Rostovtzeff 1911 on the Mosaics of Praeneste (Palestrina) – English Translation

On this page

About the translation

In 1911 Michael Rostovtzeff (1870–1952) published a long article on architectural landscapes in Hellenistic-Roman painting and mosaic. In several places in the article he discusses the mosaics of Praeneste, today Palestrina, and asks whether the buildings in the Nilotic landscape correspond to Egyptian reality. Below is an English translation of these passages, as a source. The reading of the mosaic itself is given on the page The Nile Mosaic of Palestrina.

Source: M. Rostowzew, “Die hellenistisch-römische Architekturlandschaft”, Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 26 (1911), pp. 1–186, pls. I–XI. The scanned volume at the Internet Archive.

Translated from the German for this edition. Page numbers follow the original article. Square brackets [ ] are the translator’s additions, and passages translated as a summary are marked “summary”. All passages, page and figure numbers, Greek forms and references were checked against the page images of the scan. The article is in the public domain.

1. The two oldest landscape mosaics: Stabiae and Praeneste (pp. 49–50, fig. 29)

“Before I turn to landscape as an ornamental element of third-style wall decoration, I would like to point out the close kinship that links the two oldest landscape mosaics known at present with the landscape paintings of the second and third styles. I mean the mosaic from Stabiae with the so-called Plato, and the picture from the temple of Fortuna at Praeneste. In the former […references] the figural scene is composed into its architectural setting: the schola itself, the column with the sundial and the tree all appear necessary to the composition, whereas the sacred gate and the city in the background merely indicate the place where the event happens. It is characteristic, however, that all these elements are typical of the sacral landscape of the third style. I shall return below to the conclusions that follow from this.

Equally characteristic is the mosaic from the temple of Fortuna at Praeneste, which dates from the time of Sulla, as Delbrück has shown (fig. 29, facing p. 48). Of interest here is the combination of the sacred column dedicated to Poseidon standing on the shore (within the schola already familiar to us, and with a burning altar in front of it) and a large expanse of water, which here, admittedly, is treated not naturalistically but as the background for a learned zoological display of fishes. And yet the picture involuntarily reminds us of the seascape from the Naples Museum described above. Incidentally, both the description and the interpretation of the picture by Marucchi seem to me to need closer justification before they can convince.”

Notes on p. 50: (1) Delbrück, Hellenistische Bauten in Latium I, p. 59. (2) “The relevant part was published by Marucchi, Bull. Com. 1904, pp. 233 ff. (fig. 4 and pls. VI–VII); cf. Alinari no. 27280, from which the accompanying block was made.”

Caption of fig. 29: “Mosaic in the grotto at Praeneste, detail (Alinari).” [This is the fish mosaic from the grotto of the sanctuary, not the great Nile mosaic.]

2. The Egyptianising landscape: the same building types (pp. 59–60)

“In the Egyptianising landscape, then, as in the purely sacral landscape of third-style pictures and in the large landscape of the second style, the same architectural types recur again and again. The buildings that characterise the purely sacral landscape do not occur, or only quite exceptionally; other forms prevail instead: small temples with a flat concave roof [flachgewölbt], the pylon tower, the round tower, the altar-shaped building, the ‘house with garden’, the gallery building. Apart from the cylindrical building with the portico, these types are foreign to the sacral landscape; on the other hand we meet some of them on the friezes and in the large landscapes of the second style. I mean the types ‘house with garden’, round and square tower, and in addition the peculiar building of the Farnesina (p. 37) with a lattice-like superstructure, which reappears on a pygmy mosaic in the Museo delle Terme. For the purely Egyptian buildings of the large second-style landscapes I need not look for analogies in our landscapes: the light reed buildings, as on the black wall of the Farnesina (p. 32), and the pylon gate, as on the white frieze of the Farnesina, are also familiar types in the pygmy landscapes. Now the question arises: apart from the exotic elements, such as the well with the shaduf, the crocodile stele and the like, is the rest of the architecture also really Egyptian, or are these arbitrary buildings placed into a Nile landscape?

I have no doubt that we are dealing with Egyptian architecture. First, several of our types also occur in the famous mosaic of Palestrina, whose Egyptian character, and probably also its direct or indirect Egyptian origin, nobody has disputed. Second, a whole series of the types discussed above can be demonstrated in the actual architecture of Ptolemaic and Roman Egypt, as far as it is known to us from monuments, papyri and models. In saying this I do not yet touch the question whether the genre as such arose in Egypt or in Rome; I ask only whether the pictures corresponded to reality.”

