1: \documentclass[12pt,preprint]{aastex}
2: %\documentstyle[emulateapj]{article}
3: %\received{}
4: %\accepted{}
5:
6: \begin{document}
7:
8: \title{\bf PERIODIC ACCELERATION OF ELECTRONS IN THE 1998 NOVEMBER 10
9: SOLAR FLARE}
10: \author{A. Asai \altaffilmark{1,2},
11: M. Shimojo \altaffilmark{3}, H. Isobe \altaffilmark{1,2},
12: T. Morimoto \altaffilmark{1}, T. Yokoyama \altaffilmark{3},
13: K. Shibasaki \altaffilmark{3}, and H. Nakajima \altaffilmark{3}}
14: \email{asai@kwasan.kyoto-u.ac.jp}
15: \altaffiltext{1}{
16: Kwasan and Hida Observatories, Kyoto University, Yamashina-ku,
17: Kyoto 607-8471, JAPAN}
18: \altaffiltext{2}{
19: Department of Astronomy, Kyoto University, Sakyo-ku, Kyoto 606-8502, JAPAN}
20: \altaffiltext{3}{
21: Nobeyama Radio Observatory, Minamisaku, Nagano, 384-1305, JAPAN}
22:
23:
24: \begin{abstract}
25:
26: We present an examination of the multi-wavelength observation
27: of a C7.9 flare which occurred on 1998 November 10.
28: This is the first time of imaging observation of the quasi-periodic
29: pulsations (QPPs).
30: Four bursts were observed with the hard X-ray telescope aboard
31: {\it Yohkoh} and the Nobeyama Radioheliograph during the impulsive phase
32: of the flare.
33: In the second burst, the hard X-ray and microwave time profiles clearly
34: showed a QPP.
35: We estimated the Alfv\'{e}n transit time along the flare loop using
36: the images of the soft X-ray telescope aboard {\it Yohkoh} and the
37: photospheric magnetgrams, and found that the transit time was
38: almost equal to the period of the QPP.
39: We therefore suggest, based on a shock acceleration model,
40: that variations of macroscopic magnetic structures, such as oscillations
41: of coronal loops,
42: affect the efficiency of particle injection/acceleration.
43:
44: \end{abstract}
45:
46:
47: \keywords{acceleration of particles --- Sun : activity --- Sun : flares
48: --- Sun : radio radiation --- Sun : X-rays }
49:
50:
51: \section{INTRODUCTION}
52:
53: Non-thermal electrons generated in the impulsive phase of a flare
54: are observed in hard X-ray, $\gamma$-ray, and microwaves.
55: The light curves in these wavelengths show short-lived bursts
56: with durations between 10 s and 10 min \citep{Dulk1985}.
57: These bursts include smaller pulses with shorter duration,
58: and they sometimes show periodicity.
59:
60: A good example of such quasi-periodic pulsations (QPPs) was seen in
61: a flare on 1980 June 7 \citep{Kipl1983}.
62: \citet{Naka1983} and \citet{Kane1983} examined the temporal evolution of
63: the X-ray and radio spectra and the spatial structure of the flare.
64: They suggest that the quasi-periodic pulses indicate a modulation of
65: the particle injection/acceleration rate.
66: \citet{Taji1982} showed by numerical simulation that stored magnetic
67: energy is explosively transformed to particle acceleration.
68: They also suggest that the current loop coalescence instability induces
69: the QPPs.
70: Moreover, \citet{Taji1987} showed that the period of the QPP is equal to
71: the Alfv\'{e}n transit time `across' the current loop.
72: However, these works were mainly based on the total flux,
73: and the detailed spatial configuration of the flare that shows QPP is
74: still unknown.
75:
76: On 1998 November 10 a solar flare (C7.9 on the GOES scale) occurred in
77: NOAA 8375.
78: The flare was observed by {\it Yohkoh} \citep{Oga1991} and
79: the Nobeyama radio observatory, and clearly showed quasi-periodic
80: behavior in the hard X-ray and microwave time profiles.
