Some Types of Error in Manuscripts of Aeschylus' "Oresteia" Young, Douglas Greek, Roman and Byzantine Studies; Summer 1964; 5, 2; ProQuest pg. 85

Some Types of Error in Manuscripts of 
Aeschylus' Oresteia 

Douglas Young 

W HILE conducting a seminar in Greek palaeography in the 
University of Minnesota I happened to illustrate some of 
the corruptions that occur in transmission by specimens 

taken from the Oresteia, which I was engaged at odd moments in 
turning into English verse. It having been suggested by colleagues 
elsewhere that my collection of examples might be of wider interest, 
I revised it and added a few comments. I have not aimed to set down 
every instance of every type of corruption: for example, I take here 
no account of errors in the ascription of speakers, arguably a field in 
which manuscript authority is worthless; and I have usually neglected 
singling or doubling of '\, IL, v, p, and a, and confusion of vowels 
and diphthongs, such as of 0, w, ov j 10, Tj, E"L, exL, OL, v, and L. 

My main sources for variants have been the editions of Murray 
(1955), Headlam-Thomson, and Groeneboom; and I limited my 
interest to the manuscripts M, V, F, and Tri, agreeing substantially 
with the evaluation of them by Fraenkel in the prolegomena to his 
Agamemnon. I did not have available complete facsimiles of all the 
extant manuscripts of the Oresteia, thorough collation of which would 
be needed if one were to attempt a rigorous quantification of the 
varying percentages of different types of error. But a general conclusion 
emerges, that errors involving more than one letter or one syllable 
are relatively a trifling proportion of the total of errors. This con­
clusion should be stressed, in view of the fact that very many so­
called emendations published involve changes of several letters, 
syllables, or even words, and all too often fall into the category of 
what Professor W. L. Lorimer terms C<immendations." Most of the 
innovations found in Tri, the holograph of Demetrius Triclinius, are 
such Himmendations," often motivated by his metrical notions; and 
I have not listed all of them; nor have I paid exhaustive attention to 

variants found or implied in the scholia. 
For convenience, besides the usual sigla, I denote Triclinius' holo-

85 



86 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

graph (Murray's Tri) by T. I abbreviate the plays as A (= Agamemnon), 
X (= Choephoroi), and E (= Eumenides). The assumed genuine reading 
precedes the bracket. 

Some confusions of letters in the extant manuscripts M V F T, or 
some of them, appear to derive from the uncial stage of transmission. 
Thus there are confusions involving the round uncial letters epsilon E, 
theta e, omicron 0, and sigma C. A useful mnemonic for this group is 
the word EOos. Examples: A 1655 O€pos] 0 Epws. X 56 cpP€vos] cpP€V€s. 
X 71 OLyov'nJ otYOVT£. X 74 tOvuav] lovuav. X 374 CPWV€LS· StJvauaL] CPWV€L 

dSvvauaL. X 438 dAo{fJ-av] ~Ao{fJ-av. X 718 f30VA€vuofJ-€uOa ] f30VA€VofJ-€Oa. 

E 46 AOXOS ] A€xos. E 137 uv 15' ] ouS'. E 450 V€OO7JAOVJ Vo07JAOV M, dOv€lov 
Fr. Here round epsilon has been dropped by a near-haplography in the 
proximity of omicron. The initial nu has been attached in later Mss to 
the end of the preceding verbal inflection, KaOaLfJ-agwuL (-OVULV FT). 

Some words show confusion of uncial forms of alpha, delta, lambda, 
mu, and nu, A 11 A M N. A 1014 "has] ALas F. A 1291 TauS' €ycfJ ] Tas 

A€yw. A 1418 a-T}fJ-aTWV ] A'1JfJ-fJ-aTwv. X 45 fJ-' laA)..€L ] fJ-LA)..€L. X 252 ' HA€I<­
Tpav A€yw, with loss of 1\ after N, or possibly after -AN in the form A 
with overstroke for nu. X 424 laA€fJ-LuTplasJ l)..€fJ-LuTplas. X 474 SL' clJfJ-av 
EPW ] aLwfJ-avULp€w. Here confusion of vocalisation accompanies the 
graphic confusion of uncial delta and alpha. X 566 S€gaLT' ] AlgaLT'. 

E 54 Alav (in the form I\IA) may be the original reading, leading to 
S{a M, whence f3{av of F and T would be a mere conjecture of some 
savant of the Palaeologan renaissance. E 938 7TV€OL ] 7TA€OL FT. 

Gamma and tau were liable to confusion in uncials. X 48 )..tJTPOV ] 

)"vypav. X 137 fJ-€ya ] fJ-€Ta. X 353 yasJ Tas. X 399 raJ Ta. E 398 yijv] 7"1Jv. 

A 768 Sa{fJ-ova T€ Tav of Mss may be derived from an original SalfJ-ova 

T€YWV, "a demon of the house." 
What some find odd is the occasional emergence of a kappa from 

the misreading of uncial iota and sigma juxtaposed. This certainly 
occurred at X 897, where M's clJKV must be from an uncial form of 
<[J UV, OICV. So too E 177 €lULV 00] €I<€lvov. E 862 lOptJ07]LS "Ap'1J ] lOptJ07]L 

Kap'1J Ml ] lSpV07]L Kapa M2FT. At A 985 a brilliant insight by Professor 
Denys Page shows us how F's ifiafJ-fJ-las aKaTa derived from an original 

a,s 
ifiafJ-fJ-{aLS. aTa through an assumed intermediate stage ifiafJ-fJ-las aTa. 

At E 119, for tcf>l)..OLS yap €luw OUK EfLoLS 7TpOU{KTOP€S, I would have 
Klytaimestra's ghost say, cf>lAOLS yap €luw OUK EfJ-oLS 7TpOS iUTopas, "With 
friends not mine he is going to judges" (cf 81, oLKauTas ••• €Vp7Juo-



DOUGLAS YOUNG 87 

IUV). She had seen Orestes go out with Apollo and Hermes. npoclC­
TOPAC became npOCKTOPAC and, with EICIN taken as from €lp..t 

sum, a nominative 7TpOULK'TOP€~. 
Uncial gamma, carelessly written, was apt to be confused with the 

round uncial sigma. Thus, at A 101, the original as ava~alVEt~ postulated 
by Ahrens could, with misdivision, give rise to M's ayava cpalv€ts. 

X 542 uVYK6'\'\w~ ] uvuK6'\'\w~. E 58 if 'Tt~ ala] if'Tt yata, where there is 
also some glossing mentality at work. 

