Area G and the Digging of Kom Aushim

1 Digging the Kôm

The digging of Kom Aushim is well documented in the historical record. While surveying the Fayum in January 1890, Flinders Petrie recognized that its ancient town had been the most important in the district but also noted that the necropolis to the north had been ransacked prior to his arrival.^[4] A few years later, in the winter of 1895/6, David Hogarth and Bernard Grenfell identified the site as the ancient Karanis while excavating for the Egypt Exploration Fund (= EEF).^[5] The publications stemming from their work are noteworthy in the history of papyrology for the attention they paid to archaeology,^[6] but so too do both archival materials from the dig and the excavators’ personal correspondence shed important light on the site’s condition at the time. The most interesting example of the latter is an early entry from Hogarth’s daybook, in which he reports that the kôm was “much worked” on the southeast and southwest, that many homes were untouched on the north and northwest, and that, “Ali Habibi? worked 6 days. Dimitri also, but not Farag (old ghaffir)”.^[7] This last remark evidently identifies the dealers whose agents had already probed the kôm and even Hogarth’s source.^[8]

Although Hogarth and Grenfell had intended to excavate Kom Aushim for “a fortnight or three weeks”, after only ten days of work probing its temple, houses, and necropolis they transferred the dig to Kom el Atl – the ancient Bacchias.^[9] In a letter sent shortly after this first foray, Hogarth complained that most of Karanis’ necropolis had been plundered already in antiquity, and the town itself in more recent times.^[10] At the conclusion of the season, they nonetheless returned to Kom Aushim briefly, this time for less than a week’s work whose primary goal was planning the temple.^[11] Hogarth had initially remarked that the kôm was big enough to last more than a single season^[12] – a claim subsequently justified by the Michigan excavations – but the EEF expedition’s explicit goal was papyri and, on this front, it was a disappointment.^[13]

In retrospect, a mere ten days of digging seems incongruous with so dismissive an initial assessment, but Hogarth believed that the dig’s exploration of temple, necropolis, and houses was sufficiently wide-ranging to establish the site’s potential.^[14] The disappointment was evidently prompted by several considerations, market gossip not least among them. For, as the introductory section of Fayum Towns and their Papyri makes clear, the decisions both initially to excavate for papyri at Kom Aushim and, subsequently, to cut the work short were influenced by the rumor mill of the Egyptian antiquities market:

During this period, 1887–1894, dealers’ agents were busy at other sites, especially at Kôm Ushîm, where the houses in the suburbs are said to be almost as fruitful as those at Dimê.^[15]

No doubt on the basis of such reports, Hogarth and Grenfell had hoped for better results from their work. But when the excavations were not immediately bountiful, the pair promptly concluded on the basis of the kôm’s condition that they had been beaten to the punch and that their energies would be better spent elsewhere.

Contemporary sources corroborate this account of the chatter circulating on the Egyptian antiquities market. When Edgar Goodspeed wrote to the Rev. John Alexander to inquire into the origins of a lot of papyri purchased with the latter’s assistance in 1897 – the bulk of which became P.Kar.Goodsp. – Alexander responded with the following report from an anonymous source:

(1) The papyri were found in the Faiyum at a town called Washim. (2) They were found by people digging in the heaps or “tells” of dirt and rubbish at the edge of the town. As they dug down, they found walls of houses standing in position. In the rooms of these houses or some of them they found the papyri – some were in sacks, others in piles in corners, etc. (3) The pieces are all from that one town but probably not all from the same old ruin – but from different houses.^[16]

Once more, the historical record indicates both that Kom Aushim was an important source of papyri at the end of the nineteenth century and that the edges of town were especially fruitful.

It is clear that the southern slope of Kom Aushim and other peripheral areas had been probed by dealers’ agents and local farmers prior to Hogarth and Grenfell’s 1895/6 season. But there can also be no doubt that the two Englishmen’s arrival and the commencement of their excavation did nothing to discourage further activity of this sort. Unpublished material from the dig captures the interest their work aroused in the Fayum’s residents. Hogarth’s daybook, for example, describes the difficulties of selecting and managing the local workforce that occasionally thronged in hope of employment, and it repeatedly remarks on various uninvited visitors to the site in the wake of their arrival.^[17] After the excavation had been transferred to Kom el Atl, furthermore, Hogarth learned that men were observed digging at Kom Aushim in their absence.^[18] Whether that work was the innocuous harvesting of sebakh or the targeted pursuit of antiquities cannot be determined,^[19] but the latter would not surprise. However brief it may have been, the 1895/6 season brightened whatever spotlight was already on the kôm, further advertising it as a potentially lucrative source of papyri.