81: In this paper we analyze the QPP using the high resolution X-ray and
82: microwave images.
83: Then we compare the period of the QPP with typical time-scales of flare
84: loops.
85: Finally, we discuss the effect of the magnetic structure on the particle
86: injection/acceleration rate.
87:
88:
89: \section{OBSERVATIONS}
90:
91: The solar flare occurred in NOAA 8375 (N19, W78) at 00:10~UT, 1998
92: November 10.
93: Microwave images of the flare were taken with the Nobeyama
94: Radioheliograph (NoRH; Nakajima et al. 1994) which observed the sun
95: at 17 and 34~GHz with a temporal resolution of 1.0~s.
96: The spatial resolutions of NoRH data are $12^{\prime\prime}$ for 17~GHz
97: and $6^{\prime\prime}$ for 34~GHz.
98: The Nobeyama Radio Polarimeter (NoRP; Torii et al. 1979; Shibasaki
99: et al. 1979; Nakajima et al. 1985) measured the total flux of the flare
100: at 1, 2, 3.75, 9.4, 17, 34 and 80~GHz with a temporal resolution of 0.1~s.
101: The hard X-ray images were obtained with the hard X-ray telescope
102: (HXT; Kosugi et al. 1991) aboard {\it Yohkoh}, with a spatial and
103: temporal resolution of about 5$''$ and 0.5 s, respectively.
104: The soft X-ray images were also obtained with the {\it Yohkoh} soft X-ray
105: telescope (SXT; Tsuneta et al. 1991).
106: We used full resolution images in the Partial Frame Image mode with
107: a spatial resolution of about $2^{\prime\prime}_{\,\cdot} 5$.
108:
109: Figure 1 shows the microwave, soft X-ray, and hard X-ray time profiles.
110: The top solid line is that of NoRH 17~GHz, the second dotted line is the
111: GOES 1.0 - 8.0~{\AA} channel,
112: the third solid line is the HXT M1 band (23 - 33~keV),
113: and the bottom solid line is the HXT L band (14 - 23~keV).
114: Four bursts with duration between 10 and 30~s are seen in the microwave
115: and hard X-ray emission.
116: The fine spikes in the second burst clearly show the QPPs in both
117: microwaves and hard X-rays.
118: In this paper we investigate the quasi-periodic nature of the second
119: burst.
120:
121: Figure 2 shows the images of NoRH/17~GHz, the HXT/M1~band, and SXT
122: at the second burst.
123: The flare shows a double source in the microwave data
124: ({\it gray contours}).
125: We refer to the northern and brighter source as the source ``A'' and
126: the other one as source ``B''.
127: On the other hand, in the hard X-ray data ({\it black contours})
128: only one source is seen near source ``B''.
129: This is common in the L and M1 bands of HXT, but it is unknown whether
130: it is also common in the M2 (33 - 53~keV) and H (53 - 93~keV) bands.
131: We could not synthesize the hard X-ray images in these bands at the time
132: due to insufficient counts.
133: The soft X-ray image ({\it gray scale}) shows a small bright kernel at
134: the same region at source ``B''.
135: We can also see a faint loop which connects the bright kernel and the
136: radio source ``A''.
137: %The intensity of the faint loop increased after the second burst.
138:
139:
140: \section{PERIODIC PULSATION}
141:
142: Figure 3 presents the hard X-ray and microwave time profiles for the
143: second burst, from 00:14:20~UT to 00:14:40~UT.
144: The QPPs are clearly seen.
145: In the time profiles obtained with the NoRP data the QPPs are seen in
146: 9.4 and 17~GHz.
147: The period calculated from the autocorrelation function of the time
148: profiles was 6.6~s.
149: We also calculated the period of the QPPs in hard X-rays
150: (M1 and M2 bands), and found that the period (6~s) was almost the same
151: as those in NoRP.
