You’re reading Age of Invention, my newsletter on the causes of the British Industrial Revolution and the history of innovation. This edition went out to over 12,200 subscribers. If you’ve just been forwarded this, you can sign up here:

Previously, on Why wasn’t the Steam Engine Invented Earlier? we saw that the general idea of using heat and steam for mechanical work has a long history, and that when we talk about a “steam engine” in the eighteenth-century sense we really mean a machine that exploits the weight of the atmosphere compared to the vacuum resulting from rapidly condensing steam. This sort of sucking power, rather than steam’s pushing power, is the thing we’re trying to explain.

But then I revealed that the exploitation of this suction effect — by cooling and contracting either air or water vapour — also has an extremely long and continuous history going back to ancient times. (See also my video discussion with Carbon Upcycling on the topic). The suction effect from cooling was not only known in a general sense by a handful of isolated people, but was the centre-piece of sixteenth-century debates on the nature of the weather. It was being exploited since at least the 1590s, and probably centuries earlier, by alchemists. It featured in lots of designs for solar-powered fountains. And it was the basis for Cornelis Drebbel’s perpetual motion machine of c.1606, which worked by exploiting changes in the temperature and pressure of the atmosphere, and which quickly became famous throughout all of Europe.

Long before there was any proper understanding of the vacuum or of atmospheric pressure, the underlying principle behind atmospheric engines was in the 1600s-20s already being widely exploited — including for the most exciting, highly sophisticated, and widely-lauded invention of the age. And as we’ll see in this post, it was no mere flash in the pan. A water-raising atmospheric engine should have already been just around the corner.

And yet, it wasn’t. It was almost another century after the unveiling of Drebbel’s perpetual motion machine for a technology as practical and widespread as Thomas Savery’s 1690s engine to appear. In this post I’ll delve into the details of what happened exactly happened. And I’ll finally give an answer to the question of why the steam engine wasn’t invented earlier.

This is going to be a long post. So grab a drink, settle somewhere comfortable, and let’s get stuck in.

[Public service announcement: because of all the illustrations and references, there’s a high risk that this email may end up truncated for some people. If this happens, you can always read the post online, by simply clicking here.]

The invention of temperature

As I hinted in Part II, there were still more wonders to issue from Drebbel’s workshop — many of them building on the same principles as his perpetual motion machine.

So it’s worth a very brief recap of how that device worked. Drebbel had improved upon an ancient experiment involving an inverted flask in water: that is, to heat the base of a long-necked glass flask and place it mouth-first into a bucket of water. The heated air trapped inside the flask would bubble out, and as the remaining air cooled, the water of the bucket would rise up into the flask.

The inverted flask experiment. The air bubbles out on the left. As the remaining trapped air cools, on the right, the water rises up the flask (in fact pushed up by the pressure of the atmosphere).

Drebbel’s big breakthrough was to notice that once the water was already sucked into the flask, it would continue to rise and fall even when it wasn’t being heated or cooled on purpose — movements that were the result of natural changes to atmospheric pressure and temperature.

From this continued movement — to his mind, a harnessing of the perpetual movement of the universe itself — Drebbel then constructed a machine that seemed to show the ebb and flow of the tides, as the liquid inside it rose and fell between the cold of night and the heat of day. He also exploited that same rise and fall of the liquid in order to rewind clockwork that continually showed the time, day, months, and years, along with the cycle of the zodiac and the phases of the moon.

Few have now heard of Drebbel’s perpetual motion machine, but noticing that same rise and fall of the liquid in response to changes in the weather would also serve as the basis for the invention of the thermometer and barometer. Or, more accurately, to the reinterpretation of the ancient inverted flask experiment as a device capable of measuring both temperature and atmospheric pressure. (The two different applications were not disentangled and isolated until later, as we'll see below, and so in modern terminology the initial device is often referred to as an air thermoscope.)1

The first to do this appears to have been one Santorio Santorio, professor of medicine at Padua. In 1612 he published a book on Galenic medicine, in which he mentioned applying the inverted flask experiment to measuring temperature. (Santorio’s priority is often disputed, but there just isn’t the space here to go into the details, so I’ll write about this separately next week.) What Santorio did was to mark the level of the water in the inverted glass flask when it was at some normal temperature, and then used a pair of compasses to measure any changes from that level. According to Galenic medicine, the reference point for a temperate state — the “temperature” — was the body’s healthy balance between four Galenic “degrees” of both hot and cold either side of it.2 When we talk about temperature today, we thus still use the Galenic vocabulary.

Santorio soon developed the concept further, marking the level of the water in the inverted glass flask when a candle was held to the air trapped at the top of the flask — the water would be low — and marking the level of the water when the device was exposed to the winter snow, when the water would be higher. Having thus defined both a maximum and minimum for his scale, he then subdivided it into degrees, using it “to measure the cold and hot temperature of the air of all regions and places, and of all parts of the body”.3

By 1617, Santorio’s device was being referred to by others as the thermoscopium, and in a French work of 1624 as the thermometre.4 It was a major breakthrough in visualising hot and cold, and more and more people were excited to discover that the thermoscope’s readings could often contradict the body’s senses. The thermoscope unleashed a frenzy of experimentation, data collection, and further development. Consider, for example, the problem of how to calibrate instruments, to create a reproducible scale, or to even identify reliable fixed points from which to derive such a scale. Does water always freeze at the same point? And how would you even know? Was cold a phenomenon distinct from heat, or just heat’s absence?5 Such questions would take well over a century to resolve.

Thunder Glass

At about the same time, however, in England and the Low Countries the thermoscope initially went by a different name — perhaps because it had a different inventor.

It was Drebbel, after all, who had first applied the inverted flask experiment. He had used it to run some incredibly complex clockwork, and as we saw in Part II had even made it the basis for an entire, novel theory of the weather. So it was no great leap for somebody — either Drebbel himself, or one of his many emulators — to notice that his perpetual motion could sometimes predict the weather too. When there was going to be bad weather, for example, the pressure of the atmosphere dropped below that of the air trapped within an inverted flask, drawing the water down and out of the glass. As a 1641 London advertisement put it, “the sudden falling of the water is a certain sign of rain”.6 An even more pronounced drop indicated that a storm was on the way.

A Dutch-style weather-glass, often J- or U-shaped, signalling an impending storm.

Thus, while in Italy the inverted flask experiment was reinterpreted as a thermometer, in England and the Low Countries it first gained popularity as a kind of barometer (as I mentioned above, the two countervailing effects of temperature and air pressure were not separated until later). In Dutch, it came to be known as the donderglas (i.e. thunder-glass), and in England as the weather-glass or kalender-glass (in the seventeenth century, drawing up a calendar was not always just a matter of listing dates for holidays, but of using astrology to predict the weather for the days ahead). By 1620, Francis Bacon was already even referring to it in Latin as the vitrum calendare, suggesting that the English word had already come into widespread use.7

It is unclear exactly when or by whom the inverted flask experiment’s weather-predicting properties were first noticed. My money would be on Drebbel, but the earliest mention is from 1619, in the Belgian city of Ghent, where one Ghijsbrecht de Donckere appears to have developed a simplified version of Drebbel’s perpetual motion and sold it as being able to predict storms.8 Regardless, it soon became clear that the device was practically identical to the thermometer. Indeed, it came to be known as the “Drebbelian”, or “Dutch” thermometer, often distinguished from Santorio’s “Italian” version only by its shape. The Dutch weather-glasses were typically J- or U-shaped, holding their water at the bottom of the bend, rather than being a simple straight flask placed into an open bowl.9 This was, perhaps, because they were just less pronounced versions of Drebbel’s perpetual motion machine, its not-quite-closed O shape evolving into a U and then a J. But other than in shape, the Italian and Dutch versions were essentially the same.

Art historian Floor Koeleman suggests that this detail from “Allegory of Hearing” by Jan Brueghel I and Peter Paul Rubens may be of Donckere’s device. It seems to be a not-quite-closed O, like a simplified Drebbel’s original perpetual motion.

By the 1620s, then, the inverted flask experiment had produced the technological wonder of the age, and had been repurposed as both a thermometer and barometer. And so all across Europe, people turned their minds to exploring its possibilities. An English physician, Robert Fludd, became obsessed. He saw in the thermometer the possibility of measuring all fundamental dichotomies, and not just hot and cold: light and dark, heavy and light, fullness and emptiness, sickness and health.10 Others began more practical attempts to isolate the effects of temperature and atmospheric pressure. In the 1640s the Grand Duke of Tuscany, Ferdinand II, may have been the first to rid the thermometer of the complicating influence of the atmosphere — his big improvement was to seal up its end.11 (The invention of a separate barometer was rather more convoluted. We’ll get to that in a bit.)

The inverted flask experiment was thus at the very cutting edge of early seventeenth-century science. And so the basic principles behind the atmospheric engine — that of raising water by cooling a gas above it — were being thoroughly discussed and debated. Indeed, in the 1620s Francis Bacon noted the “various uses of the motion of dilation and contraction in the air by heat” as one of the main targets for systematic investigation, and one that was already bearing fruit. Bacon listed the thermometer, solar-activated musical instruments, Hero of Alexandria’s machines, and Drebbel’s perpetual motion as practical examples of the research programme’s early results.12

But Drebbel was actually already doing so much more. Because while Santorio was trying to simply measure temperature, Drebbel had found a way to control it.

The Philosopher’s Stove

For alchemists like Drebbel, being able to control the temperature of furnaces and ovens was a valuable prize, because so much of their skill in manipulating metals and minerals depended upon it. The alchemist’s art — the intangible, tacit skill built up over years of experience — was one of sensitivity to heat, being able to judge, by feel and by look, the varying intensities of flame, and then to manipulate it so as to keep it at a constant level. The art was known as pyronomia, or as regimen ignis — the governing of fire.

At some point before 1624, Drebbel worked out that he could exploit the inverted flask experiment to radically improve furnaces. He did this in two ways. One was simply to affix a mercury thermometer to the furnace, to indicate its heat (mercury, with a higher boiling point, would be less liable than water to simply evaporate away). But the other, and more ingenious way, was to create a feedback mechanism to control the oven’s heat automatically. Drebbel placed a cork to float atop the mercury in yet another thermometer, which as it rose or fell would then cover or uncover the furnace’s air supply. He could thus choose a desired heat, and then let the oven do the rest. If it grew too hot, the air supply would be restricted. If it grew too cold, it would be increased.13 Drebbel had invented the thermostat, and perhaps one of the first widely-applied practical feedback control mechanisms. (Drebbel’s self-regulating oven is, incidentally, also the answer to the last invention quiz! Of the answers I received, Barry Devlin was definitely on the right track, but nobody quite got it.)

Drebbelian self-regulating ovens, or Philosopher’s Stoves, spread beyond England, to be adopted in France, the Netherlands, Germany, and even across the ocean in New England — they were a major source of business for the husbands of Drebbel’s daughters, the brothers Abraham and Johannes Sibertus Kuffler, to whom he passed many of his secrets. Drebbel even applied its thermostatic principles to artificially incubating eggs, for which maintaining a constant temperature was essential. To give an idea of how big a deal this was, Francis Bacon filled his techno-utopian vision of a New Atlantis with “furnaces of great diversities, and that keep great diversity of heats; fierce and quick; strong and constant; soft and mild; blown, quiet; dry, moist; and the like”, some of which, like the incubator, were able to provide even the gentle heat of animal bodies. Drebbel, in Bacon’s lifetime, was thus producing the stuff of science fiction. He was, as one admirer termed him, a true Mysteriarch.14

And the Mysteriarch did not stop there. Just before his death in 1633 he was working on improving the stoves, making them more efficient, reducing the need for people to attend the fire, and reducing their smoke. They could thus be applied to drying hops, malt, fruit, spices, and gunpowder, heating rooms in houses, and distilling fresh water from sea water.15 His heirs, the Kufflers, even made the stoves portable enough to be used for baking the bread for armies — they were allegedly used by Frederick Henry, Prince of Orange, in his various successful campaigns against Spain.16 Both the portable ovens and the seawater distilling machines were apparently used in the 1650s aboard ships headed for the Indian Ocean.17

By the 1620s, then, many of the key elements for a steam engine were already coming into fairly widespread use. Scientists across Europe, inspired by Drebbel’s perpetual motion and Santorio’s thermometer, were eagerly pursuing the possibilities from expanding and contracting gases. And Drebbel had invented a widely-used thermostatic feedback system — a general concept that would later prove extremely useful in making steam engines practicable. Feedback systems and safety valves would come to regulate the movements of engines and reduce the risks of them overheating and exploding.

Even the general idea of applying the contraction of gas to motive power was also gradually coming into use, though thus far only for moving very light weights. Drebbel had already used it for resetting clockwork in his perpetual motion machine, and for moving the cork to open and close the shutters regulating his stoves and incubators. The manifold makers of weather-glasses and thermometers were also playing around with similar ideas, including a 1630s version that moved the entire inverted flask apparatus up and down, in a way that is remarkably close to being a piston:18

A weather-glass as almost-piston, one of a wide variety likely available for sale in London by 1635. As the air trapped in the flask KL cools, it draws the water up into it, so that the falling level of the water in AB moves the whole flask and board down. When it warms and the water is expelled out of the flask, the piston rises.

Given all this excitement and investigation in the 1620s, it feels as though work on a Savery-style atmospheric engine should have been just around the corner — especially after the centuries, if not millennia, that the inverted flask experiment and its permutations had already been known and in use. Even all the incentives for a water-pumping machine had long been in place. By the early seventeenth century vast sums were being spent throughout Europe on new equipment to drain marshes and raise the water out of mines, with both England and the Netherlands alone seeing dozens of innovative designs for new water-, wind-, and muscle-driven pumps.

In fact, even Savery himself seems to have been inspired by the age-old inverted flask experiment. After his death, the scientist John Theophilus Desaguliers recounted Savery’s own story of how he had been inspired to investigate steam. It is almost word-for-word a description of the inverted flask experiment: “having drunk a flask of Florence [wine] at a tavern, and thrown the empty flask upon the fire, he called for a basin of water to wash his hands, and perceiving that the little wine left in the flask had filled up the flask with steam, he took the flask by the neck and plunged the mouth of it under the surface of the water in the basin, and the water of the basin was immediately driven up into the flask by the pressure of the air.”

(Desaguliers thought that Savery had made it all up, accusing him of having instead plagiarised another inventor, and because he unsuccessfully tried to repeat the experiment. But it sounds like Desaguliers misunderstood him and got all sorts of crucial details wrong. Rather than placing a near-empty flask near a fire to warm it, as Savery's own story went, he instead brought half a flask of wine to the boil directly over the fire and so massively over-heated it. Holding it with a thick glove, his plunging it in the cold water “beat the flask out of my hand with violence, and threw it up to the ceiling.”)19

So what was the hold up? Did any work towards a Savery-style atmospheric engine take place in the three quarters of a century between 1620 and the 1690s? Once you know to look for mentions of perpetual motion and the inverted flask experiment, the answer is almost certainly yes. And to better understand what took so long, it’s worth looking at those attempts. It’s time for me to introduce you to one of the unsung mechanical geniuses of the seventeenth century, the gunmaker Kaspar Kalthoff.

Kalthoff and Petty

Kalthoff is a strangely under-studied figure. Since the mid-nineteenth century he has often been name-checked in the history of the early steam engine, but really only as a sort of footnote to the story of Edward Somerset, Lord Herbert, who in 1646 succeeded his father to become the second Marquess of Worcester. In 1663, Worcester published a book in which he listed a hundred inventions he had been working on since the 1630s. They are described only vaguely, but number 68 mentions using fire to drive up water, and number 100 — for which Worcester obtained a 99-year patent by Act of Parliament — is described as a “water-commanding engine”.

It is tantalising stuff, and has been since the early eighteenth century. Worcester was the person whom Savery was accused of plagiarising. Desaguliers even spread a rumour that Savery had bought up copies of Worcester’s book to cover his tracks.20 Albeit baseless, the story captured the imagination of the Victorians, securing Worcester’s fame. In the 1860s, Worcester’s grave was even broken into by Bennet Woodcroft, a key founding figure for what is now the London Science Museum, because he suspected a model of the supposed steam engine had been buried with him. (It had not.)

Based on the available evidence, however, Worcester did not invent an atmospheric engine, let alone anything like Savery’s. The so-called water-commanding engine was actually mechanical, with eyewitnesses reporting that it involved buckets and being turned by a person. And as for the method of using fire to drive up water, Worcester’s own, few words are unequivocal. His method used the expansive force of steam to push the water up, and specifically “not by drawing or sucking it upwards”.21 So much for the legend.

But with all the exciting myths surrounding the Marquess of Worcester, nobody seems to have properly looked into his employee, Kaspar Kalthoff. The last paper to mention Kalthoff in the title was written in 1947,22 and pretty much anything written since only mentions him in passing or in different contexts: he also invented a very early repeating rifle, and was a talented grinder of lenses for microscopes and telescopes. Kalthoff has thus been overlooked, and undeservedly so, because he may actually have had something to do with a very early atmospheric engine — an engine that, like Savery’s almost half a century later, exploited the apparent sucking force of condensing steam.

Kaspar Kalthoff was from a large dynasty of gunmakers who were active throughout Europe. I’ve not yet been able to tell for sure whether they were German or Danish, but for most of his career Kalthoff was based in England and the Netherlands. By the mid-1630s he was working for the king, based out of what was essentially a royal military research and development establishment at Vauxhall, just south of London. He was seemingly also doing some work on the side for Worcester, as in 1638 they’re mentioned together in a letter by Samuel Hartlib, a London-based German enthusiast for all things invention, and a key node in the Republic of Letters. Hartlib reports that they had invented a perpetual motion, but that king Charles I had forbidden them from revealing it to anyone.23 The use of the term perpetual motion is immediately interesting here, given the association with Drebbel’s device, but a year later Hartlib describes Kalthoff's device as involving fifteen wheels on a single axis,24 and that it could lift no more than four pounds.25 It doesn’t sound at all like an early atmospheric engine. I suspect it was some kind of spring-based mechanism, as Worcester is next mentioned as having made a wind-up horseless carriage, which unfortunately “by reason of the intolerable slowness … is worth nothing”.26

Hartlib’s surviving records are abruptly and frustratingly lacking for the period 1642-48, but it seems as though Kalthoff fled the country. As well as being a military inventor in the service of the king, he may have had royalist sympathies. He seems to have produced a bullet-firing crossbow, or stonebow, which was considered for an assassination attempt on Oliver Cromwell.27When Kalthoff next resurfaced in 1649, however, it was while working near Dordrecht in the Netherlands on installing a water-pumping machine that used steam.

The key document is Hartlib’s transcription of a letter that was sent by one Benjamin Worsley, who had visited Kalthoff, to William Petty — a name well-known to many historians for his works on economics and statistics. Petty, it turns out, had also been working on some kind of steam engine, and both Hartlib and Worsley were trying to get the two inventors to join forces. Although this scheme came to nothing, Worsley acted as a sort of conduit for a conversation between Kalthoff and Petty about their inventions, with the correspondence allowing us to glean some valuable insights into how they worked.

For a start, both inventions certainly burnt a fuel. Kalthoff told Worsley that his invention would be most useful for pumping the water out of coal mines, as the fuel would be so cheap.28 And Petty, when asked whether his invention might be applied to grinding sugar-canes in Barbados, replied that he was unsure if there was plentiful enough combustible material available on the island.29 Kalthoff also cautioned Petty that when he scaled up his device, he’d have to consider the extra “flame” it would require, along with the added fuel costs and smoke that this would bring.

Both inventions also certainly used steam. Petty boasted that his was “the greatest art and rarity that of all the three elements (air, fire, water) has been in the world”30 — as we saw last time, air and water as elements were more akin to what we’d now call the states of gas and liquid. To the seventeenth-century mind, steam was thus a form of air, transmuted from water thanks to fire. Kalthoff also told Worsley that he had already spent some years investigating Petty’s hypothesis, “having taken notice of that merry Christmas sport of making water rise out of a pail into a basin by a lighted card” (which sounds remarkably like some version of the inverted flask experiment). Kalthoff even noted that he had failed to raise water very high using Petty’s method, and so had begun to consider ways “of raising it by strong heat, or ordinary fire applied about it (as I have seen such a kind of thermometer)”. By placing “a pipe full of water upright, into a vessel closed with water … and applying … fire to the bottom”, he had raised a continuous spout of water.

But were either of them using the sucking force of condensing steam, rather than just steam’s expansive, pushing force? The explicit mentions of both the Christmas trick and the thermometer strongly suggests to me that both had the inverted flask experiment in mind — so much so that when I first read the letter I almost fell off my chair.

And there’s the way that Kalthoff described some of his experiments. After some small trials he had made a large vessel of lead, much like Petty apparently had, which “striving to put nature to extremity, the sides also of the vessel fell in”. Although he then made larger and more robust versions of first lead and then copper — both of which were left at Vauxhall when he fled the English Civil War in the early 1640s — the mention of the sides falling in, and not out, sounds a lot like what would happen if you created a vacuum inside a sealed chamber through condensation, rather than by over-pressurising it with steam.

Kalthoff also suggested to Petty that his cistern be “divided into two pipes, that the one run out, while the other draws”, as otherwise the pumping would be “hugely interrupted and stopped”. Albeit vague, it sounds suspiciously similar to the 1690s Savery engine, with one vessel cooling and drawing the water from below, while the other is heated and pushes the water further up and out.

When I first wrote this draft I had to conclude that the available evidence was tantalising but ultimately frustrating. Indeed, Hartlib’s copy of the letter is all torn at the edges, losing us some valuable details. One of the lost lines even begins with Kalthoff telling Worsley “the hypothesis and ground of his invention” before abruptly being cut off. Agony.

But at the last minute, having decided to do just one more check through Hartlib’s correspondence for anything that may have been mislabelled, I yesterday stumbled across the smoking gun: a letter that Petty had earlier written to Worsley, asking him to “fish out of” Kalthoff some information on his invention. Specifically, Petty asked “whether he did ever raise water either in great or small pipes above 32 feet high or thereabouts. If he says he can do it, still keep him to the question whether ever he did do it, by one operation or without repetition of the same operation. Herein I find difficulty though the force be great enough”.31

It may not seem like much at first glance, but for me these lines are conclusive proof that Petty was using the sucking power of condensing steam. Because there is a physical limit to how high suction can raise water. And it just so happens to be about 34 feet (~10 metres).

The Limits to Suction

This physical limit applies to all kinds of suction devices, not just to those using condensation. Regardless of whether you use a bellows or piston or siphon to try to raise water above 34 feet in a vertical pipe, you will always, inevitably fail. The pipe could be as wide or narrow as you liked — as Petty said, “either in great or small pipes”. You could even, as Petty also said, feel like you had force to spare. But beyond that level the water will simply never budge higher. As Petty implied, the only way around it would be to repeat the operation using a separate suction device, once the water had already been raised to a new level. Or else to complement the suction with some other kind of pushing force.

For many years, the exact limit to suction was not well understood or even explored. Salomon de Caus in 1615 placed it at about 26-30 feet, and Galileo in 1630 estimated it at about 40 feet, before revising it downwards in 1638. But it wasn’t necessarily appreciated as an inviolable law of nature. De Caus seems to have blamed it on just the limitations of tools and materials, and Giovanni Battista della Porta doesn’t seem to have recognised any reasons why a suction device, at least in theory, could not be used to raise water by as much as 100 feet.32

Yet in 1630 the limit was brought to Galileo’s attention by a Genoese correspondent named Giovanni Battista Baliani, who had tried to build a copper siphon that would raise water over a hill more than 34 feet high. No matter how he tried, having filled the siphon full of water with its ends both closed, the moment he opened them up the water within would become separated, as the water would rush out down to a level of about 34 feet. But then that left the troubling question of what was now left in the sealed space above the water. Baliani, like others of his era, believed that creating a vacuum was extremely difficult, if not impossible. But there were no obvious gaps or cracks in the copper pipe, through which any air could have got in.

The trouble with Baliani’s siphon

Galileo’s reply was that the weight of the water itself must be the problem. No matter how narrow or wide the pipe, and no matter the incline, the column of water within the siphon was much like a physical rope — at a certain point it would no longer be able to support its own weight and would have to snap. Galileo then reasoned that the space left above the water must be a vacuum after all: a claim he would publish in 1638.

This revolutionary idea — that a vacuum was not just possible, but quite easily created — was soon the object of heated debate and further experimentation. By 1643, at Rome, one Gasparo Berti set up a series of experiments to test the explanation, hoping to prove Galileo wrong. He created a lead pipe over 34 feet / 10 metres high, with a glass sealed to its top through which to actually see the vacuum. Having filled it with water while sealed at the bottom, which he placed in a bucket of water, he then unsealed the bottom. But to his disappointment, and as Galileo predicted, the water fell to the limit, leaving an apparently empty space in the glass.

This was, unfortunately, still inconclusive. Some held to the idea that light could not travel through a vacuum, so the fact that they could see inside it at all suggested there might still be some particles in there. Others believed that air, or perhaps even the mysterious fifth element, quintessence, must have seeped through invisible pores of the glass itself in order to fill the space. Berti even placed a bell inside the glass, knowing that a vacuum would prevent its sound from travelling — the problem was that it could still be faintly heard, as the yoke supporting it was inevitably still attached to the glass wall.33

A little later, Evangelista Torricelli and Vincenzo Viviani repeated the experiment in miniature by using a denser liquid, mercury, allowing them to also use a much shorter pipe (just over 2 feet) entirely made of glass. They thus created a much easier way to both view and create a vacuum, and demonstrated that no matter how large they made the space at the top, the mercury would always fall to the same level.

Torricelli and Viviani demonstrate that mercury always fell to the same level AB, even though the size of the space above it varied.

But what explained the level that the mercury or water could never exceed? Torricelli and Baliani both came up with similar explanations, that it was really determined by the relative weight of the air. Suction effects, they argued, were not caused by the universe itself abhorring a vacuum. They were instead caused by the atmosphere pushing a column of liquid up until their weights balanced out. Beyond that point, the atmosphere could obviously push the liquid no higher. As Torricelli so beautifully put it, the experiment seemed to show that “we live submerged at the bottom of an ocean of elementary air”. (And in the process, by creating what was essentially a weather-glass with a vacuum rather than air trapped at the top, he invented a version that wouldn’t suffer the interference from temperature: a barometer.)34

Now, in the traditional story of the steam engine, Torricelli’s thoughts on atmospheric pressure and the vacuum are usually treated as the beginning. But as we saw in Parts I and II, the same forces were being exploited long before they were correctly understood. Petty was certainly aware of Torricelli’s experiment — in 1648, just a year before his own steam engine experiments, he had received word of it being replicated in Paris, though with no mention of the idea of the atmosphere having a weight, and couched in very sceptical terms about whether it could really create a vacuum.35 Petty may also have read Galileo’s 1638 work, which at least identified the limits to which suction could apply. Or perhaps, as I suspect, Petty simply bumped up against the limit himself when trying to create his engine. His words that “herein I find difficulty though the force be great enough” suggest he had already repeatedly tried and failed to exceed it, while holding to the idea that it might still in theory be possible.

Regardless, knowledge of the existence of the limits of suction was no real barrier to inventing an atmospheric engine — it was only a barrier to scaling it up. We can, I think, say with certainty that Petty’s steam-using engine attempted to raise water by suction. And having bumped up against suction’s limits, he was thus eager to know whether Kalthoff had found a way to exceed them.

Kalthoff clearly understood the implication of Petty’s questioning. It was why he mentioned the Christmas trick and could talk of having initially tried to follow the same hypothesis. He knew that Petty was talking about condensing steam to create suction. But Kalthoff, having already discovered suction’s limits for himself, appears to have switched his attention to exploiting steam’s expansive force — either instead of atmospheric pressure, or perhaps in combination with it, as in Savery’s 1690s engine. Crucially, Kalthoff claimed that the prototype he left behind at Vauxhall could raise water 60 feet — clear proof that it did not use atmospheric pressure alone, if at all.

As for the engine Kalthoff was trying to erect at Dordrecht in 1649, he was sure that his engine was “wholly and altogether different”. Indeed, Petty interpreted his reply the same way: he wrote back to Worsley that “I am glad to hear Mr Kalthoff’s motion is toto caelo [to the whole extent of the heavens] different from mine, supposing thereupon that there is more Art in the world than I thought there was, and wish his profits may be according to his deserts.”36 It all suggests that by 1649 Kalthoff had shifted entirely to using the expansive force of steam on its own.

Why the wait?

Kalthoff, then, at some point before the English Civil War was experimenting with condensing steam to raise water for industrial uses like draining mines. And Petty’s engine of 1649 attempted the same. Yet neither of them took off.

Petty unfortunately seems to have lost interest in the invention, focusing instead on more immediately profitable endeavours. And Kalthoff was a major victim of circumstance. He had been expecting a large reward from Charles I for the engine at Vauxhall, but when he fled he left all his prototypes behind. Having had his assets seized by England’s revolutionary Parliament, he then also lost a fortune from the collapse of the Dutch West India Company in 1645-47. And his 1649 pumping machine near Dordrecht — apparently financed by some English merchants — was then allegedly burned down by some angry locals.37 (A rival engineer suggested that Kalthoff set the fire himself, to hide the invention’s failure,38 but this seems unlikely given he immediately set about trying to rebuild it). Although he later returned to England and worked for his old patron the Marquess of Worcester at Vauxhall again, by then he seems to have moved on to other projects.

Now, one might suggest that Petty and Kalthoff simply lacked the tools or materials to make sufficiently strong and precisely fitting vessels and pipes. But I highly, highly doubt this. In 1640 Kalthoff had invented a method “of boring into brass and iron as if it were into wood”,39 which greatly impressed his contemporaries — well over a century before James Watt’s improved steam engines relied on the precision of John Wilkinson’s cannon-boring machines. He also improved minting machinery, and by 1649 had developed a device that “makes his tools or files work of themselves”.40 When reports or drawings of foreign inventions made their way to England, he was the first person Hartlib had in mind to create copies.41 Kalthoff was thus an early pioneer of machine-making, even if he’s largely forgotten today.

Nor do I see any reason to believe that the tools or materials available to Savery in the 1690s were in any way superior. If anything, Kalthoff as an inventor and maker of so many precision tools probably actually had much better access to them. As one of the best rifle-makers in Europe, he was exactly the person to manufacture precise parts for any kind of engine. And even if either he or Petty couldn’t have made some of the parts themselves, Hartlib’s records reveal London in the 1640s and 50s to already have been a major hub for engineering talent. Petty’s name in particular often crops up as an important source of information on the city’s many mechanics and inventors.

So why did it then take almost another half a century, even with the increasingly widespread understanding of atmospheric pressure and vacuums, for a widely-adopted and practical atmospheric engine like Thomas Savery’s to appear?

The answer, I think, is what it almost always is: that inventors are simply extremely rare. People can have all the incentives, all the materials, all the mechanical skills, and even all the right general notions of how things work. As we’ve seen, even Savery himself was apparently inspired by the same ancient experiment as everyone else who worked on thermometers, weather-glasses, egg incubators, solar-activated fountains, and perpetual motion machines. But because people so rarely try to improve or invent things, the low-hanging fruit can be left on the tree for decades or even centuries.

The development of the steam engine, rather than being a story of the Torricellian vacuum science unlocking a new technology, is instead just like that of textile machinery, signalling systems, and any number of other “ideas behind their time.” The rate at which such inventions appear is really down to the number of people applying themselves to improvement, and the strange-seeming delays are down to there so often being so few. When only a handful of people are inventors, is it really any wonder that the circumstantial setbacks and distractions that prevented a Petty or a Kalthoff can end up delaying things for decades? As a later steam engineer William Blakey put it, when complaining of the lack of improvement to even the eighteenth-century steam engines, it all shows “how much we are subject to keep to the routine we have been trained in, and that few men appear capable of invention.”42 Or, more accurately, that too few people even think to try.