The Brownian motion of a microscopic particle in a fluid is one of the cornerstones of statistical physics and the paradigm of a random process. One of the most powerful tools to quantify it was provided by Langevin, who explicitly accounted for a short-time correlated "thermal" force. The Langevin picture predicts ballistic motion, x 2 ∼ t 2 at short-time scales, and diffusive motion x 2 ∼ t at long-time scales, where x is the displacement of the particle during time t, and the average is taken over the thermal distribution of initial conditions. The Langevin equation also predicts a superdiffusive regime, where x 2 ∼ t 3 , under the condition that the initial velocity is fixed rather than distributed thermally. We analyze the motion of an optically trapped particle in air and indeed find t 3 dispersion. This observation is a direct proof of the existence of the random, rapidly varying force imagined by Langevin. Langevin [4], who introduced a stochastic differential equation for the motion of a particle in a thermal bath, with an explicit rapidly varying force f (t) constantly exerted on the particle by the fluid molecules of the bath.The Langevin equation governs the velocity v =ẋ of a free particle in a viscous fluid, in the absence of hydrodynamic memory effects:vwhere m is the particle mass, and τ = m/6πηa is the viscous relaxation time of the spherical particle of radius a in a fluid with viscosity η. The force f is "delta-correlated," meaning that its fluctuations are present down to arbitrarily short-time scales. This is equivalent to having a flat spectrum up to arbitrarily high frequencies; hence, it is referred to as white noise. The variance of the particle velocity is v 2 eq = τ at equilibrium, and thanks to the equipartition theorem is such that v 2 eq = k B T /m, thus, relating temperature T and damping rate 1/τ with the magnitude of the random force, a manifestation of the fluctuation-dissipation theorem.Langevin did not exhaust all the riches of his model but used it to compute the ensemble average squared position of the particle in the long-time limit (t τ ) and found, for particles all starting at x 0 = 0, that (and Fürth [7]) devised a similar model and reached the same conclusion regarding the existence of a diffusive régime [Eq. (2)] but also noted the existence of ballistic behavior, x 2 = τ t 2 , in the shorttime limit t τ . This result, however, still depends on an average being performed on an initial condition of particles with distributed velocities. We come back below to that crucial point, not singled-out as such by either Taylor or Langevin.Consider a particle released at t = 0 in a thermalized medium (i.e., a bath of uniform temperature T ), with constrained initial velocity v 0 (whether it is equal to zero or not) and position x 0 = 0. Solving the Langevin model [Eq. (1)] for the displacement x(t) gives a meanwhere ·|x 0 ,v 0 denotes an average over many realizations of the thermal force, with fixed initial conditions (as opposed to an average over the thermal distribution of...