THE IMPACT OF ARCHAEOLOGICAL RECORDING ON THE STUDY OF METAL ARTEFACTS. MYCENAE 1939: A CASE STUDY

Aegean Archaeology is one of the oldest branches of prehistoric archaeological scholarship, and many important settlements and cemeteries, such as those at Mycenae, were excavated before the development of more advanced recording techniques that we take for granted today. Nevertheless, the significance of these legacy data as a source of knowledge means we must still find ways to integrate them into our interpretations, despite their limitations. To derive the most robust results possible, it is important to understand exactly what types of impact these earlier recording strategies may have had on our perception of their findings. Yet this type of investigation is rare, meaning that in many cases we know more about the repercussions of taphonomy and the social practices of past societies on the archaeological record than those caused by the actions of our own predecessors. In preparation for a holistic study of all aspects of the use of metals at the Late Bronze Age site of Mycenae, this paper details the exploration of the recording processes in place during the 1939 excavation season. This has been identified as an ideal case study for examining recording strategies because its organisational structure gave each trench supervisor a great deal of individual freedom. Concentrating on their consequences for metal artefacts in particular, each stage of the recording process, in the field, in the museum and in publications, is discussed, as is the aftermath of the Second World War.


INTRODUCTION
Mycenae is one of the most thoroughly excavated Late Bronze Age sites on the Greek mainland. Not only did the site attract a great deal of early archaeological attention through its Homeric links to Agamemnon, the astonishing objects uncovered in the very first digging season in , the relatively high level of preservation of architectural marvels, such as the Treasury of Atreus and the Cyclopean fortifications, its clear position as one of the foremost centres of Mycenaean culture and the comparably easy access to the prehistoric remains, in contrast to Athens or Thebes, ensured that it retained that attention throughout the twentieth and into the twenty-first century. This depth and breadth of investigation makes Mycenae the ideal candidate for a holistic examination of the use of metals in Mycenaean societies, which is the basis of the ongoing 'Forging Society at Mycenae' project. Also of great importance in the selection of Mycenae for this study is the availability of the original field notes of the Helleno-British excavations team, through the Mycenae Archive based at the Faculty of Classics at the University of Cambridge. Holistic studies that integrate data derived from multiple fieldwork sources are, of course, highly sensitive to the recording strategies employed. It quickly became apparent that a different reading of events could be reconstructed if solely relying upon the field notes, the registered finds or the publications. Nor was it safe to assume that, drawn together, these different sources would provide a complete and coherent narrative. At each stage some level of pruning of information occurred, with certain elements being foregrounded and others sidelined as less significant; naturally the type of decisions made also changed over time.
The  excavation season stands out as an especially interesting example of this phenomenon. Each trench supervisor had control over their own recording strategy, both in the field and the museum, and clear differences between them and the consequences this has had on the recording and retention of metals in particular can be readily identified. The Second World War, and decisions taken at the Nauplion Museum in its wake, affected the current inventory of particularly struck his interest or were related to his wider research questions. He seems to have managed to visit and make notes on each excavation area almost everyday. Nevertheless, there are a few minor inconsistencies. For example, Wace first mentions the West Terrace Wall on  July but excavation began on  July, under the supervision of Vronwy Fisher. Wace also makes no entry for the Greek Temple on  July, but Frank Stubbings' diary shows no break. Therefore the dates given for areas that rely solely on Wace's daybook should be considered indicative rather than definitive, as the aim of this document was not to give a comprehensive daily commentary. judgement of the archaeological value of finds at a level that is simply no longer expected on a modern excavation. Over recent decades, the recording practices on archaeological excavations have become more rigorous and standardised. The adage that excavation is inherently destructive has been taken to heart, and maximisation of recording has become one of the chief aims, demonstrated by developments such as photogrammetry. The recovery and retention of all portable finds, regardless of size and preservation status, has become accepted as a standard goal for fieldworkers. This is partly because scientific advancements have enabled us to derive information from fragments previously considered to be of low archaeological value, using techniques such as lead isotope analysis and petrography. Analysis of the heavy fraction derived from flotation and investigation of soil chemistry provide new ways of studying metallurgy in the past (Boyd et al. , , ). Continued advancements in conservation play another important role in this change. Most modern field projects have access to a professional conservator who, as well as ensuring the preservation of kept remains, thereby making retention worthwhile, is also able to provide best practice guidelines for recovery that significantly increases its success.  The digital revolution has also made it easier to store large quantities of data and conduct studies using far larger datasets, making the gathering of this additional information a worthwhile investment. These changes have been accompanied by a commensurate and substantial increase in cost, time and storage needs.  The conceptual framework for archaeological fieldwork in  was in fact very similar to today: the desire to maximise the archaeological value of the work undertaken. However, the perception of archaeological value has changed rather dramatically over time. Thus, given that less information could be derived per trench, excavation took place on a grander scale. Furthermore, this early period of excavation in the Aegean generated the foundational knowledge for our discipline. Wace (b, ) listed the main results of the  season as 'the determination of the date of the temple foundations, the study of the architecture of the House of Columns, the tracing of the prehistoric cemetery outside the Lion Gate, and the knowledge gained of the construction and date of the Treasury of Atreus' and correctly described them as 'factors of the first importance for the history and culture of Mycenae'. Proving that the fortifications post-dated the Prehistoric Cemetery and did not respect its integrity (A.J.B. Wace , , ) had profound consequences for our understanding of Grave Circle A, and drove future field investigations in the vicinity to clarify its complex chronology. These examples demonstrate how different the underlying research questions were to those that stimulate modern scholars today, whose chief aims are to refine, improve, clarify and add detail to these foundations.
Thus, an absolute recording strategy was not only unfeasible in , it was also considered to be of very little archaeological value. Each trench supervisor took multiple decisions about recording in the field on a daily basis. Without aids such as standardised recording sheets, the responsibility held by each supervisor was far greater, and they would have relied upon their previous training and experience to make these judgements. However, one key way in which the  season differed from the Artemis Orthia dig at Sparta and the s campaign at Mycenae is the relative experience of the trench supervisors. Those on these earlier excavations had a level of archaeological experience that was not dissimilar to that of the dig director. This was not the case for the  season. Before attempting to analyse the evidence from the field notes, it is therefore necessary to discuss the type of excavation experience and training that each supervisor had received prior to arriving at Mycenae. This background information is vital to understanding the context within which each trench supervisor made their decisions, and, as I will discuss later,  I would like to thank Kim Shelton for drawing my attention to this fact.  A thorough description of an excavation that has sought to employ many of the currently most up-to-date methods in field archaeology, and the advantages and disadvantages this has produced, can be found in Boyd et al. , which details the most recent fieldwork seasons on the islands of Keros and Dhaskalio. In particular, the input of specialists during the planning stage, strategies for standardising recording, the use of on-site photogrammetry, the adherence to single-context recording, the employment of a digital recording system and the accessibility of data across the entire team are regarded by the authors as essential to a modern approach. STEPHANIE AULSEBROOK  traces of the practical impact that their previous experience and training had on the formation of the archaeological record are clearly discernable.
The  trench supervisors Before the establishment of the Institute of Archaeology at the University of London in , formal training programmes for excavation were almost non-existent, and students of Aegean Archaeology from British institutions were taught these skills through active participation in fieldwork projects, facilitated by the British School (French , -; Thornton , -, ). Therefore, before the  season commenced, Wace encouraged his three postgraduate students to travel to Greece to carry out their own research projects and gather experience. The following is a brief overview of the backgrounds for all six trench supervisors, with a focus on their opportunities for contact with metal artefacts.

Vronwy Fisher
In December  Vronwy Fisher assisted Richard W. Hutchinson during his excavation of a tholos tomb at Kephala, near Knossos, keeping the catalogue of small finds (Hutchinson , ; Cadogan ). These included several gold and copper alloy artefacts, such as jewellery and tools; notably even very poorly preserved metal objects were retained, the most striking example of which is a possible copper alloy vessel, which when recovered was no more than a pile of tiny fragments (Hutchinson , , no. , pl. c). She then commenced research on the contents of  Mycenaean tombs at the cemetery of Chalkis, on Euboea, based at Chalkis Museum before joining Wace at Mycenae (Hankey ; Cadogan ). This work was published after the Second World War (Hankey ). The metal finds comprise gold and copper alloy jewellery, weapons, knives and some lead fragments.  She also visited sites and viewed artefacts from around Euboea both before and after the  season at Mycenae, up until April .  Her correspondence with Wace  prior to her arrival at Mycenae for the excavation includes consideration of the manufacture of bronze sculptures, which she describes as having discussed with Homer Thompson (Fisher a, -), and specific references to metal finds from across Euboea; the only other class of material highlighted in this way is pottery (Fisher a, -). She also visited Mycenae earlier that year, meeting members of the future dig team, including the foreman, and checked over one of the areas that she would be responsible for: Schliemann's Dump (Fisher b).

Frank Stubbings
Frank Stubbings became a student at the British School in  and worked on two Mycenaean pottery projects, one on Attica and the other on Mycenaean pottery in Cyprus, Syria and Egypt (Cadogan ; The Telegraph ), both of which were published after the Second World War (Stubbings ; ). His digging experience in Greece included supervising the excavation of area L at Tris Langadas, Ithaca, from - (Benton and Waterhouse , ). Area L contained the remains of three LH III houses, and no metal finds were reported in the publication (Benton and Waterhouse , -). It is therefore difficult to gauge whether Frank Stubbings had had much experience with the recording of metal finds prior to the  Mycenae excavation. From his correspondence with Wace, it is clear that Stubbings only studied  Hankey , -, -. Although  pages of photographic plates of ceramic vessels (along with further photographs of six vessels and three detailed drawings to illustrate the range of motifs) were included in the publication, there is only one figure for metal finds ( fig. ), which comprises drawings of the six best-preserved copper alloy artefacts.   pottery whilst he was in Greece and the Near East before the  excavation began, with passing remarks in his letters and diary to a few metal artefacts, mostly recent discoveries, as well as an ivory pyxis.  He was chosen to publish the Poros Wall hoard and the winged axe mould found during the s excavations at Mycenae (Stubbings a; b). No reason for this decision was given,  and it may simply reflect the significant role he played in the publication team at the time, rather than indicating an otherwise undocumented interest. 

Helen Thomas
Wace suggested that Helen Thomas should conduct and publish a detailed study of the gold and silver objects found in the so-called Acropolis Treasure from Mycenae; this appeared in the / volume of the Annual of the British School at Athens.  This may be the paper that she discusses in her correspondence with Wace (Thomas , ; , ), but no direct reference to metal artefacts was made. This correspondence reveals that she visited Mycenae with her mother in , and took photographs of the Treasury of Atreus (Thomas , ). Wace also entrusted her with the task of researching prehistoric Laconia; her results were published after the Second World War (Waterhouse ; Waterhouse and Hope Simpson ; ; Tomlinson n.d.). For three weeks in May and June , she supervised the excavation at Stavros, Ithaca, under the aegis of Sylvia Benton (Waterhouse ,  n. ). The site consisted of post-Mycenaean graves dug into a deposit that spanned from the Early to the Late Bronze Age; in situ copper alloy, silver, iron and lead artefacts were recovered from the later graves, all fully published, and nine copper alloy and iron artefacts were recorded from the remainder of the deposit. All of the latter were recognisable but only two were published; these were apparently selected because of their greater interest (Waterhouse ). Helen Thomas also supervised the excavation of House T at Tris Langadas, Ithaca, in  and  (Benton and Waterhouse , ). Three small recognisable but broken copper alloy tools were recovered from this building and published (Benton and Waterhouse , , ).

