Previous |Contents |Next

The Three Parthenons

The terms of the controversy changed completely in 1892 when Wilhelm Dörpfeld, by examining what had been exposed by Kavvadias’ excavations, proved that the temple of Athena destroyed by the Persians in 480 B.C. was to the north of the Parthenon and not below it, as had been assumed a priori up to that time. The immediate reaction of scholarly opinion was to reject completely Ross’ conclusion that the structures found underneath the Parthenon are the remains of an earlier temple. The only scholar who opposed this trend was Penrose who in 1891 argued that the Temple of Athena destroyed by the Persians was located where there was later erected the Parthenon. The reason for which he clung stubbornly to what had been the current view up to 1885, was his incorrect analysis of the measurements of the column drums imbedded in the north wall of the Acropolis. Since he had concluded that these column drums do not belong to the Parthenon, he had to maintain that they belong to an earlier temple similar to the Parthenon, though not to the temple uncovered just north of it. Penrose, albeit for the wrong reasons, stressed the concrete and unquestionable fact that the substructure of the Parthenon must have been erected for a different temple. Scholars were then glossing over this fact because they could not accept the existence of a second temple of Athena antedating the Persian invasion—all the texts clearly indicate that the Persians destroyed only one temple of Athena.

The archaeological argument of Penrose impressed Dörpfeld, who was most careful in matters of archaeology, when this is understood in a narrow sense, but was inclined to take a cavalier attitude with historical evidence. He was able to establish that the Parthenon designed by Iktinos was erected on a substructure which reveals the existence of two earlier different plans of the temple.1 Dörpfeld spoke of a Parthenon I and of a Parthenon II, applying the term Parthenon III to the temple with which we are familiar. He not only ascertained the existence of these two proto-Parthenons, but was able to reconstruct the dimensions of their ground plans; but these achievements have been treated with some skepticism by scholars because he followed a false lead in establishing the chronology of the proto-Parthenons and as a result got involved in difficult or insoluble questions which obscured the soundness of his contribution as far as strictly tectonic archaeology is concerned.

In an article of 1892 and in another published ten years later, Dörpfeld applied himself to the identification of the characteristics of Parthenon I.2 He concluded that one could trace three steps of the platform of Parthenon I. The lowest step was constituted by the top layer of the substructure (twenty-first layer counting from the bottom of the southeast corner of the substructure), the substructure being merely the extension downwards of this step. Another step, which Dörpfeld considered the middle step, is recessed in relation to the lowest step and the rest of the substructure. Dörpfeld thought that he could identify also a third and higher step. This step is made of Karrha limestone, a material hardier than the friable poros limestone, but less sturdy than the Pentelic marble of the Periclean Parthenon. This step is covered by the lowest step of the Periclean Parthenon (Parthenon III in Dörpfeld’s later terminology, which I follow). Today there is general agreement that inside the platform of Parthenon III there are the remains of a step of Karrha limestone; but it has been argued that Parthenon I had only two steps of poros limestone. The step of Karrha limestone would belong to another, different temple, which was identified later and was called Parthenon II by Dörpfeld.

In his article of 1892, Dörpfeld did not follow the contention of Ross and Penrose that Parthenon I had been the victim of the Persian fire of 480 B.C. He followed a different and sounder method of dating the substructure and Parthenon I. It is evident that the substructure was built in the same period as the south wall of the Acropolis. One of the functions of this wall, was is called the Cimonian Wall, was to retain the earth filling that covered the substructure and formed a terrace to the south of the Parthenon. Only the four top layers of the substructure are finished on their outer surface in a manner that indicates that they were intended to be exposed, whereas the remaining lower layers show an irregular face that obviously was intended to be buried underground. From Plutarch (Life of Cimon, XIII) we learn that the south wall was built with the spoils of the campaign conducted against the Persians by Cimon, a campaign that culminated with the victory of Eurymedon (468 B.C.). Hence, Dörpfeld concluded that the substructure and Parthenon I must be dated after 468 B.C.

