Sometime before 325 AD, Eusebius of Caesarea compiled his Chronicle, in two books. The second volume exploited the new, large-size, parchment codex, and consisted of page after page of tables of dates and events, synchronising events in different kingdoms, and laying the basis for all subsequent history. Around 380, Jerome came across a copy in Constantinople, and translated it into Latin. A copy of his translation dated to 450 AD is held in the Bodleian Library, Oxford, where I have seen it; and 10 copies exist dated before 1000 AD. Eusebius’ original Greek, unfortunately, did not survive.
The manuscripts split into two families, each based on a 5th century exemplar. These are the group of 4 mss, SANP; and the group of 2, OM. (A list, explaining each letter, can be found here).
In a fascinating paper which deserves to be better known, Alden Mosshammer noticed that OM preserve errors of translation, which were corrected in SANP. One of these requires access to the Greek.
Here’s the first example, (References are to Schoene’s 1956 edition, but you can find these in the online translation fairly easily).
P.217, line 24.
- [Original] ἆθλα μ’ … (nnn ran in the contest for the birthday of Rome …)
- OM = athalamos natali romanae urbis cucurrit (currit M) = “Athalamos ran in the contest…”
- A = XL missus natali romanae urbis cucurrerunt = “40 ran in the contest…”
- PN = quadraginta missus natali romanae urbis cucurrerunt = “40 ran in the contest…”
(I don’t know Greek numerals – what is the original number in Arabic numerals?)
It seems that Jerome dictated the numeral as a proper name, and the scribe wrote it down as one. Somebody corrected it later, but OM preserve the dictation error. Access to the Greek is required to spot this one.
The following example does not require consulting the Greek, and is in fact just a scribal correction:
P.83b, lines 21-23, is a heading. It gives the name of Alcamenes, who was the 9th king of Sparta, and then the years of his reign follow below the heading.
- [Original] = θ‘ Ἀλκαμένης (i.e. “9. Alcamenes”).
- O = thalcamenes
- M = thalcamenis
- A = VIIII menes
- P = VIIII tarcamenes
- N = VIIII tharcamenes
OM think the text reads “Thalcamenes”. But the copyist SANP realised that the first letter was actually the number 9, although they still didn’t get the name right. Possibly they realised this, because all the kings have numbers, so they inserted “VIIII” (i.e. “IX”) in front.
Here I have a little personal experience to contribute.
Scribes copied the names the first, and worked down the columns, rather than across. When I transcribed the chronicle, I found that this was much the quickest and safest way. The only problem was that you might write too many numerals, and suddenly realise that after year 9 there is a new king! In the Bodleian ms (O), indeed, you see erasures of just this kind. In HTML, luckily, I could just go back.
So the scribe will have quickly realised that a numeral was missing, and added it; although he could not determine the correct spelling of the name. This correction could have occurred at any time, tho.
Mosshammer gives only these examples, and a couple of others which do not bear on this question.
Numerals in Greek are vulnerable things. The first example proves that even St. Jerome could be foxed. In this case, the lists of unfamiliar names, preceded by numerals, were a perfect occasion for error.