Introduction Presenting ion mobility results in terms of CCS (collision cross section) distributions is attractive for visually comparing various molecular systems, or results obtained on different instrumental platforms, or comparing theoretical CCS with experimental data. The equations usually found in the literature to change variables (from arrival time tA to collision cross section CCS) are however incorrect. Some imply that t0 (the time spent outside the drift tube) is constant for the entire distribution. Most procedures neglect the fact that diffusion in the drift tube contributes to the peak width in the time domain, whereas that contribution is absent in the CCS domain. Here we describe how to reconstruct CCS distributions from arrival time distributions (ATDs) in a more rigorous manner. Methods All ion mobility experiments were carried out in an Agilent 6560 drift tube IMS-Q-TOF instrument, modified to allow ion mobility measurements in helium. Test samples include small molecules, native or denatured proteins, rigid or polymorphic nucleic acids, and artificial foldamers. ATD data were processed with Origin Pro 2016. Preliminary Data The proposed method involves a step-field experiment in the drift tube IMS. Each peak in the ATDexp is fitted by a Gaussian. At each voltage, we obtain the peak center tA and the full width at half maximum of the arrival time distribution (FWHMATD). The step-field experiment (tA = f(p/ΔV)) gives t0 and K0 (then CCS) for the peak center. The peak width contains one contribution due to diffusion and one contribution to the system’s polydispersity. The contribution of the diffusion is: FWHMdiff = 4(tA-t0)*√ln2*√(kBT/zeV) The contribution of the system’s structural polydispersity to the FWHMATD is thus given by: FWHMsystem² = FWHMATD² - FWHMdiff² tA and FWHMsystem allow to reconstruct a Gaussian curve (ATDsystem), representing the theoretical ATD if there had been no diffusion in the tube. Its center is the same as in ATDexp. Reconstructing the CCS distribution means finding the CCS of the ions that drifted faster or slower than the center of the peak. This assumes that conformers do not interconvert on the time scale of the drift experiment. Let’s do it for ions with arrival times at half-maximum of the peak.The step-field linear regression as a function of p/ΔV can be done for example for tA ± FWHMsystem/2, like done usually for tA. This linear regression gives t0 ± FWHMt0/2 (through the intercept) and CCS ± FWHMCCS/2 (through the slope). Knowing the CCS of the peak center, the FWHMCCS (i.e. the full width at half maximum of the CCS distribution) is deduced. Finally, the CCScenter and FWHMCCS then allow to reconstruct a Gaussian curves which render the CCS distribution in a more accurate manner. We will illustrate our method and compare it to existing ones, for molecules containing single or multiple conformers, narrow or broad peaks or a combination thereof. Novel Aspect: Because our data treatment takes diffusion into account, the width of the CCS distribution correctly represents the ion structural polymorphism.