Dörpfeld, by turning his attention to the steps of the platform of Parthenon I, had positively proved that the substructure was built for a temple different from Parthenon III. The archeological evidence below the crépis of the Parthenon does not point to the existence of a temple, but to the existence of a different plan of a temple. Logically, Dörpfeld should have reached the conclusion, which was reached by others much later, that the plan of the Parthenon was changed in the course of execution. This would have led to the conclusion that the substructure of the Parthenon belonged to the early stages of the construction of the Parthenon itself; but his would have also led to the conclusion that the south wall of the Acropolis, which is closely linked with the substructure, must be dated to the time the Parthenon was in its early stages of construction, and this was a conclusion considered unacceptable in Dörpfeld’s time. Ross and Penrose had met with stubborn resistance when they argued that under the Periclean Parthenon there were the remains of an older temple. The same opposition continued to operate against Dörpfeld, although he could present a stronger case. Dörpfeld felt the need to find some specific argument to prove that the substructure of the Parthenon and, hence, the proto-Parthenon existed at the time of the Persian invasion. When confronted with this irrational opposition, Dörpfeld resorted to the same desperate argument which had been used by Ross, the traces of the Persian fire. Ross had spoken of traces of fire in the column drums of Pentelic marble imbedded in the north wall and in those found scattered to the east of the Parthenon; Dörpfeld claimed that traces of fire are to be seen in the poros steps of the platform of Parthenon I. But in order to claim that these traces of the Persian fire exist, Dörpfeld had to completely alter his dating of Parthenon I. In his first article of 1892 Dörpfeld, relying on the stratigraphic data gathered in Kavvadias’ excavations, had correctly concluded that the substructure, which is an extension of the steps of Parthenon I, was built together with the Cimonian Wall, that is, after 468 B.C.; but in 1902, since he claimed that the steps of Parthenon I show traces of the fire of 480 B.C., he had to date Parthenon I before this point in time. However, he had to reconcile this dating with his discovery of the Old Temple. Dörpfeld understood that the shift from Parthenon I to Parthenon II and from Parthenon II to Parthenon III must be ascribed to some events of Athenian history but, since he had little judgment in strictly historical matters, he completely failed to identify these events. He propounded the hypothesis that after the battle of Marathon (490 B.C.) the Athenians, in order to celebrate this victory, began to erect a new temple which had reached the level of the platform when the Persians came back in 480 B.C. On this occasion the Persians would have put one fire not one, but two temples, the Old Temple and the uncompleted Parthenon I.

Once he embarked upon this line of argument, Dörpfeld had to defend his case by accumulating gratuitous suppositions. The lack of historical evidence for the destruction of this second temple was to be explained by assuming that it was under construction at the time of the Persian invasion. To explain how the Persians could have put on fire a temple which consisted merely of a stone platform—the destruction by fire of Greek temples was usually caused by the conflagration of the wooden ceiling, since stone walls and columns cannot be ignited—Dörpfeld imagined that by 480 B.C. the Athenians had completed the platform of Parthenon I and were in the process of erecting the columns. Some of the drums—those intended for the bottom of the columns—would have been partly fluted, whereas the others would have been as yet unfluted; some of the column drums, those with bosses still in place, would not have been as yet put into position, and, therefore, were at an even earlier stage of execution. This would explain why the column drums imbedded in the north wall and the similar ones found scattered throughout the Acropolis are in the vast majority of cases still in the rough, as Penrose had correctly reported. Dörpfeld contended that a wooden scaffold had been set up in order to erect the columns; this scaffold would have been put on fire by the Persians, damaging the column drums and the platform. From the measurements we know that none of the drums that have been found belong to a level of the columns higher than the third drum; hence, in order to follow Dörpfeld’s theory of the scaffolding it is necessary to assume that the scaffolding was erected when none of the drums that it was intended to lift had yet been cut. But even if all of these suppositions were to be accepted, he still had to explain how columns of Pentelic marble could have been placed on top of a platform of poros limestone. This is incredible not only for esthetic reasons, but for structural reasons as well, since poros limestone would crumble under the pressure of columns of Pentelic marble, which is a harder and heavier material. Dörpfeld thought that he could turn around this difficulty by assuming that Parthenon I had a top step of Karrha limestone, which is a material somewhat sturdier than poros limestone. The highly contrived hypothesis of Dörpfeld should have been rejected:

- because it had been proved that the substructure is contemporary with the Cimonian Wall, a fact that Dörpfeld had original accepted as evident

- because a temple of Pentelic marble would not have been erected on a platform of limestone

- because the historical record shows that the Athenians would not have started the construction of a great temple of such size and cost (Pentelic marble temple) in the period 488-480 B.C.

