If you drop food on the floor and pick it up within 5 seconds, it's still safe to eat.

Food science students at the University of Illinois had a long-standing piece of conventional wisdom to work with in 2003 when an undergraduate named Jillian Clarke investigated the claim during a research internship. The five-second rule, the informal guarantee that food dropped on the floor and retrieved quickly remained uncontaminated by bacteria, was already old enough to have acquired the texture of established fact. Clarke's investigation was the first formal laboratory test of the claim. She applied Escherichia coli bacteria to floor tiles, dropped gummy bears and cookies onto them, and measured the bacterial transfer. The organisms moved to the food immediately. Contact time under five seconds did not prevent contamination. The finding earned Clarke an Ig Nobel Prize and a brief round of press coverage, and then the five-second rule continued circulating without significant disruption.

The deeper investigation came thirteen years later, when Donald Schaffner, a professor of food science at Rutgers University, designed a study systematic enough to generate publishable data. Schaffner and his team were interested not just in whether bacteria transferred, but in what variables governed how much. They introduced Enterobacter aerogenes, a non-pathogenic relative of Salmonella used as a safe surrogate in food safety research, to four surfaces: stainless steel, ceramic tile, wood, and carpet. They then dropped four food types, watermelon, white bread, buttered bread, and gummy candy, onto those surfaces at four different contact intervals: less than one second, five seconds, thirty seconds, and three hundred seconds. The experiment generated 2,560 measurements.

The study, published in Applied and Environmental Microbiology in 2016, found that bacteria transferred in every trial at every contact time, including the sub-one-second interval. Contact time did matter, longer exposure produced higher bacterial transfer, but the relationship between time and contamination was not a step function with a clean threshold at five seconds. It was a continuous gradient. The five-second cutoff was not a point at which contamination began; it was an arbitrary number on a curve that started at zero seconds.

Moisture was the more important variable. Watermelon, with its high water content, accumulated far more bacteria than gummy candy in every trial. Bacteria travel with moisture, and the wetter the food, the more readily the contact between surface and food transfers organisms along. Carpet, counterintuitively, transferred bacteria at much lower rates than tile or stainless steel, a function of the reduced contact area between the food's surface and the uneven carpet fibers.

Schaffner was careful not to overstate the practical risk: most floors are not heavily contaminated with dangerous pathogens, and the bacterial loads transferred in typical dropped-food scenarios would not cause illness in healthy adults. The risk of eating a piece of candy off a tile floor is low. What the data could not support was the claim that there was a safe window defined by time, that five seconds provide meaningful protection. The rule is less a biological principle than a social consolation, retold in the language of science. Bacterial transfer begins at the moment of contact.