We show that atomic motion leads not only to noticeable quantitative, but in some cases also to qualitative modification of collective effects in dense and cold atomic ensembles even in the case when the characteristic Doppler shifts are tens of times smaller than the natural linewidth. The observed influence is explained as a result of the suppression of the impact of sub-radiant collective states caused by the displacement of the atoms.Collective effects such as sub and superradiance, weak and strong (Anderson) localization have recently attracted keen interest. Atomic motion can significantly modify the character of these effects. The light emitted by one atom turns out to be non-resonant to the transition of another one moving with a different speed. In the case when the Doppler shifts are greater or comparable with the natural width of the atomic excited state, the mutual non-resonance of the atoms is taken into account by introducing a random shifts of atomic levels that are different for different atoms ([1, 2]).Lowering the temperature weakens the Doppler effect, therefore one of the most interesting objects for the study of collective effects is the atomic ensembles, cooled to sub-Doppler temperatures in special traps [3][4][5][6]. Thus, in modern magnetooptical and dipole traps, atomic clouds are cooled to temperatures of the order of 30-100 µK. In this case, the Doppler shifts and the inhomogeneous broadening of the lines are substantially less than the natural linewidth and the dipole-dipole interaction is practically indistinguishable from the case of fixed atoms (see for example [1,7]). Even the frequency diffusion due to the possible large number of light scattering by atoms is usually irrelevant [8]. For this reason, when describing collective effects the model of motionless scatterers is usually used. The displacement of atoms due to the low but finite temperature is taken into account by averaging the observed values over the random spatial configurations of atoms in the ensemble.In this paper, we show that such approach can lead not only to substantial quantitative errors, but also in some cases to qualitatively wrong conclusions, even when the Doppler frequency shifts are few tens times smaller than the natural width of the atomic transition. As an illustration, we consider two problems. In the first, we calculate the transmission of the atomic cloud and show that the time average transmittance obtained as a solution of the dynamic problem for continuously transient atoms may significantly differ from the results obtained by averaging over the random spatial configurations of immobile atoms with the same spatial distribution. The second example deals with the dynamics of the spontaneous decay in a cloud of slowly moving atoms. For times exceeding the natural lifetime, a significant effect of motion on decay rate is also found out.We consider an ensemble consisting of N >> 1 identical atoms with a nondegenerate ground state with an angular momentum J g = 0. The excited state is J e = 1. The lif...