Table of Contents
Thirty years ago I wrote:
The carved stucco dados from Samarra, one of which Herzfeld believed to have been derived from Mushattā, actually are translations into stucco of Late Antique stone dado revetments … .
I believe that nearly all the hundreds of decorative motifs found at Samarra were derived in this same way: dado revetment designs were transposed into the Bevelled Style (or one of the other two “Samarra styles”) for use in dado revetments.1
I was wrong. I inferred the existence of Late Antique carved stone dadoes from a panel excavated in Antioch, but in fact there are none. The Antioch panel is instead a pilaster capital or something similar.2
Retiring from my leapt-to position, I fall back on Ernst Herzfeld's analysis in Der Wandschmuck der Bauten von Samarra und Seine Ornamentik. Herzfeld organized the stucco from Samarra (221–79/836–92) by design (Ornament) in three styles, First, Second, and Third, of which the Bevelled Style was the First Style. According to Herzfeld (and simplifying somewhat) almost all the designs executed in the Bevelled Style were developed from borders (Kanten). In the introduction to the field ornament series of the section on the First Style he wrote:
One of the most conspicuous aspects of the First Style of Samarra is that its field ornament (Flächenmuster), with but few exceptions, is not proper field ornament, that is, ornament that can be extended indefinitely in two directions or instead laid out in two or more directions from a point of origin. On the contrary, it consists of simple enlarged or repeated borders (Kanten), which can be extended indefinitely in only one direction.3
Alles ist Kanten. Borders could be be stacked on top of borders: the title of one section of Wandschmuck is “Flächenmuster aus zwei oder drei übereinander gereihten und verbundenen Kanten” (field ornament of two or three superimposed and interconnected borders).4 Herzfeld published two photographs of such stacking in relief wall decoration from the Parthian palace at Assur.5
For Herzfeld the border design from which most Bevelled Style field ornament was derived was the Hellenistic acanthus frieze as it had been developed in Mesopotamian art since the Parthian period. Its simplest form, which Herzfeld called the Samarra Frieze, is a series of leaves with ridges down the center and nothing between them. In an only slightly more complex form (ornament 27) tripartite flowers rise from between the bases of the leaves and open out between their tips (see Herzfeld Archive FSA A.06 04.PF.19.048). A closely related border design replaces the leaves with flower buds; Herzfeld generally called this alternation of open flowers and buds the “Blüten-Knospen-Reihung” (flower-bud-sequence), and saw it as derived from the ancient Egyptian lotus-flower and -bud frieze, again as it had been developed in Mesopotamia.6
Herzfeld arranged the Samarra Bevelled Style border designs in a series of lettered groups (A–K), and the field ornament in another series (M–X), according to salient characteristics of composition and in order of increasing complexity.
These groups employ the Samarra Frieze and the “Blüten-Knospen-Reihung” in various compositional elaborations of their simple forms, enriched by some details in keeping with their vegetal origins. Aside from the addition of beaded (or “pearl”) bands no other motifs are used.
Herzfeld's analysis is economical and explains the variety of Bevelled Style field ornament effectively. It requires no reliance on the art of the Late Antique world, such as I imagined. Instead it implies very considerable invention in composition coupled with restraint in selection of motifs, both within a tradition that Herzfeld traced right back to Parthian Mesopotamia. I was wrong to claim that Samarra Bevelled Style dadoes are translations of Late Antique models.7
Several of the points I made about the Bevelled Style in Five Essays remain valid, I think. The Bevelled Style was used to transform Late Antique ornament in such cases as the capitals from Raqqah that Herzfeld published.8 It was an artistic horizon. It has nothing to do with Central Asian horse trappings. And there is nothing “Islamic” about it.
The chronology of the three Samarra styles remains unclear. Creswell reversed Herzfeld's sequence, in which the Bevelled Style (First Style, Creswell's Style C) came first without saying why, although he observed that Third Style stucco (his Style A) was found in the Bāb al-ʿĀmmah, which belongs to the beginning of the Samarra period. Herzfeld found Bevelled Style stucco elsewhere in the Dār al-ʿĀmmah and believed it to be of the period of the foundation of the city. The topic was revisited most recently by Matthew Saba, who pointed out that according to Herzfeld the Bevelled Style decor (in both marble and stucco) in the quadripartite block with basilical halls at the east end of the Dār al-ʿĀmmah, which Saba called the Audience Hall Complex, was of the period of construction and not a result of later renovation, the traces of which were detectable elsewhere. Saba, cautiously, drew no conclusion on the matter but I think it is very unlikely that the quadripartite block is not of the same date as the Bāb al-ʿĀmmah.9 Thus the First and Third Styles coëxisted at the outset of the Samarra period, and perhaps all three styles were in use simultaneously.