Helen Wace
Since her marriage to A.J.B. Wace in , Helen Wace had taken an active role in his archaeological projects.  She had a background in archaeology, having begun studying for a PhD on the economy of Ostia at the School of Classical Studies at the American Academy in Rome (AAR), where she is listed as a visiting student from the University of Wisconsin and Chicago for the years /-/ (AAR /, ; /, , ). The  season was the first time A.J.B. Wace had returned to excavate at Mycenae since his marriage, but Helen Wace was already familiar with the site, having visited in .
She was a strong advocate of well-organised systematic recording, introducing many important innovations to the framework used by the Helleno-British excavation team at Mycenae that have proved to be invaluable for facilitating research. However, her management of the finds kept in the Nauplion Museum in the aftermath of the Second World War, which is discussed in detail  Stubbings . This is also the case for his own personal diary that he kept for part of . The long tiring hours on excavation prevented him from continuing his diary whilst at Mycenae, with entries only available for the first few days. These do not refer to his digging or recording strategies.


He did not excavate either of these finds.


In one letter to Wace, Stubbings outlines a theory that Mycenaeans colonised Cyprus to take advantage of its mining potential adding 'I think you know the importance I attach to Myc. [sic] metalwork' (Stubbings ). Although clearly aware of and willing to acknowledge the importance of Mycenaean metalwork in a general sense, there is no evidence to suggest that he intended to pursue any studies at this time that were not directly related to pottery.


Thomas /. This volume was actually published in . Hence, the comment that her paper was written four years earlier (Thomas /, ) actually refers to  (Tomlinson n.d.).


These details were very kindly provided to me by her daughter, Elizabeth French.

STEPHANIE AULSEBROOK
 below, do indicate that, at that time, she apparently placed little archaeological value on indeterminate metal fragments.
During the s excavations, Helen Wace played a pivotal role in deciding which objects were registered and which were discarded. In  she assisted with cataloguing, which is mentioned seven times in her diary (H.C.P. Wace ); on four occasions this was carried out as a group activity. Presumably this process included the small finds catalogue, although this is never explicitly stated. However, she does not seem to have intervened in the selection of artefacts. This can be inferred from the distinct retention strategies employed by each trench supervisor and her later expulsion of metal fragments recovered during the  season from the Nauplion storeroom (both discussed in detail below).

Michael Fuller and Colin Kraay
Both Michael Fuller and Colin Kraay were Oxford undergraduate students during the  excavation season. Kraay had some limited fieldwork experience from the UK, as his sixth-form at Lancing College had an archaeological society, known as the Haverfield Society, through which he conducted a small excavation on the school grounds, visited local sites of interest and catalogued antiquities held by the Society's museum (Sutherland ). This excavation revealed a skeleton with a single grave good: a sliver of iron, which was interpreted as the remnant of a pin (Kraay -). It has not been possible to confirm whether Michael Fuller had any prior experience of excavation or object handling before his arrival at Mycenae. He attended Marlborough College, where he specialised in Classics from  onwards.  His form tutor for the summer term of  was H.C. Brentnall, who undertook excavations on the school grounds whilst Michael Fuller was a pupil (Brentnall ), although it is not possible to prove his participation. The only recorded metal find from the s was a coin, which was found during building works rather than the archaeological excavations.  The school had an active Natural History Society that conducted field trips, managed the school's museum and held lectures; whilst archaeology was one of its many subjects of interest, it is not apparent to what extent Michael Fuller participated in their activities, if at all.  Therefore it is unlikely that the latter was at all familiar with the recording of metal artefacts, and it can be safely assumed that both were far less acquainted with material culture from the Greek archaeological record than any of the four other trench supervisors discussed above.
To summarise, Vronwy Fisher and Helen Thomas had an excellent grounding in the recording of metals, having both supervised their excavation and catalogued them, although the distinctions between the type of sites that provided this introduction may have led them to draw rather different conclusions as to best practice. Frank Stubbings and Helen Wace had a good understanding of the Greek archaeological record, although they were possibly less familiar with the handling of metal finds; nevertheless, it is clear that the latter advocated well-organised systematic recording as the cornerstone of good archaeological practice. Michael Fuller had opportunities to gain archaeological experience in Britain prior to the  season, but it is not certain whether he availed himself of these, and no certain contact with ancient metalwork can be established. That of Colin Kraay was based in Britain and had included some contact with metalwork. As will be made clear in the following sections, this heterogeneity with regard to  I would like to thank Marlborough College's archivist, Gráinne Lenehan, for these details regarding Michael Fuller's school record and information about the broader archaeological backdrop at Marlborough College during the s.  Brentnall , . Leary et al. (, ) state that Brentnall himself recovered two Roman coins from the castle ditch, but this is not confirmed by the reference given and it is unclear from where this information was derived.
 His interest in the past is confirmed by his winning of the school's 'McSwinney History Prize' in . However, neither excavations nor museum studies were listed as amongst Michael Fuller's interests in his record at Wadham College, Oxford (I thank the Wadham College archivist, Jeffrey Hackney, for this information), and his significant achievements after , which saw him awarded the DSC and an OBE (Navy News ), have, quite naturally, dominated publicly available records. MYCENAE : A CASE STUDY  their previous experience and training can be recognised as a significant factor in the recording and registration practices each trench supervisor applied.

The field notes
The remainder of this paper is based upon the preserved field notes generated by the trench supervisors and the daybook kept by A.J.B. Wace. This section describes each notebook and gives a basic account of the excavation conducted by each trench supervisor so that the analysis of the metal recording can be placed into context. The notebooks are all physically very alike (those of A.J.B. Wace and Fisher are identical), averaging . cm by . cm, bound with a cover edged with red. The internal organisation varied significantly between them and was clearly left to the discretion of each supervisor. The exception to this is Helen Wace's diary, described below.
Beginning with A.J.B. Wace's (a) daybook, this consists of  pages,  of which were related to the  excavation season.  Four of those are concerned with organisational details for the immediate aftermath of the dig. The daybook had a dual purpose. The account of the progress of the excavation is recorded on the right-hand page, structured by area. The left-hand page was used for drawings, observations and the noting of various comings and goings. The progress account was a day-by-day summary of the most important activities and findings in each area, which Wace would have accumulated through conversations with both the trench supervisors and the workmen (French pers. comm.). It is clear from entries in Helen Wace's () diary that A.J.B. Wace also took an active part in excavation too,  although this is never mentioned in his own daybook. Vronwy Fisher's (c) field notebook consists of  pages worth of notes, sections, plans and sherd analysis. There are a further four pages that cover cataloguing notes, a list of photographs and the organisation of the finds for storage in the museum.
Her account begins with the laying out of three trenches positioned directly outside the fortification wall, lying between Grave Circle A and a Cyclopean terrace wall to the west ( Fig. :), which was marked on Steffen's plan (A.J.B. Wace , ). Trenches A and B were intended to clear part of the dump left by Schliemann's excavations; the third trench ran alongside the terrace wall to clear accumulated earth from its face (Fig. ). The quantity of material being removed daily during this period is indicated by the need to use a horse and cart; large blocks from the fortification wall also had to be relocated (Fisher c, -, ). The third trench against the terrace wall did not include any architecture nor was it particularly prolific, and stereo was reached across the entire trench after seven days. Once the original ground level had been reached in Trench A, it was partitioned into four separate trenches with dividing baulks, which were also later excavated  The others detail the activities and observations of A.J.B. Wace in , when he was able to return to the Argolid for a short period.  See, for example, the entry for  July, which mentions him digging with a knife at the Treasury of Atreus site.  A.J.B. Wace (, ) states that '[i]n view of the discovery of the prehistoric graves to the north, in Area , it was decided that it was not worth while [sic] to open out this trench or to try to trace the walls further'. However, the graves were discovered after work in Trench A was stopped on  July to allow for planning and photography; the same day work began on clearing around the Hellenistic House to search for grave cuttings (Fisher c, ; A.J.B. Wace a, , ). Wace (a, ) described the area as 'suspended for plan' on  July, and this may indicate that he initially intended to continue work here.

STEPHANIE AULSEBROOK
 her notebook only mentions the excavation of the Hellenistic House. It is not clear whether any field notes were kept for the Theatre and the Perseia Fountain House nor, if there were, by whom. The Hellenistic House was an area that had been previously excavated by Christos Tsountas,  which made it especially suited to the search for the Prehistoric Cemetery (A.J.B. Wace , ). Underneath it, Mycenaean walls and graves of the Prehistoric Cemetery began to appear by the end of July, and the trench was extended multiple times to reveal  graves in all (Fisher c, -; A.J.B. Wace , ).
The notebook for the House of Columns ( Fig. :), written by Michael Fuller and Colin Kraay (), totalled  pages,  although some of these were only half filled, and consists of notes, plans and drawings of objects. There are no notes related to the storage of finds or other details such as a photograph list.
The primary task was to uncover the plan of the House of Columns, a structure previously excavated by Tsountas (Fig. ). The account begins with the removal of the deposits overlaying these remains, with walls and floors being uncovered on the first day of digging (Fuller and Kraay , ). On  July, undug deposits were found in Room IX (Fig. ).  This was named the 'Pithos Room' (Fuller and Kraay , ). On  July attention moved to the area of the megaron, removing vegetation which revealed a passageway (Fuller and Kraay , ). Whilst cleaning continued and was extended to Room III, a new area, known as the 'pill-box', was investigated (Fuller and Kraay , -). From  August, the area designated as 'the cellars', around Room A and its surrounding corridors, was cleared (Fuller and Kraay , -). Further undug deposits were discovered in the 'Almond Room' on  August (Fuller and Kraay , ), and, once they had been removed, more cleaning took place around Rooms III and IV (Fuller and Kraay , ). This was all supervised by Michael Fuller. Just before  His work was not published.  The first page contains an entry for  July, when Michael Fuller was assisting Frank Stubbings with the Greek Temples. It relates solely to the architecture and has no relevance to this analysis.  Fuller and Kraay , . The present paper refers to the preliminary designations as given by the excavators, and these were subsequently altered for the final published plans. Various plans showing these original labels can be found in Fuller and Kraay .


Colin Kraay took over on  August,  focus moved to the room at the top of the stairs and back to the Pithos Room, to retrieve some finds previously left in situ and return others to where they had been found (Fuller and Kraay , ). For the first two days, Colin Kraay directed the continued cleaning of areas previously excavated by Tsountas, then removed the remaining vegetation that impeded the planning of the structure, after which cleaning was resumed (Fuller and Kraay , -). Throughout, occasional small pockets of undug deposits left by Tsountas were discovered (see, for example, Fuller and Kraay , ).
Frank Stubbings' () notebook, totalling  pages in all, has  pages devoted to the day-today account of the excavation with eight pages of sherd analysis,  pages recording levels and other  There is a clear change in handwriting, supporting the account given by Wace in his daybook, although Colin Kraay's name was never added to the notebook.

