Some Newly Identified LXX/OG Fragments among the Amherst Papyri at the Pierpont Morgan Library in New York City

by Robert A. Kraft, University of Pennsylvania

[28 May 2002; updated 04 September 2002; alignment and Greek finicky!]

The Setting

When Grenfell and Hunt published the papyri from Lord Amherst's collection a century ago, they described a number of LXX/OG fragments and suspected that there may be other scraps from the same sources among the unidentified pieces.\1/ They did not attempt to provide complete transcriptions of most of these unidentified fragments, but did include some selected lines, and some paleographic comments. The pieces that they were able to identify as from Jewish Greek scriptures (including Aquila) are as follows:

PAmherst 003(c1) LXX Genesis 1.1-5 (first half of 4th CE; Goettingen #912)
PAmherst 003(c2) Aquila Genesis 1.1-5 (first half of 4th CE)
PAmherst 191(a) LXX Exodus 19 (6th CE; Goettingen #914)
PAmherst 192 LXX Deuteronomy 32 (6th CE; Goettingen #916)
PAmherst 004 OG Job 1-2 (stichometric, 7th CE; Goettingen #913)
PAmherst 005 OG Psalm 5 (stichoi marked in line, 5/6th CE; Goettingen #2008)
PAmherst 007 OG Psalm 58-59 (stichometric, 5th CE; Goettingen # 2010)
PAmherst 006 OG Psalms 108, 118, 135, 138-140 (stichometric, 7-9th CE; Goettingen #2009)
PAmherst 193 (vellum) OG Proverbs 10 (stichometric, 6th CE; Goettingen #917)

To this list may now be added, as discussed below:

PAmherst 200 -- further fragments from PAmherst 006 of Psalms (see above)
PAmherst 191(b) and 194(a) -- OG Isaiah 58.11-14 (Goettingen #915)
PAmherst 198 -- OG "A Text" Judges (Goettingen #876)

1. PAmherst 006 + 200: Pages from the Psalms (Goettingen # 2009)

Grenfell and Hunt describe the plate labelled PAmh 200 as "fourteen small fragments ... belonging to the papyrus of the Psalms (Amh. Pap. I. 6)." The editors were not able to place any of these fragments. The larger, identified materials in PAmh 6 represent parts of four double-sided pages "from one manuscript which ... was a papyrus book. The [original] pages were of considerable size and each contained a single broad column. The handwriting is a large uncial, heavy and upright and carefully formed -- a typical example of the later Byzantine style. ... We are unwilling, in the present state of the evidence, to be definite concerning the dates of Byzantine uncials, especially of this variety which extends over a very long period. ... Provisionally, we do not think that this specimen was written before the seventh century, and it may be one or two centuries later. The ink is of the brown colour common at this period. A middle or high point is irregularly used to mark a pause. Single or double dots denote diaeresis."

They continue with a paragraph describing the format of this codex: "The stichometric arrangement of the Psalms ... is followed in these fragments. A fresh line is always begun for each sticoV, and the initial letter is considerably enlarged. When a sticoV is too long to be contained in a single line, the succeeding lines are commenced further to the right, by the space of a couple of letters, until the next sticoV is reached. By this method the divisions of the verse are sharply distinguished. Above each Psalm is written its title, enclosed within small wedge-shaped signs; and to the left of this is the number of the Psalm, above which is a horizontal stroke surmounted by a wavy flourish."

Much of this information is useful in attempting to identify the further small fragments in PAmh 200.\2/ Some of those pieces clearly preserve letters on the margin, and need to be viewed with the above formatting features in mind. It thus became apparent quite quickly that PAmh 200.1 could be fitted nicely onto the upper part of the final page, "d," of PAmh 6 at Psalm 138.21-24 and 139.10-12, although a textual variant at the start of Psalm 139.11 doubtless obstructed the original editors from making this identification. Fragment 3, and possibly also 4, 8, 10, and 13 of PAmh 200 come from the same page. The resulting reconstruction of PAmh 6(d) is thus as follows (transcription of newly identified fragments appears in upper case):

Psalm 138(139).20-139(140).9a (written against the fibers, "verso"; 37 lines, up to about 30 letters in the longest lines)

(top and left margins preserved at the top, and parts at the right)












































































