The relationship between mean daily precipitation and elevation is often regarded as linear and positive, resulting in simple "precipitation lapse rate" equations frequently employed to extrapolate daily rainfall from a single weather station over a large area. We examine the precipitation-elevation relationship in the Swiss Alps using a combination of weather radar and rain gauge data to test this common assumption, challenging it by fitting a two-segment piecewise linear model with a mid-slope break-point as an alternative. By examining data stratified by catchment, season, and weather type, we assess the space-time variability of the precipitation-elevation relationship. We conclude that a non-linear and non-stationary model seems necessary to capture the variability of the observed precipitationelevation relationship. Based on our findings, we suggest that the simplified precipitation lapse rate concept is misleading and should be reconsidered in hydrological applications, emphasizing the need for a more realistic representation of precipitation variability over time and space.