Moore et al. (2001) have reported observing bursts of high energy radiation associated produced in a 1 to 2 ms interval just before the start of the return stroke and as the stepped leader was nearing the ground.  Dwyer et al. (2003) observed bursts of energetic radiation in the last 160 μs of the dart leader and possibly right at the start of the return stroke in 31 out of 37 triggered events studied.  The shorter interval may be because dart leaders have a higher propagation speed.  My understanding is that the sensors used in both these two experiments were unable to distinguish between energetic electrons, X-rays, and gamma rays.

A NaI (sodium iodide) scintillation detector was used in both experiments.  Light is emitted when the high-energy radiation strikes the NaI.  The weak light signal is then detected and converted into an electrical current using a photomultiplier tube (source of the figure below)



A sketch of the signal produced by the scintillation detector used by Dwyer et al.,  when exposed to a Cesium-137 γ-ray source, is shown below (the control sensor was identical except it did not include the NaI scintillator). 


The exponential decay following the peak is the response of the preamplifier circuit. 

The figure below is a sketch of high energy radiation produced by a triggered lightning discharge from the Dwyer et al. (2003) paper.


The output of the scintillation detector is shown at top.  This signal began 160 μs before the start of the return stroke (the return stroke began at time 0 in the figure).  The complex signal shape during the rise to peak indicates that several energetic particles were detected.  A very small pulse is just visible on the control signal trace and may have been caused by energetic particles directly striking the photocathode of the photomuliplier tube.  The return stroke current was measured at the strike point and rose to a peak value of about 22.5 kA.  The electric field was measured at a point 260 m away from the strike point and has the characteristic asymmetric V-shape described by Rubenstein et al. (1995) for the leader-return stroke transition observed at close range.

An improved sensor was used during the summer 2003 campaign and Dwyer et al. (2004) report X-rays were measured 0 to 80 μs prior to and at the beginning of 73% of triggered return strokes studied.  Each X-ray burst usually lasted less than 1 μs.  The most intense bursts come from parts of the channel that is within 50 m of the ground.  A sketch of one of the recorded signals is shown below.


Energies for most of the x-rays that were detected are shown in the publication, some of the values are included in the sketch above (in keV).  Note how the amplitudes of the x-rays in this and the next figure increase as the leader (a dart leader in this case) nears the ground.

Dwyer et al. (2005) has measured X-ray emissions coming from the stepped leader process in natural cloud to ground discharges.  The last 11 steps of the stepped leader are shown in the sketch below (the start of the return stroke is at time = 0).



The lightning struck about 50 m from the electric field derivative antenna (the E field signal above is an integration of the measured dE/dt signal) and about 260 m from the x-ray detector. 

It seems clear that the x-rays are produced during the stepping process.   Because of the similarity between x-rays produced by dart leaders in triggered lightning and stepped leaders in natural lightning Dwyer et al. (2005) suggests that the production mechanisms are similar and that dart leaders also step, but with a frequency that isn't resolved on optical or field records.

Implications
Emissions from dart leaders in triggered lightning and stepped leaders in natural lightning are similar.  This suggests some similarities in the discharge process (dart leaders may actually step

Observations may provide some clues about leader propagation processes.

The standard "relativistic runaway electron avalanche model" might have some trouble trying to explain the lightning generated X-ray emissions.






And finally something that appeared in the local newspaper yesterday.