STEPHANIE AULSEBROOK
 measurements, sometimes accompanied by plans, two pages each for a large section and plan of the site, two pages with detailed drawings of inscribed tiles and four pages listing the Mycenae dig kit for . Many pages of the main account were only partly filled, due to his habit of beginning each day on a fresh page, although this practice was not strictly kept and was abandoned from  August.
Stubbings began by clearing around the foundations of the Greek Temples ( Fig. :) (Stubbings , ) and quickly discovered evidence that indicated their multi-phase history (Stubbings , ). He continued systematically, excavating at several points around the foundations simultaneously, whilst investigation of the terrace below the north end, known as the 'north terrace', began from  July (Stubbings , ). On  July a new trench was opened to the west of the foundations, to try to trace the continuation of a Mycenaean wall (Stubbings , ). Two days were spent exploring Schliemann's trial pit, which lay to the south, starting on  July (Stubbings , -), at which point Stubbings seems to have suspended work on the foundations until  August, when props were installed to enable digging to continue (Stubbings , ). From  July, excavation began below the north terrace, in an area known as 'Prinaria' due to the presence of prickly oak scrub (Stubbings , ; A.J.B. Wace and Porada ,  n. ). Digging continued around the foundations, north terrace and Prinaria with additional investigation of the north corridor of the palace on  August (Stubbings , ), the south terrace for two days from  August (Stubbings , -) and an area to the west of the foundations dubbed the 'Mycenaean Shrine' from - August (Stubbings , -), with further tests nearby on  August that lasted for three days (Stubbings , -). Excavation


of the foundations apparently came to a close on  August,  the north terrace on  August (Stubbings , ) and the Prinaria on  August (Stubbings , ). Helen Thomas' () notebook was arranged so that there were  pages devoted to the main excavation narrative and one single page for levels at the front; at the back are  pages of sherd analysis,  pages describing the poros blocks uncovered during the dig, with drawings and plans, and five pages with other sections and plans.
Two trenches were laid out for the first stage of investigation at the Treasury of Atreus ( Fig. :), one on either side of its dromos; as well as excavating, Helen Thomas initially organised her team to  Stubbings , . Although the notebook continues to state that men were assigned to work on the 'temple' up until  August (Stubbings , ), the areas discussed under this heading are in fact the north terrace and the Mycenaean shrine.

STEPHANIE AULSEBROOK
 search for fragments of the facade of the Treasury of Atreus.  Some work was also briefly carried out in the dromos and chamber (Thomas , , ). However, the trench to the north of the dromos became the main focus as it contained the 'bothros' (Fig. ), a large pit containing household rubbish originating from houses on the ridge above.  Another trench was opened up on  July at the corresponding point south of the dromos to find out whether the bothros extended this far, which it did (Thomas , ). It was these trenches that furnished the  Thomas , . According to Kosta Dasis, four baskets of similar material were sent to Athens, but he had continued to find further examples. Vronwy Fisher communicated this information to Wace, which she had received directly from Kosta during her visit in March  (Fisher b, ).  French . The term 'bothros' is misleading as it was not a prepared pit, but this name has stuck (French ,  n. ).

MYCENAE : A CASE STUDY
 important and incontrovertible dating evidence for the construction of the Treasury of Atreus (A.J.B. Wace b, ; , -), and multiple extensions to them were required to uncover the full extent of the bothros. The first exploration of the ridge (Fig. :) started on  July (Thomas , ), with Trial Trench Z. Further trials were made to follow the line of walls, confirm the construction-related stratigraphy beyond the area of the bothros and later to recover a significant quantity of worked poros blocks, the function of which remained shrouded in mystery until , when it became understood that they were originally part of a wall to retard erosion of the mound over the Treasury of Atreus (A.J.B. Wace , ). These trial trenches demonstrated the existence of Mycenaean remains on the ridge, which were further investigated in  (A.J.B. Wace , ).
Helen Wace's () diary consists of nine unpaginated loose leaves of paper covering the period from  July to  September. It is more akin to a daybook, mostly recording daily life on the excavation. The notes related to her trench supervision activities are limited to the first leaf. On  July two trenches, A and B, were opened in the area known as Sykies ( Fig. :), as it was believed that two visible conglomerate blocks indicated the presence of a tholos tomb. A third trench, C, targeting these two blocks may have been started on the same day.  Digging began outside the Postern Gate (Fig. :) on  July, and a tomb was discovered on  July. Sykies Trench B was abandoned on  July and a new trench, D, opened to the south-east. From this point onwards only progress in Trench C was reported, with the appearance of cross walls, a floor and a possible kiln noted on  July. The workers outside the Postern Gate were moved to a new area adjacent to the tomb on  July. Here the account of these excavations end, as her daughter was ill and it was necessary for Helen Wace to nurse her for two days, from - July. Although the Sykies trials continued on  July, it is unclear whether she resumed her supervision duties for this single day; her entry merely states that she 'returned to dig after being at hotel  days',  and the remainder of her diary does not indicate that she acted as a trench supervisor elsewhere.
With this overview of the  season complete, it is now possible to move onto the analysis of the metal recording strategies used by each trench supervisor.

METAL RECORDING STRATEGIES: FIELD, MUSEUM, PUBLICATION
This section is divided into three parts, each considering a different aspect of the recording of metal objects and their presentation to the wider academic community.
Metal recording in the field Strategies for artefact recording at Mycenae in  can be divided into two separate processes. The first is to enter a discovery into the trench notebook. When such an entry is made, the level of detail can vary quite widely from a simple acknowledgement of presence to a full description, including measurements, photographs and drawings made in the field. The second stage is registration: an object or fragment is deliberately retained for future study, assigned a number  and inventoried. Some object photography also took place at the Nauplion Museum in .  Helen Wace states that on  July the area between the two conglomerate blocks, which is marked 'C' on her sketch plan of the area, was examined, but she does not use the term 'Trench C' until  July, when she notes that work continued there. The director's daybook does not mention Trench C until  July (A.J.B. Wace a, ).  The inherent ambiguity of the word 'dig', meaning both the action of digging and the site of excavation, and her cursory style, typical for a diary, prevents a clear meaning being assigned. She uses the term in its latter sense in the entries for ,  and  August (capitalised for the last two incidences) and the word 'dug' to describe the action of digging in her entry for  July.


These numbers are placed in square brackets when referred to in this text.


Many other considerations have to be taken into account at each stage. An item can only be recorded if it is actually physically found. Although this issue is beyond the purview of the current paper, it is worth bearing in mind that differences between metals can have consequences for the likelihood of retrieval. Patination, which affects all metals in use at Late Bronze Age Mycenae except those high in gold content (Untracht , ), produces a rainbow of matte greens, reds, oranges, whites and blacks that would have been relatively visible against some soils, and almost indistinguishable against others. Corrosion encourages fragmentation, with sheet metal most at risk; therefore the type of object also affects preservation, and this influence is exacerbated by the use of different metals to make different categories of artefacts. As shown in Fig. , the excavation areas were widely spread across Mycenae. The character of the dug deposits varied from denuded architectural remains to thick midden deposits, graves and dump from previous digs. This affected the type of objects likely to be encountered and the possibility of retrieval. The nature of the deposit would have influenced decisions about sieving, which was conducted at the discretion of the trench supervisor. This could have affected the size and quantity of metal objects recovered. At the beginning of each daily entry, each trench supervisor usually listed the number of people in the team, their assignments and, occasionally, the equipment in use, thus allowing changes to the excavation strategy to be assessed in terms of their possible impact on retrieval.
As will be clear from the following analysis, on certain occasions objects were registered but their recovery was not recorded in the notebook. It is evident from the excavation narrative that each supervisor was often in charge of multiple trenches simultaneously, and these were not always physically positioned in such a way that enabled continuous close supervision. Finds made whilst the supervisor was otherwise engaged may have been less likely to enter the notebook. Furthermore, the character of some of the accounts suggests that entries were created after or near to the daily close of work, suggesting the production of a single summary of the day's output rather than a blow-by-blow account of events as they happened.
The decision to retain a metal object for registration was based upon an assessment of its archaeological value. This pruning of the archaeological record appears to have taken place in the field, unlike the more formalised process in place for assessing the sherds, known as papsing, which collected them together and examined them as an assemblage. Preservation probably played a role; excessive corrosion can cause metal objects to crumble when an attempt is made to move them, significantly decreasing their archaeological value and thus increasing the likelihood of discard. Furthermore, a full-time conservator was seldom available (French pers. comm.). That it was known that a badly decayed object was unlikely to receive the stabilising treatment it needed may have also encouraged discard. It is important to note that none of the supervisors ever explicitly refer to their discard or registration policy in their notebook, and this can only be inferred by comparing the latter to the catalogue of small finds.
These notebooks, unlike many of the later examples from the Mycenae Archive, only very rarely contain the excavation numbers given to registered objects. This has made it difficult to match up certain entries, particularly when little description is provided or the objects are no longer available for study. In some cases more than one object was registered under a single number. Whenever this issue has had an impact on analysis it has been clearly highlighted. The notebook lists the discovery of  silver coins and  gold, eight lead, at least three copper alloy and at least  iron objects and fragments; the number of fragments was not always noted. These two areas therefore seem to have produced a good range of different types of metalwork. There are, however, several clear discrepancies between the account given in the notebook and the contents of the small finds catalogue. This is especially noticeable for the metal finds from MYCENAE : A CASE STUDY  Schliemann's Dump. She lists its contents as three pieces of lead, a large-headed iron pin, an iron nail and two arrowheads, one of copper alloy and one of iron (Fisher c, -). The small finds catalogue lists three copper alloy nail fragments, two possible iron nails, an indeterminate number of lead fragments and a copper alloy arrowhead. Thus the copper alloy nails registered under [-] and [-] do not appear in the notebook, and, conversely, the notebook does record an iron arrowhead, which was not subsequently registered. Can one or more contributing motives be identified for these discrepancies?
Although these two areas included a large quantity of dumped material generated by Schliemann's excavations and some backfill and disturbance from investigations by Tsountas, the context of the metal finds does not seem to have played a strong role in determining whether or not they were registered. The majority of the metal objects recorded in the notebook as coming from Schliemann's Dump were registered despite their lack of stratigraphic context. The same seems to be the case for the previously cleared area of the Hellenistic House, although the generalised recording strategy employed here, discussed in more detail below, makes it harder to be as certain on this point. Context did play a role in the method of excavation, as the workmen changed from picks and shovels to knives in order to dig the pits of the Prehistoric Cemetery, with the material from them being sieved as well.  This may have improved the recovery rate for small objects and fragments, such as those from Grave III, although whilst using picks to remove the large quantity of material from Schliemann's dump, the workman were still able to detect very small items such as a bead fragment [-] and the nails mentioned above.
Moreover, indeterminate fragments of copper alloy, gold, iron and lead were all retained for registration even though their original form was no longer recognisable. Therefore these discrepancies cannot be attributed to a straightforward discard of non-identifiable fragments or a preference for the retention of certain metals. In summary, the discard of the iron arrowhead cannot be easily explained by its context, as other objects from the dump were kept. Nor can its discard be easily explained by its material, its form or its preservation status, since other iron artefacts, the copper alloy arrowhead and fragmented or indeterminate pieces from the same context were all kept. It is notable that a silver coin from Trench A SD  (Fisher c, ) was not registered, although the other silver coins found nearby were. It seems that this coin was discovered during the papsing of sherds, and this may account for the fact that it was not registered. However, it was the only metal object mentioned in the papsing notes, and thus this cannot explain the other discrepancies.
It does appear that the type of material did play a limited role in the way that certain objects found in the Prehistoric Cemetery were registered. Out of  registered entries for the Prehistoric Cemetery, only eight have no further contextual information provided.  Two of these were post-Mycenaean pots, one an animal figurine fragment and another a steatite conulus, as well as two iron nails and two groups of indeterminate fragments, one of iron [-] and the other of lead [-]. These latter two, unlike the rest, seem to have been generated by collecting together fragments from multiple locations around the Prehistoric Cemetery.
Iron nails or nail fragments were reported from the main, north extension and south extension trenches (Fisher c, , , ): four in total, although only two appear in the registered finds ([-] and [-]), leaving another two unaccounted for. An iron fragment and a pin were reported from the north-west extension and main trenches respectively (Fisher c, , ). Nothing matching the description of a pin appears in the small finds catalogue. It is thus likely  Sieving is not mentioned for any other type of context. Use of knives and sieves is first mentioned in the entry for  August, after excavation of the pits had started, but her wording 'continued digging with knife, & sifting earth fr[om] Pits , ' makes it clear that this strategy was implemented from the start (Fisher c, ). It is again made explicit for the entry for  August, when Fisher lists the tools in use and adds a sieving woman to the team list (Fisher c, ); she did not usually list the tools in use with the exception of barrows.


