Cold Comforts: Natural Refrigeration in the 1632 Universe

In Walden Pond, Henry David Thoreau declared, “it appears that the sweltering inhabitants of Charleston and New Orleans, of Madras and Bombay and Calcutta, drink at my well.” What he meant by this was that the ice cut from Walden Pond would travel, in Boston ships, to those far off places, and be used to chill drinks.

Mark Huston, “Refrigeration and the 1632 World: Opportunities and Challenges” (Grantville Gazette 8) addressed the prospects for artificial refrigeration in the 1632 universe. It will happen, and it will happen a lot sooner than it did in the old time line (OTL), but it will not happen right away and it will not happen everywhere at once. For one thing, all the appropriate refrigerants are in short supply, and those which are most readily available are also rather dangerous to use.

A short-range snow and ice trade existed before the Ring of Fire, and a long-range trade sprang up in the nineteenth century and was quite profitable. I believe that there will be a window of opportunity in which it can prosper in the new time line until it is finally eclipsed by modern refrigeration. Bear in mind that the same body of up-time knowledge that makes artificial refrigeration possible also enables the building of steamships and railroads, which will minimize the time necessary to transport ice over a great distance.

While the principal purpose of this article is to explore the possibilities for long-range trade in natural ice, what is said here concerning ice storage, transport and use applies to manufactured ice, too. And the discussion of insulation may be of more general interest; it’s relevant to protection of temperature-sensitive electronics, chemical reactors, and liquefied gas storage.

Culinary Uses of Ice

There are four basic culinary uses of ice: short-term chilling of food and drink to make it more palatable; long-term refrigeration, to preserve it from spoilage in storage or transport; incorporation into frozen desserts; and temperature control during brewing.

Chilling. Chilling drinks was probably the most common seventeenth century use of snow or ice; Francesco Redi wrote (1685), “Snow is good liquor’s fifth element”—its quintessence. (Redi 17).

While snow or ice could be put right into a wineglass, the ancient Romans favored putting it into a kind of strainer suspended in the vessel, so it wouldn’t be drunk along with the wine. Of course, it would still melt and dilute the beverage, but the Romans didn’t usually drink wine straight, anyway.

The alternative was to put the wine inside a larger vessel, filled with ice; a “wine cooler.” This principle could be inverted; the Grand Duke Cosimo of Florence, in 1570, had several large (25-28 pound) silver wine coolers that had an inner vessel that held snow. (David 5).

The Italians leaped from using frozen water to chill food and drink for the table to using the ice as a form of decoration. Beginning in the 1620s, fruit, ice, salt and water were placed in pyramid-shaped pewter moulds to form ice pyramid centerpieces. (David 38ff, 58ff).

John Barclay, a Scotsman resident in Rome, wrote the romance Argenis (1621), set in Mauretania, which spoke of ice-encrusted apples and wine goblets made of ice. It’s clear that this was based on observation of the Roman table, since Antonio Frugoli of Lucca reported (1631) that at the feast of the Assumption on Aug. 15, 1623, there was un monte di diaccio con diversi frutti dentro, an ice mountain with fruits frozen inside it. If an ice mountain wasn’t spectacular enough, the centerpiece was an icy volcano; it spouted orange flower-perfumed water for over half an hour. This signifies not merely an ability to preserve natural ice, but to artificially freeze liquid water inside a conical mold with an inner tube to serve as the “volcanic vent.” Artificial freezing was also necessary to make the ice bowls and dishes described in books published by Florentine stewards in 1669 and 1672, and possibly describing a practice dating back to our time period. (David 55-65).

Food Preservation. To stop bacterial action completely, you need to freeze the food, and ice alone won’t accomplish it. However, ice-based refrigeration slows down bacterial action and can keep food fresh for a week or two, assuming that it brings the temperature down at least to 40oF.

In the Far North, the Inuit are well aware that food can be frozen for later use. In winter, “fish froze whole within a few seconds of being removed from the water.” If they were caught around freezeup (September-December, with freshwater freezing sooner), “whole fish were often placed on a gravel bar to freeze overnight, and then thaw again the following day.” (Burch 146). For that matter, Inuit have been known to eat frozen mammoth meat. The Dolganes of Siberia pack fresh bear and reindeer meat into snow and come back for it when ready to eat it. The Lapps eat poronkaristys, sauteed reindeer, fat-fried slices of frozen reindeer meat. (Shephard 281-2).

The Russians, certainly, were familiar with frozen food. William Coxe’s 1787 Travels described (2:300) the “frozen market” of Saint Petersburg, and Murray’s 1838 Handbook for Travelers said that similar markets were held in all the large cities. John Bell’s 1763 Travels said that Astrakhan fish “caught in autumn are carried to Moscow frozen,” and Jonas Hanway’s An Historical Account of the British Trade over the Caspian Sea (1753) said that fish are “sent either salted of frozen to distant parts of the Russian empire” (141).

Further south, European down-timers are aware that meat can be preserved by cold; Francis Bacon died in 1626 as a result of traipsing about collecting snow for an experiment testing how long it would preserve a chicken. However, outside Russia and Scandinavia, they made no systematic practical use of this knowledge.

Brewing. Beer is made by fermenting a malted grain (typically wheat or barley), in water flavored with herbs (a variety were used in the seventeenth century, but nowadays hops is standard). The beer yeast converts the sugars in the grain to alcohol, producing carbon dioxide gas as a byproduct. “Top-fermented” beers are produced in open vessels, at around 20oC (68oF), and stored for a few days or at most weeks at normal cellar temperatures. (Cannavan).

Even with top-brewed beers, temperature control would be desirable. “When the weather was hot and sunny it was possible that fermentation would run out of control.” (Sambrook 155). “The first recorded use of a thermometer in a brewery was in 1758 . . . .” (Blocker 94).

Conventional wisdom is that before refrigeration, brewing was seasonal, avoiding summers. However, the reality is more complex, at least for top-fermented beers. In Derbyshire, England, at Calke Abbey, during 1834, ale was produced only October through May. Beer was produced throughout the year, but summer brews were smaller and less frequent. Whereas at Lilleshall and Trentham in 1646 – 7, the peak production was June for strong beer and September for weak. (Sambrook, 155ff).

Lager yeast (S. pastorianus) flocculate to form large dense clumps that settle to the bottom of the vessel, and is active at colder temperatures than are the top-fermented ale yeasts (S. cerevisiae). They have an optimal growth temperature of about 28oC, a maximum growth temperature of about 34oC, and a minimum growth temperature of about 7oC. They are traditionally employed at 7 – 15oC in order to “develop specific flavor characteristics.”

In contrast, ale yeast forms loose clumps that trap carbon dioxide, and thus rise to the top of the tank. Ale yeast has an optimum growth temperature that is above 30oC, and a maximum growth temperature of 37.5 – 40oC. Most will not grow below 15oC. They are usually used at 18 – 25oC. (Essinger 123; Robert 349 – 50).

Supposedly around 1420, brewers in the Bavarian Alps “discovered that beer lost its natural cloudy appearance when stored in mountain caves . . . .” This discovery gave rise to the lager style, “bottom-fermented” beers, which are fermented at about 8oC (46oF), “with as little air-contact as possible and cold-stored for as long as possible (six months was once considered the minimum . . . ).” (Cannavan).

With Alpine caves at their disposal, these Bavarian brewers didn’t have to worry about refrigeration. Their lager was winter-brewed and winter-stored, however.

When lager became popular outside of Bavaria in the late 1830s, the brewers had to harvest or buy natural ice from lakes or rivers and store it. (Blocker 94). In 1880s America, the brewers were the biggest customers of the natural ice companies, and sometimes cut and stored their own ice. Dissatisfaction with the seasonality of the natural ice supply, and the occasional ice crop failure, led them to be among the first adopters of artificial refrigeration (see “Competition” section).

Frozen Desserts. A dessert, by definition, is sweet, and therefore contains some kind of sugar. The source of the sugar could be fruit juice, wine, honey, sugarcane, sorghum or sugar beet.

Frozen desserts include sherbets, water ice, ice cream, ice milk, and sorbets. FDA defines a sherbet as a food produced by freezing, while stirring, a mixture of a fruit juice (or certain other sources of flavor and sweetness) and dairy ingredients, the resulting milkfat content being 1 – 2 percent. (37 CFR 135.140). A “water ice” is similar except that it contains no dairy ingredient except egg white. (135.160). And “ice cream” is similar except that it contains at least 10 percent milkfat. (135.110). In-between sherbets and ice creams, we have low-fat ice cream, also known as ice milk. The term “sorbet” doesn’t have a legal definition, but it’s often used as equivalent to “water ice.” The term “milk ice” is sometimes used to cover all the frozen desserts in which milk is incorporated.

The use of the terms “sorbet,” “sherbet,” and “ice cream” in historical literature is quite different than the modern usage. For example, in sixteenth-century France, a sorbet was a beverage, a sweetened fruit juice diluted with water. (David 46). A Turkish sherbet of the same period might be ice-diluted, and a European copy might be ice-cooled, but that doesn’t mean that either was frozen. Likewise, you cannot assume that a product called “ice cream” was one in the modern sense unless you actually can read the recipe and see that milk was used.

There are plenty of entertaining legends about the origin of ice cream (milk ice). There is some evidence of frozen milk products in early China, but it didn’t seem to have much impact on seventeenth-century Chinese culinary practice, let alone what Europeans ate. I think it unlikely that the Arabs—even the “Caliph of Baghdad”—ate ice cream, although they certainly enjoyed flavored and sweetened water ices. There’s no doubt that Italy was the first European country to enjoy ice cream, sometime after chemical freezing methods (see below) became known there, and it’s possible that milk ice was available there before the Ring of Fire (RoF). Ice cream was first served in Britain in 1671, at Windsor Castle (Durant 172). Ice cream reached France sometime before it reached Britain.

Non-Culinary Uses of Refrigeration

One of the earliest non-culinary uses of ice (and ice water) was to bring relief to fever victims. (Visser 290). This prospect, in fact, was the key selling point that permitted Frederick Tudor to obtain a monopoly for the sale of ice to the British Caribbean (Weightman 46). Ice could also be used to minimize swelling and inflammation.

The main limitation on the use of natural ice for non-culinary refrigeration is that natural ice alone can only bring temperatures down to the freezing point of water; it cannot achieve colder temperatures as can a mechanical system. However, one can use chemical refrigeration (freezing mixtures) as discussed in a later section to achieve moderate freezing.

In the OTL, air conditioning of homes, workplaces and vehicles made hot summers more tolerable. It would require enormous quantities of ice, but conceivably natural ice could be used to cool air in ventilation systems. This would be somewhat similar to its use in the late-nineteenth and early-twentieth centuries on refrigerated rail cars. It would probably only be practical, if at all, in municipalities with excellent rail or water connections.

Likewise, ice can be used to cool chemical reactors. This is actually still done, on a small scale; a laboratory might use an ice bath to slow down a reaction, to reduce the vapor pressure of a volatile substance, or to alter the equilibrium of an exothermic reaction in favor of the products. Salts can be added to the ice bath to achieve colder temperatures. Ice might also prove useful in munitions plants, to reduce the temperatures at which the explosives are made and stored.

Refrigeration, possibly with natural ice, might also be used to remove (by condensation) moisture from the compressed air used by a blast furnace.

