The Catena Manuscripts on Acts: A Revised Classification

Earlier scholarship faced a number of limitations in classifying catena manuscripts on the Acts of the Apostles. This study makes a comparison of exegetical scholia in selected text passages (Acts 2:1–16, 8:9–25, 28:19–31) in order to determine the different types of catena and how they relate to each other. This survey reveals the diversity of the tradition: some manuscripts are merely copies, which repeat the same text with only small variations, but others are unique and cannot be directly identified with a particular catena type. It is therefore necessary to expand the classification of catenae on Acts in the Clavis Patrum Graecorum so as to mark subdivisions within the individual types.


Introduction: An Overview of the Previous Research
Recent years have seen a significant increase in the study of patristic exegesis of the New Testament, especially with regard to biblical catenae.1 Detailed research on the catenae on the Acts of the Apostles, however, is yet to be Scieri 10.1163/15700720-bja10042 | Vigiliae Christianae (2021) 1-25 undertaken. At present, there is no critical edition and it is necessary to rely on the printed editions of a few individual Acts catena manuscripts.2 In 1532, Donatus of Verona published a selection of material on Acts, the Pauline epistles, the Catholic epistles, and the book of Revelation, apparently based on Paris, BnF, Gr. 219 (GA 91).3 A Latin translation was then published by John Henten in 1547. A subsequent printed edition containing both the Greek text by Donatus and the Latin translation by Henten was produced in 1631 by Frédéric Morel.4 This edition eventually found its way into Migne's Patrologia Graeca, where the compilation was attributed to Oecumenius (sixth century).5 Migne also reprinted the edition by Finettus (1755) of three different catenae on Acts attributed to Theophylact, an eleventh-century Bishop of Bulgaria.6 The first catena, from Vienna, ÖNB, Theol. Gr. 150 (GA 1524), had been previously published by Sifanius (1557). The other two texts were based on Vatican, BAV, Vat. Gr. 652 (GA 1842) and Florence, BML, Plutei IV.5 (GA 455), respectively. In 1838, John Cramer published a catena on the Acts of the Apostles as the third of his eight volumes of New Testament catenae.7 This was based 2 A summary of the editorial history is given in R. Devreesse, "Chaînes exégétiques grecques," Dictionnaire de la Bible, Supplément I (Paris: Letouzey & Ané, 1928)  In addition, the incipit and explicit of the catena edited by Cramer are given along with the first and last words of twenty-one scholia to Acts 8:9-25; finally, a list of the Church Fathers named as sources for the extracts is reported.12 Some information pertaining to Acts is found in Karl Staab's article about the catenae on the Catholic Epistles (1924), soon after expanded with minor additions by James Hardy Ropes (1926).13 Again, some of the manuscripts in Staab's major work on the catenae on the Pauline Epistles (1926) also contain catenae on Acts, although he pays little attention to these.14 The first to attempt a thorough analysis and a classification of catena manuscripts on Acts was Hermann von Soden. In the first volume of his edition of

Classification
One common method of identifying and describing a work in a manuscript is to consider its incipit and explicit. On this basis, Parpulov assigned sixty-one Acts witnesses to one of the CPG types.19 Due to the complexity of catena tradition, however, Karo and Lietzmann sampled witnesses in a selected passage in order to determine their affiliation. This has been adopted in the present study.
In addition to the passage used by Karo and Lietzmann, Acts 8:9-25, two additional passages have been selected so as to include fragments and incomplete copies: these are Acts 2:1-16 and 28:19-31. Table 1 provides a list of the manuscripts considered in this survey with details of the revised classification offered by this study (and described further below).20 The three types of the CPG have been retained, with further divisions and subdivisions marked with numerals and letters (e.g. C150.1a, C150.2a). The division into subtypes has been applied even where these consist of only one manuscript: while this may simply be a result of editorial revision of that one witness, it may also represent a broader tradition which has not survived. It should be remembered that this study considers all the evidence known at present. In addition, a new category, C155, has been introduced for the codices singuli. The Gregory-Aland number, although technically descriptive of the biblical text, has also been included here and in the discussion below for ease of reference. Thirteen of these preserve brief glosses or isolated scholia rather than a full catena: GA 203, 302, 457, 462, 617, 627, 1162GA 203, 302, 457, 462, 617, 627, , 1277GA 203, 302, 457, 462, 617, 627, , 1764GA 203, 302, 457, 462, 617, 627, , 1780GA 203, 302, 457, 462, 617, 627, , 1845GA 203, 302, 457, 462, 617, 627, , 1859GA 203, 302, 457, 462, 617, 627, and 1980 The glosses mentioned in the previous note have not been included, and the microfilms of GA 101 and 2733 are illegible. On the other hand, GA 1371 has been added as a representative of C151.1a, although the manuscript is fragmentary (see section 4 below).

