Alan:
It would be informative to see raw data for all your calibration rides to see how well you are able to duplicate final rim readings under ideal conditions.
A certifier reviewing a submission can get a good idea of correct operation of an electronic counter from calibration data. Also, if the two rides of the course differed by much less than 1 rev/5 km, correct operation would be likely. Previous demonstration that a measurer had got identical results with a Jones should also be reassuring. Note a certain amount of trust in the measurer is necessary even if a Jones is used.
Duane is correct in pointing out that counters have a limit to the number of counts/sec they can handle. As Pete has recently pointed out, it is perfectly possible to exceed the rating for the Jones counter on a bicycle. (This may be very conservative.) The Protégé has one of the best ratings I have seen for an electronic counter because I have not exceeded it with the four-magnet method during normal certification, the two-magnet method at 35 mph, and the single-magnet method at 45 mph. Moreover when counters exceed their rating many counts are lost not just one. Therefore you cannot postulate a lost count by going too fast on a downhill.
The temperature of 46 deg is unlikely to have been a problem unless maybe you have a very poor battery. Cyclocomputers are rated down to 32 deg, but I do not know how consevative this is because I have not tested below this temperature.