Monday night, Jon Stewart laid into coverage of the BP oil spill in the Gulf of Mexico, noting the disaster had dealt “a potentially crushing blow to our nation’s describing-the-size-of-things industry.” He rolled a series of news clips that variously compared the size of the resulting slick to Maryland, Jamaica and Puerto Rico, among other not-particularly-helpful geographic entities.

The satire hit on a weakness for sloppy comparisons based on suspect estimates that has plagued coverage of the spill since the story switched from a firefighting, rescue effort to a sunken rig and a tar-balling environmental concern.

The day after the Deepwater Horizon platform sank and the search for survivors ended (eleven were killed), estimates of an oil leak ranged from nonexistent—the Coast Guard’s first guess on April 23—to a potential 336,000 gallons per day, which is what the exploratory rig was pumping before the explosion. By April 25, after further reconnaissance by remotely controlled submersibles and sonar, the Coast Guard and BP reassessed, settling on an estimate of 42,000 gallons per day.

“This is a very serious spill,” the Coast Guard’s Rear Adm. Mary Landry told the New Orleans Times-Picayune. “This has the potential to be a major spill.”

The warning drew moderate concern from media outlets, but was mitigated by the underwhelming and oft-mentioned comparison to the 1989 spill of the Exxon Valdez, which dumped eleven million gallons of oil into Alaska’s Prince William Sound. With hope that the submersibles in the Gulf might succeed in shutting the failed blowout preventer at the wellhead on the ocean floor, the 42,000-gallon-per-day estimate allowed reporters to dust off their calculators and provide a back-of-the-envelope figure that would ease fears and put the leak in perspective. A New York Times story from April 26 that ran on page A11 was typical of the national media coverage, reporting that, “At the rate of 42,000 gallons of oil a day, the leak would have to continue for 262 days to match” the Valdez, the worst spill in United States history. In other words: no need to lose too much sleep over this, yet.

Although some outlets began reporting growing anxiety over the oil slick’s threat to the fragile wetlands and fisheries along the Gulf Coast, its potential economic impact, and implications for future offshore drilling, the spill did not become the story until April 28, when the Coast Guard announced that it had underestimated the leak by a factor of five, and that the well was actually gushing 210,000 gallons a day. Burned by their uncritical acceptance of the previous estimate, reporters finally started to ask how BP and the Coast Guard were coming up with these numbers.

“The leaks on the sea floor are being visually gauged from the video feed” from the remote submersibles, explained Doug Helton, a fisheries biologist with the National Oceanic and Atmospheric Administration, in an e-mail to The New York Times. “That takes a practiced eye. Like being able to look at a garden hose and judge how many gallons a minute are being discharged. The surface approach is to measure the area of the slick, the percent cover, and then estimate the thickness based on some rough color codes.”

So, gauging the rate of spill is a very inexact science. Coast Guard and BP officials have been willing to admit that, but reporters overlooked it in their hurry to report an exact figure.

The quintupling of the estimate eight days after the rig exploded changed that, and captured the full attention of the national media. With no quick solution to cutting off the leak in sight, reporters recalculated that, in fact, the disaster on the Transocean rig could very well prove to be the worst oil spill in American history, with approximately fifty-two days and counting to a new record. That is still a guesstimate, and whether or not it will stick remains to be seen. Nonetheless, reporters are in fact now asking tougher questions.

Brett Norman is a reporter for Politico.