The scope was told to read the first few hundred microseconds of data output. It appears to be similar to the observations made by blargg in the thread about polling. Other, not pictured, tests confirmed that the crystal was oscillating at 12MHz and that the filtered 5V input line was very stable.
Does anyone have any interpretations of this? I assume that this suggests that there are no problems with the circuit being used. Any problems are presumably in the coding.
I have been unable to address a previous observation that the report was not matching up with the expected report length. This was mentioned when running through USBlyzer, but my trial has since run out. I can investigate that again when I resolve the trial expiration issue.
Suggestions welcome.