The 2000 Census missed more than 73,000 Marylanders, including almost 17,000 people in Prince George's County and more than 11,000 in Baltimore, according to adjusted population figures reluctantly released yesterday by the U.S. Census Bureau.
The numbers suggest that the decennial count, on which much government funding is based, overlooked more than 1.1 million children nationwide - half of them black or Hispanic.
The data were part of a nationwide set of adjusted population estimates, released under a federal court order, that indicated more than 3.2 million Americans - about 1.2 percent of the population - were missed by the census takers almost three years ago.
The Census Bureau disowned the data yesterday. It warned that the estimates are so seriously flawed that the bureau would take no responsibility for their accuracy "for any purpose whatsoever." Bureau staff say they won't help interpret the figures.
The bureau has said subsequent research has found the actual undercounts may have been as small as .06 percent, or fewer than 200,000 people nationwide.
Democratic lawmakers who sued for access to the adjusted figures, called their release a victory.
"The new census data is essential to understanding how errors in the census affect funding for cities, counties and critical programs funded with federal dollars," said Rep. Henry A. Waxman, a California Democrat.
About one-third of the 73,376 people missed by the count in Maryland were children and 47 percent - almost 34,000 - were black.
In Baltimore, the official count missed 11,527 people, or about 1.8 percent of the population, according to the new data.
"That doesn't surprise me," said Kristin Forsyth, spokeswoman for the Maryland Department of Planning. "Historically, cities have had a greater percentage undercount than other jurisdictions."
But it was Prince George's County, the state's only majority-black county, that showed the largest undercount in Maryland, with 16,927 people overlooked, or about 2.1 percent.
Montgomery County (11,119 people missed) and Baltimore County (9,264) were close behind Baltimore City.
It is difficult to calculate exactly how much money an increase in population might generate from population-driven aid formulas.
Baltimore officials have estimated that an additional 10,000 people could mean up to $5 million in additional aid. The true benefit, however, would depend on the money available and relative population changes elsewhere.
Mayor Martin O'Malley said yesterday, "We felt pretty strongly that that the official numbers from the  Census count undercounted us. We felt it hurt us in terms of the federal aid that should be coming to the city."
He expressed hope that the adjusted numbers would "promote debate in the halls of Congress as to how we arrive at a truer and fairer formula for federal funding."
Politicians, especially Democrats, have perennially squabbled over the accuracy of the census count. The numbers are used to allocate representation in the House of Representatives, to redraw federal, state and local legislative district lines, and to distribute $185 billion a year in federal aid.
Rep. William L. Clay, a Democrat from Missouri, said, "It was unfortunate we had to fight so hard to get this information. But now that we have, I hope we can use it to its fullest advantage."
States and cities frequently press the Commerce Department to release population numbers adjusted to reflect those missed by the official count. They argue that overlooking millions of poor people and immigrants - missed because they are transient, homeless or suspicious of census takers - shortchanges communities with large populations of both.
Democrats usually take the lead, calculating that they have more to gain by counting more poor people and immigrants. Republicans generally resist use of the adjusted figures, calling the statistical manipulations unreliable.
After the 2000 Census, the Census Bureau conducted research to estimate the undercount, using sample surveys and computer modeling. It found that as many as 3.2 million Americans were missed in the official count.
But the estimate was stillborn. Commerce Secretary Donald L. Evans decided last year not to report details of the adjusted numbers to the states and localities.
He said a panel of "senior career professionals" at the bureau had "serious reservations" about the accuracy of the results.
In response, 16 Democratic congressmen led by Waxman filed suit in federal court to force the Commerce Department to release the numbers. Oregon lawmakers filed a separate request for the numbers under the federal Freedom of Information Act.
Last month, the U.S. Court of Appeals ruled in favor of the Oregon lawmakers, and ordered the data released.
The Census Bureau did so yesterday, but with a disclaimer warning that research since the suits were filed has found that the 2001 adjustments "dramatically overstate the level of undercoverage in Census 2000."
The adjusted estimates were so severely flawed that "all potential uses of these data would be inappropriate," the bureau said.
The latest research indicates much smaller undercounts for virtually all race groups, and possible overcounts among whites and Asians.
The bureau would not release the new findings, however, saying that demographers are not finished with their research.
Sun staff writer Eric Siegel and the Associated Press contributed to this article.
Adjusted population figures released yesterday by the U.S. Census Bureau were consistently higher than the official counts. Here are the data for the state and the Baltimore area.
Unadjusted Adjusted Difference Pct. Diff MARYLAND 5,296,486 5,369,862 73,376 1.4 Baltimore City 651,154 662,681 11,527 1.8 Anne Arundel Co. 489,656 494,926 5,270 1.1 Baltimore Co. 754,292 763,556 9,264 1.2 Carroll Co. 150,897 152,060 1,163 0.8 Harford Co. 218,590 220,762 2,172 1.0 Howard Co. 247,842 250,547 2,705 1.1 Largest undercounts in the nation
California 509,012 1.5
Texas 364,032 1.7
New York 202,049 1.1
Florida 195,684 1.2
Georgia 119,852 1.5