At least in the Far North, ice blocks can be used in construction, as evidenced by the Inuit igloos. Russia had the first known “palace” made of blocks of ice, constructed for Anna Ivanovna in 1739 – 40 (Wikipedia/Ice Palace). For serious construction, we will want to take a look at some form of pykrete. It’s uncertain whether it’s in Grantville literature, but there are certainly books on WW II in town, and one may have passing mention to the proposal (Project Habbakuk) to construct an aircraft carrier using ice reinforced with wood pulp, with the ice integrity maintained by artificially refrigerated brine pipes. Pykrete strength was about 7600 psi, half that of 1940s concrete. (Wikipedia/Pykrete). Additional strength could be achieved with better reinforcement (see my polymers and composites article), and insulation could be added, but it is probably wise to limit even a super-pykrete to the higher latitudes.

Trade in Ice and Snow Before RoF

The snow and ice trade is an ancient one. The Bible speaks of the refreshing nature of the “cold of snow in the time of harvest.” (Proverbs 25:8, 13). Snow certainly wouldn’t have been lying on the ground of Judea in the fall, so this was snow saved from the previous winter.

Generally speaking, those southerners, whether in southern Europe or in Asia, who had the advantage of living reasonably close to mountains that were snow-capped in winter, could enjoy chilled beverages in summer—if they could afford to pay for this privilege.

China. A poem written around 1100 BC states, “In the days of the second month, they hew out the ice . . . in the third month, they convey it to the ice houses which they open in those of the fourth . . . .” (David 228). I don’t have any specific information about Chinese ice harvests in the seventeenth century, but in the eighteenth century, the British East India Company became aware that Chinese fishermen and fish merchants had above-ground ice houses and used ice to preserve fish. The ice, in turn, came from rice paddies deliberately flooded during the winter. (229ff). Ice could also be heaped outdoors and covered with several feet of clay. (243).

India. The ice used by the Mughal emperors could be manufactured chemically (see later section), or harvested from natural sources. Beginning in 1586, natural ice was brought to Lahore, the new Mughal capital, from the mountains, about 100 miles away. It could be transported by barge, carriage or bearers. The Ain-i-Akbari reports, “all ranks use ice in summer, the nobles use it throughout the year” (Mubarak 56). It was about five times as expensive during the monsoon heat as in winter.

The price in Akbar’s time was as low as five copper dams (from which coin we reportedly get the expression, “I don’t give a dam(n)”) per ser (637.74 grams). At the time there were 40 dams to the rupee, so that’s one-eighth rupee. In 1873 Calcutta, American ice sold for the same price! (Mubarak 56).

EB11/Ice provides information about nineteenth-century Indian practice (but I would suspect that the practice was already centuries old): “In the upper provinces of India water is made to freeze during cold clear nights by leaving it overnight in porous vessels, or in bottles which are enwrapped in moistened cloth. The water then freezes in virtue of the cold produced by its own evaporation or by the drying of the moistened wrapper. In Bengal the natives resort to a still more elaborate forcing of the conditions. Pits are dug about 2 ft. deep and filled three-quarters full with dry straw, on which are set flat porous pans containing the water to be frozen. Exposed overnight to a cool dry gentle wind from the north-west, the water evaporates at the expense of its own heat, and the consequent cooling takes place with sufficient rapidity to overbalance the slow influx of heat from above through the cooled dense air or from below through the badly conducting straw.”

This was an extraordinarily labor-intensive process—two thousand laborers might hope to collect 25 – 30 tons in one night (Wightman 143)—but beggars couldn’t be choosers.

Persian Empire. In mid-fifth century AD, a member of the Chinese diplomatic mission observed that in Ctesiphon (near modern Baghdad ), “families keep ice in their houses.” (David 199). Persian methods of producing and storing ice varied from region to region. In some places it was practical to bring snow and ice down from the mountains, and in others they had to make it locally.

In 1620, according to the Marquis Pietro della Valle, ice was made in Izfahan by creating conditions in which water could freeze outside. They dug a long trench and built a three-sided shade wall around it, so the trench was exposed only to the north. The wall protected the trench from the wind as well as the sun. Beyond the trench, in the plain to the north, they dig many small, shallow channels. They flood the channels, which froze over at night. In the morning, transferred the ice from the channels to the trench. Water was poured over the old ice in the trench so the new ice would fuse to it. (David 191ff). The trenches were covered with reeds during the day. (208) Come summer, the ice was broken with pickaxes, and sold in shops or by street vendors. The practice continued throughout the seventeenth century, as attested to by the reports of Jean de Thevenot (1650s – 60s) and John Chardin (1670s).

Ice was cheap. Chardin notes that ice was sold by the donkey load, for effectively two deniers a pound. If a French denier had the same value as a Dutch denier—1/120 of a guilder—its purchasing power was probably about a third of a dollar in USE currency. Jean-Baptiste Tavernier, after returning from Persia in 1670, stated that there were charities, funded by bequests, that sent workers into the street with ice and water to provide it free to anyone who asked for it, and Chardin said that the wealthy would have ice water placed outside their homes for the convenience of passers-by. (212).

Despite the availability of this manufactured ice, some consumers preferred mountain snow for their drinks, and that too was available at the bazaar. Doctor Fryer reported that in 1676, even the poor would spend part of their money on snow.

If you were traveling, you could bring ice with you, or purchase it in a caravanserai (which probably had its own yakyal (ice house) and ice production facility). At Merv, in Turkmenistan, one can find the ruins of a fifteenth-century ice house in the shape of a stepped cone, with a shade wall nearby, and I would assume that this house was either built above the trench or alongside it. (215). An ice house of this type was still in use near Sirjan in 1975. (208).

Ottoman Empire. In Jaffa, Syria, in July 1494, the captain of a Venetian ship was gifted with a large sack of snow. (David xiii). According to the traveler Pierre Belon, who visited the Ottoman lands in 1546 – 51, the Turks “gather the snow, filling certain houses [buzchane] constructed like vaults or else like a hillock of earth,” and situated in a location sheltered from the sun, and the packed snow could last for two years without melting. (David 41).

I have not been able to locate any reference to seventeenth-century snow or ice collection for the benefit of the rulers of the Barbary Coast states, but snow does fall regularly in the Atlas Mountains. In the nineteenth century, snow was stored at the icehouse La Glaciere, for use in the summer in Algiers. (Strahan 402). But ice was also imported; in 1905, a ton of Norwegian ice sold for fifteen francs ($3).(I&R 9:236).

Pre-Ottoman Egypt. Snow was transported from Lebanon and Syria to Cairo in the thirteen and fourteenth centuries. (David xii). This took advantage of the existing postal system (destroyed by Timur in 1400), which employed relays of horses and camels. This implies that only small quantities, for the sultan and his favorites, were carried, but they were carried very quickly indeed. (David xii, James 523).

Roman Italy. Pliny complained about his effete fellow Romans who defied the natural order by using snow to cool wine in summer, and Seneca was in high dudgeon because his compatriots used ice as well as snow. (“Nothing is cold enough for some people,” yada yada yada.) Martial says that the cost of the ice or snow could exceed that of the wine it was cooling. (Forbes 113ff). The extreme example of Roman indulgence in natural refrigeration was set by the Emperor Elagabalus; “one summer he made a mountain of snow in the pleasure-garden attached to his house, having snow carried there for the purpose.” (Thayer).

Renaissance Italy. According to Cardinal Ferdinando Medici, in the 1570s, Italians packed snow into pits fifty feet deep, and twenty-five feet wide at the top. The pit was lined and covered with “prunings of trees and straw”; there was a wood grating three feet from the bottom to suspend the snow above a crude drainage space. (David xiii-xiv). In 1583, Ferdinando had a vaulted underground ice-house constructed at the Villa Medici, in Rome.

Apparently, Ferdinando Medici was not the only resident of Rome who fancied ice in summer. On July 24, 1571, he issued instructions “to the Rome chief of Police and all other personages of whatsoever rank or condition, giving notice that Ottaviano da Burrino, his muleteer, and the muleteer’s boy, are bringing two loads of snow per day to Rome, are not to be molested in any matter whatsoever, nor the snow to be taken to any other place whatsoever, ‘for it is for our use.’ ”

In 1581, Michel de Montaigne saw the pits at Pratolino, Tuscany, whose snow was delivered to Grand Duke Francesco Medici In 1598, the hydraulic engineer Bernardo Buontalenti, Francesco’s Superintendent of Public Works, was granted a monopoly over the delivery of snow to Florence each summer. It carried a pension of 210 scudi and of course the opportunity for profit. The penalty for violating Buontalenti’s rights was “a fine of twenty five scudi and two strokes of the rope.” (David 15ff).

Until Buontalenti’s time, the snow was carried down from the mountains; Florence is situated on the Arno river, and the Appenines are no more than twenty miles to the north. However, in 1603 – 5, he constructed several laghi (or peschiera) di diaccio, ice lakes. Presumably, these were artificial lakes that froze over in winter, providing a convenient source of ice. The ice, in turn, was transported to buca (or conserve) di diaccio, ice pits. It appears that the ice was harvested in late December. Buontalenti died in 1608 and the exclusive rights passed to Francesco Paulsanti.

Despite innovation in Florence, ice was carried by cart from the Lessini mountains in Friuli, to Verona, Venice and Mantua. (David 68). In Venice, John Reresby reported in 1657, “in summer the meanest person seldom drinks his wine without having it cooled either with ice or snow, which is preserved in places made for that purpose under ground, and sold publicly in markets.” (Reresby 102).

In the Spanish-controlled Kingdom of Naples, snow was stored in either ice pits, or the natural caves of Monte Somma and Monte Mauro (David 68).

Spain. Pits for the preservation of snow were dug during the reign of Carlos III of Navarre. (David xii). In 1492, not only did Columbus sail the ocean blue, the last Emir of Granada surrendered his city to Columbus ‘ patrons, Ferdinand and Isabella. The surrender was in January of that year, but Emir Muhammad XII may have remembered summers past in which he stood on a balcony of Alhambra and drank water chilled with snow from a mountain eighteen miles away, in the Sierra Nevada of Spain. (David 53). By a 1584 English account, the Sierras were “continually covered in snow.” (54).

In the late-sixteenth century, “snow was in common use at the Court of Castille by their Majesties, the Princes and princesses, and all the great Nobles and Gentlemen and the common people who reside there.” However, this was not then true in Seville, where Nicolas Monardes (1493 – 1588) penned his Tratao de la Nieve y del Bever Frio [Treatise on Snow and Cold Drinks] (1574).

The Iberian snow trade expanded, and snow was available in Seville, Valladolid, Toledo and Murcia by 1621. (David 53). In 1645, in Madrid, the right to sell snow was auctioned off.

France. During the reign of Henry III (1560 – 74), the French “began decorating their tables with carved ice sculptures, serving dishes atop piles of snow, and putting ice in their drinks.” (Qinzio x).

That ice had to come from somewhere, and, according to Monardes (1574), ice was transported 180 miles, from Flanders to Paris. If that’s correct, then it presumably is ice cut from ponds or rivers, because Flanders (northern Belgium ) isn’t mountainous. But David (43) suggests that perhaps it came ultimately from the Ardennes mountains.

England. It has been suggested that simple, unlined ice pits were used back in medieval times. The first documented icehouses in England were built by James I; at the Greenwich royal palace in 1619 and 1621, and at Hampton Court in 1625. (Durant 172).

The ice house constructed in 1660 by Charles II in Upper St. James (Green) Park inspired this 1661 verse by Edmund Waller.

“Yonder the harvest of cold months laid up,

Gives a fresh coolness to the royal cup,

There ice, like crystal, firm and never lost,

Tempers hot July with December’s frost”

Nonetheless, ice houses were a rarity in seventeenth-century England. According to David (xv), in seventeenth-century and even eighteenth-century England, “only the most wealthy could afford ice.” David indicates that the ice houses were “expensive undertakings on account of the digging,” but as we have seen, digging pits was hardly unusual. No doubt the brick added substantially to the cost.