The digging certainly continued. By the time the University of Michigan was granted a concession and had made the necessary preparations to excavate in the final weeks of 1924,^[20] several decades of informal excavation and the extraction of sebakh had dramatically altered the topography of Kom Aushim:

Twenty-nine years had elapsed since Grenfell, Hunt and Hogarth had last dug at Karanis. In the interval the mound had been subjected to the unceasing attacks of sebâkh diggers, who had operated with light railways, as well as with the usual beasts of burden. The extreme eastern and western parts of the town had practically disappeared and the southern face of the mound had been plundered extensively. A great gap had been made in the centre of the mound stretching from its southern nearly to its northern limits. Around this irregular, crater-shaped, hole the remaining portions of the mound formed a sort of rampart, in some places thirty or forty feet high.^[21]

The proprietor of the light railway in question was the Daira Agnelli estate, a company with a permit to extract ca. 200 cubic meters of sebakh daily. In the central part of the kôm where the railway’s heads lay, it had dug to bedrock, destroying all traces of the ancient village.

The obliteration of the kôm’s center via the industrial extraction of sebakh is a famous, dramatic part of Karanis lore. But at the same time, it seems only to have begun in the early 1920s:^[22] although the impact was unmistakable (and catastrophic for archaeology), it was also only a relatively recent development. More typical, rather, were the aforementioned expeditions by dealers’ agents as well as local farmers, and the extraction of sebakh by donkey. Boak’s complaint about “unceasing attacks” cannot be verified, but captures the gist of the kôm’s incremental disturbance. Earlier reports had noted that its southern face had been plundered extensively, that both extreme east and west of the village had “practically disappeared”, and that the edges (or ‘suburbs’) of town were fruitful sources of papyri.^[23] An aerial photograph, taken only days after the first season of the Michigan excavation formally began, documents the damage; both the monumental crater and the pockmarked southern slope of the kôm are clearly visible from the sky.

Figure 1:

19 January 1925. Aerial Views. 47 sqdn. R.A.F. Fayoum. Survey A3. 6000 feet. (Kelsey Museum of Archaeology, University of Michigan, neg. no. 4.1992)

2 Early Purchases of Karanis Papyri

It is against this historical backdrop that we must consider the appearance of Karanis papyri on the Egyptian antiquities market. In the final years of the nineteenth century, batches of them became available. The earliest documented items belong to the group of papyri purchased by James Andrew Sandilands Grant Bey ca. 1888, now in Aberdeen, which includes six items from Karanis.^[24] In Berlin there are 137 Karanis papyri originally acquired by Heinrich Brugsch in 1891 as part of a massive lot,^[25] as well as five each from purchases made by Rudolf Mosse and Karl Reinhardt, which came to Berlin in 1894 and 1896, respectively.^[26] Chicago and New Wilmington are the homes of Karanis papyri purchased by Edgar Goodspeed and the Rev. John Alexander in and around the year 1897.^[27] In the British Library, Karanis papyri are scattered across several lots acquired between 1891 and 1903. Especially important for my purposes is its 2 February 1891 purchase from the Rev. Greville John Chester,^[28] but there are other acquisitions from L. Paul Philip,^[29] the Rev. Chauncey Murch,^[30] and Bernard Grenfell.^[31] Via Karl Wessely, Grenfell also sold a trio of Homeric papyri from Karanis to the Vienna Papyrussamlung,^[32] and among those he sold to the Bodleian Library is a single papyrus from Karanis.^[33] As for the Karanis papyri that were excavated during the 1895/6 season, the EEF either repatriated or distributed them in the early 1900s, following the publication of P.Fay.: they are now in Alexandria, Cairo, Cambridge, the British Library, and Toronto’s Royal Ontario Museum.^[34] In sum, I count approximately 330 papyri whose provenance has reasonably been identified as Karanis and whose acquisition occurred in the years 1888–1903.^[35]