152: %
153: We analyzed the microwave spectrum using the NoRP data and found that
154: the power-law distribution index was about $-2 \sim -1.5$ and the
155: turn-over frequency was about 10~GHz.
156: These results indicate that the emission of the QPP originated from
157: gyrosynchrotron radiation from non-thermal electrons.
158: Since both the time profiles of the M1 and M2 bands clearly showed the
159: QPPs, the emission in hard X-rays was also produced by non-thermal
160: electrons.
161:
162: As described in section 2, the flare has two microwave sources (see Figure 2).
163: We found that the time profile of source ``A'' showed the QPP,
164: while that of source ``B'' did not.
165: The pulsating source is different from the hard X-ray source
166: that is located near source ``B''.
167: To clarify the relation between the microwave pulsation at source ``A''
168: and the hard X-ray pulsation,
169: we calculated the correlation between the time profiles of 17~GHz
170: and the M1 band as a function of the time lag between them.
171: The maximum correlation was found for a delay of the microwave pulsation
172: with a time of 0.6 -- 1.0~s.
173:
174: We suggest that the acceleration site lies near the hard X-ray source,
175: and that the delay time between microwaves and hard X-rays is explained
176: by the flight duration of non-thermal electrons between both the sources.
177: The length of the soft X-ray faint loop connecting the hard X-ray source
178: and source ``A'' is about $5.0 \times 10^4$ km.
179: If we assume that the velocity of the electrons is about
180: $1.0 \times 10^{5}$~{km s$^{-1}$} ($\sim$ 30 \% of the speed of light),
181: the delay time between microwaves and hard X-rays is about 0.5~s,
182: which is almost equal to the observed delay time.
183: %
184: The soft X-ray images show two parts related to the flare:
185: a bright kernel near source ``B'' and a faint large loop
186: connecting the kernel and source ``A'' (see Figure 2).
187: We believe that the flare occurred at the contact region of the two parts
188: via magnetic reconnection \citep{Hana1997,Hey1977}.
189: In such a scenario, the acceleration site would be located near the
190: reconnection site.
191: Hence, it is reasonable to suggest that the acceleration site is located
192: near source ``B'' and the hard X-ray source.
193:
194: Here, another question occurs;
195: why did the time profile of the microwave source ``B'' not show the QPP?
196: As an answer, we suggest that the emission mechanism of source``B'' differs
197: from that of source ``A''.
198: The emission mechanism of source ``A'' is optically-thin gyrosynchrotron
199: radiation from non-thermal electrons.
200: On the other hand, the dominant emission mechanism of source ``B'' is
201: optically-thin thermal Bremsstrahlung since the spectrum index of source
202: ``B'' that is derived from the 17 and 34~GHz is about 0.
203: Therefore, the time profiles of source ``B'' do not show the population
204: of the accelerated/injected non-thermal electrons.
205: To confirm the suggestion, we estimated the thermal radio flux from source
206: ``B'' by using the temperature and the emission measure of SXT bright
207: kernel (see section 4), and found that the flux was almost the same
208: as the observed flux of source ``B''.
209: %
210: We also suggest that the thermal emission was radiated from the thermal
211: plasma that was confined by the flare loops.
212: The loops were formed at the first burst via magnetic reconnection.
213: Therefore, although the fine structure of the kernel is not resolved in
214: the SXT images, we think that it consists of the main flare loops
215: (see Figure 2).
216:
217:
218: \section{TYPICAL TIMESCALES OF FLARE LOOP}
219:
220: In the previous section we showed that the period of the pulsation was 6.6~s.
221: What is the explanatin for the period?
222: To answer the question, we derive some typical time-scales of the flare loop
223: and discuss them.
224: Temperature, density, size, and magnetic field strength of the flare loop
225: are needed for the estimation.
