How Many Complete Greek New Testament Manuscripts Are There REALLY?

By Katie Leggett

For decades, the standard answer to the question, “How many complete Greek New Testament manuscripts do we currently have?” has been “about 60.”

But what, precisely qualifies a manuscript as “complete”? And do 60 manuscripts really meet that standard?

In what follows, I examine these questions, reassessing the commonly cited estimate in light of the extant manuscript evidence. As will become clear, the answer is much more complicated than it seems—ultimately, it hinges on how one defines “complete.”

I. Where Does the Number 60 Come From?

The familiar estimate of “about 60” complete Greek New Testament manuscripts derives from the 1963 Kurzgefasste Liste. There, Kurt Aland—building on the system of Caspar René Gregory—used the classification eapr as a cataloguing shorthand for manuscripts containing all 4 major divisions of the New Testament:

e = Evangelia (Gospels) 
a = Acts and Catholic Letters  
p = Pauline Epistles (including Hebrews and Pastoral Letters)
r = Revelation 

A manuscript was thus considered “complete” if it preserved at least some portion of each of these 4 sections. The 1963 Liste identified 58 such manuscripts.¹ A few later additions and reclassifications raised the total slightly, and the figure stabilized at “about 60.”

In subsequent scholarship, these manuscripts were often described as containing “the whole” or “the entire” New Testament. For instance, Kurt and Barbara Aland noted that 3 majuscules and 57 minuscules “contain the whole of the New Testament.”² J.K. Elliott similarly refers to 61 manuscripts that “contain the entire New Testament canon,”³ and Bruce Metzger writes that “only about 60 manuscripts … contain the entire New Testament.”⁴

While this round number has been widely repeated for decades, it is sometimes understood to extend beyond what the eapr classification was meant to indicate. The classification does not specify:

II. Three Meanings of Completeness

It seems clear that the term complete is used in more than one sense when applied to Greek New Testament manuscripts although this distinction has not always been clearly articulated. As a result, widely cited figures for the number of “complete” manuscripts can be misleading, since they often rest on differing—and sometimes unstated—assumptions.

At least three distinct meanings of completeness can be identified:

These three meanings correspond to different questions about classification, original production, and current preservation. Although conceptually distinct, they are frequently conflated. Consequently, the widely cited figure of “about 60 complete manuscripts” is based on categorical completeness, though it is often interpreted as referring to extant completeness.

III. Definitions Change the Numbers

These definitional differences are not merely theoretical; they directly affect what is counted as complete. Defining “complete” in terms of original completeness⁶ led, in the early 2010s, to the removal of numerous manuscripts from the eapr category in the Liste,⁷ because Revelation in these manucripts had been added later or was codicologically distinct from the rest of the codex. Thus, Ulrich Schmid counted only 52 manuscripts that he believed were originally designed to contain all 27 books and identified 7 additional for which it remains uncertain whether their parts once belonged to a single unified codex.⁸

J. Eldon Epp approached the question by working with multiple definitions of completeness. In a 2007 essay, he initially defined a manuscript as complete “in the sense of containing all twenty-seven books,” excluding 4 manuscripts that lack entire New Testament books.⁹ He then refines the category further by excluding manuscripts that were not originally intended to be complete codices, arriving at a total of 53 “likely complete” New Testament manuscripts.¹⁰

The central difficulty, then, is that the meaning of complete is not self-evident. Without specifying which sense is intended, numerical claims risk being misleading, since they may be answering a different question than the one many people actually have in mind.

IV. The Problem of Incomplete “Complete” Manuscripts

Once these distinctions are in place, the problem becomes clear: a manuscript can be classified as eapr and counted as “complete,” even though, in its current state, it does not preserve the entire New Testament.

In some cases, textual losses are minor; in others, they are extensive. GA 1384, for example, lacks 2–3 John altogether and is also missing substantial portions of 2 Corinthians, Titus, Philippians, and Revelation. GA 04 (Codex Ephraemi Rescriptus) is highly fragmentary, lacking 2 Thessalonians and 2 John completely, with considerable omissions across nearly all books, Even more strikingly, GA 886 lacks the Catholic Epistles entirely, yet it is still classified as eapr.

These examples are not isolated. A substantial proportion of manuscripts classified as eapr contain significant lacunae, often affecting large sections of text and in some cases entire books.

