UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Formidable genus armorum: the horse archers of the Roman Imperial Army McAllister, David W. 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_spring_mcallister_david.pdf [ 6.21MB ]
JSON: 831-1.0058215.json
JSON-LD: 831-1.0058215-ld.json
RDF/XML (Pretty): 831-1.0058215-rdf.xml
RDF/JSON: 831-1.0058215-rdf.json
Turtle: 831-1.0058215-turtle.txt
N-Triples: 831-1.0058215-rdf-ntriples.txt
Original Record: 831-1.0058215-source.json
Full Text

Full Text

FORMIDABILE GENUS ARMOR UM: THE HORSE ARCHERS OF THEROMAN IMPERIAL ARMYbyDAVID WILLIAM McALLISTERB.A., The University of British Columbia, 1991A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(Department of Classics)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAApril 1993Copyright^© David William McAllister, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(SignatureDepartment of  C L4C,S (C 5The University of British ColumbiaVancouver, CanadaDateDE-6 (2/88)ABSTRACTThe equites sagittarii formed a considerable force in theauxilia of the early and middle empire, but they, and particularlytheir role and employment, have received little scholarlyattention. This paper will attempt to define their place on thebattlefields of the high Empire.While horse archers are known to have been used in easternarmies such as the those of the Persians and Parthians with greateffect, they were recruited into and employed by the Roman army instrength only starting at the reign of Augustus. Units of horsearchers were almost exclusively recruited from eastern lands andwere employed there as well, not only in war, but also as theborder garrisons of Pannonia, Dacia, Africa, and the Levant.Chapter one discusses the equipment in use by horse archers.The bows in use were invariably compound recurved bows of easterndesign. Other specialized archery equipment was also eastern inorigin while their general military and cavalry equipment seems tohave been standard Roman military issue.Horse archers, like all military systems, can be defined interms of their characteristics: qualities which determine theirpattern of effectiveness, and therefore role, on the battlefield.The characteristics of horse archers are: flexibility, mobility,vulnerability, and firepower. Chapter two analyzes mountedarchers' characteristics to explain in detail the reasons for theireffectiveness.Chapter three looks at the question:^Were these unitsiiinvariably bow-armed? It appears that they had at their disposala variety of weapons (including javelins and other weapons) for useaccording to the situation.Chapter four analyzes the evidence of the role of horsearchers in all phases of war: the advance to contact, the attack,the pursuit, the defence, and the withdrawal. They were wereespecially valuable in independent battlefield operations, due totheir mobility.Appendix one is a list of known and suspected units ofsagittarii, together with a brief commentary on certain of themounted units.Mounted bowmen's firepower has a certain pattern ofeffectiveness. Appendix two analyzes arrow fire throughapplication of the Theory of Small Arms Fire in order to explaincertain paradoxes of archer employment on the battlefield.iiiTABLE OF CONTENTSAbstract^ iiTable of Contents^ vAcknowledgements viINTRODUCTION^ 1Chapter One^Weapons and Equipment^ 131. Archery 132. Bows^ 163. Arrows 194. Horses and Cavalry Equipment 235. Protective Clothing^296. Other Equipment 31Chapter Two^A Military Analysis of Horse Archery^33Chapter Three^Questions of Armament^ 43Chapter Four^Operational Employment 581. Advance to Contact^622. Attack^ 763. Pursuit 814. Defence 855. Withdrawal 866. Summary^ 89Chapter Five^Conclusions 92Appendix One^Bow-Armed Units^ 95Appendix Two^Analysis of Arrow Fire with Referenceto the Theory of Small Arms Fire^102List of Figures 109Figures^ 112Glossary 124Bibliography^ 127ivACKNOWLEDGEMENTSI would first of all like to thank my thesis advisor,Professor James Russell, for his support and advice, and for hiscriticism of the earlier forms of this work, without which it wouldindeed be vastly inferior.I also owe a debt of gratitude to Raymond Gabin, who kindlyconsented to apply his military knowledge to review some of thechapters; and especially to my fiancee, Christine Parker, not onlyfor her advice and suggestions, but also for her unflagging supportwhile I was researching and writing. All errors and omissions areof course mine alone.Of course, one needs not only scholarly assistance to write apaper of this length. I would like to thank in particular myCommanding Officer, Lieutenant Colonel Roderick Bell-Irving, forhis indulgence and understanding in allowing me a reduced workschedule for research and writing; and my fellow subalterns of theSeaforth Highlanders of Canada, for their support and constantcomments about "how the Roman archers would have done it". Thisthesis is dedicated to them. Ducimus.liINTRODUCTIONIt is not known who the first people were to combinehorsemanship and archery; probably the combination is as old asboth skills, and preceded such technological innovations as thechariotl. We first encounter the bow in the hands of a cavalryman,however, in the near East in the first millenium before Christ, inthe armies of the Scythians and Assyrians2. Although mainly knownfor its infantry phalanx, fifth-century Athens maintained a forceof horsemen, among whom were a number of mounted archers3. Thismay have been partly as a counter against the cavalry of thePersians, but already it is clear that these troops were becomingmore important. The high point in the history of the mountedarcher came with the destruction of Crassus' army at Carrhae in 53B.C. by a Parthian army composed mostly of lightly-armoured, bow-1The prevailing view is that chariots preceded cavalry, andthis is supported by Assyrian and Egyptian reliefs (see Y. Yadin,The Art of Warfare in Biblical Lands [Jerusalem 19631 4-6); but itis hard to believe that this technological innovation precedes anon-technological one. Some of the first known mounted archerswere nomads, and these peoples produced little in the way ofpermanent artwork that would record their methods of hunting orfighting. That the skills of mounted archery could be acquired ina relatively short time can be demonstrated by the fact that thePlains Lakota (Sioux) Indians of North America gained them soonafter the horse was introduced to them: see E. McEwen et al. "EarlyBow Design and Construction" Scientific American, June 1991. 76-82.2Seen on a stone relief found in the palace of Ashurnasipal IIat Nimrud: Yadin (op. cit.) 297, 384-5; G. Denison, A History of Cavalry (London 1913) 10-11.3G.R. Bugh, The Horsemen of Athens (Princeton 1988) 221-4; andXenophon, Memorabilia III, 3, 1.armed cavalry.Although by the late Republic the Romans had experienced firsthand the effectiveness of mounted archers, there is no evidence ofthese troops in a Roman military force until the time of Caesarwhen they formed part of a contingent sent to Caesar's rival Pompeyby Antiochus I of Commagene4. The equites sagittarii first maketheir appearance as regular auxiliaries in Germanicus's campaignagainst the Chatti5. From this point on, references to them becomemore frequent in the epigraphical and literary sources as they cameto assume a greater role in the Roman military of the Empire.The nomenclature assigned to the alae and cohortessagittariorum gives some evidence for the history of these units.The names show, for example, that they were raised mostly in theeast, especially in the provinces of Syria and Thrace°. Obviousexamples of these are the cohortes Commagenorum, Antiochensium,Damascenorum, and Thracum. Some come from areas outside theboundaries of the early Empire. Parthia, for example, wasobviously the origin of some of the mounted archers in the alaParthorum et Araborum. The cohors XX Palmyrenorum milliariaequitata sagittariorum seems to have been formed in the late secondcentury A.D., about the time Palmyra became part of the RomanEmpire, from the numerus Palmyrenorum that is known to have servedin Dacia in the early part of the same century. Unit titles, where4Bellum Civile III, 4, 5.5Tacitus, Annales II, 16.°See Appendix one for a full listing of known bow-armed units.2they appear, may supply an approximate terminus post quem for theemployment of the bow-armed units to which they were granted. Anumber of bow-armed units, for example, bear the Imperial titlesAugusta, Flavia, or Ulpia. These titles seem to have been assignedwhen the units were raised or reorganized7. Two such units werethe cohortes I Flavia Damascenorumafter the revolt of Civilispossible to assign preciseand I Flavia Canathenorum,in 69-70 A.D.8. It is notdates to the formation orformedalwaysdissolution of each unit.There is evidence for the existence of a total of twenty fourmounted, bow-armed auxiliary units in the Roman army from the firstto the third centuries A.D.9. As we have seen, they were raisedmainly in the east and they appear to have served mainly in theeast as well. At least three units composed of mounted archersformed part of the garrison of Dacia in the second century, as wellas at least seven units of foot archers. Pannonia had a similarnumber of horse archer units among its garrison, as did the Africanprovinces of Mauretania Tingetana, Numidia, and Egypt. At leastfour mounted archer units are known among the garrison of Syria,and from the late second century the cohors XX Palmyrenorum was7Aks far as can be determined, there is no evidence of theawarding of such titles to bow-armed units as a reward forperformance in battle. See P.A. Holder, Studies in the Auxilia of the Roman army from Augustus to Trajan. BAR International Series#70 (Oxford 1980) 14-18.8Holder (op. cit.) 16.9At least four other mounted units may have been sagittaria aswell. See Appendix one.3stationed at Dura-Europos. Only one unit of bow-armed cavalry isknown to have served in Germany: the cohors I Flavia Damascenorummilliaria equitata. While a cohors Ramiorum sagittariorum is knownto have served in Britain, no alae or cohortes sagittariorum areknown to have been raised in or to have served in Gaul, Hispania,or Italy.But what was their role? What was it about them that causedtheir commanders to use them as they did, and what made them soeffective on the battlefield? How were they used? The purpose ofthis thesis is to attempt an answer to these questions, applyingwhere relevant the techniques of military analysis and anunderstanding of military affairs based on experience.SOURCES The sources for the employment of horse archers in the Romanmilitary are primarily literary, though there are also considerableamounts of archaeological, papyrological, and epigraphic evidencefor their equipment, distribution, history, and composition.The historians of the Roman imperial period furnish thegreatest amount of information on horse archers. Many details oftheir employment, especially as a screen and guard for the army onthe march, can be gleaned from a careful reading of Josephusm.Tacitusll gives hints of their actions in attack and pursuit.MEspecially Bellum Iudaicum II, 500-1; III, 66-9; and V, 47-9."Ann. II, 16-74Arrian is an especially important source as he was in command of anarmy in Cappadocia in 135 which included horse archers12; he wasalso a historian, and his account of Alexander the Great's army inIndia13 is valuable as comparative evidence for the role of themounted archers in an ancient army. Ammianus Marcellinus had firsthand experience of military affairs, so that his history mightreasonably be expected to contain useful information on mountedarchers. He does refer to them as a "formidable arm of theservicel", but unfortunately he makes few other references to thistype of cavalry and these are too vague to be of much value.Finally, Procopius in his Bellum Gothicum frequently describes theRoman battle line of the sixth century which was mainly composed ofmounted archers. Although war in his day differed greatly from thehigh Empire, being by that time largely a contest between cavalryformations, he occasionally mentions small forces of horse archersoperating as independent units apart from the main force. Sincethis seems also to have been the function of the horse archers inthe high Empire, some indications of their employment in that rolemay be garnered from Procopius's account of them in his day.Unfortunately the writers of the Greek tactical manuals,Aeneas Tacticus (mid fourth century B.C.), Asclepiodotus (earlyfirst century B.C.), and Onasander (mid first century A.D.), are"Eirccc4; Komi AAAivow13'Avdeficcatc 'AXE4dcv8pou , e specially  V, 14-8.un ...formidabile genus armorum..." Res Gestae XVI, 12, 7.5silent on mounted archery in war. The latter two do contain someindication of ancient thinking on cavalry, albeit from an earliertime, but these references are nevertheless of little use. The fewreferences to archers and cavalry that are found in theStrategemata of Frontinus and Polyaenus likewise do little toillustrate the employment of mounted archers but do give someindication of the use of cavalry. Vegetius15 (late fourth centuryA.D.) does not, unfortunately, mention mounted bowmen. On theother hand, he is of considerable value for his many references toarchery, the training of soldiers, and the use of cavalry in Romantimes. Finally the Strategikon, attributed to the Byzantineemperor Maurice (late sixth or early seventh century A.D.),although it describes, like Procopius, a type of battle muchdifferent from that of the high Empire, contains16 importantreferences to the training and employment of bow-armed infantry andcavalry which can serve as indications of earlier practise.Most modern scholarship on the subject of the equitessagittarii has been confined to consideration of them as archers oras cavalry, not as a distinct arm of the service. Another body ofscholarship has been concerned with archery equipment and thehistory of bow-armed units. These topics are important for anydiscussion of mounted archers, but comparatively little attentionhas been paid to details of employment and their role on thebattlefield of the high Roman Empire. The most comprehensive work15De Re Militari, especially book I.MEspecially in book XII.6on the archer in Roman service is that of J.C. Coulston17. Whilehe includes much discussion on the subject of archers, Coulstonconcentrates mainly on technical details of the evidence for theconstruction of archery equipment and the physics behind theoperation of bows18 and does not discuss questions of employmentwith any great sophistication. Though dealing with a periodconsiderably later than ours, A.D.H. Bivar's19 study of the easternfrontier of the Byzantine empire and the equipment and tactics ofboth the Byzantines and their enemies provides comparative evidenceof the use of archer-cavalry in the earlier Empire. A number ofother writers28 have discussed archers in some detail; they have,however, not been concerned with the technical details of thedeployment of horse archers themselves on the battlefield.The main epigraphic evidence for horse archers in the Romanperiod is to be found in tombstones, which give indications ofwhich units were bow-armed as well as details of their equipment.The tombstones having the most value for a study of horse archers""Roman Archery Equipment" in M.C. Bishop, ed. The Productionand Distribution of Roman Military Equipment. BAR InternationalSeries #275 (Oxford 1985) 220-336.18 W.F. Paterson's article ("The Archers of Islam" Journal ofthe Social and Economic History of the Orient 9 (1966) 69-87)provides similar detail on later Persian and Arabic archery.19"Cavalry equipment and Tactics on the Euphrates Frontier"Dumbarton Oaks Papers 26 (1976) 272-291.20E. g. P. Medinger "L'arc Turquois et les archers Parthes a labataille de Carrhes" Revue Archeologique 6, II (1933) 227-234; D.B.Saddington "The Roman Auxiliaries in Tacitus, Josephus, and OtherEarly Imperial Writers" Acta Classica XIII (1970) 89-124; and H.van de Weerd and P. Lambrechts "Note sur les corps d'archers auhaut Empire" Die Araber in der Alten Welt I (1964) 661-667.7are reproduced here (figs. 18-21). Military diplomas locate bow-armed units and indicate when they were present at variouslocations and in combination with what other types of militaryunits. Important details of day-to-day activities and equipmentare found in the papyri and graffiti from Dura-Europos. Anespecially well-documented unit of mounted archers, the cohors XXPalmyrenorum milliaria sagittariorum equitata, was stationed thereafter the late second century.Columnar sculpture provides some evidence for archery andcavalry equipment. Unfortunately, Trajan's column is of littlehelp for a study of mounted archers. A careful examination of thiscolumn did not reveal any mounted archers on it, and the peditessagittarii that do appear on it are armed in an improbable andinconsistent manner21 (figs. 14, 16, 23-4). Examination of theMarcus column is only marginally more useful, as it shows at leastone bow-armed cavalryman (fig. 22). Unfortunately, his equipmentis damaged, and the foot-archers on the same column (fig. 25) areshown armed in the same improbable manner as those on Trajan'scolumn22.The equipment and weaponry used by the equites sagittarii, andindeed most auxiliaries in the service of Rome, are reasonably wellunderstood. How they dressed, armed and protected themselves, and21See Florescu, F.B. Die Trajanssdule (Bonn 1969) passim, andCoulston (op. cit.) 235. The bows are depicted with curled ears,which is a feature totally unknown from archaeological or othersculptural evidence.22See G. Becatti, Colonna di Marco Aurelio (Milan 1957) passim,and Coulston (op. cit.) 235.8equipped and mounted their horses, has generated a considerablescholarship based on archaeological finds, sculpturalrepresentations, the occasional literary reference, and exhaustiveattempts at reconstructions. A chapter is included brieflyreviewing the evidence on his bow and his mount, the two componentsthat, along with himself, made the eques sagittarius so effectiveas a weapon system, to employ the terminology of current militarytheory.Analysis of any military system, however, requires not justanalysis of technology but also analysis of its use: in order tounderstand how it "ticks" we must look at what actually affects howit operates. We must, in short, look at its characteristics.Horse archers can be fast, and they need not be within arms' reachof their enemy to kill him. This much is obvious, but much morecan be understood from a closer examination. Chapter two discussesin detail the characteristics of the mounted archer: not of theRoman version per se but of any individual armed with a bow andriding a horse.During the research for this paper I became aware of a gap inknowledge that was being continually ignored or glossed over in theliterature. Coulston23 partially recognized the problem andwondered how the mounted members of the cohortes equitataesagittariorum were armed. One of the difficulties lies in theabsence of any depictions showing them as bow-armed. This is onlypart of the problem, however. No force equipped inflexibly can230p. cit., 284-5.9operate with flexibility. The cohorts of mounted archers thatgarrisoned Rome's eastern provinces had a multitude of fighting, asopposed to engineering, tasks to fulfil, including border patrol,policing, and convoy and dignitary protection 24, as well asdefence. To limit their armament to bows would be to endanger thesoldiers and hinder them in the performance of their tasks. Sucha limitation would also be counterproductive to the original andprincipal raison d'être for the cohors equitata, namely economywith flexibility 25 . Chapter three is devoted to a closer look atthe evidence of armament.Chapter four forms the central focus of the thesis. In thischapter a number of ancient battles are examined to determine howthe Romans employed their horse archers in battle. It is not myintention to provide an exhaustive survey of all appearances ofmounted archers in the literature of the Roman period, but to coveronly those where the evidence is sufficient and reliable enough togive a clear picture of events, and where enough detail is suppliedto reach general conclusions of a comprehensive nature concerningtheir role in battle as a whole and in support of other arms, andnot just their deployment in a single action or manoeuvre.According to current military theory, battles are composed of anumber of phases. These phases, as well as the characteristics of24Border patrol, policing, and convoy and dignitary protectionare fighting tasks in that those engaged in them must always beprepared to fight even if battle never occurs.25R.W. Davies "Cohortes Equitatae" in Service in the Roman Army(New York 1989) 140-51.10the equites sagittarii discussed in chapter three,^provide aframework for analyzing their role on the Roman battlefield of thehigh Empire.No work on Roman archers would be complete without a list ofthe units they served in. The most complete list to date is thatof Eric Birley, published in 1977 by Jeffrey Davies 26 . It isreproduced here, with some modifications, in the first appendix,along with a brief commentary. The second appendix contains ananalysis of the "fire" 27 of archers in terms of modern militarypractice. This exercise may explain certain puzzling aspects ofthe employment of archers, both mounted and on foot, by quantifyingwhat would be obvious to one who has been in the position ofcontrolling or creating fire, but would be only dimly perceived byone unaccustomed to these experiences. A glossary is also suppliedto explain the many technical terms employed in the discussion:not only military terms, but especially those concerning the bow.The bow is a complex instrument with many components, some withArabic names that would be unfamiliar to the non-specialist.I have six years' experience in military matters includingcomprehensive training as a regular infantry officer in theCanadian Forces. My training has been in the command of bothdismounted and mechanized infantry forces as well as in the moretheoretical aspects of war. In addition, I have commanded an26"Roman Arrowheads from Dinorben and the Sagittarii of theRoman Army." Britannia 8 (1977) 269-70.27For the meaning of "fire" in this sense, see Glossary.11infantry platoon for one year. On the basis of this experience, Ihave applied my personal knowledge to the subject of mountedarchers in order to provide a new approach to their study. I willbe successful if I have provided some contribution to themonumental scholarship on which I base my work.12CHAPTER ONEWEAPONS AND EQUIPMENTHorse archery needs certain basic equipment: the bow, thearrow, the horse and its associated gear, protective clothing forthe archer, and related items such as the quiver and the bow-case.This chapter will cover the essential elements of horse archery andthe equipment