Article: Pergamene Palm capitals and other Foliate Fancies

In the Internet Age, I figure that articles should be more freely available. This is one I wrote years, and years ago published Quaderni Ticinesi Numismatica e Antichita Classiche XXVI 1997 - yup, I know it's obscure, and I should chose journals on grounds other than 'my friend asked me' ... The text is what I have in a Word file, and should be quite close to the published version - though I have changed my mind about a few things since I wrote it - for some reason I could not copy and paste the footnotes, but happy to email people a copy.


Pergamene Palm capitals and other Foliate Fancies[1].

When this review of ‘palm leaf’ capitals began only eight were known; it was expected that it would be largely an analysis of the form for its own sake, and to be unlikely to add anything to our appreciation of Greek architecture. The documented capitals are four Archaic, from the Treasuries of Massalia and of Clazomene at Delphi[2], a possible temple in Phocaea[3] and another rebuilt into a tomb at Arkades[4] (Aphrati) in Crete, catalogued by Wesenberg[5]; and four Hellenistic, all from stoas, discussed by Coulton[6]. The four stoas are the two built in Athens by Eumenes II and by Attalus II, the North-East Stoa from the sanctuary of Athena Polias in Pergamon built by the former, and from the South Stoa (formerly known as the ‘Bazaar’) at Assos; one notes that the first three are from examples of Pergamene architecture, and the last from a town that falls within its sphere of influence. Attempts have been made to ascertain their origins, variously assigned as Egyptian, Minoan or miscellaneous Middle Eastern. The type was a variant capital, never an order, and its analysis benefits from the addition of other examples, along with a reexamination of the forms of Egyptian palm columns. I begin with a review of the eight Greek capitals.

The earliest of the palm capitals, discovered reused at the beginning of the sixth century and dated as early as the end of the seventh century, making it one of the earliest stone capitals, is the example from Arkades in Crete, now in Heraklion Museum (fig. 1). The capital and separately worked abacus are of local limestone. The context, found built into the structures of Tombs A and B in Arkades, is of little help, but the perfect fit of the two pieces and traces of red paint on both show that they belong together. The capital and abacus may not have been architectural members, and Wesenberg interprets them as from a votive column. The capital is decorated with twenty six thin leaves, that have an overhang at the top, and are separated by projecting ridges. An astragal around the bottom of the capital brings in an aspect of trompe l’œuil, ‘binding’ the leaves to the bell of the capital, and possibly hinting at the origins of the type. The large abacus consists of a simple square block, 0.49 x 0.50, whose sides are decorated with incised spirals; the underside has incised rosettes, designed using compasses, in the corners. The top does not carry any decoration, but has a smaller concentric square step.

The Treasury of Massalia, in the Marmaria at Delphi, was a small distyle in antis structure built ca. 535 BC of a fine-grained, blue-grey marble. The restored capital[7] is a composite of the two extant (fig. 2), one of which was made, for an unknown reason, with a separate abacus, and the other having had its astragal cut off at some point in the nineteenth century. These were single-calyx capitals; a two-tier form for the capital, as proposed by Pomtow[8], was conclusively disproven by Dinsmoor[9]. As at Arkades an astragal circles the base of the calyx, but here it is beaded and holds twenty-two leaves. Wesenberg restores darts between the ridges of the leaves, presumably on the basis of an additional ridge that appears between those of the main leaves, towards the exterior of the upper and lower surfaces, in the better preserved sections. Assigned to these capitals are shafts with twenty-two crisp Doric flutes and an arris at the bottom suggesting a base, believed to be of Asiatic form[10].

Although the Treasury of Clazomene, in the Sanctuary of Apollo at Delphi, is dated earlier than that of Massalia, to ca. 550 BC[11], its marble capitals are restored in line with those of the better preserved structure. The capitals with eighteen leaves, which again would seem to match the number of flutes on the proposed shaft, survive in only two fragments. They were slightly larger, but with a lower abacus than the Massaliot examples; 0.094 compared to 0.120. A parabolic Asiatic base, whose profile matched that of the base moulding on the walls, is restored. The treasury is believed to have been destroyed during the Hellenistic period.

The fourth capital attributed to the Archaic period comes from Phocaea, modern Foça. For reasons that will become apparent, discussion of this will come later.