To the uncial period belong the confusions of round sigma and 
iota. X 183 Kap8{aL ] Kap8{a~. X 519 p..EtW] p..EUW. X 691 EJl7Ta~ WS ofM may 
have resulted from the adverb ip..7TCtLW~, formed from the adjective 
used at A 187, with internal correption. 

Confusion of gamma and pi is an uncial error, as at X 835 Avyp8sJ 

AV7Tp8s. Confusion of pi and the juxtaposition of iota and tau (either 
way) is more likely to be uncial than minuscule. There is a curious 
instance in the scholia at A 186, p..a.V'TtV ov'Ttva !fEYWV, where the scholiast 
remarks: 7TEptUUEVEL 'TO 7rVEvp..a. Clearly someone had read the com­
pendium 7Tva for 'Ttva. That is to say, tau+ iota was read as pi. The same 
mistake may lie behind X 958, where M offers Kpa'T€L'TaL 7TWS. The 
omega would originally have been an 0 simpliciter, which could also be 
interpreted as ov. Assuming the pi derives from tau+ iota, and re­
dividing, we get 95Sf thus: Kpa'T€L 'T' al'TLOVS 'T6 ()EtOV 7Tapd. 'T6 p..~ I 
lJ7TovPY€I.V KCtKOl.S, meaning: "And the divine (power) masters guilty 
persons by not subserving evils." In dochmiacs exact responsion is 
not required, and, with internal correption of ()ELOV, we get a doch­
miac in the form v v - v v v - • 

At E 1044 a misreading of iota+ tau as pi could have led to M's 
U7TOVOa~ 0' is 'TO mxv EvOCtLO€~ OZKWV. I suggest Aeschylus wrote U7TOVSij. S' 
eLUt'T' av' EVOatO' otKOV: "With a libation enter in, along the torch­
filled dwelling." He would write this in the form ~nON.ilAID.EL 
ITANEND.AID.OIKON. IT, misread as n, led from ELITAN, via 
EillAN, to the common £S 'T6 miv. u7TOVOij., dative singular, written 
illON~AI, was misinterpreted as the nominative plural u7TovoaL, 
made subject of a sentence with the relevant part of the verb to be 
supplied mentally, viZ. €law. Then the adjective in the nominative 
plural EVSCttO€S was evolved from END.AI.il which originally stood 
for EvoaLo(a). OIKON, originally meaning OtKOV, was made into a 
genitive plural OLKWV, depending on the new subject u7TovOat. A 
relatively small number of corruptions involve more than a 



88 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

one-stage evolution, as this one does; but the original cause of error 
was a simple graphical confusion. 

Turning to errors arising from graphical confusions in minuscules 
of various dates, one may note the occasional confusion of a form of 
beta with a form of kappa. Thus A 889 fl)vxflas ] KAafla~ F. X 936 flapvFH­
KO~ ] KapvoLKo~. E 110 vEflpoiJ ] VEKpOiJ FT. E 246 VEflpovJ VEKpOV. A 102.4 

&flAaflE{f[- ] aVAaflE{a F shows the reinforcement of the graphical 
confusion by a Byzantine assimilation in pronouncing the diphthong, 
with the upsilon consonantalized. 

Beta is sometimes found for mu, as at A 1420, luaajLarwvJ flLaajLarwv 
in G, a manuscript I am not here normally citing. 

Gamma occurs for delta at A 310 rooE ] rOYE M and E 75200' ] 0 y' M. 
But here it may be mere confusion of common particles. At X 989 M 
offers ,pEYW and L Myw; but I suspect Aeschylus may have written 
Aly{a8ov yap ov ,pEOW jLopOV, cf ,pEOELV . EvrpE7TELV, cppovr{~ELv in Hesychios: 
"I do not care about Aigisthos's doom." The scholiast's Myw could 
mean "I do not reckon in ... "; but so common a word is little likely 
to have been corrupted to M's ,pEyW. In minuscules the high 
gamma sometimes has a loop at the foot which makes it very 
like a delta of which the lower part is skimped and the flourish 
above is drawn to the right. But this tendency is hardly evi­
denced before the date of M, around A.D. 1000; and it may be we 
have here merely a substitution for a rare word of a commoner 
one, itself in turn supplanted in the scholia by a very common 
one. 

X 530 V€oYEVE~ ] VEOPEVE~ suggests that M's minuscule antigraph had 
a blotchily written gamma looking like a rho. 

Theta is lost after phi at A 1187 aVjLcp8oyyo~ ] aVjLcpoyyo~ F and E 371 

E7TLcp86vOL~ ] E7TLcp6vOL~. 
Theta develops into rho at A 919 flapflapov ] flapfla80v in p1 (and 

El, which I usually neglect here). 
Theta is deaspirated to tau at A 946 EjLfla{vov8' aAovpYEaLV ] EjLfla{vovr' 

&AOVPyEaLV. 
One would expect interchange of theta and delta, as possibly at 

A 1089, where we find ,pv()Yj emerging as ,pv8Yj in T; but as the form 
,pvoYj occurs at A 999A one cannot be sure that Triclinius was not 
merely conforming to that earlier place. 

At A 1595, for the Mss' &vopaKa~ Ka()YjjLEvo:; Professor A. J. Beattie 
has a brilliant, as yet unpublished suggestion, ctv8paKas Ka()' ~jLjLEVOV~I' 



DOUGLAS YOUNG 89 

"around kindled coals" (burning charcoal), which would involve 
confusion of theta and delta in Byzantine pronunciation. 

Kappa is liable to confusion with chi, as at X 35 EI..aK€ ] EI..aX€. X 39 
~I..aKOV ] D .. uxov. X 180 XU{TYJV ] KU~ T~V. X 215 E;YJVXOVJ E;YJVKOV. E 170 

/l-VXOV ] JLVKOV M, where F and T displace the word with the gloss 
uov olKov. Confusion of K/X, as of 7T/cP, is an "ear" mistake. 

Mu sometimes develops to lambda, as at E 881 KU/l-0V/l-at ] KUI..OV/l-UL 

FT. Maybe this happened at X 814, where I would read EV>">"d.f30L 
'0' , '0 I - < M I , .. ,/.. I -C' I 8 I \ \' '0' o €VOLKWS 7TaLS 0 aLas, E7TEL ,/-,opWTaTOS 7TpaSLV ovpLav E/l-€V' 7TOl\l\a 0 

u>..>..a cfoaVEL xpf! (WV KPV7TT', ... "Let Maia's son duly take a hand, for 
he is most furthering to make an operation favoured (by wind); 
and many things else he will show forth, at his will, though hid ... " 

w 

M offers O€I..EV, the suprascript omega meaning that the reading 
should be OEI..WV, doubtless influenced by xpf!(wv below in line 815. 

