1. Distribuitur autem longitudo aedis, uti latitudo sit longitudinis dimidiae partis, ipsaque cella parte quarta longior sit, quam est latitudo, cum pariete qui paries valvarum habuerit conlocationem. Reliquae tres partes pronai ad antas parietum procurrant, quae antae columnarum crassitudinem habere debent. Et si aedes erit latitudine maior quam pedes XX, duae columnae inter duas antas interponantur, quae disiungant pteromatos et pronai spatium. Item intercolumnia tria quae erunt inter antas et columnas, pluteis marmoreis sive ex intestino opere factis intercludantur, ita uti fores habeant, per quas itinera pronao fiant.
1. The length of the temple is so arranged that the breadth is half the length. The cella itself is to be a fourth part longer than its breadth, including the wall which contains the doors. The remaining three parts, that is, the portico, are to run forward to the antae of the walls. The antae ought to have the thickness of the columuns. If the temple be more than 20 feet in breadth, between the two antae two columns are to be placed and these columns are to separate the space of the pteroma and the pronaos. Between the three intercolumniations, which will come between the antae and the columns, let there be a fence of marble or of fine joinery with gates by which the portico may be entered.
2. Item si maior erit latitudo quam pedes XL, columnae contra regiones columnarum, quae inter antas sunt introrsus conlocentur. Et hae altitudinem habeant aeque quam quae sunt in fronte, crassitudines autem earum extenuentur his rationibus, uti, si octava parte erunt quae sunt in fronte, hae fiant X parte, sin autem VIIII aut decima, pro rata parte. In concluso enim aere si quae extenuatae erunt, non discernentur. Sin autem videbuntur graciliores, cum exterioribus fuerint striae (XX aut) XXIIII, in his faciendae erunt XXVII aut XXXII. Ita quod detrahitur de corpore scapi, striarum numero adiecto adaugebitur ratione, quo minus videtur, et ita exaequabitur dispari ratione columnarum crassitudo.
2. Also if the breadth be more than 40 feet, columns are to be placed towards the inner part, in a direct line with those which are between the pilasters. These other columns are to have the same height as those in front, but their diameters are to be lessened in the following manner: if the diameter of the front columns in one eight of their height, the side diameters are to be one tenth; if the front diameter is one ninth or one tenth, then proportionately. For being diminished in an enclosed space, they will not be remarked. But if they should seem too slender they may have 28 or 32 flutes against the outside 20 or 24. Thus what is taken from the diameter of the shaft will be added by the extra number of the flutes, so as not be observed. In this way the varying diameter of the columns will be balanced.
3. Hoc autem efficit ea ratio, quod oculus plura et crebriora signa tangendo maiore visus circuitione pervagatur. Namque si duae columnae aeque crassae lineis circummetientur, e quibus una sit non striata, altera striata, et circa striglium cava et angulos striarum linea corpora tangat,quae circumdatae erunt, (non erunt) aequales, quod striarum et striglium circuitus maiorem efficit lineae longitudinem. Sin autem hoc ita videbitur, non est alienum in angustis locis et in concluso spatio graciliores columnarum symmetrias in opere constituere, cum habeamus adiutricem striatarum temperaturam.
3. This effect is produced for the following reason. The eye thus touches a greater number of points, and ranges over a larger circumference of vision. For if two columns of equal diameter, of which one is fluted and the other is not, have a line measured round them and one line touches the shafts of the columns round the flutes and their fillets, although the columns are of equal diameter, the bounding lines will not be equal because the circuit of the fillets and the flutes produces a greater length of line. Now if this shall seem to be the case, it is not inappropriate, in narrow places and a confined space, to use in building more slender proportions for the columns, since we have the adjustment of the fluting to help us.
4. Ipsius autem cellae parietum crassitudinem pro rata parte magnitudinis fieri oportet, dum antae eorum crassitudinibus columnarum sint aequales. Et si extructi futuri sunt, quam minutissimis caementis struantur, sin autem quadrato saxo aut marmore, maxime modicis paribusque videtur esse faciundum, quod media coagmenta medii lapides continentes firmiorem facient omnis operis perfectionem. Item circum coagmenta et cubilia eminentes expressiones graphicoteran efficient in aspectu delectationem.
4. The thickness of the walls of the cella itself ought to be proportionate to its dimensions, provided that the antae (from pilasters) are equal to the diameter of the columns. If the walls are constructed (of common stone) they must be built of very small stones, but if the walls are built of square stone or marble, the pieces should be of moderate and equal size. For the middle of the stones in the course above will bind together the joints below them and will strengthen the execution of the whole work. Further, the raised pointing about the upright and bed joints will produce a more picturesque effect in the general view.