Given that the three styles at least overlapped each other, it is remarkable that their field ornament differs strongly in composition and motifs. They are alike mainly in conforming to the same scheme of interior decoration:
The decoration is confined to the socle of the wall, entirely naturally for men who squat on the floor. The socle is something over 1 m., that is a double ell, high. … Aside from the socles only the doorways were decorated, surrounded with decorative strips, and occasionally upright strips climb the walls, rhythmically repeating the door frames. Nowhere is the upper termination of the walls preserved. But with certainty one may assume that there was a frieze there, as terminates the wall in the Mosque of Aḥmad Ibn Ṭulun, and that the customary decoration of these friezes was the Samarra Frieze.10
Both the decorative scheme of Samarra and the motifs of the three styles differ strongly from earlier ʿAbbāsid stucco.
Michael Meinecke surveyed ʿAbbāsid carved stucco from the period before Samarra and contrasted the expansive decorative scheme of Samarra to the comparatively spare scheme employed at Raqqah (177–92/793–808), where fragments of large compositions were encountered loose, but only door frames were found in situ.11 In this scheme vertical strips of carved stucco were applied to the wall on either side of a doorway and on both faces of the wall (but not on the doorway's inner faces, or jambs). Such strips could frame all the interior doorways of, for example, a palace's reception block. Only the bottommost sections of these strips were found, so it is not known how the top of the door frame was arranged.
Meinecke saw the Raqqah stucco as derived from Antique Syrian prototypes—not the Late Antique of nearby Ruṣāfah but the second- and third-century architecture of farther-off Palmyra.12
Meinecke also considered the eighth-century carved stucco excavated at Ḥīrah, which is limited to door frames as well but employs the motifs seen at Raqqah in different proportions:
The plaited bands of the borders are almost identical to the al-Raqqa friezes, but most of the major patterns are markedly different. Though climbing scrolls of vine ornament are [also] represented in somewhat coarser variations, in most cases the friezes depict series of alternating geometric forms—square, circles, multifoils—densely filled with vine elements. Thus this can be described as a different pattern type—that of basic geometric ornamentation, while the vegetal forms only constitute secondary elements.13
Some of the Ḥīrah designs are guilloches with beaded bands. Those that are merely juxtapositions of squares, circles, and so on may be merely degenerate guilloches.
Meinecke thought that the generally contemporary carved stucco door frames from ʿAlwīyah in present Saudi Arabia14 were executed by workers from Raqqah.15 While they resemble the Raqqah door frames in some respects (the borders, the use of a vine scroll, the rendering of the leaves), they are of better quality and not closely paralleled at Raqqah. Could they have been executed by another (possibly Baghdadi) workshop?
As for the Bevelled Style, Meinecke thought it might have arisen in Raqqah shortly before the foundation of Samarra, and that at Samarra it might date earlier than has been proposed by some.16 He proposed that the Bevelled Style capitals found at Raqqah and published by Herzfeld were carved not during the residence of Hārūn al-Rashīd there but later.17 He also attributed the Bevelled Style main mihrab in the Great Mosque of al-Rāfiqah, Raqqah's companion city, to an inferred brief residence in Raqqah by al-Muʿtaṣim after his first attempt at founding a city at Samarra but before the buildings on the site eventually chosen would have been ready.18
But the Samarra dado scheme was found nowhere at Raqqah, and the closest connection of Raqqah stucco to Samarra is with the Third Style, not the First.
