Radiosonde observations can be useful to assess temperature and humidity trends occurring in the upper levels of the atmosphere. Studies of upper-air temperature trends based on radiosonde observations have shown a warming of the troposphere while a cooling was observed in the stratosphere since the 1950s. However, radiosonde instruments and observing practices have changed during the past 50 years, and it became more evident that these changes could have affected the climate trends observed in the atmosphere. This work presents a preliminary assessment of the potential biases in the radiosonde temperature and humidity time series in Canada. It attempts to identify the source of biases and to quantify their magnitude.