This does not include a Middle Helladic (MH) cup from the Prehistoric Cemetery, which is registered in the Mycenae Museum inventory (BE ) but does not have an excavation number. This may indicate that it was registered during a review of the ceramics during a later study season.

STEPHANIE AULSEBROOK
 that these objects, notwithstanding their different find locations, were registered together under a single number.
Lead wire and other fragments were reported from different areas in the south extension and north extension trenches (Fisher c, , , , ). There are no registered lead fragments from the north extension trench but at least three from the south extension trench ([-] and [-]), which is one more than reported in the notebook. Further indeterminable lead fragments are registered as coming from the Hellenistic House [-], but no such finds appear in the notebook. This raises the prospect that the general entry for registered iron objects may have also included fragments not mentioned in the notebook. Unfortunately, it is not possible to clarify this situation any further, as none of these iron and lead fragments are available for study. It is important to stress that no other class of objects was treated in a similar way.
Vronwy Fisher was a prolific drawer of plans and sections, with at least one appearing on almost every other page of her notebook. She sometimes used them to show the position of certain finds, and metal objects were marked on diagrams five times: a hoard of silver coins, the gold from Grave III (twice) and two lead fragments (Fisher c, , , , , ) (Fig. ). In one case, involving lead wire, the find was not mentioned in the text and only appeared on the diagram. Objects were less frequently drawn, and the only metal examples are sketches of one of the conical beehive ornaments found in Grave III and a bead from Grave VIII (Fisher c, , ).
Superficially, Vronwy Fisher's approach to the recording of metal artefacts may be considered somewhat haphazard. These discrepancies between the notebook and small finds catalogue are not limited to metals alone and are in fact quite common. Stone conuli are one example, as the notebook records  specimens but  were registered. Similarly, there are at least  figurine fragments in the notebook (the exact number was not always given) but  registered, and the latter include several examples of furniture figurines, which never appear in the notebook. Of course, she was well aware that the registration process afforded an additional opportunity to record a find and provide details of its location. Although the lack of artefact drawings may indicate that she felt that the recording of objects was a secondary consideration in comparison to architecture and stratigraphy, she did add some lengthy descriptions of certain items, including particulars of dimensions, colour, preservation and findspot (see, e.g., Fisher c, ). The level of detail varied, even between objects of the same type and material. The only metal objects provided with such a full description were the silver coins, a spearhead and gold beads from Grave VIII. The number of metal objects that were certainly discarded were few in number (the iron arrowhead, one silver coin and one copper alloy coin), but it is not possible to see why that decision was taken.


In general though, it seems that she registered almost everything found, including objects not mentioned in the notebook.  On many days there are multiple entries per context, and this habit of updating the notebook throughout the day may have contributed to these oversights. Appearances can, of course, be deceptive, and may simply reflect the constraints inherent in this investigation; without being there, it is impossible to tell what other factors, such as whether she was physically present at the time of the find, played a role in her decisions about recording. Furthermore, given that generally her policy seems to have been to register almost everything, the absence of these three metal objects from the small finds catalogue may be due to misadventure, such as being mislaid or disintegration upon removal.
However, it cannot be assumed that adding together the details from the notebook and the small finds catalogue provide a full picture of the metal finds from this area. There is at least one example that makes such a simple reconciliation impossible. This is a spearhead recovered on the surface of the north extension trench, which in the notebook is described as 'bronze' (Fisher c, ), although the only viable registered candidate is in fact manufactured from iron [-]. Other similar mistakes, especially for the lead and iron fragment groups discussed above, cannot be ruled out. In addition, the discrepancies discussed here prove that the notebook is not a full account in terms of metalwork recovery, and it cannot be taken on faith that Vronwy Fisher ensured that all finds not recorded in the notebook were definitely registered instead.

Michael Fuller and Colin Kraay's notebook 'House of Columns'
At first glance this area (Fig. :) does not seem to have been particularly prolific in metal finds, which is perhaps not surprising given that part of the time spent here was dedicated to the uncovering of areas already excavated by Tsountas in . Only five objects were mentioned in the text, all of which were subsequently registered: three coins, an iron ring and part of a rivet. However, it is certain that more finds of metalwork were made which were not noted in the text. A.J.B. Wace (a, ) reports the find of some melted lead in the Pithos Room that was not mentioned by Michael Fuller at all. Furthermore, three more objects were registered than were mentioned in the text. In two cases ([-] and [-]) there are two objects registered under a single number when only one was mentioned in the notebook.
The entries provided for each day usually consist of one or two short paragraphs with several clear plans and diagrams to convey information about the architectural layout and the position of important in situ features, such as the vessels found in the Pithos Room. The employment of sieves is not mentioned. Very little contextual information was provided for the objects, and in most cases this was usually limited to the room in which they were found, unless there were other circumstances considered particularly noteworthy; such detail is more commonly given for areas left untouched by Tsountas. Twice the entry for a day ends with the line 'no finds of importance' (Fuller and Kraay , , ), meaning it is possible that other metal pieces were found and discarded. All the finds registered by Michael Fuller were identifiable and of copper alloy or iron, whereas the metal artefact that he is known to have ignored was an amorphous lead mass. It is therefore not possible to distinguish whether his metal recording and retention strategy was based upon the type of metal or its level of preservation.
No metal artefacts were mentioned by Colin Kraay for the six days that he was in charge of this area, but this is perhaps to be expected given that half of that time was devoted to the clearance of vegetation. Apart from pottery, Colin Kraay did report the finding of some shell fragments and a piece of painted plaster (Fuller and Kraay , ), but fragments similar to the latter were also noted by Michael Fuller and therefore provide no indication as to Colin Kraay's likely metal recording and retention strategy. Unfortunately, no light can be shed on this by examining the registered finds from Lisa's  Keeping in mind the difficulty in matching notebook entries to the small finds catalogue, there are only two examples of non-metal artefacts mentioned by Vronwy Fisher that were definitely not registered: a loomweight (Fisher c, ) and a chip from an obsidian knife from Grave XI (Fisher c, ). It is possible that the latter became muddled with the large quantity of obsidian from Grave I, especially since she used Arabic rather than Roman numerals for the graves.


House, which was under the direction of Colin Kraay from - August. No metal finds were registered: only eight pots,  terracotta figurine fragments and three stone conuli. In the absence of field notes or any indications in A.J.B. Wace's daybook, it is impossible to distinguish whether the trial excavations at Lisa's House truly yielded no metal finds or whether they were discarded.
It is thus possible that the impression of rarity given by the account in the notebook and the small finds catalogue, with regard to metal finds from the House of Columns, is misleading, but there is no further evidence that can be brought to bear on this issue.

Frank Stubbings' notebook 'Mycenae '
The metal finds from this area (Fig. :) were mainly concentrated in two places: the terrace known as the Prinaria and stratum  of the north terrace. The character of the finds from each location was very different. The latter consisted of mostly copper alloy jewellery dated to the Geometric Period (Fig. ), in a context that otherwise only contained sherds, and the Prinaria finds included three gold pieces of jewellery and many fragments of gold foil, along with a wide variety of other objects including an ivory figurine group [--], plaster head [-] and a faïence cylinder seal [-]. Metal artefacts were also noted from the south terrace and within and around the temple foundations (Fig. ). In total the notebook lists  copper alloy, eight iron and at least  gold and four lead objects and fragments.


The level of detail provided for each object varied; some entries simply stated the material and type, but others include information about appearance and preservation status. Sketches were made of several artefacts, including metal objects (Stubbings , , , , ), and this has played an important role in matching the registered artefacts to those in the notebook.  The majority of the entries for each digging area often amount to little more than two to three sentences, with perhaps an accompanying list of finds. Longer sections usually relate to architectural observations and interpretations. Multiple plans and sections appear throughout the notebook.
Much of the area around the temples had already been cleared by Tsountas; the foundations were also cut by a trial trench made by Schliemann. Moreover, the untouched remains contained very little in situ; even the interesting range of finds from the Prinaria came from elsewhere. These circumstances are reflected in the recording strategy of Frank Stubbings, as he often specified the exact location of a find, sometimes even providing a sketch map, but gave little information about the nature of its surrounding matrix. Indeed, the entries for some days only mention finds, architecture or a description of actions undertaken;  it is possible that the dug deposits were so uniform that the repetition of contextual data was considered superfluous.
Sieving only seems to have been carried out at the Prinaria and is mentioned after the first unusual finds were discovered; it enabled the recovery of several fragments of gold foil and at least one coin (Stubbings , , , -). Surface layers excavated after this point were sieved as well. Several of the gold foil fragments were registered under a single number, despite the fact that they had been found on different days around the Prinaria terrace, some through sieving and others through excavation. However, the exact same treatment was applied to other classes of material in the same area, including bone and ivory inlays [-] and badly preserved fragments of glass ornaments [-]. This may again reflect Frank Stubbings' response to the secondary nature of the deposits in the Prinaria.
Given the number of finds discussed in the notebook it is surprising that a few registered objects do not appear. These three artefacts all came from the north terrace, above stratum , and consist of a fragment of gold foil [-] and two copper alloy discs [-]. Furthermore, although two coins were registered from the Prinaria, only one was mentioned in the notebook, and one more copper alloy ring was registered from the north terrace stratum  group than can be reconciled with the list in the notebook provided by Frank Stubbings. However, such instances can also be found amongst the non-metal finds, such as the amethyst pendant [-] from the Prinaria. There are three more steatite conuli registered than are mentioned in the notebook, whereas the number of terracotta or clay conuli mentioned in the notebook and the small finds catalogue match exactly. The wide-ranging nature of these discrepancies means that they are unlikely to stem from differential valuation by material or object type. Again, it may relate to his apparent lack of interest in recording the detailed context of finds.
Frank Stubbings appears to have changed his retention strategy for metals over time. The amorphous lead fragments he found on  July were retained (Stubbings , ; [-]) but those excavated on  July were discarded (Stubbings , ). This cannot be due to the latter being recovered from Schliemann's trial pit, as a copper alloy ring found the following day in the same trench was registered (Stubbings , ; [-]). Conversely, the first two iron fragments he excavated on  July, which were both from recognisable objects (Stubbings , ), were discarded, but all the subsequently discovered examples were retained. The type of object and context do not seem to have played a role in this decision-making process. Two fragments of copper alloy wire (Stubbings , , ) and a copper alloy fragment (Stubbings , ) were also discarded, perhaps because the original form of the objects from which they were derived was no longer recognisable. Aside from pottery, obsidian was the only other material for which a discard policy was in place. Only one obsidian object, an arrowhead [-], was retained, and all the others (Stubbings , , , , , ), which included recognisable flakes and a blade, were discarded.