Mu seems to have given rise to pi at A 1255, in F's Svu7TaOij for T's 
Svu/l-u8ij. But Verrall's hapax SVU7TVOij deserves consideration with 
reference to the foregoing 7TvOoKpavTa. 

Graphical confusion of minuscule nu (the type not "on a leg") and 
upsilon may occur, as at A 529, where F has TO£OVS€ (sic) for TOLovDE. 

So too E 77, 7TOVTOVJ 7TOVTOV. E 136 aVT{K€VTPUJ aVT{K€VTpa F. E 670 
Xpovov ] Xpovov MI. 

Pi and phi are liable to interchange. Thus X 418 cfoavTES ] 7Td.JJTES. 

E 523 avaTpE7Twv J avaTpEcfowv. I read 522-525: 

Tls SE J-tTJDEV EV cfoa€L (= EV f3lq;) 

KupSlav avaTpE7Twv, (= cfoo{3ov/l-€vos) 
" 1\ f3 I 8' < I TJ 7TOl\LS POTOS 0JLOL-

" ", IQ A I WS, €T av U€jJOL L.J LKav ; 

Comparing such phrases as Theocritus 8.90, aV€Tpa7Tero cfopEva I..V7TCf, I 
would render this: "Who that not at all in life upsets his heart (= has 
his heart upset, gets terrified)-either a city or a human likewise,­
would still reverence Justice?" 

Confusion of pi and tau is more likely to occur in minuscules than 
in uncials. A 1571 DVUTI..YJTa 7T€P ] SVU7TI..YJTa 7T€P F. X 600 a7TEpwToSJ 

a7TEpw1ToS MI. E 356 TLOaUOS ] 1TlOauos M, 1TtOauuos FT. E 914 1TpE1T­

TWV ] Tp€1TTWV FT. 
Tau evolves to psi once, at A 1566, 7TPOS uTaL ] 7Tp° ua!f;at. 

Pi + tau develops to double pi at A 590, EV l7TTW V ] EV{1T1TWV. 
G.R.B.S·-3 



90 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

Tau develops from sigma + tau. at A 143, cptAop.aaTOt~] cptAop.aTOtS 
Ml, and at A 145, where the paradosis has unmetrical aTpovOwV, the 
original reading may have been the dialectal form TPOVO(jjV. Here 
there is vulgarisation more than graphical confusion. Indeed, merely 
graphical error is less common than error involving some thought, 
or lack of thought, by the scribe, who would normally be familiar 
with some sort of Greek. 

Misdivision of the originally continuous text was a pregnant 
source of error, usually entailing subsidiary errors of non-graphic 
types, for example: A 254 aVv opOOVaVT(XL~] avvopOov atiTa,s MV, aVvap. 
Opov aVTa~s FT. A 340 avOaAo~EV av ] <Xv OaVOtEV av V, a~ OaVOtEV av 
FT. A 374 €YKovovaa TOAP:'lTWV J €yyovovs aToAp.1JTWV. A 702 aTtp.watv] 

" II , F " T A 1091 ' '0 I ] ", F ' aTtp.w~ LV ,aTtp.w~. KaKa apTa· vat KaKa KapTaVat ,KaKa 
KapTavas T. A 1392 yavE' ] yay El. A 1551 P.'ATJP.' aMYEtV ] p.'ATJp.a 
MYEtV. A 1595 avOpaKas KaO' ~p.p.'vovs ] &v8paKa~ KaOTJp.€vos. X 197 E~ 
aacp' 7]v 7} ] E~ aacpTJvfj. X 230 aKtbpat. TOP.fj ] aKE"'a£To p..;,. X 262 8' &v 
apEtasJ oavaptas. X 510 ap.Ep.cpfj T<)VO' €TEtvaTov] ap.op.cpTJTOV 8' Twa T6v. 
X 532 o~eap 7]v VTrO aTvyovs] OVXap"lV VTroaroyos. X 675 olKEta£ aayfj£ ] 

, I " X 742 ' -,,, ] ' - X 956 ' e - J' I O£Kta£S aY"l. EKE tV OV EKE WOV. Eyxpov£a E £aav EV XPOVOtS 
O - X 1021 -'- '\ '\' • " ,~ -, , , ] , '\ '\ " 't" ... ," E 224 E£aav. a/\I\ ws av E£O"lT • ov yap ul\J\oaav EL o"l TOUT up. 
o~ lluUasJ S' €1T' a,uas. E 435 ~£' aVT' €1Ta,lwvJ &,lav T' €1Ta,lwv. 
E 890 TfjaO€ yap.oplfJ ] TfjOE y' ap.olpov. 

The Oresteia exemplifies the tendency for articles to be added by 
scribes, who were habituated to Attic prose usages: e.g. A 116 ol F. 
A 140 aFT. A 145 TwvFT. X 325~. E2566. Bearing in mind this tendency 
one may take a new look at A 102f: 

''\ , " .l. I~'" '\ 
EI\1TL~ ap.VVE£ .,.,POVTW a1TI\"laTOV. 
rryv Ovp.ocpOopov At1m]S cpp'va. 

Wilamowitz thought that an iambic dimeter clausula could stand to 
conclude the anapaests, rightly deleted the article, and then printed 
for 103, with two changes, Ovp.ocp06pov At1m]s cpPEVl. Some may prefer 
to interpret the residual paradosis, by adding an iota subscript, as 
Ovp.ocpOopov AtkrJS cpp'va. "Hope wards off insatiable anxiety, soul­
destroying with griefs my heart," where the verbal compound 
Ovp,ocp06pov governs a direct object, cpp'va, cf. X 23, xoas 1TP01TOP.-
1TOS ••• Some might prefer the form AV1Ta£S. 

Particles and other small words added include: A 2 0' MV. X 87 O~. 
X 788A 8~. X 960 8' (after tXtwv, which should be kept, as the asyndeton 



DOUGLAS YOUNG 91 

is effective). A 448 y€ T. A 1418 T€ FT. E 121 y' FT. E 546 y€ T. E 378 yap. 

A 387 ws T (probably metri gratia, cf. 369). A 1340 ayav T (again probably 
from some metrical theory Triclinius had). 

