Questions about the possibility of temples with an odd number of columns on the main facade in the Northern Black Sea region and their identification with specific objects are not new in Russian historiography. Interest in this topic is caused by the fact that, as is known, an odd number of columns on the main facade in ancient Greek architecture is extremely rare and then only at the earliest time(1).

The basis for the development of this topic in relation to the Northern Black Sea region was found on the Bosporus coins of Kotis I and Eunice, issued in the late 60s of the first century AD (68/69)(2) (Fig. 1), as well as a tombstone stele with a pedimented portico (Fig. 2), dated variously from the second century. B.C.(3) to the end of the first and beginning of the second century A.D.(4) All these monuments depict the facade of a five-columned church or, as V. D. Blavatsky more cautiously noted(5), a portico. An additional, but extremely important and, as it seemed, indisputable argument in favor of the existence, in particular on the Bosporus, near the turn of the era of temples with an odd number of columns was the discovery of a fragment of the Doric architrave with a dedicatory inscription to King Aspurgus (6).

Returning to the topic of temples with an odd number of columns in this article is due to the need to take into account materials that previously practically did not attract the attention of researchers in this regard. This refers to ceramic molds from the layers of the V-IV centuries BC for casting pendants with images of the facades, obviously, of a temple with an odd number of columns, found in Olbia. Currently out of excavation

(1) For example, in the temple of Hera of the second half of the eighth century BC in Samos, in the temple of A of the seventh century BC in Prinia in Crete, in the temple of Apollo around 625 BC in Thermos, in the so-called basilica of the first half of the sixth century BC in Poseidonia, in the Greek Archaic the temple of Pompeii.

(2) Zograf A. N. Antique coins / / MIA. 1951. 16. p. 202 el. Tables 47, 4, /7; Kiryn konsky P. O. Bosporus and Rome according to numismatic data / / VDI. 1953. N 3. pp. 179-190; Frolova N. A. Copper coins of Kotis I as a historical source / / SA. 1976. N 3. pp. 103, 111; Anokhin V.A. Coinage of the Bosporus. Kiev, 1986. p. 98 el. Table 14, 370-372; Nazarov V. V. Architectural structures on Bosporus coins / / Antique culture of the Northern Black Sea region in the first centuries of our era. Kiev, 1986. pp. 153-163.

(3) Kieseritzky G., Watzinger S. Griechische Grabreliefs aus Siidrupland. T. 2. V., 1909. S. 407; Blinatsky V. D. Antique archeology of the Northern Black Sea region. Moscow, 1961. P. 155; Ivanova A. P. Sculpture and painting of the Bosporus. Kiev, 1961. p. 126 el.

(4) Kharko L. P. O pyatikolonnom khrame, izobrazhennom na bosporskikh monetakh II V. [On the five-column temple depicted on Bosporan coins of the second century]. VDI. 1950. N 1.pp. 197-205; Ivanova. Uk. soch. p. 141.

(5) Blavatsky V. D. Stroitelnoe delo Pantikapey po dannym raskopok 1945-1949 i 1952-1953 gg. [Construction business of Pantikapey according to the excavations of 1945-1949 and 1952-1953].

(6) Ibid., p. 68 el.

page 144

1. Bosporan coins with the image of a five-column facade on the obverse. / - Kotis I dupondium; 2-Eunice dupondium

2. A stele depicting a five-column pedimented portico

Olvia knows at least three molds for casting such pendants: two from the excavations of B. V. Farmakovsky in 1926 with the image of the seven-column facade (7) (Fig.

(7) Фармаковський Б.В. Розкопування Ольвп р. 1926. Odessa, 1929. p. 48. Mal. 39; Furmanski A. I Livarni formi z rozkopok Olvii // AP URSR. 1958. Vol. 7. P. 56. Tables 3, 4. A. I. Furmanskaya when describing the published foundry form cites its inventory number (A-755) and, according to the author, the number of the foundry form is determined by the author.apparently erroneously, indicates that this form has already been published by B. V. Farmakovsky. However, a comparison of the forms published by A. I. Furmanskaya and B. V. Farmakovsky suggests that these are different finds (some differences are observed in the proportions of the main volume), and have different inventory numbers (A-1545 for B. V. Farmakovsky and A-755 for A. I. Furmanskaya). Unfortunately, to examine these forms in kind

page 145

3. Casting of the pendant with the image of the seven-column facade from the mold stored in the Nikolaev Museum of Local Lore

one of the excavations of N. A. Leypunskaya in the early 70s with the image of a five-column facade (Fig. 4). Both in the first and second cases, the images of facades are inscribed in the pedimented antlered portico framing them. The production of molds from amphora pens and the rather rough drawing of images give every reason to believe that the molds were made on site - in Olbia.