The number of objects discussed in this section does not tally with Klein's  publication of this area, as a mistake in the small finds catalogue led her to double count several artefacts.


Examples include  and  July and , , ,  and  August.


In summary, there are a number of important points that can be made about the metal recording strategies that Frank Stubbings used. Three classes of metal finds, copper alloy, iron and lead, were subjected to a discard policy. For iron and lead objects his discard strategy altered over time; no clear reason for this change, such as artefact type, preservation or location, can be discerned. Copper alloy fragments were not retained if their original purpose was uncertain. These strategies differed from those used for any other material. As far as can be ascertained, all the recovered gold objects were registered. Although the physical location of many finds, both metal and non-metal, was well recorded, contextual detail was rarely given. The discrepancies between the notebook and small finds catalogue, as well as his tendency to group together similar finds by material for registration, may also be a response to the secondary nature of much of the dug deposits. More generally, sieving was employed in only one area, noted for its prolific nature and the unusual character of the finds; it did increase the number of metal finds recovered.

Helen Thomas' notebook ' Atreus'
The metal retention strategy in this area ( Fig. :,) was almost entirely based upon the specific material of the objects in question. As far as can be certain, all the gold artefacts were kept and registered, even though they were highly fragmented. Four times the name of each workman who struck gold was recorded in the notebook alongside the entry for its discovery (Thomas , , , ). This practice only appears a further two times: once for the discovery of a small silver 'Aeginetan turtle' (Thomas , ; although the material was not specified, this very recognisable type is always manufactured from silver [Milbank ]), and the second time for the find of a possible copper alloy pendant and fragments (Thomas , ). The coin was not found in the trench and this may explain why the name of the finder was noted, even though it was not registered. The reason for marking the name of the workman who discovered the possible copper alloy pendant and fragments is far from clear, as these objects were discovered during the normal course of excavation and not registered. On the inside cover of the back of the notebook there is a list of the names of some of the workmen with the words 'gold' and 'bronze' written alongside them, which match some of the entries in the notebook elsewhere.
The other metal artefacts were treated very differently. At least one iron and  copper alloy objects and fragments were mentioned in the notebook, but only one of these, a complete pin [-], was registered. Even recognisable pieces, such as a coin and fragments of copper alloy blades, rings, nails, arrowheads and pins, were discarded with no descriptive detail provided. It may well be the case that these items were too poorly preserved to be considered worthy of future study. Many copper alloy fragments were only described as 'scraps', and entries include 'pieces of bronze blades not uncommon' (Thomas , ), 'a little bronze' (Thomas , ) and 'pieces of bronze ring and others' (Thomas , ), which all provide tantalising glimpses of the unusual concentration of copper alloy fragments in this area. The same vague description without enumeration was occasionally employed for terracotta figurines (Thomas , -, , -, -) and once for stone conuli (Thomas , ), but this was not their sole record. A few lead objects were recovered, only some of which were registered. So few details were given that it is not possible to determine the basis upon which this selection was made. On  August a piece of wire, thought to be of silver or bronze, was discovered, but this was not retained for future study.
A clear description of the changing stratigraphy across this area was carefully recorded, but it is not always clear to which layer each object belonged. This issue was especially acute when the trenches were widened, as often the removal of several layers was recorded as a single entry with no attempt made to distinguish the location of finds. The positions of the first few metal finds were shown on a plan (Thomas , ), but such plotting did not occur again except to show the position of stone blocks and, in one case, a pithos. The depth was sometimes given for certain objects, but it is not clear why. It is possible that this reflects the recovery of some artefacts through sieving, although the use of sieves is never referred to explicitly. It is evident that Helen Thomas did not regard the recording of the precise location of every find as a MYCENAE : A CASE STUDY  priority, and this probably stems from the obviously secondary character of the main deposit. She did retain the distinction between the two halves of the bothros, which allowed for a comparison of the sherd material.  There were no finds mentioned by A.J.B. Wace in his daybook that were not noted by Helen Thomas. This fact, coupled with the significant quantity of entries for unregistered metal artefacts, implies that her account of the metal finds is reasonably complete, and thus a reliable guide to the range and quantity of metalwork excavated here. A comparison of all the objects listed in the notebook and those entered into the small finds catalogue is shown in Table . The treatment of copper alloy stands out. Several other materials seem to have been under-registered (gold, iron, lead, obsidian, silver and stone), yet the discrepancy is far less than that seen for copper alloy. In the case of gold this apparently arises from the registering of multiple finds together, rather than deliberate discard. This may simply reflect the density of copper alloy finds, and it is impossible to tell whether if, say, iron artefacts had been found in such quantity they would also have been treated in the same way. The significant discrepancy for ceramic is linked to the multiple figurine fragment entries without enumeration discussed above.
There are categories of finds that are also mentioned in the notebook which were apparently not kept by Helen Thomas at all, such as charcoal and tile. The former does not seem to have been retained by any of the excavators,  and the only registered example of the latter was an Archaic antefix found in the vicinity of the Prehistoric Cemetery [-]; inscribed fragments found around the ruins of the Greek Temples were sketched for posterity (Stubbings , , ). Therefore even in comparison to these types of objects, Helen Thomas' treatment of the copper alloy finds still stands out.
The decision to retain the complete copper alloy pin [-] points to it being considered worthy of exception in some way. Unfortunately the pin itself is now missing, so a direct examination to confirm this is not possible. Nevertheless, there are a few minor indicators that may bear out this hypothesis. It is one of only two metal objects mentioned by Helen Wace in her diary. The other was the gold pomegranate bead from the Prinaria, which she mentioned in a list that also contained the plaster head, faïence cylinder seal and lantern-shaped bead found from the same area. Very few other objects are considered worthy of mention; this select group consists of the ivory figures, a fragment of a large female figurine, another terracotta figurine, a whole pot from the Prehistoric Cemetery, inscribed tiles and the pithoi in the House of Columns. Reference is also made to the copious pottery from the Atreus bothros and Lisa's House. It is apparent that Helen Wace felt that the copper alloy pin was particularly noteworthy, A similar treatment of the metal finds will be attempted by the present author at a later date.  Note that this excavation was conducted before the development of scientific analyses that made the recovery of charcoal worthwhile, including radiocarbon dating, dendrochronology and species identification.


as it warranted its own stand-alone entry. The pin may have been selected for photography (Fig. ), although this cannot now be verified. It was certainly drawn by Piet de Jong in ;  other objects from the  excavation that were later drawn include an unusual figurine [-], the ivory figure group [--], various ceramics (especially those from the Atreus bothros) and four objects recovered from the Prinaria: the faïence cylinder seal and three pieces of gold jewellery. Putting to one side the pottery, the drawing of which is chiefly intended to provide evidence for discussing relative chronology, the other items selected for this treatment are clearly outstanding finds from the  season. Taken altogether the evidence does seem to support this hypothesis, but unless either the pin itself or the Piet de Jong drawing is found, the reason behind this judgement will remain unknown.
Overall, the impression of rarity for metal finds from this area given by the small finds register is highly misleading, as is the apparent predominance of gold over the other metals. However, a comparison of Helen Thomas' notebook with A.J.B. Wace's daybook, as well as the inclusion of  so many unregistered entries, does imply that the former provides a relatively reliable account of the quantity of metal excavated in this area. One likely exception is a possible crucible fragment [-], which is described as containing metal slag;  this object was registered but not mentioned in the notebook, and it may have been found during the course of papsing. The metals were all treated very differently, both to other types of material and to each other. Great diligence was taken over the recording of gold, but the others, with the exception of lead, seemed to have been subjected to a policy of automatic discard regardless of preservation, artefact type or context. Only one item was rescued from this fate: a copper alloy pin that seems to have been judged to be an exceptional find. The decision-making behind the recording strategy for lead is not possible to reconstruct. More generally, the provision of detail concerning the exact context of a find was not prioritised.
Helen Wace's  diary Helen Wace supervised the excavation of the Sykies (Fig. :) and Postern Gate (Fig. :) areas from - July. During the first day of work at the Sykies site, two copper alloy arrowheads were recovered;  their findspots were recorded on a rough sketch plan of the area. These were the only registered objects from either of these areas. The entry for each day usually refers to a small selection of the trenches underway, with one or two sentences of description for each one. It was clearly intended as a summary of the most significant findings, rather than a narrative of excavation. The only other finds considered worthy of mention seemed to have been ceramic (pots, sherds, tiles and loomweights).
The discovery of other unrecorded metal finds is plausible; although A.J.B. Wace does not mention any other metalwork from these trial trenches in his daybook, he does report the finding of a saddle quern associated with the grave outside the Postern Gate (A.J.B. Wace a, ) that Helen Wace did not mention. In her entry for  July, Helen Wace states that there were few surface finds in trial Trench D at the Sykies area, but no further details were given. No mention is made of sieving. With only two examples it is difficult to assess her intended metal recording and retention strategy. Both registered objects were readily identifiable, made of copper alloy and were not recovered in situ. The discovery and discard of other unrecorded metal objects or fragments cannot be ruled out.

Discussion
Before comparing the metal recording and retention strategies discussed above, it is important to also consider the contents of A.J.B. Wace's daybook. Although none of the supervisors had excavated with Wace prior to this  season at Mycenae, it is plausible that the recording strategies of Helen Wace and his three postgraduate students may have been influenced by his ideas. The entries concerning metalwork made by A.J.B. Wace are not necessarily a reflection of verbal reports provided by the supervisors, as he would routinely discuss progress with the workmen directly as well as making his own observations (French pers. comm.).
Wace made  references to metal artefacts in his daybook. Half of these referred to gold finds from the Prehistoric Cemetery, Greek Temples and the Treasury of Atreus bothros (A.J.B. Wace a, , , , , , , ). Another three references were to coins found below Schliemann's dump and on the Acropolis (A.J.B. Wace a, , , ) and two each to amorphous lead masses (A.J.B. Wace a, , ) and copper alloy objects (A.J.B. Wace a, , ). This apparently indicates that the type of metal played a role in his recording decisions, as copper alloy artefacts are significantly under-represented and gold over-represented. Although the gold objects mentioned by Wace included the important jewellery finds, also noted are fragments of foil ornaments, for which it is difficult to see any other reason for their inclusion other than their material. Lead was still considered worthy of mention even when no longer  This has not been verified by a specialist.  One was described as a spearhead in Helen Wace's diary.