226: %
227: The temperature $T$ and the volume emission measure $EM$ are derived
228: from the soft X-ray images which were taken with the SXT thick aluminum (Al12)
229: and beryllium (Be) filters.
230: The region of flare loop $A$ is defined as the region where the intensity
231: of an SXT Be filter image is more than 2,000~DN s$^{-1}$ pix$^{-1}$.
232: This leads to an area of 2.5$\times10^{18}$~cm$^2$.
233: Using the filter ratio method \citep{Hara1992},
234: the average temperature and volume emission measure over the flare loop are
235: calculated to be 9.4~MK and 8.1$\times$10$^{48}$~cm$^{-3}$, respectively.
236:
237: Here, we assume that the kernel is a cube, and that the volume $V$ of
238: the flare kernel is the area to the power of three halves,
239: i.e. $V = A^{3/2}$.
240: Then, the number density $n$ in the top of the flare loop becomes
241: $n = \sqrt{EM/V} = \sqrt{EM/A^{3/2}} \approx 4.5\times10^{10}
242: \;\mathrm{cm^{-3}}$.
243: Also, the acoustic velocity $c_s$ in the flare loop is given by
244: $c_s \approx \sqrt{\gamma k_B T / m_p} \approx 360 \;\mathrm{km \;s^{-1}}$,
245: where $k_B$ is Boltzmann's constant; $m_p$ is the proton mass;
246: $\gamma$ is the ratio of specific heats and is here assumed to be $5/3$.
247: %
248: Since the structure of the flare loop is not resolved in the SXT images,
249: we can not measure the length $l$ and width $w$ of the flare
250: loop correctly.
251: Therefore, we use the square root of the area instead of the length of
252: the flare loop ($l \sim A^{1/2} \sim $ 16,000~km).
253: Then, the acoustic transit time along the flare loop is
254: $\tau_{sl} = l/c_s \approx 44 \;\mathrm{s}$.
255: On the other hand, if we use the width of the faint loop instead of
256: the width of the flare loop ($w \sim $ 6,000~km),
257: the acoustic transit time across the flare loop becomes
258: $\tau_{sw} = l/c_s \approx 17 \;\mathrm{s}$.
259: Both the time-scales are much longer than the observational period of
260: the QPP (6.6~s).
261:
262: In order to derive the Alfv\'{e}n transit times of the flare loop,
263: we have to estimate the magnetic field strength in the corona.
264: However, this estimation is difficult in our case as;
265: (1) we can not measure the magnetic field strength of the corona directly,
266: (2) we can not obtain the actual magnetic field strength
267: in the photosphere at the flare region since it is located near
268: the north-west limb.
269: %
270: Therefore, we calculate the potential field of the flare region based on
271: a magnetogram on November 6.
272: The magnetogram was obtained with the Michelson Doppler Interferometer
273: (MDI; Scherrer et al. 1995) aboard the Solar and Heliospheric Observatory
274: ({\it SOHO}; Domingo et al. 1995).
275: The field strength on the apex of the flare loop $B$ is estimated to be
276: about 300~G using the potential field extrapolation by the software package
277: MAGPACK2 \citep{Saku1982}.
278: As a result, the Alfv\'{e}n velocity $c_A$ in the flare loop becomes
279: $c_A \approx B / \sqrt{4\pi n m_p} \approx 3,100 \;\mathrm{km \;s^{-1}}$.
280: Then, the Alfv\'{e}n transit time `along' the flare loop is
281: $\tau_{Al} = l/c_A \approx 5.1 \;\mathrm{s}$
282: and the Alfv\'{e}n transit time `across' the flare loop is
283: $\tau_{Aw} = w/c_A \approx 1.9 \;\mathrm{s}$.
284:
285: The values of our estimations are summarized in Table 1.
286: Although the Alfv\'{e}n transit time is sensitive to the magnetic field
287: strength estimated above,
288: the transit time along the flare loop ($\tau_{Al}$) is the most similar to
289: the observational period of the pulsation in microwave and hard X-rays.