The Late Antique stone architecture of Palmyra is a plausible source for the vine scrolls of the Raqqah door frames, but not for the Raqqah scheme of interior door frames in stucco. Meinecke's Palmyran sources for Raqqah motifs include a soffit, a niche head, panels applied to pilasters, exterior corner piers, and a figural sculpture. He claimed that architectural ornament at both Raqqah and Palmyra was confined (beschränkt) to pilasters on the one hand and door and niche frames on the other, but this formulation ignores the considerable difference between the thoroughly classical architecture of Palmyra (an architecture of parts, or members) and that of Raqqah (an architecture of wall surfaces). The doubled vine-scroll border he cited in support of this view, in the Temple of Baalshamin, framed a shallow niche holding a cult figure in a composition that does not conform to the style of the temple.19
The Raqqah scheme was realized with (at least) two different families of motifs: the vine scroll that Meinecke thought was derived from Palmyra and the guilloches (or degenerate guilloches) he pointed out at Ḥīrah.20
So the Raqqah scheme may have been independent of any particular motifs and preëxisted any borrowing from Palmyra. It does not seem to have been employed in any medium in Umayyad architecture (for the case of Qaṣr al-Ḥayr al-Sharqī see Footnote 20). In this connection, however, the carved stone frame of the door between the porch and the entrance hall of the palace at Khirbat al-Mafjar is of interest.21 Its ornament is not a guilloche but a series of square panels enclosing simple, mostly geometric motifs, some of them interlacings, within a surrounding braided border. It has the feel and something of the visual texture of the Ḥīrah door frames, and Robert Hamilton pointed out that similarity. It shows how such frames may have been finished off at the top: with a horizontal section the same height as the width of the vertical sections.
The Samarra scheme was realized in two much different ways. In the First Style the door frames are vertical stacks of panels, sometimes of different heights, filled with single flowers. These stacks are separate from the dadoes and of different visual texture. The Third Style, however, continues the dado design into the door frame, so there is no actual door frame below the top of the dado, and the decoration runs continuously, like wallpaper (except for the anachronism: it is perhaps more like tiling).22
A great deal of the corpus of Samarra stucco falls easily into Herzfeld's three styles. There are edge cases, odd variations, and anomalies, but not very many. The three styles are internally consistent and also markedly different from each other. This consistency and differentiation may reflect conditions in the construction industry of Samarra.
The business organization of construction in Samarra has barely been touched upon in print.23 There was strong demand for luxurious as well as mundane buildings from Samarra's founding through the reign of al-Mutawakkil, and an enormous area was developed.
The career of the London builder Thomas Cubitt (1788–1855) offers an example of one business response to similar conditions. London expanded greatly during his lifetime, favoring the development of large-scale enterprises.
In the first quarter of the nineteenth century Cubitt built up a construction enterprise that possessed the facilities for producing everything he needed (bricks, ironwork, and so on) and employed his own workers instead of subcontracting.24 In this way Cubitt gained economies of scale and consistently high quality of materials and workmanship; he was known for being able to execute a large undertaking reliably and was able to take on ever larger projects. In the second quarter of the century Cubitt went on to build large tracts of Belgravia and Pimlico.
At Samarra Cubitt's approach should have been similarly successful. Perhaps large enterprises capable of executing most if not all phases of a construction project already existed in Baghdad before Samarra was founded. They would have developed the ability to execute large projects reliably, they would have had economic advantages of scale, and there would have been fewer rather than more of them.
There may have been no such enterprise that maintained an in-house stucco workshop, but any large builder would have wanted to be able to rely on some workshop known to be capable of producing work of good quality reliably and on schedule. And so I think it likely that at Samarra there were also fewer rather than more stucco workshops, which may be reflected in the consistency and differentiation of the three Samarra stucco styles.
The repetitive nature of Samarra stucco in all three styles lent itself to a division of roles between few designers and many carvers (and other workmen).
So however the Bevelled Style was invented, the corpus of Bevelled Style stucco at Samarra could have been largely the output of a single designer (and perhaps associates), exercising very considerable invention within a tightly constrained method of composition and selection of motifs.
1. Terry Allen, Five Essays on Islamic Art, 1988, p. 12.
2. The Antioch panel is in the Baltimore Museum of Art. An image of it can be found by searching for accession number 1940.170a at http://collection.artbma.org. From the relevant footnote (no. 27 in Chapter One of Five Essays):
First published by Richard Stillwell in Antioch-on-the-Orontes [various editors, 5 v. in 6, Princeton, 1934–72], v. 3, The Excavations 1937–1939, pp. 169–70 and pl. 42, no. 231; more recently in Kurt Weitzmann, ed., Age of Spirituality, pp. 667–68. The piece has also been published, I believe mistakenly, as a pilaster capital by J.-P. Sodini, et al., “Dèhès. Campagnes I–III, 1976–1978”, [Syria, v. 57, 1980, pp. 1–304], fig. 302 (cf. also the useful comparative material).
