OREM — A programming mistake caused punch-card reading machines in Utah County to not count 33,000 straight-party votes cast on Nov. 2.

The machines added up the total number of ballots but didn't forward votes to the candidate columns.

"The initial counter didn't carry the straight-party votes into the individual races," said Neil Peterson, director of information systems for Utah County.

The error initially was brought to the attention of State Elections Director Amy Nacarrato by the Deseret Morning News, which first reported the abnormal totals on Nov. 4. At the time, the unofficial computer tally showed that 144,423 Utah County residents cast ballots, but that only 110,143 voted in the presidential race.

Under-votes are normal since many voters choose only to vote in some, but not all, races. But they are rarely as pronounced in a presidential race as appeared to be the case before the recount in Utah County.

"It didn't make sense because almost everybody votes for president," Utah County Clerk/Auditor Kim Jackson said Saturday.

The day after the election, Nacarrato speculated that Utah County voters might have been making a statement they were unhappy with President Bush without voting for Sen. John Kerry.

Instead, after Utah County officials re-ran the punch cards Wednesday, the results showed 145,769 ballots were cast in Utah County with 143,797 including a presidential vote. President Bush still won 86 percent, but Kerry edged past 11.5 percent, up from 11 percent.

Utah County Democratic Party Chairman Vaughn Cook was incredulous when he learned about the blunder.

"Thirty-three thousand votes?" Cook asked. "That's something we'll have to pay attention to as the Democratic Party . . . strives to create an environment where there is more political balance in Utah County. Subsequent elections could be a lot tighter, and 33,000 votes would be much more significant to us."

In the end, the recount did not change the outcome of any of the races in the conservative county where 83 percent of the 50,992 who chose the straight-party option voted the Republican ticket.

"There was no conspiracy theory," Peterson said. "If anything, when we corrected the error, it made the winner's spread bigger in nearly every case."

The mystery behind the under-vote further caught county officials' attention after a member of the Personal Choice Party called to ask why about 1,900 people had voted a straight-party ballot for Presidential Choice Party but only about 70 votes were listed for that party's presidential ticket.

It turned out only one character was wrong in the programming code.

"It was a simple problem, but a little tricky to find," Peterson said. "The readers worked fine, everything worked like it was supposed to. It was a setup mistake by our programming team that sets up the election."