290:
291:
292: \section{DISCUSSION}
293:
294: We investigated the impulsive phase of the flare whose time evolution
295: showed a clear QPP in microwaves and hard X-rays.
296: We found that the periods of the QPP were 6.6 s and
297: that the acceleration site of non-thermal electrons was located near
298: source ``B'' using correlation analysis of the time profiles.
299: Then, we estimated some typical time-scales of the flare loop from the
300: observational data and found that the Alfv\'{e}n transit time along the
301: flare loop was close to the period of the pulsation.
302:
303: \citet{Taji1987} suggest that the periods of QPP in microwaves and hard
304: X-rays are equal to an Alfv\'{e}n transit time `across' the flare loop
305: ($\tau_{Aw}$).
306: However, this transit time was much shorter than the observational period of
307: the QPP.
308: This implies that the origin of the periodic pulsation in our case is not
309: the coalescence instability in the reconnection site.
310:
311: Recently, {\it Transition Region and Coronal Explorer} ({\it TRACE};
312: Handy et al. 1999) observations have shown coronal loop oscillations
313: that was induced by a flare \citep{Naka1999}.
314: T. Miyagoshi (in preparation) performed three-dimensional MHD simulations
315: of the coronal loop oscillation and found that the period of the loop
316: oscillation is equal to the Alfv\'{e}n transit time `along' the oscillating
317: loops ($\tau_{Al}$).
318: %
319: Does the oscillation of the coronal loop relate with the particle
320: acceleration/injection?
321: \citet{Tsu1998} propose that non-thermal electrons can efficiently be
322: accelerated by a first order Fermi process at the fast shock
323: located below the reconnection X-point.
324: They suggest that the accelerated electrons are trapped between the two
325: slow shocks, and that the energy injection depends on the length of
326: the fast shock which is pinched by these slow shocks.
327: If the reconnected (flare) loop that is located under the fast shock is
328: oscillating, then the length of the fast shock probably varies with the
329: loop oscillation, synchronously.
330: %
331: Hence, we propose, under the hypothesis of the acceleration model proposed
332: by Tsuneta and Naito (1998),
333: that the origin of the QPP in microwaves and hard X-rays
334: is the modulation of the acceleration/injection of non-thermal electrons.
335: Moreover, we propose that the modulation is produced by the variations of
336: macroscopic magnetic structures, for example, oscillations of coronal loops.
337:
338:
339: \acknowledgments
340: We first acknowledge an anonymous referee for his/her useful comments and
341: suggestions, which improved this paper.
342: We'd like to thank Drs. K. Shibata, T. Terasawa, T. Saito, T. T. Ishii, and
343: T. Miyagoshi for fruitful discussions.
344: We also thank Drs. D. H. Brooks, T. Miyagoshi for their careful reading
345: and helpful comments.
346: We are indebted to Dr. T. Sakurai for allowing us to use his potential
347: field program (MAGPACK2).
348: We also thank all the members of Nobeyama Radio observatory for
349: their support in our observation and study.
350: We made extensive use of {\it SOHO} MDI Data Service and {\it Yohkoh}
351: Data Center.
352:
353:
354: \begin{thebibliography}{}
355:
356: \bibitem[Domingo et al.(1995)]{Dom1995} Domingo, V., Fleck, B., and
357: Poland, A. I., 1995, \solphys, 162, 1
358:
359: \bibitem[Dulk et al.(1985)]{Dulk1985} Dulk, G. A., McLean, D. J., \&
360: Nelson, G. J., 1985, Solar Radiophysics: Studies of emission from the sun
361: at metre wavelengths, ed.s McLean D. J. \& Labrum N. R.