In 1665, the Governor of Virginia, Sir William Berkeley, was granted “licence to gather, make and take snow and ice . . . and to preserve and keep the same in such pits, caves and cool places as he should think fit, saving the king’s loving subjects liberty to make and preserve snow and ice necessary for their own proper use.” (Visser 290). These pits, presumably, were the old-fashioned unlined pits, and thus cheaper.

Denmark. Frederick II had an ice-house at Elsinore; on a 1580 map, it looks rather like a tepee. (David plate 2). It was stocked with ice as early as 1564, and in that year, the crown engaged carpenters to make “ice-coffins.” (284).

Russia. It might not seem that Russians had much reason to store ice, but in the 1830s, Georg Kohl said that “their short but amazingly hot summer would render it difficult to keep all those kinds of provisions which are liable to spoil, if their winter did not afford them the means of preventing the decomposition accelerated by heat.” The first documented use of the term lednik (ice-house) was in 1482. (Molokhovets 41). In the sixteenth century, fish were salted, smoked or packed in ice. (Smith 10). A 1646 report on flood damage to a drink shop at Velikie Luki on the Lovat said that “the water poured over the ice-house and froze . . . .” (Smith 146). In the 1660s, “the Russian Tsar had fifteen ice cellars for storing meat and fish and more than thirty cellars for storing drinks”; the ice was changed each March. (Molokhovets 41). The archbishop’s palace also had an ice-cellar; in the attempted robbery of the church treasury in 1663, the thieves “had already broken a tunnel through the floor from the ice cellar into the palace.” (Michels 97).

Adam Olearus—who has appeared in 1632 universe canon—wrote in 1656 that the Russians “prepare ice-cellars, in the bottom of which they place snow and ice, and above that a row of kegs, then another layer of snow, and again kegs, and so forth. Over the top they lay straw and boards, since the cellar has no roof. Thus they . . . may have fresh and delicious beer throughout the summer—which is quite hot.” (Tatlock 32).

In the nineteenth century, there was a small export trade in frozen fish. “Perch would be sent from Tsaritsyn or Uralsk to Berlin and Vienna in wooden boxes with handles, packed between layers of straw and ice.” (Smith 270). Of fish exported from Astrakhan in 1897, 11% were frozen or packed in ice. (272).

Elsewhere in Europe. Monardes says that ice was also available in the Germanies, Hungary, and Bohemia.

However, despite all this interest in snow and ice, the fact remained that it was an essentially local trade. It was not until the nineteenth century that means were devised for routinely shipping ice across great distances.

Pre-RoF Freezing Methods

Natural ice is great for cooling down drinks, but it won’t freeze them. For that, you need some sort of artificial refrigeration.

The Huston article said that Thomas Cullen’s process was the “first refrigeration,” by which he meant, the first artificial method of freezing a liquid. Evaporating water was absorbed by sulfuric acid, which meant that more water could evaporate. Evaporating requires heat and the heat came from the remaining water. However, there were chemical freezing methods known before Cullen, and indeed before the RoF.

In Bengal, the temperatures usually don’t go below freezing, so they couldn’t make ice by the Persian method. Prior to 1586, ice for Akbar’s table was made by mixing Bengali saltpeter (potassium nitrate) with water. (David 246).

This works because potassium nitrate has a large positive heat (enthalpy) of solution (8,340 cals/mole, CRC 69th D122), meaning that it needs energy to dissolve (“endothermic solvation”). The heat has to come from somewhere, and so the salt takes it from the water. Common salt also has a positive heat of solution, but it’s very small: 928 cals/mole.

Now, an important point: You aren’t putting the saltpeter into the water that you’re trying to freeze. Rather, you have a vessel within a vessel, one containing the water to be frozen and the other the freezing mixture. The temperature of the latter will drop, thanks to endothermic solvation, but it doesn’t freeze itself because the salt also depressed the freezing point.

Giambattista della Porta of Naples, in the “cooking” section of his Magia Naturalis (1589), explained how wine could be frozen: “Put Wine into a Vial, and put a little water to it, that it may turn to ice the sooner. Then cast snow into a wooden vessel, and strew into it Saltpeter, powdered, or the cleansing of Saltpeter, called vulgarly Salazzo. Turn the Vial in the snow, and it will congeal by degrees . . . .”

Cornelius Drebbel, at the court of James I of England, demonstrated chemical freezing in 1620. The same year, Francis Bacon, in Novum Organum, wrote that “nitre or salt when added to snow or ice intensifies the cold of the latter . . . .”

The Machinery’s Handbook has a table of freezing mixtures (24th, 2442) featuring combinations of snow or water with common salt, calcium chloride, ammonium chloride (3533 cals/mole), ammonium nitrate (6140) or potassium hydrate (sic, exothermic!), and stating the resulting temperature change. EB11/Calcium says that “a temperature of -55oC is obtained by mixing 10 parts of the hexahydrate with 7 parts of snow.” It’s only the hexahydrate that dissolves endothermically; anhydrous calcium choride releases heat when it dissolves. (Cal-chlor). A calcium chloride brine can be cooled down (mechanically, or by being outdoors in a cold enough clime) to temperatures cold enough to “flash freeze” food. (Shephard 305, 2002EBCD/”food presrvation”).

Some old encyclopedias (e.g.,) have articles on “freezing mixtures”; New International Encyclopedia (1903) adds ammonium sulfocyanate (5400), ammonium nitrate, potassium sulphocyanate (5790), and sodium nitrate (4900) to our potential salts.


Sources of Ice

A square mile (640 acres) of ice, 12 inches thick, weighs 700,000 tons (Hall 1), and Thoreau was told that one acre of Walden Pond ice yielded 1,000 tons. So finding ice, per se, isn’t difficult (if you’re in high enough latitudes or altitudes for water to freeze), the problem is finding ice that’s convenient to transport to consumers and yet is unpolluted.

The principal sources of ice were lakes and rivers. Ice is opaque if it contains many air bubbles, which scatter light, and porous ice melts more rapidly. So, as our characters will learn, they should prefer clear ice. As a result, they will prefer a deep, gentle river to a lake, and a deep lake to a shallow one; the current and the depth tend to result in a lower air content. (Hall 8). A strong current inhibits ice formation, however.

Ice can be made locally anywhere there’s an adequate supply of drinking water and temperatures fall below freezing at night during the winter, as evidenced by Persian practice.

In the mid-nineteenth century, around Berwick-on-Tweed, a British center of salmon fishing, “local farmers . . . flooded fields for the purpose [of making ice] and . . . sold it for 5 – 10 shillings per tonne; for some years it was their most profitable crop.” (Cushing 108).

The natural ice industry had its unpredictable aspects; “ice famines” could occur if the producing regions suffered an unusually warm or short winter. New York, normally a producer state, had to import ice from Maine and Massachusetts in 1870 (Hall 21), and in 1880 it even obtained 18,000 tons ice from Norway. (Hall 3, 27). In 1898 the Norwegian and German ice crops failed, and Britain imported ice from Finland. (Blain 11).


New England (initially just Massachusetts, later Maine was also exploited) ice was exported all over the world, including Martinique (1805), Havana (1807), Charleston (1817), Savannah (1818), New Orleans (1820), Calcutta (1833), Rio de Janeiro (1834), London (1842), Marseilles, Madras, Bombay, Canton, Manila, Hong Kong, Batavia, Sydney and Yokohama. (Hall 2 – 3; etc.). In the 1880s, the Massachusetts ice companies could expect to harvest about 669,000 tons in a good year (Hall 23), and in 1880 the Kennebec region of Maine shipped out 890,000 tons. In the new time line, the French are taking over the British colony in Massachusetts, and conceivably could exploit New England ice. However, even early-nineteenth century New England had much more of an infrastructure (sawmills, ships, laborers) to support a long-distance ice trade than is the case in 163x.

The Hudson river region of New York, in 1880, had the capacity to store 2,800,000 tons.

In the 1632 universe, the French are expected to take over, forcefully, the Dutch colony of New Amsterdam, and that of course will give them control of the Hudson River.

In 1880, ice was harvested in Ohio, Illinois, Indiana, Rhode Island, Pennsylvania, New Jersey, Iowa, Minnesota, Michigan, Wisconsin, and even Kentucky, Tennessee and Missouri, but these areas weren’t colonized by the Europeans as of 1635 in the new time line and hence they aren’t useful as a source of ice yet.

Entrepreneurs still in the Old World will want to find European sources, if possible, and these are discussed below.


Scotland was of importance as a source of ice for local fishermen, but while fish frozen with Scottish ice ended up on British tables, it doesn’t appear that Scotland had a larger role in the ice trade.


Jan Baptist Van Helmont (1579 – 1644) reported that whalers in Greenland water strengthened wine by freezing out the water. (David 326). It’s not clear when the ice was first harvested for sale, but it’s known that in the late-eighteenth century, ice was brought from Greenland to Hamburg, and Greenlandic ice was brought to England as early as 1815. ( Id. ) In 1832, a 500 ton load was valued at 950 pounds for duty purposes. The same year, a ship brought in 150 tons from Iceland and the Faeroes, but it was valued at only 200 pounds. (335).


Norway proved to be a much more important competitor for New England. The Norwegian trade began, unsuccessfully, in 1822; Leftwich’s ship arrived in London with all its ice melted. “By the turn of the century, Norway exported more than 1,000,000 tons of ice each year, which vessels going to Northern Europe, the Mediterranean, Constantinople, Africa and even as far away as India.” Less than half of this went to England (Weightman 189); still, Norway held 99% of the English market.

Initially, the Norwegians harvested ice from the fjords, rivers and glaciers of its rugged west coast. However, “in many places the ice had to be carried on people’s backs.” Later, they switched to the lakes of the more heavily populated south and southeast coasts. The terrain was gentler and “ice-mining was an ideal part-time occupation for both the local farmers and the shipping crews . . . .” In addition, the local sawmills generated plenty of sawdust, an insulator. (Blain 7-8). It proved more convenient to create artificial lakes close to the fjords, on high points so ice could be slid down wooden inclines to the harbors, rather than rely on natural lakes further inland (9).


Sweden, Russia and Finland played only a minor role in the ice trade with England, because ice in the Baltic tended to keep their ports closed for a couple of months after Norway ‘s North Sea ports had opened. (Blain 12).


In the nineteenth century, Russian Alaska exported ice to California. In 1852, “250 tons of Novo-Arkhangel’sk ice were sold to the California Ice Company at $75 per ton and shipped to San Francisco.” (Black 264). Sitka proved to be an unreliable source, so in 1855 the Alaska Ice Company began harvesting ice from Woody Island. In 1852 – 9, over 7,000 tons were shipped from Alaska to points south (not just California, but also Latin America ). As volumes grew, the price fell to $7 a ton. (Carlson 58).


In the nineteenth century, after Norway, the largest producer of ice in Europe was Austria-Hungary, in particular the Vienna Ice Company. However, it serviced the German market, not Britain, and all I know is that in 1883 – 1885, it paid 20% dividends to its shareholders, but that it was liquidated in 1913.

Harvesting Ice

Timing is all. In the states at the southern border of the American “ice belt,” such as Pennsylvania, Maryland, Ohio and New York, late in the season, ice might be cut as soon as it was six inches thick. Further north, the companies could safely wait most years for the ice to be ten to fifteen inches thick, and in Maine the preference was for it to be 20 – 30 inches. You didn’t want the ice to be much thicker than that as it made it harder to handle economically. (Hall 8).