To avoid repetition of order a third type of capital was needed when Hellenistic architects began to build two-aisled two-storied stoas, and those in Pergamon used the palm capital. Corinthian would have been more obvious, but was not used. It would have involved more labour, and thus expense. The Pergamenes in any case seem to have had an aversion to that order, and only one example[12], a heavily disputed votive column from the Stadium, is known from the Hellenistic period of the city. One can also note that Pergamon and Assos were sites where the use of the Doric rather than Ionic order was popular for temples and other buildings, leading to their description as ‘Doric sites’. In their stoic incarnation the shafts of columns that bear palm capitals, as an internal order, are not fluted; they all rest on Attic Ionic bases. A table drawn up by Coulton clearly shows that the Pergamene capitals increase in height in proportion to their diameter, with the curvature at the tops of the leaves taking up an increasingly smaller percentage of this height and projecting outwards less than the Archaic examples; the capital from Assos is excluded from these developments[13]. The Athenian capitals have twenty leaves, the examples from Asia Minor twenty four. A Doric abacus, with a typically Pergamene moulding along its upper edge and leaves in the corners of the soffit, is used at Athens and at Assos. Ridges, separated by a groove, continue to be used to mark their edges. The capital at Pergamon, diam. 0.448, from the stoa built by Eumenes II towards the beginning of his reign (197-159 BC), is the earliest of these (fig. 3); the stoa Eumenes built in Athens (fig. 4), and that built by his brother Attalus II follow chronologically (fig. 5)[14]. The capital from Assos is little more than a Doric capital whose echinus has been replaced by plastic leaves (fig. 6)[15] .

The time span between the last Archaic and the first Hellenistic capital would thus be of a minimum of some three centuries. This is problematic as only one of the Archaic buildings, the Massaliot Treasury, was definitely still standing and its capital accessible by the beginning of the second century BC, and there are considerable variations in representation. Suggestions of a Pergamene revival or continuation of an Archaic type are therefore possible, but unlikely[16]. I hope to show that they are not necessary by proposing a new injection of the type as seen in the Belevi Mausoleum, whose architectural core dates to the time of Lysimachos in the early third century[17].

The palm capitals from Belevi, of which only one example has been found, lined the interior of the ‘court’. Their design is decidedly Egyptianising rather than Greek, with strong links to Ptolemaic palm columns, although with a greater overhang at the tops of the leaves which reveals their other sides (fig. 7). Inscriptions from their architrave point to a decorative scheme in which statues of the Heliades would have stood between the columns, presumably illustrating the myth of Phaethon. The inclusion of the daughters of Helios, suggesting heliolatric links, is important when one considers links to the Egyptian palm columns. These were sporadically used throughout Egyptian architectural history, but seem to have been most popular, outside the Ptolemaic and Roman periods, during the Vth Dynasty and the XVIIIth Dynasty reigns of Amenhotep III and his son Akhenaten (Amenhotep IV), both periods in which sun-worship was brought to the fore; Bubastis, a site from which a XIIth Dynasty example comes, was also linked to sun-worship. Palm columns were symbols of the sun-god Re[18], and the palm generally, notably as a branch associated with other symbols in hieroglyphics, comes to stand for lengthy, by implication even eternal, reigns[19]. One can speculate as to the reasons for the use of palm capitals during other periods of Greek architecture, but one can suggest a concrete reason for which they were included at Belevi. This reason was a treaty of 289 BC between Lysimachos and the Ptolemies, which resulted in the marriages of his daughter, Arsinöe I, to Ptolemy II Philadelphus, and of himself to Arsinöe II, respectively the daughter and sister of the first and second Ptolemies. One can thus see the strongly Egyptianising Belevi palm capitals as an outward expression of this alliance, and taken in conjunction with the motifs of the coffers, where some figures of the Centauromachy wear Thracian helmets symbolising his ‘subjugation’ of those peoples, the structure as a whole as a glorification of his achievements and his reign. Given the historical context, the interpretation of the use of palm columns at Belevi is quite straightforward, and is amongst the least subtle of the many symbolic meanings Hellenistic architecture was used to convey. The structure at Belevi served as a Mausoleum, almost certainly its purpose from its inception, but was never the final resting place of Lysimachos; upon his death in 281 his wife and children fled Ephesos, and he was buried at Lysimachia, the Thracian town he had founded in 309. It is however difficult to doubt that it was he who had initiated the construction of the mausoleum. Given these circumstances a brief review of Egyptian palm columns would not be superfluous.

