(FORTUNE Magazine) – Parents have been trying to choose the sex of their children for as long as they've been having them. In ancient Greece, Aristotle counseled men to tie off their left testicle to guarantee a son. The Talmud advises husbands to hold back and let wives take their pleasure first if they want boys. In late-20th-century America, an obstetrician named Landrum B. Shettles built a near-cult following for a best-selling book that advised timing sex to ovulation: several days before if you want a girl, at or near the event for a boy. None of these, nor a thousand other recipes that humans have cooked up over the ages, have been proved scientifically. That's another way of saying that, at least in the eyes of the medical establishment, they just don't work.

The Genetics & IVF Institute, an infertility clinic in Fairfax, Va., has a sex-selection technology that does work. By separating male from female sperm, the process, called MicroSort, gives parents the power to choose the sex of babies before they're conceived, with a high degree of success--especially, so far, with girls. "Finally someone has developed a method of sex selection that is honest and not a fraud," says Sherman Silber, director of the Infertility Center of St. Louis at St. Luke's Hospital and author of How to Get Pregnant With the New Technology. "Because everything prior to MicroSort has been clearly in error or frankly fraudulent."

The Virginia institute didn't invent MicroSort; improbable as it may seem, the technology is a creation of the U.S. government. A Department of Agriculture scientist invented it in the late '80s as a way to pick the sex of livestock. That nascent business will bring in $100 million a year in the U.S., Europe, and Japan by 2010, estimates XY Inc., the Fort Collins, Colo., company that is developing the technology. But the privately owned Genetics & IVF Institute, founded by maverick entrepreneur-physician Joseph Schulman, holds the exclusive license to apply MicroSort to humans until the government's patent expires in 2009. Without fanfare, the institute has embarked on a gradual but potentially explosive plan to take the choosing of boys and girls out of the realm of wishful thinking and into the world of available science--and lucrative business.

Sex selection is virtually unregulated in the U.S., to the distress of ethicists. That's because Congress has steered almost entirely clear of the hornet's nest of New Age baby making. The Food and Drug Administration has the power to pronounce on a machine's safety and effectiveness--the institute says the FDA has approved its sperm-sorting machine--but not on the ethics of how it's used. As for the USDA, it is loath to get in the way of licensees trying to commercialize its inventions. Its technology-transfer program, which generates $2.6 million a year in royalties from 221 licensees, was set up to move government inventions off the shelf and into commercial use. While the USDA theoretically has the power to stop a licensee from sublicensing, it has never exercised that power. So when MicroSort is unleashed in a big way, it will be controlled mainly by market forces--such as the budgets and desires of parents themselves.

MicroSort first made headlines in 1998, when the institute announced it had successfully produced infant girls for 13 of 14 couples in a clinical trial. Since then the institute has delivered 134 more MicroSort babies, with at least 90 more on the way, and it is looking into sublicensing the technology. Meanwhile, because the USDA has patented the technique only in North America, Europe, Australia, and Japan, doctors from countries like India and South Korea (where the cultures strongly prefer boys) are free to pursue the technology now. Some are already doing so. "It's an immense market," concludes Robert Edwards, the British scientist who co-engineered the first test-tube baby.

Institute founder Schulman, 59, a secretive sort who refuses almost all interview requests (including ones for this story), gave up his CEO title in 1998 but still exercises considerable power as the institute's medical director and chairman of its board. He has a long history of pioneering in ethical gray zones. Under Schulman, the 350-employee institute charged several women $11,000 each for removing an ovary and freezing pieces of it before cancer chemotherapy--a method of preserving fertility that had succeeded in only one large animal (a sheep) and that is unproved in humans. The institute also flouts federal law by refusing to report its in-vitro-fertilization success rates to the Centers for Disease Control. Schulman argues that such numbers are meaningless for an individual patient. "Joe thinks he can do anything, pretty much," says a former associate. "And a lot of what he does is motivated by the business of medicine, not the beneficence of medicine." The institute declines comment, but Schulman defender Samuel Marynick, medical director of the Baylor Center for Reproductive Health in Dallas, chalks up such comments to sour grapes. "It's hard to be at the top of the mountain," Marynick says. "People are shooting at you all the time."

Ten years ago, Schulman and Ed Fugger, a crew-cut Texan reproductive biologist who had developed the institute's huge sperm bank, visited the Beltsville, Md., lab of Lawrence Johnson, a USDA scientist. They knew that the unassuming Johnson, widely considered the world expert on sperm sorting, had spent the previous decade developing an effective sperm-sorting technique for animals.

