Ashmolean Eastern Art Online, Yousef Jameel Centre for Islamic and Asian Art

The Department of Eastern Art in the Ashmolean Museum, Oxford, holds an outstanding, even unique, collection of early Indian and Islamic textiles. The collection was brought together by one man, the Egyptologist Percy E. Newberry, during his extensive stays in Egypt between 1900 and 1930. He donated the entire collection, consisting of 2,240 textiles, to the Ashmolean Museum in 1946 [1]. All of the textiles were already in a fragmentary state when first collected, and were believed to be mostly medieval in date. As textiles only survive under specific, atmospherically stable conditions, the Egyptian finds were recognized as important, even in their fragmentary condition. The collection falls into two groups: the first contains resist- and mordant-dyed textiles, mostly block-printed and generally of Indian origin, and the second group has embroideries that were produced in the Near East, possibly Egypt itself.

This publication restricts itself to the first group, made up of the printed textiles. It presents a comprehensive survey of this entire part of the collection, with detailed catalogue entries and illustrations for each piece [2]. In addition, the collection is discussed and analysed in the light of recent research. A few words may help to illustrate why it was decided to undertake such a detailed study of this collection, in particular. To start with, we face a specific problem: from early historical times onwards we have literary and documentary references that India produced fabrics of outstanding quality and beauty, much in demand at home and in other parts of Asia, but not more than a few threads still exist as evidence of this early glory. Substantial numbers of Indian textiles only begin to survive in a late medieval, primarily Islamic context, often historically connected with the Egyptian Ayyubid and Mamluk period and its trade. The Newberry Collection of printed textiles belongs to this type; while material like it is not uncommon in museums and private collections, here the size of the collection and variety of ornament is unique.

An introduction to Indo-Egyptian textiles

Indian textiles apparently have been a major export article from the subcontinent for several thousand years. The production of fabrics was already well documented by writers of Mediterranean antiquity, and references to the cloth trade can be found in Chinese, South-East Asian, and Islamic sources, all of which predate the arrival of Europeans as participants in the maritime Asian economy by centuries. However, while text documents are revealing as historical evidence, they cannot supply any visual impression. When one finds a reference in Pliny’s Natural History about the transparent nature of Indian muslins, or reads in eleventh-century Javanese records about the sumptuous quality of certain types of cloth possibly brought from India, one cannot have a mental image that reflects anything actually seen from that distant past. Unlike stone and metal sculpture, textiles do not generally survive.

Yet it is certain that the making of cloth was a major industry on the one hand, and the source for artistic invention and creativity, on the other. The silk industry of China and the manufacturers of cotton goods in India were the suppliers for a pan-Asian market that extended into Mediterranean societies certainly by the time of classical antiquity. The combination of technical skill and aesthetic confidence assured them a market wherever their cloth was taken. With the fabrics travelled their patterns: there can be no doubt that the transmission of textiles between societies was a primary source also for the transmission of designs. Immensely portable and, in the short term at least, not as fragile as glass and ceramics, textiles of a certain quality had an international clientele. Their presence as grave gifts and in treasuries attests to the value given to them, and occasionally their survival in archaeological sites can be evidence for their function in daily life. Textiles from India in particular seem to have catered both for a luxury market and for general distribution and common consumption.

In the history of textile research, the name of Fustat (Old Cairo) in Egypt has gained a reputation as the likely source of early Islamic and Indian textiles. In particular certain block-printed fabrics have been referred to in numerous publications as the earliest surviving group of Indian textiles. Although this is strictly speaking an exaggeration, it does remain true that the textiles found in Egypt at present are the foremost surviving examples of the pre-European Indian fabric trade.

The textiles first came to light early in this century; most of them seem to have appeared in the markets of Cairo in the 1920s. Their Indian identity was already considered at the time, but the first serious, full-scale discussion was undertaken by the textile scholar R. Pfister (1938). On stylistic grounds he related the fragments he studied to north-western India, in particular to Gujarat, and he established a chronological sequence by linking the textiles to architectural designs. The dates he suggested span four hundred years, from 1200 CE to 1600 CE. He also considered briefly the evidence of the medieval Indian textile trade, as still observed by the first European eyewitnesses who travelled to the ports of Asia in the early sixteenth century. One major problem for Pfister’s discussion was the uncertain provenance of the textiles he studied. All of them were in established collections by the time he looked at them, most of them privately owned. None of the fragments had come from properly excavated circumstances, and their association with Fustat was tenuous at best.

