In 1630 a physician named Jean Rey, who lived in the small town of Le Bugue, in the Dordogne, published a book describing some chemical researches that were, in some ways, far ahead of their time.' Rey believed that the increase in weight when tin and lead are heated in air is due to air condensing and adhering to these substances. This idea brought him a long letter from Marin Mersenne in Paris, dated September 1, 1631,2 in which Mersenne very naturally objected that air is not condensed by heat, but rarefied. As an example he adduced the air thermoscope: "Then the thermoscope, making the liquid descend by the rarefaction of its air, bears witness that heat makes air more subtle, unless denser air comes down to take its place." Rey wrote an answer on January 1, 1732, at even greater length. As to the thermoscope, it seems evident that he had never even heard of the ones commonly in use, for on page 244 he writes:

there are a variety of thermoscopes or thermometers, or so it appears. What you say cannot agree with mine, which is nothing more than a little round phial with a very long and slender neck. To use it I put it in the sun, and sometimes in the hand of a feverish patient, after first filling all but the neck with water. The heat, expanding the water, makes it rise more or less according to whether the heat is great or small. If I knew the construction and use of the one you speak of, I believe that the difficulty would be easy to resolve.

Thus, with no help from outside, Jean Rey had invented a liquid-in-glass thermoscope. But Mersenne does not seem to have recognized the importance of this, and it was forgotten. Apparently, Rey's thermoscope was not sealed up so as to make a permanent instrument.

When the variability of the pressure of the air became known in the years following 1644,5 a capital defect of the air thermometer at once showed itself; namely, that it responds to changes of pressure as well as of temperature. Pascal mentioned this when he was describing the celebrated barometric experiment on the Puy-de-Dôme:

From [this experiment] there follow many consequences, such as… the lack of certainty that is in the thermometer for indicating the degrees of heat (contrary to common sentiment). Its water sometimes rises when the heat increases, and sometimes falls when the heat diminishes, even though the thermometer has remained in the same place.

This consequence does not really follow from the Puy-de-Dome experiment but rather from the variability of atmospheric pressure. However, the times were plainly right for the invention of a thermometer of a quite different sort. It is universally believed -and I have found nothing to cast doubt on this belief--that the sealed liquid-in-glass thermometer was invented by a man of a very different station from Dr. Rey, none less than the Grand Duke of Tuscany, Ferdinand II, one of the great family of the Medici. Even after a very great allowance has been made for the desire of courtiers to flatter, the scientific ability of Ferdinand, and of his brother Leopold, shines brightly in the learned manuscripts of the period. Unfortunately, Ferdinand's grasp of politics was far inferior to his scientific acumen, and the prestige of Tuscany declined sadly during his reign.

Ferdinand, we may believe, invented the sealed thermometer and it remains to assign a date to this invention. A good deal of the difficulty of doing this vanishes when we recognize that two quite different sorts of thermometer were invented in Florence before 1660. These are shown in Figs. 2.1 and 2.2 respectively, taken from Plate I of the famous Saggi of the Academia del Cimento, which will come into our story later in this chapter. The first of these consists of a tube of liquid (spirits of wine) containing a number of glass balls or similar objects in which the ratio of weight to volume varies. If the temperature is very low and the density of the spirit correspondingly high these will all float; but as the temperature rises they will sink one after an other, so that the temperature of the spirit can be estimated from the number that have sunk. The other thermometer will be immediately clear to the twentieth-century reader.

Various authors have stated that the sealed thermometer was invented in Florence not later than 1641. This statement almost certainly stems from an entry in the original diary of the Academia del Cimento, dated June 20, 1657, which reads, "A thermometer made 16 years ago was cut open with a diamond." The object was presumably to see whether the spirit had altered.

This would seem to bring the date to 1641 without any question; but what kind of thermometer was this? To find the most probable answer we have to turn to the travel diary of Balthasar de Monconys, who went around Europe at intervals between about 1645 and 1665, buttonholing whatever celebrities he couldespecially the virtuosi and making notes, often obscure but always copious, on what he heard and saw. On November 7, 1646, he was in Florence and saw Torricelli,

who told me that the Grand Duke had various thermometers for determining heat and cold, all with eau de vie and glass balls full of air, but one in which there are two balls, one at the top and the other at the bottom; when it is hot the lower one rises, and when it is cold the one at the top descends. He told me of another which has a ball half full of air and half full of water, with a hole in the bottom, and prevented from rising by a glass chain; when the air condenses, more water goes in, and so the chain shortens and the bottle descends; on the contrary, when the air expands, the water goes out, the bottle rises and the chain is longer.

The first and second of these are clearly the thermometers with the glass balls, and there is no sign of the more familiar form. The third kind is of interest in that it must be one of the earliest applications of a chain as a weighing machine, antedating by at least a year or two Pascal's suggestion for measuring atmospheric pressure in this way.

We may, therefore, be fairly certain that the termometro infingardo was invented in or before 1641. What about the Florentine thermometer of the more familiar sort?

One red herring must be helped to evaporate from the trail. In 1929 Giuseppe Boffito wrote that in 1643 Claude Guillermet de Berigard, who used the pen name Berigardus, gave the first illustration of a liquid-in-glass thermometer, calling it "a rather common glass instrument for measuring the degrees of heat and cold." But when we examine Berigardus' booki we find (pace Boffito) an illustration (Fig. 2.3) and description of what is most certainly an air thermometer. The solution to the mystery is very simple; in 1661 there was a second edition, published at Padua, in which the illustration was changed to a spirit-in-glass thermometer on a base, destroying the correspondence between text and figure. This had been noticed in 1909 by Galli.14 It would seem that Boffito must have seen the edition of 1661.

As far as I have been able to determine, the earliest date for the invention of the "Florentine" spirit-in-glass thermometer, that can be fully documented, is sometime in 1654. What is quite certain is that by December of that year a number of comparable thermometers with a scale having fifty degrees had been made and sent to observers at Parma, Milan, and Bologna, at least. For on December 22, 1654, Antonio Terillo wrote from Parma to Luigi Antinori in Florence a letter (Fig. 2.4) that has been preserved, 15 which begins,

I have received the letter from your Reverence, with the box, in which were the two ampoules for measuring the variation of heat and cold; and keeping them for some time in the same place in a room I reflected that they both moved at the same rate without any considerable difference occurring. I made the liquid unite by getting out the air that was mixed with it; and I have fastened them outside two windows, one facing south, the other north, and I am observing them three times a day.

A few sample observations follow, which make it evident that these were the so-called "50-degree thermometers. Antinori also prints letters from the observers at Milan (December 29, 1654) and Bologna (January 2, 1655)10 acknowledging the receipt of similar instruments.

Of all the observations in this, the first meteorological network, organized by Luigi Antinori at the orders of Ferdinand II, only those made at the Monastery of the Angels in Florence have survived. V. Antinori was able to publish the complete series from December 15, 1654 to March 31, 1670, with short interruptions.1 We are given the temperatures of the northwardfacing and southward-facing thermometers and the state of the sky almost every hour during the day.

Thus, the spirit-in-glass thermometer certainly belongs to 1654, and probably earlier; at any rate many were available by Decem ber of that year. It is quite likely that a search through the enormous mass of manuscripts at Florence might establish a date a year or two earlier.

The Florentine thermometer is universally associated with the Academy of Experiments (Academia del Cimento) founded in 1657 by the Grand Duke, and a few words about this remarkable institution may not come amiss. In the ten years of its existence it made an extensive series of experiments which may fairly be said to have laid the foundations of experimental physics, and in 1666 published at Florence a corporate account of these as Saggi di Naturali esperienze fatte nell' Academia del Cimento. The printing was apparently continued into 1667, most copies having this date. There were further editions published at Florence in 1691 and 1841 and a facsimile of the first edition in 1957. An English translation by Richard Waller appeared in 1634 and a Latin version, with much commentary by the translator Petrus Van Musschenbroek, at Leiden in 1731. The most useful edition is the third Florentine edition of 1841, for it ends with a long account by G. Gazzeri of the surviving diaries of the Academy and is prefaced by 130 pages of the most resounding Italian by Vincenzio Antinori. While this introduction is intended to cele. brate the glories of Italian, and especially Tuscan science--after all, the edition was being paid for by the then Grand Duke- it nevertheless contains a good deal of the history of the Academy and the lives of the members.

The Academy was disbanded in 1667. It is probable that this was due mainly to the opposition of the Church. The Medici family wanted a Cardinal's hat for Prince Leopold, and this arrived in 1667, presumably with conditions attached. But according to Antinori the dissolution of the Academy began earlier and was largely the work of the Neapolitan member Gianalfonso Borelli, 18 whose hatred of Vincenzio Vivian is notorious. However this may be, the end of the Academia del Cimento marked the end of Italian predominance in the new sciences of the seventeenth century.