Note 2 on p. 60: “The mosaic has so far nowhere been reproduced mechanically, although good photographs by Alinari are on sale. The best drawings are those published by D. Sante Pieralisi, Osservazioni sul mosaico di Palestrina, Rome 1858; cf. A. de Barthélemy, Explication de la mosaïque de Palestrine, Paris 1760, with 2 pls. (Mém. de l’Acad. d. Inscr. XXX, p. 503). Most recently Marucchi, Bull. com. 1904, p. 233.”

3. The buildings of the Palestrina mosaic (pp. 60–62, figs. 34–35)

“I shall not dwell on all the buildings of the Palestrina mosaic, since most of them do not occur in our landscapes and could not be doubted as genuinely Egyptian even if they did: I mean the great temple, the two smaller temples, the reed arbours with the drinkers, the boat with the cabin, the great pavilion (fig. 35, facing p. 61) under which the Roman soldiers feast with their laurel-crowned commander, who has been presented by the goddess with a victory fillet (beardless, therefore not Hadrian!). Most amusing is the ibiōn or ibiōn trophē [ἰβιών, ἰβίων τροφή, the place where ibises were kept and fed], now so often named in the papyri (cf. Otto, Priester und Tempel I, p. 268, 2; 416; II, 329), to the left of the great temple in the upper zone.

Of particular interest to us, however, are two complexes. On the right of the middle zone (fig. 34, facing p. 60) we see a columned building with a flat concave roof [flach überwölbt], open on two sides and closed on the other two by low parapets. A procession passes through this building; in front of it stands a statue of Anubis, represented as a jackal, on a base. From the left front of this pavilion runs a wall, behind which is a large garden. This wall connects our pavilion with the pylon tower already known to us, which stands right at the water (fig. 35, facing p. 61). It has a monumental entrance, two pilasters with a cornice. Two low walls, ending in antae, project in front of the door; between them stands a water basin on a column-like foot, and on the left a palm branch. On the tower stands a large vessel; inside the tower stands a bearded man. Still more interesting is the complex at lower left (fig. 34, facing p. 60), unfortunately only partly executed. On the right is a tall, three-storeyed, tower-like building. The upper storeys each have three windows. From this tower-like building a wall runs to the left; it is pierced by a monumental gate and then continues until it meets a dovecote. Behind the wall and the tower grow trees: a mighty palm and a broad-leaved tree. Out of the door steps a man with a pileus on his head; in front an old woman is busy with an object that stands inside a reed enclosure. There is no doubt: here again we have the complex ‘house with garden’, this time characterised as a farmhouse.”

Note 1 on p. 61: “The building itself recalls a series of halls and dining rooms with mosaic floors recently excavated on an island in Lake Mareotis. According to the excavation report the building was simply a pavilion for dining and festivities with several rooms and a colonnade; see Bull. de la Soc. arch. d’Alexandrie IV (1902), pp. 79 f. and pl. 3.”

Captions of figs. 34–35: “Barberini mosaic at Praeneste, detail (Alinari).”

4. The buildings and Egyptian reality (pp. 62–65, summary)

Rostovtzeff opens: “And now, how do these pictures relate to Egyptian reality? Unfortunately, of the superstructures of Egyptian ruins of the Ptolemaic and Roman period, apart from the temples, practically nothing has survived.” In summary: the window beside the anta in the pronaos of the small temples in the pictures is a feature of Ptolemaic Egyptian temples unknown elsewhere in the Greek world. For the other buildings he cites the superstructures of Graeco-Roman tombs found at Antinoe and Alexandria (a round building, sometimes surrounded by columns, or a square plan with a porch), stone grave altars and their clay copies, and the clay lamp-houses placed in tombs (Lichthäuschen, after S. Loeschcke), which take the form of a tower with a broad door below and windows in the upper storey, or of a round building with a conical roof. On p. 65: “These buildings lead us to the pavilion described above on the mosaic at Palestrina, and to a building represented on a fragment of gold glass (of pre-Roman or early Roman date)” (fig. 38, Golenishchev collection, Moscow).

5. The Egyptian country house, the dovecote and the tower (pp. 69–71)

“Are all the buildings that correspond to our models and appear in the landscape representations also really dwelling houses? Perhaps not in every case. For, first, the tomb superstructure often reproduces the form of the house, as for instance in Lycia; second, a dwelling house, especially a country house (which all our houses must be, since we are in the countryside), also includes certain other things, which first characterise the building as a dwelling. What these are, the landscapes teach us: a courtyard surrounded by walls, with some trees and a monumental gate as entrance. The pygmy landscape adds a tower, and the mosaic of Palestrina a dovecote.