What Johnson had done--and what the USDA had applied to patent--was adapt a sophisticated cell-sorting machine called a flow cytometer to exploit the fundamental difference between male and female sperm. Male sperm carry a Y chromosome; females an X. The X chromosome holds more DNA--2.8% more in humans and about 4% more in animals. So when sperm stained with a DNA-seeking fluorescent dye are zapped with a laser as they flow through the machine, female sperm, having taken up more dye, glow more brightly. This difference in brightness allows the cytometer to sort the sperm by sex.

Johnson walked Schulman and Fugger through his lab, where two flow cytometers were sorting bull sperm. More than 100 rabbits, cows, pigs, and sheep had been born using the technique, he told them, and they were all perfectly normal; the dye and laser didn't appear to damage the sperm or the resulting animals.

Schulman and Fugger asked about the possibility of using the technique to sort human sperm. They had a particular group of potential patients in mind: couples who carry genes for rare inherited diseases, such as hemophilia and muscular dystrophy, that occur in boys. (Nearly all sex-linked genetic diseases affect boys, not girls.) These parents could benefit hugely if sperm sorting could be moved successfully from animals to people.

Johnson had tried sorting human sperm, but had hit a daunting obstacle: His machines weren't sensitive enough to check whether sorted samples were, indeed, heavily male or heavily female. (The machines could check animal sperm because the larger 4% DNA difference makes the job far easier.) Schulman and Fugger told Johnson they had good news. They had already refined a lab technique for analyzing chromosomes that would allow them to check the sorted sperm in a petri dish. Huddling in Johnson's small office, poring over machine readouts, the three men hatched a collaboration. Johnson would sort human sperm in his lab; at the institute, scientists would check the sorted sperm for their male and female proportions. Two years later, in a groundbreaking paper in the peer-reviewed medical journal Human Reproduction, the men reported that they had successfully sorted human sperm.

In the meantime the institute, with Fugger in charge, moved quickly to launch a sperm-sorting project, buying and adapting a flow cytometer. There was plenty of incentive: Three weeks after the USDA patent was issued, in August 1992, the USDA granted the institute exclusive license to use the technology for the patent's full 17-year term. Such a long license period is standard for the USDA, which reasons that licensees need that much time to make a return on their often considerable investment. (Neither the institute nor the USDA will reveal the licensing fee the institute paid.)

The institute soon launched a trial to demonstrate scientifically that MicroSort was both safe and effective. The trial was approved by the human experimentation review board at nearby Inova Fairfax Hospital, whose parent company, Inova Health System, is part owner of the institute. In the freewheeling world of reproductive medicine, the institute wasn't required by law to seek anyone's approval beyond getting the FDA's nod on its sperm-sorting machine. But it was important to MicroSort's scientific reputation--not to mention its commercial prospects--to be seen as taking the ethical high road.

The hospital board approved the trial only for couples who carried genes for inherited diseases afflicting boys. The first MicroSort baby, a girl, was born in 1995 to a mother who had lost two male fetuses (and three brothers) to so-called X-linked hydrocephalus, a disorder that fatally swells the brain. Since then the clinic has produced daughters for six more couples with disease genes. More girls are on the way: Julie Strasser, 33, an elementary school teacher in Peachtree City, Ga., who carries the gene for Duchenne's muscular dystrophy, is expecting twin daughters in May. Her 30-year-old husband, James, a plastics-plant operations director, calls MicroSort "an answer to our prayers."

As word of the trial spread, says Fugger, his phone began ringing. It wasn't the parents of hemophiliacs who were calling; it was dozens--then hundreds--of couples who, for no medical reason, were eager to choose the sex of their next child.

Fugger pondered how he might expand the trial without causing an ethical uproar. His solution was to offer the treatment to parents trying to conceive a child of the sex found in fewer than half of the family's existing children. Unmarried couples, childless couples, and wives over 39 would be excluded. Critics could hardly blast the institute for opening an ethical Pandora's box: MicroSort would be correcting sex ratios, not skewing them. That argument found sympathetic ears at the USDA, which had licensed the method only for use against boy-linked genetic diseases. At the institute's request, officials broadened the license to include any human uses. (A USDA spokesperson refused comment.) A newly constituted institute ethics board also approved the change.

