Big cities and big states can't be blamed for slavering at the prospect of extra representatives in Congress and billions of extra dollars in federal funds.

So it's understandable that these units of government keep badgering the Census Bureau to revise its admittedly inaccurate nose-counts on which such allocations of money and political power are made.This week New York and other major cities won a big round in this long fight when the 2nd U.S. Court of Appeals ruled that the Bush administration failed to justify its decision not to adjust the 1990 Census to reflect 5 million people missed in the original count.

Though the ruling strikes a blow for greater accuracy in tabulating the nation's population every 10 years, it isn't likely to be the last word on this subject. Nor should it be. If the Clinton Justice Department doesn't appeal this case to the U.S. Supreme Court, smaller states and cities that would lose federal funds and congressional clout to their bigger neighbors can and should appeal.

When the final decision is made, it needs to temper the understandable desire for greater social justice through greater census accuracy with a healthy dose of reality.

One reality is the utter nonsense of the repeated claim that the 1990 undercount represents a plot on the part of the Republican Bush administration to deprive Democrats of political clout.

Yes, those missed in the 1990 census were largely poor urban minorities who normally vote the Demo- cratic ticket. But the simple fact is that such undercounts have been taking place for decades, during Democratic as well as GOP administrations. It's always hard to take an accurate census in urban areas with high rates of poverty, crime, drug abuse, homelessness, illiteracy and inadequate education. No court ruling can alter those sad facts of life.

Another reality is that any new adjustments in the census figures can be challenged and second-guessed just about as easily as the original count. The 1990 count, for example, tallied 248.7 million Americans. In 1991, the Census Bureau did a new calculation and decided that about 5 million people had not been counted the first time around. But another calculation in 1992 reduced the undercount to about 3 million, or about 1.5 percent of the population. The 1991 and 1992 figures were just estimates, albeit sophisticated ones. How far does the public really want the Census Bureau to wander into the foggy swamp of estimates as opposed to a flawed but hard count? With three different figures, which of them could not be challenged in court?

It's unrealistic to expect a completely accurate census count in a nation with as much individual liberty and unrestricted movement as there is in the United States. Even so, there is a way to improve the performance of the Census Bureau. This could be done by going back to basics - asking only such questions as name, sex, age and a few other details. That would speed the process, reduce costs, and still fulfill the basic purpose of the census - that of providing the numbers on which seats in the U.S. House of Representatives are allocated along with seats in state legislatures and some city councils.

For that to happen, social scientists would have to abandon the census as a tool for finding out such things about Americans as their ancestry, education, military service, disability, and transportation habits.

But with less prying into their personal lives, more Americans should be willing to fill out the questionnaires, giving the Census Bureau a more accurate count. To get such a change, the big cities and states will have to take their case to Congress rather than just to the courts.