Are these also components, and necessary ones, of a Ptolemaic-Roman country house? The evidence of the papyri answers yes. I summarise it here very briefly. The Ptolemaic house consists of two parts, oikia and aulē; the aulē is enclosed (περίβολος, P. Petr. III 20, 30, 3 f.), and a two-storeyed tower is also mentioned (ibid., ll. 9–12). The Roman papyri add much that is new. A great role is played in them by the gate, the pylōn, just as the door in general is particularly valued in houses (P. Oxy. I 104, 25 ff.; P. Lond. III 233, 8–10, cf. 268; P. Tebt. II 331, 8). The gate is still today the main thing in Nubian houses, where the courtyard plays the principal role: the finer and more ornate the gate, the richer the owner. A garden by the house is also mentioned (Aeg. Urk. 445). Mentions of dovecotes are commonplace. The upper storey occurs often; it is the hyperōon of the papyri, which is frequently the subject of transactions (P. Flor. 5, 8; Aeg. Urk. III 999 and elsewhere; Weiss, Arch. f. Pap. IV 337). Quite fittingly, the well-lit hyperōa are mostly bedrooms (koitōnes) or dining rooms (symposia).

I should like to draw attention to one more point. The tower is occasionally mentioned in the papyri not in connection with the house but independently, as the dwelling of an owner who does not live in the village, and then often in connection with a walled olive grove (BGU 298, 5; 650, 7; 889; 740). Such free-standing towers with an enclosure, in which trees sometimes grow, are also known in our landscapes. I refer to the complex at upper left in the landscape described under no. 1, and to the tower and the walled rectangle of the mosaic of Palestrina.

What, then, is the result of these considerations? Almost all the architectural types that occur in the pygmy landscapes have turned out to be genuine buildings of Ptolemaic-Roman Egypt, mostly tomb buildings, often also dwellings, country houses. We are therefore entitled to call them Egyptian wherever these types occur. Egyptian, then, are: the pylon tower, the altar building, the two-storeyed tower with an upper row of windows, the gallery building, house and garden. The Egyptian character of the round tower with the portico remains doubtful. The form may have been in use in Egypt as elsewhere; some details, such as the windows, do however point to Egypt.”

Note 1 on p. 70: “The few excavations of Egyptian farmhouses carried through so far give the same picture as the papyri (Jouguet, BCH XXV, pp. 380 ff.; Rubensohn, JdI XX, pp. 1 ff.; Wilcken, Arch. f. Pap. II, p. 298). Here too the house consists of the oikia and the aulē, the latter surrounded by walls. Traces of a large gate building have also been found (Rubensohn, loc. cit., p. 2).”

[Summary of the rest of p. 71: the same building types also appear on second-style friezes. If the pygmy landscape were a genre created in Italy, its Egyptian buildings would have to be explained by sketches brought from Egypt; for the friezes this is very hard to assume, so such a landscape “can only be an import”. Given its Egyptian buildings, the pygmy landscape too most probably developed in Egypt.]

6. The mosaics of Sousse (Tunisia) and Palestrina, and the African villa mosaics (pp. 152–153)

“Still more instructive is the series of mosaics from Roman Africa. Already in the large Egyptianising mosaics from Sousse, which represent the flooded Delta, we notice not only Hellenistic-Egyptian motifs, as in the mosaic of Palestrina, but others as well: first all the motifs of the non-Egyptianising sacral landscape (sacred gate, pavilion with tent roof and a statue beneath it, round enclosure with tree and column, combined with the thoroughly Alexandrian tomb tower, gallery building), together with the usual figures: people sacrificing, fishing, travelling, reclining and drinking, men and women; then also the villa representations, with some peculiarities that recall the African villas depicted later, such as the many horses and the like. [He dates the Sousse mosaics probably to the first century AD.]

[…] There can be no doubt that they [the African villa mosaics of the second to fourth centuries] reproduce the type of the African villa; whether in each of the cases listed above an individual portrait of a particular estate is also present remains uncertain. In any case each villa is treated individually; how far the template [Schablone] nevertheless played a part will only be decided on the basis of more material.”

Note 3 on p. 152: two mosaics, one then at Sousse, the other in Tunis in the Bardo Museum (Gauckler and Gouvet, Musée de Sousse, p. 25 no. 1, pl. VIII; Schulten, Arch. Anz. 1900, pp. 67 f., figs. 2–3).

The encyclopedia is in an ongoing process of writing, research and updating · Last updated: