1Not many drawings or images of technical constructions or machinery have been preserved from classical times1. Only in few cases is has been possible to reconstruct the functioning of a technical contraption by means of a description or illustration of the ancient engineer/designer, both for large constructions and for manual tools alike.

2The modern engineer has no choice but to extract data from the often limited and incomplete remains of the ancient construction or apparatus to decipher the code or plan of the design. This led Th. Mommsen, when confronted with an inscription on the funerary monument for Nonius Datus, describing the construction of a tunnel for the aqueduct at Saldae (Algeria), to delegate this problem, as he writes: « es wird vielleicht ein verstandiger Ingenieur unserer Epoche aus dem Bauwerk selbst dasjenige zu lösen wissen, was uns im Bericht seines romischen Vorfahren unverständlich bleibt »2.

3In the necropolis west of Hierapolis, to the left and to the right of the road towards the’Frontinus Gate’, the numerous tombs and sarcophagi are equipped with countless inscriptions and, fewer in number, with a relief. One relief, on one end of the stone cover of a sarcophagus, shows a technical contraption. Together with an inscription a water-powered twin stone saw is depicted, in schematic outline. From the inscription it is understood that the sarcophagus was dedicated to a certain M. Aurelios Ammianos. It is dated second half or end of 3^rd c. AD3.

4The Hierapolis relief shows important new information about technicalities involved in the mechanism of the water powered saw mill, which requires the conversion of rotary motion into linear motion. Undisputed remains of water-powered stone saw mills exist at Ephesos (6^th / early 7^th c. AD) and, from recent findings by Jacques Seigne, at Gerasa, Jordan (6^th c. AD). In this paper the Hierapolis relief is discussed, in view of the findings at Gerasa and Ephesos4.

The Hierapolis relief (Fig. 1-4)

5On the relief one’s attention is drawn to the large vertical wheel on the right having a long horizontal shaft extending to the left. To the right of the wheel a chute is depicted in the form of some kind of inclined piping or channel, aiming at half height of the wheel, suggesting that the wheel was driven by waterpower and employed as a breast wheel. The impact of the water onto the blades drives the wheel and would guarantee a constant rotational movement, whereby the energy needed to operate the saw is generated.

6The transfer of the rotation of the waterwheel to the stone saw mechanism has been depicted schematically. The shaft of the water wheel appears to be equipped with a gear train of two gears, one wheel mounted onto the shaft of the big wheel. This wheel drives a third wheel, apparently equipped with cranks or with crank wheels (a disk of some material fixed onto the wheel shaft and equipped with a wooden or metal eccentric pin), and connecting rods to convert the rotational movement of the wheel into a linear one, needed to operate the saws.

7Both stone saws are depicted in sufficient detail to conclude that they are representing so called span-saws, the saw blades kept in tension by means of a (wooden) frame. The image shows the saws while operating, the blades having progressed half-way through the stone blocks.

8Although the saws have been depicted correctly, a mechanism to guide the saw blades is not present. A smooth and regular saw-cut can only be achieved when the blades are kept in vertical position and movement sidelong is prevented. To this purpose a guiding frame in envisaged in which the saw, apart from the reciprocal movement by the connecting rods, may move only in strict downward direction. As in modern stone saws, several parallel saw blades may be mounted next to each other, allowing several stone slabs to be cut simultaneously. Figure 4 shows a reconstruction drawing of the stone saw of the Hierapolis relief, with a crank wheel with eccentric pin and with a crank (on either side of the shaft of the third wheel), and with a simple guiding mechanism for one of the saws.

9The conversion of rotary movement into a linear one is one of the most controversial issues in the study of ancient applications of waterpower5. One method that is ascertained is the cam or camshaft to drive tilt-hammers or trip-hammers. Cams go back to the 3^rd C BC in Greek automata, and cam-operated tilt hammers are thought to have been used in the Roman Empire6. A waterwheel moving a saw, however, requires a crank or crank wheel and a connecting rod. According to Schiøler, all historians of technology agree that the one-armed crank was unknown in Roman times’7. The Hierapolis relief obviously proves differently.

The stone saw mills at Ephesos and at Gerasa

10The mill at Gerasa and the one at Ephesos show several similarities. Each mill is located in a rectangular workshop, one of the walls equipped with a headrace some 3-4 m above floor level, millrace and tailrace in the floor of the shop. At Gerasa the positions for the bearings, to accommodate the shaft of a waterwheel, have been preserved in the millrace walls. Seigne estimated that the remains suggest a vertical overshot waterwheel about 50 cm wide having a diameter of 4-4.5 m. The outer surface of the mill race walls show circular wear marks showing that something was fixed at either end of the wheel shaft rotating with the wheel. Seigne stated that a wooden disk of at least 1 m in diameter was fixed to the wheel shaft rotating with the water wheel.

