It takes a lot of energy to grow a baby — just ask anyone who has been pregnant. But scientists are only now discovering just how much.

In a study published on Thursday in the journal Science, Australian researchers estimated that a human pregnancy demands almost 50,000 dietary calories over the course of nine months. That’s the equivalent of about 50 pints of Ben and Jerry’s Cherry Garcia ice cream, and significantly more than the researchers expected.

Previous estimates were lower because scientists generally assumed that most of the energy involved in reproduction wound up stored in the fetus, which is relatively small.

But Dustin Marshall, an evolutionary biologist at Monash University, and his students have discovered that the energy stored in a human baby’s tissues accounts for only about 4 percent of the total energy costs of pregnancy. The other 96 percent is extra fuel required by a woman’s own body.

“The baby itself becomes a rounding error,” Dr. Marshall said. “It took us a while to wrap our heads around that.”

This discovery emerged from Dr. Marshall’s long-running research on metabolism. Different species have to meet different demands for energy. Warm-blooded mammals, for example, can maintain a steady body temperature and stay active even when the temperature drops.

But being warm-blooded also has drawbacks. Maintaining a high metabolic rate requires mammals to constantly feed the furnace. A coldblooded snake, in contrast, can go weeks between meals.

Dr. Marshall set out to compile a complete inventory of the energy consumed by dozens of species over the course of their lives. He recognized that most females must not only fuel their own bodies, but must also put additional energy into their offspring.

When Dr. Marshall began looking into the costs of reproduction, he couldn’t find solid numbers. Some researchers had guessed that indirect costs — that is, the energy females use to fuel their own bodies while pregnant — might come to only 20 percent of the direct energy in the baby’s tissues. But Dr. Marshall didn’t trust their speculation.

He and his students set out to estimate the costs for themselves. They scoured the scientific literature for information such as the energy stored in each offspring’s tissues. They also looked for the overall metabolic rate of females while they were reproducing, which scientists can estimate by measuring how much oxygen the mothers consume.

“Folks were just poodling along, collecting their data on their species, but no one was putting it together,” Dr. Marshall said.

By aggregating such data, the researchers estimated the costs of reproduction for 81 species, from insects to snakes to goats.

They found that the size of an animal has a big influence on how much energy it needs to reproduce. Microscopic animals called rotifers, for example, require less than a millionth of a calorie to make one offspring. By contrast, a white-tailed deer doe needs more than 112,000 calories to produce a fawn.

The metabolism of a species also plays a part. Warm-blooded mammals use three times the energy that reptiles and other coldblooded animals of the same size do.

The biggest surprise came when Dr. Marshall and his students found that in many species, the indirect costs of pregnancy were greater than the direct ones.

The most extreme results came from mammals. On average, only 10 percent of the energy a female mammal used during pregnancy went into its offspring.

“It shocked me,” Dr. Marshall said. “We went back to the sources many times because it seemed astonishingly high based on the expectation from theory.”

David Reznick, an evolutionary biologist at the University of California, Riverside, who was not involved in the study, was also startled at how high the indirect cost could get. “I wouldn’t have guessed that,” he said.

And yet what surprised him even more was that Dr. Marshall’s team was the first to pin down these numbers. “It is disarming,” he said. “You think, someone has done this before.”

The study offers clues about why some species have higher indirect costs than others. Snakes that lay eggs use much less indirect energy than snakes that give birth to live young. The live-bearing snakes have to support embryos as they grow inside their bodies, whereas egg-laying mothers can get their offspring out of their bodies faster.

There may be a number of reasons that mammals pay such high indirect costs for being pregnant. Many species build a placenta to transfer nutrients to their embryos, for example. Dr. Marshall suspects that humans pay a particularly high cost because women stay pregnant longer than most other mammals do.

Dr. Marshall said that the new results may also explain why female mammals put so much effort into caring for their young after they’re born: because they put in so much effort during pregnancy.

“They’ve already got massive sunk costs in the project,” Dr. Marshall said.