used, in order to give a clear picture of how thesesoldiers were equipped and what they were capable of.1. ARCHERYThe mechanics of shooting an arrow consist of three actions:drawing the string, holding the string back and sighting the shot,and loosing the arrow. This section will briefly cover theseactions as well as the question of range and penetration (see figs.1 and 3).Known alternately as a "hold" and a "release", the way inwhich the archer holds the arrow on the string is clearlyassociated with the drawing of the string. The most common holdtoday is called "Mediterranean". In this hold the arrow is placedalong the left side of the bow resting on the side of the lefthand, the string is fitted into the nock, and the base of the arrowis gripped between the tips of the index and middle fingers of theright hand. The tips of the right hand index, middle, and ringfingers hook over the string. This appears to be the only hold13used by archers in the service of Rome28. When an arrow is firedusing this hold, the fletchings of the arrow brush down the leftarm. This hold thus precludes the use of a shield as it will foulor damage the fletchings. A smooth band known as a bracer is oftenworn on the left forearm to protect it from abrasion by the arrowas it moves-past29.Another common hold is the "Mongolian", where the arrow isplaced along the right hand side of the stele and both the arrowand string are held between the right thumb and the side of theindex finger. As the arrow is on the right-hand side of the stele,it does not touch the left forearm, and so a small shield can beworn strapped to the left forearm38. The Mongolian hold requiresthe use of a thumb-ring, usually made of bone, to protect the thumbfrom abrasion; these rings have not been identified before theByzantine period, and so the Mongolian hold was probably not usedby Roman auxiliaries of the high empire31.When the arrow has been placed on the string, the archer facesthe target and raises the bow with his left hand, keeping the bowvertical; at the same time he pulls the string back with his right28J•C• Coulston, "Roman Archery Equipment" The Production andDistribution of Roman Military Equipment. BAR International Series275 (Oxford 1985) 278.29Ibid., 275-6.mW.F. Paterson, "Shooting Under a Shield" Journal of the Society of Archer-Antiquaries 112 (1969) 27-8.31See below (note 29).14hand to his right cheek or ear32, and aims at the target, sightingabove or below the arrowhead, depending on the range. It is bestto hold this position for as little time as possible to reducefatigue and shaking.To loose the arrow the archer must ensure that the stringslides smoothly off his fingers without catching on clothing orequipment, and that the arrow is also unobstructed. At targetpractise he would "follow-through" by holding his firing positionuntil the arrow hit the target, but in battle he would not wastetime between shots.The range of accuracy of the ancient bow has been estimated33to be in the order of 150 m for a stationary firer and target; itwould be less if the firer or the target were moving. Since thebow used by a mounted archer would most likely have been smaller,its range would have been reduced34. More precise figures are notpossible owing to variations in individual skill and in the qualityof equipment produced.Finally, perhaps because of a scarcity of reconstructions ofRoman archery equipment, precise details of armour-piercing32Procopius (Bellum Gothicum I, 1, 15) mentions that archersof his day drew the string to the ear; he contrasts this with theHomeric (cf. Ii. iv, 123) practise of drawing the string to thechest, which would result in inaccurate aiming.33I4.E. McLeod "The Range of the Ancient Bow" Phoenix 19 (1965)1-15; and idem, "The Range of the Ancient Bow: Addenda" Phoenix 26(1976) 78-82. Vegetius mentions a target range (de re militari I,15) of six hundred Roman feet.34Coulston (op. cit.) 291; and see below, section 4 on cavalryequipment.15capability of this equipment are not known. Parthian horse archerswere able to penetrate armour with their arrows at Carrhaem.Assuming that they did not come far within the maximum range of theRoman legionaries' pila and their auxiliaries' bows, the Parthiansprobably came no closer than 30-50 T1136 from the Roman line.Because of the smaller bow used on horseback and the difficulty ofaiming while galloping it seems best to assume a distance of 30-50m as the maximum for effective penetration of enemy armour fromhorseback, and perhaps double that range for striking unarmouredtargets37.2. BOWSThere were two main types of bow used in the ancient world:the self-bow and the compound bowm. The self-bow is made of asingle piece of shaped wood (the stave), more or less straight,tightly strung with a string of leather or sinew. This form ofmPlutarch, Crassus XXV, 6mthese figures are based on an arbitrary figure for the rangeof the Roman auxiliaries' bows of 80-100 m and the fact that theywere behind the lines of legionaries.37For a more complete discussion of the application of fire todifferent types of targets, see Appendix two.mthe main works on the subject are: Coulston (op. cit.) 220-366 and especially 222-259; and W.F. Paterson "The Archers ofIslam" in The Journal of the Economic and Social history of the Orient 9. (1966) 69-87. See fig. 3 for a depiction of the parts ofthe bow.16bow, the mainstay of the English battle line at Agincourt39, wasused in Crete and Gaul since well before Roman times" and ischaracterised by slower initial arrow velocity (and thus range andpower) for a given draw weight°. Since the eastern auxiliariesthat are the subject of this study did not use the self-bow, littlemore mention will be made of it here.The compound bow (fig. 3) is a development from the self-bow,making use of horn and sinew as compound levers and springs toincrease the power imparted to the arrow on release."The thin wooden core provides adhesivestrength and the general shape but plays aminor part in the bow's physical actions.When a stave is drawn the horn belly is pulledinto a compressed curve and the back sinew isstretched. The bow is constructed to elicitthe maximum distance of curve and stretch by39P. Warner, Firepower (London 1988) 33."that Cretan mercenaries were using self bows may be impliedfrom Xenophon (Anabasis III: 3, 7 and 4, 17): the first passagedescribes the Cretans' range as less than that of the Persians'(...a yap EArec Naximpa T8V IIepaoiv kge)om...); while the secondimplies that their bows were of the same size (...geyeamaimitgaviHemmitaw...). Since it is clear that the compound bow was usedby eastern peoples including Persians (Burke, "A History ofArchery"), a significantly shorter range indicates the use of theconsiderably less powerful self-bow.°For a detailed discussion of the physics behind self- andcompound bows, see Coulston (op. cit.) 245-8, P.E. Klopsteg, "ThePhysics of Bows and Arrows" The American Journal of Physics 11(1943) 175-92, and W.F. Paterson, "The Archers of Islam" TheJournal of the Economic and Social History of the Orient 9 (1966)69-87.17making it in a "reflexed" shape so that itreverses itself when unstrung.^The extradistance from the reversed to strung positionsgives a greater potential energy storage thanwith a straight self-bow stave."42The lever action of the ears increases the pull on the stringby differing amounts: most when the string is slightly pulledback, and progressively less as the draw length increases. Thiseffect causes the firer to feel as if the "...pull on the bowstringreduces, though this is not, in fact, the case..."43. As energyis stored in the bow itself, and in the dustars and ears inparticular, and not in the arm and shoulder muscles of the firer,this bow is less tiring to draw", and so greater power can beattained for a given amount of exertion.Upon release the levers tend to accelerate the arrow more orless smoothly until it leaves the string. This acceleration is amore efficient transfer of power to an arrow than the steadilydecreasing hard push of a self-bow45, and thus greater power istransferred to the arrow on firing.42Coulston (op. cit.) 245.43Paterson (op. cit.) 79."Coulston (op. cit.) 247.45See Klopsteg (op. cit.).^Self-bows must have the stringmoving at very low speeds at the moment the arrow leaves it inorder to reduce the "virtual mass" of the limbs; a compound bowachieves this by having the limbs, in effect, "unroll", and so theyare moving very slowly (and the string comparatively quickly) atthe arrow's final contact with the string.18The ears of the compound bow can give a further useful effect.The archer's tombstone found at Housesteads (fig. 12), on theassumption that this is an accurate representation46, shows theears of the bow angled forward although the bow is at rest; such adesign causes the string, when the arrow has been loosed, to pressagainst the inside of the ears close to the dustars. The string isthus effectively shortened, which imparts a final acceleration onthe arrow.Bowstrings in the Roman era were most probably made from sinewor leather, but because of their organic and fragile nature nonesurvives, and we are limited to conjecture their actual nature.Whatever material they were made from, however, they had to be keptdry in order to prevent stretching47.3. ARROWSArrows are the most variable piece of archery equipment. Thelength of the arrow varies according to the size of the archer andhis draw length; the materials from which it is made depend uponthe type of bow used and the intended target. As is the case withbows, I will attempt to give as complete a description as possible,with emphasis on what is relevant to my study of horse archers48."As Coulston (op. cit., 236-7) believes.47See Coulston, op. cit., 270.lAs with the bow, Coulston (op. cit., 264-70) discusses thesubject in great detail; the reader is referred to his article fora more extensive treatment. See figs. 4-7 for roman arrows and19An arrow is composed of an arrowhead; a shaft or stele; andvanes or fletchings (fig. 4).Arrowheads are usually made from metal, iron being the mostcommonly used metal for war arrows because of its hardness andtoughness. Flight arrows, used in distance shooting competitions,sometimes had ivory tips49. Such tips were not designed forpenetration but only for aerodynamic qualities as they would haveshattered upon striking anything tougher than skin or dirt".Military units would have had a supply of properly made arrowheadson hand. When these ran out, or if there was not sufficient timeto prepare the metal properly, craftsmen would be forced to usesubstandard materials or techniques".Arrows had various cross-sections. Most surviving examplesare vaned, the most common would be trilobate (fig. 6), althoughoccasionally ones with four vanes are attested. Vaned arrowheadswould often be barbed. Some are flat-bladed (fig. 5); these areusually not barbed. Arrowheads with a square cross-section arearrowheads.49See Paterson (op. cit.) plate IV."Apparently the Huns used bone tips for their war arrows; uponhitting an unarmoured enemy these could splinter and causeconsiderable damage (Amm. Marc. XXXI, 2 and see Coulston [op. cit.]268)."The arrowheads of defenders of Masada seem to have beenhastily fabricated. One was analyzed and found to be made of lowquality iron without tempering; Coulston (op. cit., 270), citingKnox et al., "Iron Objects from Masada: Metallurgical Studies"Israel Exploration Journal 33 (1984) 99-100.20known as bodkin52. Finally, there are fire-arrow heads, consistingof a pointed cage of three curved bars, into which flammablematerial would be inserted, then lit (fig. 7).Obviously different forms of arrowhead have differentfunctions. The heavier the arrow, the greater momentum it willcarry, and thus greater penetration. Flat-bladed arrowheadscontain less material and so are lighter. Their shape also lendsthem to easier sharpening. They are therefore more suitable forsoft targets (animals and infantry or cavalry unprotected byarmour) than the trilobate versions, which are considerably heavierand tougher, owing to their self-reinforcing shape. Trilobatearrowheads also suggest a more sophisticated manufacturing processthan the flat examples (which are often simply hammered out) and socould be examples of prepared stocks rather than emergencysupplies. Little survives from antiquity regarding recommendationsfor use of different arrowhead designs".There were two methods of attaching arrowheads to shafts:sockets and tangs (figs. 4 and 7). The first is self-explanatory,except that either glue or a pin would hold the parts together; thesecond depended upon a hollow stele or one made hollow by drilling,and the arrowhead was pushed or hammered in and held in place by52These appear to have been used for target practise on an ox-skull found in Northumberland with several small, square holespunched in it (Coulston [op. cit.] 265). This is apparently theonly extant example of a practice target from antiquity."The Ghunyah of Taybughah, a medieval Arabic archery manual,apparently contains such recommendations for different equipment,but unfortunately it is unavailable to me at present.21friction.The shaft or stele was made from a variety of materials. Inthe Roman west, wood dowels were commonly used; in the east, hollowreed or cane was the norm. Sometimes a composite construction isfound, where the arrowhead is joined to a wood dowel, which is inturn attached to a reed shaft; Coulston believes that this was toprevent the more fragile reed from splitting upon impact with anarmoured target54 which would lessen the arrow's momentum andpenetration. At the base of the stele is the notch or nock, whichmust be wide and deep enough for the string to fit properly55 (fig.8). At the base of the stele near the nock the vanes or fletchingsare attached. The purpose of fletchings is to give the arrowstability in flight and therefore accuracy and penetrative power.All evidence from antiquity indicates that these were made frombird feathers56. They were glued in place, and sometimes furtherfastened with a whipping of sinew front and back. The exactlocation of the fletchings on surviving examples is important forreconstruction of shooting methods. It has been claimed57, forexample, that fletchings extending right back to the nock make the540p. cit., 268.55An example found at Dura-Europos has a nock depth of 0.95 cmand a width of 0.4 cm at its deepest. See S. James "Dura-Europosand the Introduction of the "Mongolian Release"" Roman MilitaryEquipment: the Accoutrements of War. BAR International Series#336 (Oxford 1987) 77-84.56The surviving arrows from Dura-Europos have featherfletchings. See James (op. cit.) 78, and figs. 3 and 4.57James (op. cit.) 78-81.22"Mediterranean release" impossible58.As different firers have arms of different lengths, their drawlengths will differ correspondingly. Arrows must therefore be madein a length appropriate for each firer. This suggests that eacharcher in an archer unit will have his own personal supply in theform of a quiverful or two of arrows sized for him, ready at hisside in battle. Once these run out, however, he would be reducedto using stock issue arrows of uniform length, or recovered enemyarrows, until he could have more made for him59.4. HORSES AND CAVALRY EQUIPMENTThe subject of the Roman cavalry horse is extremely complex,58James (op. cit., 78) is probably correct that the arrowsfound at Dura could not be shot using the Mediterranean release,i.e. gripping the stele with the index and middle finger at thestring and thus crushing the fletchings; but James fails toconsider the Sassanid release, wherein the arrow is held betweenthe tip of the middle finger and the side of the index finger,which is extended straight along the arrow between the fletchings(see W.F. Paterson, "The Sassanids" Journal of the Society of Archer-Antiquaries 112 (1969) 30; and my fig. 1). This grip wouldavoid crushing the feathers (see especially my fig. 2 for adepiction of an arrow being drawn with the fletchings extending allthe way back to the base of the fingers). The thumb ring alsofound at Dura is indeed interesting but, since little is known ofits context, it is difficult to conclude with James that the"Mongolian release" was used at Dura-Europos during the Romanperiod.59Here as throughout this chapter I follow Coulston (op. cit.,270), but he does not provide evidence to support this statement.A useful analogy is that the butt lengths of modern Canadianinfantry weapons can be adjusted as appropriate for each firer. Ifa soldier loses his rifle in battle he must be prepared to make dowith one of the wrong size until it can be adjusted for his armlength.23and only certain topics are of importance to a study of theemployment of mounted archers60. This section will attempt tocover these topics in enough detail to give the reader a clearpicture of how they affect archers. As for cavalry equipment, Iwill restrict discussion to the saddle and some bridle equipment,as these are the main items of significance for this study.The horse of Roman times was somewhat smaller than his moderncounterpart. Hyland's examination of military horse bones from theRoman eram indicates an average height of approximately 150 cm62,which is average for modern horses; the Romans, however, preferredthe largest horses available to them for war purposes, and many ofthe breeds preferred for civilian employment were much smaller.Robust horses were preferred by Roman cavalrymen as they wereeasier to ride for a rider whose saddle did not have stirrups. Thereason for this was that a a horse of robust appearance usually hada wider back, and frequently also displayed the the phenomenonknown as a "double back" characterized by ridges of muscle alongeither side of the spine. This would have ensured both a morecomfortable ride and made it easier for a rider to maintain hisbalance.The build of the horse has another effect: the "...well to60The definitive work on the subject is that of A. Hyland(Equus: the Horse in the Roman World. [London 1990]), and much ofthe information in this section comes from this source.MOp. cit., 68.621 use metric measurements rather than the technically correctterm "hands" used in measuring horses in order not to introduce yetanother definition to confuse the reader.24moderately fleshed animal usually has a good food conversion ratiowith a slower metabolism than the leaner types..."63. In otherwords, it can go longer with less feeding than a slim horse. Itis, however, "...usually not such an enduring animal as the lean,racy type of equine.. . "64. This suggests that cavalry mounted onsmaller horses would have to take greater care to ensure adequateforage for their mounts but could operate over longer distances andin tougher terrain than their heavier-mounted comrades.The disposition of the mount is also of great concern.According to Hyland, "...Arabians are sometimes too clever andindividualistic, more suited to...Oriental warfare...in a looserformation. A more phlegmatic animal is more suited to close-orderaction. ..."65. Hyland suggests as well that Oriental units wouldbe less apt to use mares than stallions, and that stallions wouldbe left uncastrated66. A list of cavalrymen and their horses ofthe cohors XX Palmyrenorum from Dura-Europos67, however, includes°Ibid., 69."Ibid., 69.°Ibid., 79-80.^This does not mean that each man was anindividual, off fighting on his own somewhere; obviously in orderto produce an effect on the enemy there must be concentration offire. The horses had to be close enough to enable their riders toproduce such an effect but leave enough space for them to be ableto pick targets and fire without hitting their comrades in theback.66Because of related cultural attitudes men in middle easternlands today do not like to ride mares. p. 81.67PDur. 97; see R.O. Fink, "Roman Military Records on Papyrus".