In that period the Athenians had no reason to replace the Old Temple, which had not yet been destroyed. The cost of the Pentelic marble Parthenon heavily taxed the resources of Athens at the peak of her power; it was made possible only by the economic exploitation of the Athenian Empire. We know of only one building constructed specifically in gratitude for the victory at Marathon—the modest structure known as the Treasury of the Athenians at Delphi. The Athenians would not have engaged in a major construction effort after the battle of Marathon, when a new attack was expected. After their disaster at Marathon the Persians, in order to recoup the prestige necessary to an imperial power, began to mobilize all the resources of the Empire for a second and decisive attack. The second attack was delayed until 480 B.C. because the preparations were extremely complex and because in the meanwhile King Darius the Great died and the new king Xerxes had to consolidate his power before assuming personal command of the greatest military effort ever undertaken by the Persian Empire. In a separate study I present precise numerical data revealing that for the second campaign the Persians doubled the army which had an official table of organization of 360,000 fighters, when fully mobilized, and doubled the fleet, which had an official first line of 600 triremes, when fully mobilized. But even without considering the numerical data which indicate that the Persian king mobilized several million people, when support troops and services are included, historians agree that the Persians began to organize lines of transport and supply drops years in advance of 480 B.C. The great miracle of history is that the Athenians did not decide to surrender when the scale of the Persian preparations became known. The Athenians decided to try their luck and invested whatever resources they had in building the fleet which turned the tables on the Persians at the battle of Salamis. The Athenians would have been utterly irresponsible if they had invested their resources in the construction of the Parthenon. As Otto Walter put it: ”There is little likelihood that after 490 B.C., when the new Persian attack was to be expected, the Athenians had the money and the leisure to start and carry through a project of such grandeur.”3

Dörpfeld was aware at least in part of all the weaknesses of his hypothesis that Parthenon I was started in the period between the two Persian campaigns. But he thought that all the objections could be removed if there is proof that Parthenon I was destroyed by the fire of 480 B.C. Hence he stated that the traces of this fire are to be seen; but although the issue is crucial, he never undertook to submit specific evidence. As to the traces of fire in the column drums, he limited himself to referring to the lines of Ross which I have quoted earlier, adding: ”These important observations can be verified by any visitor to the Acropolis.” In truth, a visitor to the Acropolis could verify, although not as easily as suggested, that the column drums are cracked, but this does not prove that they have been cracked by fire. Actually the fact that the drums in question are cracked proves that they are discarded pieces and were never part of a standing temple. Later I shall consider evidence to the effect that the cracks cannot be the result of fire, but in any case, Dörpfeld never presented any possible connection between the fissures in the drums and fire.

Dörpfeld passed quickly over the supposed trace of fire in the drums of Pentelic marble and concentrated his attention on the traces of fire in the steps of Parthenon I. Even though he asserted that his argument stands or falls on the existence of these traces of fire, he dedicated only a page to them.4 In specific terms he states only that the stylobate step of Karrha limestone is ”badly burned and damaged,” but considered decisive evidence ”the traces of fire which can still be clearly recognized” in the two poros steps below the marble steps of the present Parthenon. Dörpfeld stated that these traces of fire were a most essential proof of his reconstruction of the history of the Parthenon’s substructure. In 1908 Martin L. D’Ooge pointed out that Dörpfeld had revived Ross’ theory that the substructure had been erected before the Persian invasion and added: ”The most convincing proof of this belief Dörpfeld finds first in the marks of the fire (formerly observed also by Ross) on the marble drums and on the steps of the building.” In 1934 Dinsmoor in reviewing the entire development of scholarly opinion about the proto-Parthenon, when dealing with Dörpfeld’s contention that it was pre-Persian, stated (p. 489):

”His most important evidence was the reconstruction of the data furnished by Ross concerning the burnt Persian debris and the traces of fire on the foundation and column drums; after verifying these traces of conflagration. Dörpfeld returned to the earlier theory that the Older Parthenon had been destroyed by fire in 480 B.C.”