In Massachusetts, ice was harvested from January through March, when the ice had frozen to a depth of eighteen inches or more (Weightman 4). It was obtained from various Massachusetts lakes, including Fresh Pond, Walden Pond, Spy Pond and Wenham Lake. Fresh Pond alone could produce 90,000 tons annually. (Weightman 193). The ice companies bought the shoreline to establish ownership of the ice, but there were occasional boundary disputes.

In the nineteenth-century New England ice trade, at first ice was harvested by hand, using pickaxes and chisels to break it into large blocks, which were then cut further on shore using two-man saws. Or you could cut a hole in the ice with an axe, and then saw out a block.

The big breakthrough was made in 1825 by Nathaniel Wyeth; the horse-drawn ice plow. The horses wore spiked horseshoes, for better traction, and the ice plow was eventually refined so that as it cut its line, it also scratched out a parallel line at the right separation to mark the next line. When one set of lines was complete, a second set was drawn at right angles to the first. The ice was thus gridded with horse-drawn iron cutters, and the grooves were deepened until the blocks could be pried out with chisels, and transported to timber ice houses on the lakeshore. The size of the blocks was based on the intended destination of the ice; the further it had to travel, the larger the block. When the spring thaw arrived, wagons took the ice down to the docks, and off it went. (Weightman 5 – 6, 106ff). The only disadvantage of the horse-drawn ice plow was that “the ice had to be thick enough to support the weight of the horses and the men driving them.”

As the ice industry matured, specialized tools were developed to suit its particular needs; “eventually there were about 60 different tools used in the ice harvest for preparing the ice surface, cutting the blocks, poling blocks to the shore, breaking blocks, and getting the ice into storage.” (Howell Farm) That doesn’t mean we can’t make do with standard tools like axes and saws to get the industry going. But according to Hall (4), the tools, supplemented by steam power for lifting the ice into the ice house, increased the speed of cutting and storing by a factor of ten.

My information about labor requirements is somewhat indirect. For example, I know that in 1880, the ice houses of the Hudson River region had a capacity of 2,800,000 tons. With good ice, the houses are filled as a result of the efforts of 20,000 men and 1,000 horses, in ten to twenty days from when cutting began. (Hall 26). That implies that 7 – 14 tons can be cut and stored per worker-day, given an experienced crew with then-modern equipment. The wages paid were $1 – 1.50/day, and the cost of cutting and storing was 25 – 50 cents/ton. (27). Cooper (1905) says that if the winter was favorable and the haul isn’t more than a mile, harvesting cost 25 cents/ton. If the house was right at the shore line, half that; if the winter marginal, multiply by 2 – 4 fold. (464).

In 1844, at Wenham Lake in Massachusetts, the crop of 200,000 tons could be cut and stored in three weeks. “Forty men and twelve horses will cut and stow away 400 tons a day; in favorable weather 100 men are sometimes employed at once.” (Macgregor 988). That’s 10 tons/worker-day. Consistently, Thoreau in Walden Pond said that in winter 1846 – 7, 100 men could harvest 1000 tons in a good day.

Bear in mind that this productivity data was for the mature ice trade. Figure that novices with general purpose hand tools will be less effective. Ballard (173) comments, “In our early history [1826?], seventy-five tons was considered a good day’s work. During the past summer [1890?], several of the crews have handled in ten hours, one thousand tons.”

I am not sure how much of this harvesting technology will be known in Grantville. There is no reference to it in the 1911 Encyclopedia Britannica, probably it has a strong British bias and the British mostly imported ice. However, I have seen very detailed descriptions of the process in certain old encyclopedias, such as the New American Encyclopaedia (1872), the New International Encyclopedia (1918)(“Ice industry”), and the Encyclopedia Americana (1919)(“Ice Industry”). Some of the really old people in Grantville may have seen ice harvesting in their youth. But bear in mind that it became very uncommon after 1930.

Also, West Virginia is on the southern margin of the ice belt. For example, in the 1870s, the Kanawha River had more than six inches ice in only one of seven winters (Annual Reports, War Department). Lakes and ponds are more likely to freeze up, of course.

The best chance that someone from Grantville will have seen industrial-scale ice harvesting is on a visit to a living history farm that either still does it or has photos showing what it’s like. One such location is the Howell Living History Farm in Lambertville, New Jersey. There’s also the Longstreet farm in Holmdel, New Jersey, and the Wessels farm in York, Nebraska.

Thoreau’s Walden Pond is more tantalizing than helpful. In the chapter “The Pond in Winter,” he describes the activities of a crew of a hundred men armed with “sleds, plows, drill-barrows, turf-knives, spades, saws, rakes, and double-pointed pike-staff[s] . . . .” They divided the ice “into cakes by methods too well known to require description and these, being sledded to the shore, were rapidly hauled off on to an ice platform, and raised by grappling irons and block and tackle, worked by horses, on to a stack, as surely as so many barrels of flour, and there placed evenly side by side, and row upon row, as if they formed the solid base of an obelisk designed to pierce the clouds.”

When looking for possible sources, be resourceful. A farm-scale ice harvest is described in Laura Ingalls Wilder’s Farmer Boy (64ff).

You have the choice of harvesting slowly with a small work force or quickly with a large one, at least if you’re in an area with a long winter.

If the ice is being stored for later distribution, then it should be packed loosely (perhaps 40 – 45 pounds/cubic foot), so that the blocks can be removed as needed. But if the ice is placed directly in the overhead ice room of a cold storage house, then it’s packed as closely as possible (perhaps 45 – 50 pounds/cubic foot), and the blocks caulked together with chips, so it forms a solid mass. (Cooper 483, 490).


Since Saint Petersburg wasn’t built until 1702, it obviously wasn’t getting ice from the Neva in our time period. Still, it’s interesting to read Georg Kohl’s description of how ice was harvested there in the 1830s, since the Russian and American methods were certainly independently developed. The crew cut an inclined plane into the ice, so floating blocks could be hauled up to the rim of the quarry. As in New England, the ice was grooved, first to lay out a rectangle, and then a grid was laid over the rectangle. However, the grooves were made by hand, with an axe. A trench was dug to detach the rectangle. With this completed, workers would line up along a groove and strike it with heavy iron crowbar simultaneously. After a few knocks, the stripe would detach and they’d move on to the next one. A single laborer could cut a single stripe along the cross-grooves into individual blocks. (David 296).


What can go wrong? On the large scale, thaws and rains can melt the ice crop, and snow has to be shoveled or planed off. On the small scale, a worker can fall into the icy waters or be injured by an unexpected movement of one of the 200 – 400 pound blocks. Tools can be lost or broken.


A potentially more dangerous means of obtaining ice was to chop it off an iceberg. In August 1819, Captain Hadlock of the brig Retrieve succeeded, but the enterprise was nearly a disaster. On the first attempt, his sailors had to take shelter from a sudden storm. On the second, the inexperienced iceberg hackers caused the iceberg to topple over and damage the ship. They pumped their way to Martinique, and I hope Hadlock thought that his $1,700 fee was worth the trouble. (Weightman 92). On the other hand, in the twentieth century a Dane stated that it was “quite customary to use iceberg ice for drinking water” and “if you know icebergs, you know which ones are going to tip around.” (David 327).

In Sitka, an ice crop failure forced the Alaska Ice Company to cut ice from Baird Glacier (Carlson 58).


A small-scale operator can avoid much of the labor of harvesting ice by setting out bins of water and allowing them to freeze overnight. The resulting ice blocks are then freed from the bins and stored.

Mass Ice Storage

In the late-nineteenth century USA, ice was usually cut in January-March and consumed in May-October. (Hall 6). That was, of course, in part because ice was needed most when temperatures were high. However, the ice was not shipped south as soon as it was cut. The very conditions that made it easy to harvest ice also made it difficult to transport it. Ice was therefore harvested in the winter and stored until spring. The waste during this storage period was typically 10 – 25%. The ice would then be shipped to a distribution center and stored again until it was sold. The typical total waste, from harvest until arrival in the hands of consumers, was 40 – 55%. (Hall 9).

Natural caves. Caves are cooler than the surface in summer, but warmer in winter; a great mass of earth and stone has thermal inertia, a fancy way of saying that it changes temperature only slowly, so cave temperatures are virtually constant year-round. The depth in meters at which the annual temperature change is only 1oC is 3.18 * natural logarithm of the temperature change at the surface; that works out as 10.8 meters for a 30oC summer-winter surface difference. The average cave temperature is primarily a function of latitude and altitude; temperature (oC) = 0.6 * latitude (degrees) -0.002 altitude (meters). (A cave will be cooler than this formula predicts if it has a snowmelt-fed stream running through it during the spring.) ( Moore 27ff).

Some caves contain ice year-round; either they are at a high enough altitude so the average surface temperature is below freezing, or they are “cold traps.” The latter have a bottle shape; cold air flows in during the winter and blocks the ingress of warm air during the summer, reducing the average temperature by about 10oC relative to the predicted value. Cold traps are usually lava caves.

Underground storage is probably more advantageous in temperate regions, which have cold winters and hot summers, than in the tropics, where temperature variation is small.

Artificial caves. In Italy, Buontalenti constructed the Grotto Grande for the benefit of the Pitti Palace, and used it for ice storage. In Francesco Redi’s epic poem, Bacchus in Tuscany, Bacchus orders, “bring me ice duly, and bring it me doubly/Out of the grotto of Monte dei Boboli . . . . (Redi 18). It’s likely that Buontalenti also made use of simple ice pits.

The traditional English icehouse was mainly underground, and lined with brick or stone (Weightman 15). With regard to the 1619 “snow well” at Greenwich, contemporary accounts state that it was a “brick-lined well, 30 ft (9.23m) deep and 16 ft (4.92m) in diameter, covered by a thatched timber house with a door.” ( Pastscape Monument 761486).

Above-Ground Mound. At Walden Pond, the ice-harvesters didn’t bother to build an ice house. “They stacked up the cakes thus in the open air in a pile thirty-five feet high on one side and six or seven rods square, putting hay between the outside layers to exclude the air . . . . At first it looked like a vast blue fort or Valhalla; but when they began to tuck the coarse meadow hay into the crevices, and this became covered with rime and icicles, it looked like a venerable moss-grown and hoary ruin . . . . This heap, made in the winter of ’46 – 7 and estimated to contain ten thousand tons, was finally covered with hay and boards . . . .” Despite the lack of a proper roof, it survived the summer of 1847 and indeed “was not quite melted till September, 1848.” This was called “stacking.”

Above-Ground Ice House. The ice-houses of the American “ice belt,” used to hold the ice until it could be shipped south, were usually above-ground buildings. A medium sized one held 10,000 tons of ice, and was built using 175,000 feet of lumber. It was perhaps 30 feet high and sometimes divided into “rooms” holding perhaps 700 tons each. One ton of ice required 42.5 – 45 cubic feet. A large ice house might hold 60,000 tons and feature rooms holding several thousand tons apiece. (Hall 9).

In 1880, the cost of large ice houses, with machinery (20 hp steam engine-driven elevator chains), was $0.75 – 1.00/ton capacity for wood construction and $2/ton for brick. (Hall 10). Ice can be fed to an “elevator” by one man at a rate of 175 tons/hour, and 20 men would be needed to stow the ice away at that rate. ( Id. )

Small ice houses will be more expensive than large ones because materials cost is proportional to surface area. In 1905, for a wood creamery ice house holding 15,000 cubic feet, Cooper (506) quotes $1044.70, and the brick equivalent was about 30% more expensive.

