I would have to go along with Frantium, and nominate P.
Any metallic element would have shown in the analysis. There was an article in Sci Am about 15 years ago that showed harpsicord strings from the 17th and 18th centuries showed unusually large amounts of P in the alloys; certain characteristics of these alloys made the P actually a desireable element rather than the undesireable one it is today.
Of course, Mr. Busse was careful not to say 'element,' and rather said 'ingredient.' Thus, I hope that he is not going to avoid the advancement of scientific knowledge and indicate that it was 'brilliance,' or 'luck,' or some other quantity not subject to discovery by analysis.
Oh, well; I look forward to the answer with great anticipation. Walt