Today, we take architectural glass for granted: when we think of windows, most of us inevitably think of glass. The link between glass and windows is firmly established in modern Western architecture. Glass has also become the norm in other societies—including non-Western societies where, for centuries, it played little or no role in architecture. Even in Europe, the link between glass and fenestration was by no means as firm in the past as it is today: for millennia, windows were furnished with a variety of different sealants. What is more, glass windows did not come into use until the days of imperial Rome.

Not that glassmaking itself was unknown in antiquity. The first man-made glass dates back to around 2500 bc and originated in Mesopotamia, more specifically modern-day northern Syria and Iraq. It was produced from the same basic ingredients that, with some modifications, are still used today: a vitrifying agent, typically sand or another source of silica, was the major component; an alkali fluxing agent such as soda was added to lower the silica’s melting temperature in the furnace; and lastly, a source of calcium oxide—for instance, lime—served as a stabilizing agent.

Beads seem to have been the main product of the ancient glass industry, at least initially. Vessels were not produced until a millennium after the beginnings of Middle Eastern glassmaking. This expansion of the product range went hand in hand with the transmission of technological know-how to other parts of the ancient world: from Mesopotamia, glassmaking gradually spread to the Egyptians and later to the Greeks, who further improved it. Yet none of these societies developed glass windows.

Why was glass not used for architectural purposes in pre-Roman antiquity? Technical constraints must have played an important role. Glassblowing, the most efficient method of windowpane production in Roman times, was not yet known. Of course, glassmakers could have relied on other methods to produce panes—for instance, by way of casting. (Ancient Middle Eastern glassmakers were adept at various casting techniques.) However, it seems that the application of these skills to pane production would have been prohibitively expensive. Glass aside, windows in general were an architectural luxury in the ancient Middle East. Because of the limited availability of timber and stone, ancient Mesopotamian and Egyptian houses often were built from bricks made of clay or mud; to ensure structural stability, window openings tended to be relatively small and close to the ceiling.

This might explain why the Talmud defines the “Egyptian window” as “any window which is so small that a man’s head cannot enter through it.”

Climatic factors also influenced fenestration and made glass panes dispensable. In the Middle East, protecting the household from glare and stifling heat was, and still is, a far more pressing issue than shielding the interior from wind and rain. The design of the mashrabiya—the artfully carved wooden latticework characteristic of traditional architecture in parts of the Arab world—responds to these concerns. The exact historical origins of the mashrabiya are unclear, but the oldest surviving specimens date to the High Medieval period. It is probable that similar kinds of latticework were used in the ancient Mediterranean and Middle East as well. In fact, certain biblical verses suggest that perforated wooden screens may have formed part of the Temple in Jerusalem. We do know of wooden (and metal) grilles in Greek temples.

There are still considerable lacunae in our knowledge of fenestration in the ancient world. This holds true even for Roman times: as archaeologist Yvon Thébert has noted with regard to Roman provincial architecture, we often “do not know how many windows there were, what size they may have been, or where they were located; in most cases we do not even know how they closed.” Indeed, textual sources from antiquity are often terse or ambiguous in their discussion of windows and their sealants. To complicate matters, detailed architectural treatises from this period are scarce. Vitruvius’ On Architecture, written in the first century bc is the most comprehensive surviving treatise on Roman architecture.

This is precisely why it has been studied extensively since its rediscovery in medieval times—even though Vitruvius was hardly among the most distinguished architects of ancient Rome. In his treatise, Vitruvius discusses windows and their design at various points, but he is silent on the issue of how to seal them. This fits into a general pattern: what we know about Roman houses and their windows often is not derived from textual sources, but rather from archaeological findings. Taken together, the scattered evidence suggests that it was common among Romans (as it had been among Greeks) to close windows with wooden shutters or, alternatively, to furnish them with iron grilles. Gratings made of timber, terracotta, or carved stone were also in use. This observation, of course, does not justify the sweeping conclusion that “the form of ancient architecture was profoundly affected by the unglazed and stunted window.” Nor does it mean that residents had to “shut themselves up in darkness or stay within the brief circles of intense light cast by innumerable oil lamps.” Such belittling claims by modern historians tend to ignore the fact that windows were rarely the only sources of daylight in ancient buildings. In early Roman architecture, both public buildings and private homes often received additional light through the open courtyard, the atrium.

Well-preserved Roman homes, such as those excavated in Herculaneum and Pompeii, make it clear that even in the absence of outward-facing windows, there was enough light in the interior to enjoy the fashionable—and often elaborate—wall frescoes. The atrium’s importance for interior lighting was reinforced by the impluvium in its center: this water basin, often generously sized, served to collect rainwater, but the water surface also reflected the sunlight and thus helped to transmit it into the interior of the building.

