In 1904 a fragmentary papyrus was purchased in Luxor, Egypt by Charles Bousfield Huleatt, who presented it to Magdalen College Library, Oxford University, where it was catalogued as P. Magdalen Greek 17 and subsequently registered as Greg.-Aland P64. The "Magdalen" papyrus contains portions of Matt 26 and remained unedited until 1953 when the ed. pr. was published by Colin H. Roberts, “An Early Papyrus of the First Gospel,” HTR 46 (1953): 233–37.
At first Huleatt had assigned the fragments to the third century, whereas A. S. Hunt assigned them to the fourth century. However, Roberts questioned this late dating and on the basis of palaeography assigned P64 to ca. 200. Apparently, he had obtained the agreement on this dating by other renowned experts (Bell, Skeat, and Turner). One possible reason for the late dating by previous scholars was that it had not yet been established that the codex format itself was so early.
In 1956 more fragments from the same codex turned up and were published by Ramón Roca-Puig, Un Papiro Griego del Evangelio de San Mateo (Barcelona: Grafos S.A., 1962). These fragments containing portions of Matt 3 and 5 were cataloged as P. Barc. Inv. 1 and registered as P67 in the list of NT papyri. Roca-Puig and Roberts could determine that they came from the same codex as P64, and this judgment has remained the scholarly consensus.
In his Manuscript, Society and Belief in Early Christian Egypt (London: Oxford University Press, 1979), C. H. Roberts claimed that there could be no doubt that P4 (in Paris) containing portions of Luke also came from the same codex. It had been dated by J. Merell to the fourth century in 1938, by Kurt Aland to the third century in 1963, but now Roberts assigned it to the same late second-century date as P64+67. In a 1997 article T. C. Skeat took up Roberts' remark and similarly argued that P4 comes from the same four-gospel codex as P64+67, which would make it the oldest known four-gospel codex.
Subsequently, co-blogger Peter Head and Scott Charlesworth have argued against this identification, mainly on the basis of codicological data – it is still possible that they were copied by the same scribe. See P. M. Head, “Is P4, P64 and P67 the Oldest Manuscript of the Four Gospels? A Response to T. C. Skeat,” NTS 51 (2005): 450–57; Scott Charlesworth, ”T. C. Skeat, P64+67 and P4, and the Problem of Fibre Orientation in Codicological Reconstruction,” NTS 53 (2007): 582–604.
However, no one has to my knowledge made a comparative textual analysis of these papyri. In his 1994 publication, Skeat included a brief analysis of the text of P4, providing ”some basic facts.” Unfortunately his analysis is unsatisfactory in two ways: it concerns only P4 and it is based only on deviances from the Textus Receptus.
In a forthcoming paper which has been accepted for the SBL Annual Meeting in Atlanta this year I will attempt a textual analysis of P4 and P64+67 in order to examine the textual quality, transmission character, and the nature of the readings in these papyri. The main problem with a comparative textual analysis as I see it is the brevity of P64+67, but, hopefully, something will come out of this study to supplement codicological and palaeograhical data that will help us to better evaluate these papyri.
NB: I have not reported here on Carsten Peter Thiede's claim in several publications that the Magdalen Papyrus is from the mid 1st century, since it has received no support among scholars, but, on the contrary, has been falsified by several scholars (see e.g., Peter Head's article here)