01 lhmyontai eN[ mataiothti ]POLIS SOU
02 ouci touV misoun[taV se ke emishsa]
03 kai epi toiV ecqr[oiV sou exethko]MHN
04 t[eli]on misoV em[isoun autouV]
05 eiV ekcqrouV eg[enonto moi]
06 dokimason me k[e kai gnoqi ]THN
07   kardian mou[ (b l a n k) ] (blank)
08 etason me kai g[nwqi taV tribou]S MOU
09 kai ide i eideV a[nomiaV en e]MOI
10 kai odhghson m[e en odw aiwnia]
11 rlq yalm[oV tw da(uei)d ]
12 exelou [m]e [ke ex anou pon]hrou
13 ap[o a]nd[roV adikou rusai me]
14 [oitineV elogisanto adik]IAN[ e]n kardIA
15 [olhn thn hmeran p]aretassonto
16 [
  polemouV] ( b l a n k )
17 [hkonhsan gl]wss[an] autw[n
18 [
  wsei ofewV ] ( b l a n k )
19 [ioV aspidwn upo ]ta cilh autwn
20 [fulaxon me ke ek ]ciroV amartwlou
21 [apo anwn adikw]n exelou me
22 [oitineV elogisanto u]poskelise ta di
23 [
  abhmata mou] ( b l a n k )
24 [ekruyan uperhfanoi pa]gida moi
25 [kai scoinia dietinan pagidaV ]toiV
26 [
  posin mou] ( b l a n k )
27 [ecomena tribou skanda]lon eqen
28 [
  to moi] ( b l a n k )
29 [eipa tw kw qV mou ei su]
( blank )
30 [enwtisai ke thn fwnh]n thV deh
31 [
  sewV mou] ( b l a n k )
32 [ke ke dunamiV thV swthriaV ]MOU
33 [epeskiasaV epi thn kefalhn ]MOU
34 [
  en hmera polemou] ( b l a n k )
35 [mh paradwV me ke apo thV epi]Q[u
36 [
  miaV mou amartwlw]
37 [dielogisanto kat emou mh egkatali]


line 1: The tops of the letters at the end of the line on this side of PAmh 200.4 are difficult to decipher, and cannot with confidence be equated with the expected text of the Psalm. Without reference to the expected text, I would have read the letters as PERLI or similarly, but POLI is not impossible and traces of SOU can be imagined. The itacism POLIS for POLEIS would not be surprising. In the Rahlfs edition (also for the original editors of PAmh 6) the words eiV mataiothta taV precede poleiV sou, but this is too long for the reconstructed gap, which covers only the space of 9 letters on the other side of the fragment. The variant en mataiothti (attested by "R" and probably by several Latin witnesses) fits better, and possibly our text did not include the definite article taV (a variant not noted by Rahlfs)? There also seems to be an isolated letter or two (vertical strokes, resembling the sides of an "N") above the conjectured "PO" on the fragment, perhaps part of a page number?

line 2: Nothing of the end of this line is preserved. In keeping with the attested practice elsewhere in this manuscript, the expected nomen sacrum abbreviation probably occurred here.

line 3: Rahlfs notes the variant touV ecqrouV in R and a few other witnesses. Our fragment seems to have toiV although the word appears to be carelessly written, and the "I" is especially unusual at the top, although it does not resemble the normal form for "U." The end of the line is represented in PAmh 200.1, where the letter eta is quite clear, and the mu and nu are probable.

line 4: The original editors reconstruct telion rather than teleion, but either is possible. The end of this line is blank on PAmh 200.1.

line 5: Our fragment has ekcqrouV where ecqrouV is expected (see line 3). The end of this line is blank on PAmh 200.1.

line 6: After the me the original editors conjectured abbreviated KE (kurie) with witnesses ART and a few others. B and several other witnesses have O QEOS (in abbreviated form QS) in agreement with the MT. While the initial letter is only faintly preserved in our text, it seems to be a vertical stroke consistent with K. There would be enough room in the lacuna for a fully written KURIE. The final word of this line is reasonably clear in PAmh 200.1.

line 7: This line is slightly indented on PAmh 6, and the end of the line is blank on PAmh 200.1.

line 8: The final letters are clear in PAmh 200.1.

line 9: The letters KAIIDEIEIDES are clear, with a normal dieresis over the second iota and a single dot over the third. Rahlfs prints kai ide ei odoV, citing variants to ei odoV, which is in agreement with MT and several Greek witnesses including the R group. Most of the Greek texts for which this detail is recorded have ei (MS A and a few L MSS have a single eta), and many follow it with eideV as here; A shows conflation (ei eideV odon). Itacistic confusion and its results are obvious (ei/i/h and eideV/odoV). The end of this line on PAmh 200.1 is virtually illegible, but is not inconsistent with the expected reading.

line 10: The line begins with the marginal marker denoting the end of one Psalm and start of another. The papyrus is misaligned as mounted in this area.

line 11: There is a short horizontal stroke in the left margin between lines 10 and 11, then the number "139" (RLQ) in the margin at line 11, followed by four fillers (>>>>) and the title, which is also followed by such fillers. The beginning of the title itself is barely legible. The original editors reconstruct as above (perhaps with the name abbreviated to DAD), in agreement with a few "Lucianic" witnesses; all other witnesses represented by Rahlfs apparently begin the title line with eiV to teloV which would be too long for this lacuna.

line 12: Probably both "nomina sacra" terms (KE = kurie and ANOU = anqrwpou) were abbreviated here (judging by the expected size of the lacuna).

line 13: There is no clear evidence of writing at the preserved end of this line, despite the length of the expected line (but perhaps some traces are visible at the top of the small scrap PAmh 200.10)..

line 14: PAmh 200.3 fits nicely at the end of this line, with the letters DIA extending far into the right margin. To the left of this section, PAmh 200.10 can be placed with some degree of probability, thus attesting the variant adikian (with families R, L, and A') rather than adikiaV as preferred by Rahlfs (with the B family, Sa, and most Old Latin witnesses).

