In a recent letter from Dai, convergent-beam electron diffraction is used to determine the modulus of the Burgers vector of dislocations in icosahedral AlCuFe. In the analysis the phason strain component is neglected. We argue that this is an inadmissible simplification. We further show that a correct interpretation of the experimental data presented yields the result that the phason part of the dislocation strain field is larger than the phonon part, which contradicts the conclusions drawn by Dai.