Another point is the thing Starlight mentioned about patents and drugs of abuse. Lego doubts that there is a special unwritten rule to add an error knowingly in order to confuse the illicit cook. The abusive potential of a substance is usually discovered after its commercial breakthrough. Therefore Lego assumes that the errors in 'drugs of abuse' patents are not different than in other chemical patents.
I think the existance of such a "rule" might be true for patented improvements of older methods, i.e. newer patents - its probably the reason why older patents usually work better than newer ones - back then, when they were applied, the substance in question usually wasn't already known to be a drug - so no errors were put in...
(with newer patents, I can very well imagine they "modify" the procedure with the intention to prevent use in drug manufacture - if the inventor knows that the compound in question may related to illicit drug synthesis, or better said if a previous patent is well known for its use in clandestine chemistry, the inventor of a new procedure might want to avoid such use of "his" procedure, by confusing the clandestine chemists a bit..)
Take for example methylamine production: I BET every new patent about its synthesis is subject of DEA investigation, due to MeAm's use in meth production...
But I also bet there was NO investigation whatsoever back in 1928, when this procedure Patent DE468895 (http://l2.espacenet.com/dips/viewer?PN=DE468895&CY=gb&LG=en&DB=EPD)
was patented - there was no DEA back then... ;)
Greetz A