Although Pfister’s identification remains valid in many respects, it is time after nearly sixty years to reassess his views, in particular his method of dating the material. The size and quality of the Newberry Collection provide a unique opportunity to undertake further research. The result does not inevitably dispute Pfister’s findings, but it has been useful to step over old ground again, to look at hitherto presumed facts, and try to see the evidence with a fresh eye. In the event, the catalogue still depends on Pfister’s work, as it provides the major published comparative material. Regarding the north-west Indian origin of many of the fragments he discussed, there seems to be little reason to doubt the core of his argument, although our present understanding of trade history gives a more elaborate and expanded view. The recent finds at some medieval Egyptian sites, of related textiles but now found in an observed context, to some degree also support Pfister’s dates, although at the moment firm archaeological evidence tends to come from the later Mamluk period, historically close to the initial European involvement with Indian Ocean trade. The Indo-Egyptian textile fragments are at present still the most significant group of early examples of one of India’s main economic and artistic achievements. Although these fragments were not part of a luxury trade in the western Indian Ocean, but apparently were for general consumption, the skill of production was often good. Even where the block-printing is done with a certain lack of care, and the link between block impressions is clearly visible, one generally has to admire the quality of the dyes. Although the fabrics were used and reused until discarded as rags, their designs may nevertheless still impress us by their clarity and the intensity of their colours [EA1990.140, EA1990.1129, and EA1990.161].

The collection

With its total of 1,226 fabrics, the Newberry Collection is outstanding in size and remarkable for its variety [3]. Although similar material is found in many museums, the textile fragments are usually numbered by the dozen. For example, the Victoria and Albert Museum in London, otherwise Britain’s foremost study collection for Indian textiles, only has 19 fragments. Dumbarton Oaks in Washington has a similar number, as does the Brooklyn Museum in New York. Further small collections are in Gothenburg, Copenhagen, and Berlin. In India, the Calico Museum in Ahmedabad has only a few fragments, which have been published by John Irwin and Margaret Hall (1971). The Islamic Museum in Cairo, otherwise rich in textiles, also claims to have only about two dozen fragments. Pfister’s own collection is now in the Vatican; it was recently redis-covered and published by Georgette Cornu (1992). She included 26 printed textiles, of which only 14 are possibly Indian. Representative of a medium-sized collection is the one in the Kelsey Museum at the University of Michigan, which has 56 textiles [4]. The Textile Museum, Washington, holds just over 100 items, and a similar group seems to be in the textile study collec¬tion at the Royal Ontario Museum in Toronto. Another collection in this range, partly published by Pfister (1938), is in the Musée Cluny in Paris. Even the large collections in the Museum für Völkerkunde in Basel, Switzerland, and the Benaki Museum, Athens, each come to approxi¬mately only 300 fragments, although the Benaki collection is particularly interesting and varied in its designs. The scope of the Ashmolean Museum’s collection provides a singular possibility for comparative analysis of the material. While any comments about the representative character of the material remain hypothetical if one depends on 20, or even 100, fragments, 1,200 fabric pieces give a different range altogether.

P. E. Newberry kept all of his textile fragments in large folio books, arranged by design and technique. The fragments were fixed to cardboard sheets, each measuring 43 cm. x 35.5 cm, and often with more than one fragment glued to a sheet. The collection had come to the Ashmolean Museum in the order of the original folios, and they were stored in the same sequence in large boxes. In 1956, a typescript catalogue of the contents was made by Miss A. G. Parke, following an earlier card index compiled by P. L. Shinnie. This inventory described all items, with reference to the original box numbers and folio pages. However no accession numbers were given to the tex-tiles; each entry only described the fragments that adhered to a particular card [5].

The major part of the collection, the embroideries and printed textiles, was eventually trans-ferred to the Department of Eastern Art. The Department of Antiquities kept 36 textiles, all of which were apparently Byzantine Egyptian or Coptic. At this stage at the latest, the original sequence was no longer kept. Eventually, most of the embroideries were accessioned, but the Indo-Egyptian material remained untouched.6 The fragments remained all glued to sheets of cardboard backing, as Newberry had kept them, but no longer in the sequence of the original fold-ers. This made it impossible to use the collection, as the fragments could not be referred to prop-erly. Furthermore, their state of preservation required a major conservation project, as well as plans for appropriate storage. Thanks to funding by the J. Paul Getty Trust in 1990, forty-four years after the donation had been completed, the task of conservation and cataloguing of the printed textiles could finally start.