line 17: The lacuna from the left margin should be able to accommodate about 13 letters, but the known texts have only 10. Rahlfs records no variants here. Thus the original editors place ellipses dots at the left margin and suggest in a note that perhaps the MS had a compound verb such as exhkonhsan (see Ps 51.4, Ezek 21.11). The expected text contains several letters that are normally broad in width (eta, kappa, nu, gamma, lamda), so perhaps nothing more need be supplied.

line 19: Read ceilh (itacism). Rahlfs reads diayalma at the end of this line, and does not note any variants. Our text has no room for it (see also line 28 below and line 1 of the other side).

line 20: Read ceiroV (itacism). Rahlfs records some variations in this line, but none seem significant for our purposes. The nomen sacrum was doubtless abbreviated again here (MS S omits it).

line 21: Our MS almost certainly abbreviated anqrwpwn (see also above, line 12), and had anqrwpwn adikwn not the variant anqrwpou adikou (Bo, some of L'He, A'). The original editors place a full stop after the N, before exelou (MS A has rusai here, as in verse 2 above), but it is not obvious in the MS.

line 22: Read upokelisai (itacism). Some witnesses have tou before this word (R L' 1219'), but that is unlikely here.

line 25: Nothing is preserved to adjudicate the variant of pagidaV (B' La A) or pagida (Bo R GaHi L''-Su 1219' = MT). The original editors help us restore the original position of parts of these fragments by noting that the "t of touV is immediately under d of pa]gida in 24 and over the first e of eqen in 26 [sic! 27]." Thus as now mounted, the lower section has slipped to the left the width of about two letters, and the final N is no longer visible.

line 28: There is room at the end of the lacuna for diayalma (so Rahlfs, who notes no variant), but the original editors think it probable that our MS omitted it here (with T), as also above in line 19 and in the first line of the other side of this page.

line 29: Possibly PAmh 200.8 fits here, with the letters ATW. If so, our text would not reflect the variant eipon reported for T.

lines 30-37 are poorly preserved, and no pertinent variants are noted by Rahlfs for this section.


Psalm 139(140).9b - 140(141).5a (with the fibers, "recto")

(upper and right margins preserved at the top, and portions of the left margin)

01   PHS[ me mh pote u]ywqwsin.
02 H KEF[alh tou kuk]lwmatoV mou
03 [kopoV twn ceile]wn autwn ka[l]u
04   YEI[ autouV ] ( b l a n k )
05 KA[i pesountai ep auto]uV anqrakeV puroV
06   EPI[ thV ghV kai ka]tabaleiV autouV
07 EN T[alaipwriaiV o]u mh upostwsi-
08 ANH[r glwsswdhV o]u kateuqunqhs
09   ET[ai epi thV ghV] ( b l a n k )
10 [a]ndra a[dikon kak]a qhreusei
11 [
  e]iV diaf[qoran] ( b l a n k )
12 egnwn o[ti poihsei kV thn ]krisin
13   tou ptW[cou kai thn dik]hn[ twn
14   penhtwn
15 plhn dikai[oi exomologhsontai tw
16   onomati sou[
17 kai katoi[k]hsou[sin euqeiV sun tw p]RO
18   s]wpou s[o]u ( b l a n k )
19 r]m yalmoV tw d[a(uei)d
20 ke proV se ekkek[raxa eisakouson mou
21 prosceV th fw[nh thV dehsewV
22   mou en tw k[ekragenai me proV se
23 kateuqun[qhtw h proseuch mou
24   wV qum[iama enwpion sou
25 ep[arsiV twn ceiron mou qusia
26    e[sperinh
27 qo[u ke ]f[ulakhn to stomati mou
28   kai qur[an periochV peri ta ceilh mou
29 mh ekkl[inhV thn kardian mou
30   eiV logou[V ponhriaV
31 T[ou profasizesqai profaseiV
32    E[n amartiaiV
33 SU[n anoiV ergazomenoiV anomian
34 K[ai ou mh sunduasw meta twn ek
35 [
   lektwn autwn
36 [paideusei me dikaioV en eleei kai
37 [
 elegxei me

line 1: The line begins with the ending of the word egkataliphV, and does not include, at the end, diayalma which is found there in most witnesses (see also the other side, lines 19 and 28). A medial dot is visible at the end of the line.

line 2: The tops of the first four letters of this line are visible in PAmh 200.4 and the expected reconstruction is highly probable. The final word, mou, is unique according to Rahlfs (see MT, those who surround me!); other witnesses have autwn (their head?).

line 3: The end of the line is badly damaged, but the reconstruction is supported by the evidence of the new fragment, PAmh 200.1, for the next line.

line 4: The end is blank.

lines 5-7: This section clearly begins with KA, followed by the lower part of a vertical stroke. The textual situation for this section is confused, including how the stichoi are divided. Our text presents two sticoi, and its wording is identical to MS B, although according to Rahlfs, B has kai katabaleiV autouV as a separate sticoV. Where Rahlfs' B group has puroV, as in our text, many other witnesses have en puri (as in MT) or have both. After this, some texts such as ours include epi thV ghV (notably B) or en th gh (notably R) -- the pi is clear on PAmh 200.1 -- and some (again, notably B) continue with kai which is almost certainly the situation here as well. So in its general contours, our text is closely allied to MS B and some of its allies here. But the problem of how the section began is not solved by referring to the variants recorded in Rahlfs. It is an obvious solution to read kai at the start of line 5, but that leaves us with a very long "filler" if the expected text is simply added -- 15 letters where the other extant lines average about 12. Still, no other solution is as obvious at this point.

