Turning our attention now specifically to Roman camps, such quantities of sewage, produced every day a camp was occupied, meant there must have been restrictions on where soldiers relieved themselves, and systems for removing the faeces from camp. Literary and archaeological evidence for Roman civilian toilet practices indicates that while urinating and defecating could occur almost anywhere, especially in streets, and even in rooms that lacked toilet facilities, numerous graffiti, tombstone inscriptions and literary comments also speak against such casualness and indicate it was particularly unacceptable in religious areas and by graves, and forbidden at public statues (Hobson 2009: 142–145; Magness 2012: 82; Merletto 2023: 4–5; Jansen 2024: 244–245); the provision of numerous private and public toilets in the Roman world (including over one hundred around Rome’s Aurelian Wall alone; Dey 2024: 137) also implies that the Romans were keen to contain such activities (Hobson 2009; Jansen et al. 2011; 2024). Practicalities and military considerations must surely have forbidden casualness in camps, since the resulting slip hazards and prevention of rapid movement could seriously hamper military operations.

Despite the evidence just cited, Roman culture generally lacked the modern obsession with cleanliness and hygiene. Such notions, including definitions of pollution and dirt, vary culturally and over time (Kira 1970; Martin and Russell 2000; Hobson 2009: 79; Bradley 2012; Magness 2012: 80; Koloski-Ostrow 2015). For example, in contrast to modern views, Josephus (Bellum Judaicum 2.149) appears surprised that Essene Jews washed themselves after defecating, as if they felt defiled by what was clearly a natural ‘cleansing’ act. There were nevertheless cleaning regulations for Roman streets, but archaeological and literary evidence shows these were not always followed, and Juvenal’s (3.248) probably exaggerated comments about sewer-like and rubbish-strewn conditions of streets are likely based on a degree of reality (Jansen 2000; Hobson 2009: 92–103; Merletto 2023: 4–5). A similar picture of dirty conditions should probably be envisaged for Roman camps and forts. This somewhat differs from modern, clean reconstructions of Roman sites which inspire assumptions that the Romans always constructed neat and perfectly aligned buildings with right-angled walls. Instead, Roman camps should perhaps be envisaged something more like the shanty towns that have formed since the early modern period commonly on the peripheries of cities. This is suggested by the Numantia siege camp at Peña Redonda, where the internal layout readily suggests that irregularity and ad hoc construction were the norm (Figure 3). The area of the hilltop occupied by this camp is reasonably level (Figure 4), so the irregular building lines do not seem to have resulted from manipulating them to set the long axis of ranges more conveniently at right-angles to a slope, and several of the alignments pay no heed to the contours anyway.

Archaeological evidence supportive of this ‘shanty’ image comes from Roman forts at Carlisle (late-first–early fifth centuries AD) and Bearsden (mid-second century AD) (Howard-Davis 2009; Breeze 2016). Although the lack of debris discovered in the Carlisle building interiors suggests they may have been reasonably clean, the streets contained material that would be thought of as rubbish today, with considerable butchery debris and waste from industrial processes on the minor streets in particular; alternatively, this material could be remnants from a system of dumping and frequent removal. There were also dumps and layers of material used for levelling-up, relating to site clearance and rebuilding, which contained various types of debris implying that middens were cleared and moved around the fort (Howard-Davis 2009: 520). A number of pits contained sewage that was human or animal, or both, deposited ‘on a fairly casual basis’, by which the excavator presumably means they served as toilets or simply had waste thrown into them whenever required (Howard-Davis 2009: 527). Moss was also found in the pits, a further indication that they likely functioned as toilets. A common alternative to toilet paper prior to industrialization, moss is being increasingly recorded in Roman forts (Huckerby and Graham 2009: 929; Breeze 2016: 327–330, 384; Dickson and Dickson 2016: 234).4 Moss is ideal because of its high absorbency and coincidentally (though unlikely appreciated in the past) is anti-bacterial due to having high levels of iodine.

Beyond the sewage pits, additional evidence of uncleanliness at Carlisle was the wide distribution of bones, which may indicate that butchery occurred along the sides of streets; part- or whole carcasses may have been distributed to each contubernium, where they were butchered. Not surprisingly, environmental evidence suggests there were puddles of rotting waste within the fort which formed breeding areas for large quantities of insects (Howard-Davis 2009: 527). These include insects that can spread pathogens and parasites into buildings such as eggs of the Trichuris whipworm; there was also copious evidence of house and stable flies (Musca domestica and Stomoxys calcitrans), which are known to spread diarrhoea, salmonella, typhoid and potentially poliomyelitis (Kenward and Hall 1995: 762; Howard-Davis 2009: 527; Smith and Tetlow 2009: 925–926).