3. Forschungen zur islamischen Kunst [6 v. in 21, 1911–48], ed. Friedrich Sarre, pt. 2, Die Ausgrabungen von Samarra, v. 1, Der Wandschmuck der Bauten von Samarra und Seine Ornamentik, Berlin, 1923, p. 62. This point is hammered home in Wandschmuck, but somewhat buried in K.A.C. Creswell's translation of Wandschmuck, p. 12, in Early Muslim Architecture, 2 v., Oxford, 1940, v. 2, p. 288, sixth and seventh full paragraphs.
4. Wandschmuck, pp. 12, 69.
5. Ibid., pl. 101, upper left and upper right. Herzfeld's first experience in the Near East had been with Walter Andrae's excavations at Assur, and those photographs were later published in Walter Andrae and Heinz Lenzen, Die Partherstadt Assur (Ausgrabungen der Deutschen Orient-Gesellschaft in Assur, v. 8; Wissenschaftliche Veröffentlichungen der Deutschen Orient-Gesellschaft, v. 57), Leipzig, 1933, pl. 21 e and b (drawings on pl. 20, f and d).
6. Ibid., pp. 25–28 on the Samarra Frieze; p. 30 for ornament 27; pp. 20–21 for the “Blüten-Knospen-Reihung”.
7. The dado design from the Balkuwārā Palace, ornament 156, that I compared to the Antioch marble panel resembles it less when one views the stucco dado alongside the other members of its group, U (see FSA A.06 04.PF.23.095). I am still fond of viewing ornament 118, from House XIII (see FSA A.06 04.PF.19.106), as a transformation of the composition formed by quarter-sawn marble, which was imitated in paint at Qaṣr al-Ḥayr al-Gharbī and Bālis. It is sufficiently different from other Bevelled Style field designs that Herzfeld put it in its own group, P. For Qaṣr al-Ḥayr al-Gharbī see the article by Daniel Schlumberger, “Les fouilles de Qasr el-Heir el-Gharbi (1936–1938). Rapport Préliminaire”, Syria, v. 20, 1939, pp. 193–238, 324–73, reprinted with additional illustrations as Schlumberger, et al., Qasr El-Heir el Gharbi (Bibliothèque Archéologique et Historique, v. 120, 1986), pl. 57 a (published upside-down), 57 c–e. The Institut Français du Proche-Orient has recently put scans of photographs from Schlumberger's excavation online at Médihal, where the entire set can be found by searching for “Qasr el-Heir el Gharbi”; those corresponding to the published images I referred to, and a few others, may be found at 5226390, 8325599, 8325622, 8326030, 8326035 (upside-down), 8328876, 8328877, and 8328879 (some seem to be mirror-reversed). For Bālis see Thomas Leisten, “For Prince and Country(side)—the Marwanid Mansion at Balis on the Euphrates”, Karin Bartl and Abd al-Razzaq Moaz, eds., Residences, Castles, Settlements. Transformation Processes from Late Antiquity to Early Islam in Bilad al-Sham (Orient-Archäologie, v. 24, 2008), pp. 377–94, fig. 10, 11. The Qaṣr al-Ḥayr al-Gharbī painted imiation revetment is somewhat taller than the Samarra dadoes, and the Bālis decoration as (plausibly) reconstructed was much taller.
8. I think the wooden panel in the Museum of Islamic Art, Cairo, no. 6280/2, is probably also a transformation rather than an invention (see its entry at discoverislamicart.org).
9. Early Muslim Architecture, v. 2, p. 286. Matthew D. Saba, “A Restricted Gaze: The Ornament of the Main Caliphal Palace of Samarra”, Muqarnas, v. 32, 2015, pp. 155–95, p. 173. I do not use Saba's terminology because I do not think that audience took place in the quadripartite block.