Despite their varied destinations and discrete acquisition histories, early discoveries from Karanis occasionally overlap in interesting ways. Some were promptly recognized as belonging together, while others have been grouped together in the course of their scholarly analysis.^[36] A good example of the former is the archive of approximately 150 receipts for seed corn issued by the Karanis sitologoi,^[37] which is divided principally between the Goodspeed/Alexander collections and Berlin.^[38] The latter group is represented by the papers of Mikkalos son of Petheus (alias Ptolemaios) and of Gaius Iulius Apollinarius, respectively. Purchased papyri pertaining to Mikkalos and his family are now in London, Berlin, and Oxford,^[39] while those of Apollinarius are now in Berlin, Warsaw, and London.^[40] Common to these clusters of papyri is the aforementioned narrow window of acquisition. Only a few items could have been purchased after the turn of the century,^[41] which is to say that most – if not all – were acquired within a decade of one another.

Although their institutional and geographical distribution indicates that the papyri in these clusters were not always purchased together, the compact chronology of their acquisition is nonetheless suggestive: were some perhaps unearthed together? At first glance, this question may appear fundamentally insoluble; the reality of purchases made on the antiquities market is that dealers’ inventories as well as a given lot’s composition are arbitrary – even potentially random. Furthermore, dealers were not necessarily well informed about provenance.^[42] A single purchase, in other words, could consist of all sorts of material from different sources and origins; papyri uncovered on a single occasion, conversely, need not have come to market together. Brugsch’s massive 1891 purchase, with its approximately 3,000 items, is illustrative: only 137 of them (i. e. < 5 %) have been located in Karanis on internal grounds. But upon closer examination of even a lot such as this, patterns nonetheless begin to emerge. Of those 137 texts, papyrological analysis has distinguished half a dozen different assemblages. In addition to the aforementioned pieces pertaining to Gaius Iulius Apollinarius and the sitologoi archive, the Karanis papyri in the Brugsch lot include nine texts pertaining to Aurelius Longinus alias Zosimos,^[43] three to Sabinus son of Zosimus,^[44] two to Ninnaros alias Ptolemaios,^[45] and four to Tapetheus daughter of Ptolemaios.^[46]

It would be rash to assert that all of these texts were unearthed in a single place on a single occasion, but the hypothesis that individual clusters of related texts came out of the ground together is attractive and, in a few cases, defensible. No other papyri pertaining to Tapetheus daughter of Ptolemaios, for example, are known apart from those purchased by Brugsch: the acquisition of four items within a single lot is consistent with a common archaeological origin. The two papyri pertaining to Ninnaros alias Ptolemaios are suggestive in other respects. Two is of course too small a number to allay the objection of a serendipitous coincidence,^[47] but the acquisition history of a further item provides a modicum of ballast: P.Lond. II 182b is an extract from a census roll involving Ninnaros purchased from L. Paul Philip in 1891 – i. e. the very same year as the Brugsch lot and its Ninnaros papyri. The number of items relating to Ninnaros alias Ptolemaios remains worryingly small, but their near-contemporaneous acquisition by different collections invites the hypothesis that, although they evidently came to market separately, they might have shared an archaeological context or were unearthed in proximity to one another.^[48]

By virtue of its magnitude, the group of receipts for seed corn issued by the Karanis sitologoi makes the strongest case for purchased papyri with a common archaeological origin. Approximately 100 of them were acquired by Goodspeed (1897) and Alexander (ante 1902), but forty-three others were part of the 1891 Brugsch lot.^[49] In the early days individual pieces also came to Gothenburg^[50] as well as museums in Alexandria^[51] and Aberdeen.^[52] Goodspeed was made aware of a connection between his papyri and those in Berlin early on,^[53] and in republishing the Berlin papyri alongside his own he noted that the overwhelming majority (= 95 %) date to the autumn of the twenty-second year of Antoninus Pius’ reign.^[54] Most significantly, he also recorded how many came into his hands rolled or even stitched together: their relatively ‘intact’ state of conservation is consistent with the aforementioned report regarding their origins and comes as close as we can for purchased papyri to establishing a common archaeological provenience.^[55] Internal evidence and the physical condition of the papyri combine to indicate that they were deliberately archived, i. e. collected and deposited in antiquity, though exactly where remains unknown.^[56]

Moving beyond the 137 Karanis papyri in the Brugsch lot, texts from other early purchases are also occasionally linked in ways that indicate a common archaeological origin. Take the case of a lot of thirty-seven papyri, now in the British Library, that was purchased from the Rev. Greville John Chester in 1891.^[57] The papyri’s provenance is for the most part Arsinoite, but fourteen of them can be placed specifically in Karanis, and fully half of those (i. e. seven) pertain to one Mikkalos son of Petheus (alias Ptolemaios) (Table 1).