Human sewage at Bearsden similarly included whipworm and additionally roundworm (Ascaris) (Jones and Maytom 2016: 301–303). The same two types of worm were also found in a pit containing human sewage at the first/second century fort at Ambleside, northern Britain (Jones 1985). Evidence from a third-century drain at Vindolanda fort by Hadrian’s Wall of the same types of worm and Giardia duodenalis, an intestinal parasitic micro-organism, has recently been published (Ledger et al. 2025: 4–5). Human fleas were found at both Bearsden and Carlisle (Howard-Davis 2009: 527; Smith and Tetlow 2009: 926; Breeze 2016: 371). At Bearsden, there were also varieties of beetles which like rotting hay and horse or cow dung (Breeze 2016: 371). The overall picture is similarly squalid conditions inside these forts, and the growing evidence for such pathogens in the Roman World (Jones 1985; Ledger et al. 2024; 2025) suggests such conditions could have been common in all types of Roman fortifications.

The pathogens present in camps and forts would have affected the military abilities of infected soldiers and hence the unit’s overall capability, as the infected men could have suffered symptoms that included breathing difficulties, reduced nutrition intake, dehydration, fatigue and diarrhoea, according to the specific pathogen, and some men may have been infected by several infections. Several of the pathogens can also be fatal.

The spread of human waste and animal remains as being the cause of such disease, illness and suffering was probably not totally realized in antiquity (Hobson 2009: 147); even today, diseases related to sanitation account for 10% of health issues globally (Prüss-Üstün et al. 2008). The health-impacting conditions could easily have turned from perilous to deadly within a camp. Hence it should come as no surprise that we read that dysentery killed or caused suffering for many of Pompeius’ soldiers at Numantia during the winter of 141/140 BC (App. Hisp. 78); and when Nobilior wintered there in 153/152 BC, some of his men died ‘inside the camp from the shortage of space and from the cold’ (App. Hisp. 47). The precise meaning of the space-related comment in the latter example is unclear, but presumably means conditions were so cramped they impacted on health; the reference to cold is straightforward, as the area’s bitter winter temperatures still killed people when Adolf Schulten excavated there in the early twentieth century (some of his field notebooks in LEIZA, Mainz, record names of his local excavators who died from cold during the preceding winter).

There is, however, one piece of literary evidence for camp cleaning. Polybius (6.33.3–4) comments that ‘two maniples’ (240 men; Dobson 2008: 48) were tasked to sweep and damp down the area along the front of the row of tents occupied by the tribunes, as this was the main location of daytime troop activity. There was actually more importance to this area than Polybius indicates, as it was the camp’s principal street, running right across it, with entrances at either end (hence its name of via principalis). As a result, it had to be kept clear and allow easy troop movement. Polybius does not indicate whether other streets were cleaned, but since he shows an understanding of the military importance of free access along a street, perhaps there were cleaning requirements for other streets; his source may have been a tribune’s military handbook (commentarius) (Rawson 1971, 15), which may not have included ‘minor duties’ of individual units and thus Polybius was unaware of them.

As will have become apparent, the human waste and debris that accumulated inside fortifications impacted on human health and military capability. Such accumulations must nevertheless have been addressed by the Roman army to prevent what would otherwise have become huge, obstructive piles of material. Possible management methods are discussed in the following section.

An obvious management method for dealing with human sewage in fortifications was to provide toilets, and such things were well-known in Roman civilian culture. The impression from the limited literary references to Roman toilet practices is that attending to such bodily functions was a social thing, confirmed by remains of multi-seat toilets, sometimes even with gaming boards etched between the seats (Hobson 2009; Jansen et al. 2011; Magness 2012; Koloski-Ostrow 2015; Hoss 2018; Merletto 2021; 2023). Facilities in houses usually consisted of chamber pots and a seat over a cesspit (Varro, Saturae Menippeae 192.104; Hobson 2009: 46, 133; Petznek et al. 2011; Petznek 2018). The latter were commonly not private and secluded, but ‘open’ and ‘social’, as the cesspits were frequently in kitchens and by the stove, where they doubled as kitchen rubbish bins (Magness 2012: 82; Koloski-Ostrow 2015). Although there are indications of this practice changing from the late Republic, toilets became increasingly segregated and linked to social hierarchy from the second century AD, evidenced by latrines increasingly occurring away from kitchens. Slaves and servants did not experience such change, as their facilities stayed where they lived and worked (Hobson 2009: 79, 168; Goldwater et al. 2011; Pérez et al. 2011; Koloski-Ostrow 2015: 6). This is a good reminder that, hygiene issues aside, cultures determine their own concepts of privacy, which can be complex and changeable (Kira 1970), with ‘isolated privacy’ generally being something of the modern, Western world.

