On the foundations of temples

Contents of this chapter

  1. Text and translation
  2. Comment
  3. Philibert De l'Orme on foundations
  4. Scamilli impares
  5. Illustration of the Cesariano edition, 1521

1. Fundationes eorum operum fodiantur, si queat inveniri, ab solido et in solidum, quantum ex amplitudine operis pro ratione videbitur, extruaturque structura totum solum quam solidissima. Supraque terram parietes extruantur sub columnas dimidio crassiores quam columnae sunt futurae, uti firmiora sint inferiora superioribus; quae stereobates appellantur, nam excipiunt onera. Spirarumque proiecturae non procedant extra solium; item supra parietis ad eundem modum crassitudo servanda est. Intervalla autem concamaranda aut solidanda festucationibus, uti distineantur.


1. Let the foundations of those works be dug from a solid site and to a solid base if it can be found, as much as shall seem proportionate to the size of the work; and let the whole site be worked into a structure as solid as possible. And let walls be built upon the ground under the columns, one-half thicker than the columns are to be, so that the lower portions arestronger than the higher; and these are called stereobate, for they receive the loads. And let not the projections of the base mouldings proceed beyond the bed. Further, the thickness of the wall is to be kept above in the same manner. The spaces between the columns are to be arched over, or made solid by being rammed down, so that the columns may be held apart.

2. Sin autem solidum non invenietur, sed locus erit congesticius ad imum aut paluster, tunc is locus fodiatur exinaniaturque et palis alneis aut oleagineis (aut) robusteis ustilatis configatur, sublicaque machinis adigatur quam creberrime, carbonibusque expleantur intervalla palorum, et tunc structuris solidissimis fundamenta impleantur.
Exstructis autem fundamentis ad libramentum stylobatae sunt conlocandae.


2. But if a solid foundation is not found, and the site is loose earth right down, or marshy, then it is to be excavated and cleared and remade with piles of alder or of olive or charred oak, and the piles are to be driven close together by machinery, and the intervals between are to be filled with charcoal.
Then the foundations are to be filled with very solid structures. The foundations being built to a level, the stylobates are to be laid.

3. Supra stylobatas columnae disponendae, quemadmodum supra scriptum est, sive in pycnostylo, quemadmodum pycnostyla, sive systylo aut diastylo aut eustylo, quemadmodum supra scripta sunt et constituta. In araeostylis enim libertas est quantum cuique libet constituendi. Sed ita columnae in peripteris conlocentur, uti, quot intercolumnia sunt in fronte, totidem bis intercolumnia fiant in lateribus; ita enim erit duplex longitudo operis ad latitudinem. Namque qui columnarum duplicationes fecerunt, erravisse videntur, quod unum intercolumnium in longitudine plus quam oporteat procurrere videtur.


3. Above the stylobates the columns are to be erected as described above; whether in pycnostyle (as are the pycnostyle temples), or in systyle, or diastyle or eustyle, as it has been described and determined above. For in araeostyle there is a freedom to determine as everybody pleases. 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. For those who made double the number of the columns seem to be at fault because in the length one more intercolumniation than is necessary seems to occur.

4. Gradus in fronte constituendi ita sunt, uti sint semper impares; namque cum dextro pede primus gradus ascendatur, item in summo templo primus erit ponendus. Crassitudines autem eorum graduum ita finiendas censeo, ut neque crassiores dextante nec tenuiores dodrante sint conlocatae; sic enim durus non erit ascensus. Retractiones autem graduum nec minus quam sesquipedales nec plus quam bipedales faciendae videntur. Item si circa aedem gradus futuri sunt, ad eundem modum fieri debent.


4. The steps are to be so placed in front that they are always of an uneven number. For since the first step is ascended on the right foot, the right foot must also be set on the top of the temple steps. And the risers of the steps must be of such dimensions that they are neither deeper than ten inches nor shallower than nine. For thus the ascent will not be hard. But the treads of the steps, it seems, should be made more than two feet. Also, if steps are to be round the temple, they ought to be made after the same measure.

5. Sin autem circa aedem ex tribus lateribus podium faciendum erit, ad id constituatur, uti quadrae, spirae, trunci, coronae, lysis ad ipsum stylobatem, qui erit sub columnarum spiris, conveniant. Stylobatem ita oportet exaequari, uti habeat per medium adiectionem per scamillos inpares; si enim ad libellam dirigetur, alveolatum oculo videbitur. Hoc autem, uti scamilli ad id convenientes fiant, item in extremo libro forma et demonstratio erit descripta.