Copyists were liable sometimes to drop particles and other small 
words, e.g. A 81 0' V. A 154 yap FT. A 539 yf (restored by Enger before 
T€(JvaVat). A 741 T'. A 546 a' (by haplography after ¢p€vos). E 550 wv T. 
In A 410, for lw lw owp-a owp-a, F negligently writes each word singly, 
lw oWfLa. 

Fairly abundant are non-graphical "ear" mistakes, including 
confusions involving the vowels and diphthongs 0, w, OV, €, "f'j, €£, at, 

O£, v, £. A 87 1T€£(J0£ ] 1TV(J0£ F. A 262 €A1T{atV ] €A1T{a€tV M. A 297 1T€O{OV 

'Aaw7TovJ 1TaLo{ov dmov MV. A 312 £TOLP-O£ ] £TVP-0£ F. A 959 laapyvpov] 

€lr; apyvpov. A 1602 dMaOa£ ] dMa(J"f'j. A 1624 1Ta{aar; ] 1T~aar;. A 1652 
1TpoKw7Tor; ] 1TPOK01TOS F, 1TpoKo1TTor; T. X 26 0' lVYfLo£aL ] O£otYfLo£aL M. 
X 73 X€tpOfLvaij] XaLpofLvaij. X 74 KaOL€pOVVT€S ] KaOa{pOVT€S. X 87 T{ 

¢w ] TV¢W Ml, olp-aL TVp-!3cp M2. X 126 OWfLcXTWV ] 0' 0p-fLcXTWV. X 172 
K€{paLTo VLV ] K€{P€TO V€£V. X 291 KpaTfjpor; ] KpaT€por;. X 380 ovr; ] wr;. 

X 563 iJaop-€v] otaofL€v. X 653 av, T{r;] avOLr;. X 783 1TapaLTov/dva£ 

fLO£ ] 1TapaLTovfLEV' €fLol. X 992 €g 0.0] €K aov. E 40 O€OfLvaij ] O€ofLvaijt 

M, (J€OfLtaij FT. E 113 €YKanAAw!foar; ] €KKanAAw!foar;. E 186 0.0 KapavL­

aTijp€r;] 013K up avvaTfjp€r; T. E 257 av T€ywVJ aVTE y' oVv. E 406 Ka£­

vT]v] Kat VVV. E 409 gEVcpJ aTEvw FT. E 519 O€£ P-EV€£V] O€LfLa{V€£. E 656 
1TpoaoEg€TaL ] 1TpoaoEgatT€ M. 

At X 482 the paradosis may result from an "ear" mistake. After 
Orestes, at Agamemnon's grave, has prayed for sovereignty over his 
palace, Elektra is given by M these words (481f): 

"" I~....,,, Kayw, 1TaT€p, TOtaO€. aov xpHav €XW. 

t cpvy€£V fLEyav 1Tpoa(J€£aav Aly{a(Jcpt ••• 

Orestes then refers to Agamemnon's future participation in the 
palatial banquets, and Elektra, at 486ff, promises to give her father 
libations from her marriage portion. Accordingly, at 482, some have 
thought she had some reference to marriage. I would make 482 an 
accusative and infinitive of wish, in this form: 

.J.. I I () ~ A' I () , .,.,VEtV fLEyav 1Tpoa ELaav LyLa cp (naLV>. 

"I too, father, am of such a mind. I have need of you. May I bear a 
great (son) after inflicting on Aigisthos revenge." CPV€LV could evolve 
to the commoner word cPvy€£V by a slight aural confusion in the mind's 
ear. Further, the copyist might have obscurely in mind the notion 



92 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

c/>vy€'iv fL€ yav .•. "May I go into exile from the land after inflicting on 
Aigisthos (?)." Such a wish would not be wholly absurd in view of 
Orestes's wish at X 438 to die after killing his mother. TI:LIN might 
fall out by a near haplography after the element r1:L in AI r1:L901. 

The transmission offers samples of errors in inflection, some of 
them due to the normalizing tendency of scribes or to metrical 
theory. At A 263 and 271, F and T, to suit changed ascriptions of 
speakers, offer the participles aLywvn and c/>povova?Js with changed 
genders. At A 680 p2 and T offer the infinitive KAv€LV for the participle 
KAVwv after LaOt, probably as a construction more normal in the 
innovator's conception. Datives in -OtS or -otm(v) are interchanged 
too often to notice. The following will serve as examples of types of 
inflectional error: A 26 a?JfLaLvw M ] a?JfLavw VFT. A 48 KAa'OVT€S ] 

KAaygaVT€S FT. A 80 TpL7TooasJ TPL7TOOOS FT. A 109 iff:3as ] iff:3av. A 222 

f:3POTOVS ] f:3POTO'iS. A 236 c/>VAaKat ] c/>VAaKaV. A 336 (XTTaAAax0I.V'T€s ] 
a7TaAAayl.vT€S FT. A 736 7TpOa€Opl.cpOYJ ] 7TpOa€TpacpYJ FT. A 833 cpOovwv 
F ] c/>0ovov T, rpoyov Stobaeus. A 1146 aYJooOs ] aYJoovos. X 15 fL€tALYfLa­
Ta] fL€tALYfLaaLv. X 136 cf>€vywv] CP€VY€LV. X 144 OlKYJ ] OLKYJV. X 202 
OLKYJVJ OlKYJ. X 217 EK7TayAovfLI.VYJV ] EK7TayAovfLI.VYJS. X 222 EfLo'is ] EfL0'i­
aw. X 360 ijaO' ] ijv Ml, ijs M2. X 365 aAAws ] aAAwv. X 366 T€OacpOaL ] 

TI.OarpaL. X 480 alTovfLl.vwL ] alToVfL€VOS. X 556 KT€LVaVT€S ] -as. X 727 
(EpfLijVJ EpfLija M (? *(EpfLl.a is possible). X 867 O€'iosJ O€LOLS. X 1004 
cf>pl.vaJ cf>P€vl. X 1057 7TAYJOvovm ] 7TAYJOvovaaL. E 435 al.f:3ovaaL y' ] a€­

f:3OfL€vat P, al.f:3OLfL€V T (F2 Jere). E 445 ECP€'OfLYJvJ ECP€'ofLI.VYJ, implying 
T'J 

an intermediate stage Ecf>I.'OfL€V. 