There are fewer palm columns in Egyptian architecture[20], that antedate the Greek examples, than one would imagine, most people seeing a mass of floral capitals, usually lotus or papyrus, and assuming them to be palms: in fact this is an illusion of the Roman period, from when the vast majority of these date. Much of the misconception, at least in Anglo-Saxon perception, can be attributed to the fact that three of the four columns in the Egyptian Hall at the British Museum are palm columns. A number are represented in early reliefs and paintings of architecture, and thus restored as non-extant wooden examples, but as these are lost, as would be many other types of columns, one should be careful about restorations: they would probably not in any case alter the proportion built in relation to other types of columns. The lotus, and the papyrus were more popular column types, as they respectively symbolised the kingdoms of Upper and Lower Egypt[21]. Like other vegetable columns, the palm column may have originated in the use of palm leaves decorating the tops of wooden columns, the motif in time coming to be reproduced in stone.

Palm columns were generally used in Egyptian architecture in courtyards and porticoes, sometimes within a mix of other column types. The typology was established in the monumental granite monoliths of the Vth Dynasty and copied into the Roman period. They were vividly painted: the examples from the ‘Amarna architect’ are exceptionally elaborate, having been gilded and decorated with inset glaze. Palm capitals were decorated by eight tall leaves which moved outwards with height, but had no overhang (fig. 8). A raised middle rib was delineated, and engraved parallel curved markings sprang from it. The shaft was smooth with a light taper, separated from the capital by five fillets, with a ‘loop’ below marking the ‘front’ of the column. It is hard to speak of an abacus, but a round mass of stone, with approximately the same diameter as the base of the capital, projected above the leaves. If one can speak of a ‘canonical’ format, this would have been it[22]. At some point after the XXth Dynasty a new type of capital developed whose leaves had a bulging overhang that projected outwards almost the same distance as the lower diameter of the capital. An abacus-like disc covers much of the overhang[23]. The body of the bell does not curve in the manner of the Archaic capitals, but is ‘T’-shaped in section. It is these Egyptian capitals, if any, that would have influenced the Archaic examples, rather than the more canonical form generally illustrated by Hellenicists. The orthodox shape was revived by the Ptolemies when they wanted to validate their position by establishing links with the Golden Age of Pharaonic Egypt.

The Belevi capital is decorated by eight leaves, of which only five are fully worked suggesting that the columns were placed against the wall, and linked to baseless Doric shafts of twenty-four flutes, again not fully worked[24]. The leaves, with a strong triple middle-rib, curve outwards displaying the tips of a second row, and their bodies incline at a regular angle rather than curve sharply outwards at the top. Their design has little in common with Archaic examples, and shows the strong influence of Egyptian capitals. Even the circular slab is retained, though here only 0.009 high, placed between the capital and the proposed abacus. It is quite easy to see how the Pergamene capitals could have been born out of this type, being a simplified and Hellenised development of it: as well as the development of the profile, one can note, first the addition of the base and then the abacus. There are examples of Doric capitals from Cyrene and Aegae[25] where leaves are painted on the moulding that replaces the echinus, which can be seen as painted versions of the short palm capitals from Assos, and suggest the possibility that the palm capitals from Assos may have developed out of an early Hellenistic elaboration of the Doric order and could predate those from Pergamene architecture. The Pergamene type would thus be a fusion of these and of Egyptian prototypes as represented by the Belevi Mausoleum.

Onions sees the use of palm capitals by the Attalids as a deliberate rejection of the Corinthian order[26]. Marcus Kohl has concluded that they were copying the capitals of the Massaliot Treasury, which was by then being used by the Romans, as a propagandic way of showing support for their allies[27]; this is an appealing argument, but more likely to be part of rather than the whole reason. Early Pergamene architecture made use of andesite, a hard stone locally available. The clean lines of, for example, Doric capitals are attributed to the difficulty in carving andesite. If this was so, one can imagine how much harder it would have been to render the complex tendrils of the Corinthian capital. An early Hellenistic palm capital has recently been found, which would seem to confirm this[28]. Possibly the original capitals of the ‘Megalesion’ of Pergamon, which were replaced in the Roman period, they have incised leaves on the bell, crowned by a low, circular abacus. It is close in style to the other Pergamene examples, but technically less sure and the design thus suffers. I have noticed two similar capitals at Priene, under a bush in the so-called House of Alexander.