The social norms established in non-military settings would have influenced soldiers’ practices in the field; they could naturally have followed their familiar home ‘social/communal’ behaviour and would not seek privacy or distance for bodily functions. There would also have been military benefits from soldiers engaging communally in as many activities as possible, as this practice naturally aids unit cohesion (Kira 1970). Literary sources provide no information about where or how far away Roman soldiers in camp went to relieve themselves. It would have been unwise to allow soldiers to leave camp for this, as the proportion of men needing to defecate first thing in the morning would result in a large part of the army being outside the camp at the same time (and vulnerable!), undermining security.5 A planned surprise and successful attack on a Greek army so disposed reveals the ill-advised nature of such practice (Xenophon, Hellenica 2.4.6; Anderson 1970: 66). Strategy aside, there are nevertheless other examples from the ancient Middle East of toilet practices being outside a camp (Deuteronomy 23:9–14; Magness 2012: 83).

The frequency of evidence for latrines inside Roman forts (e.g. 40% of 137 examples of forts in Britain) suggests they were common features (Goldwater et al. 2011, 136). Although there is much evidence of the various types of fort latrine, there is no published overview despite this being pointed out some years ago (Ebeling 2006: 126; Goldwater 2011: 136). Consequently, the relative frequency of each latrine type and other statistics are not readily available.

Multi-occupancy structures are the largest form of fort latrine, effectively duplicating civilian communal latrines. A well-known and excellent example of this type is at Housesteads fort, Hadrian’s Wall (Johnson 1983: 213; Mann 1989), with several other good fort examples in Britain (including Bearsden, mentioned above regarding unhygienic conditions, which now perhaps seem unexpected since the fort did have a latrine), Germany and The Netherlands (Johnson 1983: 213–214, with references; Bouet 2009: 373–383; Hobson 2009: 33–41; Goldwater 2011: 136, with references; Breeze 2016: 70–72; Fischer 2019: 244; Dütting et al. 2020). Such latrines are usually in the space behind the fort defences (intervallum) and on lower-lying ground compared to the rest of the fort. This encourages drainage, and reduces smell and aids hygiene as the latrines are distanced from accommodation. Smaller fort latrines occur in two types of location within barrack blocks. Firstly, as trenches in areas attributed to officers’ quarters, although the whole of an officer’s unit may have used them, as they are unnecessarily large for just one person (Goldwater 2011: 137; Koloski-Ostrow 2015: 59). Secondly, as pits in the front rooms of the two rooms occupied by each contubernium, likely meaning they may have been used just by that group (Goldwater 2011: 138; Koloski-Ostrow 2015: 59). There were individual latrines for fort commanding officers, which may relate to the cultural changes in toilet practice referred to above (Goldwater 2011: 137; Koloski-Ostrow 2015: 62). Finally, there were fort urinals, with several examples formed from an amphora or large pot (Simpson and Richmond 1935: 225–226; Richmond and Webster 1951: 6; Goldwater 2011: 138).6

The number of latrines per fort suggested by the archaeological evidence seems rather limited for a garrison’s daily requirements, so other containers (e.g. pots or general refuse pits) may also have been used. Soldiers may have gone outside the defences, when possible, but this could also have been too time-consuming and also impact security.

It is easy to appreciate why latrines were provided in forts, since they were occupied for some time, but the problem of human and animal waste would still have been present in the temporarily occupied camps. It was perhaps less pressing in camps occupied for only one or two nights, but would still have had an impact. At the opposite end of the scale, it would have been a serious issue for siege camps as these may have been occupied for considerable time, with significantly accumulating sewage.

Since the layout and general form of forts were essentially just the ‘permanent’ version of ‘temporary’ camps (Dobson 2008), it is likely that camps also had latrines. Camps could similarly have used their intervallum as a good location for communal latrines. They could have resembled the shared latrine trenches created by First World War armies, for example, with a seat formed by a pole above the trench (Figures 5 and 6). This form of latrine and the sequence of frequent trench digging and backfilling are detailed in an early twentieth-century British army manual (Figure 7; General Staff War Office 1911: 57–58, plate 38). Interestingly, when excavating Roman camps at Cawthorn, North Yorkshire, Ian Richmond (1933: 68–69) recognized one pit as resembling a British military field-trench latrine, even with postholes for seat supports (Figure 8).