The findings of ceramic forms suggest that much earlier than on the Bosporus, the appearance of the image of a temple with an odd number of columns along the facade and its manifestation in a completely different place in the Northern Black Sea region significantly weakens the position of supporters of the hypothesis that the prototype of such a structure was located on the Bosporus. Moreover, the Olvian finds give rise to the question: is it the same temple that served as the basis for all the images listed above?

It seems that there are some more or less substantial grounds for a positive answer to this question. This is primarily a graphic similarity of images on Bosporan coins and Olvian pendants, and in the main features: the presence of an odd number of columns in both places; a high pediment; clearly defined corner and central acroteria, emphasized by the shape in which the facade of the temple is inscribed. The difference from the coins is the absence of the image of a stepped stereobat on the pendants and in two cases not five, but seven columns. These discrepancies, however, can be explained in the first case by the much smaller (almost one and a half times smaller) scale of the image of the temple on pendants compared to coins, and in the second - by the fact that the masters did not see the temple in nature.

Of course, pendants and coins do not have the same historical value. But nevertheless, the discovery of such images on the pendants of a much earlier time, and not in a single copy (as several forms indicate), gives reason to assume that the temple depicted on the pendants must have had a prototype in nature.

Obviously, such a prototype could hardly be any ordinary structure. Apparently, in this case, we should also exclude the possibility that the pendants depicted one of the Olbia temples. None of the excavated Olbia temples of the V-IV centuries BC (and these are the temple of Apollo Medic of the late VI-V century BC in western Temenos (9) and the temple of Apollo Delphinius of the V century BC in central Temenos(10)) there is not the slightest reason to assume the possibility of having five or

failed. According to the kind report of Y. V. Domansky, there are no such forms from the excavations of 1926 in the Hermitage. However, a similar form from the excavations of 1926 is kept in the Nikolaev Museum of Local Lore, where it has a completely different number-A-1499. It is possible that this is the third such find.

(8) Лейпунська Н.О. Ливарнi форми з Ольвii// Archeolopya. 1984. N 45. P. 69.

(9) Kryzhitsky S.D. The Temple of Apollo letros on the Western Temenos at Olbia (an Attempt at Reconstruction) // Ancient Civilizations from Scythia to Siberia. 1997. V. 4. N 1. P. 15-34.

(10) Karasev A. N. Monumental monuments of Olvi temenos//Olvia. Temenos and Agora, Moscow-L., 1964. p. 51 el.; Pichikyan I. R. Asia Minor - the Northern Black Sea Region. Antique traditions and Influences, Moscow, 1984, p. 178, Fig. 58.

page 146

Fig. 4. 1 - mold for casting a pendant with the image of a five-column facade from the excavations of Olbia in the late 70s; 2-casting from this mold

especially seven columns on the main facade. Naturally, having a real-life model right in front of your eyes, the master who made the form would hardly have fantasized, especially in such a striking feature as an even or odd number of columns.

Most likely, the sample that served as a prototype for the mentioned forms was located in some important sanctuary for Olviopolites. This could be primarily the sanctuary of Achilles on Leucus. We don't know anything about the type of temple in this sanctuary. The archaic architectural terracotta found on Levka, similar to that found in Istria, proves the existence of a temple on Levka already in archaic times." But it is possible that this early temple could have had a non-standard spatial planning solution. Such examples are known, and nearby-in Istria. Judging by the remains discovered here, the temple of Aphrodite had an aditon, and the temple of Zeus Polyeus-an intermediate room between naos and pronaos (12). It is possible that the early temple on Levka may have had certain spatial planning features - in this case, five or seven columns along the main facade - and, due to its significance, served as a prototype not only for Olvian pendants, but also for later images listed above.

In this regard, we note that there is no complete identity between the images of the temple on Panticapaean coins. Moreover, it is possible that these coins depict different temples, one of the Ionic order and the other of the Doric order (13). This, however, seems extremely doubtful, since both in one and the other cases of

(11) Okhotnikov S. B., Ostroverkhoe A. S. The Sanctuary of Achilles on the island of Levke (Zmeinoe). Kiev, 1993. p. 25. .

(12) Alexandrescu P. Histria in archaischer Zeit // Xenia. Konstanz, 1990. Ht 25. S. 47-101; Dupont P. Commentaire des documents exposees // Istros. Les Grecs au pays des Getes. Presence archeologique francaise en terre Roumaine. Chateau des Allymes. 1996. Fig. 4.