Waste varies according to how long the ice is stored, how large and well insulated the ice house is, and how warm it is outside during the period of storage. In Maine, the waste at the large ice company houses was about 20%. (Hall 23). That said, ice could last in a well-designed and operated ice house for 2 – 3 years.

A few up-timers may have their own farm ice houses or have at one time seen or read about such. I have documented the existence of an at least partially above-ground ice house on Sandridge farm, in Audra, West Virginia. (Sandridge). There was also one on Lost Creek farm in North central West Virginia, although nothing is now remembered about its construction. (Lost Creek). In Leesburg, Virginia, there was a late-nineteenth century above-ground ice house with stone walls and straw insulation. (Leesburg). The USDA Farmers’ Bulletin 913 (Dec. 1917) contained a detailed description on how to build a small wooden ice house (pp. 29 – 39). I am sure that farmers in Grantville received the Farmers’ Bulletin and I suspect that some of them are pack rats who never throw away anything unless they run out of space. Or the Grange may have the bulletin.

What about ice houses in warm climates? Tudor’s first Havana icehouse was an above-ground wooden structure, a cube twenty-five feet on each side. The sales office was directly above the ice store, and the ice was brought up through a trap door. The structure was double walled, with sawdust and peat packed in the interstitial space. This meant that the insulating material remained dry. Meltwater escaped through a drain. (Weightman 65ff). At first, Tudor was losing 56 pounds ice/hour. He found that covering the ice with blankets, rather than sawdust, reduced the loss to 18 pounds/hour. ( Id. )


What can go wrong in ice house operation? If there’s a gap in the insulation, melting may be excessive. If the ice is packed improperly, the ice can slide toward the outer walls and bring them down. (Cooper 494). Lightning can strike the house (did you remember to provide lightning rods?) and set it on fire. For that matter, a fire can originate in the insulating material, as occurred at Wenham Lake in 1873.

The Physics of Keeping Cold

The enemy of ice is heat, and heat is transferred in three different ways: conduction (molecules in contact), convection (gas or liquid molecules in motion),and radiation (molecules not involved, can even cross a vacuum). Heat travels to the outside of an ice house by convection and radiation through the air, and conduction through the ground. It then travels through the insulation by conduction (and, if air spaces are provided, convection and re-radiation).

The first thing we need to recognize is that there are definitely economies of scale. Since heat transfer occurs at the surface of an object, the rate of transfer is proportional to the exposed surface area. But the increase in temperature as a result of that heat transfer is going to be inversely proportional to the volume. Large masses of ice have a lower surface area/volume ratio than small ones, so they will lose a smaller percentage of their volume to melting in a given time, with equal protection.

Secondly, heat flows from hot to cold, and the rate of conductive heat transfer is proportional to the temperature difference. So it’s advantageous to store the ice up north and ship it down south only just before summer begins; you’ll lose less ice that way.

Thirdly, the rate may be reduced by insulating the inside from the outside. In considering insulating materials (next section), bear in mind that a small ice storage requires better insulation than a large one, and that the price of the delivered ice sets a cap on what can reasonably be spent on insulation. Also, it may be cheaper to use a great thickness of a decent insulator rather than a small thickness of a great one.

Insulation must be more than resistant to heat flow, we also have to worry about the effects of mechanical damage, ultraviolet radiation, fire, moisture, fungi, vermin, etc. Insulation is often a composite of different materials to take advantage of their different properties.

For some situations, the thermal mass (heat capacity) of a material is also important. The thermal conductivity measures the steady rate at which heat passes through a layer, i.e., the rate if the temperature difference between inside and outside is maintained. The thermal mass measures how much a given quantity of heat will change the temperature of a bulk material; the greater the thermal mass, the smaller and slower the change. This makes a difference if you’re in a part of the world in which there is a big change in temperature from daytime to nighttime. Some heat is drawn back out of the massive material before it has a chance to penetrate to the interior. The materials with the highest thermal mass include, in descending order: water (4186 kJ/m3K); concrete (2060), stone (1800), rammed earth (1675), brick (1360), adobe (1300), wood (904), and fiberglass (6.7). ( )

The advantage of underground construction is not that earth is a good insulator (conductivity ~4) but rather that it has a relatively high thermal mass.

Note that thermal mass is not necessarily a good thing. Professor Mapes warned that stone is the worst material for an ice house because, if the walls are heated by the sun, they retain the heat day and night. (Robinson 296). The key questions are whether the walls are thick enough so that the heat doesn’t penetrate during the daytime and does the heat flow reverse at night.

Reflectivity. A reflective surface can reduce heat transfer by radiation to the outside of an above-ground ice house, or across a vacuum or air gap. Tin (let alone aluminum) foil is likely to be too expensive, but American “ice belt” houses were painted a “glaring white.” (Hall 9).

Drainage. If outside temperatures are above-freezing, then, inevitably, there will be some melting. It’s essential to drain away this meltwater. The heat conductivity of water is much higher (0.58 W/m-oK) than that of air (0.024), so it becomes a corridor for the delivery of more heat to the remaining ice. (ETB). If the soil is porous (gravel or sand), that may be sufficient; if a drain is constructed, it provides a path for warm air to enter. (Cooper 525).

Condensation. Water vapor condensing on the warm side of the insulation will compromise it; so you will want a vapor barrier, such as kraft paper or asphalted felt. (Marks 19-15).

Ventilation. I have found a lot of contradictory statements made about ventilation. Weightman (65) says, “as ice melts, it releases latent heat, which creates a warm and therefore rising draft of air.” He explained that traditional ice houses had domes in order to provide ventilation, dissipating this warm air. But Weightman is wrong; it takes heat to melt ice, and latent heat is released when ice freezes.

Cooper envisions the ice house as having a loft, separated by a plate floor from the actual ice storage. Ventilation is needed, he says, in the loft area, to counteract heating by radiation. “Above the plate, plenty of air; below it, none whatsoever.” (527).

Professor Mapes says that ventilation is needed only if the ice is in a cold storage, to keep the food from fouling. If the ice is just being preserved for use elsewhere, it will last longer if there’s no ventilation. (Robinson 297).

Above-ground versus below-ground. All else being equal, below-ground storage provides more insulation. Larsen reported the results of some experiments in South Dakota. Of 15 tons stored in 12×12 foot pile above ground, with a foot of straw under and over, and boards on top to keep out the rain, only 10% could actually be used. A second lot was stored in an equal-sized pit, with similar protection, and of that 30% survived. Later experiments featured an ice house with, I believe, a partially underground construction. With straw insulation, 30% of 8 tons survived. The next winter, with sawdust insulation, 47% of 10 tons survived.

Nonetheless, according to Ballard’s 1892 report on the Maine ice trade, above-ground is better because a “cooler and dryer atmosphere can be maintained during the summer, with the use of sawdust and shavings of wood or meadow hay as dunnage. Evaporation is more gradual, which is necessary in order to keep ice from forming a solid mass in the houses.”

Cooper says, “The first commercial ice houses were built below the surface of the ground, but at present all are constructed above ground, for the reason that drainage is more easily secured, and the ice is more easily removed from the house. The protection afforded by the earth is of comparatively small value when the disadvantages of storing below ground are taken into consideration.” (491). Doors should be as high as possible to minimize the escape of cold air. (495).

In 1914, the USDA weekly newsletter warned that “excavations are expensive,” and also reiterated Cooper’s points about drainage and ice removal. It also pointed out that the thermal stability of earth was a disadvantage in winter, the stored ice having to be protected against the earth heat.

Building-Grade Insulation

What we want is low conductivity. Conductivity is the amount of heat energy passing through a unit thickness and area of the material in a unit time, under the influence of a unit temperature difference. Conductance is defined similarly, but for a specified thickness. Resistivity is the reciprocal of conductivity and resistance (the “R-values” used by builders) that of conductance. Resistance should be proportional to thickness (and conductance inversely proportional) but they aren’t exactly so.

There’s plenty of conductivity information in various standard engineering handbooks; there should be multiple editions in Grantville, covering different materials. Just looking at pre-RoF editions of the handbooks that are in my personal library or my local library, I find . . . CRC Handbook of Chemistry and Physics (69th, 1989) provides conversion factors for units of thermal conductivity (E2) and the density and thermal conductivities (BTU/ft2-hr-oF; one inch thick) of several dozen materials (E6). Machinery’s Handbook (24th, 1992)(MH) provides yet another conductivity/conductance table (2445). Perry’s Chemical Engineers’ Handbook (8th, 1963) gives thermal conductivities (BTU/ft2-hr-oF; one foot thick) for numerous building and insulating materials (3 – 211ff), formulae for predicting thermal conductivity (3 – 223), heat transfer equations (section 10), and a detailed discussion of cryogenic-grade insulation)(12 – 33ff). An even richer source is Marks’ Standard Handbook for Mechanical Engineers (9th, 1987), with a chapter on heat (4), sections on air conditioning (12.4), mechanical refrigeration (19.1) and cryogenics (19.2); several conductivity tables are provided. I will use the CRC values/units in the following discussion, unless otherwise stated.

Conductivity is not a constant for a material; it increases (slowly) with increasing mean temperature. For example, extruded polystyrene might be 25% more conductive in summer (44oC) than winter (4oC).(Kirk-Othmer, 14:656).

Standard construction materials, generally speaking, are not good heat insulation; brick and glass (3 – 6), concrete (6 – 9), stone (4 – 28). Poured concrete is worse (12; MH) than concrete blocks. Wood is the best of the lot: oak (1.02), white pine (0.78), and especially balsa (0.33 – 0.58).

Air- or kiln-dried wood is a good insulator, but a filler that traps air in tiny pores is more efficient than solid wood. This was not fully appreciated in the nineteenth century, and there are instances of “the use of from six to ten thicknesses of boards in one wall.” (Cooper 70). Note that nails are good conductors of heat.

Up through the nineteenth century, the best insulating materials were inorganic or bioorganic particles or fibers. These could be provided as a loose fill, or sealed into flexible batts or rolls, or cemented with a binder to form semiflexible or rigid sheets. Conductivity is in the 0.25 – 0.5 range.

In the new time line the materials that will be available first are probably local cereal plant chaff and straw (straw bales ~0.4 – 1), local plant leaves, sawdust (0.41), planer shavings (0.42), charcoal (0.36 – 0.39), regranulated cork (0.30 – 0.31), hair felt (0.26), “Linofelt” (flax fibers between paper)(0.28), and perhaps felts using other fibers.

The problem with the organic materials is that they initially contain water (increasing conductivity, unless they are dried, which increases cost), and are vulnerable to fire (especially if dry), biological attack (fungi like it damp) and wicking (drawing in more moisture). Insulation may be given a protective “finish” or additive, at additional cost.

Historically, the natural ice industry created a market for sawdust; it sold in Maine at $3/cord (Hall 22). It deteriorates, especially when damp, so Cooper recommends its use only as the immediate packing material around the ice blocks, inside the ice house. Shavings replaced sawdust as the preferred nineteenth century ice house insulation; they are somewhat resistant to rot as the cell structure is preserved. Wood shavings may be treated with lime and other chemicals for improved life. Charcoal was used in Europe OTL for insulation, but blackens everything.

To make hair felt, cattle hair was obtained from tanners, washed and air dried, deodorized, and put into a felting machine. A waterproof paper was sometimes applied to it.