Doors, too, played a significant role in interior lighting—and this included the front door as well as those connecting the atrium to the interior rooms. In this respect, early Roman houses stood in the tradition of ancient Greece, where doors often functioned as major sources of light. In ancient Greece, this even left traces in the language: while windows did exist in Greek architecture, there was no special term for them, so the window was referred to either as a mere “opening” or, tellingly, as a “small door.”

True, there were buildings in Greco-Roman antiquity that featured neither an atrium nor large doors—for example, the multistory apartment buildings that sprouted up in the densely populated neighborhoods of imperial Rome. Under such conditions of urban overcrowding, windows assumed unprecedented importance: every window meant more light. Still, even in these settings, shuttering the windows would not necessarily have implied total darkness. Sophisticated Roman shutters featured intricate latticework that admitted a gentle light even when shut. Simpler versions consisted of two separate, solid leaves. In Ovid’s love poetry we read about a room with “one shutter closed tight, the other just ajar,” which creates “the light the sort you often see in woods.”

Another way of sealing windows without shutting out daylight was to fit them with stretched animal skins or other organic materials, such as thin slices of shell, stretched animal bladders, or occasionally tow or other fibers. A far costlier option was to use panes made from alabaster, selenite, or mica. These minerals often were referred to by the umbrella term “lapis specularis,” a term literally meaning “mirror stone,” but perhaps best rendered as “windowpane stone.” Windows sealed with mineral materials accordingly were known as “specularia,” although this could also denote other kinds of sealants.

The use of mineral stones for fenestration may have had precedents in ancient Greece. That much is suggested by documents related to construction work in the sanctuary of Asklepios at Epidaurus in the fourth century bc. It remains unclear, however, which particular mineral stones were used at the Epidaurus site and whether this was an exceptional choice. We are much better informed about the production and use of mineral windowpanes in Roman times. As Pliny the Elder reports in his Natural History, lapis specularis was quarried in large blocks, which were then “split into plates as thin as may be wished.” This was a relatively easy operation because the blocks—due to their specific crystalline composition—often had a layered internal structure, which lent itself well to the slicing of panes. The final product had advantages over other sealants such as animal membranes: lapis panes admitted more light and provided better insulation while also being more robust and waterproof. Depending on the quality and thinness of the slices, lapis panes could even boast a remarkable degree of transparency.

The problem with lapis was its limited availability: the largest natural deposits were located in rather remote areas, and the mining process was challenging. As Pliny detailed, lapis often was “dug at a great depth by means of shafts.”

For a long time, the main sites of lapis mining were located in the Iberian peninsula. Egypt, too, was known to harbor significant deposits—and indeed, after the Romans conquered the kingdom in the first century bc, they sentenced criminals to mine lapis there. Smaller centers of lapis mining existed in Italy, Cyprus, and Anatolia.

Due to the high costs of extraction and transport, lapis was an expensive architectural material. To seal one’s windows with it was an unmistakable sign of wealth and prestige in Roman society. Indeed, in well-preserved excavation sites such as Pompeii and Herculaneum, archaeologists have found lapis windows (or what remains of them) near houses and villas that belonged to members of the upper class. A particularly important discovery was made in Pompeii’s so-called House of C. Cuspius Pansa, where archaeologists unearthed the remnants of twenty-five lapis panes. These panes, each sized about six by eight and a half inches, are thought to have been used to seal upper-floor windows. It seems that some Roman lapis panes were as long as thirteen inches.

Lapis windows were used not only in domestic but also in horticultural architecture. Growing fruits and vegetables year-round was a hobby of many wealthy Romans (including certain emperors), and the forcing houses in which plants were grown occasionally featured lapis panes. Bathhouses were another common site for specularia. These recreational facilities enjoyed great popularity at the time and were considered “hallmarks of Roman civilization.” Their construction was generously funded by political leaders vying for public favor. As a result, Roman bathhouses often displayed the architectural state of the art: they featured sophisticated technology for the circulation of water and steam while also boasting lavish building materials such as marble. It fits into this picture that costly lapis was sometimes used to seal the windows. Of course, this was also a practical choice, as lapis panes helped to preserve the heat and steam crucial for the operation of a bathhouse.

Excerpted from Transparency: The Material History of an Idea by Daniel Jütte. Copyright © 2023 Yale University. Reprinted by permission of Yale University Press.

Buy This Book