We could probably agree to split the difference. In practice I guess it is somewhere in between these two bounds. Either way it is close enough to monochromatic by any reasonable definition.
I did a quick back of the envelope sum using NIST data and reckon there are about 5 visible groups of lines in the sodium streetlamp spectrum that are around 0.5% of peak each plus the red lines from the start gas.
He doesn't have enough horizontal resolution in wavelength to show the gap, but I think both D lines are plotted adjacent with the 2:1 ratio visible. It is still quite a fun educational page. A log/lin switch would be nice.
I dunno about that. With the sodium line saturated I get very visible colour for the green line at ~567nm, weak colour for the blue doublets and also some red lines from the initial cold strike gas.
We are in complete agreement on this point. Worst case using my pessimistic estimate 95% pure and if you are right better than 99%.
Regards, Martin Brown