The federal government's reworking of the naturalization exam is an improvement. Any change aimed at testing prospective citizens on knowledge of civics and history rather than on their rote memorization skills ("How many stripes are on the flag?") deserves support.

However, the Bureau of Citizenship and Immigration Services has used a computer-driven randomization program to select the cities where the new test will be piloted. The results of that selection process are a little baffling and suggest that certain national groups — including those from Asian countries — could be underrepresented in the pilot program.

The 10 cities reportedly chosen to tap different geographic parts of the country and cities with immigrant populations of different sizes, are: Albany, N.Y.; Boston; Charleston, S.C.; Denver; El Paso, Texas; Kansas City, Mo; Miami; San Antonio; Tucson, Ariz.; and Yakima, Wash.

If geographic diversity was a factor, it's odd that two cities are in Texas. And in most of those cities, according to census data, Hispanic immigrants are the largest group.

Immigration officials say that Asians will be adequately represented because they apply for naturalization at a higher rate than other immigrant groups. We can hope they're right about this. In any event, the evaluators of this new test should watch the results and make sure they reflect national immigration trends, in which Asians figure prominently.

Otherwise, the revamping is something to applaud. All 144 questions can be read online (www.uscis.gov). But in general, everyone can agree that it's smarter that our new citizens be able to "name one right or freedom from the First Amendment" than simply recall the number of stripes on a flag.