We shall discuss acceleration of solar plasma electrons to energies of 30-100 keV by nonlinear waves (called "solitons" 1 )-the mechanism already used 2 " 4 to explain the acceleration of soft electrons of E ~ 1 keV.Solitons were discovered in the transition region between the magnetosphere and the interplanetary medium in Pioneer-I magnetic measurements. 5 The theory shows that with decreasing angle 6 between the undisturbed magnetic field H 0 and soliton velocity u 0 , the amplitude #+ of the magnetic field pulse and the maximum energy of its electrons E^ e ~mv L -\ 2 /2 increase 6 ? 3 : cos0/MY /2 >~~ COS0/M\ 1/2 TTl^ >9tt>^b ) > H *= H^-> rn/Hence to the electrons is imparted not the energy of the undisturbed magnetic field H 0 , but that of the field H 0 (M/m) 1/2 (at 0 = 0) enhanced within the soliton, H-\ 2 being three orders of magnitude higher than H 2 .Let us call the solitons running along the undisturbed magnetic field (6 = 0) parallel solitons. This solution of hydromagnetic equations was obtained by Saffman 7 and Tverskoi. 8 Experimental evidence gained in the last years allows us to consider now the hypothesis that 30-to 100-keV electrons are accelerated in parallel solitons and corresponding nonstationary pro-~ m-3 cesses.The relations (1) give (for 0 = 0) the required electron energy range. From Fig. 46 of Ness et al., 9 just before electron flux spikes H Q = 10-20 y. If ^ = 5 cm"" 3 then E^ e = 45 keV at H 0 = 10 y and E^e = 180 keV at H 0 = 20 y.But there remain two more soliton properties to be checked: the high field H^ and ion energy E-f i = E^ £>30 keV. In geostationary (3H ~7) solitons observed on Pioneer I, H^< 100 y, so we interpreted them 3 as oblique ones with 6 = 70-80°. The parallel solitons, having WL= 20-30, cannot be geostationary; they move relative to earth with the velocity 10 8 cm/sec and pass the device within a time of C/U)Q e u o~ 10" sec. The amplifier of the Pioneer-I magnetometer could not record such rapid flux variations. They could be observed by means of a simple induction coil and a broad pass-band amplifier. The 100-keV protons were detected on Explorer-12. 10 But the instrument limitations did not permit observation of real soliton pulses (because, e.g., of too high an energy range, small acceptance angle, time resolution limitations, etc.). The ion pulse flux is always much lower than the electronic one, due to lower ion velocity.The parallel alignment of plasma flow and the magnetic field H 0 can also occur in several cases: (1) On the day side-during systematic field reversals when passing from some solar magnetic field sector 11 to the next one.(2) During spontaneous interplanetary field 327