My father is a physicist who doodles atoms and formulas on his dinner napkins and who once appalled me by saying coolly, looking at our cat, Bobo, "What a remarkable machine."

I spent the first years of my adult life trying to be as different from him as I could, starting in college, when I made a U-turn away from science, read John Keats and William Blake, and wrote--or tried to write--lyric poetry. And now here I am, in spite of everything, thinking constantly about atoms, molecules and behavior. I am fascinated by a field that Keats and Blake might have hated more than any other, a field that often feels, to the romantic in me, like a threat to reduce us all to the status of remarkable machines.

The field that fascinates me is the exploration of the first links between genes and behavior, between what is often referred to in the headlines as DNA and destiny. The scientists I follow most closely in this field are not psychologists. They don't administer personality tests, and they don't interview identical twins. They are laboratory biologists who look at life from the genes up, from the molecules up. Their research stayed out of the news until recently because they work with flies, worms and mice, not with people. But lately their experiments are coming closer to home. The time I've spent hanging around in fly rooms has given me a preview of the sorts of difficult truths that this science may soon be telling us about ourselves.

If we pinpoint genes linked to schizophrenia, for example, or aggression, or the pursuit of happiness--and there are, of course, still biologists and psychologists who doubt such genes exist--are we allowing science to paint us into a corner? What might it mean if one group of people on this planet tends to carry the most aggressive versions of that aggression gene, and another group tends to carry the mildest versions? Finding a genetic basis for schizophrenia may lead to better treatments, but what advantage would there be to knowing that you are programmed to be just plain happy or sad?

These problems are in many ways the logical and inevitable conclusion of 20th-century biology. The modern theory of the gene was born in a fly bottle in 1910, when biologist Thomas Hunt Morgan noticed a single mutant male fruit fly with white eyes. Morgan paired that fly with female flies that had normal, bright-red eyes. He and his students watched for white eyes in the fly's family tree as the white gene traveled down the generations. With those white-eyed flies, biology began to move out of the sunlit, outdoor realm of 19th-century naturalists and into the interior realm of chemists and physicists.

It was a physicist-turned-biologist whom I've come to know well who took the next step inward in the 1960s--using flies. Seymour Benzer talks about watching his two young daughters play at the beach and wondering how much of their behavior was inherited. He hit on a novel and simple approach to this oldest of questions: He would go back to the point of origin of modern genetics, the fly bottle, and he would do just what Morgan had done; but rather than tracing the inheritance of mutant eyes, wings and hairs, he would trace mutant behavior.

Benzer's first big breakthrough came in 1971, when he and one of his students announced their discovery of a fly with a warped sense of time. Most flies live according to circadian rhythms: They wake and sleep on a regular cycle, just as most humans do. But Benzer's crew discovered flies with fast internal clocks, slow clocks and broken clocks, and they traced these problems to specific mutations in a fly gene they called period. Flies, worms, mice, butterflies and mustard weeds carry these genes--and so do we.

In fact, quite a few of the genes Benzer and his students found in the fly rooms--genes that shape everything from flies' sexual instincts to their memories--are closely related to the human genes that we watch in action in our living rooms and bedrooms. In the mid-1990s, one of Benzer's students, Tim Tully, found a particularly interesting memory gene called creb. Ordinary flies need 10 lessons to learn to avoid a shock. The creb mutant needs just one lesson. That fly has the fly equivalent of a photographic memory.

I visited Tully at Cold Spring Harbor Laboratories, on Long Island, just after he made his discoveries with creb. While I was there, a few fly people (as they call themselves) showed me how to inject a piece of DNA into a fly embryo. It is like playing a video game. You watch the embryo on a big TV monitor. With a joystick, you micro-manipulate the tip of a needle toward the embryo's rear end. Step on a foot-pedal, there is a tiny burst of turbulence at the tip of the needle, and the deed is done. If that particular ribbon of DNA takes--and if it takes in the right place--then a new gene will run in that fly's family, generation after generation, and give thousands of the fly's children and grandchildren a photographic memory. And yes, the creb gene shapes human memory, too.