line 7: The final nu is indicated by a supralinear line over the iota. The upsilon has a rough breathing over it.

lines 8-9: The first hand of MS B omits this material, which is then added in the margin.

line 10: The original editors read confidently qhreusei, with most other known texts (T has -sai), but I am unable to correlate those letters with the visible ink marks, which look more like HSWRISEI.

line 11: MS B uniquely reads katafqoran.

line 12: The initial large letter epsilon is supplied convincingly by PAmh 200.3.

lines 12-14: A single stichos, with R' T A(cf), but divided into two stichoi by B' Sa La\G/ Ga SyHe 1219 (Rahlfs).

line 13: Variant twn ptwcwn (L'Sy\mg/ 55).

line 14: Variant tou penhtoV (S Sa).

line 17: The original editors printed euqeiV here, and noted the variants en tw proswpw (B') and sun tw proswpw (the majority), with S having proswpou instead of proswpw. The -ou ending is quite clear in our MS, and thus they conjecture that it must have agreed with S here in what appears to be an orthographic peculiarity. Rahlfs notes only the sun/en variation.

line 19: The lower right foot of the expected number RM (140) is visible at the left side. Probably the name DAU(E)ID was abbreviated to DAD, as frequently happens. The "filler" marks noted in the previous heading do not appear here.

line 20: Read ekekraxa. Most witnesses transpose this word before the proV se (so also Rahlfs), while the order in our text is attested by B' La\G/.

lines 21-22: Our text agrees with Rahlfs and the majority of witnesses by presenting a single stichos here (line 22 is indented), where B T divide into two stichoi (mou / en).

line 25: Variant h eparsiV (R).

line 27: Variant eqou (S Sa).

lines 32-33: In Sa, these are presented as a single stichos.

2. The Exodus, Deuteronomy, and Isaiah Fragments: An Anthology?

One set of fragments that the editors thought may have come from a single 6th century codex or set of codices written in the same scribal hand, is found in their items 191, 192, and 194. The editors were a bit perplexed by the range of texts that they were able to identify in this set, since PAmh 191(a) was from Exodus 19 while PAmh 192 was from Deuteronomy 32, and they were unable to identify PAmh 191(b), which they did transcribe, although they were convinced that it also must have come from the same codex. Thus they imagined a relatively large work that may have contained the entire Pentateuch and perhaps even more.\3/

They did comment briefly on the unusual top line of the fragment designated PAmh 191(a): "at the top of the recto where the chapter [Ex 19] begins is a title of some kind." Regarding the unidentified PAmh 191(b) they noted that "it is from the lower part of a leaf, and the recto formed the conclusion of a chapter or section."

Not long after publication of the PAmherst 2 volume, A. Deissmann was able to identify the fragment designated PAmh 191(b) as from Isaiah 58.\4/ It is already included as such in Rahlfs' Verzeichnis (1914) where it received the Goettingen number 915 and it was used by Joseph Ziegler in his 1939 critical edition of Isaiah, although he seems unaware that the Amherst Papyri had already been relocated to New York City at the Pierpont Morgan Library many years before.

At least one of the three fragments designated PAmh 194 also comes from this same page of Isaiah, and that material will be transcribed below. Thus far, the other two fragments on the PAmh 194 plate remain unidentified.

The available clues suggest that these three passages -- Exodus 19, Deuteronomy 32, and Isaiah 58 -- were indeed from the same codex, but it was a codex of extensive excerpts rather than of complete biblical books. The heading at the top of the Exodus page ended with the words a]po merouV (in part), set off visually from what follows by the use of underlining and overlining. The bottom of the Isaiah page has the same distinctive features (filler marks plus under and overlining). How many other passages would have been included, and how extensive these three passages were, remain mysteries. A detailed investigation of the use of these passages in (Jewish and) Christian literature from that period might prove rewarding in this connection. The passages all relate to God's covenant with Israel, and Israel's response. This may well have been a theme deserving such special treatment among some (Jews and) Christians.

What remains of the page of Isaiah is as follows. PAmh 191(b) is the larger fragment (from the middle of the column), PAmh 194.l the smaller (from the margin, represented in upper case letters). There are no unambiguous variants from the text of Ziegler's Goettingen edition:

Isaiah 58.11-14 (end)

side 1 (against the fibers = verso)
01 ... pantoV   k]ai em[plhsqh {58.11}
02 sh       kaqap]er ep[iqumei h yu
03 CH[ sou kai ]ta os[ta sou pian
04 QH[setai k]ai est[ai wV khpoV
05 M[equwn         ]kai w[V phgh hn
06 MH[           exeli]pen[ udwr
07 KA[i      oikodo]mhq[hsontai sou {58.12}
08 ai erhmoi   a]iwn[ioi ...
(lower margin)

line 2: the ending of kaqaper is visible, precluding the well attested early variant kaqa (S*, Just., Tht., etc.)
line 4: the letters est are probable, rather than the variant esh (Ziegler's oI, L, etc. = MT)
line 5: there does not seem to be room for phgh udatoV found in many witnesses
lines 6-7: between 58.11 and 58.12 is a marginal horizontal stroke, and there is no room at the end of verse 11 for the long plus variant recorded in Ziegler's apparatus.

