In this section, we propose an answer to an idealized problem: Given that the only source of end effects are those discussed so far, how might one disentangle these two different effects from experimental data? This is tantamount to asking how the drift time for the centroid of the swarm might be measured, as (from eq. 4) this measurement is influenced only by the non-hydrodynamic process.
We assume that the collector current can be recorded as a
function of drift length and time, and that the drift velocity
and the diffusion coefficient 
are
known for the system (presumably by extrapolating the data to
infinite drift lengths). Then taking the first moment of the current, we have
where 
is the position of the centroid at time
t, and in the last line, n has been normalized. At large
times, according to (
), this should become linear in
t with slope 
. The intercept of this line with
the t=0 axis will give the value of 
.