Collators are often in doubt how far to take note of the presence or 
absence of the ephelkystic or facultative nu and of iota adscript or 
subscript. The lability of non-facultative nu is, however, worth 
observation. Sometimes it is lost internally, as at A 82, where M and 
V have ~fL€pocpaTov for ~fL€pocpavTov. E 138 KanaxvaLvovaaJ KaTLaxaL­
vovaa. E 267 laxvavaa' ] laxavaa' FT. (lxvavaa' M). E 705 €VOOVTWVJ €V­
OOTWV Ml. 

Conversely, nu is found intruding into the middle of a word: 
A 84 KAvTaLfL~aTpa M ] KAvTaLfL~aTpa VFT. A 110 Tayav] TaV yav M. 
A 310 TOO€ aK~7T'T€L ] TOO' EVaK~7T'T€L V. X 55 aoafLaTov] aoafLaVTov 

(a common variant). E 253 f:3pO'T€{wv] f:3povTdwv Ml. 
Non-facultative nu may be added at word-end: A 170 OUOEJ oUO€v. 

A 1284 ag€L ] ag€LV F (VLV follows). A 1486 7TaV€pyI.Ta] -hay F. X 764 



DOUGLAS YOUNG 93 

aTEixwJ UTE{XWV. E 230 aYELJ ayEW M. E 787 f3at..EI, ] f3at..E'iv. A 1068 au 

f.L~ M ] ou f.L~v M2FT is a dubious instance. 
More often, nu is lost at the end of a word: A 1325 4>ovwuw ] 4>0-

VEVUL. A 1419 xpfjv ] xp-ry. X 177 ~v ] ~. X 282 E7TavTEt../..E£vJ E7TavTEt../..€£. 

X 612 aAAav D' ~v ] &Ma D~. X 641 ovvJ oU. X 650 'TLV€LVJ T€{V€L. X 726 
oo;\.{av ] oot..{a. X 949 0' EV f.LIXX~ ] O€f.LaX<;!. X 1003 VOf.Lt,wv ] V0f.L{'w. 

A 304 f.L~v ] f.L~ is a doubtful case. A 984 €7T7}VJ E7TI:' F, E7Tr. T may be 
historicallv true . 

.I 

Iota adscript is confusingly dropped at A 77 avawuwvJ avauawv, and 
A 431 MfJ-w£ 'v] oOfJ-wv. 

At X 715, M offers in Klytaimestra's speech: 
, ~ ~" ., f) I I~ aLVW O€ 7TpaUUELV WS' €7TEV VVWL TaOE. 

Here I suspect an iota has been adscripted on the assumption that there 
is a dative adjective. I would print E7TEvf)vvw and render: HI bid you do 
these things as I direct." She is instructing her majordomo, rather 
fussily. 

A high proportion of errors in the Oresteia transmission involve 
only one Single letter, whether added, dropped, or altered: A 29 

E7Topf)La'ELv ] E1Topf)pLa'ELv MV. A 45 XLt..£OvavTavJ HIt..wv aUTav M yp, 

F yp. A 64 EpELOOf.LEVOV ] EpEL7TOf.LEVOV PT. A 69 iJ7TOKaiwv ] tJ7TOKt..atwv. 

A 94 xplf.LaToS'] xpluf.LaToS' FT. A 104 oo£Ov ] ou£Ov Ar. Ran. 1276 
plerique. 8S' DI,OV Ar. Ravennas, presumably from an intermediate 

a 

stage *oo£Ov. A 115 apyas ] apytaS'. A 127 t..ayooatTaS'] t..oyooatTaS' M. 
A 137 1TTaKa M ] 1TTwKa FT; V's 7TTawvKa is an odd development. 
A 141 aE1TToLS' ] aEt..1TToLS' M. A 246 EV1TOTf.L0V M2V ] EU1T(lTaf.L0v M1F, 
EV7TOTOV T. A 247 1TaLwvaJ alwva. A 292 Eupl7TOV ] Evpl7T7TOV MV. A 448 
DLar. ] DLa F, yE DLa T. A 655 7JPE£KOVJ 7Jp€£7TOV T. A 898 UTVt..OVJ UTOt..OV 

F. A 976 DEl,yf.La F ] DEI,f.La T. A 1093 EVPLS' ] EVpOS' Ml. A 1143 {3oas] 

{3ofis M, {3opas T. A 1148 aywvaJ alwva M yp. A 1166 f)paVf.LaT' ] f)av­

f.LaT' T. A 1258 Dl7TOVS' ] 3l7Tt..OVS'. A 1411 O{3pLf.LOVJ Of.L{3PLf.LOV. A 1414 TOT'] 

TOO'. A 1504 VEapOI,S' ] VEKpOI,S' T. A 1565 apal,ovJ pf!.ov. X 68 DLa4>p€I,] 
'Y/ 

DLa4>EpE£. X 182 ljiavuE£ ] ljiavoE£. X 221 Tapa] Tappa. X 224 Tap' ] TaD'. 

x 240 UE ] T€. X 443 DvaS' (lTtf.LOVS' ] DvuaT{f.L0VS'. X 553 f.LEV] D' EV. 
X 807 KT{f.LEVOVJ KTaf.LEVOV. X 1067 1TvEvuaS' ] 1TvEovuas. E 450 {30TOU ] 

{3POTOV FT. E 452 {30TOI,ULJ {3POTOI,UL F. E 907 {3oTWVJ {3poTWV. E 505 
V1ToDouw ] V1ToD7JULV F, V1ToDVULV T. E 800 0' EfiTE ] 8E T€. E 842 V1ToDvETaL ] 

v7TooETaL M. 



94 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

1TpO- and 1TpOU- in composition seem to be interchangeable: A 391 
1TpoufloAat~ ] 1TpofloAat~. A 1511 1Tpo{JalvwvJ 1Tpou{Jalvwv. X 647 1TPOXa>.­
KEVEL] 1TpouxaAKEVEL. X 805 1TpOUCP&:TOL~] 1TpocpaT'OLS Ml. E 445 1TpOU­

T'p/maLos ] 1TpOT'pb1TaWS Ml. E 718 1TPOUT'p01TatS ] 1Tp0T'p01Tat~ FT. 
Many corruptions affect only a single syllable of a word, thus: 

A 87 8VOUKEtS] 8VOUKLVEtS Mss Jere, 8VOUKVEt~ M olim, Svos KLVEtS VT, 
8VOUKOEtS LT. A 141 '\EOVT'WV] OVT'WV MV, senselessly, so that FT omit. 
A 165 >.EgET'a£ ] >.Ega£. A 229 alw T'E ] alwva. X 350 alwJ alwva. A 416 