The capital from Phocaea is dated to the Archaic period, on the grounds that the city was the Motherland of Massalia, the colony that erected the treasury at Delphi with palm capitals, its rendition is unlike the second century capitals, and because from the angle of the published photographs it betrays some, misleading, similarities to those from Delphi. A closer examination shows that, although badly preserved and not surviving to its full height, it is quite unlike either of the other Archaic capitals, and fits within the schema of the Belevi example: their profiles are similar, as is the representation of its leaves, which number ten on this capital. No abacus has been found, but the continuation of the stone above the leaves, on the upper surface of the capital suggests a low circular projection. In fact, other than having a foliate-covered bell, it is hard to see any similarities with the three Archaic capitals. The context of the capital, unearthed during the excavation of a temple, possibly that of Athena, helps little in its assignment. Wesenberg assumes that it must have belonged to the temple, but the dig also brought to light Ionic and Aeolic capitals, the latter of which one must not associate with the palm capitals. Little concerning the temple has been published, and lacking any evidence that precludes a Hellenistic phase, I would like to date the capital to that period, in the years following the Belevi Mausoleum[29] .

Although the consensus is for an Egyptian origin of the Archaic leaf capital, whether this came to Greece via the Phoenicians or the Minoans is disputed. The Armenia and Mesopotamia origins assigned by Lehmann-Haupt[30] could only be considered if one accepts the idea of a double-calyx Archaic format, and the same could be said of Achaemenid capitals. Phoenicia and Palestine are also often held up as sites from which these architectural elements came to Greece, but the evidence from Mycenae is more solid. A number of ivory architectural columns were found in houses there[31]. Among these were columns with palm capitals from the thirteenth century House of the Sphinxes, now in Athens, and since parallels can be found in ‘real’ architecture for a large number of the other types of model columns found, it would be reasonable to suggest that the palm capital may also have represented an architectural member in use at the time[32]. A stone lamp, Athens National Museum inv. 2921, similar in size to later architectural palm capitals, was found in the dromos of the ‘Lion Tomb’ at Mycenae, 15th-14th century. It is close in its design and execution to the Archaic examples, with an overhang at the top of the leaves, though the wide leaves curve far more here[33]. Aphrati is a complex site, currently being reevaluated. The similarity of the representation of the leaves and the shape of the bell of the lamp, and of the unusual nature of the abacus of the Aphrati capital suggest that it might not be necessary to date the latter in the Archaic period; its context, built into the structure of an Archaic tomb gives a terminus post quem but not a terminus ante quem, and its votive nature would only back this up. An almost identical Mycenean abacus is on display in Nauplion Museum. Thus one can identify a possible input in the Archaic capitals, but a direct Egyptian inspiration for those of the Hellenistic period.

