Eusebius’s tables show that single Matthean episodes are often divided into multiple numbered sections, spread across different canons, whereas Mark’s corresponding story remains one unified section. For example, the Last Supper in Matthew 26:26–29 is split into two Ammonian sections: Matthew section 284 (covering 26:26, the bread) is placed in Canon I (parallel to Mark 14:22 and Luke 22:19, plus John’s “bread of life” discourse), while the very next verses (Matthew 26:27–29) form section 285 in Canon II (aligned only with Mark 14:23–25 and Luke 22:17–18). Thus one continuous Matthean pericope is fragmented between two canon tables. Mark’s account of the same supper, however, remains in two natural sections (Mark 14:22 in section 165, and 14:23–25 in 166) that follow each other in sequence and stay within their respective canons. Matthew’s narrative had to be “expandable”, split to accommodate John’s unique material and then rejoined with the Synoptic parallels, whereas Mark’s shorter account was kept intact in each parallel.
Another clear example is Jesus walking on water. Matthew’s single story (Matthew 14:22–33) was divided: Matthew section 150 covers Jesus walking on the sea (Mt 14:23b–27) alongside Mark 6:47–50 and John 6:15b–21 in Canon IV, but Matthew’s additional scene (Peter briefly walking and the aftermath, Mt 14:28–33) has no Mark/Luke parallel and thus appears separately as Matthew section 151 in Canon X (Matthew-only). In contrast, Mark’s version of the event (Mark 6:47–52) was handled as a single unit (Mark section 67) covering the whole incident in Canon IV. Eusebius did not carve up Mark’s pericope further; Mark’s section retains its integrity (aside from a small epilogue in Mark 6:51–52 marked as a unique section) and remains contiguous in the tables. It is Matthew’s text that gets bisected, showing Eusebius’s willingness to break Matthew’s narrative continuity in order to line up parallel incidents, whereas Mark’s sequence stays whole.
Even in the Temptation of Jesus, Matthew’s account is partitioned. Matthew 4:1–11 is split so that the opening verse (the Spirit leading Jesus into the wilderness) is section 15 in Canon II alongside Mark 1:12–13 and Luke 4:1–2a, but the rest of the temptations (Mt 4:2b–10) become section 16 in Canon V with Luke 4:2b–13. Mark’s brief version (only 2 verses) remains a single section (6) used once in Canon II. Here again Eusebius (following Ammonius) subdivides Matthew’s pericope to match two different parallel groupings, while Mark’s corresponding section is indivisible and used as-is. These examples make it clear that Matthew’s stories were treated as modular pieces that could be split across canons, whereas Mark’s stories were treated as fixed units in the harmonization.
Matthew’s Sections Rearranged vs. Mark’s Fixed Sequence
Because Eusebius arranged the parallels largely following Matthew’s order, Matthew’s section numbers often appear out of sequence alongside Mark/Luke, highlighting that Matthew’s content was repositioned and sliced to line up with others. For instance, Matthew 5:13 (“salt of the earth”) occurs early in Matthew’s Sermon on the Mount, but its parallels are in Mark 9:50 and Luke 14:34, much later contexts. In the canon tables Matthew section 31 (Mt 5:13) is paired with Mark section 102 (Mk 9:50) and Luke section 185 (Lk 14:34). Immediately afterward Matthew section 32 (Mt 5:14–16, “light of the world”) aligns with Mark 4:21 and Luke 8:16/11:33, causing Mark’s reference to jump back to section 39. This means Mark’s sections appear out of their natural numerical order in service of Matthew’s arrangement – Mark 1:12 (section 6) was just cited for Jesus’ baptism, then Mark 9:50 (section 102) is cited for Mt 5:13, then back to Mark 4:21 (section 39) for Mt 5:14–16. Matthew’s discourse has been sliced into pieces and matched with Mark/Luke material from disparate chapters, whereas Mark’s pericopes themselves stay intact (each cited by a single section number) even though they are listed out of order. The canon tables also show Matthew’s sections repeated or grouped in non-linear ways: e.g. Matthew section 279 (the Last Supper betrayal prediction) is listed twice in Canon IV next to Mark 161, once with John 72 and again with John 121, aligning Matthew with two different Johannine passages. Mark’s section 161 covers the entire “one of you will betray me” scene and remains a single unit (reused in both rows but not split further). Matthew’s narrative, by contrast, is flexible enough that Eusebius can pair its pieces with multiple Gospel counterparts out of sequence. This unequal treatment underscores that Matthew was the “base” text to be diced and distributed for concordance, whereas Mark’s content was more static.