C150.2.
To the first type belong six witnesses to what appears to be the principal and perhaps the oldest catena on Acts, as it is the primary source of most later compilations. Their texts have the same beginning (Ἀντιοχεὺς ὑπάρχων τὸ γένος ὁ Λουκᾶς, ἰατρὸς τὴν ἐπιστήμην, …) and ending (Μαρτυρολόγιον Παῦλου τοῦ Ἀποστόλου˙ … μηνὶ Ἰουνίῳ κθ´ ἡμέρᾳ) and share approximately the same number of scholia (50 for Acts 2:1-6, 21 for Acts 8:9-25, 25 for Acts 28:19-31).21 However, GA 1895 features some additional comments and lacks others.22 Although further research is required to determine whether this preserves an early stage of the Andreas catena (as suggested by the early date of the codex: first half of the 10th century) or a later expansion, the additional material in each group, shown in Table 2, is used to identify two subtypes: C150.1a, the main tradition (GA 307, 2818(GA 307, , 610, 453, 1678, and C150.1b, the version of GA 1895. Most of the scholia in the Andreas catena are given a name or a title: in the three passages examined, more than eighty comments are attributed to Chrysostom, Epiphanius, Ammonius, Severus of Antioch, Theophilus, Didymus, Athanasius, Severian of Gabala, Isidore, Cyril, Gregory of Nazianzus, Irenaeus, Eusebius; two have the heading (ἐξ) ἀνεπιγράφου and three are simply marked as σχόλιον.23 It is possible that the original basis of the compilation was mostly or even completely made of excerpts from Chrysostom's Homilies On the Acts of the Apostles and On the Beginning of Acts (which are also the only commentaries on the subject transmitted in their entirety), and subsequently expanded with extracts from other exegetes.24 Support for this may be offered by the title of the Andreas catena, which in all witnesses reads: Ἑρμηνεία τῶν Πράξεων τοῦ Χρυσοστόμου καὶ ἑτέρων διαφόρων ('Exegesis of Acts by Chrysostom and other different [commentators]'). On the other hand, Chrysostom might have been emphasised in the title because he is the most frequently cited author throughout the catena. 21 The partial mutilation of GA 2818 and 610 does not allow us to ascertain whether these witnesses had the same number of scholia to Acts 2:3-4 and 8:9-24, respectively. 22 Here, as well as in the tables below, the source indication has been provided in brackets where it was detectable. This has not been applied where the scholia have resulted from multiple sources, which the compiler has freely combined, paraphrased and readapted (see Table 5,  The name of Andreas is placed as a subscriptio at the end of GA 307, along with an extract from the biography of Andreas the Presbyter found in three witnesses to the catena on Isaiah. There, this figure is credited with three books of the catena on Isaiah, but there is no mention of a catena on Acts.25 Staab and Ropes have reasonably cast doubt on this attribution, absent from the other five catena manuscripts on Acts.26 Karo and Lietzmann identified seven manuscripts as abridged forms of the Andreas catena, which we designate as C150.2.27 To these should be added GA 455 and 2576, which both Finettus and von Soden observed had been erroneously attributed to Theophylact.28 The degree of similarity is so high that von Soden dubbed the text from GA 455 as "die Andreas-Catene ohne Text".29 In the abbreviated versions of the Andreas catena, the compilers chose a small number of scholia from the fuller catena type. The name of each author is retained before some of the scholia, but in most cases it is omitted. The selected comments are either copied in full or slightly abridged; expansions are rare. Occasionally, scholia from different sources are joined together in a single extract as if they were written by the same author; conversely, a singleauthor scholium might be split into two comments, with the second sometimes marked as coming from another source (ἄλλoς).