side 2 (with the fibers = recto)
[about 11 lines missing]

00 kai kalesiV ta sabbata truf {58.13b}
01 era agia t]w        [qw sou ouk
02 areiV         to]n pod[a sou ep ergw
03 oude lalh]seiV l[ogon en ]OR
04 gh    ek   tou ]stom[atoV     s]OU KAI {58.14}
05 esh       pepo]iqwV[   epi      k]N KAI
06 anabiba]se epi[ ta agaq]A THS
07 ghV   kai y]wmiei[   se     thn   ]KLH
08 ronomia]n iakw[b         tou       ]PRS
09 sou to gar ]stom[a   ku   elash
10 sen       taut]a >>>>>>[ . . .
  11 .   .   .   .   .   .  ] . . . . [
(lower margin)

line 1: after the final loop of omega, there is a blank space of about three letters width before the break (perhaps following qw rather than the definite article); none of the variants recorded by Ziegler suggest an explanation -- one could conjecture that the tetragrammaton was expected, although kuriw is not among the variants.
line 4: the final word of 58.13 is not clear, and possibly there was a mark of some sort before the start of verse 14 (kai).
line 5: there is not enough room for the full form of kurion here (see also line 9), so I have assumed that the normal nomina sacra of this period were used; see also lines 8-9 below.
line 8: the word patroV seems to have been abbreviated and overlined, although the fragment is not very clear at this point; there does not seem to be sufficient room for the unabbreviated word.
line 9: again, the normal nomen sacrum for kuriou is expected
line 11: the tops of some letters appear below the filler marks in line 10, but defy decipherment. We might expect them to indicate the end of the excerpt.

3. Fragments from the "A Text" of Judges

The original editors transcribed enough of one fragment from PAmherst 198 to permit identification by searching the TLG LXX/OG texts, which include both of the main Greek recensions of Judges ("A" and "B"). Apparently the original editors' tools were not sufficiently sophisticated to allow them to identify the "A Text" of that book.\5/ In any event, what we have in PAmherst 198 are several fragments of that recension, in a codex that the original editors dated to the 5th century CE. The format of the text is interesting, in that new sections (often corresponding to modern verses) often begin on a new line, leaving the preceding line shorter than the others. Whether there were also marginal markers at those points cannot be determined. Probably there is such a marker at Judg 13.11, but that may be because the new section began with the last word of the previous line. At Judg 3.20, the marginal alignment on the left also seems disrupted, perhaps to mark the new section ("verse"). Fragments from at least two folia at Judges 16 are preserved, with two columns of writing missing from between them, and it is not immediately clear whether another folio (containing a column on each side) originally came between these preserved sections, or whether the original pages may have contained double columns (see further below).

The codex contained about 32 lines per column, with an average of 20 letters per line, judging from the best preserved sections in chapter 16. In chapter 13, there may have been only 30-31 lines, and in chapter 3 (if that identification is correct), as few as 29 lines. The outer and lower margins were fairly wide (3-4 cm), and each column of writing (without the margins) would have been about 10 cm.wide and 18-19 cm tall. If it was a single column codex, the size would have been about 15 cm wide by perhaps 26 cm tall, similar to Turner's groups 6-8; if double columned, then perhaps almost 30 cm wide by 26 cm tall (which would be virtually unprecedented). Turner mentions some similarly formatted single column codices, such as his C 7 (P.Vogliano 5, Pauline Epistles, 5th c), C 30 (PChesterBeatty, Manichaean Psalm Book, 4-5th c), Lit 2 (6th c Euchologium), P 35 (3-4 th c Matthew), P 46 (3rd c Pauline Epistles), P 47 (3rd c Revelation), among others.\6/

As has been noted, the textual affinities of these fragments are with the "A-group" for Judges, which is conveniently printed separately in Rahlfs' Septuaginta, and reproduced in the TLG electronic materials. Close analysis of the exact alignment within the witnesses to the "A-text" must await further study, and depends in part on careful conjectural restoration of some of the lacunae, especially in Judg 16. My quick impression is that the PAmherst 198 fragments stand closest to the group AGbckxz\mg/ where there are significant variations in the "A-text."\7/

Fragment 1 recto (along the fibers), Judg 3.15

This small fragment (5 cm wide x 3 cm high) contains mostly blank margin, but also some interesting features. The recto almost certainly was filled with standard nomina sacra abbreviations (for "son," "Israel," and "Lord"), and the overlining for UIOU is still visible at the end of the last preserved line in 3.15. The ink remnants from the lower part of the end of the first preserved line are consistent with the expected KN (kurion), but are by no means unambiguous. If the proposed reconstruction is correct, 3.15 did not begin on a new line (see also 13.11, 16.23).

ekekraxan oi ui Ihl proV ]KN
kai hgeiren autoiV kV ]SW
thra ton Awd un Ghr]A UU

Fragment 1 verso (against the fibers), Judg 3.20

The left margin is 2.5 cm wide at the first preserved line, but uncharacteristically, the next line seems to be indented by the width of about two letters (almost 1 cm), perhaps suggesting that the beginning of the new verse on "line one" was actually extended into the left margin here. Traces of the bottom of a letter or two are visible above the "first line," but cannot be identified with confidence.