Mark’s Pericope Integrity in the Canon Tables
In Eusebius’s system, Mark’s sections preserve their internal unity and sequential integrity to a far greater degree. Mark’s Gospel was divided into 235 sections, and these sections generally correspond to whole pericopes or coherent story units. When parallels are listed, Mark’s section numbers appear once per parallel set, encompassing the full Markan episode, even if Matthew or Luke break the material up. For example, the double miracle of Jairus’s daughter and the woman with the hemorrhage is a single continuous story in Mark. Eusebius assigns it one number – Mark section 49 covers Mark 5:21–43 in full – and in Canon II this single Markan section lines up against Matthew section 74 (Mt 9:18–26) and Luke 85 (Lk 8:40–56). Mark’s entire pericope stays intact (note: five distinct incidents in Mark 5:21–43 are not split into smaller sections), whereas Matthew’s version is slightly shorter and Luke’s more detailed, yet each is also one section. We see that Mark’s numbering is never used to “split” a story across multiple rows – if Mark recounts two events back-to-back without a clear break, Eusebius often keeps them together in one section. For instance, Mark 4:35–5:17 (Jesus calming the storm and healing the Gerasene demoniac) is treated as a single section (Mark 47) in the tables, precisely because Matthew’s parallel (Mt 8:23–34, section 69) narrates those scenes as one sequence. Rather than split Mark’s account into two parts, Eusebius lets Mark 47 span both episodes, preserving Mark’s narrative flow. This stands in contrast to Matthew, whom Eusebius did slice or merge as needed. The resulting pattern is that Mark’s section numbers always increment consistently through Mark’s Gospel, and each Markan section remains a self-contained, indivisible unit in the harmony. Even when Matthew’s ordering forces Mark’s sections to be cited out of order, those Markan units themselves aren’t broken apart. In summary, the Eusebian canons treat Mark’s pericopes as fixed points of reference, while Matthew’s pericopes were treated as malleable – expandable, compressible, and re-arrangeable to line up all the parallels. Eusebius (following Ammonius’s method) effectively uses Matthew as the framework into which material from Mark, Luke, and John is fitted, demonstrating by the canon sections that Matthew’s boundaries could be breached for harmonization, whereas Mark’s boundaries largely remained intact.
Implications: This structural asymmetry in the canon tables implies that Matthew was not regarded as inviolably sequential – editors felt free to segment Matthew’s text to compare teachings and events – but Mark’s narrative was taken as stable “building blocks.” Early harmonists like Ammonius and Eusebius thus accorded Mark (and to an extent Luke/John) a kind of textual integrity in the parallels, while treating Matthew as the expandable master-document. This supports the view that Matthew was used as the base into which others were conformed. The fragmentation of Matthew in the tables underscores how gospel harmonization worked: Matthew’s authority lay in its comprehensive content, which could be re-arranged to map all other gospels onto it, whereas Mark’s authority was in its precise pericopal form, cited whole. The Eusebian canons make this unambiguous – Matthew’s sections are freely shuffled and split across canons, but Mark’s section numbers run in order and remain whole, reflecting an early recognition that Matthew could absorb additions or be dissected without losing credibility, whereas Mark’s text was handled as a fixed sequence. Such treatment reveals the assumptions of ancient scholars: Matthew’s Gospel was the flexible reference scroll, whereas Mark’s Gospel was a stable secondary witness, a pattern that speaks to how each Gospel’s textual authority and priority were perceived in late antiquity.
Sources: The above analysis is based on Eusebius’s canon tables as compiled from the Greek gospel sections and the illustrative examples given in modern summaries of the canons, which clearly show Matthew’s section numbers repeating or appearing in multiple canons, in contrast to Mark’s one-to-one section correspondence that preserves each pericope’s unity. The Eusebian apparatus itself (as presented by the Tertullian project and noted by Stephan Huller) explicitly lacks any canon for “Mark + John” or “Mark + Luke + John” because Matthew was the base – all multi-gospel parallels were anchored to Matthew’s sequence. This confirms that Ammonius/Eusebius deliberately structured the harmony on Matthew, resulting in the frequent Matthew fragmentation demonstrated above, while leaving Mark’s order unperturbed except where it overlaps with Matthew. The canon tables thus provide concrete, cited proof that Matthew’s pericopes were not treated as indivisible, whereas Mark’s pericopes retained their integrity in the Eusebian numbering scheme.
