Abstract The lapse rates of high-resolution temperature profiles during nearly neutral, saturated conditions are compared with the saturated adiabatic lapse rate and with that proposed by Richiardone and Giusti. A good agreement between the latter and the mean value of the observed lapse rate is found, whereas the saturated adiabatic lapse rate differs significantly, confirming experimentally that it is not completely correct to assess the moist neutrality from a comparison with the saturated adiabatic lapse rate. The lapse-rate distribution supports the hypothesis that the lapse-rate statistics is a local collection of saturated adiabatic lapse rates in a background normal distribution centered around the neutrality.