-96-
Part III
Time in the Modern World
-97-
[This page intentionally left blank.]
-98-
7. The Advent of the Mechanical Clock
The invention of the verge escapement
In antiquity the only mechanical (or, more strictly speaking, quasi- mechanical) instruments for recording the passage of time were water-clocks. The fundamental difference between water-clocks and mechanical clocks, in the strict sense of the term, is that the former involve a continuous process, for example, the flow of water through an orifice, whereas the latter depend on a mechanical motion that continually repeats itself and so divides time uniformly into discrete segments. Many of the ancient water-clocks were instruments of considerable complexity, particularly as they were designed to indicate hours which varied throughout the year. Although there were no mechanical clocks in antiquity, mechanical models appear to have been constructed to reproduce the relative motions of the heavenly bodies. Writing in the first century BC, Cicero ( De republica, I. xiv. 22) referred to one invented in Syracuse by Archimedes ( 2 87)-212 BC). We know nothing about the gearing devices involved, but the associated mathematical calculations may have been contained in a lost treatise of Archimedes On sphere- making, that is on modelling the heavens. One remarkable Hellenistic geared mechanism, however, has survived from the first century BC. It was discovered in 1900 in the wreck of a Greek ship near the barren islet of Antikythera, off the south coast of Greece. In 1974, D. J. de Solla Price reported on the results of X-ray and gamma-ray radiography of the corroded remains of this bronze mechanism and concluded that it was a calendrical computing device.1 It appears to have included means of determining the positions of the sun and moon in the zodiac, and it involved an assembly of wheels with fixed gear-ratios for the mechanization of the Metonic cycle, in which 19 solar years correspond to 235 lunar months (see Appendix 2). According to our present knowledge, this machine was the nearest the artificers of antiquity came to inventing a truly mechanical clock.
-99-
Fig. 2 Reconstruction of the Antikythera geared mechanism.
This is a diagram of the differential gear assembly of the Antikythera mechanism as reconstructed by D. J. de Solla Price from the four surviving heavily corroded bronze fragments, which appear to contain the remains of 31 gear wheels. From the respective numbers of teeth in these remains, he claimed that the function of the instrument was calendrical, since it seemed to him that in the complete mechanism the numbers 19 and 235 of the Metonic cycle (Appendix 2) were involved. It is possible that this mechanism formed part of a collection of 'spoils of war' that was being conveyed by sea to Rome following the sack of Athens by Sulla's troops in 86 BC.
Until recently the Antikythera mechanism was thought to be the sole surviving example of mathematical gearing in the Hellenistic tradition. However, in 1983 four fragments of a geared instrument of early Byzantine origin, probably made during, or just before, the reign of Justinian I ( 527-65), were acquired by the Science Museum, London.2 It has been possible to reconstruct the complete instrument, which
-100-
was a brass portable sundial with a geared calendar that showed the approximate shape of the moon and its age in days and may also have shown its position and that of the sun in the zodiac. Two of the fragments involve gears of fifty-nine and nineteen teeth and of ten and seven teeth, respectively. These correspond to parts of a mechanical calendar described by the Persian scientist al-Biruni ( 9 73)- 1048) about the year 1000. A practical link has thus been revealed between the Hellenistic tradition of mathematical gearing and the medieval Islamic. The only surviving example of the latter is a calendrical mechanism yielding the shape and age of the moon in days and the positions of the sun and moon in the zodiac. It is attached to a Persian astrolabe of the early thirteenth century now in the Museum of the History of Science at Oxford. The earliest surviving gearing from the Latin West is attached to a French astrolabe of about 1300, now in the Science Museum, London.
Although no definite links have yet been discovered between the first mechanical clocks and earlier geared astronomical models and automata, the way in which a surviving late fourteenth-century clock such as that of Wells Cathedral displays the phases of the moon and figures which emerge at successive hours suggests that such clocks were the product of a continuing tradition from the distant past. There is also textual evidence to support this view. Nevertheless, the actual origin of the mechanical clock remains a mystery, although it probably occurred towards the end of the thirteenth century.
