type seen in Au22Mn6 (Hiraga et aL, 1982). In the present case, there is a sliding offset of Mn-column positions as indicated, for example, by the arrows in Fig. 7(b) and the width of Dla along the long-period axis is different. Careful observation of these images enables a schematic model for this lattice 'side-step' to be obtained, as drawn in Fig. 7(c) where the difference in brightness of spots is indicated by the lined, half-filled or solid circles (and the regions of D la type are shown in screen-tone). The structure model of Au22Mn6 is also shown in Fig. 7(d) for comparison. Some apparent nearest-neighbour MnMn pairs, the brightness of which is lower than that of other parts, are visible at the side-step. However, no nearest-neighbour Mn-Mn pairs are known to exist in this alloy. Moreover, the recorded image comes from the projection of the structure along the incident-beam direction. Thus, it can be concluded that the occupation probability of the Mn atoms in the columns is less than unity in order to avoid true nearest-neighbour Mn-Mn pairs, and this causes the lower contrast of the pairs, as discussed for the 'superstructure' image condition (Shindo, Hiraga & Hirabayashi, 1984). It is not possible to distinguish where the Mn atoms are located in the columns, whether there is perfect order along the tilt boundary or if the Mn atoms occupy the columns randomly along the beam direction. It should be noted that the superstructure shown in Fig. 7
AbstractA complete analysis of the modulated structure of NbTe4 is presented, using the newly developed formalism of de Wolff [Acta Cryst. (1974 up to the fifth harmonic in the modulation wave. However, the values found for higher harmonics do not describe the real modulation wave, but assume values to account for the variation in intensities due to experimental errors. Therefore, only the amplitudes up to the second order are considered to be reliable.