This fragment is difficult and puzzling. Vitruvius explains here the proportions of the cella of a temple.
The first sentence Distribuitur autem longitudo aedis, uti latitudo sit longitudinis dimidiae partis (The length of the temple is so arranged that the breadth is half the length) refers to the general proportions of the whole temple as is explained in book III, chapter 4: the breadth is half the length of the temple: Sed ita columnae in peripteris conlocentur, uti, quot intercolumnia sunt in fronte, totidem bis intercolumnia fiant in lateribus; ita enim erit duplex longitudo operis in latitudinem (But let the columns be so disposed in peripteral temples that the intercolumniations on the sides are twice as many as on the front. For then the length of the work will be twice the breadth). In order to give a coherent explanation for the different kinds of temple Vitruvius reckons with modules for which the thickness of the column is the unity. If we apply this breadth/length proportion to the eustyle temple as described with much emphasis in book III, chapter 3, and starting from a hexastyle temple (6 columns in front) we get following results:
Breadth: 6 columns of 1 module + 5 intercolumniations of 2 ¼ modules + 1 central intercolumniation of 3 modules = 18 modules.
Length: 11 columns of 1 module + 10 intercolumniations of 2 ¼ modules = 33,5 modules. From this follows that although the number of intercolumniations is the double (as Vitruvius prescribes) the number of modules is not.
If we apply these same rules to the dimensions of the cella, the cella becomes rather short and is not answering to the proportions we are used to see in temples. Throughout literature about this chapter it has been supposed that Vitruvius means that also the length of the cella is the double of its breadth. Since the ambulatory of a eustyle peripteral temple has the width of one intercolumniation, all around (see book III, chapter 2, 5), we get following proportions (supposed that length = 2 x breadth):
Breadth: 11,5 modules, length: 23 modules, or: the rear wall of the cella stands at the third column from the back, while in reality this wall has to stand at the second column. If we go back to book III, chapter 2, 5 where Vitruvius says that the ambulatory has the width of one intercolumniation, we must suppose that he describes a shorter type of building with only ten or even nine columns at the sides.
If, on the other hand, we follow Vitruvius’ texts in book III, the proportions of the cella are:
Breadth: 11,5 modules, length: 27 modules; or breadth x 2,347. This is far away from length = breadth x 2.
This difference has puzzled the different annotators of this text. Fleury in his edition of 1990 thinks Vitruvius is using different sources for his description: in book III he clearly uses Hermogenes while in this chapter he uses another, not mentioned source, and confuses things.
The whole thing becomes even more complicated if we read further: ipsaqua cella parte quarta longior sit, quam est latitudo, cum pariete qui paries valvarum habuerit conlocationem. Reliquae tres partes pronai ad antas parietum procurrant… (The cella itself is to be a fourth part longer than its breadth, including the wall which contains the doors. The remaining three parts, that is, the portico, are to run forward to the antae of the walls.)
Perrault, in his translation of 1684, presumes that Vitruvius is writing about temples in antis, without ambulatory and gives a beautiful design of this type of building. But his opinion is clearly contradicted by Vitruvius himself in another fragment of this period: Et si aedes erit latitudine maior quam pedes XX, duae columnae inter duas antas interponantur, quae disiungant pteromatos et pronai spatium. (If the temple be more than 20 feet in breadth, between the two antae two columns are to be placed and these columns are to separate the portico and the pteroma (= ambulatory).)
Maufras, in his edition of 1847, affirms the opinion of Perrault. He refers to the above quoted sentence but ommits deliberatley the last part of it: quae disiungant pteromatos et pronai spatium. Moreover, both editors are simplifying things when they make their design of this type of cella: The rectangle has twice the length of the breadth; the length is divided in eight parts according to Vitruvius remark that the cella has to be a fourth part of the breadth longer, while the remaining three parts are given to the pronaos. This gives a proportion of 5/8 for the cella proper and 3/8 for the pronaos. But it is impossible to construct a classical ambulatory around this cella. Given that the breadth is 11,5 moduli, the lenght has to be 23 moduli. The breadth of the ambulatory has to be 1 intercolumniation or 2,25 moduli. From these facts we can deduce the length of the temple to 30,5 moduli. It means that instead of having 10 intercolumniations on the side the temple has only 9 intercolumniations and 10 columns instead of 11.
I tried a lot of calculations myself but couldn’t come to a satisfactory result.