(Michigan 1971) 340 - 4; and R.W. Davies, "The Supply of Animals tothe Roman Army and the Remount System" Service in the Roman Army(New York 1989) 157-9.25mares (equum quadrimam or probatam). The other horses are male(equum quadrimum or probatum), but it is unclear whether thisrefers to stallions or geldings, or indeed whether any distinctionwas made.We turn now to cavalry equipment. As Coulston says, "...itmight be expected that eastern archers exhibited oriental featuresin their clothing and armour. On the contrary the tombstones ofauxiliary sagittarii suggest that they differed not at all fromother auxiliarii except in the carrying of archeryLittle can be discerned from these tombstonesabout the saddles they rode on but the other parts of the horse'stack or equipment seem to be identical to that employed by otherauxiliaries; it therefore seems most likely that the saddles werethe same as well. Such standardisation would make sense in theRoman army as myriad types of equipment only burden the supplysystem unnecessarily. The only item that seems not to have beenstandard was the bit.We shall first examine the Roman military saddle. Contrary topopular belief, imperial Roman cavalry saddled their horses. Thesaddle used has been reconstructed69 (figs. 9-11) from variouspieces of evidence and appears to have given the rider aconsiderable amount of support with its four inward-curving"horns". This support may have been sufficient to ensure that613Coulston (op. cit.) 278.69See P. Connolly, "The Roman Saddle" Roman Military Equipment: the Accoutrements of War. BAR International Series 336 (Oxford1987) 7 - 27; and Hyland (op. cit.) 130-44.26stirrups would in most situations have proved less of an advantagethan is usually believed70. Curving over the hips and thighs fromfront and back, they kept the rider in his seat, with assistancefrom the rider's thighs, during all types of manoeuvres includingturns, movement uphill and downhill, and close quarter battlen.The traditional assumption has been that the size of the bowwas limited because the rider had little support from his saddle.This meant "lack of stability", on the assumption that the drawingof a large bow (and therefore one with a large draw weight) wouldunbalance the rider because of his lack of footing72. There aretwo objections to this assumption. First, it is now clear that thecavalryman had very good support indeed from his saddle. Second,drawing a bow, regardless of its weight, does not unbalance thearcher; no actual weight is produced, and almost none is shifted.A large bow, however, is more cumbersome. It is more difficult toshift from a target on one side to one on the otherm. A large bowalso requires a correspondingly longer bow-case, which was carriedvertically on the side of the horse to the immediate rear of thesaddle. If it were too long, it would be an encumbrance, perhapseven hitting the ground or getting entangled with the horse's legsmCoulston (op. cit.) 292.nConnolly (op. cit.) 12, 16 - 7; and Hyland (op. cit.) 130-4.72Paterson (op. cit.) 85.The archer has to lift it up and over the horse's neck;clearly, if the target is in motion across the horse's front, thearcher must take his aim off it for a time. Also in a situationwhere an enemy is almost directly in line with the horse's frontthe archer will have difficulty directing fire upon him.27during some manoeuvres. Thus it seems that a shorter bow would bedesirable; however, Coulston claims that "...it is not always thecase that bows used on horseback were short...and the trend incentral Asia was for an increase in length over time..."74.Perhaps the later horsemen developed techniques and technology tocounteract the problems of a long bow. The question, however,remains open.The final piece of cavalry equipment to be covered briefly isthe bit. Usually the rider conveys commands to the horse throughthree channels: the action of the bit and the reins, kneepressure, and voice. The first and last pose problems for amounted archer, however. In order to use his bow he must drop thereins, and in the din of battle auditory commands would be oflimited use. The bow-equipped rider must therefore give allcommands to the horse through knee pressure and movement. When notin battle, however, he may elect to put away his bow, and take upthe reins for easier control of his horse. The bit used by theRomans has been tested on horses by Hyland 75 and she describes itas "punishing" and "crue1"76. It seems highly unlikely thathorsemen who have reached such a level of control over their horsesthat they do not need reins in battle would resort to the use ofsuch an extreme tool when away from action. There is, however, no74Coulston (op. cit.) 246. He does not, unfortunately, backthis statement up.750p. cit., 136-40 and plates 9-14.760p. cit., 138-40.28evidence for other, more gentle bits.To conclude on the horse used by the Oriental horse archer ofthe Roman army: it was small by today's standards, and evensomewhat smaller than its comrades in contemporary cavalry units.In temperament it would probably have been spirited and even"volatile"77. For best effect it would probably have had a wideback; it could have been either male or female, and there is noevidence for the practise of castration. It was saddled with thehorned cavalry saddle that gave its rider considerable support, andit appears doubtful that the severe bit of the Roman cavalry wasused. So much for the animal; now we turn to the rider.5. PROTECTIVE CLOTHINGUnfortunately, there are few sculptural representations ofRoman mounted bowmen from the high Empire. Those that do exist,however, tend to show them "...dressed and armoured in the samemanner as other auxiliary units..."78. Funerary representationsof equites sagittarii show them wearing tunics and either short orlong trousers (figs. 18-22), and only the tombstone of the alaParthorum et Araborum (fig. 18) seems to show one wearing ahelmet79. The horse archer on the Marcus column80 (fig. 22),"Hyland (op. cit.) 81.78Coulston (op. cit.) 280.79This tombstone is badly damaged, however, rendering detailsunclear.29however, is clearly depicted wearing lorica hamata. His head, onthe other hand, is damaged and it is impossible to tell whether heis wearing a helmet, but, as his clothing otherwise resembles thatof other auxiliaries on the column, it seems reasonable to includethe helmet among his armour. No Roman representations depict anarcher's horse with protective armour81.There exist a number of representations of Roman foot archers(figs. 14-16, 23-25). These seem to show much the same clothing asthe depictions of horse archers. All but one of the foot archerson Trajan's column "...appear in the well-known ankle-length robes,with conical "spangenhelme" and either loricae squameae orhamatae..."82. These may indeed not be Roman or Middle Easternsoldiers at all, but Sarmatian83. The other archer is dressed likeother auxiliaries&. All other representations of foot-archersshow them without armour or at most a helmet°.mCichorius's scene LVII; fig. 22.81Nor do the the horse archer grafitti from Dura-Europos, whichalso seem to show the rider either bareheaded, or with a conicalcap or helmet (F. Cumont Fouilles de Doura-Europos - Atlas [Paris1926] Plate XCVIII; and Inventaire des inscriptions Palmyreniennes de Doura-Europos. [Paris 1939] 69); these representations arebelieved by M. Rostovtzeff (The Excavations at Dura-Europos: Preliminary Report of the 5th Season of Work. Oct. 1932 - Mar. 1933 [New Haven 1934] 264; quoted in Coulston [op. cit.] 280, note 110)to be not of auxiliaries in Roman service but of irregularPalmyrenes or Sassanids.82Coulston (op. cit.) 279.^The archers are in scenes LXX,CVIII, and CXV.°Ibid.&Scene XXIV.°Coulston (op. cit.) 278-80.30Although Vegetius recommended that foot archers wear armourm,representations of archers in Roman service of the high Empire seemto indicate that the equites sagittarii wore little in the way ofprotective armour. As stated above87 the type of release mostprobably used in the Roman period precluded the use of a shieldwhen firing the bow. This seems to be confirmed by therepresentationsin sculpture, none of which depict archers usingshieldsm.6. OTHER EOUIPMENTTo carry arrows an archer needs a quiver and to keep his bowdry, and carry his bow when unstrung, he needs a bow-case89. Sincethere are no representations of bow-cases in Roman art of the highEmpire", it is difficult to determine what form of bow-case Romanauxiliaries may have used. Quivers, however, are shown suspendedfrom the right hand side of the saddle to the rear of the rider'smDe Re Militari I, 20 and II, 15.87Section one, Archery, p. 11.mCoulston (op. cit.) 281.89For a detailed description of the evidence for and use ofbow-cases and quivers in the medieval and ancient world seeCoulston (op. cit.) 270-275."The tombstone of the eques of the ala Parthorum et Araborum(fig. 18), seems to show the archer's quiver as long and curved,though it is also possible that the object depicted may be a bow-case. As the tombstone is badly damaged, it is impossible toidentify the object with certainty.31lee. They appear to be about one metre long, and are straightand cylindrical in shape (figs. 19-21). Their capacity is unknown,but the author of the Strategikon recommends that horse archers inhis day carry quivers of "thirty to forty" arrows92.It seems likely that horse and foot archers carried some sortof a weapon for use as a last-ditch defence weapon. Archers fromthe Roman period are not shown with swords as side-arms. Onerepresentation from Housesteads, however, shows a foot-archercarrying a long knife (fig. 12). This same archer is also carryingwhat appears to be a bill-hook in his right hand, used perhaps forthe collection of arrow materials93. Swords, on the other hand,would be of little use to an archer who, after all, carried noshield. The horse archer would rely on his speed to carry him fromdanger, and his dagger, if indeed he carried one, when in direstraits. They would in any case be out of sight on the left handside of the horse, which could explain their absence from thesurviving representations. It is therefore unlikely that mountedarchers carried swords.91An exception is the tombstone of a member of an alaScubulorum from Walbersdorf, Austria, where the archer carries hisquiver on his back on a strap (mentioned but not shown by Coulston[op. cit.] 271). Foot archers carry their quivers in this manner(see fig. 12), presumably to keep them from dragging on the ground.92Strategikon I, 2.93Coulston (op. cit.) 280.32CHAPTER TWOA MILITARY ANALYSIS OF HORSE-ARCHERYThe study of the military, any military, is not simply a studyof the equipment used. It comprises also, in part, the study ofhow this equipment was used, and why it was used that way. Thischapter is intended to answer the question of how this equipmentaffected the way these individuals did their job. In order toanswer this question it will be necessary to analyze horse-archeryin a theoretical sense. Before undertaking this analysis, it isnecessary first to provide definitions, using the examples oflegionary infantry and foot archers before proceeding to mountedarchers. From this, it will become clearer just how mountedarchers should be viewed, and precisely how the three armscompared.Military theorists in part explain military systems in termsof their "characteristics". A "characteristic" may be defined asa feature of a system that establishes the system's effectivenesswhen the system is operating in its normal or intended context. Inother words, the characteristic determines the system's use on thebattlefield. In a military sense everything has itscharacteristics 94 . Because of technology, contemporary infantry94To use a contemporary analogy, the modern Canadian lightmachine gun is, among other things, "...a belt and magazine fed,gas-operated weapon, capable of a sustained high volume of fire in33and armour have additional characteristics such as "communication"which their ancient counterparts could not share. Most land arms,ancient and modern, however, may be characterized under thefollowing headings:1. Flexibility:^the fluid nature of the battlefieldnecessitates rapid reorganization and the performance ofseveral different types of tasks.2. Mobility:^in order to exploit enemy weakness, or tominimize the effect of their own weakness, forces must bemoved around the battlefield guickly95. In particular, ahigh degree of mobility can aid in achieving surprise.3. Vulnerability: such factors as armour, camouflage, degreeof training, and discipline affect the ability of a soldier tostay alive on the battlefield and so the ability of a force tocarry out its tasks.4. Firepower: the effects of weapons and ammunition (range,penetration, and destructive effects), and logistic effectsbursts..." Canadian Forces Publication B-GL-317-019/PT-001 TheLight Machine Gun 5.56mm C9 (Ottawa 1987) 2-24.95It is important to realize that the mobility of a force islimited by that of its logistical "tail"; that is, no force canmove faster than that part of it that is carrying its shelter,food, and weapons' supplies such as ammunition, replacements, andspare parts.34(rate of fire and ammunition supply).Employing these rather technical concepts derived from currentmilitary theory let us now define the characteristics of Romanlegionary infantry, since they are familiar and relatively wellunderstood. Roman military commanders put special emphasis on theflexibility of their units. Vegetius describes a sort of military"cross-training" for the infantry. This includes not onlymarching96, jumping97, fighting with the scutum and g/adium98, andthrowing the hasta99, but also swimming100, throwing rock101 andleadl" missiles, archery103, and mounting horses104.Legionaries could expect to be employed in building not only theregular military works like fortifications but aqueducts and roads96Vegetius De Re Militari I, 9.97To clear ditches and other obstacles; op. cit., I, 9.980p. cit. I, 11.990p. cit. I, 14.1000p. cit. I, 10.1010p. cit. I, 16.1020p. cit. I, 17.1030p. cit. I, 15.1040p. cit. I, 18. Although Vegetius does not name the arm tobe trained in this way, and although it might seem to bespecifically a cavalry skill, I think it in keeping with the spiritof the first book to take this passage to refer to the training ofthe infantry, as most modern commentators seem to assume: forexample Davies (op. cit.) 15.35and other public projects as well105. A force so trained isflexible and provides a commander with the ability to do more taskswith the same number of resources and provides the force withgreater security. A legionary could thus be expected to beproficient at fighting a battle, at carrying messages on horsebackor even serving as a replacement for wounded cavalry106, and atdefending a camp after an unsuccessful battle107. Certainly thecavalry had the advantage herem, since they could be employeddismounted to reinforce the infantry in a situation that affordedlittle room for manoeuvre109. This would certainly be a much morecommon situation than the reverse, since, owing to the amount oftraining necessary for a cavalryman to become proficient, it wouldbe difficult for infantry to take the place of cavalrylio.Legionary infantry was extremely mobile although not very fastor, when in formation, manoeuvrable. They could go across riversby swimming or fording, through dense woods, through broken countryand over steep hills, but not for more than about fifteen or twenty10Davies (op. cit.) 64.mThere is no evidence for this, but for the opposite, cf.n.17 infra.107E.g. as an archer.108Cf. R.W. Davies's (op. cit., 146-50) opinion that the needfor flexibility in provincial garrisons led to the introduction ofcohortes equitatae.109Cf. Ostorius Scapula against the Iceni, Tac. Ann. XII 31, 4.110Although an infantry legion did have a large supply ofhorses, most of these horses were chosen for traits (eg. docility)which would be an asset to beasts of burden but a liability for thecavalry. See Davies (op. cit.) 158-61.36miles a day. They would have to stop about noon in order to givetime for the completion of the camp to give the soldiers somemeasure of protection from surprise attack at night, and this wouldfurther decrease their mobilitym.All soldiers, regardless of their type, are vulnerable all thetime. The fact that legionary infantry had to build fortificationsat every night's bivouac was a reflection of this vulnerability;and obviously in battle even armour could not protect them fullyfrom enemy weapons.Finally, legionary infantry had sword and spear with which tokill their enemies. This was the equivalent of the "firepower"characteristic of modern armies. In terms of range these are bothseverely limited, and the value of the spear was further restrictedby the number carried, usually no more than two. Once these weregone, the only hope for resupply until the end of the battle was topick up a used one from the ground"2.Archers' characteristics are the same in general, although thespecifics are different. Foot-archers frequently fought alongsideslingers or cavalrym as a component of the levis armatura. Assuch, they had to remain unencumbered by armour in order to keep upwith the horses or escape enemy pursuers. Nor could they carryVegetius (op. cit.) I, 9; and H.P. Judson Caesar's Army(Boston 1888) 63.1121 realize I am oversimplifying legionary use of missile-weapons; but the weight and size of a pilum or hasta was adisadvantage which the Romans early on had to develop tactics toovercome.113E.g. Caesar, BC III, 88, 6, and 93, 3.37shields as it is impossible to carry a shield and shoot a bowsimultaneously114. This vulnerability limited their employment tosituations where they could be adequately covered by troops, orplaced in a location that was difficult to attack by virtue ofsteepness or obstacles"5. Because of the heavy investment oftime spent in practice to maintain proficiency in archery the timeavailable to archers for training in other military skills musthave been limited. As a consequence, their flexibility as afighting force would suffer. Furthermore, because bow-strings ofsinew or leather lose their elasticity when exposed to wet, archersbecome useless when there is rain or snow116.Mobility, being an essential merit of the foot archer on thebattlefield, was clearly among the major characteristics of thisarm of the military. Foot archers were able to move rapidly froma position in reserve or on a flank to deal with an unexpectedopportunity or emergency as a member of a group. A single footarcher could approach the enemy front line with comparativeimmunity on his own to snipe, create disorder, and terrorize theenemy, then retreat to safety if threatened.The foot-archer's prime characteristic, however, was hisfirepower. He could, depending on the type of bow used, fire114see supra, p. 12.lnE.g., Arrian Trugt; KaukVámov 12-13.116This happened to Antiochus's archers in Lydia (FrontinusStrategemata IV 8, 30).38deadly and accurate missiles at a long range117 either atindividuals or broad targets such as masses of troops. He couldfire several arrows per minute until his quiver was empty, at whichpoint he would have had to return to draw more from his regimentalstores or pick up any he could find on the battlefield. Hiseffective firepower would, of course, be expected to dropconsiderably when moving over difficult terrain or at speed. If,for example, he was obliged to climb a series of hills or runfollowing cavalry or from a pursuing enemy, his ability to fireaccurately and draw the bow with force would be adversely affected.An eques sagittarius can be seen in a military sense as animprovement over his pedes counterpart in all of the abovecategories, though there were some limitations. He shared withother cavalry the ability to carry out a role dismounted. He wasable to close with and retire from the enemy at greater speed, andbypass or outflank an infantry battle line. Thus, he may have beenused in situations where the distance would have been too great fora dismounted archer to cover in the time necessary. A horse-mounted archer was faster than one on foot but, as a rule, thisadvantage could only be realized in open country; he lost thismobility when faced with steep hills or very rough terrain. He wasable to cross fast rivers more safely than foot soldiers, and couldeven facilitate the crossing of infantry in such situations, but he117Coulston (op. cit., 290-1) quotes a variety of sources, mostarriving at a figure of 200 to 300 m for accuracy and 100 to 150 mfor an effective (killing) range; this latter range would obviouslydrop for an armoured target. See supra, pp. 3-4.39had to take care to keep his bow and string dry 118. A formationof horse-archers could turn, wheel, advance, and retire like othercavalry, much faster than a similar formation of infantry, butalways assuming that the terrain was open and reasonably level. Inclose or broken country it would have been very difficult to moveat all and keep formation. Close country also presented anotherdanger to cavalry: vulnerability to ambush or trap. Infantry werealso liable to be surprised where visibility was low but were moresuited to fighting in these conditions119. Depending on theamount of armour worn by a horse-archer and his mount, he was moreor less vulnerable to attack by other horse archers or by differentcavalry. On the other hand, his speed would usually have enabledhim to outdistance a pursuer, and he would still have been able toharass him while retreatingin. Furthermore, compared to othercavalry, their inability to carry a shield obviously limited theirprotection from all forms of weapon.There were other disadvantages too. The larger the bow, thelarger the bow-case, and the need for this bow-case121 makes use"8Cf. Caesar's use of cavalry in river crossings to slow thewater for the infantry and to catch any who lost their footing inthe swift current (BG VII, 56; and see Judson [op. cit.] 70).119Horses can be wounded or killed, or merely frightened; theythen become difficult to control and the resulting confusion makesan organized response difficult. Infantry, on the other hand, canmove about more freely and, when well trained, are less likely tobe so completely surprised that they are incapable of response.1201.e. the Parthian shot.121This would be necessary to keep the unstrung bow and itsstrings safe from moisture and to facilitate their carrying whennot in use. See supra, p. 28, and Coulston (op. cit.) 270.40of a large bow on horseback an impracticability. The case neededfor a six- or even five-foot bow would not only come close todragging on the ground but might interfere with the horse's legsduring running. Finally, the need to be able to shift from onetarget to another that might require sighting from different sidesof the horse quickly militates against the use of a large bow,which would have to be lifted over the horse's neck and head122.Thus the size of bow that could have been used on horseback waslimited, and with it its range and armour-piercing capability.Horse-archers would have been able to compensate for thisdisadvantage to a degree by their mobility, but if they could bekept away from their targets by a force of foot-archers employinglonger-range bows, their usefulness would diminishm.It is conceivable that a mounted archer could carry largeramounts of ammunition simply by carrying more quivers or bundles ofarrows. There is no actual evidence for this possibility, however,although there are signs that ancient commanders recognized theproblem of archers' limited ammunition supply and attempted to dealwith it in other ways124. Horse-archers could pick used arrows122Moreover, shooting at a target directly to the front of thehorse would be almost impossible; see supra, p. 25.mCrassus's foot-archers at-Carrhae were probably armed withself-bows instead of compound bows like the Parthians; thus theywere outranged and unable to affect the Parthian horse-archers'access to the Roman infantry.124Cf. Surenas's use of a train of camels loaded with arrows atCarrhae (Plutarch Crassus, XXV). This prodigious use ofammunition, and the incredible logistical preparations necessary tosupport it, foreshadows modern practise.41from the field like their unmounted counterparts, albeit with lessease, but they held the advantage in that they could retire insmall groups to a rear location to be resupplied and return to thebattle far more quickly125.So we see that the effectiveness of horse-archers can bedefined in terms of their flexibility, mobility, vulnerability, andfirepower. In open ground they represent improvement in all ofthese characteristics over pedites sagittarii in one way oranother, except in range and armour penetration, i.e. "firepower".This disadvantage they make up for in part by their ability toclose with and retire from their enemy at speed causing casualtiesthe whole time. They were, like all cavalry, at a disadvantage inclose country, and would have to be protected by infantry. In somesituations the way ahead would have to be cleared of enemy beforethe cavalry could proceed. These characteristics would form thebasis for an ancient commander's disposition of his units ofequites sagittarii.125Cf. Plutarch, Crassus, XXV.42CHAPTER THREEQUESTIONS OF ARMAMENTCentral to the question of the employment of horse archers isthe question of their arming. Were these soldiers entirely, andexclusively, bow-armed? Or were units or individuals allowedgreater flexibility to equip themselves according to the needs ofthe situation? Furthermore, if they were permitted to chooseweaponry according to the situation, was it only during garrison orborder control duties, or did this flexibility extend to when theywere employed as part of an expeditionary army? This section willshow that, contrary to what is generally believed, the equitessagittarii had the flexibility to use bow, spear, and even club andsling according to the requirements of the situations they foundthemselves in.To begin, it is necessary to establish what was required toensure effectiveness in a unit of horse archers. Caesar126,Tacitus127, and Josephusull all refer to units of equites126Bellum Civile III, 4, 5. Amongst Pompey's forces were 200mounted archers sent by the Commagenian Antiochus I.127Annales II, 16, 12.