space Telescope nor the Keck 10-meter are represented (though they can be in a few years).
What did I find? The average telescope was responsible for about 20 papers per year, and the average paper was cited about 3.5 times per year. But the range was enormous - a factor of 10 or more in papers per telescope and citations per paper. Some correlations were obvious. Don't expect to learn much from a telescope located where it rains a lot, or from one that you can't afford to maintain. Thus bigger isn't always better, and the Russian 6-meter is much less productive than the 2.1 meter at Kitt Peak in Arizona.
There were also surprises. Within the US, the publicly-owned telescopes that are available to everybody on a competitive basis (like the Kitt Peak 4-meter) were neither much more nor much less than the privately owned ones available only to observers at particular universities or observatories (like the Palomar 5-meter and the 3-meter belonging to the University of California).
And "buy American" is not, at the moment, obviously the best strategy. The most productive telescope, with 94 papers and 467 citations, turned out to be the 3.9 meter Anglo-Australian Telescope (located in Australia and shared by astronomers from the two countries). Tied for second, depending on whether you go by papers or citations, were the 3.6 meter Canada-France-Hawaii Telescope (located in Hawaii and available to astronomers from Canada, France, and the University of Hawaii) and the Cerro Tololo Interamerican Observatory 4-meter (in Chile and available primarily to American and Chilean astronomers).
What does this mean? I am not quite sure. The AAT site is only moderately good compared to CFHT and CTIO. The potential user community for CTIO is much larger than the other two. And so forth. For all three, most users must come from a long ways away at considerable expense. Perhaps this provides some self-selection in favor of hard workers who know and love the particular facilities they are using. If so, then the results do have some applicability outside of optical astronomy, though the implications will not necessarily be popular in these days of peer review, equal access, level playing fields, and so on.
Oh, and in case you were wondering about the title of this piece, the UC 3-meter Shane telescope is, of course, located at Lick Observatory. Here are Ansel Adams images of Lick in the "Fiat Lux" exhibit. The complete data set and discussion will appear in Scientometrics 36, 237 later this year.