Annotation: The source characterisation of an ear canal probe is commonly carried out by measurements on two or more distinct loads (calibration cavities) of known admittances. In this study, we determine the source characteristics of an Etymotic Research ear canal probe and investigate the influence of the number of the calibration cavities used. The uncertainty of these source characteristics was analysed by applying this very probe for measuring the admittance of a test cavity that was not involved in the characterisation procedure. When two calibration cavities had been employed, deviations up to 2.6 dB were noted between the measured and the theoretical admittance values of the test cavity. However, this error varied for different calibration cavity combinations and decreased to 0.3 dB in the case of four calibration cavities. Monte Carlo simulations were performed in an attempt to investigate the sensitivity of the output quantities of the source characterisation process to its input quantities (acoustic length, frequency resolution, measured sound pressure). The effect of uncertainties attributed to these input quantities proved to have no significant effect on the uncertainty of the source characteristics and, therefore, on the test cavity admittance measured by means of the probe. It was found that this error mainly depended on the choice of the calibration cavity lengths. When two sufficiently different calibration cavities were used (i.e. with a length difference greater than 4 mm), an accuracy similar to that when four cavities were used, was obtained. However, less care had to be taken on the choice of the calibration cavity lengths in the case of four cavities.