Preparations for the catalogue: accession and conservation

The project divided into three parts: accession, conservation, and the writing of the catalogue. First the entire collection of over 1,200 textiles had to be individually numbered and accessioned, preferably in a sequence that made sense for the collection as a whole. Newberry himself had followed Pfister’s (1938) chronological divisions. These now seemed unreliable, and the sequence was abandoned [7]. Faced with the daunting task of putting order into an overwhelming collection, I decided first to divide the material by colour, and then to link similar designs within these divisions. As the colours of the textiles represent different dye techniques, one can argue for a rationale behind this approach. The collection divides into three groups, of which the second has been subdivided:

I.        blue and white textiles;

II.       (a) red and white textiles with a single mordant application resulting in just one red tone, (b) red and white textiles which had several mordants applied and therefore produced a variety of shades;

III.      red, blue, and white textiles which combine mordant dyes (producing red) with blue [8].

Each fragment was measured, accessioned, and a record photograph was taken. The fragments, still attached to their original backing, were then transferred to the conservation laboratory. There each textile was steamed off the cardboard backing. As far as possible, the animal glue that had been used to stick the fragments to the acidic paper was also removed, using distilled water. Each textile was steamed again to relax the fibres, and it was then pressed under glass [9]. The fibre of the uncreased fragment was analysed, and a thread-count was done. If remains of stitching were discovered, the fibres were microscopically identified [10].

A separate folder was then prepared for each fragment. The folder itself was cut from acid-free cardboard, and an internal frame of the same material was added. As it is now made up, this frame contains a single layer of vilene (pure spun polyester). The textile fragment rests on this backing without additional attachment, so that its reverse side can easily be inspected. The vilene has enough traction to hold the textile in place, yet by tapping the back of the closed folder and then lifting it, the fragment is released and will lie on the inside of the folder’s front, with its reverse side up. The conservator thus has developed a simple system for making it possible to view the textile from both sides, without having to touch the fragile material. Furthermore, by not being fastened to any surface, there has been minimal interference with the structure of the textile. A numbered label was sewn to each fragment.

The catalogue

For the preparation of the catalogue, each fragment was given a separate entry, even where two textiles were obviously formerly part of the same fabric. In general the catalogue as it now stands follows the sequence of accession numbers. It has not always been possible to keep to this, though, as with such a large collection it was inevitable that one would come across fragments that were certainly related in design and technique, but had been given widely different accession numbers. Furthermore, two additional fragments were discovered in the Department’s textile store after the rest of the collection had been accessioned. As the catalogue divides into three categories, each representative of a particular technical method, I have decided in most cases to move the textile into the appropriate category, but otherwise to rely on cross-reference.

The information found in each catalogue entry is mostly self-explanatory, but a few additional comments may be appropriate here, to remind the reader of certain conventions used. For this publication, every textile has a catalogue number.

Most of the textiles have been block-printed with either a resist or a mordant. Quite often there is a combination of both resist- and mordant-printing. The identification of the dye technique is without great problems for the resist-dyed blue textiles. For the methods used in the application of mordants, much has to be guessed by observation. The dye saturation on the reverse side of the fabric is the main source of information, and therefore this technical description is usually includ¬ed in the paragraph following the description of design. While the initial description presents a fact (the appearance of the pattern), the second paragraph of each entry gives room for speculation. It contains straightforward additional observations, such as on technical peculiarities, but where appropriate it also includes interpretations that are those of the author. These may be changed by future scholarship.

Any art historian looking at the catalogue will immediately notice that there is no separate entry for dating each textile. The chronology of the Indo-Egyptian textiles is still far from firmly established, and it is not appropriate to mislead the reader with a false certainty. The problems of dating, as well as the facts we have, are therefore discussed in the text rather than included in the catalogue. I have tried to make the information that has gone into the catalogue as factual as possible. Of course the entries do make comparisons, where appropriate, to archaeological finds that have a reliable chronology; of foremost importance here are the Quseir al-Qadim textiles excavated by Donald Whitcomb and published by Gillian Vogelsang-Eastwood (Whitcomb and Johnson 1979, 1982; Vogelsang-Eastwood 1990).

