... structures

Appendices A and B are only available in electronic form at http://www.edpsciences.org

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... source

A third way to explain BBP is the non-linear pulsating stage of instability (Trakhtengertz 1968; Zaitsev 1970; Kuijpers 1978). We don't consider this approach for our event of interest because we would obtain BBP with too short periods (milliseconds, Zaitsev & Zlotnik 1986).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... level

$\tau_{5000}$ is an optical depth in the line $\lambda=5000~$A. At the level $\tau_{5000}=1$ the conductivity is isotropic.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... exist

Electron acceleration due to magnetic field reconnection in several small regions within a flaring flux tube was analyzed by Kuijpers (1981). Such a mechanism provides sporadic acceleration of electrons.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... and

Cross section of the pulsation source at the trap apex.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... frequency

Unlike the Appendices here and hereafter the term $f=\omega/2\pi$ is used instead of $\omega$ for convenient comparison with observations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... electrons

for example $v_{\rm e}\sim (10\div 20)v_{\rm T}$ at harmonics $s\sim f_{\rm p}/ f_B \sim (10\div 30)$.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

As it will be seen later, in the source of ZP the inequality $f_{\rm p}\gg f_B$ or $s\gg 1$ is fulfilled. That is why the dependence $f_{\rm p}(h)$ instead of $f_{\rm uh}(h)=(f_{\rm p}^2+f_B^2)^{1/2}$ is plotted in this figure.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... too

Correspondingly, the negative and wave-like drift of zebra stripes given for example in Elgaroy (1961) can be associated with the increase of the magnetic field or plasma heating in the source volume.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...)

The density shouldn't be considered as an extrapolated value of electron density at h=0 because, obviously, the temperature is not constant along the loop, being much lower towards its foot; it is a local parameter without large scale importance.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...$\Delta\omega\leq\omega_B$

Winglee & Dulk (1986) considered only the first order corrections over the small parameters $k^2v_{\rm T}^2/\omega^2$ in the "thermal'' dielectric tensor. That corresponds to neglecting the important resonance term in (A.4).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...1975)

The basic reference is Zheleznyakov & Zlotnik (1975). We will not refer to it repeatedly but it is relevant for most details.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... effects

It should be emphasized that in the nonrelativistic approach the instability determined by the pole in the denominator in (B.4) is due to grouping and re-distribution of electrons over longitudinal velocities $v_{\parallel}$ under the action of a wave field, though the distribution (B.1) over $v_{\parallel}$ is an equilibrium function. In this case the reversal of the sign of the right side of (B.4) is due to $\partial f/\partial v_{\perp}$. The Doppler shift is a necessary detail of such an instability, and it was this effect in the nonrelativistic approach that was supposed as the cause of the instability responsible for the ZP. The statement that in this work the Doppler shift was neglected (as made by Winglee & Dulk 1986) is a misunderstanding.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.