Pierre Comizzoli, research biologist at the National Zoo was part of the team that artificially inseminated female panda, Mei Xiang with sperm from male panda, Tian Tian. (Ricky Carioti/WASHINGTON POST)

Update: The National Zoo panda died just hours after this story was published.

Nitrogen vapor drifts from the mouth of steel container No. 9 when National Zoo scientist Pierre Comizzoli lifts the plastic lid and pulls out the stopper.

The vapor is potentially deadly, but the liquid nitrogen within keeps the microscopic seeds of life inside safely frozen at about minus-300 degrees Fahrenheit.

Here, he says, as he raises a specimen rack from the vapor, is the “famous,” but now depleted, 2005 vintage that helped bring forth the zoo’s new giant panda cub on Sept. 16.

It is panda semen. Unromantic and a little icky, it’s the stuff behind the wondrous birth of the zoo’s first cub in seven years.

But zoo experts aren’t sure exactly why it worked. And Comizzoli and other zoo experts are reviewing the science and technique behind their success to try and figure it out.

The achievement was stunning — in the face of repeated failures in the past and with the odds of success at about 10 percent.

The zoo said Friday that the cub appears to be well. The female giant panda, Mei Xiang, had not left the newborn, even to eat or drink.

The birth, at 10:46 p.m. Sept. 16, came after a meticulous choreography over six months that tracked the rise and fall of Mei’s hormone levels and utilized state-of-the-art reproduction techniques.

But it was a process the zoo had been through often before, in almost the same way. So, why was it successful this time?

Brandie Smith, curator of giant pandas, said the zoo tweaked some of its procedures this year — limiting, for example, the amount of artificial light, noise and hubbub in the panda house during Mei’s reproductive cycle.

“This year was really a make-or-break year in terms of getting Mei Xiang pregnant,” Smith said Friday. The zoo had said that if Mei, who had not borne a cub in seven years, did not get pregnant this time, she should be replaced.

“So we were extra cautious,” Smith said. “During the end of the pregnancy, we became incredibly hands off. . . . We minimized disturbances. We stayed out of her den. . . . Once she started to build her nest we did not go in that area.”

“All of that said, are those changes the reason she got pregnant this year?” she said. “We don’t know.”

Attention was also focused on the artificial insemination process the zoo had used, to no avail, over the past five years.

Much of that focus was on the contents of Room 222 in the zoo’s Veterinary Hospital.

There, the zoo has more than 2,000 frozen semen samples from 100 different kinds of animals, some dating back decades, as part of its Genome Resource Bank.

Comizzoli, a lanky, gray-haired reproductive physiologist who was born in Paris, said frozen semen can, for the most part, be used effectively in insemination procedures to impregnate the zoo’s animals.

But, “some of the sperm are not going to survive the freezing temperature,” he said.

So-called cryosensitivity varies from species to species. But giant panda specimens survive quite well, he said: “We don’t know why.”

When Mei Xiang entered her annual estrous period last spring, the zoo decided to artificially inseminate her with frozen specimens taken from the zoo’s male giant panda, Tian Tian, in 2005.

That was the year that Mei Xiang gave birth to her only previous cub, Tai Shan, after being artificially inseminated with semen from Tian Tian.

“2005 was a good year,” Comizzoli said in an interview at the zoo on Wednesday, because the samples were of high quality.

But that vintage was tried in previous years with no success. In 2007, the zoo used specimens from a male named Gao Gao who was a proven breeder at the San Diego Zoo. But that didn’t work either.

Mei Xiang was not inseminated in 2006 because she was still nursing Tai Shan and did not ovulate, Comizzoli said.

Fresh specimens from Tian Tian have also been tried, and even a mix of fresh and frozen. Still, no cub.

The specimens are taken from Tian Tian when he is anesthetized and stimulated electrically. The semen is placed in five-inch-long clear plastic tubes, frozen and then thawed when needed.

Another factor in play may be the time of year that Mei Xiang goes into heat and can become pregnant.

Pandas normally go into heat from March to May, Comizzoli said. But in 2009, 2010 and 2011, Mei Xiang did so in January, which may have contributed to her failure to have a cub.

This year, she went into heat in April. In 2005, when she had Tai Shan, she went into heat in March.

But she also went into heat in April and March in the no-cub years of ’07 and ’08.

There are many problems that can upset the cosmic rendezvous of panda sperm and egg, Comizzoli said. “You have so many different factors, and it’s impossible for us to know what happened.”

Meanwhile, he said, he still has “good quality” frozen specimens from Tian Tian taken in 2003 and 2004, which have never been used, and poor quality material from 2007.

After so much use, there is probably not enough of the ’05 material left for another insemination, he said.

Comizzoli said the bank has specimens taken as far back as the 1980s from the zoo’s giant panda, Hsing Hsing, who died in 1999.

“Technically . . . we could use it,” he said. It is stable and well preserved in the liquid nitrogen.

Hsing Hsing and his mate, Ling Ling, who died in 1992, produced four live cubs, but they all died shortly after birth.

As for Mei’s miracle of life, it’s anyone’s guess what went right.

“Was it the timing?” asked Smith, the curator. “Or was it the conditions during the pregnancy?”

“We don’t know what made it special,” she said.

Maybe it wasn’t even science:

“Maybe it was something that told her body, ‘Hey, conditions are right. Go ahead and . . . get pregnant.’ ”