The number is freighted with loads of political baggage and can fluctuate wildly -- tripling one minute, halving the next -- depending on who's talking.

U.S. Capitol Police suggested yesterday's antiwar street march drew 30,000 to 50,000 people. Protest organizers said that the number was closer to 500,000. District police settled on "an awful lot of people."

The truth might fall somewhere in between the guesses, or it might fall somewhere beyond the edges. That's because no one really knows how many people showed up.

The methods used to determine head counts generally rest on rough comparisons to crowd estimates attributed to previous large-scale events. Those historical attributions, however, often resulted from ballpark guesses themselves.

"I know everyone is skittish about saying a number," said U.S. Capitol Police Chief Terrance W. Gainer. "But this was big. An impressive number."

Why is it important to demand specificity, to insist on a hard number? Many of those involved in the rallies say numbers are important because they're often used -- rightly or wrongly -- to convey the strength of a political movement. Because organizers often see their movements as being at odds with the government's way of thinking, they don't trust public officials to do the counting.

"The police always low-ball it," said Tony Murphy, spokesman for International ANSWER, the coalition that organized yesterday's demonstration. "They always give drastically low numbers, so it's important for us to make sure realistic numbers get out."

Early yesterday afternoon, D.C. Police Chief Charles H. Ramsey said the crowd count was comparable to the antiwar rally held in Washington on Oct. 26. In October, Ramsey suggested that that protest seemed somewhat larger than the April 20 rally to support the Palestinian people and oppose U.S. aid to Israel. In April, he had eyeballed the pro-Palestinian crowd as numbering about 75,000.

So did that mean his estimate yesterday was "about 100,000"?

Not really. Later in the day, Ramsey reconsidered and opined that the crowd was probably larger than October's, but he wouldn't commit to a number: "I wouldn't argue with the people who say that there's an awful lot of people out there," he said.

Protest organizers defined "an awful lot" as a half-million.

Immediately after the Oct. 26 protest, International ANSWER: (Act Now to Stop War and End Racism) seemed content with the 100,000 ballpark figure attributed to Ramsey. But later, as organizers conferred with other observers, they settled on 200,000.

Murphy explained that revision. He said organizers floated the 100,000 figure to a member of the media, someone from Pacifica Radio, a left-leaning network that covered the event. The Pacifica reporter offered a different perspective of that event.

Murphy said: "We were saying 100,000 and they [Pacifica Radio] said, 'Oh, come on -- it was a lot more than that. It was 200,000, at least.' "

And so, when coming up with an estimate for yesterday's crowd, the organizers stood on the stage near Third Street NW and compared the image of the densely packed crowd in front of them to their memory of the one in October.

"Everyone said this one was definitely twice as big and felt it was a little more than that," Murphy said. "It was very obviously bigger. Everyone had the same feeling about that."

Gainer agreed that yesterday's crowd seemed larger than October's, but -- in keeping with the disparity of base lines -- the Capitol police chief used much lower numbers.

"I've heard estimates anywhere from 30,000 to 50,000 from people who've seen a lot of these things," Gainer said. "I talked with some people in the crowd who said they felt October's was bigger. I didn't think that. I think this one was much larger."

For 30 years, measuring crowds rested on the shoulders of the U.S. Park Police, which divided the Mall into sections, used aerial photographs to determine the density of the crowd and issued a head count. The numbers often were disputed. After officials estimated that the Million Man March in 1995 was more like a 400,000-man march, Nation of Islam leader Louis Farrakhan threatened to sue. Congress later mandated that the Park Police, an arm of the National Park Service, bow out of the counting business.

"It was to prevent us from being sued," said Bill Line, spokesman for the Park Service. "Any political issue you can think of would have protests, and they'd demand that the Park Service count."

For those in the middle of a rally, a crowd of 20,000 can look just like a crowd of 2 million -- there are people crushing in on every side, as far as the eye can see.

Leslie Miller, a war protester from Chestertown, Md., said yesterday she chooses to believe the estimate given by the protesters, not the police. She also said she believes that the larger the number of protesters reported, the more impact it might have on observers with a vote in Congress.

"We're not heard any other way," Miller said.

Sera Morgan carried a sign that on one side read, "Count Me As A. . .," and on the flip side, "Federal Employee Against War." But when asked about crowd estimates, she reconsidered. Maybe people shouldn't worry so much about counting her, she said, but instead should listen to what she's trying to say.

"I wish we wouldn't get so hung up on numbers," said Morgan, of Takoma Park. "It's false to assume that if you have a million people on the Mall you're somehow more right than you are if you have 200,000 people on the Mall. The message doesn't acquire greater value."

Staff writer David A. Fahrenthold contributed to this report.

Bundled up even though they're inside, Shelley Finkelstein, left, her son Ben Finkelstein and Erika Wilson lend their support to passing demonstrators.Union members from New York were among the protesters. One demonstrator used her cheek to express her sentiments.A makeshift peace dove is right at home. "The police always low-ball" crowd estimates, said Tony Murphy, spokesman for International ANSWER.