This I believe.
Certainly books on exhaust theory from the 1950s that I was aware of before getting involved with Triumph 6s were making the point that for effective scavenging the paired primaries needed to be on opposite points of the combustion cycle. So when I acquired a copy of GTs book his point about 6-3-1 manifolds seemed rather obvious to me. And caused me much frustration when trying to find a commerically available 6-3-1.
Mind you a good number of the 4-2-1 manifolds seem to be made without much consideration of getting equal length primaries/secondaries even though this is relatively straightforward on the 4s.