A school playground in September, edged by a low wall and a large tree with lavender foliage. A group of child silhouettes stands there, long shadows stretching behind them. At the centre, two children in the palest lavender; the colour deepens figure by figure to the navy of those at the edge, and fine ripples spread outward from them.

Why does everyone get sick when school starts?

The short answer — What changes in September isn't the temperature but the web of contacts. Thousands of classrooms reassemble at once, giving viruses that circulated quietly all summer millions of new routes. The peak arrives about eighteen days later, after successive relays carry infection from child to child, then into homes.

Published: September 13, 2026
⏱: 4 min read

Reviewed & approved by: Nicolas

The calendar sets the signal

Mid-September. Three people in one household are blowing their noses. The eldest started, the younger one followed, a parent brings up the rear. Nobody “caught a cold” together. It’s a cascade: a virus entered through school, relayed from one person to the next.

The calendar confirms it. A Canadian study of asthma hospitalisations (1990 to 2002) places the peak among school-age children at 17.7 days after the start of the school year. Toddlers follow 1.7 days later. Among adults, it arrives roughly six days after the school-age peak.

What changes is the web of contacts

On the first day of school, the weather hasn’t shifted. What reconfigures is the contact network.

In summer, children scatter. Rhinovirus, responsible for half to two-thirds of our colds, still circulates: it tops summer sampling results. But it finds few routes.

On the first day back, classrooms reassemble everywhere at once. A European survey (7,290 participants) confirms it: students mix mainly with other students, and their contacts peak between ages 10 and 19. Yawning spreads through such crowds too, by imitation, with no pathogen involved.

Proof that school triggers the surge, not the season: school holidays, including those in midwinter, reduce the rate at which influenza is transmitted to children by 20 to 29%. This is French data. No effect was detected on adult contact patterns.

Why two and a half weeks

Two days per jump. That is the median incubation period of rhinovirus. It takes several jumps for one infected child to reach a whole classroom, then for that classroom to seed others, and for the peak to become visible at population scale. The relays stack up.

The same phrase comes up everywhere: “He caught a cold.” Wet hair, a forgotten jacket, cold feet. The cold is supposed to be the cause.

The explanation is incomplete. No virus is born from cold, and the peak follows the school calendar, in weather still mild. But cold erodes a local defence. In mouse cells, the antiviral response loses clear strength when temperature drops from 37 °C to 33 °C. In a separate experiment, volunteers exposed to 4.4 °C for fifteen minutes showed a drop of up to 6.4 °C at the front of the nose. A distinct protocol, on turbinate samples and cultured cells, showed that at 32 °C rather than 37 °C, secretion of antiviral vesicles falls by about 42%. Human trials diverge: no detectable effect in a 1968 trial where the virus was administered to volunteers around the time of cold exposure; in 2005, 13 colds out of 90 after foot cooling, versus 5 out of 90, without virological analysis. The effect remains local and partial, measured on defence indicators. What drives the back-to-school surge is the mass of contacts.

The child opens the door

The child comes home. Another network waits, smaller, tighter. One study followed 26 households for a year with weekly nasal swabs: children under 5 harbour a detectable virus far more often and for longer. Living with children adds roughly three weeks per year of nasal viral detection compared with a one-person household.

Adults arrive next, shifted by a few days. Within the household, measurable transmission passes through children and those with symptoms.

The season does not give the signal. The calendar does. The same virus circulates in July without producing a peak: nobody opens the door to it at the same time. The first day of school opens every door on the same morning.