Indeed, Rome was not built in a day. Although its close proximity to debris from the volcanic eruptions of Mt. Vesuvius allowed it to be built quicker than most cities. This paper explores the history behind mines, the mining techniques and methods of transportation Romans used to take advantage of this vast mineral wealth and use it to build their thriving city and iconic buildings. We also see the working conditions of a slave in the mines and how the demand of Rome’s elite during the Late Republic drove production. Class lectures from Adrianno of Roma Sotterranea, our visit to the tuff mines and online research are how I reached my conclusions. I learned that Romans were as practical as possible in their methods and sought to eliminate extra work when possible. Waterways were used as often as possible to transport goods rather than the land. The seemingly endless amount of slave labor Rome had allowed work to be done on a trial and error basis, deaths in the work site were commonplace. However, the treatment of a worker depended on his social status. Skilled artisans such as stone masons were not worked to death. Workplace death rate could have only skyrocketed in the Late Republic when Rome’s elite started demanding marble as a token of victory for their military victories. Greed for material wealth was a direct cause for poor treatment of slaves in ancient Rome.
It is a challenge for archaeologists to accurately determine Roman mining techniques because little recoverable evidence remains. Destructive methods such as fire-setting damaged tools and artifacts. Miners lit a fire directly in front of the rock face to be mined and quickly cooled it down immediately with cold liquids, quenching the rock by cracking and weakening it with thermal shock. As a result, many artifacts that have been found are damaged. There are also accounts written by hundreds of authors including Vitruvius, Diodorus of Sicily and Pliny have been found and offer insight into Roman mining methods. The writings do not mention any locations, only what was observed. The difference between mining and quarrying is the material sought after and whether it’s under or above ground. Mining is done for minerals while quarrying is done for stones.
Figure 3: Dolaucothi gold mine adit (Google Images)
This method involved extracting ore from the ground’s surface by using flowing water to wash away organics, or mining the side of an open rock face. It was the easiest and safest method and did not require miners to face the dangers of underground mining. Opencast was done to obtain gold ore from ‘placer’ deposits which are found in the slower areas of streams. Gold sinks to the bottom of a river due to its high density and will accumulate where the water is slowest because it spends more time there. Pliny notes the high quality of opencast ores in rivers and streams: “No gold is more refined, for it is thoroughly polished by the very flow of the stream and by wear” (Duncan, 1999).
Opencast was also performed on dry surfaces, Roman miners simply brought water with them since it was so helpful with this method. In the case of the Dolaucothi mines, the nearby Cothi river was used to bring water by aqueduct. Tanks were built directly above the placer deposits to hold enough water to wash away the top organic layer of soil and expose the gold ore in a technique known as hushing. The running water also helped Romans by giving them a convenient way to wash away impurities from the ore before sending it to the smelter (Greene, 1986). Romans dug tunnels only when ore was far underground and it was considered worth it, the hazards of tunneling were learned on a trial and error basis.
Figure 4: Simple deep-vein diagram (Duncan)
Figure 5: Iron axe pick marks in a tuff mine (Author)
This technique involved using mainly iron tools to chip ore deposits underground, and lift them out with a basket. Workers used this technique to mine a majority of their minerals including copper, lead, silver and tuff. To begin, vertical shafts were dug at equally spaced distances in order to speed up the process by having numerous dig sites. To avoid the need to bring ladders, miners dug handholds into the shaft walls. These were used both for climbing in and out and for raising baskets of ore. Once at the desired depth, miners began digging horizontal tunnels called galleries towards one another. Adits were also used occasionally to prospect a dig before investing the manpower to mine, and for drainage which is discussed later. In the case of tuff tunnels, only iron axes were needed. This worked well because of the stone’s soft material characteristics. Usually, stone hammers and wedges were used first on harder rock. Lifting stones out of the mines required a large amount of effort and sometimes were the cause for a mine to be abandoned if not for drainage; both are discussed later on.
Writings mention common risks to deep-vein miners and their observations of the effects. For example, Pliny writes, “when well shafts have been sunk deep, fumes of sulfur and alum rush up to meet the diggers and kill them” and “The fumes from silver mines are harmful to all animals” (Duncan, 1999). 3 issues were challenging in particular: lack of lighting, poor ventilation and drainage.
Figure 6: Hole for miner’s lamp in an aqueduct (Author)
Figure 7: Typical Roman lamp (Google Images)
Without electricity, lighting was an issue for Roman miners. Aqueduct channels sometimes meandered from side to side; we observed this during our aqueduct visit and learned that this was due to lack of light at that point in the dig. This is not always the case in mineral mines because the miner’s goal is to follow the vein rather than try to follow a straight line. In order to hold lights in place and move forwards, miners dug small holes in the walls and used them to hold oil lamps similar to those found in Roman homes. Pliny wrote that the candles were also used to measure a day’s work, shifts completed when the candle’s flame died (Duncan, 1999). Although necessary, the use of fire had to be limited in the tunnels to not deprive the air of oxygen and suffocate workers.
Before entering a shaft, a lamp would be lowered into the tunnel to determine if the air was safe to breathe. If there was no oxygen, the flame would die out and additional shafts would be excavated on either side to improve air flow by convection. Galleries which met also helped improved air flow. A more risky method of increasing air flow was to set fires, but as mentioned before this had to be done carefully to not deplete oxygen levels. Workers could also wave linen cloths in the mines to increase air flow. Heat worsened working conditions; for every 30 meters deeper, the temperature increased by 1 degree Celsius (Duncan, 1999). In the case of the Paphlagonia mine along the Black Sea Coast, slaves were constantly replaced due to the poor ventilation. Diodorus notes that “there are more than 200 workers but they are continually consumed by sickness and death” (Sherwood, 1998).
Figure 8: Water table inside a tuff mine (Author)
The volume of leakage into a mine determined whether it would be usable or not. Miners who dug near the water table and subterranean rivers often found the infiltrating water to be too much and abandoned it. During my visit to a tuff mine, we saw for ourselves where the water table was. If the leakage was manageable, workers had methods of draining the water from the site including mechanical devices.
After the mine was excavated, small diversion channels called adits were dug which diverted water away from the work site. Similar to aqueducts, adits have a very slight slope that allows gravity to move water. If the miners were lucky, a naturally occurring hole nearby could act as a water tank and hold all or some of the drainage after excavating the adit. Before it was dug, water would have to be handled by other means.
If the incoming leakage was relatively small, workers could form a fireman’s line and bail out water from the mine. Bottoms of buckets were pointed so that the containers could be easily tipped and filled. Lifting them required a strong arm as they could hold up to 150 litres (150 kg) of water plus the weight of the metal container itself.
To remove larger volumes of water, miners used mechanical devices powered by humans walking on pedals or turning cranks. The Archimedean or Egyptian screw was constructed by fitting a metal augur with a crank at the receiving end inside a hollow wooden cylinder. The device worked by trapping small volumes of water between each vane and bringing them up. Multiple screws could be used in succession, lifting water as high as was necessary. Vitruvius specifies an optimal angle of 37 degrees to remove the maximum volume of water, about 35-40 gallons per minute (Duncan, 1999). It is interesting to see the same type of augur used in today’s construction industry to remove soils from the ground.
Figure 9: Archimedean Screw (Google Images)
The water wheel was introduced not long after the augur screw. This system is similar to modern water wheels, except they were powered by workers who stepped on it like a staircase instead of with flowing water. Before the wheel could be used, a sump in the mines would have to be excavated to contain water and make space for the wheel. Rectangular containers along the wheel’s circumference filled when they passed the bottom of the rotation and then emptied at the top, raising the water. Each wheel had to be designed with a diameter large enough to bring water up to the desired level, and were often used in series of pairs. At the Rio Tinto, 8 pairs of wheels were found which raised the water a total of 30 meters (Duncan, 1999).
Figure 10: Human powered water wheel (Duncan)
As each wheel rotates it creates turbulence, which makes it more difficult to catch water in the containers. To remedy this, each pair of wheels was situated so that they rotated in opposite directions. This counter-acted the turbulence of each wheel and improved efficiency. This was important because the sump had to be slightly sloped towards the wheel to return water to it, increasing the depth of each container dip and maximizing water removed per scoop. Positioning the pairs of wheels like this reduced the slope needed and made construction easier. Wheels that have been recovered average “4-6 meters in diameter with 20-24 compartments” (Duncan, 1999). The advantage of the water wheel was that it could raise water higher, but at a slower rate than the Egyptian Screw. The performance of the wheel depended mostly on how far the buckets could be dipped into the water and where the water was emptied in the rotation. Controlling turbulence and designing the wheel with the right diameter were important considerations when using this machine. Drainage was arguably the most challenging problem for Roman miners as all removal processes required a large amount of resources and was not guaranteed to work.