SJ: ] YdP T. A 714 1TafL1Top8fj] 1TafL1Tpbu8'Y/. A 766 on] oT'av. A 1414 ovSJ:v] 

ov uvv fl. A 1430 T'vfLfLan T'EtUaL ] rVfLfLa T'lUaL. A 1471 KapoLbo'Y/KT'OV] 

Kapolf[- O'Y/KT'OV FT. X 8 1Tapwv <f}fLwga ] 1TapwfLwga the scholiast on 
Eur. Ale. 768, cod. Vatican. gr. 909. X 160 EtU' dropped out after 
OOPVU8EvryS. X 233 fL~ 'K1TAayijLS ] fL~K1TAaYLij. X 590 1TEoaopo£ ] 1TEoafLapo£. 
X 607 KaT'al8ovuaJ K' al80vua (perhaps the preposition had been con­
tracted). E 225 'At1TW 1TOT"] A{1TW 1T(i)1TOT'E Ml. E 702 ounsJ OWOT" FT. 
E 746 vVv] vav M. E 754 uwuaua] uwua Ml. E 798 0 xp~uas ] dp8~uas 
M, 0 8~ua~ M2. E 802 uT'a,\aYfLaT'a] uT'EvaYfLaT'a M yp. E 812 aV7'£1TEv8ij ] 

aV7'£1Ta8fj M. E 816 E1TLUVfLEVOS ] E1TEuuvfLivos. E 849 KalT'OL T'ei: fLJ:V] KalT'o£ 

fLJ:V M, KalT'o£ yE fL~V FT. E 908 Ev8EVOVVT'a ] EVUT'EVOVVT'a FT. E 954 KPVE­
pov] KPVWV MF, oaKpvwv T by mere conjecture, I fancy. My own con­
jecture makes better sense and explains the corruption better. In the 
above examples some are simple cases of haplography or ditto­
graphy, or simple negligence; but others reveal a more or less sub­
conscious ecdotic attitude. 

Metathesis and anagrammatism in various degrees occur, thus: 
A 234 Aa{JetV ] {JaAEtv T. A 1088 fL~ T'bO' ] T'O fL'Y/O' F, T'b 1TEp fL'Y/o' T. 
A 1205 a{JpvVET'a£] flapVVET'a£ T. A 1594 XEPWV ] XPEWV F. X 23 KT'V1TlP ] 
KV1TT'WL. X 232 €U£OE] Els oJ:. X 391 Kpaolas] Kapolas. 

There are examples of the Byzantine tendency whereby words 
are transposed to make a verse end with a paroxytone word: A 5 8lpos 

{JpOT'Ot~] {JP0T'ots ()Epos FT. A 1064 K'\VE£ CPPEVWV ] CPPEVWV K'\VE£ flE. A 
1106 1TO'\LS {Joij. ] {Joij. 1TbA£s FT. A 210 1TlAas {JwfLOV] {JwfLoV 1TlAas, in a syn­
copated lyric iambic trimeter, may be influenced by the same factor. 

Glosses have sometimes been added to a line or have replaced the 
original word in it: A 111 oop~ Ka~ XEP~ 1TpaKT'OpL ] oop~ olKas 1TpaKT'Op£. 
(Possibly, however, here SlKas is a stopgap, not a gloss, to make up 
for Ka~ XEP~ lost by homoeoteleuton. It would make up ten syllables, 
which Byzantine editors might think a correct responsion, regardless 
of quantities, to 129 as they had it). At A 153, V has uVfLfLEVEt CPVT'bV, 



DOUGLAS YOUNG 95 

where other Mss have aVf-upvTOV. (Read rather V€LKEWV TEKTova aVfLrpv­

TWV, a glyconic in astrophic rhythm.) V has conflated aVfLcpvToV with a 
gloss fL€V€t written over fLLj-tV€t in the line below, 154. A 1143 rptAOL­
KTOtS ] TaAaLVatS M, rptAOLKTOtS TaAaLVatS F, cptAOLKTOtat T. A 1174 T{(}TJat 

8aLfLwv ] 8aLfLwv 7TOt€L T. A 1454 Ct,7T€cpOta€V {3LOV appears in the Mss (F 
and T only here), but the antistrophe would then lack a final iamb, 
though its sense is complete as it stands. At A 857 Aeschylus uses 
Ct,7TOcpOLVW intransitively, and he could do so again at A 1454 (making 
the clausula there a hypodochmius, or anaclastic dochmius). I sug­
gest {3tov is an intrusive gloss by a scribe who took the verb as causal, 
as it mostly was. At X 32 the gloss lPo'i{3os has been unmetrically 
intruded in front of opOoOptg 8oj-twv OV€tPOjLCXVT£S, having originated 
in some mistaken process of reasoning. At E 560 the right reading is 
clearly O€pj-tcp, and the glossing process has been at work, as appears 
from the collation: OEpj-tOEPYCP MF] OEpj-tcP, ijyovv OEpj-tOVpycp T. Simpler 
examples of gloss-intrusion include: A 198 alV€LvJ €l7TELV FT. A 282 
ayyapov ] ayyEAov. A 400 €S 8oj-tov ] €ls 0 IKov T. A 549 KOtpaVwv ] TV­

pavvwv F. E 170 j-tvxov ] aov olKov FT. E 448 acpOoyyov ] acpwvov FT. 
E 934 Ct,7TAaK~jLaTa Vtv ] Ct,jL7TAaK~jLaTa VtV M, ajLapT~jLaTa FT. 

At X 160-163 the text probably ought to run: 

" I~ 0' 1" "I 'W, T£S oopva €VTJS (Eta> aVTJp, 
, \ \ c:,' n e ' , , ~ aval\vTTJP OOj-tWV, £JKV tKa T €V X€POtV 

, , "" , ", "A 7Tal\tVTOV €V €PYo/ 7Tt7TaI\l\WV P'Y/'i 
I~ I , , I ~ {3 I, aX€OLa T aVTOKW7Ta VWj-tWV €I\'Y/ ; 

It is an astrophic run of dochmiacs, variously resolved, and means: 
"Ho! What man will come, strong with the spear, liberator of the 
household, and what warlike force brandishing in its hands Scythian 
weapons bent back in action (= bows), and wielding at close quarters 
weapons grasped by the hilt (= swords)." The (3EATJ at the end of 163 
goes equally with L'KVOtKa ..• 7TaAtvTov(a) and with aXE8ta •.. aVTO­

KW7Ta. In the paradosis an unmetrical {3EATJ has been intruded before 
'7Tt7TaAAWV probably from an interlinear gloss, and some editors, 
following Pauw, have extruded the final word {3EATJ in favour of gtCPTJ, 

which is part of the scholiast's explanation of Aeschylus's phrase. 
{3EATJ can mean either missile weapons, as arrows, or others, as 
swords. If the paradosis' {3EATJ in 162 be not from a gloss, then it is 
an example of influence from a word in the following line, of which 
I can see some more in the Oresteia. 