KAI AW[d eishlqen proV au
    TO[n kai autoV ekaqhto en

Fragment 2 verso (against the fibers), Judg 13.10b-11a

This fragment measures 3.7 cm wide by 5.2 cm high, with a preserved left margin (doubtless damaged) of only 1.5 cm and a lower margin of 4 cm on the verso. It is noteworthy that although the practice of starting new "verse" sections on new lines is not found here, there is a marginal mark to the left an slightly above line two, which may serve to indicate a sense break in the previous line, where our "verse 11" begins.

[moi o anhr o elqwn proV]
ME TH[ hmera ekeinh {11} kai
ANES[th Manwe kai eporeuqh
(lower left margin)

Fragment 2 recto (along the fibers), Judg 13.16(middle)
The final word of the last line in this column extends well into the margin and the final two letters are much smaller than usual. The margin for the preceding line is 1.5 cm, but for this line only 0.5 cm; the lower margin is about 4 cm, as on the other side. Rahlfs has anarthrous kuriw here, but if it was abbreviated to KW, as expected, the reconstructed last line would be unusually short (18-21 letters is normal, plus the overrun into the right margin), which suggests that tw k(uri)w appeared here, in agreement with the "B-text" and most witnesses to the "A-text" as well (not Aglnow Cyr2/3).

[ou fagomai twn artwn sou]
kai ean poihshV o]LOKAU
twma tw kw anoi]SEIS AUTO

(lower right margin)

Fragments 3 and 4 recto (along the fibers), Judg 16.6-8a

These fragments were once part of the same column, but there is a break of about 4 lines between them. The first and smaller piece is 3.2 cm wide by 2.3 cm high, and shows no margins; the second piece comes from the bottom of the column and is about 9 cm wide by 7 cm high, with a lower margin of 4.2 cm (the writing at the outer margin is not well preserved, but that margin seems to have been 3 cm or slightly more). The final line of 16.6 has space to the right, indicating a break between "verses." That may also be true at the end of 16.7, although the surface of the papyrus is too badly abraded to be sure. This material shows ample evidence of representing the main "A-text" in general.

00 {6} [kai eipen Dalida proV Sam]
01       ywn anaggei]LON[ moi en tini
02       h iscuV s]OU H M[egalh kai
03       en tini de]QHSH TO[u tapei
04       nwqhnai ]SE
05 {7} kai eipen p]ROS AUT[hn Sam
06       [ywn ean dhswsin me en Z]
07       [neuraiV ugraiV mh hrh]
08       [mwmenaiV kai asqenhsw ]
09       [kai esomai wV eiV twn]

10       [anwn]
11 {8} kai anhnegka]N AUT[h oi
12       satrapai tw]N AL[lof]U[lwn
13       Z neuraV ]UGRAS[ m]H H[rh
           (lower and right margin preserved)

line 0 {16.6}: in most Greek texts, Dalila (or Daleila) has become Dal(e)ida.

line 1: the bottoms of two or three letters are visible, but not easily decipherable; possibly LON.

line 4: after SE the remaining line is blank (about 4-5 letter widths).

line 5 {16.7}: the spacing is consistent with a new section starting here.

lines 6-8: nothing is preserved of lines 6-8, but assuming consistency with Rahlfs' A-text, these lines would be approximately as displayed above, although the number seven (line 6, epta) probably was represented by overlined Z (see also the final line of the column); traces of the bottom of line 9 and part of line 10 are preserved on the second fragment, but the latter is either badly abraded or was blank at that point (after the final word of the section, presumably an abbreviated form of anqrwpwn).

line 11 {16.8}: probably begins a new section.

line 12: these letters are very damaged, and cannot be read with confidence.

line 13: the letters UGRA are fairly clear (dieresis over the U), but the piece is badly damaged (and needs to be remounted) for the last part of the line; it is unlikely that the entire word hrhmwmenaV would fit on the line; the number seven (epta) at the start of the line probably was represented by overlined Z (see also above, line 6).

The lower margin including probably the right lower corner is preserved

Fragments 3 and 4 verso (against the fibers), Judg 16.11b-13a

00         [en oiV ouk egenhqh ergon]
01          kai asqenh]SW KAI[ esomai
02          wV eiV tw]N ANWN[
03 {12} kai elabe]N AUTW[ Dalida
04          kalwdia ka]INA. KA[i edhsen
05         [auton en autoiV kai eipen]
06         [proV auton oi allofuloi]
07         [epi se Samywn kai to ene]
08         [dron ekaqhto en tw tami]
09         [eiw kai diespasen auta]
10         [apo twn bracionwn autou]
11         [wV ramma
12 {13} ka]I EIP[en Dali]D[a proV Sam
              (lower and left margin preserved)

line 1 {16.11b}: traces of the bottom of a letter or two are visible above this line, but cannot be identified

line 2: the final word of the section (16.11), anqrwpwn, is clearly abbreviated and overlined

line 4 {16.12}: there seems to be a high dot punctuation mark after ka]INA

lines 5-7 are missing completely (as with the recto fragments)

lines 8-12 five lines have left traces on the badly damaged and abraded lower fragment, but hardly any letters can be read with confidence; possibly the last line began with kai eipen {16.13}, and the last visible letter on the line appears to be a D (probably part of Dalida or Daleida)