If a miner was a slave, how he was treated depended on his social status; if he was privately owned, it was in the owner’s best interest to keep him healthy to maximize profits. State owned slaves (and sometimes their families) on the other hand were often treated terribly because they were mostly condemned criminals or prisoners of war sentenced to be worked to death in the mines. Diodorus of Sicily notes the poor working conditions in Egypt where workers were always chained together and worked night and day “under the hard supervision and blows of an overseer” (Sherwood, 1998). Guards that were assigned to watch slaves could not speak their language to ensure that no communication could happen that might make the guard have pity for the slave. Slaves were used in large numbers for their physical strength and not for their trade skills. On the other hand, stone masons were skilled artisans in charge of making the final product look good and be stable. They were not worked to death.
Stone Masons and Quarrying Techniques
The Stone Mason
Figure 11: Porous tuff not suitable for construction (Author)
Before the invention of standard size brick masonry, a majority of stones were custom cut by skilled artisans called stone masons. The material strength of stone was split into 6 quantitative categories: very soft, soft, semi-firm, firm, hard and cold. Stones in the soft range often included less compact tufa which were widely used in construction for their ease of cutting while marble and granite were considered cold and used by the elite class and to decorate the facades of buildings. Stone masons did not categorize stones by material make-up, but by the purpose they would serve. This system worked best because the hardness of stones is not consistent across the board, there are “plenty of limestones that are harder than certain marbles, and some marbles have to be worked a little like granites” (Russell, 2013). For this reason, color was not used often to identify stones because it did not help masons identify its purpose. Categorization by purpose rather than composition is evident even today in the case of Purbeck Marble in Britain, where it is actually is a limestone. The modern name comes from long ago when Romans invaded Britain for its mineral wealth and classified it according to its purpose.
Ability to take a finish was an important consideration for stone masons. Softer stones could not hold as much detail because their surfaces wore away easily while colder and harder stones could hold a high amount of detail. The Latin word ‘marmor’ refers to a stone’s property of being able to take a polished finish, stone masons often used this word when classifying stones.
If the local supply of stones was diverse, masons could designate certain types of stone for specific applications. For example, it was important to not use porous tufa as a structural element if it would be exposed to weathering. Vitruvius notes the pros and cons of soft tufa: “the stones which are not hard have the advantage that they can be easily cut and are good when used in covered places, but placed out of doors, the frost and rain turn them to dust…” (Adam, 1994). In other words, the water that infiltrates the pores causes freeze-thaw cycles which rapidly wear the stone. Vitruvius recommends quarry sapping as a factor of safety before using tufa. This method involves stocking quarried stone in an outside and covered location for a couple years. The quarrying had to be done in the summer so that the stones had minimal moisture content. Stones that wore too much did not make the cut and were crushed and used as aggregate.
Figure 12: Modern touchstones, karats increasing to the right (Google Images)
When a vein of metals was found, a touchstone was first used to determine the quality of precious metals present. The touchstone is made of a fine-grained black siliceous rock that retains the streaks of other metals when rubbed against them. Impurities are then removed with acid treatment, and the contents of the vein are determined. This method is still used today to find how many karats are in a gold sample.
Figure 13: Active quarry site (Adam)
To begin quarrying, the top layer of organic soil is removed to expose the rock face. Next, quarry men outline blocks of stone according to the size needed with iron picks and drive metal wedges in with mallets to form cracks until the stone comes loose. An alternative technique was to force dried wooden wedges wrapped in cloth into the cracks and expand them with water, forcing the stone loose. If the quarry men were lucky, a fault line might help them by acting as one of the cracks. Stone masons had to consider both the quality of the strata and if there were any natural faults to speed up the quarrying process. From our visit to the travertine quarry, we learned that a straight strata with no curves indicates a strong stone while curved strata are signs of weakness. To make sure tunnels did not cave in, miners left one or more strata above them to provide enough strength. Workers also carved out stone pillars at intervals along the length of the tunnel to increase stability.
