Since Annibaldi's diplomatic edition (Cesare Annibaldi, La Germania di Cornelio Tacito nel ms. Latino n. 8 della biblioteca del conte G. Balleani in Jesi. Edizione diplomatica. Critica a cura di Cesare Annibaldi, Leipzig 1910, p. 27; cf. already Cesare Annibaldi, L' Agricola e la Germania di Cornelio Tacito nel MS. latino n. 8 della biblioteca del Conte G. Balleani in Iesi. Citta di Castello 1907, p. 155) of the manuscript, generally called E (back then Jesi, Biblioteca del Conte Balleani, MS. Lat. n° 8, now Roma, Biblioteca Nazionale Centrale Vittorio Emanuele II, Ms.Vitt.Em.1631) the gloss to Tacitus' Germania, c. 1,2 Arnobę is unanimously read as arbonę ł norie.
Of these forms, Arnobę is the closest to the rightly supposed original Abnobae (which only occurs as a scholarly form in Ms. h; for the readings of this passage in the manuscripts cf. still Rodney Potter Robinson, The Germania of Tacitus, Middletown, Connecticut 1935, p. 104f.). The mistake traces back "auf Verlesung der zugrunde liegenden Majuskelvorlage" (Rudolf Till, Handschriftliche Untersuchungen zu Tacitus Agricola und Germania, mit einer Photokopie des Codex Aesinas. Berlin-Dahlem 1943, p. 89). The generally accepted explanation of the reading arbonę is the following one: "Dem Korrektor schwebte wahrscheinlich der Ort Arbon bei St. Gallen am Bodensee vor, der als Überfahrtsort der Pilger bekannt war" (Till 1943: 89). For norie, no solution has been found up to now, cf. Till 1943: 89: "Seinen Vorschlag noriae vermag ich nicht zu deuten."
In the last few days I became more and more suspicious of the reading norie, thinking that it should be read as none. I was convinced that the sign that was read until now as ri is indeed the letter n, when I found an exact parallel form in c. 10,1: prohibuerunt. And none would indeed make much more sense because it is the mixed form between Arnobę and arbonę combining the two n-signs. So the gloss seems to be meant to be read as arbonę ł (ar)none. This last form was in my opinion not made up by the medieval corrector but rather seems to date from the 15th century.
But how to proof this assumption? Then it came to my mind that there is one manuscript, that is normally completely neglected, simply because it is just a transcript of E and therefore offers no additional information concerning the text form of Tacitus' Germania (cf. e.g. Jacques Perret, Tacite, La Germanie, Paris 1949, p. 62: "Nous avons négligé T dont Annibaldi ... , et Robinson ... nous semblent avoir démontré qu'il est une copie de E"). This is the manuscript T (Toledo, Biblioteca Capitular 49.2). Luckily there is a collation of this manuscript that was already published in 1903 by Frank Frost Abbott, The Toledo Manuscript of the Germania of Tacitus, with Notes on a Pliny Manuscript (Chicago 1903). Here, on p. 6 the relevant passage is given as: "Arnobe (ał Arbone ał none in margine)". So the person, who wrote the manuscript T by copying it from E, read the norie of modern times as none.
The conclusion therefore is that the gloss in Ms. E should be read as arbonę ł none.