In Roman camps occupied for any considerable time, potentially hundreds of such latrine trenches may have been dug and backfilled. In situations where the ground was too difficult to dig, an alternative may have been to mark specific latrine areas, perhaps by a circuit of earth or stones, which would also confine the sewage. The piles of faeces in these latrine areas would need periodic disposal, simply to create space for more or to remove a source of stench and troublesome flies. If ground conditions did facilitate latrine trenches, the camps with a longer occupation may still eventually have run out of space within the intervallum for rows of trenches. Unpleasant cleaning duties are consequently to be expected. Such could be indicated by a papyrus of the first century AD that lists some of the duties of soldiers of Legio III Cyrenaica in Egypt, which has often been interpreted as showing that M. Longinus was detailed for cleaning latrines (Fink 1971: 9.32g, 114; Johnson 1983: 214; Goldwater 2011: 138), but Kai Juntunen (2018) demonstrates that here, ad stercus more probably indicates Longinus was on dung-related duties at a stable or external dump.

It is likely that the rows of tents in camps could have had similar latrines to those found in barracks, as well as the intervallum latrines. Pits are common archaeological features in Roman camps. They could easily have been used as rubbish/sewage pits, like the pits at Carlisle. It is easy to imagine each contubernium conveniently having its own pit in the area in front of its tent, equivalent of the front room of barracks (Figure 9).

Although soldiers would also have prepared and cooked food in these areas in front of their tents (Roth 1999: 59), a very convenient toilet/rubbish pit here would merely reflect the arrangement in their kitchens at home, and so they would have been accustomed to the situation. The impracticality of creating pits in hard ground could have been solved by piling the material on the surface, perhaps with it wisely confined by stones or in a slight hollow (the latter may have formed naturally through periodic clearing, and moisture from the waste could have softened hard ground). Containers could also have been created in camps from amphorae or suitably large pots, like the urinals in forts (above). An example may be the large Iberian pot (dolium) standing by a wall in the camp at Travesadas, Numantia — although the pot could have been for other types of storage (Figure 10).

The pits, containers or designated surface areas by the tents for rubbish and sewage would periodically require emptying. Unknown at the time, this would have presented inherent risk of faecal-oral parasite infection for the soldiers emptying the faeces-contaminated contents and then spreading the infection to their colleagues (Mitchell 2015; Merletto 2021: 2; Ledger et al. 2024: 108). Faeces or surfaces touched by faeces seriously risked the spread of disease and illness. At least 120 known types of viruses can be transmitted through the faecal-oral route, including Salmonella, typhoid and norovirus; bacteria in faeces can additionally cause infections such as boils and ulcers (Merletto 2021: 2).

The smell from these sewage receptacles must have been strong (certainly to our modern noses) and probably lingered in the soil and pots after emptying. The contribution of the faeces to the smell was likely significant, for although it generally dissipates over a couple of days, since latrines were used daily and perhaps infrequently emptied, the overriding smell would have continued. It has been suggested that the latrine/rubbish pits acted like modern composting toilets (Koloski-Ostrow 2015, 50; Nissin 2022: 642), and potentially had no smell (Nissin 2022: 642). Their mixed contents would certainly have been suitable for aerobic composting, as any carbon added to the pits, such as wood ash from clearing fires, would help reduce moisture and add porosity, and the aeration that is vital to the process would have been aided by rubbish such as pieces of pot, bone, etc (see Otterpohl and Buzie 2013: 262 for discussion of aerobic composting); the whole resembles composting methods used by modern gardeners. Both the carbon and aeration would also contribute to reducing smell (Nissin 2022: 642). The composting action could have been prevented, however, if urine and other liquids went into the pits, as although water is required for aerobic composting, the ideal moisture content is only 50–60% (Otterpohl and Buzie 2013: 262). The addition of urine and other liquids could easily have taken the contents above this level. This would drown the vital composting bacteria and render the material anaerobic. Urine can also inhibit microbial decomposition by the ammonia that is released as urine ages, causing the material’s pH to rise above the microbes’ acceptable range (6.5–8.0) (Otterpohl and Buzie 2013: 262). The urine problem is well-known in modern composting toilets, which accordingly have urine diverters.

It would be natural for us to readily dismiss the Roman army’s sewage as having been useless waste, but it should be considered whether the soldiers regarded it as a useful resource. The main use of both human and animal sewage in the Roman world was as fertilizer (see below), but such usage would have been irrelevant to the army. Aged urine was used by the Romans as an antiseptic7 for animal wounds (Columella, De re Rustica 6.7, 6.11, 6.32, 7.5.9) and for some internal animal complaints (Columella, Rust. 7.5.15, 7.5.18), sometimes with specifically human urine being recommended (Flohr 2011: 150), so the soldiers may have collected some of their urine for tending horses and mules (animal urine was probably not used as it would have been mixed with bedding and unviable to collect when animals were away from their stalls/tethers). There would not have been bulk collection of urine, as often suggested in urban contexts, by pots and amphorae standing along streets for use in tanning, fulling and dying (see Flohr 2011: 150–153 and Jansen 2024: n. 72 for the history of this notion), as not only have such urban practices and industrial methods now been discounted (Flohr 2011: 150–151; Jansen 2024: 254), but these particular industrial processes did not take place on military sites, as leather and cloth were purchased by the army or acquired via taxation (P.Oxy 2230, 2760; Fink 1971: 63; van Driel-Murray 1985; Roth 1999: 159, 226–227; Kehne 2011).

