Normally by the second week of life, RP has extensively damaged the eye, covering the entire retina with black specks of pigment. But the photoreceptors that had taken up PDE genes were still clear one month after birth.

It wasn't until the fifth week that specks began to appear, apparently because - for unknown reasons - the transplanted genes stopped working. Total destruction of the treated area didn't occur until after the eighth week.

Bennett found the results "amazing" and wondered if the treatment would last longer in humans afflicted with a less virulent form of RP.

Bennett didn't immediately tell Wilson about the results. Instead, she spent three months doing more studies to make sure that the data were solid. She still wasn't ready to say anything when Wilson's office called, requesting an update on her work, last July. With little comment, she sent slides showing the treated and untreated retina.

Two days later, she received a handwritten note from Wilson asking her to call him.

"This is spectacular," he said when she phoned.

A few weeks later, Bennett found herself standing before the institute's research and development committee and talking about human trials.

As the date for the executive committee meeting approached, Wilson huddled with Bennett, Maguire and Jacobson and pressed them with questions about the incidence of the disease.

Wilson wanted to know how hard it would be to recruit patients for the trial.

Bennett pointed out that Jacobson was a national authority on RP, with a large number of patients who came to him from all over the world.

But how many have the PDE defect?

That's difficult to say, Maguire said because the precise genetic defect usually can't be identified with a routine eye exam. That requires costly lab studies, which most ophthalmologists don't bother with since all forms of RP are incurable.

Only a small percentage of RP patients have been identified as having the PDE defect, Jacobson said. How many? Wilson wanted to know.

They knew of five.

Wilson was shocked.

Couldn't they go after a different form of RP with more identified patients, he asked.

Maguire said that more patients were known to have the form of RP called choroideremia, because it can be picked up with routine eye exams, but that no animals were known to have it.

Bennett thought it was essential to test out prospective treatments on animals before going to humans, so they had to do PDE.

Wilson didn't want to abandon the project. He was impressed with Bennett's animal studies and considered diseases of the eye, which is easier to reach than other organs, ideal for gene therapy.

He wanted Penn and his institute to lead this research, but he did not want to spend hundreds of thousands of dollars preparing for a clinical trial without the assurance of test subjects.