The issue is not that earlier scholars were unaware of such gaps. In the Liste, for example, missing material was sometimes indicated with a cross, but this mark gave no sense of scale: a manuscript so marked may be missing as little as a single folio or as much as several New Testament books. It was also not implemented consistently, leaving the impression that manuscripts not marked with a cross had no lacuna. This symbol has since been discontinued, though traces remain in the NTVMR.

To summarize: a manuscript can labeled as “complete” in the Liste  while lacking substantial portions of the New Testament. The eapr classification remains useful for broad cataloguing, but it does not tell us which manuscripts actually preserve the complete text today.

V. Recounting the Manuscripts in Light of Extant Completeness

To pursue this question, I’ve begun an evaluation of manuscripts traditionally classified as eapr focusing on extant completeness—that is whether they preserve the full text of all 27 books without substantial lacunae. To my knowledge, the eapr corpus has not yet been systematically assessed from this perspective. This post offers an initial step in that direction.

More than 60 years after the publication of the first Liste, and in light of subsequent scholarship, digitization, and the development of digital tools for indexing and transcriptions, it is now possible to approach this question with a new level of precision.

A. Defining the Corpus

A search in the NTVMR currently returns 60 manuscripts classified as eapr, but 2 of these (GA 296, 2136) have recently been removed after being identified as copies of a printed edition.¹¹

An additional 3 lack images and thus cannot be categorized in any meaningful way right now.¹²

Of the remaining 55 manuscripts, 40 are still only available on microfilm.¹³ Of the 55, 10 eapr manuscripts have color digital images available in the NTVMR. 

A further 5 manuscripts (GA 69, 367, 757, 808 and 1424) were digitized by CSNTM and have been made publicly available here.

B. Categories for Evaluating Extant Completeness

To begin assessing the extant completeness of these 55 manuscripts, I propose three categories:

1. Incomplete Manuscripts

25 manuscripts (45%) are demonstrably incomplete:

GA 02, 04, 69, 175, 218, 339, 498, 506, 517, 522, 699, 757, 886, 922, 935, 986, 1094, 1384, 1424, 1617, 1626, 1652, 2200, 2201, 2495.

This group represents the largest clearly identifiable group and illustrates how frequently “complete” manuscripts do not preserve the full New Testament text in practice.

Below some brief observations:

1. Lacunae are more extensive than indicated 

2. Extent of loss typically substantial 

3. The beginning and end of the codex are most vulnerable

4. Some lacunae are caused by scribal error, not physical loss

Examples

GA 339 (extreme fragmentation)

GA 339 demonstrates that the extent of a manuscript’s preserved text cannot be determined from the Liste alone. Although it is recorded as containing 200 folios and is not marked as incomplete, a note indicates that it is fire-damaged and survives only in fragments. Around 2010, staff at the INTF began working through the Turin fragments, identifying and indexing them in the NTVMR. 

There are now 91 indexed fragments containing New Testament text. Although this reconstruction is a significant achievement, only 10 New Testament books are extant, with most of the rest apparently lost. While this is an extreme case, it highlights how fragmentary some eapr manuscripts are and how a reliable picture of their preserved contents often emerges only through detailed analysis.

GA 1424 (minimal loss)

GA 1424, a ninth-century commentary manuscript, is the earliest of the eapr minuscules. It was removed in 1917 from the Kosinitza Monastery near Drama in Greece, later entered private hands, and eventually became part of the collection at the Lutheran School of Theology at Chicago.  In 2010, it was returned to Kosinitza. Although nearly complete, I noticed it lacks the folio containing Matthew 1:24–2:15. (See images of 1424 here.) The loss predates foliation and it is unclear when or by whom it was removed.

These two examples illustrate opposite ends of a spectrum on which manuscript incompleteness ranges from minor loss to extreme fragmentation.

2. Preliminarily Complete Manuscripts

21 manuscripts (38%) may be considered preliminarily complete:

GA 201, 205, 367, 386, 582, 680, 808, 824, 905, 1072, 1075, 1248, 1597, 1637, 1678, 1704, 1780, 2352, 2382, 2494, 2554.

These manuscripts evince no obvious lacunae, but firm conclusions can only be made after complete indexing is undertaken. 

A few brief observations:

1. Unclear whether gaps reflect defective microfilm or missing folios

In several cases, it is difficult to determine whether apparent gaps are due to physical loss or problems with microfilm

2. Unregistered New Testament material

Some manuscripts appear to contain New Testament content not yet recorded in the Liste:

3. New additions to the eapr category 

While no new eapr manuscripts have been discovered since GA 2554 was added in the late 1950s, two manuscripts have now been added to this category after research showed that material previously catalogued separately in the Liste in fact belong to the same codex.