Within a few years we are likely to also have the opportunity to choose “Cabots quilt” (0.25 – 0.26), rock wool (0.26 – 0.29), mineral wool, glass wool (0.29), powdered diatomaceous earth (0.31), powdered gypsum (0.52 – 0.60), corkboard (0.25 – 0.34), “Insulite” cemented wood pulp (0.34), and perhaps the tropical plant-derived “Dry zero” (kapok between paper) (0.24 – 0.25) and “Celotex” cemented sugar cane fiber (0.34). There’s also the possibility of “air-entrained” (foamed) concrete (0.4; Wikipedia/Building Insulation Materials).

In the Pierce House (1635), in Dorchester, Massachusetts, eel grass (Zostera marina) was used as an insulation. Cabot’s Quilt (1891) was an eel grass mat, sandwiched or stitched between sheets or waterproof paper. Despite its organic origin, thanks to its iodine content it’s resistant to fire and vermin.

To make rock wool, granite and limestone are crushed, mixed with coke, and fused in a furnace at about 3000oF, creating a slag. The slag is run out through a high pressure steam blast, which blows it into a fibrous form. (Or you can use the slag from a blast furnace and add limestone to it, to make mineral wool.) It’s not subject to decay, but it is brittle (shouldn’t be packed more closely than 12 pounds/cubic foot) and should be handled with gloves. It’s possible to have the mineral wool compressed into slabs or sheets that can be used much like lumber. (Cooper 57). A natural mineral wool was collected by Hawaiian islanders from volcanic deposits, and used as insulation. The artificial version was “first commercially produced as a pipe insulator in Wales in 1840, and artificial rock wool was first produced in 1897. (Bynum 4).

Glass wool fibers are made by extruding molten glass through an array of holes, or by drawing out molten glass.

The solid (unfoamed) rubbers and plastics likely to appear first in the 1632 universe (see Cooper, “Industrial Alchemy, Part 5: Polymers,” Grantville Gazette 29) include natural rubber, rayon, cellulose acetate, cellulose nitrate, phenol-formaldehyde, and polyacetal. Polystyrene and polyurethane will come somewhat later. Rayon is the best of the lot; 0.054 – 0.07 W/m-oK; 0.37-0.48 BTU/ft2-hr-oF, inch). This isn’t good enough to justify its use! But the up-timers might not know this in advance . . . .

Synthetic polymers may be foamed to improve their performance. Foaming takes advantage of the relatively low thermal conductivity of air, especially air confined to small voids; closed cell foams are preferred. Of the “expected early” polymers, the ones with lowest conductivity foams are phenol-formaldehyde (0.0290.040 W/m-oK; 0.2 BTU/ft2-hr-oF, inch), urea-formaldehyde (0.0260.030; 0.18 – 0.21), and natural rubber (0.0360.043; 0.25 – 0.30). Nice, but still not better than the materials cited earlier. Also, the various formaldehyde based foams were banned because improper installation could result in over-exposure to formaldehyde.

The lowest conductivity foam is polyisocyanurate (0.0120.02; 0.08 – 0.14), a polyurethane (0.016 – 0.040; derivative. I forecast lab quantities of polyurethane (0.0160.040; 0.11 – 0.28) in 1635 – 37; it’s dubious whether Grantville Literature says how to make polyisocyanurate. It’s also uncertain whether it reveals that polyurethane foam can have a conductivity lower than 0.25.

As of the early 1990s, foamed polymers were more expensive than traditional fibrous insulation, although perhaps stronger and more durable. (Kirk-Othmer 14:659).

Further improvements are possible by replacing the air with a lower conductivity gas (e.g., trichlorofluoromethane), and by giving the polymer a foil facing to stave off radiation. Still, I suspect that by the time we develop these fancy foamed polymers, and get the production costs down enough to be competitive with sawdust etc., the window of opportunity for the natural ice industry will have closed.

Air is an excellent insulator— but only if the air was still. The problem is that if a large volume of air is trapped between hot and cold surfaces, convection currents will develop. Cooper “considered the one half inch air spaces formed by battens and paper . . . to be efficient until practical experience and the tests conducted by him proved otherwise.” (92). He concluded that “any space over one-half inch in width, if it can be kept dry, will be of greater value if filled with an insulating material as good as mill shavings, than if left as an air space.” What granulated and fibrous “fills” do is confine the air into many tiny air spaces, small enough so that the viscosity of air inhibits convection (45).

Cryogenic-Grade Insulation

The best possible insulation is a vacuum—it blocks both conduction and convection. It’s overkill for an ice house, of course, but I think this article is as good a place as any to talk about high-performance insulation.

Vacuum insulation was pioneered by Dewar (1893); his first Dewar flask has a high vacuum between double glass walls, with the walls joined at the top of the flask. One of Dewar’s papers notes that the problem of constructing glass that “would withstand the atmospheric pressure and bear considerable oscillations of temperature without cracking” was “only gradually overcome by . . . improvements in the blowing and adequate annealing of the glass.” (Dewar 1115). The evaporation rate of liquefied gas in the vacuum vessel was only one-fifth that in an ordinary vessel. Dewar then silvered the inner wall, to reflect away heat radiation, and that reduced the rate to one-sixth of that before silvering. (EB11/Liquid gases).

To significantly reduce heat flow, the vacuum must be good enough so that the mean free path of air molecules is greater than the distance between the facing surfaces. For air, the mean free path at atmospheric pressure is 93 nm (0oC), and the path length will be inversely proportional to the pressure. For a vacuum of 10-3 mm Hg (torr; normal atmospheric is 760) the mean free path is one micron. (Perry 12 – 34). In 1960s, a typical Dewar flask would use 10-4 or 10-5.

In the Thermos bottle, the glass was protected with a metal shell and spacers were placed between the walls. Still, glass is brittle and of limited compressive strength (needed by the outer wall) and so people naturally wanted to replace it with metal. That didn’t work, and in 1905 it was realized that the problem was that gas molecules are “occluded” inside metals, and released for a long time if the metal is exposed to hard vacuum. That, of course, ruined the vacuum. In 1906, the problem was solved by placing charcoal (an excellent gas absorbent) in a recess communicating with the “interspace.” As noted by EB11/Liquid Gases, that meant that Dewar flasks could be “formed of brass, copper, nickel or tinned iron, with necks of some alloy that is a bad conductor of heat.”

The use of vacuum insulation for an ice house was suggested as early as 1903 (Cooper 46); the proponent figured that if we could construct a single-walled steam boiler to resist 400 – 600 psi, a double-walled vacuum-insulated refrigerating room could be built to resist normal atmospheric pressure. I don’t see this as cost-effective, but we will certainly find uses for vacuum insulation.

Sometime before 1911, it was recognized that one could reduce the effective spacing—and thus achieve low conductivity with a poorer vacuum—by filling the space between the walls with a powder before evacuating the air. (The “fill” would also allow use of thinner walls, as they would help resist outside air pressure.) Tests showed that “silica, charcoal, lampblack and oxide of bismuth all increase the heat insulations to four, five and six times that of the empty vacuum space.” (EB11/Liquid Gases). This evolved into the modern “evacuated powder” insulation with fills of silica, silica aerogel, expanded perlite, diatomaceous earth, fused or laminated alumina, expanded mica, lampblack, synthetic calcium silicate, and even charcoal peach pits. Typically, the evacuated powders are 10 – 30 fold better insulators than the powders alone. For perlite and diatomaceous earth, the conductivity curve levels off, as the vacuum approaches 10-2, at a little above 10 uW/cm-oK (0.0069). (Perry 12 – 38ff). Evacuated lampblack conductivity is more sensitive to pressure but at 10-2 isn’t much inferior to perlite and kieselguhr and lampblack of course is universally available. I suspect that many other particulate and fibrous fills will work, too.

Perlite-in-vacuum is “the commercially accepted standard in the liquid hydrogen industry,” and you can get away with a 0.1 torr vacuum, yielding a conductivity of 0.018. (Brewer 327).

Expanded perlite is a volcanic glass that contains water and therefore pops like popcorn when quickly heated above 1600oF. In 2009, the principal producers of perlite were Greece *, United States ( New Mexico *, Nevada *, California *), Turkey *, Japan, Hungary *, Italy, Mexico, Georgia, Armenia, Iran, Slovakia, Australia, the Philippines and Zimbabwe. There is also perlite in South Africa, Algeria, Bulgaria, China, Cyprus, Iceland, Morocco, Mozambique, and Russia. (Index Mundi). A few of these sources (asterisked) are mentioned in the modern Encyclopedia Britannica, which draws particular attention to the island of Melos and to Kremnica in central Slovakia.

Diatomaceous earth (kieselguhr) is fossilized diatoms, and was discovered in 1836 on the Luneburg Heath, north Germany (Wikipedia/Diatomaceous earth). EB11/Diatomaceae remarks on the size of the deposit at Richmond, Virginia, and other entries mention deposits at Llyn Arenig Bach (Merioneth, North Wales ) and between Logie Coldstone and Dinnet ( Aberdeenshire, Scotland ). Modern EB mentions Old World deposits in Iceland ( Lake Myvatn ), Denmark, France, Russia and Algeria.

Evacuated foam rubber and plastics have also been developed, but those mentioned by Perry aren’t as good insulators as the evacuated powders already discussed.

In 1951, “superinsulation” was developed; this inserts a multitude of “floating” (minimal wall contact) reflective panels (aluminum foil or aluminized Mylar) into the hard (10-4 required) vacuum space, reducing radiative heat flow. Conductivity at cryogenic temperatures may be as low as 0.0003. (Perry 12-36, Marks 19-35).

Transporting Ice

Ice is a bulk commodity (low price per unit volume) and hence is economically transported for long distances only by water or by railroad.

By water. When Frederic Tudor (1783 – 1864), the “Ice King,” first proposed that one could make a profit selling ice in the Caribbean, the reaction he received was one of almost universal incredulity. To make matters worse, he wasn’t able to simply rent cargo space on someone else’s ship, because the shipowners feared that the melting ice would sink the vessel. Hence, he had to buy his own ship, a brig, for $4750 (Weightman 37).

There were, in fact, two problems with carrying ice by sea. First, the ice would initially act as ballast, but as it melted and drained away, the ship would become lighter and harder to handle. This wasn’t just because of the loss of ballast, but because the ice could shift around once its volume was reduced. You want a stout ship and an expert crew. Secondly, the meltwater could ruin other cargo. (Weightman 27, 33; Blain 8). Obviously, the problems are minimized if you keep melting to a minimum, which is also desirable from the standpoint of having cargo to sell at the destination.

In the New England ice trade, shipments to the West Indies, which could be reached in just ten or fifteen days, didn’t require extensive insulation. The hold was lined with insulation, say, four inches of tan bark [a waste material from tanner’s pits] on bottom and sides, and then the ice was packed in and covered with hay. The hatches were kept closed until voyage’s end. (MM).

We have several descriptions of how Frederic Tudor had the Tuscany loaded for its lengthy voyage to Calcutta. Tudor first laid “a sheathing of boards one inch from the skin of the vessel.” I think that means that he left an air space. Then he covered that with “6 inches tan [bark].” And that in turn was covered with one foot boards of lumber. Note that the result was a double-walled bottom. On this he loaded 180 tons of ice. His crew rammed in first 6 inches of hay and then one foot of tan bark on the sides. And finally, “a foot of tan or perhaps 20 inches” was laid on top “to make an unbroken stratum on top ends, sides and bottom.” (Weightman 126). His expectation was that the ice “would melt at a rate of about fifty pounds an hour over the four-month voyage,” so that two-thirds would make it to Calcutta.

