Home Page Issues Publishing Principles Order Form Links Contact Türkçe
Detailed Search  
Click here to visit web site of Kaleii Museum
If you would like to get announcement mails about Akmed activities, please subcribe to our mailling list.
First name:
Last name:


Stone Sawing Machines of Roman and Early Byzantine Times in the Anatolian Mediterranean

When, in times of the Pax Romana, wealth increased throughout the empire, about every town gained its aqueducts supplying public and private thermae that became standard elements of city life. The walls and floors of these bath houses and of many other buildings were adorned with marble slabs, numerous examples to be found in the Antalya region. These marble slabs must have been produced in ever greater quantities. Recent findings in the Anatolian Mediterranean now show that water-powered stone saw mills were already known in Roman times. A relief on the cover of a 3rd c. AD sarcophagus at the north necropolis at Hierapolis in Phrygia (Pamukkale) shows a technical design - an extraordinary state of affairs on its own - of a water-powered twin stone saw mill. The mill is equipped with the crank and connecting rod system, previously thought to have been a medieval invention. With the crank and connecting rod system the rotary movement of the water wheel is transformed into a reciprocal linear movement, enabling mass production of marble slabs.

The sound of grain mills and marble saws driven by the waters of the river Erubris (Ruwer) is described Ausonius’ poem Mosella, written about 370 AD: ‘Turning the stones in headlong rotation and drawing the creaking saws through shining marble, it (the Erubris) hears an incessant noise on both its banks’. The poem had long been considered proof for the existence of sawing machines in Roman times, but in the 1960’s the notion that the crank and connecting rod system could not have existed in Roman times had led to the idea that the Mosella poem in reality was a 10th c. addition to his works. Only in the early 1980’s this opinion was proven false and the poem accepted as genuine. Yet, it could still not be imagined that the Romans knew the crank and connecting rod, for which indeed proof was lacking, and to meet with the Ausonius poem scholars proposed alternative mechanisms as circular saws and continuous wire saws which these all meet severe technical problems. The Hierapolis relief has now shown that the Romans knew the crank and connecting rod system and applied this system for stone sawing machines, proving  that Ausonius had it right in the 4th c. with the sound of water driven machines sawing marble north of the Alps.

Remains of a stone saw mill at Ephesos, with two parallel saw blades, and a similar machine at Gerasa (Jordan), with even four parallel saw blades, both dating from the 6th -7th c. AD, show that the mill machinery improved since the Hierapolis mill had been designed. These findings of Bysantine times together with the Hierapolis relief have led to a number of reconstructions of ancient stone sawing machines which are discussed in this article. All reconstructions envisage the crank and connecting rod system, but all differ on technicalities of the saw frames and of the guiding system for the saw blades. For the sawing machines the frames in which the saw blades are mounted may be either horizontal or vertical. The Hierapolis relief undoubtedly shows vertical frames, but from the findings at Ephesos and at Gerasa it is not clear what kind of frames were applied in Byzantine saw mills. Both vertical and horizontal saw frames are envisaged for the reconstructions of these ancient sawing machines. Furthermore scientists argue that some  kind of guiding system was necessary for the saw frames on their way down to ensure straight cuts, although not all agree on this.

The 2009 reconstruction of the Ephesos mill, by Fritz Mangartz (Römisch Germanisches Zentralmuseum, Germany), has no guiding system for the saw blades. Mangartz proposes horizontal saw frames for easier mounting of the blades and for the fact that the center of gravity of the frame lies along and not above the saw blades, reducing the risk of misalignment. The horizontal saw frames are suspended on all four corners from ropes with counter weights over pullies. Mangartz discusses the principle of pendular saws stating that the saw frame will move up and down continuously when pushed and pulled to and fro, contacting the stone for a short but sufficient distance of the swing only if the ropes are relatively long. The up and down movement of the  saw would guarantee that the mixture of abrasive and water deposits underneath the saw blade required for efficient sawing. In a 2010 publication Mangartz discusses experiments conducted on a 1-1 model of his reconstruction.

The Gerasa reconstruction by Jacques Seigne (University of Tours/CNRS, France) is equipped with four parallel saw blades mounted in vertical saw frames, resembling the Hierapolis situation. A 1 to 1 reconstruction on the original location was realized in 2007. Contrasting Mangartz Seigne does not apply counter weights.  But the heavy saw frames that Seigne envisages, however, do not readily allow for the lifting of the saws needed for the abrasive material to enter underneath the blades. Seigne envisages simple wooden guiding frames for the vertical saw frames to go down in a straight line. The heavy saws require sturdy and strong guiding frames, and as a consequence heavy cranks and connecting rods (for which counterweights are now thought necessary). This contrasts the manual stone saws of the pendular type and with counter weights, already known for ages, that a single or maybe two men must have been able to operate, sawing marble blocks of similar size as found in the Gerasa work shop, be it one cut at a time. In the above article it is argued that, although the Gerasa multiple blade system does require a more extended and thus sturdy vertical frame, reducing size and weight as much as possible and adding compensating weights may be an vital issue. Future experiments planned at Gerasa may resolve this matter.  

A third reconstruction, of the Hierapolis stone saw mill, by Klaus Grewe (Bonn, Germany) has both a vertical and a horizontal guiding system for  the saw blades. Grewe refers to a machine proposed in 1718 by a German engineer named Sturm. One frame is equipped with wheels or rollers to move horizontally over some kind of railing system fixed on a platform or table on which the block to be cut is positioned. This carriage is pushed and pulled by a connecting rod from a crank driven by a water wheel. A second frame, equipped with the saw blades, is able to move vertically up and down within the frame on rollers. Grewe argues that a similar system must have been applied for the Hierapolis machine. Such machine indeed prevents all misalignments of the saw blade and will produce flat cuts. However, Sturm himself claims that he invented the machine because he had seen several saw mills that in his opinion did not function properly. Inventing a machine that would have already been in use for one and a half millenium, since Ammianos’ time, does not seem quite probable. Sturm obviously invented a machine that was not known before,

Manually cutting large blocks with vertical saws carrying toothless blades measuring up to 4 m and more in length and suspended from ropes with counter weights over pullies, operated by one or two labourers, has been longstanding practice from ancient times on into the 19th and early 20th century. If the Ammianos machine is imagined with vertical frames and with a system of counter weights as for manual saws, guaranteeing lifting of the saw frame so that abrasive material deposits below the saw blade, the labourer, relieved of his eternal task of pushing and pulling the saw, adding water and abrasives and preventing misalignments of the saw, a realistic and practical model of the Hierapolis water powered stone saw mill, and later mills, may be envisaged.

* Dr. H. P. M. Kessener
c/o L&L, van Slichtenhorststraat 13, 6524 JH Nijmegen, Holland
E-mail: lenl@euronet.nl

Article List