In 1995 the institute began offering the service it calls, delicately, "family balancing." Fugger says he's thrilled that he is not just helping provide the children of parents' dreams, but also helping them avoid the financial strain of having a passel of children before the boy or girl they want appears. "It's just a win-win situation for families who want to use this," he says. Karen and Bill, an upstate New York couple with three young sons, enthusiastically agree. Karen gave birth to daughter Meghan, via MicroSort, last July. "My boys are my everything," says Karen, "but I still wanted to experience that mother-daughter bond. Now I'll get that chance to see her go on her first date, be there on her wedding day. I am still on cloud nine!"

While MicroSort is straightforward as science-assisted-baby-making procedures go, it also involves considerable time and money. Couples must submit to a battery of preliminary blood and semen tests. Women often take ovulation-inducing drugs, which insurers may or may not pay for. (They never pay for the sperm sorting.) Couples then must travel to Fairfax, where most women go through several days of ultrasound monitoring before doctors determine that they are ovulating. On the appointed day, the husband produces a sperm sample in the morning. A machine sorts it, and that afternoon doctors inject the male or female fraction into the wife's womb by catheter--a procedure known as intrauterine insemination, or IUI.

The odds of getting pregnant by MicroSort are now 21% a try, the same chance that fertile couples have of getting pregnant the old-fashioned way in any given month. The average MicroSort couple tries three times, at $3,200 a pop, before getting pregnant or dropping out. (Last month the institute started offering a cheaper $2,300 option for couples who pinpoint the woman's ovulation time themselves, whether by having her own gynecologist monitor her or by using over-the-counter ovulation predictor kits.)

Of the MicroSort babies and current MicroSort pregnancies where the gender has been confirmed, the success rate for conceiving girls is currently higher than that for boys--90% vs. 70%. Among the reasons: X sperm glow more brightly than Ys in the sorting machine, so it's easier to pick them out. The chances of conceiving a boy are climbing steadily, though, as the technology improves. "I would anticipate that ultimately the two would be equal," says USDA scientist Johnson, who recently retired.

Despite the drawbacks, the institute continues to field between 250 and 500 calls a month from interested parents, who often must wait several months to enroll. "Yesterday I had a call from Nigeria," said Fugger in late October. "The day before I had a call from India. Last week I had a call from Germany. They are coming from everywhere." In response to the demand, Fugger has doubled his laboratory staff and is upgrading his three $300,000 sperm-sorting machines so that they can sort a sperm sample in two hours instead of four. He'll soon add another sorter. The changes, he anticipates, should double the lab's capacity. Right now, operating seven days a week, the clinic handles 60 couples a month.

Fugger says that the trial will end after it has demonstrated unequivocally that MicroSort is safe. He estimates that will take as many as 750 babies because genetic abnormalities are so rare in the general population, against which the MicroSort babies have to be compared. The trial has revealed no problems so far, and MicroSort's animal-safety record is spotless.

However, the institute's quest for impeccable safety data in no way limits its freedom to do business, and to sublicense, today. "Literally we could open a site tomorrow anywhere we wanted," says Fugger. "We are developing a strategy to provide access to this to patients not only in the U.S. but in many other countries." He says that MicroSort will probably be available outside Fairfax in three to five years, with providers eventually making profits of 30% or so, but he won't say exactly where.

Some fertility doctors are skeptical about why the institute is taking so long to sublicense MicroSort, wondering if the technology really works. "If you make a claim in a peer-reviewed journal that you can do something that nobody else can, then it's a bit unusual not to be sharing the technology so that others can replicate its efficacy and safety," says Alan Copperman, the director of reproductive endocrinology at Mount Sinai Medical Center in New York City. Fugger responds that the institute hasn't yet sublicensed because of the complexity of the technology, the need for highly trained people to run sperm sorters, and the fact that most of his energy has been focused on the clinical trial.

Of course, outside North America, Europe, Australia, and Japan, anyone can start MicroSorting today, without a sublicense. "It's something that anybody that really knew what they were doing could do," says Sue Sharrow, manager of the flow-cytometry facility at the National Cancer Institute.

If that isn't already happening, it's likely to very soon--and Asia is the most probable place. "The preference for male children is universal in Asia," says Dr. Indira Kapoor, the South Asia regional director at the International Planned Parenthood Federation. According to a December United Nations report, amniocentesis and ultrasonography have made abortion of female fetuses "a booming business" in Asia, especially in India, China, and South Korea. South Korea leads the world in ending girls' lives before they begin--115 boys are born there for every 100 girls.