^Mounted archers were part ofGermanicus' force against the Chatti and accompanied his legions atIdistaviso.1282e11um Iudaicum II, 500, 5: Antiochus IV of Commagene sent2000 horse- and 3000 foot-archers along with Cestius Gallus; Gallusalso took six cohortes of auxiliary infantry and four a/ae. Ifthese units were quingenary, the Commagenian and Roman totals wouldbe approximately the same; Josephus may have confused the two here.D.B. Saddington, "The Roman Auxiliaries in Tacitus, Josephus, and43sagittarii, these being ethnic units supplied en masse by client-kings129. Their ability on horseback and with the bow would haverequired them to begin their training at an early age; Sarmatians,for example, received training in horsemanship and archery startingat an early age and kept it up at a high level throughout theirlives. This phenomenon of training in specialist skills from anearly age had good parallels. Children of the Balearic Islands,for example, were believed in antiquity to have been allowed to eatonly when they had secured their food by hitting it with a missilefrom their slingm. Whether this was true or not, other ancientcivilisations certainly placed an emphasis on training their malechildren in military skillsm.Horsemanship, especially at a level of skill where a bow canbe used effectively while galloping, is extremely difficult tomaster and would require daily training for many years. For manypeoples their safety and livelihood depended on an effective horseother Early Imperial Writers" in Acta Classica XIII (1970) 117-8does not seem to notice this and accepts that the auxiliary forcesconsisted of both Roman and client-king forces. A more detaileddiscussion would be outside the scope of this paper.129Caesar and Josephus say exactly this; it is in any eventperfectly natural for these units, whose specialist skills were notfound in the Roman world itself at the time, to be recruitedoutside the boundaries of the Empire.130Strabo III, 5, 1 in Griffiths "The Sling and its Place inthe Roman Imperial Army" in van Driel-Murray, ed. Roman MilitaryEquipment: the Sources of Evidence. BAR International Series #476(Oxford 1989) 263; and Vegetius, De Re Militari I, 16.mE.g. the ancient Spartans and Athenians; wrestling, javelin,and discus throwing have obvious military applications and probablyhad a military origin.44archer force capable of providing defence: the Palmyrenes andNabataeans, especially, depended on horse archers to secure theroutes through the desert that brought their cities trade andsupplies. They also revered the skills of the mounted archer tothe extent that some of the gods they worshipped were mountedarchers132. They must have devoted considerable effort to theraising of horses suitable to carrying these mounted archers and nodoubt spent a comparable effort in archery training. Withoutquestion, mounted archery requires a considerable amount oftraining for it to be effective in battle.In practise it proved difficult for mounted archers tomaintain their skills at a high standard for several reasons. Oncea unit has been removed from its place of origin, it did not takelong even in peacetime, before disease, accidents, and retirementwould take their toll and replacements would have to be found.Obviously a replacement for a specialist unit would have to be ableto carry out the tasks required of him. Reinforcements for horsearcher units must therefore have required comparable training ifthey were to equal the performance of their comrades already inservice. These replacements must have come from either the sameplace as the units in which they were being enrolled or from someregion where horse-archery traditions were equally strong.The evidence for the origin of the auxiliary sagittarii, as132Coulston (op. cit., 237) and see M.I. Rostovtzeff "TheCaravan-Gods of Palmyra" JRS 22 (1932) 107-116 for others.45distinct from their officersm, is quite well documented. Van deWeerd and Lambrechts134 show with reasonable certainty that theywere recruited in Syria and the east, the traditional lands of themounted archer. This suggests that military ability and specialistknowledge, at least in the case of the sagittarii, were to Romancommanders the crucial factors for recruitment rather than ease ofrecruitment. Moreover, there must have been some provision for thetransportation of replacements from the east to units that werestationed far away. Units such as the cohors I Hemesenorummilliaria equitata, the numerus Palmyrenorum Porolissensium, andthe cohors I sagittariorum, stationed in Pannonia Inferior, Dacia,and Germania Superior respectively, were the beneficiaries of sucha systemlm. The considerable expense and trouble of finding andrecruiting men in the east, and transporting them to the locationsof their new units to receive their initial military training,underscores the importance placed on the effectiveness of bow-armedunits.The evidence for this practise, however, is far frommThere are, on the other hand, signs that officers ofauxiliary sagittarii were, for a time, exclusively Italian inorigin at a time when many auxiliary units were commanded byofficers of eastern origin. H. van de Weerd and P. Lambrechts("Note sur les corps d'archers au Haut Empire" Die Araber in derAlten Welt [Berlin 19641 661-677), relying on prosopographicalevidence, believe that this evidence, if pointing to a regularpractice, indicates a fear in the Roman high command of theloyalties of their equites sagittarii.'340p . cit., 667-70.135G.L. Cheesman, The Auxilia of the Roman Imperial Army(Oxford 1914) 82-4, H. van de Weerd and P. Lambrechts (op. cit.661-2), and J. Davies (op. cit. 261-2).46conclusive. Kennedym points to a number of Parthian units thatformed the force of horse-archers in the Roman army before Syrianunits were recruited which may have lost their function as bow-armed units since there was presumably no mechanism for recruitmentand replacement from skilled sources. As the supply of skilledrecruits from the original source dried up, it became more and moredifficult to train new members recruited locally to a decentstandard. This resulted in the gradual transformation of thecharacter of these units from sagittarii to more conventionalauxiliary regiments"7. Coulstonm discusses the tombstone of136,D.L. Kennedy, "Parthian Regiments in the Roman Army" J.Fitz, ed. Limes: Akten des XI Internationalen Limeskongresses (Budapest 1976) 530.137Kennedy bases his conclusion that Parthian units stoppedbeing bow-armed mainly on the lack of the -sag suffix ininscriptional evidence. While he may find it "surprising" thatonly one text referring to a Parthian unit describes it as - sag, isit not more surprising that any would describe a Parthian unitas such? As he himself says, "Parthian soldiers are notedfor.. .above all.. .horsemanship and archery" (op. cit., 527) andindeed, those Parthians most likely to become mercenaries and fightfor Rome are from the lower classes. These lower classes suppliedthe horse-archers to the Parthian armies, while the upper classessupplied the armoured lancers. The "Parthian shot" was a leitmotifin Roman literature (Coulston [op. cit.] 292). Since the Parthianswere so universally known as mounted archers, it would be asredundant officially to title them "archers" as it would be todayto title the Seaforth Highlanders "infantry". All highland unitsare infantry. As he notes, however, "...there is no evidence toshow continued oriental archers being drafted to the regimentsafter their formation..." (527). We must be careful to distinguishbetween recruitment and membership on the one hand, and tacticalfunction and armament on the other. My point is that although thefirst may, and in the case of archers almost certainly will, havean effect on the second, it is not the only factor.1381985: 289. The tombstone is discussed by N. Benseddik inLes troupes auxiliares de l'armee romaine en Mauretanie Cesarienne sous le Haut-Empire (Algiers 1979), 38-40, fig. 11, which isunavailable to me.47a soldier of the ala I Augusta Parthorum, stationed in Mauretania.This depicted the deceased armed with a spear and javelin, evidencethat could be construed to indicate that this regiment had ceasedbeing bow-armed at some stage. Since there is no archaeological orother evidence whatsoever to suggest that it was ever bow-armed,however, we must conclude that it had lost whatever function it mayoriginally have had as a regiment of archers early onm.Other units were not kept so far from home and, in Coulston'swords, "...the strategically important Eastern regiments would nothave been allowed to decline in skill... nuo These would havebeen units such as the cohors XX Palmyrenorum, whose situation wasvital in controlling an important trade route in the desert and inpatrolling a large section of frontier territory. Palmyra is closeto Dura Europos and recruiting, selection, and transport of newtroops was thus much simpler. This is not to say that these unitswould not train with other weapons as well. Clearly they wouldhave been obliged to use a variety of weapons in order to carry outtheir tasks for which the bow was unsuitableul.Evidence of the difficulty in maintaining a high standard ofability in a unit of horse-archers comes from El-Kantara, Numidia139"None of the known personnel of the ala Parthorum were oforiental origin and perhaps two centuries of isolation inMauretania resulted in a change of weaponry." Coulston (op. cit.)289.1400p. cit. 289 (e.g. the coh. XX Palmyrenorum).ulConvoy protection, for example, requires that soldiersengaged in it be armed with both short- and long-range weapons.Those protecting a convoy must be able to fight attackers who havepenetrated a missile screen and have come to close quarters.48in the form of a tombstone (I.L.S. 9173). A Palmyrene centurion,Agrippa, of the cohors III Thracum Syriaca equitata undertook thetraining of the Palmyrenian archers of the cohors I Chalcidenorum(curam [e]git Palmyr. [s]ag.). The cohors I Chalcidenorum, with anumerus Palmyrenorum, was in the area long enough to construct anamphitheatre and other public works, and during this time perhapsthe archery skills of the numerus had begun to deteriorate. In anycase, a senior soldier from a unit of archers stationed in an areawhere archery traditions were strong was brought in, and theassumption may reasonably be made that it was to maintain theskills of the archers and not simply take care of the soldiersthemselves142.A further example of an expert moving about to train archers,and particularly horse archers, is Barsemis Abbei of Carrhae143.His career included tours with a numerus Hosroruorum, the cohors IHemesenorum milliaria sagittaria equitata, and the ala firmaKatafractaria. As Coulstonl" says, "To judge from this series ofunits Barsemius must have been an equestrian trainer and/or anarchery expert... (perhaps) a skilled horse-archery campidoctor."Shooting skills, it should be noted, were not the only skillsrequired in an archer-unit that would begin to deteriorate with142For further discussion of this case see J. Carcopino "Lelimes de Numidie et sa garde Syrienne" Syria VI (1925) 119-22.1431, ...Barsemis Abbei dec(urio) ala firma Katafractaria exnumero Hosro [en] orum, mag ( i s ter) coh (ortis mil iariae )Hemes(enorum), n(atione?) d(omo) carris..." CIL III, 10307.1440p. cit., 289.49time spent away from where the unit was recruited. The members ofsuch a unit would also need to have considerable skill to buildbows, and ample time to acquire and practise the skill145. Likearchery itself, the knowledge of making bows requires a lifetime oftraining and practisel". The climate in which bows areconstructed might also have been a factor in the efficiency ofarchery units removed from their original home in the east. Theapplication of the layers of glue-soaked sinew requires lowhumidity and warm temperatures, features not normally associatedwith northern European climates. It is thus possible that adecline in the quality of bows coming from the regimental storespersuaded commanders to diversify their weaponry.This is not to say that a mounted unit that began to use otherforms of weaponry would lose its function as horse-archers, asclearly archery training was a part of life for all Roman soldiers.Indeed a unit of mounted archers would serve as a repository ofknowledge of archery and horsemanship second to none and could bethe source of skilled trainers for the other arms. This isspeculation, it is true, but the sequence of Barsemius Abbei'scareer leads to the conclusion that an expert in one form offighting could be sent to units of different function as acampidoctor to broaden their training and increase theirflexibility.1451i period of up to a year for bow construction is recorded inthe Ghunyah of Taybughah (Coulston [op. cit.] 249).1461g.F. Paterson, "The Archers of Islam" Journal of the Social and Economic History of the Orient  9 (1966) 69-85.50It is generally assumed that a unit of mounted archers must bearmed exclusively with the bow. This assumption, implicit in mostwriting on auxiliaries147, however, requires some qualification.The bow-armed cavalry of other ancient peoples is frequentlyrepresented carrying other weapons in addition to the bowm. TheStrategikon attributed to the Byzantine emperor Maurice describesthe Byzantine battle line as consisting of horse-archers with thoseinexpert at the bow carrying light javelins149. The Byzantinebattle line of the sixth century described in Procopius was armedwith the bow but certainly could hold its own in a hand-to-handmelee; it too used sword or spear as requiredm.There is plenty of evidence for the cross-training of otherbranches of early imperial auxiliary cavalry in other forms ofweaponry. Arrianl" discusses it in detail in his manual oncavalry training. There exists, moreover, the text of an imperialin Kennedy (op. cit.), Cheesman (op. cit.), andSaddington (op. cit. 91, 95, 117-119).mCoulston (op. cit., figs. 33, 44), Y. Yadin The Art of Warfare in Biblical Lands (New York 1963) 295-7, A.D.H. Bivar"Cavalry Equipment and Tactics on the Euphrates Frontier" DumbartonOaks Papers 22 (1976) 282.149Strategikon II, 3. In the case of the Byzantines we aredealing with a main battle line intended for close contact with theenemy and not a unit of chasseurs, but my point is that differencesin ability and level of training were recognized and dealt with bya modification of armament; individual soldiers were expected to beflexible enough to use different weapons according to thesituation, and this manual prescribes training sufficient to bringthis about.150.Bellum Gothicum I, 1, 12-3.151Teritl tarrucT1 43, 1.51address given by Hadrian to units of cavalry given after a parade-ground demonstration. In it the Emperor praises the troopers ofthe cohors VI Commagenorum for, among other things, their vigoroususe of slings and javelins152. The use of slings points to agreat deal of training for both men and horses. Not only isslinging a difficult skill to master153 but unless horses becomeaccustomed to the sound of an object whirling over their head, theymay rear or bolt154. Infantry units of the Imperial period werecertainly expected to be generalists in weapon skills. Vegetius inhis De Re Militari describes training in the use of the bow as wellas the slingm for regular legionary infantry, while infantry arealso to be taught to ride horses156. Birley's belief 157 that allof the cohortes Comma genorum were primarily and originally bow-armed seems correct, but, for the reasons already stated, there is152” verum vos fastidium calore vitastis strenue faciendo quaefieri debebant. addidistis ut et lapides fundis mitteretis etmissilibus confligeretis..." CIL VIII, 18042, 11. 8-11.1531K.B. Griffiths, "The Sling and its Place in the RomanImperial Army" C. van Driel-Murray, ed. Roman Military Equipment: the Sources of Evidence. BAR International Series #476 (Oxford1989) 264. He observes that it is not clear from the text whetheror not the equites were mounted when they used the slings; however,all of the other exercises described in the text are cavalryexercises (frequens dextrator, Cantabricius densus...saluistisubique expedite "the right wheels were in quick succession, theCantabrian gallop was done tightly together, and you jumpedeverywhere promptly"), so it seems most likely that the slingingwas also done on horseback.154Hyland (op. cit.) 168.1551, 15-161560p. cit., I, 18.157In J. Davies (op. cit.) 270.52no reason to believe that they, or any other bow-armed unit forthat matter, used the bow to the exclusion of all other forms ofweaponry158.If, however, the cohors VI Commagenorum was indeed bow-armed,we must explain why it is not referred to as such in the address ofHadrian. The text in fact describes them as fightingmissilibusm but this word, when it can be defined withcertainty, refers to javelins or other weapons rather than toarrows168. Coulston161 suggests that the horsemen of the cohortesequitatae sagittariorum may have been armed with javelin and spear,and bases this on the evidence of Hadrian's address. However, ina unit where the career pattern went from pedes to eques162, itseems unlikely that a soldier would be made to re-arm and re-trainon his promotion; rather, he would stay with the weapon or weaponsthat he had trained in for many years. The commanders of bow-armedunits had made a considerable investment in recruiting and158Certainly the fact that they, and the Ituraeans andPetraeans, were well known as archers may help to explain why thereis so little reference to them as such in the epigraphical sources.Of the ala Augusta Ituraeorum and the cohortes II FlaviaCommagenorum, I Augusta Ituraeorum, and III Ulpiae Petraeorum (allattested as sagittariorum) only the coh. II Flavia Commagenorum isso attested more than once. Cf. my comments on the Parthians, note13 supra.1591. 10.1601 know this from my own survey of the literature but I havebeen unable to find any corroborating discussion in scholarlywriting.1610p. cit., 285. He does not make any conclusions, however.162R.W. Davies (op. cit.) 145 and note 31.53selection from lands where horsemanship and archery skills were ofa high standard. This system was intended to ensure an effectivecorps of archers. It is scarcely credible that they would squanderthis investment by forcing these archers, when they had beenpromoted equites, to acquire the skill of, for example, spearmen.In conclusion, the reference to the cohors VI Commagenorum inthe address by Hadrian does not necessarily mean that it was notbow-armed. Clearly it was armed with a variety of weapons, and itdemonstrated its prowess in the use of the spear and inhorsemanship to its Emperor in the parade commemorated by the aboveinscription. This suggests either a series of formal, standardizedparade-ground rehearsed formations for show and training, or thatcavalry units were cross-trained in weaponry. Indeed, it is likelythat both were the case.In addition to the epigraphical and literary evidencediscussed above, tombstones and other archaeological finds providefurther evidence for the arming of mounted archers. It is clearthat many of the units thought to have been composed of archershave been identified as such only on the basis of a singletombstone showing a soldier engaged in that activity163.Furthermore, ...no equites cohortales appear as tombstonefigures..."1", so we have no indication of how they were armed163The ala I Augusta Ituraeorum, the cohors I Ascalonitarumequitata, and the cohors II Flavia Commagenorum equitata areexamples of units only once described as bow-armed. See Appendix1 for further examples.164Coulston (op. cit.) 284.54from this direction. Archaeological evidence of laths andarrowheads is not taken as evidence of an archery function for aunit since many units had a supply of bows and arrows for muraldefence and general training as a matter of course165. Indeed, itis known that a soldier from a non-archer unit could takesufficient pride in his accomplishments with the bow to commemoratethem in stone166. It should not, therefore, be surprising thatanother could do the same, and show himself riding or shooting thebow when his unit was not officially designated as bow-armed. Thussome of these funerary depictions of archers may represent nothingmore than pride in the deceased's skill as an archer, and not thathe belonged to a bow armed unit. Some of the tombstones showingmounted archers, therefore, may be taken as evidence not of theproliferation of units armed in such a way, but as further evidencefor the detailed cross-training of the Roman military167.As far as a unit such as the cohors XX Palmyrenorum wasconcerned, normal patrol and border control duties would surely165J• Davies (op. cit.) 265; and Coulston (op. cit.) 285.1660L III, 3672 (= ILS 2558), and Dio LXIX, 9, 6. A soldierof the cohors Batavorum milliaria equitata swam the Danube in fullkit, then fired an arrow straight up and hit and broke it withanother before it hit the ground, all under the gaze of the emperorHadrian.167MY comments above on units from known areas of horse-archerytraditions notwithstanding. For example, the ala Celerum is onlysuspected to be bow-armed on the basis of an inscription describinga trooper as highly skilled in archery (...vir sagittandiperitissimus...: CIL III, 4832). Although this unit did serve inArabia and may have been recruited there (J. Davies [op. cit.] 269)I think that this is hardly reason to assign such a role to theentire regiment. The trooper of the a/a Batavorum (see note 43)was also very skilled in archery.55require considerable flexibility in weaponry. In a largeexpeditionary army a general could always call upon different unitsfor different tasks, but auxiliary alae and cohortes garrisoned inisolation often had large areas under their control. In thesesituations, a variety of possible situations might arise callingfor flexibility in armament. The use of a cohors equitata for riotcontrol is attested168, and for escort duties169 a smart commanderwould ensure that his men were prepared for both short- and long-range fighting. Indeed the caravan-gods of Palmyra, gods ofconvoy-escort, are depicted mounted and armed with both bow andspearlm.Our evidence for the armament of archer units in large-scaleexpeditionary armies is somewhat less equivocal. Arrian, in hisINn414 imacCALA6vwv, groups his units according to armament. Horsearchers are employed in the advance guard precisely because theyare horse archers: their characteristics such as mobility andfirepower suit them for this task, and their use of weapons other168E.g. a cohors equitata was taken to Jerusalem by Florus inA.D. 66 to carry out riot control; the cavalry was used for crowdcontrol but later, when the crowd became too unruly the infantrycontained them with clubs (or "riot-batons" to use Davies'modernistic euphemism) and the cavalry moved them on (cf. J. Davies[op. cit.] 88, 147).169Escort duty is attested in the papyri for cohors XXPalmyrenorum (for which see J.F. Gilliam, Roman Army Papers [Amsterdam 1986], and R.W. Davies [op. cit.] 146-7) and in the Actsof the Apostles 23.23, 31-33 Paul was escorted to Caesarea bycavalry and infantry, presumably members of the same cohorsequitata.lmSee p. 45 supra.56than bows would render them much less effective.