362: (Cambridge: Cambridge University Press), 53
363:
364: \bibitem[Hanaoka(1997)]{Hana1997} Hanaoka, Y., 1997, \solphys, 173, 319
365:
366: \bibitem[Handy et al.(1999)]{Han1999} Handy, B. N., et al., \solphys, 187, 229
367:
368: \bibitem[Hara et al.(1992)]{Hara1992} Hara, H., Tsuneta, S.,
369: Lemen, J. R., Acton, L. W., and McTiernan, J. M., 1992, \pasj, 44, 135
370:
371: \bibitem[Heyvaerts, Priest, \& Rust(1977)]{Hey1977} Heyvaerts, J., Priest,
372: E. R., and Rust, D. M., 1977, \apj, 216, 123
373:
374: \bibitem[Kane et al.(1983)]{Kane1983} Kane, S. R., Kai, K.,
375: Kosugi, T., Enome, S., and Landecker, P. B., \& McKenzie, D. L.,
376: 1983, \apj, 271, 376
377:
378: \bibitem[Kiplinger et al.(1983)]{Kipl1983} Kiplinger, A. L.,
379: Dennis, B. R., Frost, K. J., \& Orwig, L. E., 1983, \apj, 273, 783
380:
381: \bibitem[Kosugi et al.(1991)]{Kosu1991} Kosugi, T., et al.,
382: 1991, \solphys, 136, 17
383:
384: \bibitem[Nakajima et al.(1983)]{Naka1983} Nakajima, H., Kosugi, T.,
385: Kai, K., \& Enome, S., 1983, \nat, 305, 292
386:
387: \bibitem[Nakajima et al.(1985)]{Naka1985} Nakajima, H., et al.,
388: 1985, \pasj, 37, 163
389:
390: \bibitem[Nakajima et al.(1994)]{Naka1994} Nakajima, H., et al., 1994,
391: in Proc. of the IEEE, 82, 705
392:
393: \bibitem[Nakariakov et al.(1999)]{Naka1999}
394: Nakariakov, V. M., Ofman, L., DeLuca, E. E., Roberts, B., \& Davila, J. M.,
395: 1999, Science, 285, 862
396:
397: \bibitem[Ogawara et al.(1991)]{Oga1991} Ogawara, Y., Takano, T., Kato, T.,
398: Kosugi, T., Tsuneta, S., Watanabe, T., Kondo, I., \& Uchida, Y.,
399: 1991, \solphys, 136, 10
400:
401: \bibitem[Sakurai(1982)]{Saku1982} Sakurai, T., 1982, \solphys, 76, 301
402:
403: \bibitem[Scherrer et al.(1995)]{Sch1995} Scherrer, P. H., et al.,
404: 1995, \solphys, 162, 129
405:
406: \bibitem[Shibasaki, Ishiguro, \& Enome(1979)]{Shi1979}
407: Shibasaki, K., Ishiguro, M., Enome, S., 1979,
408: in Proc. of the Res. Inst. of Atmospherics, Nagoya Univ., 26, 117
409:
410: \bibitem[Tajima, Brunel, \& Sakai(1982)]{Taji1982}
411: Tajima, T., Brunel, F., and Sakai, J., 1982, \apj, 258, L45
412:
413: \bibitem[Tajima et al.(1987)]{Taji1987}Tajima, T., Sakai, J., Nakajima, H.,
414: Kosugi, T., Brunel, F., \& Kundu, M. R., 1987, \apj, 321, 1031
415:
416: \bibitem[Torii et al.(1979)]{Tori1979}Torii, C., Tsukiji, Y., Kobayashi, S.,
417: Yoshimi, N., Tanaka, H., Enome, S., 1979,
418: in Proc. of the Res. Inst. of Atmospherics, Nagoya Univ., 26, 129
419:
420: \bibitem[Tsuneta et al.(1991)]{Tsu1991} Tsuneta, S., et al.,
421: 1991, \solphys, 136, 37
422:
423: \bibitem[Tsuneta \& Naito(1998)]{Tsu1998} Tsuneta, S., and Naito, T.,
424: 1998, \apj, 495, L67
425:
426:
427: \end{thebibliography}
428:
429:
430: %\clearpage
431:
432: \begin{table}
433: \begin{center}
434: Table 1: Physical values of flare loop \\
435:
436: \begin{tabular}{lrr}\tableline\tableline
437: area ($A$) & 2.5$\times 10^{18}$ & [cm$^2$] \\
438: $\;\;$volume ($V$) & 4.0$\times 10^{27}$ & [cm$^3$] \\
439: $\;\;$length ($l$) & 1.6$\times 10^{9}$ & [cm] \\
440: $\;\;$width ($w$) & $\sim$ 6.0$\times 10^{8}$ & [cm] \\
441: temperature ($T$) & 9.4$\times 10^{6}$ & [K] \\
442: number density ($n$) & 4.5$\times 10^{10}$ & [cm$^{-3}$ ] \\
443: magnetic field ($B$) & $\sim$ 300 & [G] \\ \tableline
444: acoustic velocity ($c_s$) & 3.6$\times 10^{2}$ & [km s$^{-1}$] \\
445: Alfv\'{e}n velocity ($c_A$) & 3.1$\times 10^{3}$ & [km s$^{-1}$] \\
446: \tableline
447: acoustic transit time \\
448: $\;\;$along the loop ($\tau_{sl}$) & 44 & [s] \\
449: $\;\;$across the loop ($\tau_{sw}$) & 17 & [s] \\
450: Alfv\'{e}n transit time \\
451: $\;\;$along the loop ($\tau_{Al}$) & 5.1 & [s] \\
452: $\;\;$across the loop ($\tau_{Aw}$) & 1.9 & [s] \\ \tableline
453:
454: \end{tabular}
455: \end{center}
456: \end{table}
457:
458:
459: \clearpage
460:
461: \begin{figure}
462: \plotone{f1.eps}
463: \caption{
464: Temporal evolution of the 1998 November 10 flare.
465: From top to bottom: radio correlation plot observed at 17 GHz with NoRH
466: (scaled arbitrarily);
467: soft X-ray flux in the GOES 1.0 - 8.0 {\AA} channel ({\it dotted line});
468: hard X-ray count rate measured in the M1 band (23 - 33 keV) and
469: L band (14 - 23 keV) of {\it Yohkoh}/HXT (scaled arbitrarily).
470: Four bursts are identified by the numbered top bars.
471: \label{fig1}}
472: \end{figure}
473:
474: \begin{figure}
475: \plotone{f2.eps}
476: \caption{Left: coaligned images of the flare.
477: SXT (Be filter) image is shown by gray scale at 00:14:32 UT.
478: NoRH (17 GHz; {\it gray contour})
479: image of 00:14:37 UT is plotted at brightness temperatures of
480: 150,000, 200,000, 300,000, 400,000, and 500,000 K.
481: HXT (M1 band, 23 - 33 keV) image at 00:14:33 UT is drawn in black contour.
482: Contour levels are 90, 70, 50, 30, and 10 \% of the peak intensity.
483: Right: a simple sketch of this region.
484: The NoRH sources, the HXT source, and SXT faint loop are displayed in
485: light gray, dark gray, and black solid line, respectively.
486: There probably are flare loops (black broken line; see section 3) in
487: microwave source ``B''.
488: \label{fig2}}
489: \end{figure}
490:
491: \begin{figure}
492: \plotone{f3.eps}
493: \caption{
494: Light curves of the second burst (scaled arbitarily).
495: From top to bottom: radio brightness temperature observed at 17 GHz by NoRH
496: ({\it solid line});
497: hard X-ray count rate measured in the M2 band (33 - 53 keV)
498: ({\it dash dot line}) and
499: M1 band (23 - 33 keV) ({\it dotted line}) of {\it Yohkoh}/HXT.
500: The vertical lines show the peak times of the microwave emission.
501: \label{fig3}}
502: \end{figure}
503:
504:
505: \end{document}