For the purpose of dye analysis, samples were taken from 34 textiles. They were analysed by George Taylor at the Textile Research Laboratory in York. The results have been incorporated into the respective entries. I regret that a more comprehensive dye analysis, including an analysis of mordants used, could not yet be carried out on the collection. At some time in the future, an extensive analysis of the collection may be an attractive research project for someone with a particular interest in natural dyes and mordants. We have been able to submit 15 of the fragments to radiocarbon dating, and the results are discussed in Chapter 4 and are reproduced in the Appendix. They also are included in the relevant catalogue entries. For reference purposes outside the catalogue, each entry has a departmental accession number, which consists of the year of accession and a number indicating the place of sequence the object takes in that year’s departmental acquisitions. The accessioning began on 1 February 1990, and by that date eight objects had been acquired by the Department of Eastern Art for 1990, so that the first textile has the number EA1990.9. Following from there, a continuous sequence of accession numbers is used.

When giving the measurements of a textile, it is convention to record the warp before the weft. This has been followed here where possible. However, because warp and weft have been combined in a balanced weave, the position of either one can only be identified in these textiles when a selvedge is present, so that many of the fragments record the measurements of two sides at right angles, following the two thread systems, but without reference to warp and weft. The maximum measurement is given first and is referred to as length, the lesser measurement follows and is called width. All measurements are in centimetres and represent the maximum extent of the fabric. The thread-count is per square centimetre; it is abbreviated to TC in each entry. Again the convention of the thread-count is to give the number of warp threads before the weft threads, and this is followed here where possible. Otherwise the thread-count is for length followed by width. The twist direction of the spin also follows the sequence of warp before weft, length before width.

Sometimes catalogues of textiles also record the thread diameter and the thread ply. Representative checks on the thread diameter indicate that the collection does not significantly differ from the results published recently for the collection in the Kelsey Museum at the University of Michigan (Barnes 1993). The thread used throughout is single (i.e. not plied) for both the weaving threads and for the threads used in stitching, unless otherwise stated. Therefore for reasons of economy this information was not added to each separate entry, but is summarized in the relevant technical discussion of Chapter 5.

A few of our textiles have been published previously; this is indicated in the appropriate entry. Where a textile is very similar to another one published elsewhere, this is separately referred to towards the end of the entry. Finally, the catalogue publication is based on a database file that is in the Department of Eastern Art. This file has, of course, additional information, mainly relating to the housekeeping aspects of looking after a museum collection. These are generally of no rel-evance to the publication. They also record, however, the old box and card numbers, so that one can refer back to Newberry’s occasional pencil marks and suggested dates. These markings use Pfister’s published examples for dating and represent no independent research. While they did not seem relevant to the preparation of this catalogue, for future research one may wish to return to the original sequence. The information is available through the Department of Eastern Art.

The approach to analysis

Finally it is appropriate to add a few words about the structure of the text volume that accompanies the catalogue. Writing about a collection of this size has been at least as daunting a project as preparing a catalogue of it. The wealth of material provides a major challenge in itself, but in addition it was clear from the outset that because of its sheer size, a publication of the collection must also set certain standards for further research on historical textiles. After considerable deliberation and consideration over the sequence of topics to be addressed [11], it finally seemed best to let the material itself take the lead. This is primarily a study of historical Indian printed textiles, with the Newberry Collection as its focus. It is not a study of Indian or Islamic trade, nor is it an analysis of textiles as social commodities, although these aspects are touched upon.

A note on the spelling used throughout is also required. As the material is treated in relationship to the Islamic world, to India, and to South-East Asia, and therefore may appeal to a readership of widely different expertise and interests, it was decided not to use diacritic marks. Place-names are spelled as they appear in The Times Atlas of the World, except where international conventions have recently changed. The specialist in any of the geographic areas and languages referred to may find this irritating, but generally it makes for easier reading.

We know from the historical sources that the Indian textiles came to Egypt on the maritime route via southern Arabia. The survey of the Indo-Egyptian trade therefore focuses exclusively on this link, to the exclusion of connections that existed between India and other parts of western Asia and the Near East. Also excluded for our purposes is the importance of the Indian trade to Armenia and the Caucasus. There can be no doubt that these relationships were important for the history of textile production in many regions, in particular in Syria and Iran, but I do not feel competent at present to extend the study in that direction.