When the tuff vein pointed downward, vertical shafts were dug to continue operations. Mining would continue until the vein ended, drainage became an issue, or the effort of lifting stones from deep in the ground became too much. Immediately after loosening the stone from the quarry, the stone mason roughly shaped it to the size that was needed. It was important to leave enough excess stone so that it could be moved without damaging it, but not so much that transportation costs were too high (Farr, 2014). After removing the stone and making sure it was strong enough, it was taken to the construction site to be used.
Romans used workers and machinery for light and heavy loads, respectively. In the case of lighter loads, they were easily lifted by groups of slaves on scaffolding. On the heavier and more technical end, machinery was required whenever lifting was done. The power and type of machine changed depending on the weight of the object. Romans took pride in their ability to construct large scale buildings and viewed them as technical achievements.
Similar to how workers could bail out water from mines in a fireman’s line, the same could be done for transporting light loads of materials. Another method was to use a water wheel to raise tools and materials up instead of water. The pragmatism of Roman construction is shown here where two different scenarios use the same methods. In the case of quarried stones, workers used wooden rollers to help them overcome friction before they had to be lifted. Ropes were also used to help the workers get a grip. Access ramps were used when the stone needed to be transported up or down slightly so that it did not need to be lifted. When lifting was finally needed, Romans used Greek lifting methods which were sufficient according to Vitruvius (Adam, 1994).
Figure 14: Idealized Pulley Multiplication (Google Images)
If the load was less than the weight of the worker, a simple pulley could be used to bring up objects. If the load was heavier, a modified pulley could be used to lift it instead. By combining multiple pulleys and distributing the rope’s tension, it was possible to reduce the effort required to lift heavier objects. The benefit of this is that the force required to lift the object is proportional to the number of pulleys, the trade off is that the load will be lifted more slowly because there is also proportionally more rope to pull to achieve the same height. Romans further improved this system again by taking advantage of worker’s weight. Levers attached to a hollow wheel with steps replaced the manual winch, which allowed workers to step on the wheel and cause rotation just as with the water wheel. More workers could be added as the size of the load increased.
Figure 15: Simple Winch (Google Images)
The winch was the first appearance of reduction ratios (Adam, 1994). This system has the same trade offs as the pulley system, but works differently. The force required to lift loads decreases as the difference between the crank handle length and roller drum circumference increases. Torque is created on the roller drum as the crank is turned, bringing the object up. The torque will increase with handle length. The downside to this is that the handle has a longer distance to travel as it gets longer, meaning the load is picked up at a slower speed. Romans combined the pulley and winch together to make their version of a crane.
The versatile Roman crane is described by Vitruvius and confirmed by paintings found by archaeologists. Two beams of thicknesses depending on the size of the load were separated at the ground and joined at the top like an upside down V to form the jib. A pulley system was attached to the top of the V to lift objects, and the winch was cranked from the bottom. To swing loads back and forth, pulley systems were attached as supporting legs that could move the jib (Adam, 1994). Romans outfitted these cranes with walking wheels as well and used pulley multiplication if the load was heavy enough. It was important to make sure that the two ground supports of the beams were stable, as the whole system would fail if the jib became undone.
Figure 16: Typical Roman lifting connections (Adam)
Romans took after two of the Greek methods of lifting; handling bosses and lifting pins, and invented the use of grips. Evidence for these methods can be seen today at archaeological sites. During our visit to the Flavian Amphitheater, we saw holes in many of the stones in the walls and learned that lifting pins used to be there. The metal was later stolen by thieves.
Handling bosses were symmetrical protrusions on two opposite sides of the rock face where rope slings could be attached for lifting. Occasionally four bosses have been found on larger stones, probably for added stability when lifting. This method was discovered on sites where work was interrupted because the protrusions were removed after the stone was placed. Greeks carved U shapes into stones instead of making protrusions so that the extra work of removing handling bosses at the end of the project was eliminated.
Lifting Pins (Lewises)
This system consists of 3 separated metal rods with cavity holes which are all anchored into the stone. A horseshoe shaped hook with the same size cavities is put in between the metal rods such that the cavities line up. Lastly, a bolt is passed through all of the holes to form the lifting pin. When considering the connection’s strength, it was always the stone that broke first. Therefore, the depth to which the metal rods were anchored into the stone was increased with load size rather than the thickness of the lifting pin itself.
