Chapter 6 Finding Information on the Verso

In: Material and Digital Reconstruction of Fragmentary Dead Sea Scrolls
Jonathan Ben-Dov
Search for other papers by Jonathan Ben-Dov in
Current site
Google Scholar
Asaf Gayer
Search for other papers by Asaf Gayer in
Current site
Google Scholar
, and
Eshbal Ratzon
Search for other papers by Eshbal Ratzon in
Current site
Google Scholar
Open Access

The task of editing fragments is not complete without examining their verso, where important clues may be found for the reading and restoration: folded parts of the recto, additional ink marks, unnoticed layers, imprints of seams from adjacent layers, etc. The LLDSSDL provides images of the verso upon request for the entire DSS corpus, and the SQE platform now provides them for the entire corpus. Such images were carried out for only a few scrolls in the PAM collection, when ink signs were evident or in the case of an opisthograph.

Since the marks on the verso are frequently difficult to spot, and are sometimes also covered with Japanese paper as part of the restoration process, it is essential to enhance the verso images with digital filters in order to extract maximum visual information from them (for the correct way to perform this process, see chapter 3). It is also important to compare these finds with the signs on the recto. This chapter surveys the types of signs that can be found on the verso, most of which have been discussed in previous scholarship. However, here we concentrate much of the information in one place, treating the fragment as a three-dimensional artifact rather than a two-dimensional text. In addition, we offer new methods for distinguishing the origins of textual information that was found on the verso, which bears consequences for the material and textual reconstruction.

1 Modern Stamps

One of the most conspicuous ink marks on several scrolls’ verso is Latin letters, stamped in modern times. These letters include: G, S, A, R, H, B, V, T, J, and E, which indicate the institution that had purchased the fragment before submitting it to the PAM.1

2 The Title of the Composition

In some scrolls the title of the composition was written by ancient scribes on the verso of the very beginning of the scroll, conveniently indicating the content to the user, who would see it while searching for a specific scroll at the library. Two preserved examples are: the name מדרש ספר מושה written in square script on the verso of 4Q249 1 and the title דברי המאורות written on the verso of 4Q504. 4Q257 is another possible example.2

The following marks are crucial for the scroll’s reconstruction.

3 Opisthographs

While most scrolls were written only on one side of the surface, usually the hair side of the skin or the side of the papyrus on which the fibers run horizontally,3 some scrolls were inscribed on both sides. In such cases, each side of the scroll received a separate catalogue number, for example 4Q414–4Q415 (skin) and 4Q503–4Q512 (papyrus). Such scrolls are named opisthographs. In order to write the opisthograph the scroll was flipped either horizontally or vertically.4 In these cases the verso is as important as the recto. When suggesting a material or textual reconstruction, one must take into account the text and material of both sides of the scroll.5 Algorithms that help the scholar reconstruct both sides together already exist and may be included in the SQE platform in the future.

4 Evidence for Additional Layers

The phenomenon of fragments preserved in wads occurs more frequently than previously acknowledged in scholarship. While the original scholars did their best to separate the fragments from their piles, some layers were left unnoticed underneath another fragment, or separating them was not possible without damaging the artifacts. Thus, what at first sight seemed like cracks in the skin, may upon closer examination turn out to be evidence for additional layer(s) attached underneath the fragment. Here we bring one small example for such a layer visible on the verso from 4Q397 6, which is part of a copy of Miqṣat Maʿase HaTorah (figure 9). Another example for such layers in the copy 4Q418a of Instruction is frag. 22 (figure 10, discussed in chapter 15).

Figure 9
Figure 9

Left: 4Q397 6 verso. Right: A close-up of the encircled area. Signs of a small additional attached layer are visible on the verso. The identification of the layer attached to this fragment was done as part of the preparation of a new edition for MMT: Vered Noam, with decipherment and reconstruction by Eshbal Ratzon, 4QMMT: Some Precepts of the Torah (Oxford: Oxford University Press, forthcoming).

© IAA, LLDSSDL, Shai Halevi
Figure 10
Figure 10

4Q418a 22 verso. Left: parts of the ink signs are covered. Right: raking light (left) image demonstrates that the ink is covered with another layer of skin (red circle).

© IAA, LLDSSDL, Shai Halevi

A thorough discussion of this phenomenon and its implications can be found in chapter 7; here we briefly mention it as part of the marks found on the verso. Identifying such layers, even without being able to read the text written on them, is crucial for the reconstruction of the scroll as it suggests how much text is missing in certain places. The identification requires examining every available image, especially the older PAM images. With regard to the new multispectral images, color (composite) and raking light images are particularly helpful. When the possibility arises that more than one layer existed, it is advisable that the scholar examine the actual fragment under a dino-lite microscope that allows viewing the fragment from different angles.

5 Stitching Impressions

Most scrolls constitute several sheets sewn together. When rolled tightly, the stitches are pressed against the next layer. Finding their impressions on the recto or verso of a fragment may help in reconstructing its place in the original scroll if the actual stitches were preserved on another fragment. Conspicuous examples for this phenomenon include 11Q10 (Targum Job) and 11Q19 (the Temple Scroll), where the stitching and its impression are clearly visible.6

6 Mirror Writing

Unlike in the case of opisthographs, at times a text is found on the verso in mirror writing. Such writing can stem either from bleeding from the recto or from an imprint of an adjacent layer. To be precise, three scenarios are possible:

Option 1: Bleeding of ink from the recto of the same fragment.

Option 2: Bleeding of ink from the recto of another layer, still attached to the verso of the upper fragment.

Option 3: Imprint of ink from the next layer, no longer attached to the current fragment.

Physically, the first two options are similar. In many cases – particularly in thin skins – the ink penetrates through the skin from the recto to the verso. This phenomenon is called bleeding. Many of the fragments of 4Q417 and 4Q418a can serve as examples for this phenomenon. The third option is different, created by the attachment of ink from a subsequent layer. This phenomenon is called imprint. From the reconstruction point of view, options 2 and 3 are similar because they mean that the text on the verso does not belong to its immediate recto but rather to the preceding layer(s).

Despite the difficulties in identifying the origin of the ink, such identification is crucial for the reconstruction of the scroll and for the correct positioning of its text in the right order. Here we offer some methods for differentiating the three kinds of mirror writing on the verso. Identification of the first option, i.e. bleeding of ink from the recto, is the simplest. One should compare the verso image to the recto. Since in this case the script is seen as mirror writing, the image must be flipped on the horizontal axis with digital means. In option 1, the signs on the verso correspond to the writing on the recto (as can be seen in figure 11). This correspondence is helpful, for example, in cases when the script on the recto side is not visible or is illegible, either partly or entirely, while the verso preserves a better view of the letters.

Figure 11
Figure 11

4Q418a 12 recto and verso. Left: 4Q418a 12 recto IR image; Middle: 4Q418a 12 verso IR image enhanced; Right: 4Q418a 12 verso IR image enhanced and flipped horizontally.

© IAA, LLDSSDL, Shai Halevi

If a correspondence between the signs on the recto and the verso was not found, one should consider either option 2 or option 3: either it is bleeding from another layer, still attached underneath the original fragment, or it is an imprint of ink from a layer that is no longer attached. Option 2 can occur when two layers of skin remain attached. In this case, the recto shows the writing of the upper layer, while the verso attests to the writing of the bottom layer (see chapter 7). Option 3 occurs when the mirror writing on the verso appears not through bleeding, but rather because the letters imprinted from the next layer, a layer which is either lost or preserved separately. A classic example for this phenomenon is 11QTa, which shows a large amount of imprinted text on the verso of many of its columns.7

In both options 2 and 3, the ink signs on the verso should be compared with the preserved text on the next layer. If it exists there, the ink is definitely an imprint. Full correspondence is a rather rare occasion, however.8 When the text of the next layer is not known, several other methods can be used to distinguish ink bleeding from imprint:

  1. When there are other fragments of the same scroll that show clear bleeding or impression of ink, it is likely (but not necessary) that the other cases of ink on the verso are due to the same phenomenon.

  2. The ink signs which bled from the recto tend to appear in a pale color, with blurred boundaries of the letters. In contrast, the signs attached from the next layer show a darker color of the letters with sharper boundaries.9

  3. When signs of additional layers can be traced on the verso, ink signs that do not correspond to the recto are most likely bleeding from a still attached layer, rather than imprints from past attached ones.

Among the copies of Instruction, 4Q417 presents ink traces on the verso that show both bleeding from the recto and imprint from the adjacent layer.10 In order to check whether the signs on the verso correspond to those on the recto, the image of the verso (after being horizontally flipped) should be superimposed on that of the recto as layers on a canvas. Decreasing the opacity of the verso will thus enable easy comparison of the layers. Those signs on the verso that correspond to the recto will prove to be bleeding signs, while the rest of the signs must have originated from the adjacent layer.

Figure 12 shows the result of the complete procedure, accompanied by auxiliary notation. As this image shows, a series of horizontal marks (marked blue in figure 12) on the verso of 4Q417 1 does not correspond to the lines of the recto and must therefore have been attached from the next layer. These marks all stand sequentially in a clear vertical line, indicating the lines of that next layer. The attachment of ink from the outer layer occurred only in a certain area of the scroll, possibly due to humidity or pressure affected at this point. These points of pressure may serve as significant marks for the material reconstruction of the scroll.

Figure 12
Figure 12

4Q417 Fragment 1 recto + verso, a close-up on the ink marks. Lines which correspond to those of the recto side (diminished opacity) are marked red. The remaining signs, which do not correspond to the recto, show consecutive signs, marked in blue. These lines originate from the next layer of the scroll. Graphics: Anna Shirav

© IAA, LLDSSDL, Shai Halevi

The two kinds of writing on the verso (bleeding, imprint) may sometimes overlap, thus disturbing the reading. Imaging and digital tools have been used in the past for enhancing the reading in such cases.11 To sum up, at times the verso in no less important for the reading and reconstruction of the scroll than the recto, and is now made available by the IAA on the SQE platform. In the present chapter we offered some methodology for analyzing the verso’s finding and using them in the reconstruction.


Tov and Pfann, Companion Volume, 16.


Tov, Scribal Practices, 118–22. For a fuller discussion see Jonathan Ben-Dov and Daniel Stökl Ben Ezra, “4Q249 Midrash Moshe: A New Reading and Some Implications,” DSD 21 (2014): 131–49. We do not discuss here the many cases of titles written as the first few words of the composition. Nor do we discuss 1QS, 4QGenh, where the title is written on the recto of the handle sheet.


See Mordechai Glatzer, “The Book of Books-From Scroll to Codex and into Print,” in Jerusalem Crown: The Bible of the Hebrew University of Jerusalem, ed. M. Glatzer (Jerusalem: Ben-Zvi Printing Enterprises, 2002), 61–101; Tov, Scribal Practices, 32–33. For example, the Temple Scroll 11QTa (11Q19) was written on the flesh side of what is probably split skin; see Roman Schuetz et al., “The Temple Scroll: Reconstructing an Ancient Manufacturing Practice,” Science Advances 5.9 (2019). doi: 10.1126/sciadv.aaw7494.


See Perrot, “Reading an Opisthograph at Qumran,” 101–14. According to Perrot, horizontal flipping is carried out by the same author of the recto, who aims for both sides to be read continuously, while vertical flipping is carried out by a later scribe.


For reconstructions that consider both sides, see, for 4Q503, Francis Schmidt, “Le calendrier liturgique des ‘Prières quotidiennes’ (4Q503), en annexe: L’apport du ‘verso’ (4Q512) à l’édition de 4Q503,” in Le temps et les temps dans les littératures juives et chrétiennes au tournant de notre ère, ed. Christian Grappe and Jean-Claude Ingelaere (Leiden: Brill, 2006), 55–87; for liturgical papyri, see Daniel Falk, “Material Aspects of Prayer Manuscripts at Qumran,” in Literature or Liturgy? Early Christian Hymns and Prayers in Their Literary and Liturgical Context in Antiquity, ed. Clemens Leonhard and Hermut Löhr (Tübingen: Mohr Siebeck, 2014), 33–87, and for 4Q433a/4Q255, see Aksu, “A Palaeographic and Codicological (re)assessment,” 170–88. For more information about opisthographs from Qumran see Tov, Scribal Practices, 68–74.


Examples for clear impressions can be seen on the Temple Scroll on the intercolumnar margin between columns XXVIIXVIII and on column XXXVI, see Stegemann, “Methods for the Reconstruction,” 195. For examples of using this information, see Florentino García Martínez, Eibert Tigchelaar, and Adam S. van der Woude, Manuscripts from Qumran Cave 11.II (11Q2–18, 11Q20–30), DJD XXIII (Oxford: Clarendon Press, 1998), 101, 163, etc. Other examples for stitching impressions can be found in Tucker and Porzig, “Between Artefacts” and Matthew P. Monger, “4Q216 – A New Material Analysis,” Semitica 60 (2018): 303–33. Monger attributes the straight boundaries of the 4Q216 fragments to pressure caused by the stitching, but these breaks could have been caused by other reasons as well.


See Yigael Yadin, The Temple Scroll (Jerusalem: Israel Exploration Society, 1983), 6–7; and more recently Qimron, The Dead Sea Scrolls, 1.137–38; Torleif Elgvin and Emanuel Tov, “422. 4QParaphrase of Genesis and Exodus,” in Qumran Cave 4.VIII: Parabiblical Texts, Part 1, DJD XIII, ed. Harold W. Attridge et al. (Oxford: Clarendon Press, 1994), 417–18.


For example, the identification of the imprinted text on the verso of 4Q377 remains debated. James VanderKam and Monica Brady, “377. 4QApocryphal Pentateuch B” in Wadi Daliyeh II: The Samaria Papyri from Wadi Daliyeh / Qumran Cave 4.XXVIII: Miscellanea, Part 2. DJD XXVIII. Ed. Douglas M. Gropp (Oxford: Clarendon Press, 2001), 205–17, esp., 205 read it as matching text from the recto of another fragment, but Qimron (The Dead Sea Scrolls, 3.141) doubts it. The verso of 4Q377 preserves both text seeping from the recto and imprints from the adjacent layer.


Compare, for Genizah documents, the discussion by Eric D. Reymond, “New Hebrew Text of Ben Sira Chapter 1 in MS A (T–S 12.863),” RevQ 27.1 (2015): 83–98, esp. 83–84.


Eibert Tigchelaar prompted us to examine the verso of these fragments, and the work was carried out together with Anna Shirav. For additional examples see Tov, Scribal Practices, 37–38.


Keith Knox, Robert Johnston, and Roger Easton, “Imaging the Dead Sea Scrolls,” Optics and Photonics News 31 (1997): 30–34.

  • Collapse
  • Expand