› firstname.lastname@example.org 
Why, umpteen zillion years into the AIDS era (I used to volunteer for Gay Men's Health Crisis in the 1980s), is there still no useful data about the risks of oral sex for men? Have we really learned nothing since the first appearance of "Low risk but not no risk"? With the understanding that not letting someone come (or precome) in your mouth is a start (but also loses a lot of the appeal), is there any sensible way to assess and reduce the risks of the common American blow job?
Loyal East Coast Reader
Actually, the relative risks of the Great American Blow Job have been much on my mind of late. I'm working on an article about whatever happened to the heterosexual AIDS epidemic and what straight, middle-class ladies should do about HIV when they start dating again after their marriages break up. (Quick answer: nothing. They're not going to encounter any, but while they're taking unnecessary precautions against HIV they're incidentally protecting themselves from real menaces like human papillomavirus and herpes.) Not that this applies at all to your question or your demographic; what's sauce for the goose, after all, is not necessarily sauce for gander and gander.
Back when you were first volunteering in New York and I was out here gearing up to become a sex educator, nobody knew nothin', and the safest thing to do was to lump everything that might possibly be dangerous into "Thou shalt not" and try to get people to take a "100 percent safe" pledge. I suspect that then, as now, the people most likely to achieve 100 percent safety weren't at much risk to begin with, while the hard partiers continued to party hard-ly, no matter what their T-shirts said. I know for a fact that politically aware womyn at the time would not shut up about woman-to-woman transmission, which turned out to be so much poppycock or poppyhen, as they might have had it. Likewise, the much-ballyhooed heterosexual AIDS epidemic never made it off the cover of the news magazines and into the bedrooms and bloodstreams of straight America.
So, your question. If there were a definitive answer to that, it would be coming out of a few labs here in San Francisco. But of course, HIV being a shifty bugger and human behavior being even worse, there isn't. There are animal studies (using simian immunodeficiency virus, which is similar but by no means identical) demonstrating that you can easily spread the virus by swabbing monkey tonsils with an infected Q-tip. Then there are the epidemiological studies like HOT, the HIV Oral Transmission study, dedicated to finding those cases in which a guy gave blow jobs but never, ever, ever had unprotected anal sex and seroconverted anyway, and that is so complicated a business I'm going to let one of the researchers explain it:
"I'm going to conclude with the HOT study, in which, again, we interview men who we screen and rescreen to ascertain that, in fact, their only risk is oral sex. So they are a special population, and they are screened and rescreened, and they get their HIV test, and eventually we do another very in-depth interview, and after three corroborating screenings, or two screenings and one interview in which they say they've only had oral sex, 25 percent later report a higher risk exposure anal sex in the same time period after we get them in another environment with a different questionnaire and a face-to-face interview, and this is after they've been told that, in fact, they're negative. And so we see this working many ways, and they're, like, 'Whew! Well, now I can tell the truth.' But in fact, of those 363 men, we estimate that up to a quarter of them probably weren't having only oral sex, and so I think that we have huge problems in terms of self-reported risk behavior."
That was from a very informative experts' roundtable discussion I found on HIV Insite (hivinsite.ucsf.edu), a UC San Francisco site I have just declared required reading for the interested. The good news is that the best work currently being done is readily available to us for free. The bad news is that, due not to bad science or lack of science but to the vagaries of human memory and human motivation, they still can't really answer your question. How many new HIV infections are caused by fellatio to ejaculation? I'll let the above experts answer that. It's funny but not, you know?
JK I think we agree it's less than 5 percent, don't we?
SB Uh, ... yes, I'd probably say it's it may be less than 5 percent. I'd say 5 percent or less. But I wouldn't say 1 percent either.
JK Well, 1 percent is less than 5 percent ... [Laughter]
KS Well, I wouldn't say "5 percent or less."
SB So I don't know that we're going to come to consensus on that.
And what's the best way to reduce whatever risk there is? Not going down on HIV-positive men. Easy for me to say, sure, and awfully glib, but you can't say it ain't so.