Streaming motions of galaxy clusters within 12 000 km s-1- V. The peculiar velocity field

We analyse in detail the peculiar velocity field traced by 56 clusters within 120 h^-1 Mpc in the `Streaming Motions of Abell Clusters' (SMAC) sample. The bulk flow of the SMAC sample is 687 +/- 203 km s^-1, toward l= 260°+/- 13°, b= 0°+/- 11°. We discuss possible systematic errors and show that no systematic effect is larger than half of the random error. The flow does not drop off significantly with depth, which suggests that it is generated by structures on large scales. In particular, a Great Attractor as originally proposed by Lynden-Bell et al. cannot be responsible for the SMAC bulk flow. The SMAC data suggest infall into an attractor at the location of the Shapley Concentration, but the detection is marginal (at the 90 per cent confidence level). We find that distant attractors in addition to the Shapley Concentration are required to explain the SMAC bulk flow. A comparison with peculiar velocities predicted from the IRAS Point Source Catalogue redshift (PSCz) survey shows good agreement with a best-fitting value of β_I=Ω^0.6_m/b_I= 0.39 +/- 0.17. However, the PSCz density field is not sufficient to account for all of the SMAC bulk motion. We also detect, at the 98 per cent confidence level, a residual bulk flow of 372 +/- 127 km s^-1 toward l= 273°, b= 6° which must be generated by sources not included in the PSCz catalogue, that is, either beyond 200 h^-1 Mpc, in the zone of avoidance or in superclusters undersampled by IRAS. Finally, we compare the SMAC bulk flow with other recent measurements. We argue that, at depths ranging from 60 to 120 h^-1 Mpc, flows of order 600 km s^-1 are excluded by multiple data sets. However, convergence to the cosmic microwave background frame by a depth of 60 h^-1 Mpc is also excluded by multiple data sets. We suggest that a bulk flow of 225 km s^-1 toward l= 300°, b= 10° at depths greater than 60 h^-1 Mpc is consistent with all peculiar velocity surveys, when allowance is made for sparse sampling effects.