STEPHANIE AULSEBROOK
 identifiable as an individual object. Lead has often been undervalued by archaeologists (Mossman , ), although Wace was an exception to this and lead finds were always noted in his publications, even when they had been melted or crushed beyond recognition (e.g., A.J.B. Wace et al. /-/, -, , , , , -, , , , , ). The coin hoard was certainly an important find, especially because of its dating potential. The other coin noted by Wace was from the House of Columns. It is not so clear why he specifically mentioned it; that it was the first object to be recovered from here, combined with its unexpected appearance, given that this structure was previously cleared by Tsountas, may have prompted its inclusion. One of the copper alloy references is to the arrowheads recovered from the Sykies area. Their discovery at the time may have seemed quite significant, given that the main purpose of these trial trenches was to locate a possible tomb. The other copper alloy artefact mentioned was the earring found on the Prinaria terrace; it was mentioned alongside a bone pin and MH pottery, and was probably included to demonstrate the range of material being recovered from this prolific area.
Turning now to the comparison of metal recording and registration practices between the supervisors, it will be clear that this discussion mainly draws upon evidence derived from the study of Vronwy Fisher's, Frank Stubbings' and Helen Thomas' strategies; this is simply a reflection of the quantity of data. There are  important points that must be emphasised: ) The recording of finds was considered to be one of the primary purposes of the notebook, but the entries were not necessarily exhaustive: The recording of find discoveries, both metal and non-metal, plays a prominent role in all the notebook entries. However, the level of detail conveyed varies enormously, not just between supervisors but also between finds, even when they are of the same type or the same material. They can range from a simple observation of presence to include full contextual details, information concerning preservation, material, form and size, as well as pictorial aids such as a plan to show the find location or even a drawing of the object itself. Furthermore, it is not unusual to find objects registered in the small finds catalogue that were not mentioned in the notebook, nor is this restricted to metal artefacts alone. These discrepancies can also vary, from a minor difference in the number of a given artefact to the complete absence of the finds from an entire stratum. There does not seem to be a general tendency to record metal finds at a rate that is higher or lower in comparison to finds manufactured from other materials.
) The level of detail provided in the notebook and the decision to register a find are not correlated: The supervisors did not ensure that more information was given for objects that they did not plan to retain, to compensate for the lost opportunity for future study. Neither did they go into greater detail for the finds that they did choose to keep; as stated above, none of the supervisors explicitly explained which objects they planned to retain or discard, and thus it was not necessary for them to use the notebook to convey the justification for their decisions. Of course it is important to acknowledge that the process of registration gave supervisors the opportunity to provide more information about each find following a more structured system of entry, and the knowledge that this opportunity existed may have contributed to their individual determination of how much information ought to be mentioned in the notebook.
) The level of contextual detail given for an artefact is most closely associated with the type of context from which it was derived, rather than any characteristics of the object itself: Considering contextual information specifically (soil colour and texture, productiveness, features such as ash or mudbrick, stratigraphical position and thickness, etc.), it is evident MYCENAE : A CASE STUDY  that the quantity divulged in the notebook related to the perceived archaeological value of each context. There is no relationship between the level of contextual detail and the material, preservation or form of an object. Many of the deposits dug at Mycenae in  were clearly secondary or had been previously disturbed, meaning that they were judged to be of poor archaeological value. Their description is often relatively vague and incomplete. This awareness seems to have percolated into registration practice, as Vronwy Fisher, Frank Stubbings and Helen Thomas were all content to collect together similar finds from disparate locations within a single context to register together, thus reducing the granularity of the contextual detail available. Generally the supervisors provided more data when excavating primary deposits. Yet still in most cases a specific description of the surrounding matrix appears only infrequently; it is usually necessary to infer such details from a previous entry, and sometimes this is not possible. There was an apparent reluctance to knit these two elements together; thus in general when an artefact was derived from a well-described context its position within said context was not noted, and vice versa. Two examples are particularly instructive. Helen Thomas provided a very detailed stratigraphy for the trenches containing the bothros, yet sometimes only recorded the trench in which a find was recovered despite digging multiple strata in it on a single day. Frank Stubbings' (, ) initial entry for the Prinaria simply states that he was clearing away earth, and, despite including a plan to show the findspots for various artefacts, the first description of the dug deposits does not appear until  August, the penultimate day of excavation, when the level below the Mycenaean remains had already been reached (Stubbings , ). A small number of the graves in the Prehistoric Cemetery provide a strong counterexample to this general pattern, integrating contextual data with details of find location. These are, of course, primary deposits.
) The verbosity of the object description given in the notebook is neither strongly related to the characteristics of the find nor its find context, although metal artefacts are often more vaguely described: The quantity of detail given for each artefact was highly inconsistent, as discussed under point . Not only did registration play no role in this variability, the same is true for any other possible discernable factor. For instance, the high archaeological value placed on gold (see point  below) did not necessarily result in lengthier descriptions; examples include 'piece of gold' (Thomas , ) and '[fragment] of gold leaf ' (Stubbings , ). Vronwy Fisher notes the find of two glass beads on  August, one of which has dimensions, description of form and findspot as well as being drawn; the other just has the colour and findspot noted. It is the latter that was recovered from a grave (Fisher c, ). However, metal artefacts are more likely to be referred to in quite general terms as a 'piece' or 'fragment', often without any further specification. Although it is true that poor preservation can mean that little descriptive value can be provided for metal objects, even basic indicators of form, such as sheet or lump, are informative yet rarely used; it is more usual for a supervisor to either provide no detail or assign an artefact type.
) Discard practices vary considerably between supervisors, although material always appears to be the main criterion and metal artefacts are more prone to discard than any other object, with the exception of ceramic sherds and loomweights: All the supervisors used the process of papsing to discard uninformative sherd material, and terracotta loomweights were never retained. Beyond this, it can be demonstrated that four of the supervisors also applied a discard policy to certain categories of small finds. Vronwy Fisher is one exception, as although not every object mentioned in her notebook was subsequently registered, such instances are exceedingly few in number and do not seem to STEPHANIE AULSEBROOK  follow any pattern; misadventure or oversight seem to be more plausible explanations. The other is Colin Kraay, who did not report the recovery of any small finds in his brief tenure as supervisor for the House of Columns, and therefore it is not possible to tell if he had instituted a more rigorous discard policy.
A.J.B. Wace's mention of the saddle quern from the Postern Gate area that was neither recorded nor registered by Helen Wace indicates that she did discard at least one small find, but there is too little evidence to say anything further. The comparison between A.J.B. Wace's daybook and the notes made by Michael Fuller confirms that the latter did follow a discard policy, as he did not register at least one metal find. However, there are only five non-metal objects mentioned by him in the notebook, four of which were terracotta loomweights; the fifth object was registered. Given the paucity of the evidence it is impossible to determine whether his discard policy was more wide-ranging. Frank Stubbings routinely kept all small finds except for those of obsidian, of which he retained only one, and a small number of metal finds. Helen Thomas discarded metal, obsidian and stone objects; in all cases this practice was not absolute and a proportion of each was retained, although metal was the worst affected.
) The type of metal was the factor most likely to influence recording and retention practices for metal artefacts: Not all metal artefacts were treated in the same way and the predominant axis of differentiation was the type of metal. These differences appear at both the recording and the registration stages, and the way in which they were manifested varied between supervisors; of the four supervisors for which there is sufficient evidence, one seems to have implemented this distinction when recording, two when deciding which objects to retain and one during the registration process itself. Michael Fuller did not record the presence of a lead artefact in his notebook. It is not clear whether this was solely due to its material or whether its preservation played a role in his decision, as all the other metal artefacts recorded by him were not only non-lead but also recognisable objects or fragments of objects. The discard patterns of Frank Stubbings and Helen Thomas are far more clear-cut. Both retained all gold finds but selectively discarded those of copper alloy, iron and lead; Helen Thomas also discarded at least one silver object. However, their patterns of discard were not the same. Vronwy Fisher treated the iron and lead finds differently from all the others by grouping them by trench and registering them together.
) The recording and registration practices for gold artefacts are the most consistent between supervisors: All three supervisors who discovered gold objects seem to have recorded each instance of their appearance and ensured they were all subsequently registered; this was not the case for any other metal and occurred despite the fact that the three supervisors (Vronwy Fisher, Frank Stubbings and Helen Thomas) otherwise pursued rather different discard strategies. A.J.B. Wace also highlighted the discovery of gold more than any other metal. These trends are best explained as reflective of the prevailing contemporary value system.
) Previous experience may underlie differences in metal discard practices: The supervisor who was most familiar with the excavation of prehistoric remains was Vronwy Fisher, and it may be no coincidence that she was the only one to adopt a policy of total retention. In contrast, a significant proportion of Helen Thomas' prior experience came MYCENAE : A CASE STUDY  from the excavation of metal finds from later periods, which predominantly came from wellpreserved, undisturbed primary contexts. Her later publication of the Stavros finds from other types of context follows the same pattern as her treatment of the metals she recovered at Mycenae in ; only a small number were selected based on her level of interest in the object itself. The same can be said for the Acropolis Treasure, because the circumstances of its deposition still remain unclear and the objects themselves are valued primarily for their craftsmanship, material and distinctive character.  Frank Stubbings' change in retention strategy for iron and lead artefacts may reflect his relative lack of experience in excavating metal finds. There is not enough evidence available for the other three supervisors for a judgement on this possibility to be made.

)
There is no evidence to support an exchange of ideas concerning best practice between the supervisors: At least two supervisors, Vronwy Fisher and Helen Thomas, show great consistency in their metal recording practices over the duration of the season. The same is probably true for Michael Fuller, although with less data available it is not possible to be as certain. There is too little evidence for Helen Wace and Colin Kraay to make any definite statement. Vronwy Fisher and Helen Thomas pursued very different discard strategies, yet despite being in daily communication for two months no practical implementation of any exchange of ideas took place. Frank Stubbings' change in strategy for iron and lead artefacts stands in contrast. It has not been possible to reconstruct why he made the decision to start retaining iron objects and discarding lead ones, and given the approaches to registration used by the other supervisors it seems unlikely that this was a result of an exchange of ideas concerning best practice. A further example is provided by group registration of objects by material; although both Vronwy Fisher and Frank Stubbings employed this strategy they did so in two different ways, with the former specifically targetting lead and iron finds, whilst the latter applied it to a comparatively broad range of materials. In contrast, the Sparta excavation team led by Dawkins, discussed above, appear to have exchanged many ideas about practical fieldwork (French , ). Yet, the exchange of knowledge is otherwise known to have taken place between Wace's three postgraduate students. For example, Frank Stubbings helped Vronwy Fisher to develop her pottery analysis skills whilst they were in Rhodes earlier in  (Fisher b, ).
) The implementation of sieving had an effect on the quantity of metal retrieved: Frank Stubbings explicitly states that several small objects were recovered through sieving, and therefore it undoubtedly had an effect on the quantity of metal retrieved. How significant this impact was is difficult to ascertain, especially as even very small metal objects, such as nail fragments, were found whilst using picks. Given that sieving was targeted at contexts that were believed to or had been found to contain small and relatively rare types of material culture, even a small increase in metal retrieval rate could provide a misleading impression of metal distribution.  The Acropolis Treasure consists of a unique group of four gold goblets adorned with dog heads, another gold cup of a more typical 'teacup' form, a small gold lion figurine attached to a curved bar, two gold signet rings, several plain gold and silver rings, granulated gold beads and spirals of gold wire (Thomas /). The exact function or original appearance of several of these objects remains enigmatic, as does its context of deposition. Found by Stamatakis and Drosinos in  just outside Grave Circle A, it has been interpreted by various scholars as the remains of a partially disturbed shaft grave, as plunder from a looted mortuary context and as a hoard (Thomas /, ).