GA 94 features a different selection of scholia and has thus been identified as an individual subtype of abbreviated catena, C150.2a. Six witnesses (GA 606, 641, 103, 607, 2576 and 455) have the same abridged version of the Andreas catena, and shall therefore be marked as C150.2b. Nevertheless, a few additions shared by three manuscripts allow us to split the cluster in two groups, the second of which seems to be an expansion of the first. The evidence is shown in Table 3. Another single instance of an abbreviated catena is GA 1424 (C150.2c), which is absent from Karo-Lietzmann's catalogue and not mentioned by von Soden. A few scholia were copied in the margins by the same hand responsible for the biblical text (9th-10th cent.), whereas the majority of comments was added by a later scribe (12th cent.). Unlike the prologue (Ἀντιοχεὺς ὑπάρχων τὸ γένος ὁ Λουκᾶς, ἰατρὸς τὴν ἐπιστήμην, …), the incipit and explicit of the scholia differ from the mainstream tradition, simply because these have been abridged. Most comments have been created by reducing the exegetical material in C150.1, and the names of authors are retained. Nonetheless, the diversity of content in certain scholia suggests that the compiler may have employed additional sources. A similar situation can be observed in the catena from GA 82 (C150.2d), which also features two layers of scholia. However, unlike in GA 1424, the comments are anonymous and indicated by symbols. A curious abridgment from Andreas is GA 608 (C150.2e), which seems to contain a mixed catena, partly related to C150.2b but also apparently dependent on C151.3, one of the Oecumenius subtypes.30 The connection with the other abbreviated subtype of Andreas was detected by Staab, who observed that the commentary from this manuscript agrees with that in PG 125 derived from GA 455 (Expositionis in Acta textus tertius).31 There are notes by three librarians in the front matter that mention Andreas as the author of the catena (in Italian, Greek and Latin). Moreover, the incipit of the first scholium from GA 608 (Ὁ μακάριος Λουκᾶς Ἀντιοχεὺς ὑπάρχων τὸ γένος, ἰατρὸς …) and the beginning of the prologue from C150.2b (Ἀντιοχεὺς ὑπάρχων τὸ γένος ὁ Λουκᾶς, ἰατρὸς …) are almost identical. Nevertheless, the Latin attribution explicitly suggests a possible influence from the Oecumenius catena (C151.3): « Cod.
Finally, another abbreviated subtype is preserved in the second catena from GA 886 (C150.2f). This manuscript contains two fragmentary catenae on Acts, copied by different hands (13th and 14th cent., respectively). As noted by Karo-Lietzmann and Staab,32 the second compilation (on Acts 2:14-7:59) is an abridgment of the Andreas catena with a particular preference for comments from Chrysostom: all the scholia in this catena are found in C150.1, which contains others not present in this manuscript. Nevertheless, Devreesse believed that this shorter text might instead reflect an early stage of the Andreas catena, when this was mostly made of extracts from Chrysostom's Homilies.33 30 See Section 4 below. 31 The erroneous attribution of this catena to Theophylact has already been noted above. However, Staab claimed that the title, despite the faded red ink, disclosed τω βουλ … Θεοφυλακτω (Staab, Die Pauluskatenen, 223-224). Nevertheless, von Soden states that "Der erste Abschnitt ist unleserlich" (Die Schriften, 1:688) and nothing can be made out on our microfilm. 32 Cf. Karo and Lietzmann, Catenarum Graecarum Catalogus, 595; Staab, Die Pauluskatenen, 219-220. 33 Cf. Devreesse, "Chaînes exégétiques grecques," 1205-1206.  (GA 91, 1371(GA 91, , 1933 and C151.1b (GA 916). GA 1371 is a fragmentary copy of GA 91 in Acts 1:1-16 (but not in the other New Testament extracts), as indicated by the gaps left where this codex is damaged. On the other hand, two of the three representatives of C151.2 (GA 056, 0142, 1066) are "sister manuscripts": GA 056 and 0142 share the same the layout, marginal notes, number of pages, script and even paratextual features.39 Both subtypes exhibit a similar incipit (C151.1: Λόγον πρῶτον εἶπεν καὶ οὐκ εὐαγγέλιον, C151.2: Πρῶτον λόγον εἶπεν καὶ οὐκ εὐαγγέλιον) and share nearly all the comments on Acts 2:1-16 and 28:19-31. However, the scholia to Acts 8:9-25 and the conclusion are notably different, as shown in Table 5.