I would not go so far as to claim that all fourteen of the Karanis papyri in this lot came out of the ground together, but as regards the Mikkalos items, at least, the situation resembles that of the texts pertaining to Tapetheus daughter of Ptolemaios acquired by Brugsch: the coherence of a handful of texts within a single lot belies the likelihood of an accident. But the Mikkalos papyri in the Chester lot also resemble those pertaining to Ninnaros alias Ptolemaios for their external corroboration: a trio of pieces in Berlin and an additional one in Oxford are also connected to this same Mikkalos.^[58] Although two of those four items’ origins are not documented,^[59] the others were purchased in 1896 and ca. 1897–1899, respectively. As far as we can tell, in other words, the chronology of acquisition remains quite compact, reinforcing the likelihood of a common origin for the entire group. Like Goodspeed’s sitologoi receipts or the later case of Aion son of Sarapion’s papers,^[60] members of which ended up in a handful of collections, these papyri could easily have been divided after discovery and before they were brought to market.

3 Collating the Evidence from Field and Museum Archaeology

Recent years have witnessed an outburst of research into the Mikkalos papyri, including both editions of new texts and reassessment of the whole. Especially notable for my purposes are the two copies of a will drawn up for Mikkalos’ father Petheus, neither of which is the original.^[61] Unlike the aforementioned texts in Berlin, Oxford, and London which were acquired on the Egyptian antiquities market in the last decade of the nineteenth century, these wills were excavated decades later by the University of Michigan, as part of a large cache of papyrological material. The dig’s record of objects assigned this cache the field number 26-B17F-A, an annotation which indicates both the season of excavation (1926/7) as well as the findspot: Room F of House 17, in the ‘B’ level.^[62] For the purposes of Museum archaeology, the addition of papyri acquired via a controlled archaeological excavation to a larger group acquired by market purchase makes the Mikkalos papyri exceptional: in Karanis, as elsewhere, the manner of acquisition for groups of papyri tends to be exclusive – one or the other.^[63]

The complicated, compound acquisition histories of the Mikkalos papyri open a new vector for research, namely, the possibility that archaeological data can be brought to bear on market acquisitions. House B17, the excavation records indicate, was situated on the northwest corner of the intersection of Streets BS1 and BS2, in the western neighborhood of Karanis that the Michigan team dubbed ‘Area G’.^[64]

Although the testimony of this single findspot is insufficient to forge a meaningful connection, the Museum archaeology of London’s 1891 Chester lot reveals another significant intersection. Among that group of thirty-seven texts (which included fourteen papyri from Karanis, seven of which form the core of the Mikkalos papers) is British Library Papyrus 155, whose two sides preserve a fragmentary surgical treatise on methods for reducing dislocations of the jaw and choliambic verses attributed to Cercidas, respectively.^[65] Although there is no internal indication of provenance, we know that this papyrus comes from Karanis because the Michigan excavation uncovered an additional fragment of it, published in the second volume of Cairo papyri from the Michigan excavations.^[66] I have no doubt that the purchased and excavated fragments belong to the same roll: the two texts’ hands are identical.^[67]

Inasmuch as the Mikkalos papyri and the surgical treatise were both acquired by a combination of excavation and a single purchase, in concert they provide compelling testimony regarding provenience. The excavated fragment of the surgical treatise was assigned the field number 26-B12L-A in Michigan’s record of objects, which indicates both the season of its discovery (1926/7) as well as its findspot, namely, Room L of House 12, in the ‘B’ level – that structure’s courtyard. When one consults the plan of Area G, the overlap with the excavated wills for Mikkalos’ father is clear: the papyri were excavated during the same season, and their findspots are closely related. Houses B12 and B17 are located on corners diagonally opposite to one another, at the intersection of streets BS1 and BS2 in Area G (see Figure 2). Archaeological data regarding the excavated fragments, in other words, is consistent with the hypothesis of the purchased lot’s cohesion. Eight of the fourteen Karanis papyri in London’s 1891 Chester lot now have ties to two separate structures, located within meters of one another in Area G.