A few other capitals from around the Mediterranean show links to palm capitals. A poros capital, found in Kanapitsa, near Thebes, but not dated[34], shows certain similarities to the Archaic leaf capitals. Its leaves are delineated by red paint rather than carved, on a smooth-profiled calyx, and it carries a round abacus. Another capital in the British Museum, no. 2571, from a Cypriot tomb, believed to be of the fourth century BC, recalls Greek leaf capitals; it has palm-like leaves, with a round-profiled moulding, decorated with a zig-zag pattern, above and below them[35]. On the facade of the Hellenistic tomb W 48, Cyrene, a very unusual column with a capital whose enlarged echinus has been replaced by a ring of palm leaves (fig. 10), is to be found flanked by two Ionic columns[36]. Its abacus is also idiosyncratic, curving upwards at the corners. Although falling within the Ptolemaic sphere of influence, the design is quite unlike either Egyptian or Greek models. The short leaves, with darts in between, resemble those on the echini of the Ionic capitals, enlarged and with their volutes removed; they lean slightly outwards at the top, but have no overhang. Only in the representation of the leaves can one see similarities to Greek palm capitals, and then only to those of the Archaic period. A Punic capital from the Quartier Mago, Carthage, may have been of similar type; engraved vertical markings on the bell suggest leaves.
Suggestions have surfaced occasionally for a connection between Archaic palm capitals and the origins of Corinthian capitals[37]. One can note that in Hellenistic Pergamon they take the place of the Corinthian order, but the Archaic argument is largely based on Pomtow’s theory of the double-calyx at Delphi. A more likely origin would be a development out of leaf-ringed Doric capitals, with residual influence being seen on early Corinthian examples in the necking of simple leaves below the acanthus. The two capitals types share a geographical sphere: Corinthian originated in the Peloponnese, where many examples of this Doric variant can be found[38].
Lotus capitals, should be noted as being a Pergamene type of leaf-capital that predate the use of the palm capital in the city, and are to be found on the Stoas and Propylon of the Sanctuary of Demeter[39], built by Queen Apollonis during the reign of her son Eumenes II, in the early second century (fig. 9). They are akin to Doric capitals, with a square abacus, but with a high bell in place of the echinus decorated with two overlapping rows of long thin leaves[40]. The sixteen from the outer row reach up to the abacus, curling outwards just below it; only the tips of the second row of leaves can be seen, except for elongated leaves that reach into the corners of the underside of the abacus. The abacus has a moulding along its upper edge, similar to the one found on palm capitals. The lotus leaves are more detailed than the leaves from palm capitals, with botanically accurate markings that distinguish the two rows, one showing the obverse and the other the reverse of the leaves. Unlike palm capitals, they are used with base-less shafts of twenty Doric flutes, and the height of the column in relation to its lower diameter, 7:1, follows the Vitruvian prescription for Doric. All those with entablatures preserved, point to a use in conjunction with an Ionic entablature. These are the capital type to be found used as the inner order of the South Stoa, the outer orders of the West and Upper North Stoas, and on the Propylon. They may also have been used externally on the South Stoa, the Ionic capitals being Roman replacements. The Lower North Stoa antedates the others, being of the mid-third century, and a Doric architrave suggests the use of those columns on the facade, otherwise one could see a unifying theme within the facades of the stoas surrounding the temple. It is hard to find direct antecedents for this capital type, although one can note similarities to the head gear made of lotus leaves worn by the Egyptian deities Nefertem, and Shefton has recently published two interesting petaliform stone capitals from a fourth-century temple at Sairkhe in Colchis, now in Tbilisi Museum, Georgia[41]. Although they are closer, as he has shown, to silverware than to the Pergamene capitals, I include them as an example of capitals, from the fringes of the Greek world and under Achaemenid influence, decorated with lotus petals arranged in three rows of decreasing size. Their shape, with a height including the abacus of only 0.483 as opposed to a width of the abacus of 1.011, is also quite unlike the Greek capitals.

Coulton sees the capitals from the Sanctuary of Demeter as having developed into a type known as lotus and acanthus, with a row of tall lotus leaves ringed at their bases by half their number of acanthus leaves[42]. These become popular in the Roman period, used particularly during the reign of Hadrian[43], under the Severans[44], and in Late Antiquity[45]. Although attempts have occasionally been made to find Hellenistic examples[46], the first certain use of this capital was on the Tower of the Winds[47] at Athens. They seem to have been most popular in Greece, although examples are known from around the Mediterranean, often used in association with Corinthian capitals[48]. The only link one can between the capital type and stoas, is that the capitals of the upper storey of the Hellenistic South L-shaped Stoa on the North Market at Miletus were replaced in the Roman period with lotus and acanthus capitals; there is however no evidence to show that this was the original type[49].

