I've just spent a couple of hours of unbridled excitement (?) measuring inlet tracts on 67-1066 cylinder heads. What I've found seems to confirm what I suspected. 1) all square type iron heads, as used on sw. arm models and the last of the plungers, are marked 67-1066 though some markings are difficult to read. 2) These things were mass produced, rough cast, then finished off on some fairly inaccurate tooling so there are some variations but the heads fall into two distinct categories a) inlet tracts measuring somewhere between 26.5 and 27 mm and b) inlet tracts measuring in excess of 28 mm, typically about 28.3 or 28.4 mm. There are variations in the castings and shape and profile of the inlet tracts but nothing consistent which suggests, unsurprisingly, that there were a number of moulds and processes used at the same time and/or over a period of time. I have only come across one rogue head which seems to fall between the two categories; I suppose this may have been modified by a P.O. or it could be one of BSA's Friday afternoon jobs. The general difference between the sizes is about 1.5mm which is equivalent to 1/16". This could account for a number of fine tuning issues which have been attributed to carburettors when the actual problem is a slight mismatch of inlet tract to carb. The valves are all the same. Problems with cylinder heads and carbs were often resolved, back in the day when these bikes were worth nowt, by simply swapping items from the spares box under the bench. I suspect many of the lads fixing their bikes in the coal shed back in the sixties might not have measured inlet tracts assuming all the heads were the same, or even if they had spotted a slight variation would have attributed it to sloppy machining.