Also a comparison between the text and archaeological remains of different temples of Hermogenes and his colleagues and the period of Vitruvius stayed without result. As an example I took the temple of Athena Polias at Priene built in the fourth century B.C. He has the general layout as described by Vitruvius: 6 columns on front and back with 5 intercolumniations, 11 columns on the sides with 10 intercolumniations (the double of the width) and an ambulatory with a breadth of 1 intercolumniation. The intercolumniation = 1,75 x the lower diameter of the column (i.e. between the pycnostyle and the systyle of book III, chapter 3). If we take the diameter of the column as the module we can calculate the breadth of this temple to 14,75 moduli and the length to 28,5 moduli (nearly a 1 to 2 ratio). Within this ambulatory the cella has a breadth of 9,25 moduli with a length of 23 moduli or a ratio of 1 to 2,5. This is far away from the presumed 1 to 2 ratio as prescribed in this chapter.
Despite all the difficulties we find in this passage it is clear from the rest of the paragraph that Vitruvius is writing about a peripteral temple and not, as is supposed in some editions, about temples in antis (i.e. without ambulatory). This appears from remarks as quae antae columnarum crassitudinem habere debent (The antae ought to have the thickness of the columns). It is beyond doubt that Vitruvius refers to the columns of a peristasis. Further we read …duae columnae inter duas antas interponantur, quae disiungant pteromatos et pronai spatium (….. Between the two antae two columns are to be placed and these columns are to separate the space of the pteroma (i.e. the peristasis or ambulatory) and the pronaos).
In sentence 2 Vitruvius introduces a second set of columns in the interior of the pronaos, when the breadth of the cella is more than 40 feet. These columns are made more slender than the columns in front of the pronaos. When the columns of the pronaos have a ratio of 1:7 these in the interior of the pronaos should have a ratio of 1:10. The more slender aspect is counterbalanced by their position in the interior and by the increasing of the number of flutes.
This increasing of the flutes reflects the theory of the ancients about sight. They believed that the eye transmitted rays that were reflected to the eye when they touched an object. In this kind of perception the structure of the objects played a primary rôle: the greater the surface the longer the eye needed to see. Thus by increasing the number of flutes the surface of the reduced column was considerably increased; and because the eye needed more time to see, the column would give a taller impression.
Throughout the different editions many speculations have been made about the purpose of these additional columns. Some editors thought that these columns were necessary to support the ceiling and roof of the pronaos, but they forgot that these columns were missing in the cella proper. It seems more appropriate to think that these columns had the only function to fill up the void of a proportionately larger pronaos in these broader temples and that they had to mark a processional entrance to the temple inner.
The fourth sentence of this chapter gives us interesting details of the way of building the walls of the cella. The remark Et si extructi futuri sunt, quam minutissimis caementis struantur (If the walls are constructed (of common stone) they must be built of very small stones). This remark refers clearly to book II, chapter 8, 1-2 where Vitruvius gives a description an appreciation of the opus reticulatum and the opus incertum. In II, 8, 2 he uses the same words as in IV, 4, 4: Utraque (i.e. opus reticulatum and opus incertum) autem ex minutissimis (caementis) sunt instruenda (both kinds of walling (i.e. opus reticulatum and opus incertum) are to be built with very minute stones). From this remark we could deduce that Vitruvius prescribes the typical Roman masonry for hellenistic temple types. Unfortunately he gives no detail wether he uses opus incertum or reticulatum or is this only a description of the core of the wall, made in opus caementicium with a revetment of small (regular?) stones?
The rest of the sentence is less ambiguous. With the remark sin autem quadrato saxo aut marmore, maxime modicis paribusque videtur esse faciundum (but if the walls are to be built of square stones or marble, the pieces should be of moderate and equal size) he refers clearly to the opus quadratum isodomum for which rectangular blocks of equal size were used. The wall was constructed in a way that each vertical joint ended in the middle of a stone of the higher layer.
To enhance the picturesque effect in this kind of wall Vitruvius describes a kind of broadened and deepened joint. An example of this work can be seen at the Maison Carrée at Nîmes.
Les dix livres d'architecture de Vitruve, Corrigés et traduits en 1684 par C. Perrault, Paris, 1684.
Vitruvius, De Architectura libri X, ed. F. Granger, London, 1962.
Ton Peters, Vitruvius, Handboek bouwkunde, Amsterdam, 1999.
L’architecture de Vitruve, traduction nouvelle par M. Ch.-L.Maufras, Paris, 1847
Fleury, Vitruve, De l’architecture, Livre IV, Collection belles lettres, Paris, 1992
W.B.Dinsmoor, The architecture of Ancient Greece, London, New York, 1950
B.Wesenberg, Beiträge zur Rekonstruktion griechischer Architektur nach literarischen Quellen, Berlin, 1983
H.Knell, Vitruvs Architekturtheorie, Darmstadt, 1985
Suzan Bayhan, Priene, Milet, Didyme, Ankara, 1993
Back to contents Book IV
Suggestions and remarks?
don't hesitate to send me a message