The truth is, however, that the alleged verification of the existence of the traces of fire had not taken place. One could expect that, concerning such an essential element of his evidence, Dörpfeld would have provided a detailed description, identifying the location, size and the appearance of the traces of fire, but he never stated where these traces are to be found and what they consist of. Instead of presenting facts he appealed to authority: ”At my request numerous experts tested and verified these important observations.” But most probably Dörpfeld was merely bluffing, because when later the existence of these traces of fire was questioned, he was not able to bring forth even one of the opinions or reports of these alleged numerous experts. He never explained what was their supposed field of expertise: were they chemists, mineralogists, fire insurance investigators, or heads of fire departments? Many years later, when he was directly challenged, Dörpfeld, instead of submitting the evidence, tried to disclaim that he had made of point of the existence of traces of fire in the poros steps.

Dörpfeld tried to gloss over the superficiality of this part of his archaeological documentation by a show of accuracy on a detail. He raised the question of how it would be possible to put on fire a temple which allegedly consisted of the stylobate platform with a few column drums on top of it, and answered it by declaring that, since the temple was under construction the columns were surrounded by a wooden scaffolding. It is amazing that for the following thirty years scholars accepted the existence of this makeshift scaffolding as supporting the entire theory.

The assertion of the traces of fire was accepted so readily by the body of scholars that those who denied their existence were saddled with the burden of submitting a negative proof. It is difficult to provide proof for the non-existence of something the location and nature of which is not defined. Only in 1934 Walther Kolbe dared to question the existence of these traces of fire: he obtained the opinion of a mineralogist to the effect that the reddish streaks that appear in the drums identified as part of the columns of the proto-Parthenon, have nothing to do with fire. In 1940 Tschira proved that these drums are rejects from the construction of the Parthenon, but as a preliminary step to his demonstration showed also that they do not reveal traces of fire; not only are the reddish streaks mentioned by Kolbe a normal feature of Pentelic marble which can be noticed already at the quarry, but it can be shown also that the drums do not reveal any of those alterations that are to be found in the marble remains of temples known to have been destroyed by fire. Other similar arguments had been presented by A. Sophianopoulos in 1938, so that it can be stated outright that the column drums of the north wall and the similar ones found throughout the Acropolis were not damaged by fire. But, even supposing that there were traces of fire damage on the steps of the Parthenon, once it is proved that there is no trace of fire in the column drums, there is no longer any argument for ascribing the alleged damage on the steps to the evens of 480 B.C., rather than to the many vicissitudes of the temple through the centuries. Such traces of fire do occur in other parts of the temple, for instance, quite clearly on the inner side of the western pediment; these particular traces prove that Pentelic marble exposed to fire turns gray, and not red.

But there remained to be explained the shift from Parthenon I to Parthenon II; Dörpfeld concluded that this shift had to be connected in some way with the end of the rule of the tyrants and the establishment of democracy, dating thereby Parthenon I around 506 B.C., so that only the period between the battle of Marathon (490 B.C.) and the battle of Salamis (479 B.C.) remained available for Parthenon II. In 1934 W. Zschietzmann carried Dörpfeld’s research to its logical conclusion: Parthenon I must have been initiated by the tyrant Peisistratos (Zschietchmann chose the date of 540 B.C.) and have been replaced by Parthenon II after the expulsion of the tyrants and the establishment of democracy. This chronology is the only reasonable one if one follows Dörpfeld’s approach, but it shows the weakness of this approach, since it leads to an impossibly high date for Parthenon I. This is the reason why the details of Dörpfeld’s chronology remained vague and obscure.

