NASA yesterday formally opened its investigation into a major error in the construction of the $1.6 billion Hubble Space Telescope as shaken scientists and engineers continued to analyze the problem and speculate on possible causes.

Many expressed disbelief that such a flaw could pass unnoticed through what is described as a rigorous system of checks and tests. Members of Congress expressed dismay and the chairman of a key Senate subcommittee called a hearing for this morning.

One of the telescope's two mirrors appears to have been made in the wrong shape, engineers said, and the result is that the telescope's view of the heavens, designed to provide unprecedented clarity, is blurred.

NASA officials said they hope to correct part or all of the defect by putting what amounts to a set of prescription eyeglasses on the orbiting observatory, but that cannot be done until 1993 at the earliest, when shuttle astronauts are scheduled to mount a replacement camera on the telescope.

Until that happens the $40 million Wide Field and Planetary Camera, which was to have served as the telescope's workhorse, providing some 40 percent of the scientific return, may not be useable.

James Westphal, a professor of planetary science at the California Institute of Technology and the principal scientist in charge of the camera, said it now appears that he and his colleagues will be able to do "little, if any science" with it. "It means sitting around for another three years."

Westphal has been involved with the Hubble telescope for 13 years.

Senator Barbara Mikulski (D-Md.), whose state holds two space telescope facilities and who chairs the committee that funds the space program, responded with unusual anger. "I am very outraged at what has happened to Hubble," she said. "They have had 10 years to put this together. They spent 2.8 billion dollars to be able to get it right."

She said the incident has called into question future NASA programs, such as other planned Great Observatories and a mission to Mars. "Are we going to keep ending up with techno-turkeys?"

Sen. Albert Gore Jr. (D-Tenn.) called a hearing of his science, technology and space subcommittee for this morning at which NASA and contractor officials will be asked to explain the foulup.

NASA officials said the investigation could take months. It will be led by Lew Allen, director of the Jet Propulsion Laboratory, which is operated for NASA by the California Institute of Technology. Other members of the panel, including some who are to be independent from the space agency, have not yet been named, officials said.

A team of about 40 people at Hughes Danbury Optical Systems Inc., in Connecticut, where the mirrors were manufactured, was gathering the voluminous documents, computer tapes, engineering notebooks and other records dating back to 1977 when development of the mirrors began, officials said. The material is to be impounded and a team of NASA engineers was to arrive at the company soon.

Jack Rehnberg, chief of the company's space science office, said "something inherently, fundamentally was not done right."

The telescope's optical assembly, which includes its high-precision 94.5-inch primary mirror and a 12-inch secondary mirror, took about three years to grind and polish and was completed in 1981, according to Hughes spokesman Thomas Arconti. "Testing was done. The mirrors were checked, all along the way. And the tests showed good results. That's where the mystery lies."

"We're talking about an apparent problem large enough that you can't imagine it wasn't found," said Sam Keller, associate deputy administrator of NASA, who was responsible for a major shakeup in the telescope project's troubled management at the end of that period.

Scientists said yesterday that although it is possible that both mirrors are flawed, it appears that only one is faulty. Westphal said that during polishing and measurement of such mirrors, special test lenses and mirrors are often used to calibrate the accuracy of the polishing.

With the large primary mirror, for example, a small lens about the size of a fist would have been used to examine the mirror for flaws. By looking through the test glass and using laser light, the mirror-makers could examine their work and search for bright or dark splotches indicating flaws. These flaws would then be polished out.

Westphal said it is possible that the test lenses were flawed, which would in turn have produced "perfect but wrong" mirrors. It is also possible that instructions were bungled as they were transferred from computers, which were used to design the mirrors, to the workshop, where the mirrors were made.

"The mirror could be designed right, but somewhere from a computer to a sheet of paper, an eight got changed to a six or a six got changed to an eight and the mistake propagated through the system," Westphal speculated. "You could make a mirror that looks absolutely right, but it's the wrong shape."

NASA officials said yesterday that instead of testing the mirrors together before launch, computer simulations were run to save money.