WHILE ARGUMENTS AND court cases continue over whether the Census Bureau undercounted the citizenry of various metropolises, scant attention is being given to an equally potent issue: how the numbers are used to allocate political power in Congress.

To put the matter bluntly, the current method of apportioning House seats has cheated the larger states, given undue representation to the smaller ones, and violated both the Supreme Court's one-man, one-vote rule and the intent of the Founding Fathers.

If the method is not changed now, it is almost certain to swell the seats lost by larger states in the Northeast and Midwest, and give an extra advantage to smaller Sun Belt and Western states beyond that warranted by population shifts.

Indeed, of the estimated 14 House seats that may be transferred from urban industrial states to rural agricultural ones after this year's count, six or more -- nearly half the potential switch -- could conceivably result just from the mistaken apportionment formula. Although it is difficult to tell for sure, because almost all the "lossing" states have a preponderance of Democratic representatives, this might well help Republican House prospects in the 1982 election and beyond.

The problem with the apportionment formula, which has vexed American politics for nearly 200 years, is what to do with the fractions that result when the population is divided by congressional seats. The Constitution says, "Representation . . . shall be apportioned among the several States . . . according to their respective numbers." But because no state can have one-quarter or three-quarters of a House seat, things don't work out that precisely.

In 1970, for just one example, Oregon's exact share of the 435 House seats was 4,500. Should it have gotten four or five seats? If the number were rounded up to five seats, giving its citizens a larger voice than they deserved, another state -- say, one with a 4,450 share -- might get only four seats and therefore less representation than it deserved. What is the fairest way to deal with this?

This is no small question. The results of different formulas on Congress' composition can be dramatic. This can be seen by comparing the first formula used, one advanced by Thomas Jefferson, with another backed later by John Quincy Adams, and applying them to the Census Bureau's estimated U.S. population in 1979. The Jefferson method would give California 47 seats, while the Adams method would give California only 43. Overall, the results would differ in 27 states and involve 34 seats.

Not surprisingly, then, American history is filled with bitter conflicts over this issue, with all sides seeking advantage. When Jefferson proposed his formula after the first Census in 1792, he got into his first head-on clash with Alexander Hamilton. Jefferson's method triumphed and gave his state of Virginia 19 House seats, rather than the 18 it would have received under the Hamilton method.

By 1830, it has become apparent that Jefferson's method decidedly favored larger states, and former president John Quincy Adams launched an attack. But Adams formula, which would have saved his state of Massachusetts one seat, was never used. Adams' crusade was then taken up by the famous constitutional lawyer Sen. Daniel Webster of Massachusetts, who advanced what is the most natural and fairest formula -- and the one most used in Anmerican history.

Webster wanted first to compute the exact share of seats for each state. For example, by the 1979 estimates, New Jersey's would be 14,534 and New Mexico's 2,461. The natural impulse is to "round" these in the usual way, giving New Jersey 15 and New Mexico 2. Unfortunately, though, the result may not add up to 435. For example, what if most or all states had remainders greater than one-half?

Webster proposed a simple solution. First, calculate the ideal number of persons per congressional district, meaning the total population divided equally into 435 districts. Then let the population of each state be divided by this divisor and, "in addition to the number of members resulting from such division, a member shall be allowed to each State whose fraction exceeds a moiety [a half] of the divisor."

An ideal congressional district for the 1979 estimates is about 504,500 persons. To obtain a Webster apportionment, divide this into the population of each state and round the results to the closest whole number. If more than 435 seats are allotted, then the divisor should be adjusted upward, if less than 435 seats it should be adjusted downward, until exactly 435 seats are apportioned.

Webster's method prevailed in 1840, and from 1880 until it was upset in 1941. That was when political appetite prompted Congress to adopt a flawed formula first proposed in 1911 by Joseph Hill of the Census Bureau. Contrary to a report of the National Academy of Sciences recommending Hill's formula, it is notably biased in favor of the smaller states. Unfortunately, it is still in use today.

Hill's method works the same way as Webster's -- until you get to Hill's peculiar way of rounding fractions. Instead of taking the midpoint of fractions as the threshold for rounding-up, Hill takes the square root of the product of the whole number and the next highest one as the threshold. For example, a state with quotient 2.450 would receive three seats because 2.450 is greater than the sqaure root of 2 times 3.

Hill's method unquestionably discriminates against larger states. One needs only compare each person's representation in small states with each person's representation in large ones under the two methods. Simply split the states into three groups -- large, medium and small -- and divide the number of seats given to the large states by their total population, and then compare this with the similar number for the small states. Using the 1979 population estimates, Hill's method gives each person 4 percent greater representation in the small states than in the large ones.

By contrast, Webster's method is almost perfectly evenhanded in its treatment of large and small. The difference can be seen by looking at the effects of the formulas on two of the large and two of the small states: (TABLE) State(COLUMN)Exact share(COLUMN)Hill(COLUMN)Webster Texas (COLUMN)26,523 (COLUMN)26 (COLUMN)27 New Jersey (COLUMN)14,5345 (COLUMN)14 (COLUMN)15 Colorado (COLUMN)5,496 (COLUMN)6 (COLUMN)5 New Mexico (COLUMN)2,461 (COLUMN)3 (COLUMN)2 (END TABLE) When Hill's method is applied to all 19 censuses in American history, the average bias in favor of the small states is 3.4 percent. When Webster's method is applied, the bias is only 0.3 percent for the small states. This translates into a consistent difference of several seats, often a matter of vital political significance.

The Hill method not only violates the Founding Fathers' intent; it also violates the Supreme Court's redistricting standards. The High Court has held in redistricting cases that a good-faith effort must be made to achieve precise mathematical quality -- and it has rejected congressional district plans having a worst deviation from the ideal of as little as 2.4 percent. The Hill method's deviation is clearly worse than that.

The 1980 census figures are not yet known -- and, of course, are under serious challenge in some major cities -- so the precise outcome is impossible to predict. But the mathematical odds are about 4 to 1 that if Hill's method is used again, the small states will be favored. The time has come to restore Webster's formula as the only acceptable method of apportionment that meets the constitutional standard of one man, one vote.