^LikewiseJosephusln seems to describe an expeditionary army advance guardcomposed in large part of innat64otat. A commander could, of course,re-arm his mounted archers for an upcoming battle in which he sawlittle use for them; we lack evidence for this sort of flexibility,however172.Since it is clear that skilful use of the bow on horsebackrequires years of training to acquire and constant practise tomaintain, a horse-archer unit without a source of skilled recruitsto draw from, would, after a period of time, begin to lose itsproficiency. A unit could obtain a skilled instructor such asAgrippa or Barsemis Abbei to help resist this tendency, to be sure,but his students' generally lower standards of skill at the time ofrecruitment would begin to limit what they would be capable ofafter training. Eventually it would be necessary to arm some oreven the majority with other weapons in order to give them someeffectiveness on the battlefieldim. The fact that some unitsdid, however, maintain a high level of skill in archery in order tocarry out their mission does not mean that these units did nottrain with other weapons; in fact, on account of the variety oftasks required of them, they were most probably able to usewhatever weapon they needed.171Bellum Iudaicum III, 115-126, and V, 47-9.172Evidence is only lacking for mounted archers, however. Forother cavalry, see Chapter 2 supra.173As in the case of the Strategikon.57CHAPTER FOUROPERATIONAL EMPLOYMENTThe equites sagittarii undoubtedly formed an important part ofthe Roman field army of the high Empire. Their employment on thebattlefield, however, has received little attention. Coulston174believes that their employment has already been studied in detailbut the studies to which he refers either discuss horse archersonly amongst other arms175 or confine consideration of them to alimited periodlm. These studies also tend to be more concernedwith historical issues, with limited attention to purely militaryconsiderations1". Those scholars who have an interest in themilitary tend to confine their interest to broad strategic andpolitical topics relating to auxiliary forces, to terms of service,or to archaeological questions. Such inquiry is vital, to be sure,but an understanding of the soldiers themselves, why they fought,and especially how they closed with their enemy and killed him, isthe most certain way to discover what made the Roman military'740p. cit., 220. Coulston is referring to the employment ofarchers in general, not specifically horse archers. Obviouslythere are significant differences in the role of the two arms.175E.g. D.B. Saddington, "Roman Auxiliaries in Tacitus,Josephus, and Other Early Imperial Writers" Acta Classica 13 (1970)89-124.176Saddington (op. cit.), and P. Medinger, "Les archers Partheset l'arc Turquois A la bataille de Carrhes" Revue Archeologique serie 6, no. 2 (1933) 227-234.Somewhat of an exception to this rule is E. Darko, "Note surles corps d'archers au haut empire" Altheim and Stiehl, eds. DieAraber in der Alten Welt (Brussels 1935) 287-310.58machine so effective. Politics and strategy do not win battles.What makes groups of armed men so effective in violent combat does.Such a study of the military with emphasis on its lower levelshas been ably done for the Greek world by Victor Hanson178 butRome has so far gone without this kind of treatment. MordecaiGichon's discussion of the Roman military in the Jewish revolts";is highly realistic and obviously based on experience of soldieringand especially soldiering in desert conditions. His treatment,however, views war from the perspective of the general, not of theprefects, tribunes, and centurions. These junior officerscommanding the a/ae, turmae, and centuriae had to decide how bestto employ men and carry out their tasks at regimental and companylevel. How precisely these commanders employed them, however,remains unclear. In this chapter, therefore, I intend to explore,through analysis of the relevant sources, literary andarchaeological, and of scholarly literature, why the equitessagittarii were so important, how they fought, and how theircommanders employed them.Before addressing these questions it is necessary to providea few definitions. War, on an operational or tactical, as opposedto a strategic or political level, is considered by militarytheorists to take place in a number of phases. These phases differ178v Hanson, The Western Way of War (New York 1987).179M. Gichon, "Aspects of a Roman Army in War According to theBellum Judaicum of Josephus", P. Freeman and D. Kennedy, eds., TheDefence of the Roman and Byzantine East, BAR International Series297(i) (Oxford 1986) 287-310.59by virtue of the different requirements they impose on commandersaffecting disposal of troops, logistical preparations, and spiritof operations. On an individual level war is simply attack anddefence. At sub-unit level (i.e. company or squadron in modernterms, or centuria or turma for a Roman auxiliary unit) and unitlevel (i.e. battalion or regiment, or cohors or a/a), however,there are five different phases of operations, which fall into thebasic categories of attack and defence. It will be convenient totake these five phases of war in the order presented below, if onlybecause that is the order in which Western military schools teachthem. "Attack" is composed of:1. Advance to contact: movement of a force towards aknown or suspected enemy location or through enemy-heldor -dominated country. Enemy reaction is unknown and soall-round defence is essential, as is thoroughreconnaissance. Economy of force is often achieved bythe use of highly mobile, lightly equipped units forreconnaissance and protection.2. Attack: the process of closing with and destroyingthe enemy including preparatory fires. Maximum combatpower and violence applied at the "right" point in timeand space is prerequisite for successmcl.180The definition of the term "right" as applied to time andplace varies over time and even between and within armies at thesame time. Until the Second World War it was common to attack only603.^Pursuit: the pursuit of a defeated enemy force inwithdrawal. Never allowing a force to regroup, rest, orprepare for further operations will weaken it and hastenits defeat. Highly mobile forces are well used in thisphase to keep up constant pressure.Defence is divided into:1. Defence proper: holding of ground against an enemyforce intent on its capture.^A fortification is adefensive position and we may also consider a battle linereceiving an attack to be in a defensive posture. Mobilereserves detailed to plug gaps, launch counterattacks,and relieve worn out troops are as essential to thesuccess of the defence as is dogged persistence inmaintaining position.2. Withdrawal: the movement of a force, which might bein or out of contact, with the intention of removing itfrom circumstances unfavorable for battle. Withdrawals,of course, can themselves be designed to lead an enemyinto a larger trap or onto ground unfavorable for him.strongpoints; now it is common to bypass strongpoints and exploitareas of maximum weakness.61What does all this have to do with mounted archers? As withall troops, mounted archers' actions in battle vary according tothe phase of war their leader is fighting at any given moment. Itis necessary that the reader understand what they are and whathappens in them in order to understand why the equites sagittariido what they do.1. ADVANCE TO CONTACTWe turn first to the advance to contact. In antiquity thisoften consisted of an army moving from one fortified camp toanother in hostile country.^Two sources contain our mainreferences to such movements, Josephus 181 and Arrian182.^Bothauthors must be regarded as extremely important since they both hadfirst-hand experience as generals. One was a leader of suchmovements, while the other, before becoming an observer of militaryactions from the Roman side, had commanded troops in several majoractions against them183.Josephus has two notable digressions on the organization of181For ref., see below.182Arrian, INmoc4; icate/dahNow, especially 1-2.183With proper regard for their limitations as sources we canmake considerable use of them. "...Josephus' accounts must.. .beread with due criticism. This must necessarily be most severe whendealing with his opinions of political motives...and...when dealingwith the narrative of their (definite persons) deeds. On the otherhand, descriptions of...institutions and their functioning (italicsmine)...have been largely sustained by all outside evidenceavailable." Gichon, op. cit., 287.62the Roman column of march. The first describes the army underVespasian marching into Galilee in 67184; the second that of Tituson its way to Jerusalem in 70185.auxvO 6 Kai Napa COW 13ocauliew ovvitten avRgaxixOv,'Avrthxou Rev Kai 'Aypirna Kai Zoaigou napaaxogevow &vacSuyxkliovc ice4oi); to4Otac xai )ca,iou; tivirciç, toi 8e "Apapo;Waxen) xaiouc negwavroc inneic ivI itcoc nevtaxtcrxtliotc,frrav46tat...tolic Rev ye xrul,olic tóveinxo6ixov Kaito4Ocac poayetv exe)teuoev, thc ecvaxOrtotev Tac e4a7ttvaioucOv noXeRiew entSpoptic; icai. Stepeuvev tac ikrOrtovc xaiXoxdoOat truvagevag 0A,a; . . . (BJ I I I, 68, 116 ).../cp6; oic at te 're6v aaiAwvauggaxiat nai) rleiou; xaiovxvoi Vint (loth Ivpiac trixoupot auviWov....npotOvn 8i eicnoXeRiav TIN npoilyov Rev oi 13ccathxoi Kai nay toavpRaxtx6v.... (B,7 v, 42, 47)"Also, a great allied force had beengathered by the kings Antiochus186,Agrippa187, and Soaemus188, each providingtwo thousand foot-archers and one thousandcavalry, while the Arab Malchus sent athousand cavalry and five thousand foot. Themajority of the latter were archers.... He(Vespasian) ordered the light-armed troops ofthe auxiliaries and the archers to go inadvance, in order to beat off any sudden enemyattacks and search suspicious woods suitablefor ambushes."...In addition to these came the kings'forces in much greater strength, and a great184Be11um Iudaicum III, 115-126.185 j V, 47-49.16.Antiochus IV of Commagene.ivAgripp a II of Chalcis.lmSoaemus, king of Emesa.63number of allied troops from Syria.... Thekings' forces and the entire auxiliary forcepreceded Titus on his march into enemyterritory."The advance guard of the army of Vespasian consisted of "theauxiliaries and archers" whose task was reconnaissance and forwardprotection; that of Titus's army was made up of the client-kings'forces with the remainder of the auxiliary force. Auxiliaries arenot referred to as being in any other part of the order of march,except as the rear guard of Vespasian's army. The client kings'forces with Vespasian's army consisted largely of archers bothmounted and on footm; Titus also received auxiliaries from thekings. We may conclude that a large portion of the advance guardof both armies consisted of mounted archers.Let us take a closer look at the organization of this portionof the army. It is the light-armed troops and the archers who goahead of the main bodyl". If Josephus's statement that the"entire force"'m of troops contributed by the kings to Titus'army was deployed in the advance guard was true also forVespasian's march, and assuming Josephus's figures of the kings'contribution to Vespasian's army are correct, then there were up to189What Antiochus, Agrippa, and Soaemus sent is not madeperfectly clear. Josephus says that they each contributed twothousand foot-archers, and one thousand cavalry (BJ III, 68) but Ithink that cdtenac certainly refers to both infantry and cavalry.Malchus of Arabia, however, clearly sent a large force of both footand horse archers (BJ III, 68).190BJ III, 116191BLT^42, 47.64eleven thousand foot archers and four thousand horse archerspreceding Vespasian's legionsm. The composition of Titus' armyis less clear, but it seems reasonable to assume that the client-kings made similar contributions. There was as well a considerablebody of regular auxiliaries accompanying Vespasian's army193.The point of this is not a mere exercise in "bean counting".Fifteen thousand is a very large body of troops with which to find,mark, and guard a routem. Gichonm on the evidence ofpassages from Caesarm, believes that Roman advance guards wereecheloned in depth (i.e. divided into two bodies: the larimiantecursores and the antecursores) as is the practice in more192Since Antiochus, Agrippa, and Soaemus each supplied threethousand troops, and Malchus sent six thousand. However, I believethat Josephus is giving very rounded numbers here, rather thanexact parade or paper strengths. He is writing a history, not amilitary manual, and thus may reasonably be expected to giveapproximate figures or pass over details of interest toprofessional soldiers in the interest of producing a readable text.193There were twenty-three cohorts (ten infantry, thirteencohortes equitatae: " TiOv 8g amp& ai &Ica ggv glxov avec )0cruc 7ce0i)c,8g Xourai, Senapeig ecvec gaicoaiouc ggv ICE4oi)c, iniceic 8g &at& Erman; . . " BJ III ,67.) and six a/ae of regular auxiliaries with Vespasian's army. Ofthese, five cohorts and one ala were from Caesarea, and five a/aewere from Syria. Some of these units were probably bow-armed.Josephus does not describe precisely where these were located inthe order of march, but merely says that "light infantry...and alarge amount of cavalry..." ("...7m0A..xaiteiyvtangwvolynA..." BJ III,126.) formed part of the rearguard. For a more detailed breakdownof Vespasian's army according to the evidence of Josephus, seeGichon (op. cit., 303-8).194This figure does not include any regular auxiliaries thatmay have been included in the advance guard as well.1950p. cit., 291.196BG II, 17 and 19; BC III, 75.65modern armies. Evidence for this tactic is absent from all ancientsources197, as is the essential provision of flank guards.Gichon198 suggests that this omission from the sources may beexplained because it seemed self-evident and not worth mentioning.Josephus, like Caesar and Arrian, was a general, and so hisinterest in military matters can be expected to spill over into hiswriting. On the other hand, he was writing for a wider audience,who could be expected to have little interest in this degree ofdetail. This will explain why he omits some information ofinterest to historians and soldiers in the interest of producing areadable text. However, even without explicit reference tothem, it should be possible to establish some information about theadvance and flank guards. Their sheer number, and the speed withwhich they could cover ground compared to the remainder of theforce, suggests that they probably both operated far in front ofthe main body and, even if echeloned in depth, covered a greatlateral distance.Gichonm calculates the overall length of Vespasian's forceas between 30 and 35 kilometres.^The conclusion is obvious:...the head of the troops entered camp for their overnight rest,197Except for Arrian; see pp. 70-72 infra.198Op. cit., 290-1.199For example, Josephus, when describing the marching order ofthe legions, says simply that "a centurion" ( "... ucticateArcapxoc..."BJ III, 124) kept the legionaries in order; at six thousand men perlegion, I think more than one would be necessary.aw0-.p cit., 307.66before the last of the troops were able to leave the site of theprevious overnight-stay... "a". The length of an average day'smarch for a Roman army has been calculated202 at a little over 30kilometres. The army would leave camp around dawn203 and marchuntil about midday. Probabably the antecursores would leaveshortly before the remainder of the force in order to leave a largeenough space204.The enormous length of the train consisting not only ofsoldiers but of equipment, pack animals, and baggage as well willrequire it to have considerable protection along its entire length.Their number suggests that the antecursores did not merely scoutand clear forward of the main body but well off to the flanks aswell, thus providing flank protection. But what sort of forces arerequired to perform such a task? Unless the ground ahead isthickly wooded or full of defiles, clearing it and guarding it isa relatively simple matter of searching possible enemy ambush sitesand occupying locations dominating the march route and approachesto it. Small parties can search large areas, and a force of a fewhundred would be sufficient to hold almost any key point at least201 ibid.NmH.P. Judson, in Caesar's Army (Boston 1888) 67, arrived atthis figure with reference to American army staff tables.203Josephus, BI V, 51.204There must be a space between the advance guard and the mainbody so that both elements will have space for manoeuvre in case ofcontact with the enemy.67temporarilym. Josephus provides us with an example of a limitedcavalry force, though not specifically identified as composed ofhorse archers, conducting reconnaissance: Titus in 70 scoutedJerusalem with a force of six hundred206. Such a use of smallparties is more economical and effective in clearing a route thanemploying large forces, which are more difficult to manoeuvre aswell. The more sub-units into which an advance guard can bedivided, the more ground it can cover.If an advance guard was organized along these or similar lineswe might account for the lack of references to flank guards in thesurviving literature as follows. The advance guard would coversuch a wide area that it would take on the responsibilities offlank guard by clearing and occupying areas far to either side ofthe march route. Flank guards per se would therefore not exist inthe Roman army as they would be unnecessary. The job that theyperform in more modern armies would be done by the advance guard.An enemy attack through the latter would be unlikely given the timerequired to penetrate them. Indeed, the large numbers of troopsthat we have seen preceding the Roman armies in response to theJewish revolts would require an enormous area to either side of the205E. g. Suetonius, Divus Iulius 68, 3, where an infantry cohortholed up in a turret held four legions at bay for a few hours,despite having 130,000 arrows shot at them.aldi.RAJ V, 52-65. The six hundred tmAirmw buiwv_u areobviously forming a bodyguard for Titus. Were this a lessimportant reconnaissance (i.e. a route reconnaissance, which acommander does not normally carry out) the number would likely bemuch smaller.68main body unless they were grouped in several echelons. Such anorganization in echelons, however, would be clumsy and dangerouswithout provision for wide reconnaissance.One point to be kept in mind is that cavalry are suited forscouting and skirmish; conversely, they are useless at capturingand holding ground207. Even foot archers are vulnerable whenfaced with an attack by determined infantry. Germanicus coveredthem with auxiliary infantry at Idistavisom for this veryreason, and Arrian planned to protect his foot archers with theinfantry of a cohors /ta/ica209.The mobility of horse archers and their inability to holdground seem to support the view that they would form part of theantecursores or primi antecursores rather than a picket force.Their characteristics make them more suited to reconnaissance ofselected locations and pinning any enemy there until the arrival ofan attack force proper. The speed with which they were capable ofmoving suggests that they would not have travelled as part of aforce including infantryvo. Unfortunately, Josephus does not207This point is axiomatic and a truism for the student of war;cf. Caesar, BG VII, 80, 7.208Tac, Ann. I, 16.209.Arrian, "Eicugtgicaul'AXemov 13.mArrian (see pp. 69-71 infra) does not include any infantryin his advance guard; rather, the cohortes equitatae even seem tobe divided, the cavalry in the advance guard, the infantryfollowing with the remainder of the auxiliaries. Interestingly,the cohortes equitatae are divided for battles as well; for more,see Davies, "Cohortes Equitatae." in Service in the Roman Army (NewYork 1989) 141-152.69provide more explicit information on this subject.Our most informative source for the actual deployment of horsearchers in the field is our second major authority, the secondcentury Roman writer and general Arrian. While commanding the armyin Cappadocia in 135, Arrian issued a set of orders to his army fordealing with an attack into the province by the Alani. His orderof march against the Alani seems to have included horse archers, inan arrangement similar to that supposed for Vespasian's army, as aprotective guard in front of the rest of the advance guarel. Heactually says that the horse archers are behind the forwardreconnaissance screen, in two groups under the command ofdecurions:`Hydaeou Rev TIN than; oval-Lac TC4); icoctaalcOnovcflat:ea; êiri Svoiv tecawivouc av t oiniop hyr.p.Ovt.toinotc St to); innoto4enac 'CO'66 Ilerpaiouc, ica toótouç tni,Svoiv. ecy•fprcow St akoi); o SencSecpxat. tni^toircot;trozroixecov oi anO tfjç en.