The solid matter of the sewage probably had no use, although it could theoretically have been used as fuel.8 Faeces were used as fuel for cooking by some ancient cultures, but there are no indications the Romans did this (Ezekiel 4:12–15; Wilson 2011: 147; Magness 2012: 85). Logistically, it may have been challenging for the Roman army to use faeces in this way, as the faeces could have been mixed with general rubbish, as seen at Carlisle. One Roman army is known to have used cow dung for fuel (Livy, Ab Urbe Condita 38.18.4), so if an army was encamped long enough to dry manure, their animal dung heaps at least could have been recycled as fuel.

Having assessed the amount of human waste generated by the Roman army in temporary camps and more permanent forts, we now turn to a similar problem of animal manure, before discussing the environmental consequences of such waste.

Sewage would also have been generated by horses, baggage mules and the various numerous animals accompanying the army, such as draft and sacrificial oxen, other sacrificial animals and cattle for food; some armies also included elephants, who were present at Numantia (App. Hisp. 85; Roth 1999: 83; Murphy 2017; Milne 2024: 78–80). The animals generated a much greater amount of daily sewage than the soldiers, and it would have accumulated much faster, resulting in the army having to expend significant manpower on ensuring the debris did not obstruct the functioning of camps and forts. Even though the animal sewage would mostly be from herbivores, an exception being omnivore pigs (pork was an important part of Roman army diet; Roth 1999: 29–30), it still presented health hazards for the men, though perhaps this was not realized at the time. The many horses and mules could have caused numerous types of human infections not only through their sewage, but also by being in close proximity to humans; the pigs would have contributed similar problems, exacerbated if the meat was undercooked when eaten.9 The only positive thing regarding the animal manure compared to the human was that its smell may have been regarded as relatively more agreeable; though it is hard to imagine anyone liking the smell of pig manure.

Horses produce varying quantities of dung and urine, depending on their size, breed, diet and the climate, and, relevant to Roman army animals, very active horses produce slightly larger quantities than sedentary ones (ASAE 2005: 2). There are numerous modern studies of quantities and composition of horse manure/wet manure (both terms are used for combined dung and urine) from which we can deduce a reasonable idea of Roman horse manure quantities (e.g. Lawrence et al. 2003; ASAE 2005; Fabian and Zajaczkowski 2019; Chastain 2022). Skeletal remains suggest that cavalry horses in the Roman army were generally 10–15 hands (102–152 cm). An average of 14 hands (142 cm) is often cited, but we should keep the full range in mind, and there could have been fluctuations in the general size of horses over time (Dixon and Southern 1992: 165–167). Acknowledging this, 14 hands is a reasonable average to presume for assessing the quantity of waste produced by the Roman horses.

One of the main factors influencing the amount of manure produced by a horse is its body weight. Currently, a 14-hand horse is approximately 350–420 kg (Equine World 2026). By taking the middle of this modern range, reducing it slightly to account for an inferior diet in antiquity and the horses being in operational military conditions, we arrive at an estimated value of 380 kg for calculating the amount of manure produced.

The most frequently cited quantities of manure in relation to horse weight seem to be those of Fabian (previously Wheeler) and Zajaczkowski (2019).10 The metric equivalent of their figures means that for each kilogramme of weight, a horse creates manure consisting of approximately 31 g of faeces and 20 ml of urine per day.11 The 20 ml for the urine may need slight adjustment, however, since for example, Rumbaugh et al. (1982) state 29 ml and Horse and Hound (2019) gives a range of 15–30 ml per day per kilogramme of body weight. A more representative value may consequently be slightly higher than 20 ml, so the mid-point within the 15–30 ml range, 22.5 ml, will be used below for calculations.12 The resulting estimated average quantities of manure produced each day by a 380 kg horse are shown in Table 2.13

Soiled straw bedding should also be included in an assessment of the amount of debris produced by horses. Modern practices indicate the volume of bedding requiring daily removal is about double the manure (Fabian and Zajaczkowski 2019). How often Roman army stables were cleaned is unknown, but it would have been best to do this daily for the sake of the horses’ hooves, which suffer in wet conditions and would be seriously affected by the ammonia produced by decomposing urine (Bachrach 1988: 182; Hyland 1990: 124; Flohr 2011: 149). If the Romans adopted the same practice as today and removed twice the volume of bedding compared to the manure, it means the 0.021 m³ of manure had 0.042 m³ of associated soiled bedding, producing an overall volume of 0.063 m³. This equates to a 39.79 cm cube of material per day, though it may be reduced accordingly by periods spent outside the camp. This may seem insignificant, but a single horse would produce 23 m³ of soiled bedding in a year, enough to fill a 3.5 m square stall, for example, 1.88 m deep!

Returning to the example of the camps at Numantia, the location and form of the cavalry buildings are uncertain, but they were probably combined stable/barrack blocks and similar in layout to the stable/barrack blocks being increasingly identified in timber- and stone-built imperial forts (Rubel and Mischka 2023). These rectangular stable/barrack blocks consisted of two rows of rooms. One row was the accommodation for the men and the other row formed neighbouring stables for their horses. The size and number of the rooms suggest that each neighbouring pair of rooms accommodated three men and three horses (Rubel and Mischka 2023: 99–100). Stable/barracks can be similarly reconstructed at Numantia, with the row of rectangular stalls each being 3–5 m across and accommodating three horses (Dobson 2008). The amount of manure and bedding proposed above means the smaller of the three-horse stalls would each have accumulated 2.1 cm of combined debris every day if spread evenly across the floor (= about 15 cm a week or 77 cm a year!). The stalls must consequently have been cleaned daily, or nearly daily, not simply for the good of the horses, but to keep areas that soldiers needed to walk over clear of debris which could easily have become more than ankle deep.

The soiled straw bedding and manure would have been removed from the camp to prevent its bulk from quickly obstructing streets and forming areas of hazardous slippery material. The weight of this material was substantial; the equivalent amount of modern bedding material would be 4.04–5.7 kg, according to the composition of the bedding (Fabian and Zajaczkowski 2019). If the mid-point of this (4.87 kg) is used for calculation purposes, combining it with the weight of the manure (20.74 kg) means that 25.61 kg per horse would require removal, becoming a not insignificant 76.83 kg for the stalls accommodating three horses, and all or most of this would have to be transported out of camp (e.g. by buckets or carts).

Increasingly vast amounts of manure and soiled bedding would have been produced by a camp that accommodated cavalry. As commented above, the quantity of cavalry at Numantia is uncertain and probably differed for each army. Typical (i.e. consular) armies at the time theoretically had an overall complement of 2,400 cavalry (Dobson 2008). There would have been more horses than the actual number of cavalry, since extra mounts would have been required to cover death, injury and illness. Roman army remount practices are unknown, but an absolute minimum was presumably one for each squadron (turma) of 10 men. That results in a consular army having 2,640 cavalry horses, and potentially producing about 67.61 tonnes, 166.3 m³ of combined manure and bedding a day. There would also have been material from the horses of senior officers and potentially non-combatants. This would be variably reduced, though, by horses being away from camp on duties and for watering and grazing (Joseph. BJ 6.153; Vegetius, De re militari 3.8; Roth 1999: 128).

Pack animals would also have generated a significant amount of material. Each contubernium had at least one mule (App. Hisp. 85–86; Roth 1999: 77) and junior officers probably also had them. A consular army would consequently have at least 3,220 mules if each contubernium and junior officer had just one mule (for calculations, see Dobson 2022: 71), but increasing exponentially if they were allocated more than one. There is also an unknown number for senior officers and non-combatants. The amount of manure and bedding produced by mules and horses is similar (Chesapeake 2023), so the army mules could have daily generated approximately 82.46 tonnes, corresponding to 202.9 m³.

This results in the overall combined material from the horses and mules totalling 150.07 tonnes, 369.2 m³ every day (but subtracting material for when animals were away from camp). Using football again to visualize this, 1 m of material would be dumped on a pitch about every 19 days. If the urine were to be isolated (50.1 m³ a day), an Olympic swimming pool would be 1 m deep in 25 days. The rate would likely be higher than estimated here, because of the other animals associated with the army such as cattle.

If these statistics are extended to combine the sewage quantities of men, horses and mules, but not including the animal bedding to isolate the ‘pure sewage’, the resulting daily totals are:

In other, visualizing terms, every day the football pitch has 1.08 cm of faeces added to it and the swimming pool has 7.42 cm of urine added (only 13.5 days to be 1 m deep in urine!).

The quantity of manure and soiled bedding produced by the animals, the solid human sewage and general rubbish inside the camp, whether in pits/latrines or strewn around, means that most of it must have been frequently removed if only to ensure the army could move around the camp easily. The consequence of such clearance must have caused significant environmental damage in the surrounding landscape; given the effort of moving the material any distance, even if animal-drawn carts were used, it was likely that they instead dumped it near the defences, creating potentially vast manure and rubbish dumps.

The resulting concentrated smell from these external dumps and any debris still inside the camp must have travelled far. The issue of camp stench and the ill-effects this could have on an army were well-known in the ancient world. Vegetius (Mil. 3.8) remarks on this: ‘If a large number of soldiers stays too long in autumn or summer in the same place, then drinking-water contaminated by a polluted water-supply and air tainted by the general foul smell give rise to a most deadly disease’ (that disease likely being typhoid or cholera). Sallust (Bellum Iugurthinum 44.4) states that one Roman army involved in the Jugurthine War only moved camp ‘when the stench or need for fodder’ forced it. Onasander (Strategikos 9.1) also remarks that the smell of camps, and particularly those occupied in summer for any significant time, is apparent throughout the locality due to the concentration of sewage. It is perhaps not overly flippant to suggest that ancient military scouts could find their enemy simply by sniffing the air.

Since siege camps would inevitably have been occupied for some time, it is easy to imagine the smell and the nearby rubbish-dumps steadily increasing, and being more than merely locally ‘apparent’. For any active conflict or siege zone, there would also have been the growing stench from decomposing dead combatants and probably also horses, from both sides, accumulating between the city and the siegeworks. The scene is clearly expressed by First World War soldiers who frequently remarked on the ever-present and all-pervading stench of death coming from no-man’s land; a smell so strong and mephitic they could recall it years later (Giles 1977; Lynch 2006). Some Roman bodies may have been recovered if it was safe and practical to do so, perhaps being burnt on pyres near the camps (Noy 2000; Milne 2024: 186–189), which too would have created associated and far-reaching smells. The enemy dead near the siegeworks may have been left by the Romans, due to a lack of interest or to avoid unnecessary danger or labour; the same occurred during the First World War, with enemy dead decomposing on barbed wire entanglements months after an engagement. If the besieged did not recover their dead from outside their defences, these and the deprivations and casualties inside the city would also have contributed to the general pervasive stench.

Given how great this stench may have been, it seems surprising that the smell of warfare and particularly of the pungent sieges is not mentioned in classical sources. Perhaps writers regarded it as uninteresting to readers or maybe it simply went unnoticed. We can turn to conflicts from the nineteenth century to get an idea of the level of potential stench. British visitors to the siege of Sevastopol during the Crimean War highlighted the smell as the first thing they noticed when they arrived in the nearby harbour, and they comment on it more than once during their visit to the camps and siegeworks (Money and Money 1856: 64, 117–118). This sentiment is amplified by those living through the American Civil War, in which the siege of Vicksburg produced a ‘stifling’ stench, an unendurable ‘effluvia’ and the battle of Gettysburg resulted in nearby residents describing the odour becoming an ‘atmosphere’ not merely a smell (Smith 2015: 80, 113).

In addition to the smell, the sewage would also have caused significant water pollution. There would have been natural seeping of liquid waste through soils and rain-aided surface runoff from decomposing material in and around camps. This seeping waste — ancient leachate — could have been too strong to fertilize neighbouring vegetation, but instead kill it, and if concentration levels did not destroy the vegetation, any toxic compounds present would. If this liquid flowed into rivers or lakes, the resultant pollution could have killed fish, or high levels of potassium from the human sewage could have caused algal blooms and choked river life (Delowar et al. 2024). Significant local water fouling and pollution were also likely to result from the horses and mules being taken to the same water sources two or more times a day to drink; such water sources would have been the only viable method of providing the large amount of water that horses in particular drink at a time (General Staff War Office 1911: 53; Dixon and Southern 1992: 206–207).

Consequently, drinking such polluted water by either men or animals could have been hazardous (e.g. Ascaris, Trichuris and duodenalis pathogens, mentioned earlier in Roman forts, can all be transmitted by faecal contamination of drinking water; Ledger et al. 2025: 5), or it was simply undrinkable. Thankfully, the ancient world seemed aware of the dangers of foul water, as indicated for example by Vegetius (Mil. 3.2), who indirectly emphasizes this: ‘[an army] must not use bad or marsh water, for bad drinking water, like poison, causes disease in the drinkers’. Vegetius (Mil. 3.8, quoted above) also remarks on this in reference to armies being in one place for a long time. To avoid this he recommended often moving camp; siege armies unfortunately had no such choice.