(13) Nazarov. U k op. p. 158 el.

page 147

in other versions, the temples have common generalizing features - a five-stage stereobath and a five-column portico, a high pediment and clearly defined corner and central acroteria. Both are quite unusual for the architecture of temples around the turn of our era and, in our opinion, clearly indicates that both versions of coins show the same object. The only difference is in the presence on one of the images, apparently, of column bases and lighter proportions, in particular an entablature. In this regard, we note that the image of the order, which presumably allows us to separate these options, is not clear enough due to the small scale and primitive nature of the drawing, not to mention the fact that the possible presence of bases may indicate in favor of not only the Ionian order, but also the Tuscan, i.e., in fact Doric. In addition, the stepped stereobat is also not typical for the Ionic order.

From the above, it is possible to draw the following conclusions. Monetarians who made the image of the temple on coins probably did not see this temple in nature, otherwise there would have been no discrepancies in the display of the order. There is no sufficient reason to assume that in this case it could have been two different temples, especially given the chronological proximity of the coins. This, in turn, makes it very likely that the temple was located outside the Bosporus. In other words, the prototype for the image on the Bosporan coins of the temple, according to most researchers, Jupiter Capitoline, could hardly serve as a local (14) - Bosporan temple. In turn, the presence of five columns and a stepped stereobath, rather than a podium, contradicts the reality that the prototype could serve as a Roman(15) temple of Jupiter Capitoline. In other words, we have much more reason to assume that coins (as well as pendants) show the same object in different versions than vice versa. And in general, if we assume that due to these differences in all images (both on coins and on pendants), we are dealing with four different objects, we will have to admit that temples with an odd number of columns on the main facade were widely distributed in the Northern Black Sea region. The extreme improbability of such an assumption hardly needs to be proved.

The results of the analysis of the above-mentioned finds showed that it is advisable to once again pay attention to the block of the architrave with the dedication to Tsar Aspurg, on the basis of which V. D. Blavatsky proposed the reconstruction of the five-column facade of the church (Fig. 5). This led to some rather unexpected conclusions.

First of all, it should be noted that the author of the reconstruction did not proceed from the fact that the architrave fragment had any specific design features that indicate the presence of an odd number of columns (there are no such features), but from the restoration of the inscription proposed by him and the analogies mentioned at the beginning of the article. The core of the justification for the reconstruction was practically only the restoration of the inscription, which in the future did not cause any fundamental objections from the publishers of the Corpus regarding the number of additional letters in the first line of the inscription [16]. But it is precisely in this case that a problem arises, the solution of which radically changes the conclusion about the possibility of having an even or odd number of columns.

The point is that the vertical axis of symmetry of the reconstructed inscription passes through the center of the third (from the left end of the architrave) metope (Figs. 6, 7), and not the fourth triglyph. In other words, in this case, it turns out that above the central column

(14) Blavatsky V. D., Pantikapey. Ocherki istorii stolitsy Bospora [Essays on the History of the Capital of the Bosporus], Moscow, 1964, p. 162; Pichikyan I. R. Antique order architecture of the Northern Black Sea region of the first centuries of our Era. Uk. soch. p. 110.

(15) Zograf. Uk. op. p. 202; Karyshkovsky. Uk. soch. p. 188; Tsvetaeva G. A. Bosporus and Rome. Moscow, 1979. p. 69.

(16) See KBN. N 39. Yu. G. Vinogradov agrees with approximately this number of letters (or even slightly less) of the restored inscription, as he kindly informed me about it.

page 148

Fig. 5. 1-facade of the church according to the reconstruction of V. D. Blavatsky; 2 - drawing of the inscription according to the photo published by V. D. Blavatsky

6. 1-reconstruction of the facade of the architrave, made in accordance with the restoration of the inscription proposed by V. D. Blavatsky; 2-reconstruction of the facade of the architrave, made based on the assumption that there are an odd number of columns in the portico

there will be a metope, and not a triglyph, as it should be, given the entire experience of ancient Greek architecture; Vitruvius also writes about the need to place triglyphs above the columns (IV. II. 4; III. 2), That is, if we follow the assumption that there are an odd number of columns in the portico, then we should abandon the proposed restoration of the inscription, since in it, to reconstruct the most reliably restored first line with an odd number of columns, an additional two dozen letters will be required (even taking into account the uneven width of letters and their arrangement). The latter, given the specific historical circumstances surrounding Aspurg's personality, is hardly possible. Such a large increase in the number of letters is explained by the fact that if the axis of symmetry moves under

page 149

the center of the fourth triglyph (Fig. 6, 2), then the inscription in the un-preserved part should end in the middle not of the fifth, but of the sixth metope. The latter, contrary to the actual reconstruction of the inscription, is erroneously shown in the drawing of the temple facade (Figs. 5, 7).