Oecumenius' compilation is generally regarded as a reworking and abbreviation of the Andreas catena, with the greatest difference being the absence of author's names.46 A preliminary comparison with the Andreas catena, however, reveals a fair degree of independence from the mainstream tradition and a different compilation practice. In commenting on Acts 8:9-15 the compiler of C151.1 draws on different sources, not necessarily related to the Andreas catena. He usually selects the extracts most appropriate for the exegesis and arranges these in long comments. At the same time, he paraphrases sentences, introduces linking words or phrases of variable length and adds something of his own, following a technique of compilation known as résumé.47 This consists of rewriting the sources ex novo, retaining only a few words and sentences in their original form. The result is that we have scholia unique in appearance, whose original sources are not easily detectable. Nevertheless, in other sections (such as Acts 28:19-31), the compiler employs the same technique used by "Andreas", consisting of a more faithful reproduction of the original patristic extracts, either in full or abbreviated. Gilles Dorival has suggested that compilations of this kind, based on catenae but lacking source indications, are better labelled as commentaries.48 Nevertheless, given their origins in catena 43 The incipit is different: Ἀναμιμνήσκει τὸν Θεόφιλον τοῦ εὐαγγελίου ὥστε τὴν οἰκείαν ἀκριβείαν ἐνδείξασθαι κτλ. 44 These two manuscripts are likely to be genetically related: not only the layout, but also the content is almost identical, line by line. 45 In GA 621 the order of some scholia is reversed. This is likely to be a development from scribal errors rather than deliberate editing. 46 Cramer, Catenae, 3: iii: "Oecumenii enim opus nihil aliud est nisi Catenae nostrae epitome, in qua omnium auctorum nomina tacentur, et eorum excerpta in unum quasi corpus rediguntur"; and cf. Devreesse, "Chaînes exégétiques grecques," 1206. 47 Cf. C. Curti, "La tradizione catenaria e il recupero dei commenti greci alla Bibbia: validità e limiti", in C.  tradition, no proposal has yet been made to change the traditional classification. One could rather suggest that these works are representatives of an alternative type of catena.

C152: The "Theophylact" Catena
Five manuscripts transmit a catena on Acts under the name of Theophylact (GA 254, 455, 1524(GA 254, 455, , 1842(GA 254, 455, , 2576 and Finettus edited three recensions based on as many exemplars.49 However, as seen above, three of these seem to be witnesses to different catena types. Therefore, only the texts from 254 and 1524 are potentially related to the Archbishop of Bulgaria, and yet scholars have failed to reach consensus on the attribution. In Finettus' opinion, the catena from GA 1524 lacks the ratio commentandi characteristic of Theophylact as it emerges in the commentaries on the Gospels and Pauline Epistles.50 His doubts were shared by von Soden, who suggested that the author indication before the prologue might be a subsequent addition by a later scribe trying to make up for the omission of a commentary on Acts among Theophylact's exegetical works.51 The authorship therefore remains to be ascertained. The text of GA 1524 contains a significant proportion of similarities to the first of the two catenae on Acts preserved in GA 437 (ff. 1r-180r), which in turn seems to be partially connected to GA 1871, at least with regard to the extant passages from this incomplete witness (only Acts 25:9-28:31). Table 7 makes it apparent that GA 1871 and 437 are closely related, as they both split scholium 1 into three segments and lack scholia 2 and 10. On the other hand, GA 1871 includes additional comments (5, 7) and shares scholium 6 with GA 1524 but not with GA 437. This latter concludes the exegesis with a scholium (11) absent from GA 1871 and 1524. Moreover, all three have different endings: … ἐν ᾗ κατέμεν ἐν τῇ Ῥώμῃ (GA 1524); … μέχρι τῆς τελειώσεως, λε´ (GA 437, first catena); … ἐπιστρέψωσιν, ἰάσεται αυτούς (GA 1871). Although the genealogical relationship between the catenae of these codices is yet to be traced in detail, the data suggests that the three manuscripts might preserve different stages of