10. Wandschmuck, pp. 2–3, elaborating on Herzfeld's Erster Vorläufiger Bericht über dir Ausgrabungen von Samarra, Berlin, 1912, p. 15, which was the source for Creswell's translation, Early Muslim Architecture, v. 2, p. 283. Herzfeld omitted mihrabs from consideration here, as do I. The remark about “men who squat on the floor” leads into a discussion of Assyrian orthostats. It occurs to me that one of the aesthetic functions of a dado of this height is to elevate painted decoration on the wall above it closer to eye level, as seen at Samarra in House XIII (see FSA A.06 04.PF.19.131 and FSA A.06 04.PF.21.076).
An arched doorway the inner sides of which were decorated was found in House XII. See FSA A.06 04.PF.19.014 (more picturesquely, FSA A.06 04.PF.23.170), and FSA A.06 04.PF.19.096. It is unusual in being framed by a continuous border rather than a stack of panels. In some cases the stucco door frame stops short of the door openings, the inner faces of which are unadorned, as in House XII (see FSA A.06 04.PF.19.013 and FSA A.06 04.PF.19.121). Such doorways may have been clad in wood to accomodate wooden door leaves.
Upright strips rising above the top of the dado and also clearly unassociated with a door frame seem to occur only in House I: ornament 152, referring to FSA A.06 04.PF.19.139; the line drawing illustrating this ornament, fig. 151 (folding), shows another, clearer, example. Perhaps these tied into the ceiling decoration.
As for the frieze at the top edge of the wall, Herzfeld attributed the Third Style fragments he found in the Bāb al-ʿĀmmah to such friezes, as well as to a frieze around the vaulted area of the central bay: “Fries um die Gewölbeflächen der großen Tonne” (Wandschmuck, pp. 196–201). Samarra Friezes in both marble and stucco were found in the Dār al-ʿĀmmah and in stucco in the Balkuwārā Palace (ibid., pp. 28–33), but marble is an implausible material for friezes in such locations. Perhaps they were also executed in wood, and in other designs as well, as at the Bāb al-ʿĀmmah. The Bāb al-ʿĀmmah deserves a visual reconstruction, as so much of its stucco decoration was discovered.
Exceptionally, in the palace at Ḥuwayṣilāt identified by its excavators as the Qaṣr al-Jaṣṣ, built under al-Muʿtaṣim, carved stucco extends across the whole wall above the dado (Iraqi Government, Department of Antiquities, Excavations at Samarra 1936–1939, Baghdad, 1940, 2 v., v. 1). I obtained a PDF scan of this publication in August 2014 from http://digital.library.stonybrook.edu/cdm/ref/collection/amar/id/123417 .
11. “Early Abbasid Stucco Decoration in Bilād al-Shām”, Bilād al-Shām During the Abbasid Period, ed. Muhammad Adnan al-Bakhit and Robert Schick, Amman, 1412/1991, pp. 226–67, p. 229. See now Verena Daiber and Andrea Becker, eds., Baudenkmäler und Paläste I (Raqqa III), Mainz, 2004; and Ulrike Siegel, Die Residenz des Kalifen Hārūn ar-Rašīd in ar-Raqqa/ar-Rāfiqa (Syrien) (Raqqa 4), Berlin, 2017.
12. “ʿAbbāsidische Stuckdekorationen aus ar-Raqqa”, Barbara Finster, et al., eds., Rezeption in der islamischen Kunst (Beiruter Texte und Studien, v. 61), Beirut 1999, pp. 247–67, pp. 258–61. Detailed comparisons of ornament from Raqqah and Palmyra can be found in Meinecke, “Palmyra und die frühislamische Architekturdekoration von Raqqa”, Syrien von den Aposteln zu den Kalifen (exhibition catalogue), Linz, 1993, pp. 352–59.
13. “Early Abbasid Stucco Decoration”, p. 230. For the date see D. Talbot Rice, “The Oxford Excavations at Ḥīra”, Ars Islamica, v. 1, 1934, pp. 51–73, p. 61; recapitulated by Creswell, Early Muslim Architecture, v. 2, pp. 164–65.
14. Terry Allen, “An ʿAbbāsid Fishpond Villa Near Makkah”, Occidental, 2009, http://www.sonic.net/~tallen/palmtree/fishpond/fishpond.htm .
15. “Early Abbasid Stucco Decoration”, p. 230.
16. “ʿAbbāsidische Stuckdekorationen”, p. 263, n. 38.