Example: GA 905

GA 905 was previously divided across three locations with separate GA numbers

These dispersed parts have now been recombined into a single manuscript. While a few leaves appear supplemented with replacement leaves,¹⁵ no obvious lacunae are evident. Full indexing, however, is still required. Microfilm images of the Gospel portion are available in the NTVMR under GA 905, while the remaining images are still visible under GA 1795.

Example: GA 2382

Similarly, GA 2382 has been identified as belonging together with GA 2723 from Dousikou Monastery.¹⁶ The latter has now been merged under GA 2382. The portion housed at the Morgan (MS M. 340) contains the Gospels, while the Dousikou Monastery portion preserves the remaining books. At present, images of the Dousikou portion remain visible under the former GA number in the NTVMR.

3. Verified Complete Manuscripts

Thus far only 9 manuscripts (16%) have been verified as complete:

GA 01, 18, 35, 61, 141, 149, 664, 1503, 2886.¹⁷

These 9 manuscripts can be viewed here, for each the folio with the beginning of Romans

These manuscripts have been fully indexed and confirmed to contain the entire New Testament without lacunae.

1. A clear chronological pattern emerges 

So far these are heavily concentrated in the later Byzantine period. Most are written on parchment; only GA 61 and GA 664 are on paper. None of these are commentary manuscripts. All have a single-column text-format, except for GA 01 (4 columns) and GA 1503 (2 columns).

2. Extant completeness requires careful qualification 

Relatively few manuscripts in this group correspond to what we typically imagine as a “complete” New Testament 

Example: GA 141

GA 141 at the Vatican was registered without noting lacuna—but on closer inspection, it is actually missing 7 folios (Matt 7:3–11:21). Further investigation shows that GA 866 (Vat. gr. 1882, fols. 10–16) seems to be a match for these missing folios.

It is clear that these two manuscripts were originally part of the same codex. GA 141 and GA 866 should therefore be combined in the Liste, which would effectively yield a complete manuscript with no textual gaps.

This single case illustrates the importance of precise attention to individual lacunae, both for determining what is extant in each manuscript and for identifying and reuniting scattered fragments.

VI. Conclusion

This post has shown why the seemingly straightforward question about complete Greek New Testament manuscripts is more complicated than it initially appears, arguing that the traditional answer of about 60 complete manuscripts is based on the number of eapr entires in the Liste and cannot be equated with how many witnesses actually preserve the complete New Testament. 

When the question is framed in terms of manuscripts that preserve the full New Testament text, the number is considerably smaller. At present, of the 55 eapr manuscripts being evaluated, 25 are incomplete, and only 9 are fully extant. An additional 21 show no obvious lacunae but have not yet been systematically indexed. Given the patterns observed here, it is likely that 30–40% of these 21 will be found to be lacunose once they are fully indexed.

So, to answer the question posed at the outset, there are likely only about 20–25 manuscripts that preserve the whole Greek New Testament. While this remains a provisional conclusion, it represents the best answer currently attainable in light of the manuscript evidence.²⁰

The relative scarcity of complete New Testament manuscripts (considering these comprise less than 1% of the 5,711 total number of Greek New Testament manuscripts) can be explained by a convergence of historical and material factors. In the earliest centuries, New Testament writings did not typically circulate as a single, fixed collection, but rather as smaller groupings. The 4 categories of eapr (Gospels, Acts and Catholic Letters, Pauline Epistles, and Revelation), broadly reflect how these books were transmitted. Revelation, in particular, often circulated separately, perhaps reflecting its more limited—and at times contested—use in liturgy, as well as its place at the end of the canon.²¹

In the early centuries, this pattern of New Testament books circulating independently reflects the gradual formation of the canon and, for a much longer period thereafter, the practical realities of book production. The creation of a full twenty-seven-book codex required substantial resources, including materials and skilled labor, and was therefore undertaken only in rare circumstances. Taken together, these factors help explain why complete New Testament manuscripts represent only a very small proportion of the surviving evidence.

We are now closer to answering the question about just how rare these manuscripts are. Nevertheless, a definitive total depends on continued indexing and detailed study of the remaining eapr manuscripts. Until that work is complete, an exact figure will remain elusive.

Even so, the overall conclusion is clear: fully extant Greek New Testament manuscripts are far fewer than has often been assumed. The task then, is not only to refine their number, but also to study them more closely—as rare and indispensable witnesses to the textual history of the New Testament.