% finvt Abinocvoi 6voga.ovvrenixecov abtoic oi %fig =tip% Tric tealinn; tv 'Panay,fç amccov Mohr% Kopivetoc. tni tokotc St of. dcne ç ERIN fiOvolux Katovoi. auv<teyrecxecov St ainotg Trupaiot Kai KypivaiotKai oi ecnO ç nOrric Tatruci16. augntivrow St wimpy dcpxtuoAnuirptoc. iti TO6T01.; St ot Kat01: 1,7C71Eig, icat abtoi tni 6o, Kaiwimpy inTia0co ticomentrapxoc, 15anep èiviowatoniSou. (licta4t;'Kate( 'Ail.avow 1-2)"The mounted reconnaissance troops are tolead the entire army grouped in twoformations, each under its own commander.Following these will be the Petraian horsearchers, also in two formations under thecommand of decurions. Following them are tobe grouped the men of the ala Auriana. WithmArrian, "Ertcc4t; -Kat& 'Abivoyv 1.70them are to be deployed are the troops of thecohors 1111 Raetorum under the command ofDaphnes of Corinth. Following these will bethe men of the ala Colonorum; let the Ituraeiand the Cyrenaians and the men of the cohors IRaetorum be deployed with them. Demetrios isto command all of them. The Celtic cavalryare to follow them, also in two formationsunder the command of the centurion who is incharge of their camp."Clearly the horse archers follow the icataaminot to protect themand cover any deployment of the main force or the advanceguard212. Does the phrase bd. tcrinot; imply that the Petraians areright behind the iconaminot, or that they follow at some distance?The same phrase when used later in the same passage seems toindicate that the Colonoi follow closely the Raeti. Do theinnovgama take on any of the reconnaissance tasks? Again, Arrianis not completely clear, but their position close behind the212A problem lies in the chain of command and deployment. WhenArrian says that the decurions are in charge, does this mean incharge of individual turmae? Or does he mean of the two divisions,and would decurions, normally commanders of turmae, be put incharge of such larger forces? The Ilerpcciovc have been identified asthe cohors III Ulpia Petraiorummilliaria (see E. Birley, quoted inJ. Davies, "Roman Arrowheads from Dinorben and the Sagittarii ofthe Roman Army" Britannia 8 (1977) 269; and R. Davies, op. cit.,143.). If this identification is correct then there were up to 125eguites (assuming a strength of approximately 250 for the equitescohortales of a milliary cohort) in each of the two groups intowhich the lead mounted archers are divided, or about three or fourturmae. Given the rigid hierarchy within the rank centurio, it isreasonable to suppose a similar system for decuriones. Having thesenior decurio in each group in charge would then not seem sostrange. Arrian is telling the tribunus cohortis that he wants himback ready to command the rest of his soldiers if contact is made,and to leave the less demanding task of commanding the cavalrytroops to his junior officers.71reconnaissance troopsm would indicate that they are there forprotection of these tconotaicentot; they could perhaps have taken partif a reconnaissance in force was required, but there is noindication of this.What is the organization of the rest of the advance guard?Arrian does not describe it in detail, but it is reasonable tospeculate that the other mounted units that follow the Petraians -the a/a Auriana, the equites of the cohors IV Raetorum, the alaColonorum, and finally the equites of the cohortes I (?) Ituraeorum(sagittariorum), III Cyrenaica (sagittariorum), and I Raetorum - doso in a body, ready to deploy where they are needed214. As to whyoiWomiinmeic follow these last in two detachments, it may have hadsomething to do with the fact that the horse of the four cohortesequitatae which they precede are working as flank guards. Thetotal number of horse in the advance guard would be around 2000 -an effective striking force in any situationm.Apart from the references in Josephus and Arrian discussedabove, there is little other evidence for mounted archersparticipating in an advance guard. Caesarm used mounted213A numerus exploratorum: Ritterling (1902), quoted in theTeubner edition of Arrian (Vol. II, Scripta Minora, A.G. Roos, ed.[Leipzig 1968] 177).214The fact that they are under the command of a singleindividual (Daphnes of Corinth) may support this supposition.2151 partly follow R. Davies in this paragraph; see op. cit.,143.216Caesar, BG II, 1972reconnaissance detachments sent far in front of the main body ofhis armies. Ammianus Marcellinus also makes passing reference tosuch forces217. Of course equites sagittarii are not the onlymilitary arm suitable for such a role; any light armed cavalry willdo, but those armed with the bow have the advantage of range, soimportant in a skirmish, which is, after all, a hit-and-run affair.But, as we have seen, the evidence of mounted archers in theadvance to contact points to their use as a fighting force.The evidence of Josephus and Arrian considered thus far pointsto a fighting role for mounted archers in the advance to contact,but furnishes no evidence of their actually fighting. Evidence forthis is available only in an oblique form through Arrian's accountof the battle which occurred between the forces of Alexander theGreat and the Indian king Porus upon Alexander's crossing of theriver Hydaspes. Our main source for this battle is book five ofArrian's ',04.1)(cOpm)%1Nat pocatc218. Admittedly, Alexander was not aRoman, but Arrian, being a Roman general as well as an historian,was in a unique position to analyze and comment, and it isreasonable to suppose that his interpretation of events was basedin part on his own personal experience.2'17E.g. Amm. Marc. 31, 12, -for exploratores and 27, 2 forspeculatores. Both these references are to reconnaissancedetachments operating in front of the main body; for a detaileddiscussion of such operations in Ammianus, see N. J. E. Austin,Ammianus on Warfare, Collection Latomus 165 (Brussels 1979) 117-139.218Dioodorus (XVII 87-91) also has an account of the battle butit is extremely sketchy and omits the river Crossing entirely.73Sg innoto4Otac it; irticr% trnou npoetate....TaiNxovi Se te) to4cpxti irpocytta4c toO6 to4Otac enartvbump._ Kai gni toiyrouc tà v npika tiatip.wat 'A)14av8povpin(); AiyEt tok irnot64otag, akOv 8g olyetv tokitpoaayetv yap oilleilvat 116pov 4iw tti that! &wallet... 66 OgKatOakv &pad; to ickfieoc TO tin/^evtaii0a OVcogbruccaelv akoic 41:w tij ap.4). akeyv taro?... ('Aveliccatc'AAE4dcv8poi) v, 13 - 15)"(Alexander) stationed the horse archersin front of the entire force of cavalry....(When he advanced) he ordered Tauron, thearcher-commander, to lead them against thecavalry.... (Ptolemy) says that Alexanderfirst sent his horse archers against these(enemy that had appeared), and that he wasbringing up the rest of the cavalry, since hebelieved that Porus was pressing forward withhis entire force.... When he learnedaccurately the strength of the Indians,immediately he made a violent attack on themwith the cavalry that was with him..."After crossing the river Alexander arrayed his battle line asfollows: various infantry on the left and bringing up the rear,his cavalry on the right, his archers and other missile-armedtroops on either flank, and his mounted archers in front of hisentire force of cavalrym. He advanced his cavalry and light-armed foot quicklym; his infantry followed more slowly. Firstcontact was made with a force of chariots and possibly cavalry.Though unsure about the strength of this force, Arrian reports thatAlexander sent his cavalry against it, led by his horse archers,and destroyed it. In the above passage, Arrian seems to suggestthat Alexander's horse archers were used to pin a large opposingm'Aveeriam; —v (Vol. I, Anabasis. A.G. Roos, ed. 1967: Teubner[and all further references unless otherwise indicated]), 13, 4.22014, 1.74force before the battle. Evidently Alexander wished to destroythis force in detail to prevent their returning to Porus withinformation on the attacking force, or to prevent them from beinga problem in the upcoming main battle. In either case, the horsearchers are clearly not believed by Arrian or his source Ptolemy tohave been able to defeat such an enemy in detail or quickly, andfor this reason they were replaced in the attack by the remainderof the cavalry, a force which was capable of accomplishing this.Alexander seems to have intended that by engaging his mountedarchers he would gain time for his infantry to catch up and moveinto battle formation. It became obvious, however, that it was notnecessary to commit his entire army, so he sent just the cavalry.Alexander himself was at this point countering the mobilit y221 ofthe chariots with the far greater mobility and firepower of hismounted archers. He could have used any of his cavalry as aforward guard, but his employment of horse archers for this rolesuggests that they were only lightly armed for melee purposes222.Alexander thus employed his horse archers in this battle first asa covering force, to pin down the Indian vanguard before effectingtheir destruction in detail by regular cavalry.To conclude on the role of horse archers in the advance tocontact, the only explicit evidence we have for their use in the221Not so great, as it turned out; the chariots got stuck inthe mud (15, 2).222.A covering force pins its enemy most effectively by notbecoming involved in a pitched battle (and thus getting killed); itdelays and deceives the enemy by hit and run tactics: mobility andfirepower.75Roman army is in Arrian and Josephus. In both authors they seem toform part of the very first element of the advance guard andperform there a role of skirmishing and perhaps reconnaissance tolocate and pin down small enemy forces. The advance guard orantecursores, according to the evidence of Josephus, probablycovered a wide enough area to protect the main body from flankattack. Although from very different circumstances, thecomparative evidence of Alexander the Great's use of horse archerscomes from Arrian's pen and seems to corroborate the evidence ofRoman times. Horse archers could also make up part of a largerfighting force advancing in front of the main body.2. ATTACKThe next phase of war we turn to is the attack. If a Romanarmy meets an enemy force the commander can either decide to stopand fortify, to withdraw, or to attack. In this section we shalllook at the role of the horse archer in the attack.We turn again to Arrian's account of Alexander's final battlewith the Indian king Porus."11811tc vrOc fieXouc tyiyvero icai t4ticev tri, to ictpactO 66vvilov Thy ivribv toi); i7riroto4errac, Ovrocc tc ,caiovc, thgtapgat TO6C Torkti t.earidecac Tib V 7COX.E4LiOni It] NurvOriti 'ccT,65v to4eugaixov icat T6jv triton/ It brelAican. icat a*.ctc St TO1‘);ttocipovc Exow toi); t7téaçnapfiXavvev 64toK tri tO 66vugovTiOv pocpriipow...CAvalkcat; 'AX4dcv6pou v, 16)"As he was already within range Alexandersent his horse archers, a force of about a76thousand, against the Indians' left wing, inorder to disorder those of the enemy arrayedthere by the storm of arrows and the horses'charge. He himself, with the cavalry of the"Companions", violently charged the Indians'left wing..."After advancing on the Indian army, Alexander kept his armyout of range until the infantry caught up, and began a series ofmanoeuvres which both gave time for the infantry to rest anddeceived Porus about his intentions for the battle223. Alexanderattacked the Indian left flank with his cavalry, having ordered hisinfantry commanders not to attack until the enemy was plainly inconfusion from the Macedonian cavalry attack.Horse archers once again led the attack on their enemy. Theirpurpose was plainly not to destroy the Indians in detail but todamage their morale and command structure, thus paving the way forthe main attack by the regular cavalry224. Arrian says thatAlexander planned to hit the Indians once they were thrown intoconfusion, and since up to that point only his horse archers hadbeen in contact, it is only logical to see them as the agents ofthe Indians' confusionm Arrian's phrasing suggests certainthings about Alexander's employment of horse archers in thismArrian is explicit about the first point, but the second isimplicit from his description in 16, 1.224It is not entirely clear from the text whether Arrianbelieved that the horse archers were attacking infantry or cavalry.I believe that Arrian means "cavalry" here, but whichever arm theywere attacking, the conclusions are the same.mIt is preferable to regard them as agents for confusing theinfantry on the flank, as it makes little sense to weaken the enemyat one point and then attack elsewhere.77battle. That the Indians would be confused by the "density" of thearrows suggests volley fire rather than individual sniping; it alsosuggests a formation compact enough to produce a dense storm ofarrows. The fact that the charging of the horses would also havean effect also suggests a somewhat compact formationm. Such aformation would facilitate getting out of the way of the remainderof the cavalry when it charged immediately following the horsearchers' attack227.In the next phase the main infantry forces joined battle andthe cavalry fought a somewhat detached fight, although the Indianswere driven back upon their own forces. Little can be made of thepart of horse archers here except that, as their casualties werelightm and it seems that they were at best lightly armed withmelee weaponsm, they likely played at most a supporting role.According to the evidence of Arrian, therefore, Alexanderused his horse archers to throw the Indian force into disorderbefore charging it with his regular cavalry. It seems that theMacedonian horse archers fought from a distance and avoided closecombat, and that they employed volley fire and formations denseMu Somewhat" compact because if the horsemen were too closetogether they would not have been able to shoot without hitting themen in front of them; for this very effect, see Procopius, BellumGothicum: V, 27, 47.2271.e. in order to exploit the effects of the arrows best,just as Roman legionary infantry charged immediately upon throwingtheir Iona; and as modern infantry try to attack at the earliestpossible time following cessation of artillery fire.nEiTen killed from an original force of a thousand: V, 18, 3.229see above, Chapter one, p. 31, under Other Equipment.78enough to wreak significant damage on their target.The role of horse archers in the attack may be furtherillustrated by Tacitus's account of the battle of Idistavisomwhere both mounted and unmounted archers played an important role.noster exercitus sic incessit:auxiliares Galli Germanique in fronte, postquos pedites sagittarii; dein quattuorlegiones et cum duabus praetoriis cohortibusac delecto equite Caesar; exim totidem aliaelegi ones et levis armatura cum equitesagittario ceteraeque sociorum cohortes....Visis Cheruscorum catervis, quae per ferociamproruperant, validissimos equitum incurrerelatus, Stertinium cum ceteris turmiscircumgredi iubet, ipse in temporeadfuturus.... praemissus eques postremos aclatera impulit. ...Arminius...incubueratquesagittariis, illa rupturus, ni RaetorumVindelicorumque et Gallicae cohortes signaobiecissent... (Tac. Ann. II, 16-17)"Our forces advanced in the followingorder: Gallic and German auxiliaries infront, followed by foot archers; next camefour legions with Caesar (Germanicus),accompanied by two Praetorian cohorts andpicked cavalry; following them, an equalnumber of other legions, light-armed troops,the mounted archers, and the rest of theallied cohorts.... Having seen groups ofCherusci which had rashly begun their attack,Germanicus ordered his elite cavalry to attacktheir flank, and Stertinius with the remainderof the cavalry troops to circle around to therear. He himself would follow at the righttime.... The cavalry that had been sentforward attacked the (enemy) flanks and rear....Arminius ...had pressed his attack on thearchers, and would have routed them if theRaetian, Vindelician, and Gallic cohorts hadnot stood in his way..."amTac. Ann. II, 16. The text is the Oxford Classical Text,C.D. Fisher, ed. (Oxford 1985).79Foot archers, covered by auxiliary infantry, led the Romanarmy against the Germans drawn up on the plain next to the riverWeser. One German tribe, the Cherusci, failed to maintaindiscipline and began the battle by attacking the auxiliary and bow-armed advance guard231. These auxiliaries fought back, supportedby Germanicus's picked cavalry force which attacked the Germanflank232.Germanicus also sent the "remainder" of his cavalry underStertinius around to attack the enemy's rear. Tacitus, however,does not make clear whether his horse archers are included. As noother mounted troops are explicitly mentioned, it seems best totake them as forming part of this force. The auxiliary force atthe rear of the Roman legions was clearly the rear guard and it isunlikely that Germanicus would have let himself be exposed tosurprise attack from this directionm. Moreover, since Tacitus231Saddington (op. cit., 92: note 26) unaccountablymisrepresents Furneaux (The Annals of Tacitus. 2nd. ed. [Oxford1907] ) as taking the sagittarii here as the equite sagittario ofII, 16. Furneaux says in fact that they are pedites sagittarii.222 Furneaux (ad loc.) evidently believed that the archers andthe Gallic and German auxiliary force deployed into line upon beingattacked, with the archers on the right. This is an overlymechanistic view of Roman tactics and battle drill. Clearly forthe archer commander to deploy his troops between his coveringforce and the enemy would have been insanity. It is not whathappened anyway. Since the "...Raetorum Vindelicorum et Gallicaecohortes signa obiecissent..." and prevented the Germans frombreaking through, they must have been in front of the archers.MA provision probably resulting from his experience ofcampaigning in Germania, with its natives' propensity for guerillawarfare (as he encountered following this battle [Tac. Ann. II,19]), and since he had already received one attack from the flank,it is likely he would be on his guard for others.80makes no further mention of the rear guard, it seems most likelythat they stayed at the rear and formed a reserve, taking no partin the battle until perhaps the very end.Thus the evidence of Arrian and Tacitus indicates that thefunction of mounted archers in the attack was to support otherforces. Arrian's account of Alexander's army suggests that mountedarchers would support an attack by preceding it and "softening up"the enemy. The battle of Idistaviso as recounted by Tacitus givesan example of how mounted archers could be used with other cavalryto relieve the pressure on a force under attack by engaging theenemy from another direction. In this use they are littledifferent from other cavalry, but as their exact function is notrecorded, we cannot be certain whether they preceded other forms ofcavalry in the flank attack, as was the case in Arrian's account ofAlexander's battle against Porus's Indians.3. PURSUITThe next phase of war is the pursuit. Unfortunately there isno direct evidence for actual actions of mounted archers in thepursuit, but Arrian, in the 1Nma4t; xata'Aulavani, provides an explicitdescription of their intended role in this phase.TO & tianlaiv N.ucccv Kat& eDtac xai A,Oxov; Oxite4vvrerayggvov t4CCIT6C1X0 VA; 14*, to gtv 'PA; xipaatvtxatepat,c, npol3arjv Exov toi); &aim; np$5 xai toi);to4errac, A,Oxot •51)o, TO 8i Tij giom 44.Xayyt, A,Oxot E4 * fillit0V.TIZIKYCCOV Sg /Sam Rev ix/cc/T*1m 700)61.0V tfic 050=nm;81g4)EaTrpcgtoxrav, cbc ôirEptocSav Ikcip a*rfic. Oaot 8gX01704)Opot f icovrapOpot f gaxatpoOpot f zaacoOpot eic telcrdluiyuit icaTipokev...ecncoaeevrov 8g ei giv 4rn ?Lam&ygwycat, Staxaveiv M AZ; zeCticecc Tec4Etc icat 71;FACC6VELV toi);newrac toi); A,Oxou; (kW TO1‘); iplacac. cenixEloct 8gnixinou; awe; -Kai /yawl, bratiacmatv. toi); 8g •IllaucVaal; EncaOat jtv TOiC gnEX.oviwouatv, ev vi4Et 8g Kai=wad rti 8uget xpwivouc, 66 ei Rev (Puri1 icaprrEpec icatixot,gic8e4aaeat cv npforriv 8iogtv docgittotc TOT.; tn./wig, ci 8g TLCbacrcpainj icatoacciAtivot, enttiecaket toic einorp4ovatv...("Erca4; Iona 'Maven/ 20-21, 27-28)"The entire cavalry force, organized by theregiments and the eight squadrons2254, is tosupport the infantry. Part, consisting of twosquadrons, (having formed up) on each flank,is to keep the infantry and the archers infront of itself as a guard; the other part,consisting of six squadrons, (is to form up)at the middle of the infantry battle-line....All of the mounted archers are to form up nearthe infantry battle-line, so they can fireover it; while all of the lance-, pike-,sabre-, and axe-bearing (cavalry) will form upon each flank... If when the enemy have beenrepulsed and their flight should becomeapparent, not all, but only half of thecavalry turmae must move through the infantrybattle-line and charge on; whichever troopsdrive through first are to form up and charge.The other half is not to make a wholeheartedcharge but must follow the pursuers information, so that if, on the one hand, astrong retreat should ensue, (those following)might take over the lead in the pursuit withtheir horses fresh; while, on the other hand,if a reversal should occur, they might attack(the enemy) who are forcing the retreat..."234The "iancoug ?oath" seem to be the cavalry of the cohortesequitatae accompanying Arrian's army. Although there were nine ofthese cohorts, one (the cohors III Cyrenaica) is only present as avexillatio, and its cavalry seems to be grouped with the cohorsIturaeorum (A.B. Bosworth, "Arrian and the Alani" Harvard Studies in Classical Philology 81 [1977] 249-50).82Apparently only the Xáxot of immycogama, of which there werefivem, were to take part in the pursuit. The remainder of thecavalry were drawn up on the flanks, and the pursuit force wasclearly instructed to pass through the infantry. The other cavalrywere instructed to wait for a signalm, presumably to joinbattle, while during the battle itself the inircragozat would provideonly fire support. The iniccrth4grat thus seem to have been employedprimarily as pursuers. The horse archers were divided into twoparts for the pursuit in order that some would remain fresh ifneeded, one half to maintain close contact with the enemy, and theother to remain properly arrayed in formation to take over thepursuit with fresh horses when the occasion demanded it. Thoseactively pursuing were expected to lose their formation when doingso, and thus might have been more easily put to flight in theirdisorganized state.Arrian's plan for the conduct of the pursuit made full use ofthe mobility and firepower of mounted archers, and also providedfor possible reverses or problems. The fact that only theinnougamnwere to pursue the enemy is significant, since they were the onlyforce capable of maintaining pressure on the enemy while still notcoming into close contact with him. If the enemy were to turnabout and fight, the pursuers were still separated from them, and,owing to their lack of encumbering equipment, could outdistance the235See Bosworth (op. cit.) 237..6iXintlX0V rt -Kai TO 4iivertta npocrixevOvrav. . . " 21.83enemy; moreover, they could harass the enemy even if put to flightthemselves. They also had a backup force for relief should thishappen. If the enemy should turn about in strength, they coulddelay his approach to the infantry line and thus give it a chanceto re-form or redeploy.Tacitus237 provides one possible example of mounted bowmenpursuing a defeated enemy in an action that may be more properlytermed a rout....quidam turpi fuga in summa arborumnisi ramisque se occultantes admotissagittariis per ludibrium figebantur, aliosprorutae arbores adflixere. (Tac. Ann. II,17)...