Nobody knows exactly how it happened, but somehow, between February 6, 1898, and February 7, 1898, the issue numbering for The New York Times got a little ... off.
It's easy enough to imagine the scene: A worker, late at night, setting the paper's front-page type. He takes out the type from the preceding day's paper. He looks at the issue number—14,499—and adds one. He gets 15,000.
Perhaps he misread the number, and thought he saw 14,999 in its place. Or perhaps he'd had a long night, and just wasn't thinking straight. Who knows? What we know is that he put 15,000 as the issue number for the next day, and nobody noticed.
And nobody noticed the next day, nor the next day, nor the next.
The error came to light recently when Aaron Donovan, a news assistant, became curious about the numbering, which he updates nightly when working at the news desk. He wondered about the potential for self-perpetuating error. Using a spreadsheet program, he calculated the number of days since The Times's founding, on Sept. 18, 1851.
Through the newspaper's archives, he learned that in its first 500 weeks, The Times published no Sunday issue. Then, for 2,296 weeks from April 1861 to April 1905, the Sunday issue was treated as an extension of the Saturday paper, bearing its number. In the early days, the paper skipped publication on a few holidays. No issues were published for 88 days during a strike in 1978. (During five earlier labor disputes, unpublished issues were assigned numbers, sometimes because catch-up editions were later produced for the archives.)
Finally, by scanning books of historic front pages and reels of microfilm, Mr. Donovan zeroed in on the date of the 500-issue gap.
“There is something that appeals to me about the way the issue number marks the passage of time across decades and centuries,” the 24-year-old Donovan wrote in a memo. “It has been steadily climbing for longer than anyone who has ever glanced at it has been alive.”