Pincer grips required the least amount of preparation, it was only necessary to make sure that the end of the grip dug into a small hole in the stone. Lifting action further increased friction between the grip and stone because the pincers acted like a pair of scissors: the weight of the stone forced the forceps together, digging the grips further into the stone. If the grips were put on the faces of the stone that would not be part of the exposed surface, it isn’t possible to tell that this lifting technique was used as the holes are hidden from sight.
Modes of Transportation
Figure 17: Quarries used in construction of the Baths of Caracalla (DeLaine)
In addition to selecting the right materials for the job and handling them correctly, they had to be sourced as closely and conveniently as possible to minimize construction delays. Many Roman quarries have been found near near rivers so that materials could be shipped along them. For example, the figure below shows possible sources of building materials used in the Baths of Caracalla, all located near a river (DeLaine, 1997). Vessels could be rowed by slaves, which was much easier than land transportation.
Figure 18: 60 pairs of oxen bring Mussolini’s monolith to be carved into his obelisk (Adam)
Figure 19: Finished obelisk (Google Images)
Romans used carts and animals to travel on land. For lighter loads, caravans of donkeys and mules were used to distribute weight. Smaller donkeys could carry about 55 kg balanced on either side while larger mules could carry double that. For even heavier loads, oxen were yoked together in increasing number depending on the size of the load. It took about 12 pairs of oxen to pull a 10 ton load. One disadvantage to using many oxen was maintenance, if one animal ‘broke down’ the entire operation had to stop. As the number of oxen increases, the probability of breakdown is higher and more frequent (DeLaine, 1997). Overall, use of the waterways clearly beats the strength of any animal. This advantage is observed in the construction of Mussolini’s 560 ton obelisk where it took only 60 pairs of oxen to transport it (Adam, 1994). On land, more than ten times that amount would have been required.
River and Sea Transportation
Rivers allowed Romans to move goods with relatively low cost (although still pricier than sea travel) and without risk. The challenge rivers sometimes presented was their shallow depth. In the case of the Aniene River, flatter vessels had to be constructed so ships didn’t scratch the bottom of the surface and capsize. In more extreme cases it was necessary to wait until the rainier seasons of Winter or Spring to use certain rivers so that the water level was high enough. Transporting vessels upstream was an issue, although still cheaper than land travel. One economical way to get around this was to make temporary rafts only meant to go downstream and burn them as fuel or use them as timber after unloading them (Farr, 2014). If this was not possible, slaves or oxen were used to tug vessels back to the material source to be reloaded.
The Market for Stones
Figure 20: Exposed brick, Herculaneum column (Author)
The rapid expansion of Rome during its Late Republic put a large demand on quarrying stones for building materials and decoration. Romans saw stone buildings as a proper way of building, following the Greeks (Russell, 2013). Decorating these buildings with marble and granite was costly. To get around this, many cities used economic construction methods. For example, it was common practice to build a column’s cylindrical core out of cheap bricks and then finish it with an aesthetically pleasing material. I observed this in the cities of Pompeii and Herculaneum where construction was brought to a halt after the eruption of Mt. Vesuvius in 76 A.D.
Figure 21: Brick core construction, Pompeii (Author)
Figure 22: Ostia Antica domus with mismatched marble panels (Author)
I also learned from my visit to Ostia Antica that marble was scavenged as the Roman Empire was falling to supply the material needs of decorated houses called domuses. People did not care if the color matched, but only that it was marble. We saw at least three of these houses that had mismatched marble to decorate the interior.
Marble in the Late Republic
Greeks constructed their buildings with local materials as often as possible, only shipping them long distances when needed due to high shipping costs. Over time, the architectural and sculptural building style of the Greeks became so popular in Roman culture that stones would be imported to build with even if there were none available in the area. Building this way was an expectation and was “indicative of a level of cultural and economic connectedness” (Russell, 2013).
The demand for marble exploded around 200 B.C. when the Republic was constantly conquering new land, any marble that the victims had was taken and brought home as the spoils of war. Those in power sought to build as much with marble as they could to display their power and wealth. For example, the Temple of Jupiter Stator built in 146 B.C. after a Roman victory in Macedonia was constructed with imported Greek marble and even designed by a Greek architect. Construction projects like this show how heavily Greek architecture influenced the Roman elite in the Late Republic. Plutarch, a Greek historian at the time wrote that “the conquered should give place to the conquerors’, and for the individual conquerors, the various triumphatores of this period, keen to assert their supremacy over their political rivals, eastern marbles became potent tokens of victory”.
Tuff Mines Today
Rome’s mines have served many purposes over the millenia; a source of building material, a place of refuge for Christians under persecution, a resting place for the dead, and currently a popular tourist attraction.
Figure 1: Quarried tuff stones (Google Images)
Beginning with Rome’s former rulers, the Etruscans dug the first tuff mines to gather soft volcanic rock to build with. These excavations were later reused as a refuge for Christians in fear of being sentenced to horrible deaths by Emperor Nero.
In the summer of 64 A.D, a massive 6 day fire consumed nearly three quarters of Rome. Afterwards, the public blamed the emperor for setting the fire for his own amusement. Nero dodged this bullet by using Christians as a scapegoat and blaming them. This worked well because they already had a bad reputation in Rome for proclaiming the existence of a new king, which was taken as a threat to those in power. Nero first rounded up a few Christians and used interrogation techniques to expose the identities of many more. Nero then put a large amount of Christians to death as a form of public amusement, sometimes lighting them on fire and using them as evening street lights (Carrington, 2000). St. Cyprian, an early Christian writer wrote of Christians being worked to death in the mines as well (Sherwood, 1998). After the fire ended, a Twelve Tables law (Table 8, Law 1) was written requiring a 2.5 foot gap between buildings to minimize how quickly fires could spread.
Figure 2: Underground tuff mine with calcium carbonate deposits (Author)
Christians may have been confused with Jews and labeled rebellious and lazy as a result. These labels were a result of uprisings Judean Jews had against Roman provincial government and the fact that they did not work on the Sabbath; this may have added some fuel to the fire (Brians, 1998). In accordance with Rome’s resourceful re-use of infrastructure, the catacombs found another use when the stench of decaying bodies became too much for Rome to handle. Another Twelve Tables law (Table 10, Law 3) was written which forbade the burial and cremation of bodies within city limits. Christians of that time practiced burial rather than cremation, so they began digging out body sized holes in the tunnel walls.
Today, the bodies have been removed from the excavated portions and curious tourists are invited in only with tour guides to avoid getting lost in the seemingly endless maze of tunnels. In a world without power tools or knowledge of the dangers of underground mining, Romans faced many unknown dangers and overcame them with trial and error. The disposable amount of slaves they had surely helped them accomplish this.
The magnificent buildings and roads of Rome we see today were a daunting task to build without modern technology. Slave labor helped Romans develop awareness to the dangers of mining overtime through trial and error. The writings of ancient authors tell us about the practical techniques that were developed to overcome these problems. During the Late Republic, Rome’s marble economy exploded due to rapid expansion of the empire. The materialistic nature of Rome drove the mining industry, which in turn drove demand for slave labor.
Brians, Paul, Mary Gallwey, Douglas Hughes, Azfar Hussain, Richard Law, Michael Meyers, Michael Neville, Roger Schlesinger, Alice Spitzer, and Susan Swan, eds. Reading About the World. Vol. 1. N.p.: Harcourt Brace Custom, n.d. Print.
DeLaine, Janet. The Baths of Caracalla: A Study in the Design, Construction, and Economics of Large-Scale Building Projects in Imperial Rome. Portsmouth, RI: Journal of Roman Archaeology, 1997. Print.
Duncan, Lynne Cohen. “Roman Deep-vein Mining.” Roman Deep-vein Mining. N.p., 9 Dec. 1999. Web. 16 Sept. 2015.
Farr, Jason Michael. “Lapis Ganibus: Tufo and the Economy of Urban Construction in Ancient Rome.” Diss. U of Michigan, 2013. Print.
Greene, Kevin. The Archaeology of the Roman Economy. Berkeley: U of California, 1986. Print.
Humphrey, John William, and John P. Oleson. Greek and Roman Technology: A Sourcebook: Annotated Translations of Greek and Latin Texts and Documents. Lonson: Routledge, 1998. Print.
“Nero Persecutes The Christians, 64 A.D.”, EyeWitness to History, www.eyewitnesstohistory.com (2000)
Russell, Ben. “The Economics of the Roman Stone Trade.” Oxford University Press, 2013. Web. 18 Sept. 2015.