Returning to palm capitals, the resurgence of the type in Byzantine architecture, most notably on rood screens[50], following relatively few examples of their use during the Roman period[51], has long puzzled me. It has been noted that there was a Hellenistic Stoa in Athens whose architecture suggested Pergamene influences[52]: substantial elements were found near the Roman Agora suggesting an original location there, and more pieces were found on the Acropolis, where parts of it are believed to have been used in a reconstruction of the interior colonnade of the Parthenon cella, following a fire in the Roman period. While reading about the later reworkings of the Parthenon that I came across a document, which would seem to offer a solution[53]. The anonymous Viennese manuscript dated to ca. 1458-60, is the oldest extant description of this version of the interior of the building. Towards the end of the passage the writer speaks of “capitals hewn like palms”. Though one should always be wary of Mediaeval travellers’ accounts, the number of coincidences and the great difference between Doric and palm capitals suggests that this is not a mistake. Thus these words are important for they are our only source for these palm capitals, an unusual order not regularly used, and add weight to the argument that the stoa, although later and not of as high a standard as those of Eumenes and Attalus, which has been described as ‘the third stoa of Pergamene type’, may well have been Pergamene. Four Asiatic Ionic bases dating to the Hellenistic period can still be seen on the Acropolis, and may be linked to the capitals. The Parthenon when it had been reincarnated as a Christian Metropolis would have been amongst the most famous and influential in the Christian world, and one can note a similarity between the concepts of a double tier of columns running around three sides of the cella and the rood screen. Therefore I would like to suggest that Byzantine rood screens which incorporated palm capitals were imitating those within the Parthenon, a building already standing proud in architectural history, that would have increased in importance to the early Christians through the act of its metamorphosis from the shelterer of a pagan cult into one of their holy buildings.

The naming of the capital type has caused confusion; they were generally referred to as Aeolic until around 1970, but this is the name given to a different form of capital. Wesenberg, p. 43, in 1971, refuses to call them palm capitals, instead referring to them as capitals of the ‘Arkades type’: he lists and dismisses the other appellations, but I shall not repeat his arguments here. To conclude I believe I have shown that the Hellenistic examples, which are closely derived from the Egyptian, are schematised representations of palm trees, and that these are a separate type from the foliate capitals of the Archaic period.

Appendix - The distribution of Egyptian palm columns.

Dyn. date

V Abusir, court of pyramid Dynasty ca. 2494-2345: Sahure, second ruler of nine.
V Nyuserre sixth ruler of nine.
V Saqqara[54] Unas, last ruler of nine.

XII Deir el Bersheh, tomb[55] Dynasty ca. 1991-1786
XII Tomb
XII Bubastis[56]
XII? Ehnasya

XVIII Soleb, Sudan, courtyard[57] Amenhotep III: 1417-1379
XVIII Sesebi, Sudan, courtyard Akhenaten: 1379-1362
XVIII Tel el Amarna[58] Amarna period: 1372-1350
XVII? Tanis, temple of Amun
XVII? capital, believed to be from a tomb, in the gardens of the École Biblique, Jerusalem

XIX Bubastis Rameses II: ca. 1250[59]
XIX Heracleopolis Rameses II: ca. 1250[60]

XXVI Memphis, Palace of Apries Saïte: 664-525

Ptol. Kom Ombo 305-30
Ptol. Edfu 305-30
Ptol. Philae The columns from the courtyard of the temple of Isis are mostly third century, built under Ptolemies II & III, but some are of Roman date.
Ptol. Alexandria, two tombs.
Roman Alexandria, tomb.

Fig. 1 - a) underside of abacus, b) axial view of capital and abacus, Arkades. Drawn by the author, after Levi.

Fig. 2 - elevation of capital from the Treasury of Massalia, Delphi. Redrawn by the author after Dinsmoor.

Fig. 3 - elevation of capital from the Stoa in the Sanctuary of Athena, Pergamon. Redrawn by the author after Coulton.

Fig. 4 - elevation of capital from the Stoa of Eumenes, Athens. Redrawn by the author after Coulton.

Fig. 5 - axial view of capital from the Stoa of Attalos, Athens. Drawn by the author.

Fig. 6 - elevation of capital from the Stoa in Assos. Redrawn by the author after Coulton.

Fig. 7 - a) elevation of capital, b) underside of capital, Belevi. Drawn by the author, based on Ephesos VI.

Fig. 8 - diagram illustrating the Egyptian canon for palm columns. Drawn by the author.

Fig. 9 - a) elevation of capital, and b) view from below, Lotus capital from the Sanctuary of Demeter at Pergamon. Redrawn by the author, based on Altertümer von Pergamon XIII.

Fig. 10 - capital from Tomb W 48, Cyrene. Drawn by the author after Stucchi.