---

1. The majuscules were: GA 01,02,04; the minuscules: GA 18, 35,61,69, 141, 149, 175, 180 ,201, 205, 205abs, 209, 218, 241, 242, 296, 339, 367, 386, 498, 506, 517, 522, 582, 664, 680, 699, 757, 808, 824, 886, 922, 935, 986, 1072, 1075, 1094, 1384, 1424, 1503, 1597, 1617, 1626, 1637, 1652, 1668, 1678, 1704, 1780, 1785, 2136, 2200, 2494, 2495, 2554. ↩︎
2. Kurt Aland and Barbara Aland, The Text of the New Testament: An Introduction to the Critical Editions and to the Theory and Practice of Modern Textual Criticism, 2nd ed., trans. E. F. Rhodes (Grand Rapids: Eerdmans, 1989), 78. (The first German edition of their book (1981), 91, gives the same figures.) They do not specify which manuscripts they included in this total, but it appears to correspond to the 58 already registered in the 1963 Liste plus two additions: GA 2201 and GA 2352. Subsequent microfilm analysis showed that GA 2201 included Revelation, while GA 2352—previously listed as containing only Revelation—also preserves the other New Testament sections; both were accordingly reclassified as eapr. See Daryl Schmidt, “The Greek New Testament as a Codex,” in The Canon Debate, ed. Lee Martin McDonald and James A. Sanders (Peabody, MA: Hendrickson Publishers, 2002), 469–484 for some of the history of additions and removals in the eapr category. ↩︎
3. Elliott added 3 entries not found in the 1963 Liste: GA 1040 and 1352 (where the Revelation portions have since been given their own GA numbers) and GA 1248 which the microfilm confirmed to have Revelation. Elliot also placed 886 and 1384 in brackets noting that they never contained some of the New Testament writings. J.K. Elliott, “The Distinctiveness of the Greek Manuscripts of the Book of Revelation,” JTS 48 (1997): 116–24, here pg. 124. ↩︎
4. Bruce M. Metzger, The Text of the New Testament: Its Transmission, Corruption, and Restoration, 4th ed. (Oxford: University Press, 2005), 51 n. 80. ↩︎
5. I recognize that “completeness” is, to some extent, a heuristic category when applied to Greek New Testament manuscripts and does not always correspond neatly to the historical conditions under which these artefacts were produced and transmitted. Nevertheless, for the purposes of this study, I define “extant completeness” as a witness that preserves the text of all 27 canonical books of the New Testament without substantial lacunae. Importantly, this designation does not imply the absence of textual variation, including the omission of individual verses or passages, whether due to scribal error or differences in the textual tradition. (The longer ending of Mark, the Pericope Adulterae as well as other later textual additions not generally considered original are typically excluded from consideration). However, where a scribal omission results in a substantial loss of text, defined here as the absence of more than ten verses, I have classified this as a noticeable lacuna. ↩︎
6. Using original completeness as a criterion for the Liste raises several methodological challenges: First, such judgments depend on reconstructing a manuscript’s initial design and form, which is not always possible. Second, it is unclear how to classify artefacts that include substantial non–New Testament material (e.g., Psalms, apocrypha), are transmitted across multiple volumes, or were produced by multiple hands. In some cases, individual components may have been copied separately and incorporated later; given the fragmentary state of many manuscripts, it is often difficult to determine whether they once formed part of a complete codex. Finally, this approach does not fully align with the Liste’s own cataloguing practices: In the printed Liste, a cross symbol (†) was used to indicate either that a manuscript was lacunose or that it has been supplemented by a later hand. Together with the initial inclusion of many composite codices, this suggests that items were not excluded from being identified as “complete” (eapr) even when their “completeness” depended on later supplementation. ↩︎
7. For example, GA 60, 180, 209, 1140, 1668, and 1857 have been removed from the eapr category and Revelation was given its own GA number. ↩︎
8. These are: GA 35, 61, 757, 1072, 1704, 2494, and 2554. See Ulrich Schmid, “Die Apokalypse, überliefert mit anderen neutestamentlichen Schriften—eapr-Handschriften,” in Studien zum Text der Apokalypse, ed. M. Sigismund, M. Karrer, and U. Schmid (Berlin: De Gruyter, 2015), 436. ↩︎