Many would argue that using MicroSort to choose a boy before conception is a better idea than aborting a girl after conception. Still, it is hard to overstate the outrage and indignation that MicroSort prompts in people who spend their lives trying to improve women's lot overseas. Wide availability of MicroSort technology, if it comes, "will be disastrous for society," says Kapoor. "People will go for boys. Girls will not even be conceived."

The technology will certainly be too expensive for most Asian couples, at least initially. But in a population of billions, appealing to a sliver of affluent customers can deliver up hefty profits. So it's not surprising that doctors from India and South Korea, Thailand and Hong Kong are calling the institute, clamoring to know when they can start MicroSorting. Mervyn Jacobson, chairman of Cytomation, a Fort Collins, Colo., manufacturer of animal sperm-sorting machines, says that he's had inquiries from groups in China, Turkey, Taiwan, and Brazil wanting to buy his equipment for human use. When first contacted by FORTUNE in December, Jacobson said the company was preparing to ship a sperm sorter to South Korea for human use this year. The South Korean importer, Jacobson said, is "a very sophisticated professional scientist" in the midst of negotiating a sublicense with the institute. But, says Jacobson, the scientist won't agree to an interview and has since put the deal on hold. The institute declines comment.

Ethical concerns aren't limited to Asia. In surveys, a consistent one-third of Americans say that they have deep moral problems with sex selection--even before conception, and even for families that already have a large number of children of one sex. The objections don't seem to lie in concerns about sexism but in a deep-rooted antipathy to meddling with nature.

U.S. ethicists, meanwhile, worry about the social effects of widespread MicroSorting. Americans have a long-standing preference for first-born sons; will we become a nation of little sisters, conferring selectively on sons the purported advantages of first-borns? (Fugger says that more than half the couples who've approached him want girls.) Will MicroSort children face unforgiving pressure to be just the kind of boy or girl their parents dreamed of? "What happens if you end up having a girl who wants to climb trees and play with mud? Who doesn't have any desire to go to ballet class?" worries Lori Knowles, an associate for law and bioethics at the Hastings Center, a medical ethics think tank in Garrison, N.Y.

Most of all, ethicists fear that a broad embrace of MicroSort would open the door to yet more troubling practices, like jiggering genes to produce designer babies, with everything from hair color to IQ ordered up. MicroSort would "accustom people to getting what they order. A child becomes something that is increasingly a commodity," says Dorothy Wertz, a social scientist at the University of Massachusetts Medical School-Shriver Division in Waltham, Mass.

Worried about being ostracized, most MicroSort patients will not agree to be publicly identified. "Many people who go to MicroSort don't tell others about it. People don't agree with the procedure," a Florida mother of two boys wrote in an e-mail to FORTUNE just before getting in her husband's private plane to fly to the institute. "My husband and I are not telling our friends or family."

How big is the potential market? The back-of-the-envelope calculation goes something like this: Each year, some 3.9 million babies are born in the U.S. In surveys, a consistent 25% to 35% of parents and prospective parents say they would use sex selection if it were available. If just 2% of the 25% were to use MicroSort, that's 20,000 customers. If they, like current MicroSort patients, average three tries at getting pregnant at today's $3,200-per-try price tag, that's a $200-million-a-year business in the U.S. alone. "We estimate a U.S. market of between $200 million and $400 million, if [sperm sorting] is aggressively marketed," says Samuel Isaly, an analyst at OrbiMed Advisors, a New York City asset-management group that owns a $50 million stake in Swiss fertility-drug company Serono.

Cost could be the chief obstacle to MicroSort's becoming as common as, say, amniocentesis. "It's expensive enough that the pregnancy rates will have to be very high before MicroSort is widely used by fertile couples," says Richard Scott, director of Reproductive Medicine Associates of New Jersey in Morristown.

But many experts predict that costs will fall. Consider the flow cytometers used for sperm sorting--big, complex instruments that cost up to $500,000 and are designed for dozens of tasks, from diagnosing leukemia to discovering new genes. "If there's enough market, somebody will make a little tiny machine designed only to do [human sperm sorting]," says the National Cancer Institute's Sharrow. That will bring down the price, she says, along with the level of expertise required in operators.

For some, MicroSort may be the stuff of dreams; for others it's more like a nightmare. But for all of us, this technology will soon force intense examination of what we believe about having children, about being parents, about our kids and their place in the world. As Maria Bustillo, a Miami fertility specialist and former institute doctor, puts it: "Just because you want a particular kind of baby, should you have it? Just because you can afford it, should you have it? It's what keeps me awake at night sometimes."