As Weightman comments, the crew would be instructed not to open the hatches (and thus expose the ice to the air), and it would continuously pump out the meltwater (and of course any seawater that made it into the hold). However, we don’t have any details about the pumping arrangements.

A somewhat different description is given by Dixwell (MM). This says that the ice hold was fifty feet long, beginning at the after part of the forward hatch, and that it was essentially an ice house, with two walls of one inch deal planks and one foot of tan in-between, on bottom, top and sides. “The pump, well, and main-mast, were boxed round in the same manner.” The blocks were packed inside, without intervening spaces; the top was laid down, and then a foot of hay was placed over it. The underside of the deck was planked with deal and any remaining air space was filled with tan.

An interesting feature was an “ice gauge,” “a kind of float” embedded in the ice so its subsidence could be measured (presumably only on arrival, because the hold was supposed to remain closed en route). This didn’t work too well, because the ice melted between the blocks as well as on top. Nonetheless, the supercargo estimated that the loss in the four month seven day passage was 55 tons. Note that 6 – 8 more tons were lost ascending the river and another 20 in porting the ice from the ship to the local ice house.

It is very unlikely that anyone in Grantville in fact knows anything about how ice was shipped, so they are in fact starting from scratch. They will certainly realize that it’s important to keep the ice insulated, and they will also recognize that despite best efforts, the ice will melt and they will have to get rid of the meltwater. It’s less certain that they will have foreknowledge of the effect of melting on ship stability; there might be an interesting plot twist there.

The loss of ice in transport will be dependent on the same factors as for stored ice. For shipments from the northern USA to the Gulf States, it typically ran 25 – 30% (Hall 35). Even in New York, 25 – 33% loss was expected (27), and Cincinnati suffered 33% (31).

By rail. Ice was also, occasionally, transported by rail. In 1878, “four freight trains of 20 cars each were loaded with [ Maine ] ice at $1.50 per ton and dispatched to Saint Louis, where they delivered in good condition 672 tons out of a total shipment of 1,275. The freight cost $5.85 per gross ton, and the ice sold for $10.50 per net ton. Owing to the waste of 50 per cent, the venture netted a large loss, and was not repeated.” (Hall 21).

However, inland cities without a good river connection would have to be serviced by rail. Northern ice was shipped by rail from Savannah to Macon and Atlanta (Hall 35). Ice was also carried by rail from Quincy, Illinois to points in Texas, 5 – 7 days journey. ( Id. ), and from Sandusky to Columbus, Springfield and Cincinnati (32).

There were two basic problems with rail transport. First, while it’s cheaper to ship by rail car than by truck, wagon or pack mule, the river barge, sailing ship, or steamship is cheaper still. See Cooper, Hither and Yon (Grantville Gazette 11).

Secondly, a rail car might hold 40 tons of ice, as compared to the 100 or more tons in the hold of a ship. The surface area/volume ratio is higher for the rail car, so the percentage daily loss by melting will be higher, too, unless it’s more heavily insulated.

Refrigerated Transport of Perishables

The ice, instead of being the shipped commodity, may simply be a means of keeping perishable foods cool and therefore fresh until they reach the destination. While refrigerated transport still needs insulation to protect both ice and commodities from ambient heat, they must be designed to encourage heat flow from the perishables to the ice . . . preserving the former at the cost of the latter.

If natural ice isn’t cold enough, brine can be frozen instead. That doesn’t necessarily mean using artificial refrigeration, the brine could be set out in bins overnight in a place and season where the temperature falls low enough to freeze it.

Rail Transport. The first successful use of natural ice refrigeration on a railroad car was in 1842, by the Western Railroad of Massachussetts. This was, strictly speaking, a climate-controlled double-walled car; it used ice in summer and powdered charcoal in winter.

Perishable cargo could be carried in individual iced containers on a standard boxcar. Parker Earle’s 1866 strawberry express trains were loaded with ice chests having ice on the bottom and berries above. (White 273).

But as I said previously, bigger is better, and so there are advantages to refrigerating the entire car. In 1883, a car cost about $500 (124), whereas a reefer cost $750 – 1200.

Several different layouts were experimented with. Lyman began reefer operation in 1853, and his 1860 cars had ice bunkers with roof hatches (for re-icing) at both ends, and an inclined floor with pipes to carry away meltwater. (272). The ice of course was crushed ice, that could be poured in through the hatch, rather than the large blocks of the ice trade.

In 1877, Tiffany patented a car with “a full length overhead ice bunker.” This had a V-shaped bottom, so that meltwater was conducted into drainage pipes. These overhead bunker cars reportedly “used only half the amount of ice required by conventional end-bunker cars.” (274 – 5).

Wickes’ popular late nineteenth-century design (275 – 6) had a number of interesting features. His overhead ice tanks were surrounded by metal basketwork with projecting leaves; the latter increased the cooling surface. Also, wire was strung underneath the tank, making a close mesh. The theory was that the drip water would strike the wires and be broken into a fine spray, which—since the water was “practically as cold as the ice itself”—would further cool the perishables.

A third layout was introduced by Hanrahan (1890); his bunker was in the center of the car. That reduced the number of roof hatches needed, but the four side doors he provided ensured plenty of opportunity for warm air to leak in. (283).

It has proven surprisingly difficult to obtain reliable details of the design adopted by the meat packing king, Gustavus Swift, and designed by Andrew Chase. Many sources say that Chase’s innovation was an overhead ice bunker, but of course that’s wrong. One possibility is that Chase’s “cold blast” refrigeration featured ice bunkers in the overhead corners. (Fields 105). Cold air would sink down the sides, and warm air rise in the center, thus providing greater air circulation than in the Tiffany or Wickes systems.

However, Chase U.S. Patent 346,354, REFRIGERATING BUILDINGS AND VESSELS (1886) depicts a system featuring a full-length overhead cooling chamber but with air gaps at the sides. The floor of the ice bunker has a downward projection at one side, creating a full-length passage for cold air to sink, and an upward projection on the other side, creating a passage for warm air to rise. As drawn, the cooling chamber was not an ice bunker, but rather contained coolant pipes that were connected with a mechanical refrigeration system.

In the nineteenth-century reefers, insulation was fairly primitive, with paper, wood, sawdust, cork, and dead-air space being typical. The early refrigerator cars had wooden bodies, and therefore numerous minute cracks. Ayers proposed the use of air-tight rubber sheets to block the entry of warm air and moisture. (278).

There are several model railroaders and railfans in Grantville, and it’s difficult to predict just how much they would know about refrigerated cars. There are several truck drivers in town, and some may have driven reefer trucks.

Water Transport. The down-timers realize that fish are extremely perishable, and have tried to cope with this in a variety of ways. In sixteenth-century Holland, some fishing vessels were equipped with “wet wells,” enclosures so that the captured fish could be maintained alive in circulating seawater. However, this didn’t work well for deepwater fish. (Kurlansky).

The Dutch also had buizen, “factory” ships on which captured fish were gutted, salted and barreled; these were periodically visited by fast ventjagers (“sale-hunters”) which would bring the barrels back to market so the buizen could remain at sea. (Sommerville). Tobias Gentleman described the Dutch system in 1614, but I think it dated back to the sixteenth century. (Poulsen 110ff).

While we aren’t interested in wet wells or salted fish, the point is that there are going to be people who are keenly interested in preserving the catch.

Nonetheless, it wasn’t until the 1780s that the British realized that fish could be chilled, with natural ice, for transport from Scotland to London, a six day journey on average. Until then, fish were carried in carts or on packhorses, and only during the cooler months; a sea passage by sailing ship could be delayed by winds and deliver only a rotten cargo. First, salmon were shipped in ice, then herring. Naturally, the introduction of the steamship made this even more efficient. (David 231ff).

British fishing vessels also started carrying ice, so that they could remain at sea longer and still return with fresh fish, which sold at a premium over salted fish. (238). Reportedly, the first use of this stratagem in the American fishing industry was by a Gloucester smack in 1838, to preserve halibut that had died inadvertently. (Stevenson 359).

Freshwater ice can only be used to bring fish just down to its freezing temperature, 0oC. But fish will last longer, with good taste, if superchilled—brought to a temperature below zero but above the temperature at which fish muscle freezes. The critical temperature varies by fish species: haddock, -0.8 to -1oC; halibut, -1 to -1.2; herring, -1.4. Superchilling can be achieved by freezing seawater ice or an artificial brine. (Huss).

You need 1.2 kg ice to chill 10 kg fish from 10oC down to 0oC, 2.3 from 20oC, 3.4 from 30. More ice is needed to keep them cold. How much more depends on the length of the trip, but “it is a generally accepted ‘rule of thumb minimum’ to use an ice to fish ratio of 1:1 in the tropics.” (Shawyer 43). For temperate waters, a one week trip will require 0.35 – 0.5 tons ice for 1 ton fish. (FAO).

Note that round, small, fatty, thin-skinned or damaged fish spoil faster than their opposites (Shawyer 5). Typical shelf lives (on ice) are 2 – 24 days (marine temperate), 6 – 25 (marine tropical), 9 – 17 (freshwater temperate), and 6 – 40 (freshwater tropical). More specifically, we have 9 – 15 days for cod, 7 – 18 for flounder, 21 – 24 for halibut, and 3 – 8 for temperate water sardine. (6).

As with rail transport, there’s a choice between placing fish and ice in insulated boxes (typical for small open boats even today) or in an insulated hold. Insulation considerations are similar to those for ice houses and refrigerated cars, but bear in mind that a fishing vessel is likely to be at sea longer than a train is in transit.

Long-distance shipment of frozen fish from the fishing fleet’s home port to customers is also possible. However, the customers who are willing to pay the most for fish are likely to be those who reside inland, who otherwise couldn’t enjoy seawater fish at all, not those in some other port (at least until people can afford to pay for exotic fish).

“The first attempt to ship meat across the oceans was carried out by Bell in 1877. The refrigeration process was achieved by ice containers filled with natural ice, circulating a current of air through the ice by means of a fan.” Mechanical refrigeration was first used on the Australia-Britain route in 1879. (Blain 26).


Ice was transported to a port city or railroad nexus, and held in a large ice house for sale and distribution to the ultimate consumers. There were two different distribution strategies. In the “will call” system, the consumer came to the ice house and carried off the ice purchased. In the “delivery” system, the customers placed orders, and the ice was delivered by ice wagons or trucks to the customers’ doors.

Long-Term Cold Storage of Perishables

Marks (19 – 17) provides optimal storage temperatures for many different products. There are three problems with long-term (months to years) storage of perishables: (1) natural ice doesn’t chill them enough to halt decay, (2) cooling with ice tends to lead to high humidity conditions, which encourages the growth of mold, and (3) foods release odoriferous gases which can contaminate other foods. (These considerations apply, to a lesser degree, to icebox design.)

Now, in the nineteenth century, people tolerated food spoilage to some degree. Fish would be sold after a month’s storage at 1 – 2°C; nowadays fish would be thrown out if stored more than a few days at that temperature. (Freidberg 30). The down-timers might initially have a similar attitude to preserved foods, but of course the up-timers are going to be quite strident on the issue of food quality.

With regard to humidity and odor, it helps to put the ice overhead, as on “reefer” cars, and encourage air circulation, and also to conduct it over drying agents such as calcium chloride. Still, having direct circulation of air between the ice and the food guarantees that there will be a constant battle on the humidity front.

