MacKinnon

Mechanisms: ei bs (plus ee bs); synchrotron; IC 
 - discussion of Korchak (1967) figure comparing spectra
McClements-Brown (1986) and Kaufmann (1986)
Conclusion: BS dominates except in possible special cases
Evidence for relativistic electrons: > 1 MeV
  - Microwave
  - > 10 MeV X-rays
  - Pi-zero flattening
  - SEP relativistic electrsons (Posner-Droge recent work?)
Grant May 24, 1990 spectra
Vlahos: Do flares make MeV electrons?
  - yes, both early and late
Pion gamma-ray overview
Synchrotron overview - Silva-Share work on THz in preprint form
IC overview - plot showing basic spectrum (mention of Riemann zeta function)
  - a wonderful formula for I_ic/I_bs
  - in principle one might upscatter optical photons: bright patches of 
    coronal X-ray emission?
So: no synchrotron contributions, but a possible window of opportunity for 
  inverse Compton for a hard electron distribution, with directivity 
Discussion of applications...

Battaglia

Wheatland-Melrose 95:
Simple model calculations showing relationship of thick and thin with
  index change of 2 (steady state, no PA distribution, etc., etc.)
RHESSI observations with "Case A" and "Case B" depending on FP/LT spectral
  difference
MacKinnon: diffusion effects important below 10*kT (Emslie paper was not
  correct)
Competition in Case A with thermal emission
Poor agreement with model suggests modifications
  1) electron beam modified non-collisionally? e.g., return-current field?
  2) misunderstanding of density/critical energy?
Vlahos: there are other possibilities already in the literature
Fletcher: What about Valentina's work?
  - Observationally inconsistent
Hudson: SHS in both high and low-energy ranges?
Fletcher: perhaps we can conclude that the coronal sources really do
  differ from the impulsive-phase sources
Krucker: but be careful about systematic errors in the fitting
Improvements in W-M type of model can produce a better reconciliation
  with data (just add an E-field) in a steady-state model

Masuda

Survey for coronal HXR sources from Yohkoh/HXT
April 5 2001 Nitta event, nice composite movie
  General discussion, how to relate this to RHESSI framework, 13 degrees
Superhot thermal: 30-50 MK, Kosugi et al. 1994 6-Feb-92
  Not occulted (Tomczak). Suggest the two peaks come from different
  locations (White) or from Veronig-Brown dense corona (Hudson)?
Nitta et al. 2001 analysis of motions
Tsuneta analysis of Masuda flare
Tsuneta analysis of 21-Feb flare

Tomczak

  - caveats
Development of Yohkoh HXT flare catalog 
  (dx.doi.org/10.1007/s112-7-006-1831-5 1 GB)
Examination of catalog for occulted events: 98, or 7.6% of catalog
  - centroids above limb
  - limb appears to cut SXR (bias against impulsive SXR brightenings)
  - longitude > 90 degrees
Longitudes by two methods: extrapolation of H-alpha locations, spot records
  via 13.45 - 3*sin^2
Example of 8-May-98 event and region
Plot of gamma from M1/L ratio vs apparent longitude for 94 events
  - extreme case, both hardness and longitude, 18-apr-01!
Remove exceptions (no maximum or ambiguous) => 52 events
  - only 3 events for which HH/M2 ratio available 
Gamma distribution strongly shows soft coronal sources (6.7+-1.4 vs 4.0+-1.1)
  in M2/M1
Gamma distribution for HH/M2 much more comparable, but only 3 occulted events
Gamma vs M1 counts scattergram, showing occulted and non-occulted events
M2/M1 vs M1/LO color-color
Summary statistics suggest inverse Compton!?
Centroid shifts vs height, also SXT vs HXT/LO 
Closer looks at 3 H-channel events:
  - 18-apr-01
  - 23-apr-98 (Tomczak 03, 04, 05; Sato 01). Type C?
  -  9-may-98 nice SHS+SHH development at high energies, but not lower.
    Thermal interpretation of HXT at lower energies is difficult
More on 9 May
  Nice hard X-ray ejection seen in movies, but < 100 km/s
  Nice graphic of SXT ejecta, starting earlier than HZXR
Another possibility: 27-feb-94
Another possibility: 1-apr-01
Conclude that HXT also could see SHH category ("type C")
  - often masked by other emissions
  - close connection with ejections and SEPs
White: so we're happy with thin target in the corona
Krucker: consistent with RHESSI data, note energy correlation
Hudson: trapping time scale determines limit

Vlahos discussion on theory

Defining the problem: 
  - location within the loop or above it?
  - motion (only ever up?)
  - energetics, thin & thick target
  - sizes
  - number of particles
  - phase of flare
  - correlations with CME, microaves...
  - height
  - spectra
  - SHS in corona and in footpoints
  - time profiles
  - co-spatiality of source and accelerator?
Theory summary
  - Fletcher-Martens paper
  - high density (10^10)
  - energetics - need 10^28 ergs/s > 25 keV
  - current-sheet properties
  - evacuation of corona via collapsing trap?
  - populations of electrons the same?
  - transport of particles not a negligible problem
  - distribution function and its evolution ditto
Theory of loop-top coronal sources
  - collapsing traps 
  - injection of particles needed 
  - combination of collapsing trap and turbulent current sheet?
  - Petrosian-Donaghy model puts the accelerator at the loop top
  (Discussion of Veronig-Brown event)
  (more " ")
A better idea
  - "whistler turbulence"
  - stochastic E-fields, rather than super dense
  - self-sustain acceleration and trapping from "statistically interesting
    stressed fields" (Rim Turkmani)
  (much discussion)
  (discussion of stochastic trapping vs turbulent trapping)
  (Fletcher: preflare mass from filament? Much discussion)
Above-the-loop-top sources
  Masuda source being different...
  Drake et al. or Onofri et al.: behind the CME, large current sheet 
    collapsing to small-scale ones => Masuda source?
Opinions
  AR has multiple loops
  preflare stressing can accelerate particles
  Masuda source from local trapping
Late phase
  maybe distinct from CS(s)?
  rareness of these sources may be due to faintness