(after the battle of Idistaviso) Somewho had shamefully fled by climbing to thetops of trees and hiding amongst the brancheswere shot as a lark by archers. Others weredashed to the ground when the trees werefelled."Tacitus does not specify to which force these archersbelonged. Archers mounted on horses, however, certainly had themobility with which to hunt down a scattered enemy, and sinceTacitus informs us that the forest floor was clear beneath highbranchesm, cavalry could have operated there. It thereforeseems plausible that the sagittarii who hunted down and killedGerman fugitives following the battle of Idistaviso were mounted,237Tac. Ann. II, 17-18.2381, ...editis in altum ramis et pura humo inter arborumtruncos..." Tac. Ann. II, 1684probably the same ones that took part in Stertinius's flank attackearlier in the battlem.The evidence from Arrian and possibly Tacitus, therefore,points to an important role for equites sagittarii as a pursuitforce following a successful battle. Their horses gave them thespeed to run down infantry and other cavalry, and their bows gavethem the long range, accurate firepower with which to harass andkeep pressure on their enemy and still maintain the separation thattheir vulnerability made necessary. It is significant that in theone certain reference to their role in the pursuit they are theonly cavalry that pursue. Their characteristics obviously madethem best suited for this task.4. DEFENCEApart from one reference in Arrian there is no evidence for arole for mounted archers in the defence. Mounted archers aremobile, and lightly armouredm, and, like other cavalry, areunable to hold ground against a determined infantry assault. Theycan, however, contribute firepower. This is exactly the role thatArrian envisaged for them in support of his infantry battle lineagainst the Alans. Arrian seems to have intended that the storm ofarrows from his archers would provide his main line of defence239see pp.79-80 supra.240See above, chapter one, pp. 35-7, under Protective Clothing.85against the armoured Alan cavalr^He placed his innoAamabehind the line of infantry "...so that they can fire overit...1,242. They are engaged in providing missile support to theinfantry along with foot-archers to the rear and flanks andartillery to the flanksm. However, since Arrian has irECoitogencctimmediately behind the infantry, the mounted archers seem almost tobe an afterthought. The foot-archers are presumably placed wherethey are with the task of engaging the enemy, as they have noability to support a pursuit; their fire would be masked by theirown pursuing forces244. The horse archers are located behind theinfantry in order to make a speedy transition from defence topursuitm. Their height and weapon range merely gives them theability to assist with fire support.Apart from this passage from Arrian, there is no evidence,Roman or otherwise, for the employment of mounted archers in thedefence. Nor, given their lack of defensive armament, can a commonactive role be readily envisaged.5. WITHDRAWALN°See Bosworth (op. cit.) 236-7.242-Extot4; nevi ,Axa' vwv 21.mTragu; icauk 'Aakirvwv 18-19.24Although the cohortes NUmidarum, Cyrenaica, Bosporanorum,and Ituraeorum are formed up behind the infantry (1Nmgtgicaui'MAvwv18), only the archers formed up on the right flank and covered bythe cohors Italica are ordered to support the pursuit (29).245 27.86The final phase of war is the withdrawal. An active role formounted bowmen in this phase can be envisaged but unfortunatelythere is no evidence for it from the high Empire. There is,however, a brief mention in Procopiusm about mounted archersconducting a feigned withdrawal, from which some information may begleaned.In A.D. 537 Rome was under siege by the Ostrogoths. The Romancommander, Belisarius, encouraged by the arrival of reinforcements,decided to take the initiative against the besiegers. According toProcopius, he sent a force of two hundred horse archers to alocation near the enemy camp where they were to wait in full enemyview, and, if attacked, to hold off the enemy until out of arrowsand withdraw at full speed to Rome, where an ambush was set. This,according to Procopius, is precisely what happened, and "...notless than one thousand Goths are said to have died in thisaction.. ,247Does this passage, however, provide an accurate account ofevents? The Strategikon prescribes for Byzantine troops of thesixth or early seventh century quivers holding up to fortyarrows 248. Skilled archers can fire up to about fifteen arrowsper minute. At this rate, and assuming a similar ammunitionsupply, the Byzantine archers had between two and three minutes ofmBellum Gothicum V. 27.24711^Aiyovrat Si reneot ateaaov f iltot v 1:4 tpyq? wimp ducoeaveiv. "•Bellum Gothicum V, 27, 11.248Strategikon I, 2.87shooting before they exhausted their ammunition and were forced toretreat. The Byzantines would hardly have begun firing until theirtargets were in range249, but even a dismounted manm could havecovered the distance of a bowshot in less than two minutes. TheGoths were mounted31 and would therefore have closed with theByzantines well before the latter had run out of ammunition. Itthus seems more likely that the Byzantines withdrew before alltheir arrows had gone, and fired at their pursuers whileretreatingm, and in this way incited them to follow to theambush at the city wall.This tactic, the so-called "Parthian Shot", is known from thedefeat of Crassus by the Parthians at Carrhae in 53 B.C.253.There too, mounted archers turned and fired at their pursuers,causing considerable damage. Arabic archery manuals from themiddle ages also describe a variety of shots, some of them to therearm. Such a skill "...was common to all good horse-archersfrom the Scythians to the Crimean Tatars...pm. Although there249Less than 200 m; see chapter one, p. 22, under Archery.250Running at 10 mph, or 4.5 m/s: not a difficult speed, evenwhen burdened by equipment.251Be//um Gothicum V, 27, 9.mProcopius himself says in another context they are adept atdoing this : " . . .814mcovu5c; 're 13aaetv tai); 7COXEILiC/U; xai (peliyovrac . . " BellumGothicum I, 1, 14.mPlutarch, Crassus 24, G.mThe Ghunyah of Taybughah, cited in Coulston (op. cit.) 292.mIbid.88is no direct evidence of mounted archers in Roman service usingthis tactic, it is probable that they were well acquainted with it,and therefore were able to make use of it if necessary.Unfortunately no evidence survives from the high Empire ofRoman use of horse archers in an organized withdrawal. Thepreceding example from Procopius shows how bow-armed cavalry wasable to withdraw from a pursuing enemy while continuing to engagehim. It is possible that a retreating army would use their horsearchers in a similar manner to delay and disrupt the enemy pursuit.6. SUMMARYThe equites sagittarii of the high Empire seem to have beenabove all a support arm on the battlefield. In the advance tocontact they seem to have played a very important part; theevidence of Josephus and Arrian indeed suggests that they formed alarge part of the advance guard and flank guard which was necessaryto protect the army from surprise attack while on the march. Afighting role is all that can be certainly deduced from theevidence, but pure reconnaissance would certainly not have beenbeyond their capabilities.When an army attacked, the comparative evidence of Arrian'saccount of Alexander suggests that mounted archers may have led theattack. Their job was to cause casualties and confusion among theenemy and to weaken them for the main attack which would soonfollow. If part of the main army was in difficulty, a commander89could send a force including horse archers to relieve them, ashappened at Idistaviso. The mobility of horse archers allowed themto pursue a withdrawing or scattered enemy most effectively, and tokeep them from regrouping. Arrian in his Ilicricicacci'ALMNwv providesdetailed evidence of a Roman commander's intended use of mountedarchers in the pursuit.Owing to their difficulty in holding ground, mounted archerscould contribute little to a static defence except their fire.Although one can imagine an important role for mounted archers inthe withdrawal because of their ability to manoeuvre and run fromharm's way while maintaining a steady fire upon their attackerswhile ammunition lasted, no direct evidence of such a role survivesfrom the high Empire. Only comparative evidence from the Byzantineperiod allows us to glimpse how the ancients employed them in thisphase of war.In conclusion, the employment of any arm depends on thesituation and on the experience and knowledge of the commander, andthis is especially true for the deployment of bow-armed cavalry.Mounted archers were clearly a support arm on the battlefield ofthe high Empire. Their light armour meant that they were toovulnerable to remain in close contact with an enemy for long or tohold up against a determined attack, but their mobility andfirepower enabled a hit-and-run role which was of greatest value inthe mobile phases of war, viz, the advance to contact, the attack,the pursuit, and the withdrawal. Judging from the evidence of thehistorians, they seem to have been an important arm, despite the90scarcity of evidence for their use in battle.CHAPTER FIVECONCLUSIONSThe equites sagittarii fought in the service of Rome from theearly Empire onwards. That they were an important arm of the Romanmilitary is evident from their numbers: twenty four units armedwith the bow can be certainly identified as being mounted, and atleast four other mounted units may have been sagittariae256 .Further discoveries of inscriptions, especially diplomas, may yetincrease this number.Although they were raised and to some extent trained in theeast, Rome's archers seem to have used much the same dress andequipment as other auxiliary units, differing mainly in theancillary equipment needed with bows. Their bows were of easterndesign. Compound and recurved, made of wood, sinew, bone, andglue, they were extremely powerful weapons, capable of punching aniron arrowhead through armour plating. The arrows fired were of anumber of different designs, each perhaps with its own specificuse; unfortunately the intended uses have not entirely come down tous today.Units of mounted archers seem to have ridden small, fastArabian horses, bred in the same areas where the skills of mountedarchery flourished. These horses were best suited to the type oflooser, faster warfare that the mounted archers engaged in. Theirsaddles and other equipment, however, seem to have been the same as36see Appendix one.92other auxiliary units used. When it was necessary to use weaponsother than the bow, they would probably have used standard issue aswell.As a military system, horse archers can be understood in termsof their characteristics: flexibility, mobility, vulnerability,and firepower. They were flexible enough to use differentequipment according to the situation, and even operate dismountedif necessary; mobile like all cavalry, and vulnerable, since theycould not use a shield; but above all they had the firepower oftheir bow and arrows. It is these four things that determinedexactly how horse archers were effective, and thus how theircommanders would use them.As mentioned above, the equites sagittarii seem to have hadthe flexibility to use different weapons if the situation demandedthem. Since they often formed provincial garrisons, they wouldhave to have had access to a variety of weapons in order to carryout the multitude of tasks required of them. Some inscriptionaland literary evidence points to this conclusion as well.As can be expected from cavalry, mounted archers seem to havebeen most used in the mobile phases of war: the advance tocontact, the attack, the pursuit, and the withdrawal. Of these,only their actions in the pursuit are little attested in literarysources. Horse archers seem to have had been of little use in thedefence, and may have been dismounted, especially for muraldefence. Horse archers used their combination of mobility and longrange, penetrative firepower to harass and soften up an enemy, or93pin him down until heavier forces could be brought to bear.This much is known about mounted archers, and much can besurmised, but many questions remain. The detailed history of thePalmyreni sagittarii is not clear. What was the sequence of eventsfrom their initial raising as a unit in Roman service to theestablishment of the cohors XX Palmyrenorum milliaria equitatasagittariorum? As stated above257, it seems unlikely that theharsh bit used by regular Roman auxiliary cavalry would be used byhorse archers as well. Did they use another form of bit?Furthermore, the military works of Vegetius and Arrian need up-to-date commentaries and translationsm These will facilitatefurther research into the study of all aspects of the Romanmilitary.The Roman military machine of the Empire is a natural objectof study because of its tremendous successes and failures, in whichthe equites sagittarii played a part. Although they were littlereferred to in literature, their part seems to have been animportant one.257See chapter one, pp. 28-9, under Cavalry Equipment.mThere is a recent English translation of the INmgt; by B.H.Bachrach (A History of Alans in the West [Minnesota 1973] 126-132;from Bosworth [op. cit.] 217), but apparently it is "... riddledwith errors and useless for historical interpretation." (ibid.).94APPENDIX 1BOW-ARMED UNITSIn 1977 J. L. Davies259 published Eric Birley's list of thebow-armed units of the Imperial Roman military of the first tothird centuries. I reproduce it here, together with a briefcommentary on some of the units known or suspected to have beenmounted.a. Alae:1. I Batavorum2. Celerum2613. I Hamiorum Syrorum (Mauretania Tingitania)4. I Augusta Ituraeorum259"Roman Arrowheads from Dinorben and the sagittarii of theRoman Army" in Britannia 8 (1977) 269-70.260This unit is only believed to be bow-armed from CIL III,3676, which however identifies it as the coh. I Batavorum. The alahas been placed at Razboieni in Dacia during the reign of Hadrian.See I. B. Cataniciu, The Evolution of Defence Works in Roman Dacia,BAR International Series 116 (Oxford 1981) 24, 30, and notes 200and 371.NOSpeidel ("The Roman Army in Arabia" Aufstieg und Niedergangder ROmischen Welt II, 8 [1977] 702-3) makes a case for the originof this unit as under Philip the Arab; apparently it originallyserved in Arabia but moved to Noricum sometime after the mid thirdcentury, as shown by a tombstone at Virunum.262This unit is only known to be bow-armed by the evidence ofCIL XVI, 99, and perhaps from a diploma of 154-161 (M. Roxan Roman Military Diplomas 1978-1984 [London 1985] no. 182). This unit mayhave been stationed at Micia in Dacia before the reign of Hadrian(Cataniciu [op. cit.] note 122).955. Parthorum et Araborum2636. (I Sag)ittariorum Surorum milliaria2647. I Thracum veteranorum2658. III Augusta Thracum sagittariorum (Pannonia)b. Cohortes:1. I Antiochensium2662. I Apamenorum equitata (Eg^) 2673. I Ascalonitarum equitata'4. I Bosporanorum equitata263The only reason to believe that this unit should be includedhere is a single sculpture of a mounted archer (fig. 18; for thetext, see AE 1959, 188: "...ala part(h)o(rum) et araborum...");but see chapter three and Kennedy (op. cit.) for a closerdiscussion of the problem of Parthian units in Roman service.2 See infra, note 25.266This unit has been tentatively identified from threediplomas of 154-161 (Roxan [op. cit] nos. 102, 103, and 110, p.168-70, 172-3, and 182). It seems possible, based on the numberthat are known to be composed of archers, that most units ofThracum were bow-armed. This is merely speculation, however.266This unit is only known to be bow-armed from a diploma of161. They may have some connection with the forces sent byAntiochus IV of Commagene to Vespasian's army in Galilee in 67; seep. 59 supra. This unit apparently constructed a stone camp atDrobeta in Dacia in the early part of the second century. Amongthe later occupants of this fort were the coh. I Sagittariorum.See Cataniciu (op. cit.) 11-12.NWCIL VI, 3654; and III, 600 (= ILS 2724).268This unit is only known to be bow-armed from CIL XVI, 106.This unit is mentioned elsewhere (in III, 600, for example) but isnowhere else called sag(ittariorum).269Evidence that this unit exists comes only fromTicra4t; 3 and18; Ritterling ("Zur Erklárung von Arrians 'DerAx4Kicatd'A),av6iv" WienerStudien 24 (1902) 363: quoted in Roos' Teubner edition of Arrian)believes that it is equitata, based on Arrian's wording (3. . . Bocriropavoi ag iti 'MA'AM; 140i ia•vrav. . . " , suggesting that the Boanoptavoihad falai; as well as infantry); he also believes that this unit ismilliary, a concllusion not supported by the text. I followRitterling in assigning it the ordinal I, keeping in mind Arrian's965. I Flavia Canathenorum milliariam6. I Flavia Chalcidenorum equitata2717. I Cilicumm8. II Classica9. II Flavia Commagenorum equitatam10. III Cyrenaica11. II Cyrrhestarum12. I Flavia Damascenorum milliaria equitatam13. I Hamiorum14. I milliaria Hemesenorum equitata civium Romanorum15. I Augusta Ituraeorum (Dacia Superior)16. I Ituraeorum (Mauretania Tingetana)habit in this work of mentioning a unit's designation only if it isother than the ordinal I. Either of the possible names(Bosporanorum or Bosporiana) can be supported by the text, and Ibelieve Birley hesitates to assign the unit a name for this reason;but an ala I Bosporanorum is known at Cristesti in Dacia (Cataniciu[op. cit.] 22, 24, and notes 187 and 371) in the reign of Hadrian,and so I assign the cohort the same name.mRoxan (op. cit.) nos. 51/104, p.174. Perhaps this unit isthe same as the cob. I Augusta Canathenorum equitata known to be inMotha, Arabia before 125. See Speidel (op. cit.) 709.mThis unit is known to be sagittaria only from CIL III, 6658.mThis unit was stationed in Moesia, and it is known to bebow-armed only from a diploma of 145/146. It has been, however,tentatively identified as such in a diploma of 112 (Roxan [op.cit.] 85, p.146). This unit is not to be confused with the cohorsI Flavia Cilicum equitata, which formed part of the garrison ofSyene. See Speidel (op. cit.) 785.mThis unit may have constructed and certainly garrisoned theearthwork fort at Micia in Dacia during the reign of Hadrian. SeeCataniciu (op. cit.) 14-15, 22, 43, and notes 184 and 380.mThis unit was stationed in-Germania Superior and was one ofthe few units of mounted archers to serve outside of desert areasor areas with a mounted archer threat. I believe that theplacement of this unit is evidence of a desire of Roman commandersto maintain a flexible force throughout the Empire. Beyond theneed for dry conditions for the construction and storage of bows,however, there seems to be no explanation why this would not holdtrue for Britannia, where we have no evidence whatsoever for unitsof mounted bowmen.9717. I Numidarum27518. XX Palmyrenorum milliaria equitatam19. III Ulpia Petraeorum milliaria equitata27720. I sagittariorum (Egypt)21. I sagittariorum (Germania Superior)22. I sagittariorum milliaria equitata (Dada)27823. I Aelia sagittariorum milliaria equitata (Pannonia Superior)24. I Ulpia sagittariorum equitata25. III sagittariorum26. I milliaria nova Surorum equitata (Pannonia Inferior)27. I Syrorum27928. II Syrorum milliaria equitata (Mauretania Tingitana) NWmRitterling finds it necessary to label this unit asequitata, apparently because of its placement with the Boanoptavoi inTircattc 18. Nothing in the text, however, supports this: footarchers from this unit are indeed placed in the battle line alongwith the foot archers from three cohortes equitatae, but when it ismentioned before, in INmgt; 3, there is no suggestion that it ismounted. As with the cohors Bosporiana (see above, note 10) Ifollow Ritterling in attributing to it the ordinal I.mThis fascinating cohort deserves a book all to itself. Thereader is referred to M.P. Speidel ("Europeans' - Syrian EliteTroops at Dura-Europos and Hatra" in Roman Army Studies 1[Amsterdam 1984] 301-309), and the rebuttal, D.B. Campbell ("WhatHappened at Hatra? The Problem of the Severan Siege Operations" inFreeman and Kennedy, eds. The Defence of the Roman and Byzantine East i, BAR International Series 297[i] [Oxford 1986] 51-8) for oneinteresting problem pertaining to the Palmyrenes.277The equites of this cohort were at the head of the regularunits of Arrian's army: INmgt; 1.mSee supra, note 8; they may also have occupied the forts ofTibiscus and Zavoi for a time (Cataniciu [op. cit.] 22). Mentionedas sagittaria in AE 1959, 311 but not as mill. This unit mayperhaps be identified with the -cohors I Cretum sagittarior(um)mentioned in a diploma of 110 from Porolissum: CIL XVI, 163.mThis unit is known to be bow-armed only from AE 1961, 358;and 1962, 304 (mil (es) coh(o)rt(is) Syro(r)um sagit(tariorum)).mThis unit is known to be bow-armed from CIL XVI, nos. 181(II Syror(um) saggit(ariorum)), 170, and 182; and perhaps a diplomafrom 109 (Roxan [op. cit.] no. 84, p. 144).9829. I Thracum sagittaria (Dacia Superior) 213130. III Thracum Syriaca equitata31. I Tyrorum (Dacia Inferior) N232. II Ulpia equitataC. Numeri:1. Hemesenorum2. Palmyrenorum Porolissensiumm3.^OsrhoenorumProbable Units284a. Alae:1. I Commagenorumm2. I Augusta Parthorum3.^SebastenorummEvidence of diplomas from Dacia shows that it was presentthere in the early second century, although its location has yet tobe identified. See Cataniciu (op. cit.) note 160. It was there atleast until 179, when it is attested in a diploma (Roxan [op. cit.]no. 123, pp. 196-7).282This unit is known to be bow-armed from Dacian diplomas (CILXVI, 1934 = ILS 2685; and AE 1962, 264: I Tyr(orum)sag(ittariorum)) but is as yet unlocated (Cataniciu [op. cit] 30,45).NO ...It is probable that the numerus PalmyrenorumPorolissensium c. R. at least contained horse archers because ofthe likely 3rd century formation of an a/a and a cohorsPalmyrenorum from it...." Coulston, [op. cit.] 285. This unit isgenerally identified as the Palmyreni sagittarii mentioned in earlysecond century diplomas of Dacia Superior (Cataniciu [op. cit.]note 181, and see J.C. Mann "The "Palmyrene" Diplomas" in Roxan[op. cit.] Appendix II, pp. 217-9).284J. Davies passes over the following probable alae andcohortes without comment. Presumably he regards them as bow-armedon the basis of their origin.285This unit is suspected to have been in Egypt and Nubia from83 to 96 (Speidel [op. cit.] 784-5).994.^Palmyrenorum286b. Cohortes:1.^I Chalcidenorum equitata2871. II Chalcidenorum2. I milliaria Hamiorum3. II Hamiorum4. I Sebastenorum5.^Seleuciensiumc. Numeri:1.