) The date of the recovered metal finds does not seem to have had a bearing on their subsequent recording or retention: Even in areas where the research question was primarily directed towards a specific period, such as the investigation of the Treasury of Atreus and the Greek Temples to find dating evidence for their construction, this does not seem to have had a direct influence on the metal recording and retention policy of any of the supervisors. Modern finds such as a Napoleonic-era coin from the House of Columns (Fuller and Kraay ,  [-]) and a gold and seed pearl brooch from the foundations of the Greek Temples (Stubbings ,  [-]) were recorded in the respective notebook and subsequently registered. The only possible effect is indirect; if a supervisor had a policy of discarding either copper alloy or iron artefacts, then this would differentially impact Bronze and Iron Age strata.
The museum storage strategy for the  season Once the decision had been made to retain an object, it was registered with an excavation number and placed in storage. The register was compiled by the supervisors, although it is clear from Helen Wace's diary that she also played a role in the production of the  catalogue. Ninety-eight excavation numbers were issued for metal objects recovered during the  season, and, as discussed above, it was possible to register more than a single artefact under one excavation number. The objects were then split between the National Archaeological Museum of Athens and the Nauplion Museum. The objects registered under  excavation numbers were sent to Athens for storage; it was common practice at the time to send objects considered to be of especial value or interest to the National Archaeological Museum of Athens for conservation and storage. These were probably all transferred together in a single consignment, as it would have been necessary to arrange for an archaeological guard to accompany them (French pers. comm.). The list of objects sent there in  should, therefore, provide some indication as to A.J.B. Wace's ideas concerning archaeological value and the wider framework within which he operated ( Table ).
Table  reveals that there were several factors involved in this decision. The majority of the metal artefacts sent to the National Archaeological Museum of Athens were from the Prinaria, and this was the case for all the objects recovered from this area, with the exception of the terracotta finds. A similar pattern exists for the finds from the Prehistoric Cemetery. Of the objects that can be traced,  all the non-ceramic finds from the grave pits, except Grave I, were sent to Athens. It is not clear whether the artefacts from Grave I were excluded due to their context or material. The registered small finds from Grave I were an iron nail [-] and obsidian fragments [-], whereas the non-metal finds from the other graves that were sent to Athens consisted solely of glass and precious stone beads. However, A.J.B. Wace (, ) believed that the original contents of this grave had been removed, possibly by Tsountas, as indicated by the iron nail and a sherd of modern china. It is no longer possible to judge which was more likely; a glass and a carnelian bead found in two other disturbed graves were sent to Athens,  and this may be because they were believed to be part of the original grave goods (A.J.B. Wace , -).
Yet the same is not true for the gold foil fragments from the north terrace. Only four of the  other registration numbers given to finds from this area can be traced, and in all these cases they were sent to Nauplion, not Athens. The traceable finds are two copper alloy pins ([-] and [-]), two fragmented iron nails [-] and an iron ring [-]. Since the known finds sent to Nauplion included complete objects but no gold, it seems likely that the decision was probably based on the type of metal, although this conclusion is tentative, as the copper alloy discs and rings from the same area cannot be accounted for.


It is also interesting to consider what was not sent to Athens: none of the finds from the Atreus bothros, not even the gold objects or the remarkable copper alloy pin, were transferred there. It is difficult to understand why one group of gold foil fragments from the north terrace was sent to Athens, yet the similar assemblage of gold from the Atreus bothros stayed in the Argolid. Context may lie behind these decisions, given that the objects from the bothros were all unstratified, but as Frank Stubbings did not mention the gold in his notebook, and the catalogue gives its findspot simply as 'n terrace (upper)', it is not clear whether the former did come from a stratified context.  Given that the main factor determining which objects were sent to Athens seems to be whether they were found in situ and formed part of a coherent assemblage, then the two other puzzling omissions are the hoard of silver coins uncovered by Vronwy Fisher and the copper alloy jewellery from stratum  of the north terrace. It is possible, as discussed above, that material was the criterion that stopped the latter being sent to Athens, and only assemblages containing gold were prioritised. Alternatively, the date of the objects may have been a factor. The artefacts sent to Athens, with the exception of the north terrace gold foil (which seems to have been an anomaly), were all Mycenaean, but the silver coin hoard and the copper alloy jewellery were later in date. In general, the post-Bronze Age finds from Mycenae are rather meagre in comparison with other sites throughout Greece, and, thus, although they are relatively unusual at Mycenae, they may have been considered rather too ordinary to send to Athens.
As the likelihood of an invasion of Greece during the Second World War increased, the finds kept in Athens were packed up and hidden; A.J.B. Wace took part in this process. However, the remainder of the finds had been sent to the Nauplion Museum. During the war its basement, in which many of the finds from the s and  were housed, was used as an air-raid shelter; Wace describes how this led to the destruction of all the human and animal remains held there, as well as the loss of the stratigraphical record because the ceramics were removed from the boxes and trays in which they were classified.  French (, ) has highlighted that this  (nor in Frank Stubbings' notebook), and if he gave an instruction to send all the gold finds from the Greek Temples to Athens, meaning those from the Prinaria, these foil fragments may have been mistakenly sent too.  A.J.B. Wace , viii. The exhibits from the museum were safely hidden in a nearby cave (French , ), but these would not have included the objects excavated in .

STEPHANIE AULSEBROOK
 information must be second-hand, as A.J.B. Wace did not return to the Nauplion Museum until , but confirms from her own personal experience that the sherds were emptied from their storage boxes and trays, and separated from their labels. Neither mention the impact of this on the metal finds. Nevertheless, it is clear that they too were affected.
The process of reorganising the stores of the Nauplion Museum took place during the s. Helen Wace, with the help of Linda Witherill, matched up the remaining objects with the inventory records. In some cases when the objects could not be found or were in very poor condition the excavation number was rescinded; this occurred  times. With the exception of the Napoleonic-era copper alloy coin from the House of Columns, these were all small lead and iron fragments. It is also possible that some of these objects were deregistered immediately after the excavation, after Helen Wace looked over the catalogue.  No objects of any other material were officially cancelled in this way.  The excavation numbers remained on record in the small finds catalogue, with a note of explanation. A second reorganisation took place in the mid-s, when a subset of the objects in the museum's keeping, which were housed on the ground floor of the museum, were moved to the current storage area in the Leonardo building to make way for offices (French , ). The small finds catalogue does not mention whether this affected any of the metal artefacts, and it is also possible that deregistration took place then as well.
Of the  remaining excavation numbers for metal artefacts stored in the Nauplion Museum, the contents of just  of them were located and given new inventory numbers during the process of moving and cataloguing material from Nauplion Museum to Mycenae Museum that took place from  to . In addition, a copper alloy earring [-] from the Prinaria originally sent to Athens is now held by the Mycenae Museum. The current whereabouts of the remaining objects is unknown; they do not appear to have been entered into the inventory system of Nauplion Museum. It is possible that some of these, especially those from the Prinaria terrace, were also sent to Athens and await rediscovery, as has happened for other objects (Papazoglou-Manioudaki , ). At least one of them, the copper pin [-], was mislaid after the war, as it was drawn by Piet de Jong in  but could not be located again in . The same mystery surrounds many other artefacts excavated in , and it is by no means a problem restricted to metalwork alone.
This overview of the museum storage policy for the  finds is especially interesting as it both contradicts and reaffirms several of the trends observed in the recording and registration practices of the supervisors. The type of context from which a find was recovered was already the most important factor with regard to the level of contextual detail provided. It also guided the division of the metal artefacts between the two museums. The discard policy employed by some supervisors and the process of rescinding registration numbers also show a similar pattern; in both cases metal objects were disproportionally affected and particular types of metal singled out. On the other hand, gold was not treated as consistently as during recording and registration, because the gold finds were divided between Athens and Nauplion. Furthermore, there is a possibility that the date of the artefacts may have played a role in choosing which museum they were to be sent to, which was not used as a criterion during recording and registration.
The devastating effect that the Second World War had on the  finds and the repercussions for interpretation are clear. However, the war had a similar impact on study collections held in museums across the globe, and this must be factored into our use and interpretation of all material excavated prior to its outbreak. It is important to note that not all contexts were equally affected; it has had the most deleterious effect on object assemblages that are only partially preserved, as research of the surviving artefacts is hindered now that they can no longer be considered together with the rest of the original group.  French pers. comm. As discussed above, Helen Wace does not seem to have intervened in the selection of artefacts for registration prior to their entry into the small finds catalogue.
 The excavation number [-], which referred to beads, was cancelled, but unlike the metal finds they were re-registered under a different excavation number [-]. A second more detailed report of the  season was published in , in the Annual of the British School of Athens (A.J.B. Wace ), as the first in a series of preliminary reports that would cover the Helleno-British excavations of  and -. The majority of the text is dedicated to the finds made in the Prehistoric Cemetery. According to Wace, the first draft was produced during  using his own notes and those of Vronwy Fisher, and was subsequently revised and added to in  by Sinclair Hood, Wace himself, Helen Wace and Frank Stubbings. At the time that this final draft was produced, the objects then stored at the National Archaeological Museum of Athens were not available for examination, and those in the Nauplion Museum were in disarray, as described above, although Sinclair Hood was able to examine at least some of the pottery (A.J.B. Wace , -).
Further publications were produced of selected material. The architectural findings from the exploration of the Treasury of Atreus are discussed in the light of further discoveries along the Atreus Ridge in , but the metal finds are not mentioned (A.J.B. Wace , -). The Atreus bothros was published by French in ; the core of this paper was the important information that the ceramics and terracotta figurine fragments provided concerning the definition of the ceramic period LH IIIA. The presence of gold and copper alloy artefacts within the deposit, as well as other non-ceramic objects, is acknowledged but not discussed in detail (French , ). The lead artefacts are not mentioned.
Some The other gold objects recovered and registered from the Prinaria, all fragments of foil ornaments, and a copper alloy earring from the same area were not published. Some of the gold foil pieces were recovered through sieving, but it is not clear whether this lay behind the choice to omit them; other excluded examples were apparently found whilst digging. Although the paper may give the impression that the objects discussed were all found together, a sketch plan provided by Frank Stubbings in his notebook clearly shows that they were spread across at least two separate locations. Therefore proximity to the faïence cylinder seal does not seem to have acted as a deciding criterion either. Nor can this selection process be linked to the interpretation of the published finds as the contents and remnants of a small wooden box, as the omitted objects could also be considered in a similar vein (A.J.B. Wace and Porada , ). The most likely explanation is that the unpublished artefacts were regarded as of lesser importance for making the argument about the possible ritual associations of the finds from the Prinaria. The ivory figurine from the Prinaria was published by Helen Wace in .
The coins recovered during the  excavation season were published by Dengate (), as part of an article that included all the Greek coins excavated between  and . Dengate (,  n. ) states that  Greek coins,  silver and one bronze, were discovered in , but this does not tally with the small finds catalogue, which lists  silver coins. The two missing STEPHANIE AULSEBROOK  entries, [-] and [-], both came from the area of the Greek Temples, and according to the catalogue they may have been sent to the National Archaeological Museum of Athens, which possibly explains their omission. According to Dengate (,  n. ), the coins were examined by Josephine Shear,  which has meant that, although two other silver coins could not be located after the Second World War, their details could still be included in Dengate's catalogue. Dengate specifically mentions the Napoleonic-era coin [-], the inventory number of which was subsequently rescinded (Dengate ,  n. ), and briefly discusses the hoard of silver coins found in the Hellenistic house below Schliemann's dump (Dengate , -). He does not make any reference to the existence of five coins that were recorded in the notebooks of Vronwy Fisher and Helen Thomas but not kept, and he may not have been made aware of their existence. Although small in number, the finds from the Atreus Ridge negate his statement that all the coins discovered came from the area within the Hellenistic circuit walls (Dengate , ).
An account of the  excavation of the area of the Greek Temples was published in  by Klein, who had access to both the objects and field notes, and who was also able to discuss the  season with the excavator, Frank Stubbings (Klein ,  n. ). Her paper concentrates on the finds that were of direct relevance to the Greek Temples and details both the registered and unregistered artefacts by area,  as well as providing an appendix listing all the relevant small finds (Klein , -). It is worth noting that this appendix only includes objects that were registered, although this is not explicitly stated. In many cases Klein was able to match the registered finds to entries in the field notebook, although a mistake in the small finds catalogue led to an instance of double counting. The importance of the north terrace and the Prinaria find groups is stressed.
Alden () published all the results from the Prehistoric Cemetery in a single volume as part of the 'Well Built Mycenae' series. The main findings from the trial trenches outside the Postern Gate and Areas  and  are summarised, and images of the graves and the pottery included, as well as a plan of the area of the Prehistoric Cemetery outside the fortification wall (Alden , , , , -, , fig. , pls , b, b, , a). The  graves are then republished in full in the accompanying material, including the metal finds, which are illustrated when possible (Alden , -). Alden follows Wace's interpretation that the iron nail in Grave I was from Tsountas' clearance of this grave (Alden , ). The publication does not discuss the other metal finds that were found scattered around Areas  and .
There has never been a single publication drawing together the excavation narrative for the  season that integrates all the recorded data. The diverse range of dug deposits may be thought to discourage such an approach, although this was successfully achieved for the similarly structured s campaign at Mycenae. However, the intervention of the Second World War effectively made a comprehensive account impossible to realise, as outlined above. Indeed, the delay in publication as a direct result of the war meant that out of the six supervisors only Frank Stubbings and Helen Wace contributed to the accounts that appeared in print. From the preliminary reports onwards it is understandable that artefacts discovered in situ or considered particularly noteworthy have been highlighted in publications; in terms of metal artefacts this has led to a concentration on the gold objects and the kept assemblages from the Prehistoric Cemetery and the Prinaria and north terraces of the Greek Temples excavation.
Through these publications the archaeological record as revealed in  has been effectively pruned further in two different ways. It seems certain from his article that Dengate was not made aware of unregistered finds; thus his absolute statements have to be treated with some caution as they are based on partial data. The publications by Klein and Alden, which carefully integrate evidence from all available sources, are therefore more reliable in that regard, as well as more informative. The other method of pruning comes from the selective use of material to support a specific line of archaeological reasoning, as seen in the Prinaria article (A.J.B. Wace


