A previously unknown bacterium has been identified as the cause of the so-called Legion fever that killed 29 persons infected during an American Legion convention in Philadelphia last summer, scientists at the federal Center for Disease Control in Atlanta announced yesterday.

The CDC director, Dr. David Sencer, said scientists believe the same organism might have been responsible for a mysterious pneumonia outbreak at St. Elizabeths Hospital here in 1966 in which eight persons died.

While the isolation of the bacterium represents a giant step forward in the attempts to solve the Philadelphia mystery, scientists still do not know where the bacterium came from, why it appeared when and where it did, why it infected the persons it infected or how it was spread.

But without the bacterium, said Sencer, there was no chance of answering those questions. Now, he said, scientists at least know what they are looking for.

CDC officials said that while they do not know how the bacterium was spread, they are sure it was not spread by personal contact and does not appear to be contagious.

Yesterday's announcement casts further doubt on speculation that the outbreak resulted from a mass poisoning, was caused by nickel carbonyl, had something to do with cigarette smoking or was a result of some bizarre germ warfare plot.

A bacterium is a minute organism, larger than a virus. Bacteria are associated with such common human diseases as pneumonia - which the Philadelphia disease resembled - strep and staph infections.

The scientists came upon the bacterium while looking for rickettsia, microorganisms of the type that cause typhus and Rocky Mountain spotted fever. "In the routine process of making sure we had not missed a rickettsia, we found the bacterium," said Dr. Joseph E. McDade, the research microbiologist who isolated the organism.

Scientists at the CDC isolated the bacterium, which has yet to be named, by injecting guinea pigs with lung tissue removed from some of the victims of the Legion illness.

Material was then removed from the spleens of the animals - who died, apparently from the disease - and was injected into the yolks of fertilized eggs.

The chicken embryos also died and were found to contain many of the previously unknown bacteria.

The next step in narrowing down the cause was checking blood samples from some of the 151 persons who survived the disease for antibodies to the bacterium.

Of 33 blood samples on file in Atlanta, "29 . . . had at least a four-fold organism (antibody) rise," said Sencer, who added that researchers didn't find "anything like this" in those who did not suffer from the disease.

The researchers then checked for the same antibodies in blood samples, stored since 1966, from 14 of the victims of the St. Elizabeths outbreak. Thirteen of the 14 samples showed a marked increased in antibodies, said Sencer.

During the 1966 outbreak 62 patients at the federal mental hospital in the District contracted acute pneumonia symptoms similar to that of the legionnaires and eight of them died, the CDC said.

According to Dr. Charles Shepard, chief of the CDC's leprosy and rickettsai branch, the isolated bacterium is a peculiar one.

"It grows in yolk sacs and we aren't sure it will grow in artificial medium," he said. "Bacteria by definition grow in artificial medium and we aren't sure this does. We aren't sure of its place in the bacterial world."

The scientists' next step is attempting to find out where the disease came from and how it spread.

The second step, he said, is figuring out how to look for the bacterium in the environment, as well as in humans.

The next step, he said, is making sure the CDC has in its laboratories blood specimens from "all outbreaks of pneumonia in Philadelphia at that time," and the fourth step is obtaining specimens from other persons who had penumonia in Philadelphia to see if the bacterium caused it.