The fact that these papyri belong to otherwise unrelated groups of texts cinches the connection hypothesized by Museum archaeology, which wondered whether at least some of the fourteen Karanis papyri in the Chester lot were unearthed together. Buttressed by the excavation data, I can only conclude that we should locate these London papyri in the archaeological record: the odds of fragments from the same two unrelated sets of texts turning up in both a single purchase as well as during the controlled excavation of a neighborhood are minuscule, unless the purchased items also originated in that neighborhood.^[68]

There are two important corollaries to this conclusion. The first is that the archaeological record in Area G was being disturbed already by 1891, when the Chester lot was acquired. Above and beyond more general reports about the kôm’s suburbs and southern slope, then, we can pinpoint at least one instance of historical digging in Karanis with some precision. The other corollary involves the coherence of the papers pertaining to Mikkalos son of Petheus. Since the London portion of this group of texts can now be located in the archaeological record, the likelihood that we are dealing with an otherwise unrelated and arbitrary group of texts reconstituted only via scholarly analysis is diminished. Instead, they more probably constitute a single assemblage of materials, gathered at some point in antiquity before eventually being deposited in Area G.^[69] That assemblage plausibly includes the texts now in Berlin and Oxford as well, whose acquisition history – where documented – is compatible with the ones in London. The contingency of purchases made on the antiquities market prevents us from determining whether the purchased papyri were uncovered together on a single occasion or gradually, over time, but there is good reason to posit proximal findspots in Area G, prior to the papyri’s division and appearance on the antiquities market.

Figure 2:

Area G Map

4 Mikkalos in Light of the Papyrology and Archaeology of Area G

Advancing the archaeological contextualization of the Mikkalos papyri beyond these important conclusions is challenging. For one thing, the excavation data is imperfect. From the start the Michigan dig was plagued by mistaken inferences about stratigraphy,^[70] but when its third season of excavation was about to get underway in the autumn of 1926, the surveys and maps from the previous seasons remained incomplete.^[71] Worse, the triangulation chart that was prepared during this season was subsequently found to be faulty and an unreliable basis for further surveying.^[72] Only a preliminary plan of Area G (= Figure 2) was ever completed, and drawings were drafted for only a few “of the more important houses”,^[73] which included neither House B12 nor B17.^[74] It was not until the subsequent 1927 season that the surveying and mapping of the site was set on solid foundations, which produced the more familiar (and professional) plans of Karanis that have since been frequently (re)published.^[75]

The shortcomings and limitations of the documentation are partially explained by the difficult conditions in the field,^[76] but also by the enormity of the task – logistical, financial, methodological, managerial, and bureaucratic – that coordinating and overseeing the kôm’s excavation posed to a team of seven principals in the autumn of 1926.^[77] By the dawn of that third season of work, moreover, the dig was in crisis. The previous season had been marked by financial irregularities and dysfunction in both management and administration.^[78] Worse, its dumps of sebakh had proven inadequate to supply the Daira Agnelli through the summer, as a consequence of which the company not only had resumed digging the kôm of its own accord but was also threatening to resort to artificial fertilizer in the future, which would have only compounded the crisis.^[79]

To right the ship, Michigan replaced James Starkey as field director with Enoch Peterson, who would oversee excavations until their conclusion, and the team promptly fixated on Area G for reasons both practical and archaeological. Providing the Daira Agnelli with sufficient stores of sebakh via excavated fill was a priority, but so too did the team believe that this neighborhood would yield well-preserved houses from Karanis’ latest period of occupation.^[80] The latter expectation would be disappointed, but the area indeed proved bountiful in sebakh, with some structures buried by up to three or four meters of debris.^[81] The quantity of fill proved as much a curse as a blessing; in the course of excavation, the dig’s workforce removed an average of 200 tonnes per day,^[82] constructing a series of massive dumps alongside the excavated structures to supply the Daira Agnelli’s light railway (Figure 3). But so great was the work that not every structure could be cleared completely to its foundations, which is another reason why only rudimentary plans for the area and its buildings were completed.^[83]