As with the other side, the lower margin and probably the lower left corner is preserved

[one leaf, recto and verso (or two inner columns), are missing here]

Fragment 5 recto (along the fibers), Judg 16.21b-23a

This piece contains the largest amount of writing among the fragments, and probably for that reason was called "the largest" by the original editors -- 6.5 cm wide by 4.5 high. The outer margin is about 4 cm wide on the verso (beginnings of lines), and varies from 3.5 to 4 cm on the recto (ends of lines). This fragment contains many readings characteristic of the "A-text," which deserve detailed treatment elsewhere. A short line clearly ended 16.21, but 16.22-23 are run together.

              [kai kathgagon auton eiV]
01          gazan kai edhsan a]UTON
02          en pedaiV calkaiV] KAI
03          hn alhqwn en oik]W THS
04          fulakhV
(blank)] (blank)
05 {22} kai hrxato h qrix thV] KEFA
06          lhV autou anateil]AI HNI
07 {23}  ka exurhqh kai oi sa]TRAPAI
08          twn allofulwn sunhcq]H
09          san tou qusai qusian ...
               (about 6 lines more to bottom of column)

line 4 is blank in the preserved fragment, presumably as the end of the previous section (16.21)

line 5 {16.22} seems to have an overline above the letters EF (perhaps an elongation to the left of the top stroke of the F)

Fragment 5 verso (against the fibers), Judg 16.26

01 {26} KAI EIPE[n Samywn proV to
02          PAIDA[rion ton ceiragwgoun
03          TA AUT[on epanapauson me
04          DH KA[i poihson yhlafh
05          SAI ME[ epi touV stulouV ef'
06          WN O O
[ikoV epesthrikta
07          EP AUT[wn kai episthriso
             [mai ep autouV . . .]
              (about 6 more lines to bottom of column)

line 1 is badly damaged, but probably was the start of 16.26 as reconstructed here; approximately 17-18 lines would have preceded it in this column

Fragments 6 and 7 (unidentified)

Two small fragments also mounted in the PAmherst 198 glass plate remain unidentified. I will call the largest fragment 6 -- it is 2.4 cm wide by 3.1 high, with portions of 5-6 lines from the center of a column (no margins visible). The hand is very similar to the Joshua fragments, with possible differences in the shape of upsilon and epsilon. I have not been able to identify either fragment by searching the TLG electronic texts of LXX/OG.

On the recto (along the fibers) of fragment 6:

line 1: the bottom of a vertical stroke, perhaps iota or nun or eta (among the numerous possibilities)
line 2: UO -- a clear upsilon followed by a curved left part of a letter (sigma, omicron, omega; probably not epsilon or theta)
line 3: ARM -- alpha, rho, and a damaged letter or letters resembling upper parts of a mu
line 4: TAE -- faint strokes (upsilon?) followed by tau, alpha, and a rounded letter (epsilon, theta, perhaps omicron or omega)
line 5: LIPA -- part of an alpha or lamda, iota, pi, alpha
line 6: traces of ink from the top of the next line

verso (against the fibers), fragment 6

line 1: nothing visible in the topmost position
line 2: M -- a large mu
line 3: MEM -- mu, epsilon, and the start of another mu (on nu or eta) partly covered by a small scrap
line 4: SEST -- very clear sigma, epsilon, sigma and a trace of the top stroke of the next letter, possibly tau
line 5: NHQ -- nun, then a less clearly written nun or mem or eta, followed by epsilon or theta

Fragment 7 is smaller still, 1.2 cm wide by 2.2 cm high, with no margins showing.

recto (along the fibers)
line 1: IS -- a vertical stroke (top missing) and the start of a rounded stroke (top and right side missing)
line 2:     -- apparently blank, perhaps the end of a section of Joshua?
line 3: T -- possibly the right side of a kappa, and/or the left part of the top stroke of a tau
line 4: WM -- three angled tops of letters, such as mu (or even the right side of omega) followed by nun or mu

verso (against the fibers)
line 1: AO -- the lower right of possibly alpha or lamda followed by omicron (or theta)
line 2: ES -- epsilon, and almost certainly sigma (or omicron)
line 3: H -- badly damaged, but there seems to be a horizontal mid stroke as in epsilon or eta
line 4: M -- the top of a large mu


For the scholar, young or old, who likes puzzles and the thrill of discovery, there are numerous fragments in the various collections that can now be identified by using the available electronic texts and tools. Many of these are listed by Joseph van Haelst in his 1976 Catalog.\8/ Obtaining legible images, digitized at a suitable resolution and in color if possible, is becoming easier, but is still the main obstacle to making significant progress. Hopefully projects such as the Advanced Papyrological Information System (APIS) will help solve even that problem!\9/