96 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

Thus, at A 512, for Ka~ 7TatWVtOs we find Ka~ 7TaywVtOS F, Ka7Taywvws 

T. Some scribe's eye had strayed to the line below, ending with T' 

aywvlovc; OEotJs. At E 507 the unmetrical OE ns, deleted by Schwenk and 
Pauw, arises from a scribe's eye having caught in the next line the 
same letters in IJ:YJOE ns. (The intrusion is not from scholia, as sug­
gested by Groeneboom and Murray.) 

The influence of an adjacent line may be negative, thus: at X 832f 
the Chorus, inciting Orestes to slay his mother, say, as I supplement 
the lacuna, 

nEpaEws T' Jv c/>PEaa~v (with synizesis making a dochmiac) 
(7TEp(JE> KapOlav aXE(Jwv, ••• 

"Keeping in your midriff the heart of Perseus, destroy (her) ... " In 
uncials with round sigma to which the theta of nEP9E was assimi­
lated, the imperative fell out below uncial nEPCEOc. This is not the 
place to argue what should be read at 819ff in the strophe; but in 
819 the paradosis can be interpreted as a dochmiac, Ka~ TOTE O~ 7TAWTWV 

(original nJ\OTON taken wrongly as 7TAOVTOV). 

The influence of lost words in a strophe may cause loss in its 
antistrophe, as the lacuna in A 1006£ induced Triclinius to eject at 
A 1031 the syllables (JV/LaAY~C; TE Ka~ ovoev J7T. Fortunately the Oresteia 
paradosis seems not to have suffered substantially from such free 
ecdotic interventions before Triclinius, who was about as irresponsible 
as many scholars of the nineteenth and twentieth centuries, the 
Iron Age of Aeschylean corruption. 

Influence from the line below seems to have been the root of the 
trouble at X 64, the only corrupt place in lines 61-65, which can be 
read thus: 

• ,~., A ~, 

p07TT} 0 E7TtaK07TEL OLKav, 
~ .... , , ,L' 

TaXELa TOLS /LEV EV 'jJaEL, 
\ ~, , " 

Ta 0 EV /LETatX/LLctJ aKOTOV 
, 'Y a' /LEVEL XPOVL,=>OVTa t'PVEL, 

65 ,~, " ", (; 
TOVS 0 aKpaVTOS EXEL VVs. 

«A turn of the scale controls justice, swift for some in the light; and 
other events in the frontierland of darkness abound in power as they 
delay; and other men night without fulfilment holds." The chorus is 
oracularly discussing the varying rates and manners in which guilty 
persons are punished. The above text is exactly as in M, our sole 
authority, except that in 64 M wrote /LEVEL XpOVl~OVT' C1XEL (altered to 



DOUGLAS YOUNG 97 

a)(l7) {3pV€t. I suggest the scribe's eye had taken in the X€£ of €X€L in 
the line below, and, having mistaken the dative of the noun p.,EVOS 

for the third singular of the verb p.,EVW, he was expecting a noun 
in the neuter to go with the preceding Ta, and thus arrived, via ax€£, 
at aXT], which made nonsense of metre and of the ensuing {3pV€L. 
Line 64 in the restored form was adopted by Hermann (at one time), 
Weil, Verrall, and Groeneboom. 

The foregoing brief discussion is a reminder of the infinite guerrilla 
of interpretation that would face anyone who should attempt to 
evaluate in rigorous percentages all the types of error in the Aeschy­
lean paradosis; for it is often impossible to satisfy oneself, much less 
anyone else, what the true reading can have been from which a given 
manuscript variant has deviated. 

To conclude this paper I list a few miscellaneous errors in the manu­
scripts that involve more than one syllable. A 23 4>aosJ vvv 4>ws FT. 
A 119 eptKvp.,ovaJ eptKvp.,aTa M. A 119 4>Epp.,anJ 4>Ep{30VTO FT. A 217 
8Ep.,tS. €D yap €I.TJ ] BEp.,ts yap €D FT. A 1030 fJM7T€L p1 ] fJpEp.,€L PT. 
A lO41 oovAtas p.,aSTJs fJtav ] OoVA€tas p.,aSTJS fJta F, Kat svywv BtY€LV {3ta 

T. A 1356 TijS p.,€AAOVS KAlOSJ TijS I.t€AAOVa7]S KAlOS F, p.,€AAOVa7]S KAlOS 
T, TijS p.,€AAoVS xapw Trypho. Aeschylus' holograph may have had 
variants. X 164 ya7Tc5TovsJ ebb TOV. X 319 aVTtp.,OtpOv] laoTtp.,oLPOV (from 

Lao 

aVTtp.,OtpOV as the presumed intermediate stage). X 797 KTtaa£ ] TtS ~v. 

X 896 0' al.OmaL ] O~a€TaL. X 954 €7Twpeta~€v ] €7T' 8X(l€L a~€v. E 259 7T'€P~ 

{3P'T€L 7T'A€XB€tsJ 7T€pt{3M7T€t 7T'AayxB€ts FT. E 286 YTJpaaKwvJ y€ oLoaa­
KWV PT (influenced by a parallel at PV 981). E 343 7T'apa4>opa] 7T'apa-
4>pova (330 7T'apa4>pova FT). E 476 OVK e(mlp.,7T'€Aov ] OVKOVV €V7T€7T'AOV FT. 
E 567 1'" I ~ ~ I ]'" l' '" 'M" , l' '" I €£T ovpavovo€ ULaTopos TJ T ovv OLaTOpos ,'T} T ovv oLaKTopos 

'FdA€' FT. ovv for ovpavbv by compendium was the source of the 
trouble. E 832 Kotp.,a ] Kat Kvp.,a FT. Kvp.,aTos later in the line influenced 
the start. In many corruptions one can readily see the adjacent 

a a 

influencing factor. E 845 np.,av oavatavJ np.,wv oap.,a£ *wv M, np.,av 
oap.,'av FT. In the repetition at E 879, M switches to np.,av oap.,tav, and 
F to Ti:Tav oap.,Eav. 