Given the limitations of the excavation data and the challenges the Michigan team faced in the 1926/7 season, the analysis of Area G depends overwhelmingly upon the Kelsey Museum’s invaluable record of excavated objects. Here one encounters the second major obstacle to comprehending and contextualizing the Mikkalos papyri, namely, the circumstances of their deposit. House B17 yielded vast amounts of papyrus in addition to the fragmentary wills for Mikkalos’ father Petheus: no fewer than 142 individual fragmentary papyri share the field number 26-B17F-A.^[84] This cache of material, which dates overwhelmingly to the second and third centuries, has been the subject of rich scholarly discussion. Building upon prosopographical analyses first conducted by Silvia Strassi, Peter van Minnen famously identified House B17 as the home of the local tax-collector Sokrates son of Sarapion, a conclusion he premised on the idea that the structure’s papyrological finds were a primary deposit consisting of Sokrates’ personal archive.^[85] Even though van Minnen has revised aspects of this thesis in several significant ways,^[86] his portrait of Sokrates is so vivid that the idea he resided in the neighborhood persists. Its influence can be discerned in the not uncommon assumption that the papyri from this structure and its environs were once his.^[87]

Figure 3:

Dump heaps, season 1926–27, from north to south, taken in the middle of February, 1927. Credit: J. Anthony Chubb (Kelsey Museum of Archaeology, University of Michigan, neg. no. 5.2370).

Since the excavated wills share the field number 26-B17F-A with the core of Sokrates’ papers, he is the elephant in the room during the contextualization of the Mikkalos papyri. What does one have to do with the other? Despite the influence of van Minnen’s seminal analysis, which marked a watershed moment in papyrology’s interdisciplinary turn, the archaeology is far from clear-cut and it is dangerous to subordinate everything to Sokrates. The most significant criticism of his argument is grounded in archaeological theory, which objects on that grounds that it confuses the papyri’s archaeological context (i. e. the circumstances of their excavation) with their systemic context (i. e. their use in antiquity).^[88] Put simply, the fact that excavation unearthed papyri in House B17 (for example) does not itself indicate that the papyri were used or read there in antiquity, or that Sokrates once lived there or nearby. It only indicates that they eventually came to be deposited there.

The composition of the deposit similarly warns against jumping to conclusions. Very little in the latest volume of Cairo papyri from the Michigan excavations, for example, has any obvious relationship to Sokrates or his family, which is consistent with the papyri from House B17, more generally. It yielded several papyri like the Mikkalos wills that do not obviously pertain to Sokrates, his business, or the individuals thought to comprise his family.^[89] The same diversity is evident in the finds from Street BS1, too, on which the house was situated.^[90] Instead of proceeding inclusively on the assumption that archaeology has uncovered a unified archive of which Sokrates is the protagonist, then, the opposite approach is called for: each papyrus’ candidacy for inclusion among his papers should be evaluated on a case-by-case basis. Disabused of the assumption that the papyri ought to relate to one another, one observes instead a mixed batch of material from multiple sources dating to the second and third centuries. In this respect, Cornelia Römer is correct to note that we have a part of the archive, but not the archive itself.^[91] Although they share a field number, in other words, the excavated Mikkalos papyri do not pertain to Sokrates. Nor do they need to.

Another reason to decenter Sokrates from the discussion is the evident confusion in the archaeological record, characterized by the dispersal of papyri across Area G. Sokrates’ papers, for example, were not restricted to House B17, but were also unearthed in several other contexts, as well.^[92] Such ‘wandering’ is a feature of the neighborhood, more generally.^[93] The literary papyri resemble those of Sokrates in this respect: there are fragments of a grammatical treatise which were excavated in separate contexts,^[94] and three separate locations yielded the half-dozen fragments of Menander’s Epitrepontes that have been identified to date.^[95] P.Cair.Mich. II 9 + III 6 offers a particularly extreme example: these fragments of Demosthenes’ On the Crown were excavated in entirely separate regions of the kôm.^[96] There are, of course, several potential causes for the scattering of papyri in Area G’s archaeological record, but my analysis of the Mikkalos papyri and the 1891 Chester lot provides at least a partial explanation. Because this neighborhood was already being disturbed by 1891 (to say nothing of any subsequent incursions), it should come as no surprise that potentially discrete deposits of papyri in Area G were stirred up and confused in the decades prior to their excavation by the Michigan team.