\1/ The Amherst Papyri, being an account of The Greek Papyri in the collection of the Right Hon. Lord Amherst of Hackney, F.S.A., at Didlington Hall, Norfolk, by Bernard P. Greenfell and Arthur S. Hunt. Part I: The Ascension of Isaiah, and other Theological Fragments, with nine plates (London 1900) and Part II: Classical Fragments and Documents of the Ptolemaic, Roman and Byzantine Periods, with an Appendix containing additional Theological Fragments; twenty-five plates (London 1901). I would like to express my appreciation to Dr. William Voekler and the staff at the Pierpont Morgan Library in New York City for helping me to obtain photos and electronic images of pertinent items in the collection. For a more general survey of some of those materials, see my forthcoming article "The Amherst Papyri Revisited: Fragments of LXX/OG MSS" in a 2003 volume of Orientalia Louvaniensia Analecta
from Peeters Press (also on my Amherst Papyri web page, which also includes some images: ).

\2/For the Psalm materials, textcritical information is dependent on Alfred Rahlfs, Psalmi cum Odis (Goettingen Septuaginta 10; Vandenhoeck & Ruprecht, 1967\2/). I am indebted to the careful work of Bryan Harmelink, a doctoral student at the Westminster Theological Seminary, in attempting to place more of these small fragments. He is almost certainly correct in locating fragment 6 on the original plate as also from the page represented by PAmherst 6(d), just above the larger fragment from PAmherst 200 noted below. These are items 1 and 4 on the rearranged images of PAmherst 200 found at -- top and bottom of the left hand vertical row. Once that page of the original papyrus could be reconstructed visually, some other small pieces also seemed to fall into place, as can be seen at

\3/Regarding PAmh 191(a) [Ex 19]: "The fragment is from a very handsome MS., written in large round uncials of calligraphic type, of about the sixth century. The following fragment of Deuteronomy (192) is in a similar handwriting, and probably it as well as the unplaced piece printed below [191(b)] and three smaller fragments (194) belonged to the same manuscript, which may have extended to several volumes including the whole of the Pentateuch. The ink is of the brown colour common at this period." To PAmh 191(b): "The following fragment is certainly from the same MS., and should be readily identified, but we have not succeeded in doing so." On PAmh 192 [Deut 32, two fragments]: "The large upright hand in which these fragments are written is apparently identical with that of the preceding fragment of Exodus, and they probably belonged to the same manuscript." Finally, for PAmh 194: "Three small fragments from a papyrus book ... in a large uncial hand resembling that of 191 and 192, perhaps forming part of the same manuscript."

\4/I am indebted to Detlef Fraenkel of the Goettingen Septuagint Project for this information. See A. Deissmann, Beilage 251 of the Allgemeinen Zeitung Muenchen from 31 October 1901. Deissmann reconstructed the Isaiah fragment, and this information was included in C. Schmidt's "Referat Christliche Texte" in APF 2 (1903) 382. On this basis, Alfred Rahlfs listed the fragment in his Verzeichnis der griechischen Handschriften des Alten Testaments (Berlin 1914); Rahlfs quotes the original editors about the relationship of the two fragments included under the 191 label and comments "that a papyrus book had contained both Exodus and Isaiah is not very probable." He does not note Schmidt's observation that in a Christian liturgical context, Isaiah 58 may have connections with the Exodus passage (ibid). For Isaiah, textcritical information is based on Joseph Ziegler, Isaias (Goettingen Septuaginta 14; Vandenhoeck & Ruprecht, 1939).

\5/"Six fragments, the largest measuring 4.2 x 6.2 cm., from a papyrus book. ... About the fifth century A.D." The present plate numbered 198 contains seven fragments, one of which is too small to place with any confidence, and another of which does not seem to come from the A text of Judges (see below, "fragment 6"). Textcritical information for the book of Joshua is based mainly on Alan England Brooke and Norman McLean, The Old Testament in Greek 1.4 ("Cambridge Larger Septuagint"; Cambridge University Press, 1917); see also Alfred Rahlfs, Septuaginta (Privileg. Wuertt. Bibelanstalt, Stuttgart, 1952\5/ [1935]).

\6/Eric G. Turner, The Typology of the Early Codex (University of Pennsylvania Press, 1977), especially "Table 1: Papyrus Codices Grouped by Dimension," and "Table 16: Consolidated List of Codices Consulted."

\7/For a brief and convenient summary of the textual situation in Judges, see Walter Ray Bodine, The Greek Text of Judges: Recensional Developments (Harvard Semitic Monographs 23; Scholars Press, 1980) 2-3: he notes that the group A(G)abc(k)x has been identified by several scholars (e.g. Pretzl, Billen, Soisalon-Soininen) as the primary witness to the "A-text."

\8/Joseph van Haelst, Catalogue des Papyrus Litte/raires Juifs et Chre/tiens (Paris 1976), chapter 10 ("Textes non identifie/s" = ##1082-1190; see also ##1214-1215 [Latin] and 1223-1226); this provides an excellent starting point from the mid 1970s. I hope to summarize this information and supplement it on an appropriate internet folder on my own papyri page. The temporary location of this material is gopher://

\9/For information on the APIS project, see