One method of improving the drinking water for the soldiers was posca, their normal drink, and a staple of military culture (Merrow et al. 2020: 33–35). It was made from sour wine, almost vinegar, mixed with plenty of water and perhaps a little flavouring such as honey (the exact recipe is unknown). The sour taste helped disguise any bad-tasting local water and consequently encouraged the soldiers to hydrate more; intoxication was not a risk as posca contained hardly any alcohol. The high level of acetic acid in the drink also meant that bacteria in local water would be killed; a benefit that presumably was not realized at the time (Merrow et al. 2020: 33–40).

As well as an army dumping large amounts of waste and polluting water supplies, it also consumed large quantities of local resources extracted from increasingly distant locations as occupation continued. In addition to essential food and water for both men and animals (Erdkamp 1998; Roth 1999; Stallibrass and Thomas 2008; Merrow et al. 2020), each tent-group needed firewood for cooking and warmth (Caesar, Bellum Civile 1.84; Roth 1999: 59; Merrow et al. 2020: 84). Using a typical Numantia Roman army as an example, the approximately 3,000 tent-groups meant a huge overall daily quantity of wood would be extracted. Fulfilling such needs, or being unable to, could have fatal consequences, as illustrated at Numantia where soldiers on firewood duties were killed and deaths occurred simply from cold (App. Hisp. 47). Deforestation would also have occurred during camp construction, clearing the site and if the defensive stakes carried by soldiers (Caes. BGall. 7.11; Joseph. BJ 5.502–509) needed to be supplemented to create effective chevaux-de-frise fortifications (Figure 11) (Caes. BCiv. 1.42; BGall. 5.39; Livy 33.5.4–12; Polyb. 18.18.1–18). If more heavily fortified camps or lines of defences were needed, such as for sieges, timber requirements would have risen significantly. Additional local materials could also have been extracted to adapt tents for protection against the elements (Figure 12; Caes. BGall. 8.5), or to construct more comfortable and resilient timber or stone buildings in longer occupied camps (Dobson 2014: 81–84). The overall result was effectively environmental warfare by the army, despoiling the locality both through the extraction of resources and the leaving behind of a rubbish-strewn, polluted landscape.

The piles of sewage, manure and general waste produced by a camp would inevitably grow and spread into the immediate area as occupation continued. Over time, the army’s impact would steadily spread beyond the immediate area as nearby resources ran out and supplies were transported in from increasing distances. After the camp had been abandoned, the waste and manure dumps were probably welcomed by the locals as conveniently located fertilizer for their fields, since animal and human excrement-based fertilizers were common in antiquity (Cato, De agricultura 7.3; Columella De arboribus 8.5, 10.80–85, 23.1; Rust. 4.8, 5.9, 5.10, 11.3.12; Varro, De re rustica 1.38.2–3) and Varro (Rust. 1.13.4) says that latrines for slaves are sometimes even placed over manure pits (Flohr 2011: 149; Wilson 2011: 147; Nissin 2022: 643–644). There would have been risks, though, since sewage in the fertilizer was likely to include viable eggs of intestinal parasites and disease bacteria.14 These pathogens could enter the human food chain when mixed into the soil of the fields, becoming transferred to crops or ingested by grazing animals, but the farmers collecting and spreading the material would have suffered more quickly due to these pathogens entering their bodies directly via skin, cuts, or a casual touch of the mouth or eyes (Hobson 2009: 150–151; Aspöck et al. 2011: 163; Heirbaut et al. 2011: 16–17; Koloski-Ostrow 2015: 86; Mitchell 2015; Merletto 2021: 2). A modern demonstration of the latter type of transfer occurred in Denmark, where children suffered roundworm infections after their school garden was fertilized with pig manure (Ledger et al. 2025: 6).

Unbeknown to the local population or the Roman army, the camp had created a form of long-lasting, invisible germ warfare. They would also probably not have known about the dangers of using ‘fresh’ fertilizer and that human and animal excrement only become safe once the material has fully composted and lost any smell, which usually takes at least six months (Jensen et al. 2008: 437). The numerous leaching cesspits, rubbish dumps and latrines would also have polluted the area inside camps, regardless of how frequently they were emptied, and most were presumably also left uncleared when the camp was abandoned. Neutral conditions are unlikely to have been reached for several months, through natural composting of the pit and latrine contents.

The consequences of the stench, pollution and environmental warfare were that wherever the Roman army encamped, everyone in the locality would have known about it. The soldiers may even have had the worst deal, as they were right on top of the source of the problem, with their continued presence making it worse every day and their health suffering because of it. The military then took their problem and associated suffering to the next locality, but they left behind lingering devastation and cause of illness for those living nearby.15