W is the awkward letter of the set. It was not a letter of the Latin alphabet the cipher was built on, and in the original chart it stands after Z rather than in alphabetical position, as a recent addition. Later copyists appear to have misread that presentation, and the printed sign for W differs between sources. Which of the circulating forms is faithful to the 1518 chart has never been established, so the atlas cannot say which sign for W it is showing.
What would resolve this A collation of the surviving copies and early impressions of Polygraphiae against the Theban charts in Agrippa's De occulta philosophia and the later grimoire reprints, reading the 1518 sign for W from the earliest surviving impressions directly rather than from any redrawing, and tracing where the divergent forms enter the tradition.
OpenAnalytical bibliography and description of early printed books, palaeographyWikipedia contributors, Theban alphabet (reference article)