Warren Ellis points out that the print prozines' circulation is declining.
Sadly, this has been true for a long time. In 1991, Analog had a circulation of 83,000, and Asimov's had nearly 80,000. The number have been in more or less steady decline for at least that long. In fact, Analog's circulation in 1976 was 110,000; I don't have numbers for the '80s, but I would venture a guess that, with some brief reversals (including a couple of slight upticks in the late '90s), the numbers have been mostly declining for thirty years. (Numbers in this paragraph come from a Truesdale piece in Steven Silver's Argentus (PDF file), issue 1, fall 2001, 'cause it came up in a web search and I was too lazy to go downstairs and look up numbers in Gardner's Year's Bests. I later did go look up some numbers there for the years around 2000; I owe some of the discussion below to Gardner.)
And this is not news to the sf world. Pretty much every major con has a panel on the magazines and/or the state of short fiction, and the circulation decline is always one of the topics. There are many theories about the decline; the one that I tend to focus most on is that reading of short fiction (regardless of genre) has been in decline for a long time.
In 1937 or so (if I'm remembering right), Theodore Sturgeon could make a living, living in NYC, by selling one short story a week to the national newspapers. Back then, Damon Runyon was selling stories for a dollar a word and was a household name all over the country.
Things have changed since then. These days, people aren't so interested in reading short stories. In my experience, most people who read fiction want to read novels.
On the other hand, outside of sf there are still some fiction-oriented magazines that have larger circulation. Wikipedia says Granta had a circulation of about 46,000 copies in 2004, for example; it doesn't publish only fiction, but I think fiction is its main focus. So there may be all sorts of other reasons for the decline in circulation of sf print prozines.
For example, editors on those panels tend to mention that the magazines are shelved in hard-to-find places in bookstores, while people who dislike what's published in the magazines tend to claim that there's a vast readership hungry for good short sf (by which is meant "stories I like"), "unlike the crap that gets published in the magazines." There's occasional discussion of distribution issues, especially the implosion of distributors in the late '90s. There's discussion of the fact that reducing the number of copies you send to a newsstand that doesn't sell many copies can actually increase your sell-through (and your profits) even though it makes your print run go down. Someone usually mentions that even magazines that are doing quite well and/or turning a profit can get cancelled; witness SF Age and Omni Online. There's much worry that short sf is too insular, that it requires too much background in the field, that kids can't get into it, that it's not focused enough on providing science education, that it's too dry. And so on. Someone invariably says that the solution is to move all publishing online; someone else invariably responds that they hate reading from a computer screen.
Anyway, I don't think anyone knows the definitive answer as to why circulation has been declining for so long. But I'm heartened to see a lot of people trying a lot of different approaches.
There are people publishing print anthologies. There are people engaging in a wide variety of online publishing models; for example, I was told recently that Escape Pod, the weekly podcast of short speculative fiction read aloud, is doing quite well. (And it pays for both original stories and reprints.) It seems to me that sf is becoming more and more acceptable in mainstream venues, by and large (usually without being called sf). And so on.
And, honestly, I don't see the print prozines disappearing in the near future. Their circulation has, it's true, been declining, but (for example) Analog's current circulation is, I think, about what Asimov's was seven or eight years ago; that suggests to me that even if numbers continue to go down, Analog probably has at least another decade left. And as Gardner has noted at various times, one nice thing about digests (and quasi-digests) is that they're very inexpensive to produce, so it's still possible to turn a profit at relatively low print runs.
I feel a little silly writing all this out. Gardner has written much the same thing in his Year's Best intros for at least ten years now, and every year there continue to be panels and lamentations about the imminent death of short sf, so I'm sure this entry won't have much effect.
But I should've thought of that before I wrote it. Now that I've written it, I may as well post it.
And now I'm gonna go pack for WorldCon, where I'm sure we'll be discussing all this again.