For the reaction 290 MeV 2~197Au the number of neutrons emitted in coincidence with e-particles is used to discriminate between two equally important but qualitatively different mechanisms for the emission of fast e-particles. They are connected with low and high inelasticity, respectively, and the resulting e-particles differ in mean velocity and velocity distribution.All the time since 1961 when Britt and Quinton [1] for the first time observed the emission of fast forward-directed e-particles in heavy-ion collisions at energies around 10 MeV/nucleon, their origin has been the subject of much theoretical and experimental effort. The explanations proposed range from preequilibrium emission [2], emission from a hot spot [3,4], a hot moving source [5,7] or a fireball, over sequential projectile decay (sequential breakup) [8-12] to projectile fragmentation (direct breakup) [13][14][15], uncorrelated emission 1-17-19], and incomplete fusion [20][21][22] or massive transfer [23,24]. In all these investigations, only one mechanism was considered explicitly, a possible second mechanism appearing only sometimes through an unexplained remainder in the inclusive cross sections [10,25]. However, assessing all existing evidence, one has to anticipate the virtual coexistence of (at least) two mechanisms in a heavy-ion reaction above 10 MeV/ nucleon: the sequential decay of excited projectilelike fragme~.ts as a more familiar process, and besides that a type of fast, non-equilibrium emission. Here we present for the representative case of the 290MeV 2~ system an explicit experimental decomposition of inclusive e-particle cross sections into two components resulting from different production modes. The decomposition was achieved by counting with high efficiency the neutrons emitted simultaneously with each c~-particle. The neutron number is a measure of the energy converted into internal excitation. This method was previously applied [26] to projectile-like fragments from the same system. There it allowed to discriminate between projectile residues the formation of which implied the transfer of an e-particle to the target, and the same residues emerging after emission of an e-particle. In that experiment results were also obtained for e-particles which shall be discussed here. Some experimental details are given in [26]. The previous information is resumed in Fig. 1. It shows the inclusive e-particle spectrum measured [25] at 22 ~ which is characterized by the well known intense high-energy contribution superimposed onto the evaporation component. The latter is derived from a measurement at backward angles using the symmetry of its c.m.s, angular distribution about 90 ~ A second component shown is the calculated contribution of e-particles accompanied by a projectile-like fragment, i.e. the contribution of projectile break-up. The calculation is based on a measurement of in-plane coincidences [9] of e-particles with projectile-like fragments. All (or at least about 95 ~) of these coincidences were due to the sequential edecay ...