For most of the narrative, the extant MSS of the Roman d'Alexandre (RdA) do not present substantially different versions of the text. There is no definition of a 'long version' or a 'short version', as there is for the Prose Lancelot, nor does the core text in the cyclic redaction, with interpolated narratives, differ substantially from that in the pre-cyclic one. The exception is Branch II, the Fuerre de Gadres, in which two radically different accounts divide the MSS between two versions: version alpha or 'common version', and version beta or 'Gadifer version'. In the three other branches, the Princeton team of editors recognized at least two groups of MSS, some of which are relatively stable. Other MSS switch between the two groups or betray the influence of copies from other MS sets. Especially with relation to Branch IV, the description of Alexander's death and funeral, the extant copies appear to 'resist clear-cut schematization into groups' (Armstrong 1937, p. xx).
The introductions and notes to the various volumes of the Princeton edition demonstrate, moreover, that the texts in the surviving copies do not only vary on the micro-level, but that there are considerable differences in the number of stanzas and the order in which these are presented. In some copies, entire laisses appear to have been omitted, while the text in other manuscripts expands or reorders information. In our descriptions (see: Segmentation), the stanzas which are particular to a (limited set of) MS(S) have been marked as 'interstitial'. This allows for a comparison on the basis of the presence or absence of these specific stanzas. Together with an inventory of missing or omitted stanzas, and the reordering of laisses, this provides a first rough idea of the textual relations between MSS. For each branch of the text, the Princeton edition gives a useful overview of the stanzas found in the copies of that branch. This is also our point of departure for the description of the text in each MS.
A considerable number of MSS include added narratives, surfacing in varying positions and combinations. These interpolations include two, probably contemporaneous, but nonetheless independent, vengeance 'sequels'. The Venjance Alixandre by Jean le Nevelon (Arras or Reims, c. 1181?) has Alexander's son, Alior, avenge his father. In the Vengement Alixandre by Gui de Cambrai (Arras or Arrouaise, before 1191), the capture and punishment of the assassins are initiated by Alexander's brother (Armstrong 1926; Ham 1931; Ham 1935). Interestingly, the MSS with the Vengement are all dated to the late 13th c., and their provenance is mainly from the Picard region. By contrast, the Venjance is only found in later MSS, the earliest of which may be Paris BnF f.fr. 24365, dated to c. 1315-25. In general, the MSS with the Venjance also have one or more texts from the Peacock cycle integrated into the text of the RdA or copied after it (see below). The exception is Parma Bibl. Palatina 1206, a 14th-c. Italian MS which has a unique version of Alexander's vengeance and combines portions of both texts (Edwards 1926, 4-12; Ham 1935, 13-6).
Circa 1250-60, two more episodes were added to the verse cycle and interpolated at the end of Branch III. Linguistic traits suggest that these poems were also composed in Picardy (Peckman 1935, xxvii, lxi). These new interpolations always appear in MSS also comprising a vengeance sequel, whether the Vengement or the Venjance. The Prise de Defur describes how, after his departure from the Queen of the Amazons, Alexander sets out for Babylon, lays siege to the city of Defur, and meets Queen Candace, the future mother of Alior. A second interpolation, the Voyage au Paradis terrestre, brings Alexander to the gates of the Earthly Paradise, after which he is confronted with his approaching demise (Peckman 1935, xxxviii). Intertextual references, and the fact that in extant MSS the Voyage is never found without the Prise (which, by contrast, does appear without the Voyage), suggest that the second addition was conceived as a sequel to the first interpolation. The renewed interest in the RdA in Northern France, as demonstrated by the insertion of these episodes, is contemporaneous with the flourishing production of Alexandre manuscripts in the region.
The final stage in the development of the Alexandre cycle is formed by the three so-called Peacock texts. The Voeux du paon, with a description of the siege of Epheson and the eponymous ceremony of vows, was written c. 1312 by Jacques de Longuyon and dedicated to Thibaut de Bar, bishop of Liège. The exceptional success of this text is underlined by the number of extant MSS, of which over 40 copies survive. Initially conceived as an amplification of the Prise de Defur, the Voeux were copied separately as well as in MSS of the RdA. In the latter case, the text is found inserted in two different places: either at the end of the text, after Branch IV (as in the Parisian copies Paris BnF f.fr. 791, Paris BnF f.fr. 1375, Paris BnF f.fr. 1590, and BnF f.fr. 24365) or integrated into the text of the Prise, albeit at slightly different points depending on the MS. A first continuation of the Voeux, the Restor du paon, written c. 1338 by Jean Brisebarre, survives in only one extant MS that also has the RdA - the famous Oxford Bodleian 264. Additionally, it appears in numerous MSS of the Voeux du paon. One of these, Paris BnF f.fr. 12565, includes the Prise de Defur, but compared to the Bodley MS, the Voeux have been integrated in a different position. BnF f.fr. 12565 also has the final addition to the Peacock cycle, the Parfait du Paon, written in 1340 by the Hennuyer Jean de le Mote for the Parisian goldsmith Simon de Lille. The Parfait has survived in only one other, incomplete MS: Oxford Bodleian Douce 165. The complete sequence Voeux-Restor-Parfait appears in none of the known RdA MSS. The same is true of an alternative 14th-c. offspring of the Voeux, the monumental Perceforest, which transfers the Macedonians to the British Isles, thus connecting Alexander to Arthur and the matière de Bretagne.