In A 1493, F had aa€{3€1. BaVaTlfJ, but in the repetition at A 1517 changes 
to €va€{3€1. 8aVaTlfJ, whether by a monkish thought about a pious 
death, or from a graphical confusion whereby a form of alpha was 
read as the Byzantine ligature for €v. (This same confusion happened 
in the paradosis at Theognis lO44, in XPD and some later Mss.) 



98 TYPES OF ERROR IN MANUSCRIPTS OF AESCHYLUS' ORESTEIA 

X 247 ylwav €OVLV ] ylwavLv Ml, with EVV lost by near haplography 
after av, may be an example of the same graphical confusion of 
alpha with a medieval ligature. But this type of corruption seems to be 
extremely rare in the Aeschylus tradition, which is distinguished, 
on the contrary, by errors deriving from uncial confusions. 

Uncial confusion may be at the back of the crux in X 649, which 
involves, as so many problems do, strophic responsion. The mildest 
cures being applied to M's readings, we find this: 

,~, " \ , C',I,. ~ TOO aYXL 1T1\€VP-0VWV s L'(-'O<; O't'p. 0 

640 ~ "c \, ~ ~ \ OLav-raLav 0SV1TEVKE<; OVTl(- OLaL 
A' ~ 0' \.,. 

..:.I LKa<; T0P-7J. EP-L<; yap OVV. 

/tag 7Tl80L 1TaTOlJp-€VOV 
\ ~ A \ 't:) 

TO 7Tav ..:.ILO<; aEJ-'a<; 1Tap€K-

645 f1cY.VT' Eafi- O€p-LaTw<;. 

A I ~" I~ () , ..:.ILKa<; 0 €p€LOETaL 1TV P-7JV. «v't'. & 
-- \ , ~'A" ,I,. ,,~, 7TPOXat1.K€VEL 0 Laa '(-'aayavovpyo<;' TEKVOV 0 

E7T€wcplp€L 8L' &""ay~<; 
owp-aTwv 7TaAatT€pOV 

650 I , , \' 
nV€LV p-vao<; XPOvcp KI\VT'7] 

{3vaabcppwv 'EPLVV<;. 

"Here near the lungs the sword, keen-timbered, strikes a straight­
through (blow), through, by the surgery of Justice. For it is right. 
Trampled underfoot to the ground entirely, the majesty of Zeus 
fills with Ruin a transgressor, righteously. The anvil of Justice is firm 
fixed; sword-worker Destiny continues her smithy-work; and the 
famous deep-minded Fury in time brings in a child to requite in turn 
the older pollution of the household." The Chorus's sentiments are 
suited to the moment where Orestes is about to enter the palace 
and avenge his father. Changes from the paradosis are these: 639 
TO 0' M, T68' Young. 640 aOtlTaL M, OVTfi- Hermann. 641 TO p-~ M, 
Top-ii Young; ov M, ovv A. Ludwig. 645 7TapEK{3&v-r€<; &()€p-taTw<; MI, 
7Tap€K{3&vT' Eafi- ()€p-LaTw<; Young. €la&w would be a correct formation 
from the Homeric verb &&w, which LSJ attest in the aorist from 
Aeschylus and Sophocles. It would mean "fill with Ate," cf. €laav8pbw, 

"to fill with men." When a redivision of the paradosis offers acceptable 
sense, no other remedy should be sought for a crux. Here all that is 
needed is the addition of an iota subscript. 647 1Tpoaxa/tK€VEL M, cor­
rected by Jacob metri gratia, resulting in bacchius+palimbacchius+ 2 



DOUGLAS YOUNG 99 

cretics. 648-9 3Ll-'aa~ oWI-'(frwv M (a in rasura) is an unlikely corruption 
from 861-'0ts" ail-'eXTWV, the original assumed by editors who follow the 
readings of Schuetz (after Pauw) and Stephanus, who got the idea of 
aLfLeXTwv from the scholiast, who misunderstood the passage. Murray 
reports an epsilon in M above the Ow of OWp'&.TWV, which he takes to 
have the value at, and to imply alfLcfrwv. Even if this were by the first 
hand and ink, it would have no more authority than the uncompre­
hending scholiast's comment E7Tf:tacpEp€t O€ TO'iS OLKOLS TEKVOV 7TaAauJJv 

aLfLeXTwv,o Ean, T{KT€t <> cp6voS" aAAov cpovov. The child brought by the 
Fury is Orestes, just as at A 1607 Aigisthos claimed to have been 
brought home by Justice to avenge his murdered brothers and sisters. 
M's 8tfLaa€ is meaningless, but looks like a conscientious effort to 
reproduce a difficult antigraph, for M corrects a letter. Putting otfLaa€ 

into uncials we get ~[MACE, and we need a diiamb for responsion. 
The sense is suited by 8t' aAAaY1JS", which originally would have been 
~IAAAArEc. I suggest that AA became M (cf. Thomson-Headlam on 
X 995 [their 1001]), and the uncial gamma was taken for sigma, as at 
A 101 and X 542 (with the converse phenomenon at E 58). Then we 
have loss of uncial sigma after round uncial epsilon (standing for eta). 
The suprascript epsilon in 8wfLeXTwV may be someone's alternative 
interpretation of the first 0 of an old ~OMATON somewhat illegible, 
which had been primarily interpreted as for omega. At 649, on my view, 
M's 7Ta'AatTEpwV should be re-interpreted as 7Ta'Aa.{T€pov. 650 T€LV€£ M, 
T{V€tV Lachmann. M's form KAvTTJ at 650, where editors change to the 
lyric alpha, supports the eta forms of 641 TOfLfi, and 648 8,' aAAaY1JS", 

as do M's forms at 22, 386, 430, 467 and, with suprascript alphas, at 
383, 388, 468, 623 and 646 (cf. Fraenkel on A 1535f [III, p. 727 n. 3]). 
Contrariwise, there are Doricisms in anapaests, e.g. at A 1569. Why 
suppose Aeschylus more pedantically consistent in dialectal purism 
than Homer, Pindar or Burns? 

In sum, though in this passage, as in many, there are several slight 
adjustments of the paradosis to be made, most of them are routine 
remedies of slight normal errors. Even the less obvious remedy, 
8t' &AAaY1JS" for M's 8tl-'aa€, will not appear very difficult to those who 
have familiarized themselves with what has actually happened in the 
transmission of the Aeschylus text to us. 

UNIVERSITY OF ST. ANDREWS