11In the workshop at Ephesos as well as at Gerasa two large stone blocs have been found which are partly cut into slabs. At Ephesos the blocs are rectangular and some 2. 2m long. They are still in their original position, parallel to each other on either side of the mill race. The blocs at Gerasa consist of limestone drums 1. 51 and 1. 67 m. long and about 1 m. in diameter, their original position are thought to be parallel to each other left and right to the mill race as well. The Ephesos blocs show 2 parallel saw-cuts, the Gerasa drums show 4 grooves each, penetrating to the same depth even though the outer surface is curved. Evidently the saws worked on both blocs simultaneously, with parallel blades for each block.

12In the reconstruction of Schiøler the waterwheel is envisaged to drive a large horizontal frame in which two sets of saw blades were fixed. The weight of the heavy rectangular frame, which is to be guided on its way down to prevent deviation from vertical, is thought to be compensated by counterweights exerting an upward force by means of ropes over pulleys8. Seigne, however, envisages two separate vertical saw frames, as in the Hierapolis relief, one for each stone bloc, moved to and fro by connecting rods that are driven by wooden disks equipped with a metal pin and fixed to the shaft of the water wheel (crank wheels)9. As the frames are smaller and less heavy than the horizontal frame in Schiøler s reconstruction no compensatory weights would be necessary.

The Hierapolis relief and the mills at Ephesos and at Gerasa

13In the Hierapolis relief the waterwheel appears to be applied as a breast wheel, the sloping chute pointing to the wheel at half height10. Two blocs are sawn simultaneously, by two span-saws, the saw blades fitted into vertical frames. Such frames may accommodate two or more parallel saw blades. Thus, from the Hierapolis relief, a reconstruction of the stone saw mills at Ephesos and at Gerasa11, with crank wheels or cranks and connecting rods, and with vertical saw frames, appears realistic.

14According to Ritti, the relief and the inscription must be dated to the 2^nd half or end of the 3^rd c. AC, preceding the attested Ausonius poem on the marble saw mills at the Ruwer (371 AD) a hundred years, and three centuries before the mills at Ephesos and at Gerasa. Ritti states that the inscription tells us that M. Aurelius Ammanius is referred to as an ingenious and inventive man (δαιδάλεος), not just a labourer or an official, a man who maybe invented this type of stone saw or introduced some new element to the saw mechanism.

15The Hierapolis relief shows us beyond doubt that vertical saw frames have been applied for water-powered stone saws, and, of great importance, that the crank wheel /crank was known in the 3^rd c. AD. There are, however, remarkable differences between the situation at Ephesos and Gerasa and the one depicted on the Hierapolis relief. At Gerasa and apparently also at Ephesos, the shaft of the water wheel is short, not much longer than needed for the wheel itself and the two bearings on either side of the wheel. In the Hierapolis relief the water wheel’s shaft is extended, long enough to accommodate a smaller, toothed wheel that serves as a driver for a third wheel, to the shaft of which the crank wheels are mounted. This gear train reminds of the machinery of vertical water mills applied for driving millstones, the Vitruvian mill, where a vertical cogwheel is mounted onto the shaft of the water wheel, driving a horizontal wheel to which the rotating millstone is fixed. The cogwheel and the horizontal wheel were usually accommodated in a separate space into which the shaft of the water wheel extended, the shaft’s total length usually being between 2.9 and 3.6 m.

16Furthermore, in the Hierapolis relief the saw frames and the stone blocs that are sawn are positioned to the left and to the right from the gear train. This contrasts the situation at Ephesos/ Gerasa, where the blocs are positioned next to each other on one side of the water wheel. The arrangement at Ephesos/Gerasa, the water wheel being positioned next to the wall with the headrace, does not allow an alternative set-up. Maybe it was the artist’s impression, in the Hierapolis relief, to depict the saws and the stone blocs on either side of the driving wheel, but the configuration with the gear train no doubt allows the positioning of the saws in the relief. It may have been the merit of M. Aurelios Ammianos that he changed the combination of vertical cogwheel and horizontal wheel into a gear train with two vertical wheels for the purpose of driving some non-rotational machinery. He may actually have used an existing water mill to experiment, as the only thing he had to do was to replace the horizontal wheel of this gear train by a vertical one and position it below the cogwheel, adding crank wheels, connecting rods and span-saws. This would explain the long shaft of the water wheel and the two extra wheels in the Hierapolis relief.

17By the gear train that Ammianos applied for driving his saws, energy is inevitably lost by friction. One may envisage that, at a later stage, the gear train was abolished to the favour of crank wheels mounted directly onto the water wheel’s shaft, thus simplifying the design, reducing the length of the shaft, and enhancing efficiency of the mill. This would lead to the mill configurations known from Ephesos and Gerasa. One may only speculate that Ammianos was the inventor of the crank or crank wheel, to convert a rotational movement into a linear one, and, that he also was the inventor of the water-powered stone saw. It would have been a revolutionary and tremendous invention. The Hierapolis relief shows us that, whatever the case may be, Ammianos, and the Romans, knew and applied this technique, in his time, the second half of the3^rd c. AD.