It was decided early on that the catalogue publication should include the entire collection, even where there is repetition, or the origin of the textile is in doubt. Most of the fragments are Indian, but there also are some that are Mamluk or Ottoman. Others could even be imports from Europe. To some degree the collection may be representative of the printed cloths available in Egypt throughout the centuries. My own interest is in the early part of the collection, and specifically in the Indian material. However, there also are some examples of Islamic printed cloth, as well as later material that may be of importance to someone with a different focus of interest. The comprehensive catalogue makes the entire material available, so that future research may be done on all aspects of the collection.

At the beginning of the study of a great collection, however, must come the collector. As probably anyone who has had the opportunity to work through a large collection brought together by a single person has experienced, the cataloguer sooner or later has some sort of rapport with the collector, even when there is no chance of personal contact. This may be positive or negative, but indifferent it cannot be. Collecting with a purpose involves choosing, and one inevitably begins to wonder about the person who once responded strongly enough to particular objects to acquire them. The Newberry Collection of textiles is clearly not an art collection, but is primarily significant for historical reasons. There can be no doubt that this was why it was brought together. But Newberry must also have had a particular interest in the history of design, to gather as large a group of fragments as he did. He and his wife cared greatly for their collections, studied them, and published about them [12]. Apart from that, however, Percy E. Newberry has remained surpris-ingly elusive, despite the extensive correspondence addressed to him that is now in the Griffith Institute in Oxford. One reason may be that as far as is known he left no descendants, so that the physical link has been broken. It also must be due to his scrupulous separation of his personal and professional life. Very few personal letters survive among the archival material. It is difficult, therefore, to elaborate in this case on the relationship between the collector and the collection. What is left to us, though, is material that remains unique in significance.

By way of introducing it, a summary of previous research on Indo-Egyptian textiles seems appropriate. As there is a certain amount of confusion about the provenance and dates of the textiles, this might help to define the issues involved. At this point, the most current reasons for dat-ing, in particular the C-14 results, are also discussed at some length. In the production of artefacts, technical considerations are of foremost importance, as they determine to a large degree the eventual appearance of the object; furthermore, the technical aspects of textiles are not always clearly defined even by experts in their fields. The subject also still suffers from the problems of an unresolved classification. This publication must address that particular issue, and for that reason the discussion of the collection begins by looking at the technological aspect, rather than with the more conventionally art historical design analysis. This is not an attempt to revive Gottfried Semper’s dictum that ‘form follows technique’. I am not comfortable, though, with an interpretation of art and material culture that does not try to understand fundamental aspects of the production process.

After establishing the technological ground and discussing the difficulties of definition, the design characteristics are assessed and where possible interpreted. Initially the discussion focuses on the evidence that is immediately offered through the collection: common motifs and pattern arrangements are identified, and a statistical analysis of their numbers and arrangements on the textiles can help to produce a picture of the type of Indian textile appropriate for the trade to Egypt. It is inevitable that these design elements will also be sought after in other media, for comparative purposes.

Also of interest is the function of the textiles once they entered a different society. The role of dress and house furnishings, and the possible role the Indian trade goods had for both, are aspects to be looked at. A particular issue to be followed up is the possible identification of our printed fragments with a type of textile that is referred to in certain documents.

The textiles are of course remarkable in their own right, as examples of generally well-designed fabrics produced for common consumption. They would have been of historical interest if found in India itself. But as they all were collected in Egypt, their interest for us goes beyond their intrinsic appearance: to discuss them one must look at the issue of maritime trade between India and Egypt. Furthermore, publications of the last decade have illustrated very similar or even identical Indian textiles found in South-East Asia and Japan [EA1995.61]. These textiles do not come from archaeological sites, but from personal collections or communal treasuries, and we have detailed evidence for their functions in the South-East and East Asian context. Generally they are unlikely to be as old as the earliest of the textiles brought to Egypt.

In other words, the Indo-Egyptian textile fragments have become primary evidence in our understanding of the Indian Ocean trade network, with clearly comparative examples coming from as far afield as Egypt and Japan. Historically this is their particular interest, and an analysis of the collection would not be complete without considering this aspect. By presenting the Newberry Collection of block-printed textiles in its entirety, it is hoped that a foundation is provided for further, fruitful research in the history of the Indian Ocean trade.