The error of Dörpfeld’s second position, by which the substructure was separated chronologically from the Cimonian Wall, was soon exposed by the publication in 1906 of the report of the excavations conducted in 1885-1889 by Kavvadias and Kawerau. These excavations, which removed the entire earth filling between the substructure and the Cimonian Wall, established with certainty that the strata in front of the substructure are contemporary with it. The method used in erecting the substructure consisted in placing one layer of blocks in position at a time and then piling up earth in front of it in order to raise the level of the ground to the height of the next layer. In the earth filling one can distinguish the several strata of earth covered with chips of poros from the cutting of the corresponding layer of the substructure. The strata of earth filling that were gradually piled up in front of the substructure were retained by the Cimonian Wall, except for the very lowest strata. The very lowest strata were retained by a wall to which scholars have given the name S2, meaning that it was the second wall on the south side. This wall ran in front of the substructure, almost parallel to it, at a distance of more than 10 meters at the southwest corner and more than 12 meters at the southeast corner. The wall S2 substituted for the Pelasgic Wall in the area south of the future Parthenon. Whereas the Pelasgic Wall followed an irregular curved course, the wall S2 straightened this course more to the inside, joining with the Pelasgic Wall in front of the southwest corner of the Parthenon. From the manner in which the new wall joined the Pelasgic Wall it can be established that at the time the wall S2 was built the Pelasgic Wall had been violently thrown down. It may have been thrown down either by the Persians or by the Athenians themselves after the failure of the attempted oligarchic coup.

Whereas the filling piled up behind the Cimonian Wall was loose earth, the filling piled up behind the wall S2 was of a more solid nature. It consisted of remains of constructions destroyed by the Persians. The excavators, Kavvadias and Kawerau, reported that the strata between wall S2 and the substructure contained ”fragments of building blocks and sculptures of poros” or, in other words, ”remains of destroyed buildings, votive offerings, and other sculptured works.” They further remarked:

In the masses of earth there wee found, besides many sherds of black-figured vases and even older vases, a greater number of sherds of red-figured vases. The first group of findings indicates that the fillings date after the destruction of many ancient buildings and statues, whereas the fragments of red-figured vases demonstrate that the mentioned destruction occurred at the earliest at the end of the sixth or at the beginning of the fifth century. Therefore, it must be considered a certain fact that here we are dealing with the so-called ”Persian rubble,” and that, accordingly, the piling up of the filling and with it the erection of the substructure took place in the first half of the fifth century.

The authors of the report could not have been more explicit in dating the substructure after 480 B.C. But they did not feel that it was their task to enter into polemics with Dörpfeld who had dated it before 480 B.C., most probably because polemics did not belong in an architectural report which aimed at being a strictly factual presentation of data. Hence, after listing all the empirical data according to which the substructure must be dated after the Persian destruction, Kavvadias and Kawerau added a paragraph about the theory of Dörpfeld, pointing out that his theory depends exclusively on the contention established ”through his own observation” that ”there are really strong traces of fire on the steps of the older Parthenon.” Diplomatically they contended that ”if this were to be proved, one would have to agree with Dörpfeld.”!

Kavvadias and Kawerau were leaving to other scholars the task of deciding whether the traces of fire allegedly seen by Dörpfeld really existed; by implication they stated that they had not seen them. But other archaeologists did not take up the challenge and preferred to continue to accept the figment of the traces of fire which had become part of the folklore of Acropolis archaeology.

The excavators had insisted that the filling between the wall S2 and the substructure was made up of ”Persian rubble” (Perserschutt in German), that is, of the fragments of buildings and sculptures destroyed by the Persians in 480 B.C. The vase sherds found within the filling confirmed that the rubble came from constructions which had been standing up to 480 B.C. As to the purpose of the wall S2, the excavators concluded that it had been erected as a provisory retaining wall designed to retain the filling while the Cimonian Wall was being erected. I propose a slightly different explanation for the wall S2, even though I agree that it was erected somewhat earlier than the Cimonian Wall. The wall S2, behind which there was piled a filling of heavy stone rubble instead of loose earth, served two purposes: it reinforced the foundations of the substructure and, at the same time, it reduced the pressure of sliding materials against the Cimonian Wall.