the same catena type (C152.1). For these reasons, we shall mark them as C152.1a (GA 437, catena one), C152.1b (GA 1871), C152.1c (GA 1524  a In the Andreas catena (C150.1) this scholium is entitled τοῦ αὐτοῦ and comes after a comment from Chrysostom. However, since there is no trace of this text in Chrysostom's works, the attribution to Severian of Gabala is probably correct. the use of a common type as a source. Another departure from GA 1524 is the indication of authorship in the margins of GA 254, next to the first words of each comment, which is mostly absent from the representatives of C152.1. The most relevant discrepancy, though, seems to be in the compilation practice: while the compiler of GA 254 generally separates individual authors' extracts, the catenist of GA 1524 tends to combine multiple sources in a single scholium. Unlike the Oecumenius catena (C151), however, he does not 'rewrite' the mixed comments: in most cases, these are merely abbreviated and assembled while preserving the author's style. This could be defined as a "cut and paste" technique.54

C155: Codices singuli
This category has been introduced as the next available number in the Acts catena sequence in CPG for manuscripts whose catena does not correspond to any of the types so far illustrated. Three manuscripts have already been assigned a number in the Clavis Clavium based on Parpulov's catalogue: GA 605 (C155.1); GA 920 (C155.2); Patmos, Ioannou, 263 (C155.3).55 The next manuscripts to be included are: GA 437, catena two (C155.4); GA 886, catena one (C155.5); GA 1839 (C.155.6). Further detailed research is required to determine the nature of these compilations, but there is already sufficient evidence to exclude any close relationship between the individual witnesses: each presents a different selection and arrangement of scholia, as displayed in Table 9.

Conclusion
To some extent, the present survey may appear to have confirmed Charles Kannengiesser's opinion that the study of the genealogy of catenae is a "a bewildering task".56 Nevertheless, this attempt to classify the manuscript tradition in greater detail than has hitherto been the case has brought a degree of order to the confusion which may arise both when approaching individual manuscripts and in understanding the existing printed texts. The forty-six catena manuscripts on Acts considered in this study disclose great diversity: not only are there discrepancies between the selection and transmission of 54 Cf. Curti, "La tradizione catenaria", 280. 55 https://clavis.brepols.net/clacla/OA/Link.aspx?clavis=CPG&number=C155. 56 Kannengiesser, Handbook of Patristic Exegesis, 978.
scholia, but even within groups it is rare that two copies are exactly the same.
Determining the boundaries between one type of catena and another is a challenge, because during the process of compilation, existing collections were reduced, extended or mixed with others, quotations by Christian exegetes were expanded or abbreviated, and often simply removed. It is apparent that the Andreas catena (C150) exerted a consistent influence on all the other types, yet there is sufficient evidence to adduce a certain degree of flexibility in following the main model and the utilisation of multiple sources (this is the case in C150.2e, C151, C152.2, and most likely some codices singuli). As observed by William Lamb, "the rather chaotic manuscript tradition characteristic of catenae suggests that a catena is an 'open book' . Material was added and amended with the production of each new copy."57 In the light of these considerations, the deployment of specific categories of identification (type, subtype, group), albeit flexible, will permit scholarship to gain a better idea of the diversified tradition of catenae on Acts. This account provides a point of reference for future research and lays the groundwork for specific exploration of matters yet to be addressed, such as reconstructing the stages in the development of the recognised catena types and identifying the manuscripts which merit close examination (such as the codices singuli). Only by proceeding in this carefully defined way will it become possible to reach a fuller appreciation of the nature and value of the catena tradition on the Acts of the Apostles.