Exactly when this happened is not clear.  A fragment of lead from the south terrace is missing from her account, but as she mentions the similarly unregistered lead fragments from Schliemann's trial pit, this appears to be an oversight rather than a deliberate omission.
MYCENAE : A CASE STUDY  and Porada ) and the publication of the pottery from the Atreus bothros (French ). This is a perfectly legitimate and reasonable use of data, and indeed forms an essential part of analysis that enables the incorporation of such information into wider discussions. However, in the absence of a comprehensive publication of an area or context, it does act to further prune the archaeological record, even when this is not intended.

CONCLUSIONS
The purpose of this paper was to utilise the  excavation season at Mycenae as a case study to investigate and compare metal recording practices used on early digs, and the effect that these have had on our present-day understanding of the results. It has done so by considering all possible elements of relevance, such as the organisation of the team and their prior experience, and then effectively tracing the biography of the excavation itself, from the initial act of digging, through to the processes of recording and finds registration, and finally to storage decisions and publication strategy. By choosing to examine an excavation conducted as a 'round table', it has been possible not only to conclusively demonstrate that these choices made at the trench side have had significant repercussions on our interpretation of the  season, but also to assess the differential impact of divergent approaches. The aim was not to pass judgement but to raise awareness of these issues and their complex nature; multiple metal recording and retention strategies have been identified through this research, and the differing implications of each need to be understood if we wish to successfully integrate data from this and other similarly early excavations.
Beginning with the digging process itself, it is of course impossible to determine the success rate for the retrieval of metal finds during early excavations. The workmen were able to recover very small-scale finds even when using picks, but the employment of sieves did demonstrably increase the retrieval rate. Sieves were selectively used in areas where it was believed that small archaeologically valuable objects may be present or had already been found. Such contexts should be expected to be over-represented in any analysis of the spatial distribution of metals. Although further investigation into retrieval rates was beyond the scope of this paper, any assessment of these should consider matrix colour as an important factor, due to the varying corrosion products produced by metals.
With each supervisor responsible for their own recording strategy, more variation in the type of data collected and in the consistency in note-taking, compared to a modern excavation, is to be expected. The concept of using their notebook to record the discovery of finds was shared by all the supervisors, but the exact nature of how they chose to implement this had the greatest effect on metal artefacts, because of the application of selective discard policies. Entries for metals were more likely to be framed non-descriptively, using expressions such as 'fragment' or 'piece'.
With the extent of location information given dependent on context, and no relationship between the level of detail disclosed and the decision to register a find, metal objects not found in situ are generally the most poorly represented in terms of data quantity. This problem is compounded by the way in which selective discard strategies were employed for metal artefacts. Discard is not, in itself, a problematic practice, and it is in fact a cornerstone of good finds management, particularly for sites that yield large ceramic assemblages yet where storage is limited, as exemplified by the papsing process in use at Mycenae in . However, when performed well, the papsing of sherds has two main objectives: to filter out sherds of low archaeological value in order to facilitate greater efficiency of storage and to keep a representative sample from each assemblage for future study. The discard policies in place in  for metals did not attempt to retain a representative sample for future study. The metal artefacts were assessed individually for archaeological value and not as an assemblage. Thus, when Helen Thomas began to unearth a large number of copper alloy artefacts of similar type, each was judged on its individual archaeological merits and her strategy of discard remained unchanged, despite the fact that it would have been equally reasonable to consider them together as a single assemblage, just as the pottery was treated.


In general, the base metals (copper alloy, lead and iron) were more at risk of discard. The only metal that was apparently never treated this way was gold. However, silver was not always immune despite the fact that, with the exception of the shaft graves, it is relatively rare at Mycenae (Kelder , ), and therefore it cannot be assumed that discard strategies invariably reflect the distinction between precious and non-precious metals. Thus the likelihood of disproportional representation of gold specifically needs to be borne in mind when handling data from early excavations. That this practice of discard could also take place in the museum, by rescinding allocated registration numbers, must be taken into account as well.
Only one supervisor appears to have committed themselves to retaining all metalwork, and this may be related to their previous experience of excavating prehistoric remains. There was not enough data available to confirm this point, but the possibility of such a relationship indicates that the type of prior fieldwork experienced by participants ought to be considered when analysing material from early excavations. This is especially important given that there seems to have been no practical implementation of any exchange of ideas concerning best practice between the supervisors whilst the excavation was underway.
The Second World War had a profound impact on the  season in a number of different ways. The most obvious of these was the loss of material and information from the use of the Nauplion Museum basement as an air raid shelter. The upheaval caused by the war has meant that any further records kept by other participants are unlikely to come to light. Yet there are other less visible effects that have, perhaps, had equal or even greater consequences on our understanding of the achievements of this excavation season. The war effectively recast the  season as a stand-alone dig rather than the beginning of a renewed campaign, as originally intended. Not only was research on the finds and in-depth publication of the results postponed, by the time it was possible to restart these studies there had been a change in personnel. Of the six supervisors only Helen Wace and Frank Stubbings contributed to the post-war publications of the  season. A.J.B. Wace used the contents of the notebooks as well as his own daybook, but further input from the other supervisors is not mentioned, and it must be presumed that they made no contribution to the publications.
In the absence of a comprehensive publication of the  excavation season, the reporting of the recovered metal finds has been left incomplete. Those found outside of the main deposits of interest have been especially neglected, but the pruning process inherent in the construction of archaeological arguments has also affected artefacts that were otherwise treated as valuable during the stages of recording and/or registration. This issue highlights the importance of returning to the original field notes, as it can be observed that the impression generated by them differs from that given by the catalogued finds, which in turn also differs from the picture presented by the publications.
In terms of understanding metals at Mycenae, the most profound implications relate to the Atreus bothros. Although it seems that a record of every metal find from the bothros was kept, only a select few objects were chosen for registration. This rigorous discard policy was not implemented without justification. Based on the prior experience of the trench supervisor, Helen Thomas, these discarded artefacts were of low archaeological value: many were fragmented or otherwise poorly preserved, they were not found in situ, and in quantity they were reasonably comparable with what she had seen at Stavros. The greatest importance of the Atreus bothros remains, without doubt, its pottery assemblage and the fact that the dromos of the Treasury of Atreus cut right through it, thus laying to rest one of the most contentious arguments concerning the relationship between Crete and the Greek mainland and establishing the chronology of the monuments at Mycenae. The preliminary reports and other publications written in the immediate aftermath of the excavation clearly reflect this. The significance of the pottery assemblage for chronological refinement was realised slightly later and resulted in its detailed publication (French ). The metal finds made no contribution to this study, and their limited appearance in this paper is understandable.
Without any deliberate intention, the metal artefacts from the Atreus bothros have been effectively written out of the archaeological record. The seriousness of this situation can only be appreciated when the bothros is compared against the rest of the settlement remains at Mycenae; MYCENAE : A CASE STUDY  it represents one of the densest concentrations of metal finds outside the citadel, alongside two of the five recognised 'hoards' at Mycenae (Schliemann , -; Stubbings a). Although, when placed in this wider context, its treatment may seem extraordinary, it must be remembered that in  very little was known about the settlement itself. The major excavations outside the citadel had not yet taken place, and the clearances carried out by Tsountas within the boundaries of the fortification wall were unpublished; no-one could yet fully appreciate the relative paucity of metal finds from the habitation areas of Mycenae, nor the potential importance of the Atreus bothros for understanding wider questions pertaining to metal use at the site.
Of course ultimately what this paper provides is just a snapshot of a single excavation season. Its findings are not necessarily applicable to all early digs, nor is it possible through a single study alone to exhaust every possible permutation of metal recording, retention and publication strategies, and their consequences. Nevertheless, the nature of the  excavation season at Mycenae presents an opportunity to understand at least some of the diversity of strategies employed during early fieldwork. Particularly within the Aegean, it is not possible to shy away from the use of the information generated by early excavations, especially as studies based on 'big data' are becoming more prevalent in archaeology. By removing broad assumptions about the way that metalwork was treated by these early excavators from our interpretations and replacing them with concrete examples of the actual sources of potential bias, grounded in the analysis of their working practices, it is possible to integrate data from such fieldwork with more confidence.