So far, most studies of human genes and behavior are nowhere near as solid as Benzer's studies of flies, although you wouldn't know that from the stories in the newspapers, which tend to wear a sort of cocky know-it-all smirk. Front-page headlines about the discovery of the "Happiness Gene," the "Gay Gene," or the "Novelty-Seeking Gene" make it sound as if all this human behavior is now understood and nailed down. Actually these studies are extraordinarily difficult, the scientific results are usually tentative, and they are often retracted. (News of those retractions usually is buried on page 37.)

But within a year or two, every letter in the human genome will have been transcribed. Then biologists will be able to explore the links between DNA and human behavior as easily as Benzer explored those links in fruit flies. This is the kind of research in which knowledge leads to power, sometimes terribly fast. The same foot pedal, joystick and needle I played with at Cold Spring Harbor can inject a human gene into a fly or a mouse, and create what ancient Greeks and young molecular biologists call a Chimera, like the lion-goat-serpent hybrids of legend. And someday sooner than we think, these same tools may allow biologists to inject genes that shape memory, or musical ability, or even (who knows) a poetic or a mathematical temperament into the early embryo of a human being in a laboratory dish.

Last week I had lunch in Pasadena with Benzer and asked him if he ever feels queasy about helping to launch the human species on this adventure. Benzer brushed the question aside. He belongs to the old school of investigators who believe it is their job to explore, and society's job to decide what to do with the territory they open. He has always done his best work at the very beginning of scientific revolutions, and he pointed out that he could not have predicted at the outset where any of them would go. His tinkering with germanium alloys in 1943 helped lead to the invention of the transistor and to the electronics revolution, which is now exploding in directions that no one could then foresee. On the one hand, he said to me, electronics have given us guided missiles that can kill millions in a matter of minutes. On the other hand, they have given us the World Wide Web, in which Benzer delights. So he refuses to join in the general agitation over the direction genetics research is leading us.

Benzer's studies of genes and behavior are helping to unravel the genetic causes of neurodegenerative disasters like Lou Gehrig's disease and Alzheimer's (a disease that Benzer's wife, Carol Miller, a neuropathologist, is studying with the help of discoveries her husband made in his fly room). Benzer himself, at 77, has gotten interested in the mystery of aging. Recently he identified a mutant fly that lives 100 days or more, about 30 days longer than the other flies in the bottle. Benzer calls that mutant methuselah, and is searching for the human equivalent of methuselah genes.

I wouldn't want Benzer and his school to stop working on methuselah, or any of the other genes he studies, because the medical possibilities are too valuable. And I think Keats and Blake would approve of some of the metaphysical possibilities as well. This is the point at which we connect, for the first time, the outer, objective world of the physicist with the inner, subjective world of the poet. This is where we take what we've learned about stones and rocks and trees--the universe of matter--and connect it with everything inside us and everything we do, which is, for the poet in us, the universe that really matters.

On both fronts, we may be at the very beginning of something extraordinary. The whole of the 20th century may come to look like what astrophysicists call a singularity--a point of origin, a big bang, as new fields open up for medicine, perhaps for lyric poetry, and certainly for psychology. After all our thrashing around on the couch, after thousands of years of self-analysis and tragicomical self-misunderstanding, this millennium will go out with the punch line that Philip Roth gave to "Portnoy's Complaint." "So," said the doctor. "Now vee may perhaps to begin. Yes?"

But I do still feel queasy. Science has now put us all into the fly bottle and we may spend the 21st century trying to get out. We all want to feel that we are free agents--not only chips off the old block but sculptors of our particular chip. Will we lose that possibility--which is the possibility of possibilities? Or as my father put it when I started hanging around with fly people and mutant flies, "Have they found the free-will gene yet?"

Jonathan Weiner is the author of the recently published "Time, Love, Memory: A Great Biologist and His Quest for the Origins of Behavior" (Knopf).