5. But if a platform is to be made round the temple on three sides, it is to be planned in such a way that the plinths, bases, dados, cornices and cymatium conform to the pedestal which is under the bases of the columns. The stylobate must be so levelled that it increases towards the middle with unequal risers; for if it is set out to a level it will seem to the eye to be hollowed. The method of making the risers suitable to this will be set out with a figure and demonstration at the end of the book.


The first care of every architect is of course the search of a solid ground for laying out the foundations of a building. Vitruvius describes different circumstances and manners to find this solid base. It is remarkable that in his theory only the points which shall have to carry the load of the building are put on a firm soil. The spaces between these points (or walls)are or arched over or filled again with earth which has to be tamped down.
The method with the arches had the advantage to save a lot of labour. Moreover, the pressure of the arches on both sides of the walls involved the immovability of these foundation walls. The method is illustrated in the drawing I found in the recently published Dutch translation.
Throughout the architectural history and the construction methods through the ages we can find these method with arches mostly used in loose soils where it is too difficult and too labourious to dig the whole foundation trench to the solid ground. In that case smaller pits were dug in which piles were built; these piles where subsequently linked up by arches to form a continuous base for the wall to be erected.
If the walls were not connected with arches, the spaces between were filled with earth wich had to be tamped down. For this purpose was used a fistuca. Vitruvius uses this word twice: the first time in this context, the second time in book VII, chapter 1, where the fistuca is used to level pavements.
It isn't very clear what Vitruvius meant by this fistuca. Throughout Latin literature the word has two meanings: stick and rammer. In my opinion we must assume that Vitruvius means a stick or a kind of cudgel. Maybe this tool is represented on the column of Trajan where I found two representations of a soldier working with a kind of cudgel in the building of a camp. The first is clearly levelling stones of a wall while the second, in a rather confused scene, is working at the foot of a wall in construction ... ramming earth.

In the second sentence a more complicated method for marshy soils is described. First a coffering is to be made out of two rows of posts rammed in as close together as possible. Here again Vitruvius refers to the use of engines (machinis) but without giving us any idea about the kind of engines he meant. Maybe it is the same fistuca but used in the meaning of rammer. We find this meaning in Caesar, De bello gallico, IV, 17, 4 where he describes the construction of a bridge across the Rhine. A model of this engine is presented in the Museo della Civiltá Romana in the EUR at Rome. Then the earth between these two rows is dug out and the space filled with a solid material. Vitruvius uses the word 'carbo' for this material. In the three translations I possess this word is rendered by charcoal, charbon (French), or houtskool (Dutch). In my opinion these translations are wrong since charcoal can't resist heavy loads. Another possible translation for carbo is ash, but this poses the same problem. Consequently it isn't clear which material exactly is meant with that carbo. I have the same problem with the filling of the foundation above the carbo: structurae solidissimae. The English translation I use renders the text literally: very solid structures. The French translation gives maçonnerie très solide and the Dutch translation gives solid brickwork.