Thus, if we proceed from the correctness of the restoration of the inscription by V. D. Blavatsky, the length of the architrave will be approximately 256 cm, and not 3 m, as shown in the above-mentioned drawing of the facade of the temple (17). Naturally, in this case, the possibility that the block of the architrave with the inscription could consist of two parts, and the portico, respectively, could have an odd number of columns, is removed, since the junction of two adjacent blocks of the architrave would fall on the metope (18).

All this indicates that the Panticapaean architrave could not possibly belong to a portico or a temple with an odd number of columns. Moreover, familiarity with this monument in kind (19) gives grounds to reject the possibility of linking this architrave with an ordinary columned portico at all.

In this regard, we will briefly discuss the main features of this detail, since its publisher limited himself mainly to specifying only the general dimensions and did not mention some details that are important for solving the question of the location of the architrave in the structure of the structure to which it belonged(20).

Of particular interest are the presence and location of pyrons in the left part of the architrave (Figs. 7, 2, 3), in the holes for the installation of which you can see the remains of iron rods sunk in lead. In this regard, we note that the architrave in cross-section is not symmetrical, its rear facade has rough processing, and the thickness of the part is only 17.5 cm. In other words, if it were located in a columned portico, its thickness would have to be twice as large or it would have to consist of two parallel blocks. Meanwhile, a pyron on the lower plane of the architrave and one on the upper plane, judging by the direction of the grooves,

(17) Blavatsky. Construction business of Panticapaeum... p. 70.

(l8) This raises the question whether, in principle, such a marble architrave can withstand a span of about 2.16 m in length in purity (assuming that 0.2 m from each end of the architrave falls on the abacus of the capitals). Unfortunately, the strength characteristics of this architrave are unknown to us (to determine them, it would be necessary to conduct tests before the sample is destroyed). The variation of their values in comparison with the standard values of marble strength taken into account in calculations can be very significant. Based on this, and understanding the precariousness of such an approach, it is still possible to turn to analogies. In this regard, it should be noted that the architrave of the Temple of Aphrodite in Chersonesos (Pichikyan), made in one block with the frieze, had similar dimensions (height 0.32 m with a length of 2.09 m). Uk. soch. p. 249). About 2.5 m in the axes had the spans of the portico surrounding the sanctuary of Athena in Pergamum, with the height of the architrave about 0.35 m (Universal History of Architecture, Moscow, 1949. Vol. 2. Book 1. Fig. 281). Thus, in principle, since the working conditions of the architrave in the construction of the entablature are to a certain extent close to the beam with pinched ends, a span of 2.16 m in size is obviously possible with an architrave height of 0.31 m, although it is certainly at the limit.

(19) I would like to take this opportunity to sincerely thank V. P. Tolstikov for his assistance in examining and measuring the architrave currently located in the Pushkin State Museum of Fine Arts.

(20) The length of the block on the upper plane is 115.3 cm, on the lower plane - 118.5 cm. The thickness in the lower part is 17.5 cm, in the upper part (taking into account the shade) - 22 cm. Height 31 cm. The right part of the block is broken off. The shade is partially preserved, its left edge is destroyed for 27 cm from the left end of the block. There are two regula with six gutta each. The width of regula (and, accordingly, triglyphs) is 22.8 and 23.0 cm; the width of the metope is 28.3 cm. The back side of the architrave in the upper part has a sub-rectangular projection (shelf) with a distance of about 3 cm and a height of 14-15 cm, which does not reach the left end of the architrave by about 25-28 cm. The latter feature, taking into account the presence of a hole and groove near the end of the ledge (shelf), i.e. a nest for installation (there is a similar nest on the lower plane of the architrave, but almost next to its end) of the pyron, indicates that in this place another perpendicular block adjoined the architrave. Another socket for the pyron is located on the upper plane of the architrave at its left end. It was intended for fastening to the end of the neighboring block.

page 150

Fig. 7. Dimensional drawing of the architrave. / - facade; 2-left end of the architrave; 3-plan of the architrave (fragment)

they were used to fasten the architrave with a stone detail (or parts) that approached the architrave from the back side perpendicular, and not parallel. This is supported by the absence of the above-mentioned shelf on the back of the architrave in this place. The presence of anaphyrosis on the left end of the architrave (Fig. 7,2), the nature of the location of the second pyron on the upper plane of the architrave, as well as the arrangement of triglyphs indicate that this architrave could not be angular. All this practically excludes the possibility of installing this architrave in a conventional columned portico.