The existence of these problems doesn’t mean that this niche can only be filled by artificial refrigeration. The trick is to combine chemical generation of sub-zero temperatures with indirect cooling. In the Cooper gravity brine system (453ff), you had a freezing tank on the upper floor, that was filled with broken ice and a suitable salt. Brine was circulated through primary coils in the freezing tank, and secondary coils in the cooling room on the lower floor. Since the brine in the primary coils was colder, it was denser, and would sink down to the secondary coils. The brine here would cool the room by absorbing heat, and thus would become less dense, and rise up to the primary coils. Of course, this density-driven circulation could be assisted by a pump.

Brine temperatures of 5 – 20oF were obtainable, and the room could be cooled to 10 – 15oF. Because the ice and the brine do not come in contact with the air in the cooling room, they don’t increase its humidity, and indeed condensation of moisture on the secondary coils will tend to dry the air there. Air circulation is encouraged by creating separate paths for air to rise and fall, as in the Chase reefer system or the Dexter house system (448). A fan may be used to increase air circulation.

I very much doubt that there’s any description of this system in Grantville Literature, so the question is whether it will be re-invented.


While I have shown that there were plenty of ice users in the seventeenth century, please note that there was medical opposition to drinking chilled beverages both before (Piero Nati, 1576) and after (Louis Lemery, 1702) our time period.

In the nineteenth-century Caribbean, the inhabitants didn’t quite know what to do with ice, and sales were slow. One of Tudor’s suggestions to his icehouse keepers was to “offer bartenders free ice and persuade them to serve customers cold drinks at the same price as those that were unchilled.” Once the customers were hooked, he could start charging for the ice. (Weightman 93).

By the turn of the nineteenth century, stored local ice was a commonplace in England. In 1804, Cassandra Austen wrote to sister Jane, complaining of an ice famine in Weymouth, to which Jane responded tartly, “for every other vexation I was in some measure prepared . . . but for there being no ice what could prepare me.” (David 328).

The Wenham Lake Ice Company began its marketing barrage by sending a block of ice to Victoria and Albert (who became high-profile customers) and then displaying a massive block of ice in the window of their store on the Strand. This block of ice appeared indestructible . . . because in off-hours it was replaced by fresh ice from underground storage. (David 339).

The goodwill that developed for the “Wenham Lake” appellation became something of a disadvantage when that lake (located in Massachusetts) proved incapable of satisfying all the British demand; Wenham Lake itself never produced more than 10,000 tons/year.(Wightman 192) The Company had the Norwegian Lake Oppegaard, near Kristiania, rechristened “Lake Wenham” for commercial purposes. (347).

Nineteenth-century France and Italy were tougher sells. In the cities, consumers were accustomed to buy fresh and eat everything they bought the same day. (Freidberg 27ff).


In India, the ice made by the nineteenth century evaporative process I described earlier was known, rather derogatively, as ” Hooghly slush,” and in the 1830s it was available only at a price of four pence a pound. Thanks to Frederick Tudor’s entrepreneurship, New England ice could be obtained in Calcutta in 1833 at a delivered price of 3.5 pence a pound. (David 258).

According to Weightman (49), what made possible the ice trade to India was that Boston ships had little in the way of cargo to convey there; the voyage was paid for by the sale of Indian goods back home. Hence, the ice could travel, very cheaply, “in ballast.” In 1870, almost $100,000 worth of ice was exported to India, but artificial refrigeration eventually became practical and by 1880 that trade had declined to only about $5,000. (David 272).

Curiously, warm countries were not the biggest market for ice. Rather, it was cool countries that had short but hot summers. In 1890, the biggest ice consumers were the United States and Russia.

At its peak, the natural ice trade was big business. In 1879, the American harvest was eight million tons, and the ice sold for about $4/ton. The biggest market was New York, and the cost of shipping ice there from Maine was $1.50/ton (breakdown was 20 cents cut and store, 50 cents transfer to ship, 50 cents ocean freight to NY, 30 cents unload—Hall 21). In 1875, 50,000 tons of Boston ice was exported to India, China, Indonesia and Japan. (David 279).

The per capita consumption of ice is dependent to some degree on whether it’s priced to be affordable to all, and also on whether ice is simply used at home or also by businesses. In 1880 America, which may be considered a mature market, the twenty largest cities (total population 5,930,000) consumed 3,961,000 tons of ice, which works out to two-thirds ton per person. Another 200 smaller communities, located in the “ice belt,” are believed to have consumed about one-quarter ton per person, with less in the way of industrial use. (Hall 5).

Over the long term, prices dropped, thanks to increased supply, but profits rose, thanks to increased demand. In the 1840s, Macgregor (987) wrote, “Formerly, ice sold in New Orleans for six cents (threepence) per lb., and now sells for one cent (one halfpenny) per lb.; but more money is made from the increased consumption at one cent than was made at six cents.” Tudor first sold ice in Charleston at 8 1/3 cents/pound. (Weightman 78).

In the short term, the price of ice varied a great deal, depending on the severity of the previous winter and the size of the lots purchased. Ice shortages led to high prices, which encouraged construction of new ice houses and vigorous harvesting of ice the following year, and that in turn to an ice glut and low prices. While profits were then low, the low prices encouraged increased consumption, and some of the increment remained even after prices rose again.

During the Hudson ice famine of 1880, despite a supply of “old” ice in the storehouses, the wholesale price rose from $1.50/ton in January to $12/ton in August for large consumers (families briefly paid $20/ton). The next winter was severe and the wholesale price even in the summer was $1.50 – 2/ton, with small consumers paying $4 – 8/ton. (Hall 27).

In Cincinnati, a canal-boat load of 60 tons cost $4 – 6/ton on average in summer; $2 – 2.50 in a good ice year; $10 in a bad one. In 1880, if you bought 100 pounds at a time, as might a hotel or saloon, you paid $14 – 16/ton, and a household paid $20. In an abundant year the latter might pay $6 – 7 in winter and $10 – 12 in summer. (Hall 31). In Chicago, in a high supply year, ice sold for $1.50 – 2/ton for the RR carload, $2 – 3 for the wagon load; $4 – 6 for a 100 pound lot. “At retail the price in summer is always a little more than double the one which rules in the winter.” (34).

Transport costs and associated ice losses also affected the delivered price. In the Gulf Coast states, the wholesale price on the coast was $10 – 30/ton in 1880, but the price to consumers inland was $60 – 75/ton. (Hall 35). At San Francisco, the price of Alaskan ice was initially $100 – 120/ton retail, but after the Southern Pacific Railroad opened up the Sierras, ice could be had for $40/ton. (Hall, 37).


Between Natural Ice Companies. Nothing illustrates the sometimes cutthroat nature of ice industry competition better than the “Wisconsin Ice War.” The Wisconsin Lakes Ice and Cartage Company had long harvested ice from the Milwaukee River. With great fanfare, the Pike and North Lake Company built large ice houses on the shores of various remote lakes, and advised the public that their ice would be clean, unlike that of the WLICC. Much to their dismay, the PNLC discovered that agents of the WLICC had acquired the rights of way between the PNLC lakes and the railroad, and refused to allow the PNLC’s ice to pass. The PNLC management outfitted a steam launch as an icebreaker, and advertised “river excursions” on the Milwaukee River. For six weeks, to the oom-pah-pah of a brass band, the launch plowed back and forth through the ice that the WLICC had intended to harvest, breaking it into useless fragments. The “excursionists” on board the vessel were husky laborers, armed with pike poles to repel WLICC boarders. The result was the ruin of the WLICC ice crop. ( Lawrence 264).

From Artificial Refrigeration. The natural ice trade came to an end because its costs rose (the harvesters had to venture further and further afield to find unpolluted ice) while the cost of artificial refrigeration dropped. (Weightman 10).

Early refrigeration equipment was massive, totally impractical for household use. Hence, artificial refrigeration was used to manufactured blocks of ice, which would then be distributed and used in the same manner as natural ice. The advantage that artificial refrigeration had was that the machines could be in the same cities as the customers, so shipping costs were minimized.

But cost was the critical factor; in America, artificial refrigeration factories needed to be able to supply ice at $2 – 3/ton in order to compete. (Hall 20). In remote destinations, they had more of an edge; the American ice trade to India dried up after 1880.

Artificial refrigeration made inroads whenever the supply of natural ice was interrupted. The American Civil War stopped the shipment of New England ice to the heat-bedeviled South, which accordingly slipped two Carre ammonia-based refrigeration machines through the blockade. (David 274). The Southerners learned to appreciate the manufactured product, and continued to use it after the war was over.

In Life on the Mississippi (1883), Mark Twain says, “in Vicksburg and Natchez, in my time, ice was jewelry, none but the rich could wear it. But anybody and everybody can have it now” (Chap. 39). The change was the result of the local ice-factory; the one in Natchez made thirty tons a day, using an ammonia-based refrigeration system, and the ice sold for six or seven dollars a ton. That price wasn’t low enough to shut out natural New England ice, but it did set a ceiling for the pricing of the northern product.

German WW I submarine activity disrupted the British-Norwegian natural ice trade, encouraging expansion of the artificial refrigeration capacity. (Blain 31). Artificial refrigeration had a stabilizing effect on the price of Norwegian ice in Britain, at least in locales that had their own icemaking equipment to compensate for the seasonality of the Norwegian supply. (Blain 27).

In order to break into California market, an artificial ice machine company paid the Alaska Ice Company on Woody Island not to ship its Lake Tanignak ice to California. To make sure the moola kept coming, the Woody Islanders cut and stored ice each year. (Carlson 60).

However, the natural ice trade survived a surprisingly long time. In England, the “crystal ice” of Norway had an aesthetic edge, at least for table use, over the bubble-clouded artificial ice (Blain 40); some manufactured ice also looked “more like frozen milk than pure water.” (David 278). The bubbles could be eliminated by water agitation but this added considerably to the cost.

Early icemaking machinery was unreliable, and leaks, fires and “explosions were not unusual.” (Blain 41) In 1920, ammonia-based refrigeration was still considered a major risk factor by business insurers. (Freidberg 38).

The equipment was also large and expensive (relative to production capacity). The first household refrigerators had to have the machinery in the basement, piping the refrigerant up to the box in the kitchen ( Id. ).

Even at the time of WW I, there were serious problems with artificial ice: “It was still often tainted by ammonia; it still contained occasional drops of oil; and it could be no purer than the city water from which it was made.” ( Lawrence 260). Norwegian ice was still exported to England after WW I; there was a public distrust of manufactured ice, in part because of the use of ammonia as a refrigerant (David 243).

The expansion of artificial ice production capacity was slow (no doubt because of the initial cost and the public misgivings), which in turn meant that one couldn’t rely on it alone. In 1911, the daily British consumption of ice was 2000 tons, and the manufacturing capacity only 500 (Blain 30).


There was once a saying, “as rare as snow in Egypt.” (Forbes 115). But if a long-distance ice trade can be established in the new universe, the comforts of cold might not be rare there, or elsewhere in the civilized world, at all.

Author’s Note:

Bibliography will be posted to, Gazette Extras.



About Iver P. Cooper

Iver P. Cooper, an intellectual property law attorney, lives in Arlington, Virginia with his wife and two children. Two cats and a chinchilla rule the household with iron paws. Iver has received legal writing awards from the American Patent Law Association, the U.S. Trademark Association, and the American Society of Composers, Authors and Publishers, and is the sole author of Biotechnology and the Law, now in its twenty-something edition. He has frequently contributed both fiction and nonfiction to The Grantville Gazette.


When not writing (or trying to get an “orange blob” off his chair so he can start writing), he has been known to teach swing dancing and folk dancing, or to compete in local photo club competitions. Iver adds, “I can’t get my wife to read my fiction, but she has no trouble cashing the checks.”

Iver’s story “The Chase” is in Ring of Fire II