In a feat that could point the way to the growth of any mammal - including a human being - outside the womb, a scientist here is raising embryo mice in the laboratory until they are halfway through gestation and their heart cells begin beating.

The scientist is Dr. Yu-Chih Hsu of the Johns Hopkins School of Hygiene and Public Health.

The mouse fetuses developing in plastic dishes in the laboratory are not yet complete test-tube babies.

"I don't know whether or not it will ever be possible" to bring a mouse or any other mammal, including a human, completely to "birth" or maturity outside the womb, he said last week.

"It might never be done. Or it might be done in the mouse by someone in a very short time.

"In the human being? I hope never. That is not necessary, and it raises too many ethical questions."

Still, Dr. Hsu has been recognized in scientific circles for an outstanding achievement in growing mammalian embryos to so advanced a stage.

If there is ever complete test-tube development of any creature, for good or for ill, he will be remembered for taking a long step on the way.

Hsu does not try to fertilize his mouse mothers' egg cells in his laboratory himself, the way Britain's Patrick Steptoe and Robert Edwards fertilized the egg cell that two weeks ago became baby Louise Brown.

Instead, he finds it easier - and cheaper, since he has little help and sparse funds - to remove normally fertilized embryos from mother mice at the day-and-a-half-old, two-cell stage, the stage where the fertilized egg has just divided.

Other scientists have fertilized mouse egg cells with mouse sperm successfully. After just a few days, they did successfully implant their embryos into mouse mothers.

Hsu says he does not consider either of these steps necessary to accomplish his main goal, which is simply to observe his embryos' development for as long as possible.

As long as possible means 11 or 12 days, which is just over half of a normal mouse's 19 days in the womb. It is the equivalent of nearly five months for a human fetus.

"We can't yet construct the maternal environment completely," he admitted. And no, he said, he does not really have "artificial wombs" in his laboratory, only standard laboratory Petri dishes a few inches across in which his little embryos grow. Still, they do grow.

He is happy to demonstrate.

Getting ready to approach a newly pregnant mouse, he heats the end of a glass syringe to make a pipette, or suction tube, with a fine point.

He opens the oviduct of the mouse, the duct that carries the fertilized egg to the uterus. He gently removes the mouse's usual seven to 10 speck-like eggs.

He places these in a broth in his petri dishes. These go into an incubator kept at a constant 98.6 degress Fahrenheit in an uncomfortable 98.6-degree room in which he concedes that he spends too much of his time, seven days a week, month in and month out.

As the embryos grow, they ultimately attach themselves to the Petri dish, just as they might attach to a womb. And he gives them a daily change of environment: the broth, made at various stages of ordinary serum, fetal calf serum or serum from human spinal cords, as well as standard laboratory culture media rich in proteins, enzymes and vitamins.

After 11 to 12 days the specks become round islands a tenth of an inch across. Under the microscope these islets come alive.

For inside each is a plainly developing creature with two hemispheres that are the first signs of a brain, an elongated structure that is the growing spinal cord, a dark mass of future vertebrae and, most dramatic of all, a clump of vigorously beating cells that are the primitive heart.

In his words, "the slightest deviation" in reconstructing "the delicate and precise 'in vivo' process" - the process as it takes place in the living womb - "may cause complete arrest."

After 11 to 12 days in any case, he says a bit sadly, the embryos "become disorganized and fuzzy and gradually disintegrate. This is so far as long as I've been able to take them."

Still, this has been long enough for him to show that this mammalian embryo does not need to be attached to a mother to develop to a fairly advanced stage.

It has been long enough for him to refute the belief that contact with the uterus is needed before an embryo can start to "differentiate" into brain, spinal cord, heart and other specialized tissue.

Maturing is a little more slowly than they would in the womb, his mice reach almost the halfway point in their physical development.

Six years ago, Hsu was first to bring any mammalian embryo past the crutical point of "implantations," the stage at which it normally attaches itself to the womb. After working for nine years, he has learned only in the past six months to bring fully 50 percent of his incipient mice to the 11-to-12-day stage.

One day, he said, he hopes science may be able to use trays of mouse embryos as a valuable and sensitive test system to learn whether new drugs or chemicals cause birth defects or cancers.

Meanwhile he is struggling to get renewed support from the National Institutes of Health, which once gave him $45,000 a year. He said he was cut off after a visit from a review committee which "didn't know what they were seeing."

He is supported currently by Johns Hopkins. But he is awaiting a visit by a new NIGH committee. He hopes that in the aftermath of the birth of a human "test-tube baby" - though a baby that spent only its first 50 hours in the laboratory - money for work like his will be more available.

"I think the atmosphere is different now," he said. "I think people who want babies will tell Congress they want such research into human fertilization."

What did he think when Britain's Steptoe and Edwards finally delivered a child conceived in a laboratory?

"I thought it was very good news for a couple who so much wanted to have a baby," he said.

"No, I don't think the religious and ethical objections to having a by this way are just nonsense. But the parents' happiness may be more important."

He paused. A stolid man, he stood among his delicate laboratory tools and charts of developing fetuses.

"It is instinct for people to have children," he said. "If you can't have a child the usual way, this is better than adoption. I don't think you can stop it."