Early that century the market for water-clocks was such that a guild of clockmakers is known to have existed in Cologne who by 1220 occupied a special street, the Urlogengasse, or Clockmakers Street.3 Since, in northern climes, water-clocks must have been a nuisance in winter when they froze, in the fourteenth century sand-clocks were invented. This invention was made following the introduction of a new and finer 'sand' made of powdered eggshell. Coarse sand cannot be used for this purpose, because it soon enlarges the hole through which it flows. Sand-clocks proved suitable only for measuring short periods. They were principally used on board ship to measure its speed by counting the number of knots paid out on a line tied to a log floating astern, while the sand-glass measured a given time that was usually half a minute. They were also used for timing the length of sailors' watches. Incidentally, it was not until the end of the fifteenth century that the sand-glass was depicted as the attribute of Father Time.4
The incentive to develop the mechanical clock may well have been fostered by the need for it in medieval monasteries, where punctuality
-101-
was a virtue that was rigorously insisted on and late arrival at divine service or meals was punished. The need for punctuality was not due to any desire for 'saving time' but because the strict regulation of time was needed to help maintain the discipline of monastic life. In any case, it seems inevitable that the development of the mechanical clock should have been primarily due to the Church for, although the transmission of power by rope and pulley had long been known to craftsmen, the mathematics of gear-trains (particularly astronomical trains) was known only to the highly educated, and their education was provided only by the Church.
The English word 'clock' is etymologically related to the medieval Latin word clocca and the French word cloche, meaning a bell. Bells played a prominent part in medieval life, and mechanisms for ringing them, made of toothed wheels and oscillating levers, may have helped to prepare for the invention of mechanical clocks. Possible evidence for this view can be seen in the only surviving thirteenth-century picture of a Western water-clock, which seems to have been used about 1250 in Paris at the court of Louis IX. It was essentially a device for ringing the hours. The only visible wheel appears to have twenty-four teeth, which may signify that it rotated daily. The driving power was provided by a slowly descending weight hanging from a cord wound round the axle, this being the earliest instance known of a weight drive in a clock. It was followed about twenty years later, in 1271, by the forecast of a purely mechanical chronometer by Robertus Anglicus ('Robert the Englishman') in a commentary that he wrote on the Treatise on the Sphere of Sacrobosco. He envisaged this as a well-balanced wheel driven by a lead weight suspended from its axle so that it would make one revolution between sunrise and sunset. Nevertheless, he said of those clockmakers who were attempting to make such a timepiece that 'they cannot quite perfect their work'.5
Under the patronage of Alfonso X (el Sabio, 'the wise') of Castile a set. of improved astronomical tables, known as the 'Alfonsine Tables', was compiled by Rabbi Isaac ben Sid of Toledo and published in the Libros del saber de astronomica in 1277. In Volume IV of this work, republished in Madrid in 1866, various inventions are described including a 'mercury clock'. This is a weight-driven clock equipped with a brake consisting of a drum divided into twelve compartments with small holes in the dividing walls. The lower six compart
ments are filled with mercury. As the driving weight causes the drum to rotate, the mercury is raised until it counter-balances the weight, which can then fall slowly as the mercury
-102-
flows through the dividing walls. The uniform motion of the drum depends on the viscosity of the mercury. The motion can be regulated by varying the weight and/or the size of the drum. The essential feature of the mechanical clock missing from this interesting device was the 'escapement'. The invention of the mechanical clock probably occurred after 1277, since if it had occurred earlier it is almost certain that it would have been included in Volume IV of the Libros del saber de astronomica. It would seem that the date of the invention of the mechanical clock is probably some time between 1280 and 1300.
The crucial invention that made the mechanical clock possible was the 'verge-and-foliot' escapement. A horizontal bar, or 'foliot', was pivoted at its centre to a vertical rod, or 'verge', on which were two pallets. These engaged with a toothed wheel (driven by a weight suspended from a drum) which pushed the verge first one way and then the other,
Fig. 3 The verge-and-foliot mechanical clock.