It must not pass unmentioned that the excavation data occasionally bear witness to disturbance of this kind. In a room in House B17 adjacent to the one in which the large cache was uncovered, the dig excavated another dozen or so papyri. These were assigned the field number 26-B17E-A in the record of objects, to which was added the annotation “papyrus very high in debris above the room”.^[97] The excavators, one observes, were sufficiently perplexed by the archaeological context of these papyri to make a note of it: implicit in the annotation is the question of how the papyri came to be deposited so high. Two explanations come immediately to mind: either they were deposited at a much later date – e. g. during the structure’s post-abandonment phase – or the house’s archaeology was disturbed before excavation. Some combination of the two cannot be discounted, but my analysis certainly corroborates the latter. In either case, this is no primary deposit.

Unfortunately, the archaeology of Area G prevents us from ascertaining how the papyri in Area G were deposited or how many deposits there were. There may well have been a massive dump of mixed material like that found in the Cantina dei Papiri at Tebtynis,^[98] or there may have been a number of individual deposits in different locations across the neighborhood. Some combination of the two is possible, as well. But it seems clear that the material is heterogenous, and to press that point, I would offer an additional example drawn from the literary evidence. Thanks to the second and third volumes of Cairo papyri from the Michigan excavations, a sufficient number of Homeric papyri from Area G have been published to illustrate in simple terms their diversity and distribution.^[99] On Figure 4 are mapped eleven fragments, all from the Iliad, and none of them from the same roll.^[100] Only a few were found in the same context: no fewer than eight separate structures in Area G yielded Homeric papyri. Several books are represented in multiple copies, with Iliad 1 being especially well attested. Although the phenomenon of wandering papyri is attested elsewhere in Area G, the likelihood that these derive from a single library kept in a single house is low; in concert, the multiplicity of copies and their distribution across the neighborhood, rather, is evidence of multiple archaeological deposits or – like the Cantina dei Papiri – a larger dump of material that was amassed from various sources and subsequently deposited as part of some indeterminate process.

The evidence of the Homeric papyri dovetails nicely with the conclusions regarding Area G, more generally. There is a great deal of material, and although the circumstances of its deposit(s) are unclear, its diversity is more indicative of a multiplicity of origins than a single source, unless that single source was itself a composite. The disturbance of the archaeological record this article has identified provides a specific explanation for how it all came to be confused, but there could well be others, especially given the decades-long interval between London’s 1891 acquisition of the Chester lot and Michigan’s 1926/7 season in Area G. The finds as excavated need not represent the state of the neighborhood as it was inhabited; equally plausible is that they reflect a later habitation phase whose archaeology was further confused. The confusion limits the kinds of contextualizing conclusions that can be drawn: as regards the Homeric papyri, assigning the fragments to particular owners or individual libraries would be a fool’s errand.^[101] The unclear circumstances of deposit make it similarly difficult to assign the structures in Area G to particular residents.

Figure 4:

Area G Map, with findspots of Homeric papyri indicated

Despite the challenges posed by the archaeology of Area G, there is nonetheless opportunity. Because they permit legacy data from the Michigan excavation to be brought to bear on papyri purchased on the market, the Mikkalos papyri are unique. Historical purchases are not always transparent about provenance, and as the records from Area G testify, there are a variety of reasons why the excavation data from even relatively careful early digs can challenge interpretation. It is therefore encouraging that the conclusions yielded by the interdisciplinary analysis of this group of texts are significant. By pinpointing on the map at least one historical occasion of digging on the kôm, they go some way towards explaining the confused states in which the vast papyrological output of Area G (and House B17, in particular) were preserved in the archaeological record.

Neither the earliest acquisitions of Karanis papyri nor the digging of Kom Aushim were single, momentary events, but were rather complex processes undertaken over a number of years by a variety of different actors in a variety of different guises. The scholarly reconstruction of those processes and their consequences accordingly requires a similarly sophisticated approach combining multiple disciplinary approaches – from history, papyrology, Museum archaeology, and field archaeology – to piece together what was happening in the field and in the shops of the Egyptian antiquities market decades ago. This paper has attempted one such analysis of the Mikkalos papyri, one that synthesizes these approaches to offer a brief but rich glimpse of one group of papyri’s history. From their findspot(s) in Area G to the late nineteenth century market, and from their acquisitions by various hands to their eventual institutional homes in London, Berlin, and Oxford, much more is clear about this group of papyri than is ordinarily true for purchases made on the market alone. And although my conclusions take issue with the tendency to discern Sokrates son of Sarapion throughout Area G, that too is an important step in the right direction for the scholarly understanding of the neighborhood.