A later article (1892)[32] was more conclusive, as here he attempted to prove that no one forger could have adopted the variety of attitude towards both the Senate and Christianity which we find expressed in the various sections of the “lives,” while the presence of geographical names and official titles, lost before the beginning of the fourth century, point to earlier authenticity, not later forgery. Woelfflin[33] in 1891 supported Mommsen on textual grounds. He traces the differences of style to the fact that certain authors had used Suetonius, others Maximus, while others again had trusted to their own retentive memories, not altogether a safe historical criterion. He states that the traces of similarity running through the works are due certainly to a reviser, but that the reviser was Vopiscus,[34] which either puts Vopiscus at a much later date than had ever been done before, or resigns the idea of a late reviser in the Mommsen sense. Dessau[35] in 1892 replied with a scathing attack on this same Vopiscus, from the point of view of his age and the impossibility of his having seen and heard all he claims to have done. Seeck[36] in 1894 published a second article supporting Dessau with six points culled from titles and names not known till after the reputed dates of the Scriptores. He now considers that plurality of authors, or forgers, as the case may be, is certain, and that they wrote, or forged, as Diocletian and Constantine gave command, using for their work many sources, including the Imperial Chronicle. But it is an inconclusive article. In 1899 an American, Dr. Drake[37] of Michigan, published some studies in detail on the life of Caracalla, which tended to establish the genuineness of certain portions which had been thought spurious. Heer[38] of Leipzig followed in 1901 with a critical survey of the life of Commodus, dividing it into two parts, the first chronological, the second biographical, and came to the conclusion that, though the chronological part was trustworthy, the biography was derived from very poor sources, and was only in part contemporaneous. Schulz[39] in 1903 applied the same methods to the lives from Commodus to Caracalla, in 1904 to the life of Hadrian,[40] and in 1907 to the lives of the house of Antonine,[41] unfortunately leaving out Elagabalus.