When Dörpfeld developed his theory he believed that the proto-Parthenon had a crépis that consisted of the two layers of the poros substructure. A new and important contribution to the history of the Parthenon was made by the American archaeologist B. H. Hill when in April-May 1910 he discovered, inside the marble steps of the Periclean Parthenon, the steps of another temple, smaller in size. According to him the step of Karrha limestone that Dörpfeld had taken as the highest step of Parthenon I was in reality the lowest of the three steps of another temple (which I shall call Parthenon II, following Dörpfeld’s terminology). The three steps of Parthenon II are covered by the three steps of Parthenon III. Parthenon II is much narrower than Parthenon III, but almost the same length. In relation to Parthenon I, Parthenon II is narrower and shorter, but its platform is exactly centered in relation to the upper edge of the substructure (lowest step of Parthenon I). In other words, Parthenon II is recessed in relation to the substructure by an equal space on the two flanks and by an equal space on the two fronts. The key discovery of Hill was that Parthenon II is so small in relation to the substructure that this must have been planned for another temple, namely, Parthenon I: ”That the great podium had not originally been designed to receive so small a temple seems certain...” (p. 556).

As soon as Hill communicated his findings at the December 1910 meeting of the Archaeological Institute of America, Dörpfeld revised his position. Dörpfeld claimed that even earlier he had been inclined to recognize the existence of two proto-Parthenons: One (Parthenon I) for which the substructure had been planned; and the other (Parthenon II), that discovered by Hill. By the time Hill published a preliminary report of his survey in the American Journal of Archaeology of 1912, he and Dörpfeld had reached substantial agreement. If there was any significant disagreement between the two, it concerned the steps of the Parthenon. According to Hill Parthenon I had a platform with only two steps of poros limestone, whereas Parthenon II had a platform with three steps, of which the lowest was of Karrha limestone and the other two were of Pentelic marble. According to Dörpfeld, Parthenon I had a platform with only two steps of poros limestone, whereas Parthenon II had a platform with three steps, of which the lowest was of Karrha limestone and the other two were of Pentelic marble. According to Dörpfeld, Parthenon I had a crepis of two steps of poros limestone with a stylobate of Karrha limestone, which is a material harder than poros limestone, whereas the step of Karrha limestone became the bottom step of the crépis of the second proto-Parthenon. The blocks of Karrha limestone would have been later moved and in some cases recut to be used as the lowest step of Parthenon II. This was reasonable, since it was hardly credible that columns of Pentelic marble would have been placed on a stylobate of poros. Hill did not exclude completely the possibility that Dörpfeld’s interpretation was the right one. According to Dörpfeld it was the second proto-Parthenon that was destroyed by the Persians. But he never felt the need to review his contention about the traces of fire, although there are no traces of fire in the Karrha step and on the marble stylobate blocks of the second proto-Parthenon.

As in other cases, Dörpfeld proved sound in matters of tectonic archaeology, but weak in matters of historical reconstruction. Having established that there had been three Parthenons, he proceeded in a heedless manner when it came to dating them. In 1882 and 1892 he had dated the substructure and Parthenon I in the Cimonian age; then, in 1902, he had dated them before 480 B.C., because of the supposed traces of the Persian fire, but had dated them as late as possible because of the chronological datum provided by the Cimonian Wall. Now, in 1912, when confronted with the existence of Parthenon II, he tried to solve the chronological difficulty merely by shifting the traces of fire from the steps of Parthenon I to the steps of Parthenon II. By this legerdemain he transferred the assumed date of Parthenon I to Parthenon II. It would have been Parthenon II that was under construction in 480 B.C. and was covered by a makeshift wooden scaffolding which would have been put on fire by the Persians. Dörpfeld never explained how the traces of fire, which he had claimed could be clearly seen on the steps of Parthenon I, now should be seen on the steps of Parthenon II. It is a fact that Hill, who discovered and described Parthenon II, did not mention any traces of fire on the remains of this structure. Since the date previously ascribed to Parthenon I was ascribed to Parthenon II, the former had to be dated earlier. Parthenon I would have been started in the age of the tyrants and left unfinished at their fall. Dörpfeld never truly explained how this new date of the substructure could be reconciled with the date of the Cimonian Wall and with the arrangement of the strata between this wall and the substructure.


  1. Penrose, however, contributed to later confusion of thought by referring to the substructure by the word “subbasement,” which is a barbaric rendering into English of the word subbaisement used by French archaeologists, a word which in French is the equivalent of the English “foundation.”
  2. Athenische Mittheilungen, 17 (1892), 188-189; Ibid., 27 (1902).
  3. Athenische Akropolis, pp. 68-72.
  4. Athenische Mittheilungen 27 (1902), 494.

Previous |Contents |Next