WASHINGTON (AP) _ X-rays applied at the right time and in the right dose may allow some healing of severed spinal cords _ and the partial restoration of use of paralyzed limbs _ according to laboratory research with rats.

Nurit Kalderon, a scientist at Memorial Sloan-Kettering Cancer Center in New York, said extensive additional research is needed before the technique could be applied to human patients. But she said it does suggest a new way of treating injuries now often considered hopeless.

Kalderon, lead author of the study, said the X-rays apparently halt the action of cells that block the regrowth of injured nerve fibers. These cells, which she called reactive astrocytes, are made by the body, for reasons unknown, after a spinal injury.

She said that the studies on rats found the technique had to be done in the third week following the injury.

A report on the study is to be published Tuesday in the Proceedings of the National Academy of Sciences.

Dr. Michael Walker, director of the division of stroke and trauma at the National Institute of Neurological Disorders and Stroke, said ``Dr. Kalderon has performed a terribly interesting series of studies.''

``She is using a therapy (X-rays) normally used to kill cells to cause cells to survive,'' he said. ``We don't know yet what the side effects would be, so I am cautious about it. I would like to see the work repeated in another lab.''

Spinal cord injuries often result in total paralysis of limbs. Researchers say this occurs because nerve fibers in the central nervous system, which includes the spinal cord, do not regenerate as do most nerve tissue elsewhere in the body.

``The tissue in the area of the injury will actually decay after a certain point,'' said Kalderon.

Early in the recovery from a spinal cord injury, she said, the body makes an effort to regrow nerve fibers. But at some point this halts and the injury to the spinal cord is irreversible.

That point in rats, said Kalderon, is about three weeks, the time when the reactive astrocytes cells appear. X-rays, she said, prevent the formation and action of the these cells, but the radiation has to be precisely timed.

``At four weeks, the radiation is not helpful,'' she said. ``It has to be during the third week.''

Kalderon tested the X-ray treatment on a series of rats whose spinal cords had been completely severed.

At three weeks after the injury, some of the rats were exposed to X-rays at doses similar to that used to treat cancer. A group of control rats received no X-rays.

After four to five months, about a third of the rats who had been exposed to X-rays showed some sign of spinal cord regeneration.

Of 11 rats exposed to the X-rays in one group, ``six definitely had some recovery of function,'' said Kalderon. ``We could see them move their back legs and to support their weight. They could plant their feet on the ground.''

She said none of the rats recovered fully, however. This, said Kalderon, would require extra therapy, a treatment not practical with rats.

In other groups of animals, the regrowth of nerve fiber was tested electrically. When the brains of test rats were stimulated electrically, sensors placed near the hind legs of the test rats recorded nerve impulses, proving that nerves had grown together inside the spinal cord. In some cases, the impulse reached only a single muscle, said Kalderon, but there were no such impulses detected at all in the control rats who received no X-rays.

And in still another study, the nerve connections were tested with a dye that moves along nerve bundles from the spine. Kalderon said the dye test showed that nerve connections were restored in about a third of the rats tested.