According to Russian Opera Singers 1750 - 1917, vol 2 book, Alexander Rosanov recorded for Beka in 1906 and 1907, for Odeon in 1912, and for Anker (no year provided) solely in Moskov.
Sorry, but none of them says that it is recorded in Berlin on the label...
This one says that it is recorded in Moscow and it matches data from the Pruzhansky's book.
Sorry, but the label speaks of Moscow rather as place of the artist's origin than of recording.
Concerning his Ankers:
Matrixes 1016 (Arelie Révy) and 1037 (Alfred Herbert) were both undoubtedly recorded in Berlin. Between lies Rosanov. It is very unlikely, the Anker engeneers went to Russia to record only 21 sides...
I support Raoul's opinion that this disc was recorded in September 1906 in Berlin. Recording engineer was Willy Bielefeld. No Russian recording tour is reported for Beka during this year. Before you ask, St. Petersburg was visited in June 1905, not 1906!
By the way it wasn't entirely uncommon for the Gramophone Co. to name the artist's residence instead of the place of recording. For example, Wiesbaden is named on recordings of the violinist Hermann Irmer, although his records were taken in Berlin and Frankfurt on the Main.