Recount or not, it's unlikely Gov. Siegelman got enough votes to win last week's gubernatorial election.

But from the way I look at it, neither did U.S. Rep. Bob Riley.

Neither one of them received a majority of the vote, obviously the most definitive way to select a leader.

It also should be the American way.

States with open primaries, such as Louisiana, already elect their leaders in this manner, holding runoff elections in which the two top vote-getters in the primaries compete. Because there are only two in the runoff, the winner always gets a majority of the votes.

In last week's gubernatorial race, Siegelman and Riley each accounted for about 49 percent of the vote, with the Libertarian candidate, John Sophocleus, garnering the remaining 2 percent.

The situation is similar to the 2000 presidential election in which George W. Bush and Al Gore each picked up about 48 percent of the vote (Gore getting about a half million more than Bush), with independent candidates, largely the Green Party's Ralph Nader, getting the remaining 4 percent.

In both cases, a runoff election would have avoided a lot of turmoil, and have ensured the choice of a majority of voters for these important positions.

The traditional argument against runoffs is their cost. Not only do governments have to foot the bills for new elections, they also force the runoff candidates to spend more money on their extended campaigns, not a happy topic amid current calls for campaign finance reform.

The ideal way would be to hold runoff elections at the same time as the general elections, and the technology exists to do that right now.

It's called IRV, or instant runoff voting, and it's used to decide elections in some foreign countries, including Australia and the Republic of Ireland.

It's also being examined by a growing number of states and municipalities interested in improving the election process. (San Francisco, for example, will implement the system next year.)

Under the system, voters choose not just their preferred candidates, but second and third choices (and maybe even more) for the elected post up for grabs.

If no candidate receives a majority of first-place votes, the candidate with the least amount of first-place votes is eliminated. The second-place votes of the voters for this candidate are then counted as first-place votes for the remaining candidates. The process is repeated until one candidate gets a majority and is declared the winner.

In the case of last week's gubernatorial election, the more than 23,000 votes for the third-place Libertarian candidate, Sophocleus, would have been redistributed to Riley and Siegelman according to the second-place preferences of the voters who cast them.

Who would have been the winner?

Who knows?

It's a safe bet, though, that Professor Sophocleus' party would have done far better under an IRV system, perhaps even preserving the Libertarians' "major party" status in Alabama by garnering at least 20 percent of the vote in at least one statewide election.

That's because voters are more apt to support underdog candidates in primaries if they know they'll have another chance to vote in the runoff if their candidate doesn't make it.

The IRV system instantly provides that. And by doing so, it encourages more parties and candidates (and presumably more voters) to get involved in the electoral process.

Most importantly, though, it ensures that winning candidates have the support of the majority of the voters.

And that should be the American way.

Mitch Chase is a DAILY copy editor. For more information on instant runoff voting, check the Web site of the Center for Voting and Democracy (chaired by former U.S. presidential candidate John B. Anderson) at www.fairvote.org.