I have the following opinion:
Vitruvius is clearly referring to alluvial (congesticius) and marshy (paluster) soils. In these soils it is impossible to dig a trench for the layout of a foundation. To this purpose piles are driven in to form a coffering to allow workmen to dig out safely the ground between the two rows of piles; without these piles the incoming water could cause the crumbling down of the walls of the trench. But, once the trench is dug, in these watery conditions it is also impossible to make a good masonry. Therefore concrete or a comparable material is used which has the property that it dries and hardens even under water. I refer for this material to book II, chapter 6 which discusses the properties of the Pozzolana - or vulcanic ashes.
If we accept this interpretation we have a satisfying explanation of this fragment. However, the only problem remains the fact that Vitruvius didn't use the proper words 'calx' or 'caementum'.
This point of view is strengthened by a fragment of Philibert De l'Orme's architecture where he describes in his second book, chapter XI, the same technique. I quote the original text in ancient French:
Doncques le fondement estant prest à maçonner, s'il est grand et large,&que vous ne puissiez trouver de grandes pierres pour mettre au fond, soit pour édifices, ou pour fonder un port de mer, ou faire ponts sur une rivière ou bastir dans un palus, ou encore dans la terre; la meilleure chose & plus prompte sera de préparer le mortier ainsi que la chaux vient du four (i.e. carbo) (comme nous l'avons declaré au premier livre) avecques du sable qui soit de rivière,&portant plusieurs sortes de cailloux de telle grosseur qu'il se trouveront, pourveu qu'ils n'excedent la grosseur du poing pour le plus, ou la grosseur d'un oeuf,&soient accompagnez de plusieurs autres petits cailloux & gravois, comme on les trouve dedans les rivières. Telle matière destrempée & meslée avecques la chaux sert de pierre & de mortier: pour autant que tel gravois porte du sable quant & soy,&se jecte tout à une fois dedans les fondements, sans que les Maçons ayent peine d'y besonger avecques leur truelle, car il suffist le dressant uniment avecques la paelle. L'ayant ainsi respandu, jusques à un demi pied d'espesseur, vous y pouvez jecter & entremesler par-cy, par-là, plusieurs grosses pierres seules, ainsi qu'on les peut trouver à propos, sans toutefois qu'elles se touchent: les plus dures y seront les meilleures, comme sont roches ou cailloux. Après cela vous rejectez encores par dessus dudit mortier faict de cailloux & gravois, comme vous l'aviez faict auparavant. Il faut ainsi continuer jusques à ce que le fondement soit plein, jectant le tout d'enhaut avecques toutes sortes de cailloux. Telle matière ainsi disposée s'endurcit & reserre si fort dedans les fondements, qu'estant accumulée & liée ensemblement, devient comme une seule masse & roche, laquelle nature auroit faicte toute d'une pièce, estant si forte & massive, quand elle est seiche, qu'on ne la peutrompre avecques le pieu, ou autre instrument, n'aussi arracher les cailloux du fondément, qu'ils ne se mettent en pièces.

In this fragment De l'Orme advises the use of a carbonated material (mortar) as the first layer in marshy conditions. This layer is half a foot thick. Above this first layer a kind of concrete (=structura solidissima) is poured in from above with the purpose to give this foundation the maximum strenght. This material has to fill up the whole foundation.

Scamilli impares

Throughout the different editions of Vitruvius the 'scamilli impares' of sentence 5 have always been a point of discussion. Vitruvius refers to a design at the end of his book, but this design is lost. The French translation of 1684 is completely wrong: there the scamilli impares are projecting pedestals on which the columns rest.

In fact this sentence is about optical effects. If a stylobate is completely set out to a level it will seem hollow or concave. To counteract this effect the middle is raised a little which results in a (for the human eye) straight line. This effect was already known in Greek architecture and we can find it in many temples. From the survey of these temples following rules have been deduced for these scamilli impares.
The system consisted in deciding first the maximum increment of curvature desired, and then any convenient arbitrary number of equal intervals between the corner and the middle of the building. The maximum increment of curvature was next divided by the square of the above-mentioned number of intervals, thus determining the size of the fractional parts of the height, sot that the curve could be set out either in diagrammatic form as a parabola with the equal spaced ordinates descending on either side of the apex as successive square numbers of the fractional parts, or on the actual temple as a colossal arc by replacing the ordinates with corresponding levelling blocks (the scamilli impares).
An example can make this clear. A stylobate with a length of 70 m. might have 12 intervals of 5,83 m., 6 on either side of the centre. If the desired curvature were 12 cm it would be divided by 6², into 36 parts of 3,33 mm. each. The first ordinate on either side of the apex would descend 1 part or 3,33 mm lower than the apex, the second 3 parts or 9,99 mm., the thirth 5 parts or 16,65 mm, the fourth 7 parts or 23,31 mm, the fifth 9 parts or 29,97 mm, the sixth finally 11 parts or 36,63 mm. The difference between the successive ordinates were the successive odd numbers, thus the levelling blocks by means of which these rising curves were executed became scamilli impares in two senses: they were not only unequal but they differed by odd numbers.


Philibert De l'Orme, Architecture, Rouen, 1648
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.
H. Knell, Vitruvs Architekturtheorie, Versuch einer Interpretation, Darmstadt, 1985.
W.B. Dinsmoor, The architecture of Ancient Greece, London, 1950

Chapter 5
Back to contents Book III

Suggestions and remarks?
don't hesitate to send me a message