If the air screws on the carbs are all set at about 1 turn out...then, two things come to mind: first, try a different condensor on the 1-4 points. A weak condensor will make the coil charge poorly (I'm presuming points-and-coils, here, no electronics involved).
Second:
You might try a different coil, if you can nab one or borrow it. While this is rare, I have seen it on several (3) bikes (including my own brother's, in 1979). The story behind it:
If there is a break in the coil wire inside the high-voltage side, the plug will spark when it is laid against the head, even on both plugs (although the weak one will be more yellow than blue). Usually this break is in the form of a carbon trace inside the coil where the wire has eroded away from the ends where the break is, so the ohmmeter tests for it are inconclusive. In action, the break is usually more toward one "side" of the coil than the other, and the "good" side gets a spark started, draining the coils' magnetic energies, then the "bad" side suffers less energy and poor spark, especially in the RPM range of 0-2000 and 5500+ RPM. It often runs in the midrange, until the coil gets hot, then quits altogether until the coil cools again.
I have one of these rare failure coils here as a curiosity. I used it to test my Transistor Ignition during the design phase, to make sure it would not damage the electronics. It generates enormous back-kick voltage to the points. On his bike, it ate 3 condensors in 200 miles during a very hot trip across Kansas in 1979, while #4 would not fire unless the engine was stone cold. It would fire for a minute or so on startup, then only run at midrange RPM, until hot, then it would stop #4 completely, leaving a wet sparkplug (not fouled, just wet).
