The imperfect enumeration is mentioned by Cinnamus and confirmed by Odo de Diogilo apud Ducange ad Cinnamum, with the more precise sum of 900,556. Why must therefore the version and comment suppose the modest and insufficient reckoning of 90,000? Does not Godfrey of Viterbo (Pantheon, p. xix. in Muratori, tom. vii. p. 462) exclaim?
—Numerum si poscere quaeras,
Millia millena militis agmen erat.