This is also confirmed by the nature of the surface treatment of the architrave. Its facade is finished under the Trojan and sanded. The lower plane, which was also treated as a Trojan horse, has no traces of grinding, i.e. it was not designed for viewing from below.

Finally, we emphasize the absence on the lower plane of the architrave - at its left end-of any traces of the abacus of the capital. Apparently, V. D. Blavatsky, speaking about the presence of a trace from the capital, was mistaken in this case, since usually the presence of such a trace is expressed in a different treatment of the lower plane of the architrave adjacent to the abacus, there may be a risk, but neither of them is present in this case(21).

This raises a question, the solution of which can help narrow down the range of structures where this architrave could be located. This means determining the probable height of the architrave relative to the base of such a structure. We follow in this case Vitruvius (whose treatise appeared at the end of the first century BC, and the architrave dates back to 23 AD), who believes that the height of the architrave and the width of the triglyph should correspond to one module each (Vitr. IV. III. 4). In our case,

(21) We will note some details of the architrave processing that are not relevant for the topic under consideration, but are of interest from the point of view of marking techniques. These are, first, the marking points applied between the gutts of the regul, and, secondly, the risks for marking the inscription. In the latter case, in addition to the main figures, there are additional ones: above the first and fourth lines, as well as under the second line. Additional risks are 4-5 mm away from the main ones. Not a single letter touches additional figures. It is possible that additional risks arose earlier than the main ones and were intended for placing a different label.

page 151

the difference between the height of the architrave (31 cm) and the width of the triglyph (23 cm) is 8 cm, so it seems logical to take the first and second digits as the limit values of the module. Hence, since, according to Vitruvius (III.4), the height of the Doric column should be 14 modules, the height of the location of our architrave could most likely be in the range of 3.22-4.34 m. The latter, even taking into account the possibility of heavier proportions, makes it very unlikely to use an architrave in the altar (in the Northern Black Sea region, such monumental altars are still unknown). Its placement on the facade of the naos temple is also doubtful - such inscriptions were usually made on the main facade, i.e. the columned portico.

The considered features allow us to conclude that this architrave was used, apparently, in some kind of wall structure, either a memorial or a utilitarian structure of relatively small dimensions, but not in a colonnade, especially a pedimented portico with an odd number of columns. And this, taking into account all the above, in turn removes one of the important arguments in favor of the possibility of the existence of temples with an odd number of columns on the main facade on the Bosporus in the first centuries of our era. It seems more likely that such a temple appeared in archaic times. In the Northern Black Sea region, it was the only one located outside the Bosporus in any sanctuary of general Hellenic significance, most likely on the island of Levke.

ON THE PROBLEM OF BUILDING TEMPLES WITH ODD NUMBER OF COLUMNS ON THE MAIN FACADE IN THE NORTH BLACK SEA REGION

S.D. Kryzhitsky

The interest to the problem is due to the fact that odd number of columns on the main facade in ancient Greek architecture was a rarity even in the earliest times, though they occur among the archaic temples. The research was stimulated by the discovery of Cotys I and Eunice's coins (68-69 AD) and of a grave stele with pediment portico (2nd с. ВС - early 2nd c. AD) in Bosporus. All these finds have on them images of a five-columned temple (or a portico). A Doric architrave fragment with a dedication to Aspurgus seemed to give another important argument in favour of existence of a temple with odd number of columns in Bosporus.

Ceramic moulds of pendants with an image of a temple with odd number of columns found in 5-4th ее. ВС layers in Olbia (and obviously produced there) made the author of the article turn to this problem again. The temple on the moulds was the same as on the Bosporian finds and represented some sanctuary important for the Olbiopolitans, probably that of Achilles on the island of Leuce, which had existed there since the Archaic times.

The author believes that the Bosporian coiners who made the coins with the temple had not seen the temple itself, otherwise they would not have depicted it in two different orders. The temple must have been far from Bosporus.

A thorough architectural, constructive and epigraphic analysis of the Panticapaean architrave revealed that it could not belong to a temple with odd number of columns; in fact it cannot be associated with any usual column portico whatsoever. This being proved, the hypothesis of existence of a temple with odd number of columns in Bosporus loses one point.

It is quite reasonable to suppose that such temple could have been built in the Archaic time. It was the only temple of this kind in the North Black Sea Region and was situated in a sanctuary of pan-Hellenic importance, obviously on the Island of Leuce.

Permanent link to this publication: