The relationship of electron plasma oscillations to type III radio emissions and low-energy solar electrons

An extensive study of the IMP-6 and IMP-8 plasma and radio wave data has been performed to try to find electron plasma oscillations associated with type III radio noise bursts and low energy solar electrons. This study shows that electron plasma oscillations are seldom observed in association with solar electron events and type III radio bursts at 1.0 AU. In nearly four years of observations only one event was found in which electron plasma oscillations are clearly associated with solar electrons. Numerous cases were found in which no electron plasma oscillations with field strengths greater than 1 V/m could be detected even though electrons from the solar flare were clearly detected at the spacecraft.For the one case in which electron plasma oscillations are definitely produced by the electrons ejected by the solar flare, the electric field strength is very small, only about 100 V/m. This field strength is about a factor of ten smaller than the amplitude of electron plasma oscillations generated by electrons streaming into the solar wind from the bow shock. Electromagnetic radiation, believed to be similar to the type III radio emission, is also observed coming from the region of more intense electron plasma oscillations upstream of the bow shock. Quantitative calculations of the rate of conversion of the plasma oscillation energy to electromagnetic radiation are presented for plasma oscillations excited by both solar electrons and electrons from the bow shock. These calculations show that neither the type III radio emissions nor the radiation from upstream of the bow shock can be adequately explained by a current theory for the coupling of electron plasma oscillations to electromagnetic radiation. Possible ways of resolving these difficulties are discussed.     

Solar radio bursts of spectral type II,coronal shocks,and optical coronal transients

The association of solar radio bursts of spectral type II and coronal shocks with solar flare ejecta observed in H, the green coronal line, and white-light coronagraphs is examined. Rather than identifying fast-moving optical coronal transients with outward-travelling shock waves that generate type II radio bursts, as has been suggested in some earlier papers, we suggest that, for the most part, such transients should probably be identified with piston-type phenomena well behind the shock. We then discuss a general model, consisting of three main velocity regimes, in which we relate type II radio bursts and coronal shocks to optically-observed ejecta.     

Formation Of Coronal Mhd Shock Waves – I. The Basic Mechanism

The formation and evolution of a large amplitude MHD perturbation propagating perpendicular to the magnetic field in a perfectly conducting low plasma is studied. The perturbation is generated by an abrupt expansion of the source region. Explicit expressions for the time and the distance needed for the transformation of the perturbation's leading edge into a shock wave are derived. The results are applied to coronal conditions and the dynamic spectra of the radio emission excited by the shock are synthesized, reproducing metric and kilometric type II bursts. The features corresponding to the metric type II burst precursor and the moving type IV burst in the case of kilometric type II bursts are identified. A specific radio signature that is sometimes observed at the onset of a metric type II burst is found to appear immediately before the shock wave formation due to the associated growth of the magnetic field gradient. Time delays and starting frequencies of bursts' onsets are calculated and presented as a function of the impulsiveness of the source-region expansion, using different values of the ambient Alfvén velocity and various time profiles of the expansion velocity. The results are confronted with the observations of metric and kilometric type II solar radio bursts.     

Electron plasma oscillations associated with type III radio emissions and solar electrons

An extensive study of the IMP-6 and IMP-8 plasma and radio wave data has been performed to try to find electron plasma oscillations associated with type III radio noise bursts and low-energy solar electrons. This study shows that electron plasma oscillations are seldom observed in association with solar electron events and type III radio bursts at 1.0 AU. In nearly four years of observations only one event was found in which electron plasma oscillations are clearly associated with solar electrons. For this event the plasma oscillations appeared coincident with the development of a secondary maximum in the electron velocity distribution functions due to solar electrons streaming outwards from the Sun. Numerous cases were found in which no electron plasma oscillations with field strengths greater than 1 μV m^?1 could be detected even though electrons from the solar flare were clearly detected at the spacecraft. For the one case in which electron plasma oscillations are definitely produced by the electrons ejected by the solar flare the electric field strength is relatively small, only about 100 μV m^?1. This field strength is about a factor of ten smaller than the amplitude of electron plasma oscillations generated by electrons streaming into the solar wind from the bow shock. Electromagnetic radiation, believed to be similar to the type III radio emission, is also observed coming from the region of the more intense electron plasma oscillations upstream of the bow shock. Quantitative calculations of the rate of conversion of the plasma oscillation energy to electromagnetic radiation are presented for plasma oscillations excited by both solar electrons and electrons from the bow shock. These calculations show that neither the type III radio emissions nor the radiation from upstream of the bow shock can be adequately explained by a current theory for the coupling of electron plasma oscillations to electromagnetic radiation. Possible ways of resolving these difficulties are discussed.     

Type II–IV radio bursts and compact and diffuse white-light clouds in the outer corona of December 14, 1971

The radio observations of type II–IV bursts on December 14, 1971 are analyzed. These radio events were associated with a H-spray or eruptive prominence, and later followed by several compact moving clouds observed with the NRL white-light coronagraph aboard OSO-7. There was also observed a diffuse expanding cloud behind the compact moving clouds.From the comparison of the interferometer observation of the bursts with the optical observation, it is strongly suggested that the compact moving clouds were likely to be the optical counterparts of the sources of moving type IV radio emission. This fact suggests that the magnetic bubbles were really produced in the flare process. The frequency-drift of the first group of type II bursts was so rapid, that we could neither identify the type II shock with the leading edge of the diffuse expanding cloud nor interpret it as the piston-driven shock of the latter. Because of the uncertainty of the velocities of the compact clouds due to the projection effect, the possibility that the type II shock was the piston-driven shock of the compact clouds cannot be excluded. Nevertheless we suggest that the type II shock was a blast type MHD shock and had no direct physical relation to the flare-associated mass-ejection processes. The relation between the type II–IV bursts and the interplanetary shock is also discussed.     

The structure of the turbulent shock wave propagating in the solar atmosphere across the magnetic field

A consistent account of plasma turbulence in magnetohydrodynamics equations describing transport processes across the magnetic field is presented. The structure of the perpendicular shock wave generated in the solar atmosphere, as a result of either local disturbance of the magnetic field or dense plasma cloud motion with a frozen-in magnetic field, has been investigated. The region of parameters in the solar atmosphere at which the electron-ion relative drift velocity u exceeds the electron thermal velocity V _eand generation of radio emission becomes possible, has been determined. The plasma turbulence inside the front has been shown, under conditions of solar corona, not to cause the oscillation structure of shock front to break down. Under chromospheric conditions, the shock profile is aperiodical. Then, the condition u > V_ecan be satisfied and shock waves having an Alfvén Mach number M which exceeds the critical value M _c 3.3 for aperiodical shock waves can exist (Eselevich et al., 1971a). Arguments are given in favour of the fact that perpendicular shock waves are generated in the Sun's atmosphere when dense plasma clouds, with a frozen-in magnetic field, are expanded.     

The relation between the surges and solar radio emission

The 120 limb surges which have been observed by means of Wrocaw Observatory coronagraph from September 1966 to November 1977 are investigated. The evolution of surges was compared with the radio data during the surges. A correlation between radio bursts and the surges was found, particularly with chains of type I radio bursts, which is the first reliable correlation found of these bursts with non-radio events. The type I correlation only applied for surges without accompanying flare, of which 43% are correlated with this type of radio emission. In 23 of 30 associated events the start of a surge coincided within 5 minutes with the start or an enhancement of the type I storm. If flares were present, the association was not significant.We also compared the maximum height reached by a surge with the frequencies of the radio bursts emitted at the same time and the maximum velocity of the rising surge with the frequency drift of type I chains. No such a correlation was however found.We discuss the possibility that surges are the result of a sudden energy input into the chromosphere related to the type I source in the corona.     

A statistical study of moving type IV bursts based on Culgoora radioheliograph observations

Thirty-one moving type IV (IV(M)) bursts recorded with the Culgoora radioheliograph are examined to deduce their characteristic features, such as spatial distribution, projected velocity, etc., and their relation to other phenomena. The distribution of the projected velocity suggests that less than 15% of the total IV(M) bursts have fast velocities (>1000 km s^–1), almost equal to MHD shock velocity, and that the remaining IV(M) bursts have slower velocities (400 km s^–1) and are probably not associated with MHD shock waves. Most of the slow IV(M) bursts (and 70% of the total IV(M) bursts) are of an isolated plasmoid type. Even if they are associated with minor H flares, IV(M) bursts of the isolated-plasmoid type have 10³¹ ergs in the form of magnetic energy. They are in many cases closely associated with extended flare-continuum sources; this seems plausible if the flare continuum is interpreted as an interaction of a plasmoid with a large-scale magnetic arch.The association of IV(M) bursts with energetic proton events seems to be poor - contrary to expectation.     

Dynamics of a cloud of fast electrons travelling through the plasma

A simulation of normal type III radio bursts has been made in a whole frequency range of about 200 MHz to 30 kHz by the usage of the semi-analytical method as developed in previous papers for the plasma waves excited by a cloud of fast electrons. Three-dimensional plasma waves are computed, though the velocities of fast electrons are assumed to be one-dimensional. Many basic problems about type III radio bursts and associated solar electrons have been solved showing the following striking or unexpected results.Induced scattering of plasma waves, by thermal ions, into the plasma waves with opposite wave vectors is efficient even for a solar electron cloud of rather low number density. Therefore, the second harmonic radio emission as attributed to the coalescence of two plasma waves predominates in a whole range from meter waves to km waves. Fundamental radio emission as ascribed to the scattering of plasma waves by thermal ions is negligibly small almost in the whole range. On the other hand, third harmonic radio emission can be strong enough to be observed in a limited frequency range.If, however, the time integral of electron flux is, for example, 2 × 10¹³ cm^–2 (>5 keV) or more at the height of 4.3 × 10¹⁰ cm ( _p = 40 MHz) above the photosphere, the fundamental may be comparable with or greater than the second harmonic, but an effective area of cross-section of the electron beam is required to be very small, 10¹⁷ cm² or less, and hence much larger sizes of the observed radio sources must be attributed to the scattering alone of radio waves.The radio flux density expected at the Earth for the second harmonic can increase with decreasing frequencies giving high flux densities at low frequencies as observed, if x-dependence of the cross-sectional area of the electron beam is x ^1.5 or less instead of x ², at least at x 2 × 10¹² cm.The second harmonic radio waves are emitted predominantly into forward direction at first, but the direction of emission may reverse a few times in a course of a single burst showing a greater backward emission at the low frequencies.In a standard low frequency model, a total number of solar electrons above 18 keV arriving at the Earth orbit reduces to 12% of the initial value due mainly to the collisional decay of plasma waves before the waves are reabsorbed by the beam electrons arriving later. However, no deceleration of the apparent velocity of exciter appears. A change in the apparent velocity, if any, results from a change in growth rate of the plasma waves instead of the deceleration of individual electrons.Near the Earth, the peak of second harmonic radio flux as emitted from the local plasma appears well after the passage of a whole solar electron cloud through this layer. This is ascribed to the secondary and the third plasma waves as caused in non-resonant regions by the induced scattering of primary plasma waves in a resonant region.     

71 GHz (4.2 mm) solar radio bursts in the period July 1967 to December 1969

The results of 2 1/2 years (July 1967 – December 1969) monitoring of solar radio bursts at 71 GHz ( = 4.2 mm) at the Radio and Space Research Station, Slough are presented. During this period only seven events were positively identified as 71 GHz bursts. One of these events (6 July, 1968) is among the largest solar bursts ever recorded anywhere in the microwave-millimetre wave band (47000 × 10^–22Wm^–2Hz^–1), and the associated magnetic field may possibly have exceeded 7200 G. Another event (27 March, 1969) has demonstrated that bursts at 71 GHz can be both intense (4700 × 10^–22Wm^–2Hz^–1) and complex. On other occasions, the absence of any detectable event at 71 GHz helps to define the high frequency spectrum of the burst, this being an important factor in determining the initial energy distribution of the electrons ejected by the associated flare. On one such occasion (21 March, 1969) the derived energy distribution index is 8, in contrast with the more usual values of 2–4.1969–1970 NCR-OAR Senior Post-Doctoral Research Associate at Air Force Cambridge Research Laboratories, L. G. Hanscom Field, Bedford, Mass., U.S.A.     

Solar microwave bursts recorded at 2.84 GHz with millisecond time resolution

A millisecond recorder for solar observation at 2.84 GHz was put to work in 1981. From April 1981 to September 1982 it recorded 250spike pulse events, which are here listed together with information on the associated solar flares, hard X-ray bursts and radio bursts. In defining these spike events, particular attention has been paid to ensure the reliability of the recorded data. Statistical analysis of the data has given some interesting results. The spike pulses have shorter duration and higher flux density than were previously known, and the basic units of the spike pulse events are single spike pulses crowded together forming separate clusters. Many spike pulses observed are not yet resolved at 1 ms. Fast spike pulse events are closely correlated with complex magnetic field regions, and are often accompanied by hard X-ray bursts and fast drifting radio bursts. Some such events showed no correspondence with the radio bursts at the same frequency, 2.84 GHz, but corresponded to those occurring at dm and short cm wavelengths. Some theoretical investigations of the spike phonomena have been carried out, the details of which will be discussed elsewhere.     

Behavior of the flare-produced coronal MHD wavefront and the occurrence of type II radio bursts

The propagation of the weak MHD fast-mode shock emitted into the corona by flares at their explosive phase is computer-simulated. It is shown as the result that the shock wave is refracted towards the low Alfvén velocity regions pre-existing in the corona, and the strength of the shock, which is otherwise weak, is drastically enhanced on encountering low- V _A regions due to the focussing effect by refraction and also due to the lowered propagation velocity of the shock in such regions. It is expected that electron acceleration takes place in such a drastic strengthening of the shock, leading to the local excitation of plasma waves and eventually to the occurrence of radio bursts at such locations. Such locations of shock strength enhancement, when computed by using HAO realistic models of coronal density and magnetic field of the day of certain type II burst events, actually coincide roughly with the observed positions of type II bursts. Peculiar configurations of type II burst sources as well as their occurrence even beyond the horizon of the responsible flare are explained consistently by the large scale refraction and the local enhancement of the shock due to the global and local distribution of Alfvén velocity in the corona. A unified interpretation is given for the occurrence of type II bursts and Moreton's wave phenomena, and also the relation of our MHD fast-mode disturbance with other flare-associated dynamical phenomena is discussed.     

Unstable chromospheric dark features and type-III bursts association

The temporal association between the kinematic parameters of chromospheric dark features (DF) and the production of radio type-III bursts is investigated during a period of five months. The Doppler shifts inside six different DF are measured by means of the Meudon Multichannel Subtracting Double Pass Spectrograph (MSDP) during periods of some minutes around 24 type-III bursts. The position of the radio bursts has been checked to be associated with the same active region observed by MSDP, by using the Nançay Radioheliograph. It appears that 23 out of 24 bursts take place when the DF is totally or predominantly blue-shifted. In 18 out of 21 cases, a maximum of the outward velocity is observed in the optical image closest in time to the radio burst. The following peculiarities are also shown by the analyzed DF:

1. All of them present a lengthened shape, in most cases pointing toward a sunspot: a bright region coinciding with a parasitic polarity is observed in between.
2. Horizontal velocities along the DF major axis are often observed, always in a direction opposite to the sunspots.

This association is interpreted by assuming that a shock wave, generated by the moving chromospheric material, is able to accelerate coronal electrons (according to a mechanism proposed by Leroy and Mangeney, 1984) thus producing type-III bursts.     

Solar microwave bursts as indicators of the occurrence of solar proton emission

A study has been made of the relation of 19 GHz( = 1.58 cm) solar radio bursts to solar proton emission, with particular reference to the usefulness of relatively long duration bursts with intensities exceeding 50% of the quiet Sun flux (or exceeding 350 × 10^–22 W m^–2 Hz^–1) as indicators of the occurrence of proton events during the four years from 1966–69. 76 to 88% of such bursts are directly associated with solar protons and 60 to 85% of the moderate to large proton events in the four year period could have been predicted from these bursts. The complete microwave spectra of the proton events have also been studied, and have been used to extend the results obtained at 19 GHz to other frequencies, particularly in the 5–20 GHz band. The widely used frequency of 2.8 GHz is not the optimum frequency for this purpose since proton events have a minimum of emission in this region. Most of the radio energy of proton events is at frequencies above 10 GHz. The radio spectra of proton events tend to peak at higher frequencies than most non-proton events, the overall range being 5 to 70 GHz, with a median of 10–12 GHz and a mean of 17 GHz.On leave from the Radio and Space Research Station, Slough, England, as 1969–1970 National Research Council-National Academy of Sciences Senior Post-Doctoral Research Associate at AFCRL.     

Studies on metric–decimetric Type II bursts in relation to soft X-ray flares

The relationship between metric type II radio bursts and soft X-ray (SXR) flares is studied. Type II bursts are highly associated with SXR flares. The duration and drift rate of type II bursts are found to depend on the duration, asymmetry in duration (ratio of rise time to duration), as well as on the peak flux of SXR bursts. Important results obtained are: (i) the durations of type II bursts are linearly correlated with the durations of associated SXR bursts in case of long-lived events (duration >40 min), whereas in short-lived flares such a correlation is not found, (ii) the durations of type II bursts do not depend upon the SXR peak flux, (iii) more durable type II radio bursts are correlated with more symmetric SXR bursts, (iv) average drift rates of type II bursts are larger in the events associated with more powerful and more symmetric SXR bursts.     

Statistical Analysis of Solar Events Associated with Storm Sudden Commencements over One Year of Solar Maximum During Cycle 23: Propagation from the Sun to the Earth and Effects

Taking the 32 storm sudden commencements (SSCs) listed by the International Service of Geomagnetic Indices (ISGI) of the Observatory de l’Ebre during 2002 (solar activity maximum in Cycle 23) as a starting point, we performed a multi-criterion analysis based on observations (propagation time, velocity comparisons, sense of the magnetic field rotation, radio waves) to associate them with solar sources, identified their effects in the interplanetary medium, and looked at the response of the terrestrial ionized and neutral environment. We find that 28 SSCs can be related to 44 coronal mass ejections (CMEs), 15 with a unique CME and 13 with a series of multiple CMEs, among which 19 (68%) involved halo CMEs. Twelve of the 19 fastest CMEs with speeds greater than 1000 km?s^?1 are halo CMEs. For the 44 CMEs, including 21 halo CMEs, the corresponding X-ray flare classes are: 3 X-class, 19 M-class, and 22 C-class flares. The probability for an SSC to occur is 75% if the CME is a halo CME. Among the 500, or even more, front-side, non-halo CMEs recorded in 2002, only 23 could be the source of an SSC, i.e. 5%. The complex interactions between two (or more) CMEs and the modification of their trajectories have been examined using joint white-light and multiple-wavelength radio observations. The detection of long-lasting type IV bursts observed at metric–hectometric wavelengths is a very useful criterion for the CME–SSC events association. The events associated with the most depressed Dst values are also associated with type IV radio bursts. The four SSCs associated with a single shock at L1 correspond to four radio events exhibiting characteristics different from type IV radio bursts. The solar-wind structures at L1 after the 32 SSCs are 12 magnetic clouds (MCs), 6 interplanetary coronal mass ejections (ICMEs) without an MC structure, 4 miscellaneous structures, which cannot unambiguously be classified as ICMEs, 5 corotating or stream interaction regions (CIRs/SIRs), one CIR caused two SSCs, and 4 shock events; note than one CIR caused two SSCs. The 11 MCs listed in 3 or more MC catalogs covering the year 2002 are associated with SSCs. For the three most intense geomagnetic storms (based on Dst minima) related to MCs, we note two sudden increases of the Dst, at the arrival of the sheath and the arrival of the MC itself. In terms of geoeffectiveness, the relation between the CME speed and the magnetic-storm intensity, as characterized using the Dst magnetic index, is very complex, but generally CMEs with velocities at the Sun larger than 1000 km?s^?1 have larger probabilities to trigger moderate or intense storms. The most geoeffective events are MCs, since 92% of them trigger moderate or intense storms, followed by ICMEs (33%). At best, CIRs/SIRs only cause weak storms. We show that these geoeffective events (ICMEs or MCs) trigger an increased and combined auroral kilometric radiation (AKR) and non-thermal continuum (NTC) wave activity in the magnetosphere, an enhanced convection in the ionosphere, and a stronger response in the thermosphere. However, this trend does not appear clearly in the coupling functions, which exhibit relatively weak correlations between the solar-wind energy input and the amplitude of various geomagnetic indices, whereas the role of the southward component of the solar-wind magnetic field is confirmed. Some saturation appears for Dst values \(< -100\) nT on the integrated values of the polar and auroral indices.     

Solar X-ray bursts at energies less than 10 keV observed with OSO-4

Using data from a proportional counter spectrometer, sensitive in the wavelength range 1–20 Å, on OSO-4, X-ray bursts in the energy band 3.0 to 4.5 keV have been studied. 150 events have been identified between October 27, 1967 and May 8, 1968, mostly of an impulsive nature. Some gradual rise and fall bursts occur, but there is a selection bias against such long-enduring events. A study of the profiles of these events reveals no basis for identifying different types of impulsive event.Single frequency radio bursts and H flares of class > 1F are almost always accompanied by X-ray enhancements. For the sample of X-ray events, only 25% are correlated with radio bursts and 46% with flares. Only 11% of the sample events are associated with type III radio bursts. Microwave burst peaks occur an average of two minutes earlier than the X-ray burst peak, but the first observation of X-ray activity is usually before the start of the corresponding microwave burst.Impulsive bursts, although differing widely in fall time, are due to the heating of a volume of plasma from a temperature of 10.0 to 30.0 × 10⁶ K. Differences infall time probably indicate different electron densities in the source. Observation of an iron line at 1.9 Å suggests that a non-thermal mechanism may be operating during some of these events since the temperatures are too low to permit thermal excitation of the 1s ²-1s 2p transition in Fe⁺²⁴. It is also possible that, in spite of the low temperature, most of the iron ions have been stripped to the Fe⁺²⁴ stage. Collisional excitation and dielectronic recombination processes would then be able to provide the observed flux in the resonance line of Fe⁺²⁴. A gradual rise and fall event and event precursors have also been studied.     

Magnetic-cloud chamber effects behind flare-generated shock waves

A mechanism explaining the generation of the helium-enriched plasma-condensation colud (HAE-events) behind the front of shock waves associated with mass-ejecting flares is presented. The mechanism is based on the occurence of physical conditions, analogous to those in a Wilson cloud chamber in a magnetic field, behind the front of a flare-generated shock wave propagation out into interplanetary space. Consequently, if the solar atmosphere above the flare active region is saturated with ejected helium plasma, conditions are created for the forming of the helium-enriched plasma-condensation colud in the temperature-depressed region behind the shock wave front.     

A model for solar radio pulsations at short centimetric band

In this paper, the observed solar radio pulsations during the bursts at 9.375 GHz are considered to be excited by some plasma instability. Under the condition of the conservation of energy in the wave-particle interaction, the saturation time of plasma instabilities is inversely proportional to the initial radiation intensity, which may explain why the repetition rate of the pulsations is directly proportional to the radio burst flux at 9.375 GHz as well as 15 GHz and 22 GHz. It is also predicted that the energy released in an individual pulse increases with increasing the flux of radio bursts, the modularity of the pulsations decreases with increasing the flux of radio bursts, these predictions are consistent with the statistical results at 9.375 GHz in different events. The energy density of the non-thermal particles in these events is estimated from the properties of pulsation. For the typical values of the ambient plasma density (10⁹ cm^–3) and the ratio between the nonthermal and ambient electrons (10^–4), the order of magnitude of the energy density and the average energy of the nonthermal electrons is 10^–4 erg/cm³ and 10 kev, respectively. It is interesting that there are two branches in a statistical relation between the repetition rate and the radio burst flux in a special event on March 11–17, 1989, which just corresponds to two different orders of magnitude for the quasi-quantized energy released in these five bursts. This result may be explained by the different ratios between the thermal and the nonthermal radiations.     

Are ‘negative bursts’ due to absorption?

Optical activity near the time of two microwave negative bursts is distinctive enough to demonstrate a real relation, but does not occur simultaneously with the radio events. Such a loose association between microwave flux decreases and H prominence or filament activity is typical of a sample of 23 negative bursts. A model of a microwave occulting cloud differs from that of an H absorber: the H dark flocculus is relatively low, dense, and compact; the microwave occulter is higher in the corona, larger, and more completely ionized. The two types of absorption are not expected to be seen simultaneously, although they could be separate phases of an ejection of chromospheric material into the corona. The association of microwave decreases with H activity in some cases, and their usual non-simultaneity, is consistent with interpretation of the decrease as absorption, but it does not rule out alternative interpretations such as an intrinsic change in the emission of the microwave source.The location of decrease-associated activity has a suggestive, though not statistically significant, bias toward east limb that leads one toward an absorption interpretation, with asymmetry introduced by a tendency for ejected material to move from the leading part of an active region toward the following part, agreeing with Liszka's observation of asymmetry in the line-of-sight velocities of prominences.