9. Eldon Jay Epp, “Are Early New Testament Manuscripts Truly Abundant?” pages 77–117 and 395–99 in Israel’s God and Rebecca’s Children: Christology and Community in Early Judaism and Christianity: Essays in Honor of Larry W. Hurtado and Alan F. Segal, ed. David B. Capes, April D. DeConick, Helen K. Bond, and Troy A. Miller (Waco, TX: Baylor University Press, 2007), here 90 n. 16. He excludes GA 218, 498, 1352, and 1384. Strangely, Epp only lists half of the eapr manuscripts that are missing whole books of the New Testament. ↩︎
10. Three of the manuscripts he lists have long been removed from the eapr category because Revelation was added later (GA 180, 209, 1668). Interestingly, ten years earlier, Epp gave a higher number: 59 complete manuscripts, in “Textual Criticism in the Exegesis of the New Testament, with an Excursus on Canon,” pages 45–97 in Handbook to Exegesis of the New Testament, ed. by Stanley E. Porter (Leiden: Brill, 1997 [rep. 2002]), 76. ↩︎
11. GA 296 was removed in 2021 after Hugh Houghton identified it as a copy of the 1534 Greek New Testament by Colines. This conclusion had already been reached about Revelation by Darius Müller, “Abschriften des Erasmischen Textes im Handschriftenmaterial der Johannesapokalypse”165–268 in Studien zum Text der Apokalypse, ed. Marcus Sigismund, Martin Karrer und Ulrich Schmid (Berlin: De Gruyter, 2015). Likewise, GA 2136 was removed last year after Peter Montoro confirmed it was also copied from a printed edition. ↩︎
12. These are: GA 1785, formerly located at Kosinitza Monastery in Drama, has been missing for over a century; GA 242, currently held at the State Historical Museum in Moscow (only indexing and a transcription of Revelation are offered in the NTVMR, no images). GA 241, formerly housed in Dresden and long presumed destroyed during World War II, has recently been rediscovered in the Russian State Archive of Ancient Acts in Moscow. At present, only images of the book of Revelation are available. Indexing seems complete according to Bondach, Ф. 1607 ДРЕЗДЕНСКИЕ РУКОПИСНЫЕ КНИГИ, Опись 1 27 ед. хр. X–XVIII вв. (F. 1607 Dresden Manuscripts, Inventory 1, 27 items, 10th–18th centuries) (Moscow, 2023), 14–16. ↩︎
13. It should be noted since the cyber-attack on the British Library, digital images of a few previously digitized manuscripts are unavailable for the time being. ↩︎
14. Manuscripts that lack one or more complete New Testament books include: GA 04 (lacks 2 Thessalonians and 2 John); GA 218 (lacks Titus and most of Philemon); GA 339 (lacks numerous books, including Matthew, several Pauline Epistles, the Catholic Epistles, and Revelation); GA 517 (lacks the Gospel of John); GA 886 (lacks the Catholic Epistles); GA 1384 (lacks 2–3 John); and GA 2201 (lacks 2 John). An eighth could be added to this number because in GA 1652 Revelation is essentially lost (only 1:1–3 survives). ↩︎
15. E.g., Matt 1:1–2:9 and 5:19–6:5 ↩︎
16. This was noted in Kavrus-Hoffmann, Catalogue of Greek Medieval and Renaissance Manuscripts in the Collections of the United States of America. Part IV, 1: The Morgan Library and Museum (2008), 71–85. ↩︎
17. In the future, it may be useful to include a feature in the NTVMR for tagging manuscripts that are completely extant (e.g., with the siglum κδ for Καινή Διαθήκη) to offer users better transparency. ↩︎
18. Cf. Brent Nongbri, “The Date of Codex Sinaiticus,” JTS 73.2 (2022): 516–34. ↩︎
19. Cf. Schmid, “Die Apokalypse,” 432–3. ↩︎
20. The number of fully extant eapr manuscripts may increase if missing folios are identified—as in the case of GA 141—or if dispersed portions of the same codex are located and combined, as with GA 905 and GA 2382. It may also change as additional images become available. For instance, it is likely that GA 1040 and GA 2041 originally constituted a single, complete New Testament codex, a possibility already noted in the 1963 Liste. To date, however, this has not been conclusively demonstrated.  ↩︎
21. For more on the history of the codex, see D.C. Parker, An Introduction to the New Testament Manuscripts and Their Texts (Cambridge: Cambridge University Press, 2008) and Harry Y. Gamble, Books and Readers in the Early Church: A History of Early Christian Texts (New Haven and London: Yale University Press, 1995). ↩︎