The foliot (from the Latin word for 'leaf') was a horizontal bar (or balance) with a weight (or regulator) at each end (Fig. 3a). At its mid-point the bar was fixed to the verge (from the Latin word for 'twig'), a vertical rod on which were two pallets (or flanges). These engaged with a toothed wheel, which pushed the verge first one way and then the other (Fig. 3b), thereby causing the foliot to oscillate. The wheel itself advanced by one tooth for each double oscillation. The rate of oscillation could be adjusted by altering either the weights or their distance from the verge. This ingenious mechanism was robust, almost impervious to wear, and capable of. ticking away ceaselessly so long as its moving parts were kept well-oiled. Its main disadvantage was that, unlike a pendulum, the balance controlling the oscillations had no natural period of its own.
-103-
causing the foliot to oscillate. The wheel advanced, or 'escaped', by the space of one tooth for each to-and-fro oscillation of the foliot. The foliot carried two weights (regulators) on each side, and the speed of the oscillation could be adjusted by altering either the weights or their distance from the verge. (In Italy the foliot was sometimes replaced by a balance wheel with a similar reciprocating action.) The system also involved a mechanism for counting the oscillations. No one knows who first made this ingenious invention, although as already indicated it was probably towards the end of the thirteenth century. According to C. F. C. Beeson, the earliest European record of a clock with a mechanical escapement is that of 1283 in the Annals of Dunstable Priory in Bedfordshire.6 He also cites records from Exeter Cathedral ( 1284), old St Paul's, London ( 1286), Merton College, Oxford ( 1288?), Norwich Cathedral Priory ( 1290), Ely Abbey ( 1291), and Canterbury Cathedral ( 1292). As J. D. North, who has drawn attention to these, remarks, 'Taken singly, the records are easy to view with scepticism, but taking them together, and noting especially that relatively large sums of money are involved in payment for the materials used, they persuade us that the mechanical clock had indeed arrived on the scene.'7
Although it is generally supposed that the first truly mechanical escapement was of the verge-and-foliot type found in various church clocks throughout Europe, the earliest escapement of which we have definite detailed knowledge is that of the clock designed for the Abbey of St Albans c. 1328 by Richard of Wallingford (c. 1292- 1336), the son of a blacksmith, who became Abbot in 1327. (His father's occupation is particularly significant, because the invention of the mechanical clock must presumably have depended on the co-operation of the learned man, probably a monk, who first thought of it and the blacksmith who actually constructed it.) It was an oscillating mechanism involving an extra wheel, as compared with the verge-and-foliot system. J. D. North has succeeded in reconstructing the St Albans escapement from the purely verbal description given in the surviving manuscript.8 It was in some ways superior to the verge type. North has also found that a similar escapement was known more than a century and a half later to Leonardo da Vinci. Drawings of it are given in his Codex Atlanticus of about 1495, but Leonardo can no longer be regarded as its inventor. The St Albans clock had two similar escapements, one to control the going-train and one to ring the bell each hour on a twenty-four-hour system with the number of strokes equal to the hour. According to North, it is not inconceivable that such an oscillating striking device, triggered at suitably
-104-
chosen intervals by a hydraulic clock, pointed the way to the first mechanical escapement proper.
The oldest surviving clock in England is the Salisbury Cathedral clock that was made not later than 1386. It has no dial or hands but strikes the hours. The verge-period for a half-swing is four seconds. The clock was restored to its original condition in full working order in 1956, after a lapse of seventy-two years.9 Another more complete clock, believed to be by the same craftsman, that was in Wells Cathedral from at least 1392 is now in the Science Museum, London. (Both clocks were in due course converted to pendulum clocks.)
The accuracy of all early mechanical clocks was low, because the foliot and wheel had no natural periods of their own and also because of the effects of friction. It was, however, an age when civilization was becoming more vigorous and the number and skill of metal workers were increasing. A tremendous craze developed for the construction of elaborate astronomical clocks. As a leading historian of medieval technology has remarked, 'No European community felt able to hold up its head unless in its midst the planets wheeled in cycles and epicycles, whilst angels trumpeted and countermarched at the booming of the hours.'10 Outstanding among these 'clocks' was the astrarium of Giovanni de' Dondi of Padua, designed between 1348 and 1364. This complex instrument with its finely cut teeth and intricate gearing was made of brass and was smaller than the clumsy early English clocks that were made of forged iron. It was only incidentally a timepiece. Primarily it was a mechanical representation of the universe, a kind of planetarium. It was much more elaborate than the first of the famous series of astronomical clocks in Strasbourg Cathedral that was installed about the same time, 1350. The original Strasbourg clock probably contained, besides moving figures, an annual-calendar dial, and possibly a lunar dial and an astrolabe, but the instrument designed by Giovanni de' Dondi incorporated a perpetual calendar for all religious feasts, both fixed and movable, and also indicated the celestial motions of the sun, moon, and planets, including even the motions of the nodes of the moon's orbit, which take over eighteen years to make a complete revolution around the ecliptic.
It is not surprising that this remarkably complete astronomical clock attracted the attention of princes. It was acquired in 1381 by Duke Gian Galeazzo Visconti, an intellectual who has been described as 'a sedate but crafty ruler with a great love of order and precision'.11 He moved it to his palace in Pavia, where in 1420 it was recorded as being in the ducal library. It was very difficult to keep in working order, and when the
-105-
Fig. 4 A drawing of de' Dondi's astronomical clock.
This drawing, dated 1461, of a part of de' Dondi's astronomical clock is from a manuscript in the Bodleian Library, Oxford. ( MS. Laud Misc.620, fol. 10v.) This complex weight-driven instrument, which was completed at Padua in 1364, was regulated by a horizontal balance-wheel shaped like a regal crown. Two pallets were fitted to the verge of this wheel. The upper one was made to move by one of the 24 teeth of the escape-wheel, which turned the verge and balance-wheel so that this pallet then escaped and at the same time the lower pallet became engaged. The latter then turned the verge and balance-wheel the opposite way, and so the process continued. The balance-wheel had a beat of 2 seconds. In recent years a number of models of this clock have been made; one is in the Museum of the Smithsonian Institution, Washington, DC.
-106-
emperor Charles V saw it in Pavia in 1529it needed repair. Charles, who had a taste for mechanical devices, commissioned Giona
llo Torriano of Cremona to repair it, but owing to corrosion he found that this was impossible and he agreed to make a similar instrument. When Charles retired to the monastery of San Yuste in 1555, with a large collection of clocks and watches, he took Torriano with him. After Charles V died in 1558, Torriano entered the service of his son Philip II and moved to Toledo, where he died in 1585. A few years ago manuscript evidence was discovered that Torriano's copy of de' Dondi's clock was still at his house in Toledo in the seventeenth century, and it is therefore unlikely that, as formerly thought, it perished when the convent of San Yuste with its art treasures was set on fire by the French in 1809. In recent years a number of working reconstructions of de' Dondi's clock have been made.
In the course of the fourteenth century mechanical clocks became progressively more numerous in Europe, most of those that were not installed in churches being public clocks. One such was a striking clock designed by Giovanni de' Dondi's father Jacopo, on whom the surname ' del Orologio' was conferred. It was erected in the entrance tower of the Carrara Palace at Padua in 1344, but was destroyed in the assault on that town by the Milanese in 1390. Although they were expensive, public clocks were generally regarded as being very useful. Whereas church bells announced the times of the various religious offices, the communal clock was a secular instrument that struck the hours, and by the end of the fourteenth century some were made that struck the quarters, although this did not mean that they were any more accurate. They were often unable to keep time to within fifteen minutes a day and were frequently out of order. This is not surprising since all geared wheels had to be cut by hand.
Time in History: Views of Time From Prehistory to the Present Day Page 14