17. During the reign of al-Ma'mūn, “Early Abbasid Stucco Decoration”, pp. 232–33; or around midcentury, “ʿAbbāsidische Stuckdekorationen”, p. 263.
18. “Early Abbasid Stucco Decoration”, pp. 233–34. The mihrab is illustrated in Daiber, et al., Baudenkmäler und Paläste I, pl. 5 b. In “ʿAbbāsidische Stuckdekorationen”, pp. 263 and fig. 4, however, Meinecke remarked how closely it resembles Herzfeld's reconstruction of a frieze from the main mihrab of the Great Mosque of al-Mutawakkil.
19. See the reconstruction by M. Gawlikowski and M. Pietrzykowski, “Les sculptures du Temple de Baalshamîn à Palmyre”, Syria, v. 57 [not v. 56 as often cited, including by Meinecke], 1980, pp. 421–52.
20. Meinecke wrote of a “shift in emphasis from more naturalistic vegetal patterns to geometric arrangements [that can] be ascribed to the period of Hārūn al-Rashīd”, “Early Abbasid Stucco Decoration”, p. 230, but if one family of motifs displaced the other it would have been the vine scroll displacing the guilloches, as Ḥīrah is fairly well dated before Raqqah by coin finds. Meinecke also cited the Large Enclosure at Qaṣr al-Ḥayr al-Sharqī, where, he argued, the stucco door frames found in situ belong to “a later phase of occupation” after Raqqah (op. cit., p. 231 and n. 20). While the stucco of Qaṣr al-Ḥayr al-Sharqī is not all original, I do not see the “meander borders [and] ‘Sāmarrā' friezes’” he cited in support of a late date. See Oleg Grabar et al., City in the Desert: Qasr Al-Hayr East, 2 v., Cambridge, Mass., 1978, v. 2, figs. 133, 139, 140, and 141.
21. R.W. Hamilton, Khirbat al Mafjar: An Arabian Mansion in the Jordan Valley, Oxford, 1959, pp. 25, 136–37, pl. 3, 16.
22. From the Qaṣr al-Jaṣṣ examples are Excavations at Samarra 1936–1939, pl. 5, 8 top, and 11, drawings pl. 19 and 24 (with a border between the dado and the door frame strip).
Herzfeld published only one example of this wallpaper or tiling scheme, which he described: ornament 171, which covered the walls of the “Kuppelsaal des Harem” in the Dār al-ʿĀmmah. It was executed in a unique technique: the design was cut in low relief and then molded vine leaves and perhaps stems were attached (see FSA A.06 04.PF.22.078, FSA A.06 04.PF.22.077, FSA A.06 04.PF.22.073, FSA A.06 04.PF.22.074, FSA A.06 04.PF.19.254, FSA A.06 04.PF.19.255, FSA A.06 04.PF.23.164, FSA A.06 04.PF.19.143, and FSA A.06 04.PF.19.257).
Herzfeld assigned ornament 171 to the First Style, but in its composition it matches much better some examples of the Third Style (ornaments 271 and 273). They are best appreciated from Herzfeld's reconstruction drawings, but see FSA A.06 04.PF.19.197 and FSA A.06 04.PF.19.195. Herzfeld analyzed these patterns as meanders, which I think is incorrect topologically. I see them as exploded (or shrunken) tiling patterns, in which the floating polygons can be pushed together (or expanded) to fill the plane.
The preserved lintel of the western entry at Qaṣr al-Ṭūbah is similar to these patterns in its L-shaped polygons filled with vine scrolls and floated on a background of similar visual texture (Creswell, Early Muslim Architecture, v. 1, pt. 1, pp. 607–13, pl. 137 c, 138 a; Antonin Jaussen and Raphael Savignac, Mission archéologique en Arabie, 3 v. in 6, 1909–22, v. 3, pl. 17).
23. Marcus Milwright, “Fixtures and Fittings. The Role of Decoration in Abbasid Palace Design”, A Medieval Islamic City Reconsidered: An Interdisciplinary Approach to Samarra, ed. Chase F. Robinson (Oxford Studies in Islamic Art, v. 14), Oxford, 2001, pp. 79–109, considered some relevant topics, but not construction as a business.
24. Hermione Hobhouse, Thomas Cubitt, Master Builder, New York, 1971, an excellent book.