^Surii Sagittariiam286This unit is my own addition to Birley's list. Evidencethat it may be bow-armed is the well-known function of its parent(the numerus Palmyrenorum Porolissensium) and sister unit (thecohors XX Palmyrenorum).287This unit is my own addition to Birley's list. It seemslikely that this unit was composed of archers in view of theinscription (ILS 9173) naming the Palmyrene Agrippa who was put incharge of a force of Palmyrene archers attached to this unit. Inthis I follow J. Carcopino, "Le Limes de Numidie et sa gardeSyrienne" in Syria 6 (1925) 119-20.amThis unit is my own addition to this list. Cataniciu (op.cit.) mentions such a unit in her index; at the pages she sites,however, there is only mentioned a "numerus Syrorum". This unitapparently formed part of the Flavian garrison of Romula along withthe cohors I Flavia Commagenorum. The origins of both of theseunits suggest that they were composed of archers. It may be arguedthat garrisoning a location with archers alone would giveinsufficient flexibility to deal with problems; Romula is, however,situated on an interior defence line in Dacia, closely flanked byother, differently armed units: the cohors II Flavia Bessorum atCincsor, the a/a I Asturum (which at least constructed the fort atHoghiz, and possibly was stationed there), and the cohors IINumidarum at Feldioara (Cataniciu [op. cit.] 30 and n. 72). Italso faces the Roxolanian Sarmatians to the southeast across theCarpathian mountains: this fact suggests that at least one of thetwo units was mounted. The a/a I sagittariorum Surorum milliariais only known from the early third century (R.W. Davies [op. cit.]269); it may be possible to suggest a link similar to that betweenthe numerus Palmyrenorum Porolissensium and the later cohors andala Palmyrenorum (see note 21 supra).100All the cohortes Commagenorum, Ituraeorumm, and UlpiaePet raeorum.289A cohors II Ituraeorum equitata is known to be in Nubia from99 until at least 131 (Spiedel [op. cit.] 786, 788).101APPENDIX 2 ANALYSIS OF ARROW FIRE IN TERMS OFTHE THEORY OF SMALL ARMS FIRE Frequently in ancient accounts of battle foot-archers andother missile troops such as slingers are spoken of as massed onthe flanks of the infantry290. Exactly why they were placed thereis never mentioned, but modern commentators have been quick toexplain their placement as a tactic to cover or protect the flankof the infantry formation291. It seems likely that this was notthe case, however, and that their placement is to be explained, notas any sort of a protective force, but simply as mass missilesupport for the main fighting force. Their exact location would bedetermined by how best they could take advantage of the effect ofthe fire of their weapons and the pattern that multiple missilesform when fired at the same target or a general area. There aregood reasons to question this explanation of the placement of foot-archers on the flanks. The vulnerability of archers to infantry iswell known today, and was certainly understood by the ancients292.290Most notably Arrian, Mgt; 12-14; also (Pseudo) Caesar, deBello Africo 60 and 81.mCoulston Hop. cit.] 292-0, however, mostly avoids analysisof reasons, and contents himself with a straightforward rehearsalof evidence.292the most striking examples are Caes. BG VII, 80, 7 (Gallicfoot archers were killed when they were attacked by Romaninfantry); and BC III, 93-4. Maurice's Strategikon (XII: trans.Dennis, 1984. 127-169) never prescribes that foot-archers beplaced where they are vulnerable to the enemy. They are always to102It is absurd that a body of troops would be protecting othersagainst troops to whom they are especially vulnerable. By theirvery nature they are most effective with their enemy at a distance.Their opponents understood this, and would attempt to close with abow-armed enemy as fast as possible in order to nullify the effectsof the arrowsmModern commanders analyze a situation, and decide how best toemploy their projectile weapons, by applying the theory of smallarms fire. This theory applies to all projectile weapons, ancientand modern, and will be a convenient means to understand whyweapons are employed as they are. The effects of weapons areobvious to one who fires them or directs their fire; these effectsare, however, not at all obvious to one who has not used theweapons. I will first describe the theory (see figs. 26 and 27).No weapon, no matter how accurate, can hit precisely the samespot with a missile every time it is fired, even if it is set on aperfectly unmoving base and fired at precisely the same point ofaim every time. This fact is due to many factors. Smallvariations in missile mass and aerodynamics, propulsion variations,and perturbations in the air over the flight path of the missileincrease or decrease range and deviation left or right of and aboveand below the target. The decrease in number of these variationsbe protected by other arms. Where they are not in or behind theinfantry or cavalry battle line, they are on the flank, but eventhere they are to be protected by heavy cavalry, or by heavyinfantry when acting as a flank guard for the cavalry (p. 144).mTac. Ann. VI, 35.103is inversely proportional to their magnitude; in other words, thelarger the deviation, the less likely the deviation is to occur.A large number of missiles fired from a single weapon at agiven point of aim, then, will tend to describe a cone, bent intothe parabola of the trajectory, with its axis the optimumtheoretical path from the weapon to the target. The intersectionof this cone with the ground is known as the beaten zone, and takesthe form of an ellipse with its long axis parallel to the line fromthe weapon to the target. Its total area is directly proportionalto the distance the missiles are fired.Factors affecting the shape of the beaten zone are the shapeof the ground on which the missiles fall, the aspect of the firerto the ground, and the angle at which the missile is travellingwhen it hits the ground. If any obstacles such as ditches, mounds,or walls lie in the beaten zone, they will create dead ground orground which is free from the missiles' impact. As well, if theground on which the missiles fall is tilted toward the firer, or ifthe firer is higher in elevation than the target, the beaten zonewill be smaller, less elongated, and denser in impacts; this typeof fire is known as plunging fire, since the missiles hit theground or their target at a high angle: they plunge down into it.Plunging fire is also created when missiles are fired at a highangle or at the extreme end of their range; in both cases themissile has lost most of its forward momentum and is dropping morethan moving forward.While the beaten zone is the area on the ground where the104missiles actually hit, the missiles' trajectory creates a muchlarger area where targets can be hit depending on their heightabove ground. That is, the missile is descending for a whilebefore it hits the ground; a standing human can be hit at anylocation where a missile descends below his height. For obviousreasons this is known as the dangerous zone and forms an ellipsethe same width as that of the beaten zone but much longer. Itsshape too is affected by the shape of the ground and the firer'saspect to the ground at his target.A single weapon firing a single shot at a single target isknown as a point weapon and its fire is known as point fire. Thisfire is used against individuals or single objects. When a weaponor group of weapons is fired at a broad target or area - a mass oftroops, for example, or the space on the ground between wallturrets - it becomes an area weapon and its fire area fire. Thisfire is used for harassment, attrition, or area denial and isrepresented by machine gun fire today.In order to maximize the effectiveness of the fire of alimited number of weapons or of an area weapon the firers placethemselves so that the long axis of their weapons' beaten zone iscoincident to the long axis of the target. In terms of battle,both ancient and modern, this means placing missile weapons on theflanks. Fire from the flank is known as enfilade fire.The value of plunging enfillade arrow fire against a battleline of armoured infantry lay in the fact that the infantry wererelatively well protected in the direction of the threat that is105perceived as the greatest, usually the front, by their shieldswhile parts of their arms, legs, and necks were bare when seen fromthe sides and from above. Arrows fired from a distance could notpenetrate the armour of a line of men directly but were certainlyable to wound if they dropped from the sky or hit the men from theflank. Arrows fired from close range, however, had the power topenetrate thicker armour; but if the enemy was within a hundredmetres or so the archers would have to spread out in order to givemore than just the men in the front the opportunity to shoot. Thearrows would have had to be fired at such a low trajectory thatonly the front men in a compact formation could fire, but a looseformation was extremely vulnerable if attacked by infantry. Thisfact required that the archers find cover294 or height above theirenemiesmWhen considering the effects of ancient weaponry it isabsolutely essential that one consider what it would actually belike to face these weapons in battle. Imagine a soldier in theranks in an armoured infantry formation before combat. All thatwould be visible to him would be the helmets, shields, and spearsof those around, and ranks of the enemy to the front seen asglimpses between his comrades' heads. The fear of every manincreases the closer he gets to the enemy ranks, and faltering or294Arrian (IN=4; 13) placed light infantry in front of hisarchers on the flanks to protect them.295horse archers could be used behind the line to fire over it;cf. Arrian ('amg14 21).106breaking ranks now would mean weakening the line and encouragingthe enemy. Consider the effect upon the soldier of arrows fallingthickly around him at this moment. He is surprised; the missileshave come from an enemy the soldier has probably not yet seen. Ifhe himself is not wounded, he has seen his comrades hurt, heardthem screaming, seen their blood, and watched them falling andbeing trampled by those behind them. This is an experiencesoldiers cannot be trained for but can only be inured to byexperience296 .Regardless of their experience, however, a sudden deadlymissile attack from an unseen enemy terrified and demoralisedsoldiers and made them vulnerable to the enemy to their front atprecisely the worst time. Such an attack would also havedevastating consequences if it was delivered during a crucialmoment after battle had been joined, causing either a loss ofmomentum for those getting the upper hand or the "final straw" forthose nearing rout.Archers give a commander the flexibility to launch this kindof attack at the most opportune time and place on the battlefield,and horse archers have even greater utility because of theirgreater speed, being able to move to wherever they are needed andthen depart at speed if threatened. Foot-archers, because of theirlocation on the ground, when firing their weapons en masse at alarge target such as an infantry battle formation, create plunging296Romans threw their pila and then immediately charged theirenemies with drawn swords to gain exactly this effect; before theenemy had time to recover from his shock they were upon him.107fire with a large dangerous zone. The large dangerous zone and itsalignment with the enemy's battle line make them most valuable onthe flanks.Thus it appears that the use of foot-archers and other missiletroops on the flanks can be explained by application of the theoryof small arms fire: the bow used en masse creates a pattern offire on the ground that is exploited most efficiently when it isfired from the flanks. The same theory is also a convenient way oflooking at the usefulness of any missile troops in any givensituation.108LIST OF FIGURES Fig. 1Fig. 2Fig. 3Fig. 4Fig. 5Fig. 6Fig. 7Fig. 8Figs. 9-10Fig. 11Fig. 12The Sassanid hold.^From W.F. Paterson "TheSassanids". Journal of the Society of ArcherAntiquaries 112 (1969) fig. 1.Different representations of the Sassanid hold fromPaterson (op. cit.) fig. 2.The parts of the compound bow. From Coulston 1985:fig. 1.The parts of the arrow. From Coulston (op. cit.)fig. 8.Roman flat-bladed tanged arrowheads. From Coulston(op. cit.) fig. 42.Roman trilobate tanged arrowheads. From Coulston(op. cit.) fig. 46.Roman arrowheads found at Bar Hill, Scotland. Notethe fire-arrowheads in the centre. From R.W.Davies, Service in the Roman Army. (New York 1989)Fig. 4.6.Arrow steles found at Dura-Europos. Note how thefletchings on the far left example extend all theway to the nock. From S. James, "Dura-Europos andthe Introduction of the Mongolian Release" RomanMilitary Equipment: the Accoutrements of War. BARInternational series #336. (Oxford 1987) fig. 4.A reconstructed Roman saddle in use. Note the widerange of movement possible. From P. Connolly "TheRoman Saddle" Roman Military Equipment: theAccoutrements of War PlatesSide view of a reconstructed Roman saddle. Notehow the "horns" seem to hold the rider in thesaddle. From A. Hyland: Equus: the Horse in the Roman World (London 1990) Plate 8.Tombstone of an archer of the cohors I Hamiorumfound at Housesteads. Note the strongly recurvedbow with set-back handle, and the unusual hook-shaped device in his right hand, and the long knifetucked into the belt on the archer's left side.From Coulston (op. cit.) fig. 26.109Fig. 13Fig. 14Fig. 15Fig. 16Fig. 17Fig. 18Fig. 19Fig. 20Fig. 21Fig. 22Detail of the bow in fig. 12. From Coulston (op.cit.) fig. 27.Bow detail from Trajan's column, scene CXV. Notethe strong recurve, and the unusual curling of theears. From Coulston (op. cit.) fig. 23.The tombstone of Monimus, an archer of the cohors IIturaeorum. His bowstring seems to be slacklystrung. From Davies (op. cit.) fig. 4.5.An auxiliary archer from Trajan's Column, sceneXXIV. His dress seems to be the same as that ofthe other auxiliaries on the column, but his bowhas the same strange shape seen in fig. 14. FromCoulston (op. cit.) fig. 19.A Palmyrene representation of caravan-escort gods.Both are carrying what appear to be combinationbow-cases and quivers slung from their saddles; theleft-hand rider seems to be holding a bow in hisleft hand. From Coulston (op. cit.) fig. 33.The tombstone of a member of the ala Parthorum etAraborum. The rider may be carrying a combinationbow-case and quiver but the representaton is toodamaged for certain identification. The horsearcher seems to be flocking three arrows on thestring at once! From Coulston (op. cit.) fig. 31.The tombstone of an archer of the a/a I AugustaIturaeorum from Gyor, Hungary. He is shooting at around target from which three arrows protrude.From Davies (op. cit.) fig. 4.3.Another member of the ala I Augusta Ituraeorum,from Tipasa, Algeria. From Coulston (op. cit.)fig. 32.The tombstone of Flavius Proculus, an equessingularis Augusti from Mainz. From Coulston (op.cit.) fig. 29.The sole surviving columnar representation of amounted archer. From the Marcus Column, sceneLVII. Note the lorica hamata. From Coulston (op.cit.) fig. 25.Figs. 23-24 Foot archers from Trajan's Column, scenes CXV andCVII respectively. Because of their unusual dressthey may not be regular auxiliaries but Sarmatiansin Roman service. From Coulston (op. cit.) figs.11020 and 22.Fig. 25 The^cone^of^fire.^From^Department^of NationalDefence Publication B-GL-317-019/PT-001 The LightMachine Gun 5.56 mm C9^(Ottawa 1987)^fig. 2-2.Fig. 26 The beaten zone. From idem,^fig.^2-4.Fig. 27 The dangerous zone. From idem,^fig.^2-6.1 11FIGURESFig. 11^2^3^4FIG. 21. After Morse, Additional Notes, Fig. 30. Sassanid c. A.D. 400.2. After Morse, A. & M. Methods, Fig. 57. Sassanid 5th cent. A.D.3. Ardashir I (?), from silver plate, Archaeological Museum, Teheran, item No. 7700. cf. Fig. I.4. Firoz 1 (A.D. 458-484), Hermitage Museum, Leningrad.Fig. 2112Fig. 31131Fig. 4ARROW TERMINOLOGYFig. 51144.6) Arrow-heads and bone terminals from composite woodand bone bows found in the well in the headquartersbuilding of the fort at Bar Hill on the Antonine Wall,Scotland.Fig. 71Fig. 8115116Fig. 11Fig. 12 Fig. 13117Fig. 16Fig. 17119.t^gtof '"77^E447t.Fig. 20Fig. 18Fig. 19120Fig. 21Fig. 22Fig. 23121Fig. 24Fig. 25: Cone of Fire122Fig. 26: Beaten ZoneFIRST CATCHBEATEN ZONEDANGEROUS ZONEDANGEROUSSPACEFig. 27: Dangerous Zone123GLOSSARYBecause this paper deals with a technical topic, in order toavoid confusion some definitions are necessary. I hope I am notinsulting the reader's intelligence but it occurs to me that thedistinction between, for example, "tactics" and "strategy" may notbe completely clear to someone without experience of the military.Other terms are certainly obscure and I include them here in ordernot to burden the chapters with constant definitions.Back: The generally convex surface of the bow facing the target.Belly: The surface of the bow facing the firer.Bodkin: A type of arrowhead without blades, with a square ortriangular cross-section.Braced: The bow, when strung; also called "at rest".Characteristic: an attribute which determines the effectiveness ofa weapon or weapon system in battle. As this paper will show,the characteristics of horse archers are flexibility,mobility, vulnerability, and firepower.Dustar: The "limb" or flexible section of the bow; between the"grip" and the "ear".Ear: The stiff outer section of the dustar; usually stiffened withhorn and designed to act as a lever. Also known as the"siyah".Fire: the application of weapons or ammunition to a target.Fires can be short- or long-range, and are usually coordinatedor timed for maximum effect. As this paper will show, onefunction of horse archers in the attack was to prepare theenemy with their fire before the main attack.124Fletching: The feathers attached at the base of the stele tostabilize the arrow in flight.Grip: The handle area in the middle of the bow where it is heldand the arrow is rested during firing.Guard: a protective force placed between one's army and the enemy.The job of the guard (be it advance, flank, or rear) is togain information and to delay, deceive, and destroy the enemy;in other words, a guard does its job by fighting.Knee: The recurved junction between the ear and the dustar.Lathe: The bone piece used to stiffen the ear; also refers to bonestiffeners for the grip.Loose: To release the string and fire the arrow.Nock: The notch on the back surface (facing the target) of theear. Also refers to the notch on the end of the arrow intowhich the string is fitted.Reflexed: The bow, unstrung.Screen: a reconnaissance force whose job is merely to locate theenemy and gain information about him. A screen does not fightexcept in self-defence.Stave: The bow itself, excluding the string.Stele: The shaft of the arrow, made from reed or wood.Strategy: the art of forcing upon the enemy one's own plan ofaction, usually by disposing armies. The allies' strategy in1944 was to create a third European front to divide the Germanarmies.Tactics: The art or skill of disposing men and machines on theground, especially in contact with the enemy. When confrontedby an enemy with a weak flank, a commander may decide toattack that flank. The flank attack is his tactic.Tang: A spike on the rear end of the arrowhead, designed to fitinto a hollowed out section of the stele.Trilobate: Formed of three "vanes" or blades.Weapon: a device with which to kill. A weapon may or may not bethe actual means of killing: for example, a bow is a weaponeven if it does not have arrows with it.Weapon system: the entire means by which a weapon is brought to125bear in battle. A horse archer, including bow, man, horse,protective clothing, extra ammunition, and other equipment, isa weapon system.126BIBLIOGRAPHY1. Ancient Works Ammianus Marcellinus. Res Gestae. vols. I-II. W. Seyfarth, ed.1978: Teubner, Leipzig.Arrian. Anabasis Alexandrou. Vol. I, Anabasis. A.G. Roos, ed.1967: Teubner, Leipzig.^ • Ektaxis kata Alanon and Techne Taktike. Vol. II,Scripta Minora. A.G. Roos, ed. 1967: Teubner, Leipzig.Caesar. Bellum Civile and Bellum Gallicum. (Oxford ClassicalText) R. du Pont, ed. 1962: Clarendon Press, Oxford.(Caesar) Bellum Africum. (Oxford Classical Text) R. du Pont, ed.1962: Clarendon Press, Oxford.Frontinus. Strategemata. (Loeb Classical Library) trans. C.E.Bennet. 1961: Harvard University Press, Cambridge, Mass.Josephus. Bellum Iudaicum. (Loeb Classical Library) trans. H. St.J. Thackeray. 1961: Harvard University Press, Cambridge,Mass.Mauricius. Strategikon. trans. G.T. Dennis. 1984: University ofPennsylvania Press, Philadelphia.Procopius. Bellum Gothicum. J. Haury, ed. 1963: Teubner,Leipzig.Tacitus. Annales. (Oxford Classical Text) C.D. Fisher, ed. 1985:Clarendon Press, Oxford.Vegetius. Epitoma de Re Militari. C. Lang, ed. 1967: Teubner,Leipzig.Xenophon. Anabasis. C. Hude, ed. 1969: Teubner, Leipzig.1272. Modern WorksAnderson, J.K.^Military Theory and Practice in the Age of Xenophon. 1970: University of California Press, Berkeley andLos Angeles. pp.111-140Aricescu, A.^"Auxilia limitis Scythici." in D.M. Pippidi, ed.:Actes du IXe Congrês International d'Etudes sur les Frontieres Romaines. 1974: Editura Academiei, Bucuresti, Romania.pp.117-122Becatti, G. Colonna di Marco Aurelio. 1957: Editoriale Domus,Milano. passimBirch, C.M. (ed.) Concordantia et Index Caesaris.^1989:^Olms-Wiedmann, New York. pp.656 and 1316Birley, E.^"Alae and cohortes milliariae." in Corolla Memoriae Erich Swoboda Dedicata. 1966: Hermann Bohlaus Nachf., Graz -Köln. pp.54-67Bishop, M.C. "The Military fabrica and the Production of Arms inthe Early Principate." in M.C. Bishop, ed.: The Productionand Distribution of Roman Military Equipment. BARInternational Series #275. 1985: Oxford. pp.1-43.^"Cavalry Equipment of the Roman Army in the FirstCentury A.D." in J.C. Coulston, ed.: Military Equipment andthe Identity of Roman Soldiers. BAR International Series#394. 1988: Oxford. pp.67-196Bivar, A.D.H.^"Cavalry Equipment and Tactics on the EuphratesFrontier." in Dumbarton Oaks Papers 26. 1976: The DumbartonOaks Center for Byzantine Studies, Washington, D.C. pp.272-291Bosworth, A.B.^"Arrian and the Alani" Harvard Studies in Classical Philology 81. 1977: Harvard University Press,Cambridge, Mass. pp.217-255Bowersock, G.W.^"Nabataeans and Romans in the Wadi Sirhan." inA.M. Abdalla, S. al-Sakkar, and R. Mortel, eds.: Studies inthe History of Arabia, vol. II: Pre-Islamic Arabia. 1984:King Saud University Press, Riyadh. pp.133-6Bugh, G.R. The Horsemen of Athens. 1988: Princeton UniversityPress, Princeton, New Jersey. pp.221-224du Buisson, M.^Inventaire des Inscriptions Palmyreniennes de Doura-Europos. 1939: Librairie Orientaliste Paul Geuthner,128Paris. p.69Burke, E. The History of Archery. 1957: William Morrow and Co.,New York. pp.40-137Callies, H. "Die Fremden Truppen im rOmischen Heer des Prinzipatsund die sogenannten nationalen Numeri." in Bericht derROmisch-Germanischen Kommission #45. 1964: Philipp vonZabern, Mainz am Rhein. pp.130-224Campbell, D.B.^"What Happened at Hatra?^The Problems of theSeveran Siege Operations." in P. Freeman and D. Kennedy, eds.:The Defence of the Roman and Byzantine East. BARInternational Series #297(i). 1986: Oxford. pp.51-8Carcopino, J. "Le limes de Numidie et sa garde syrienne." in Syria 6. 1925: Librairie Paul Geuthner, Paris. pp.118-149Cataniciu, I.B. Evolution of the System of Defence Works in RomanDacia (E. Dumitrescu, trans.). BAR International Series #116.1981: Oxford. passimCheesman, G.L.^The Auxilia of the Roman Imperial Army.^1914:Clarendon Press, Oxford. passimConnolly, P.^"The Roman Saddle." in M. Dawson, ed.:^RomanMilitary Equipment:^the Accoutrements of War.^BARInternational Series #336. 1987: Oxford. pp.7-27Coulston, J.C.^"Roman Archery Equipment." in M.C. Bishop, ed.:The Production and Distribution of Roman Military Equipment. BAR International Series #275. 1985: Oxford. pp.220-366. "Roman, Parthian, and Sassanid Tactical Developments."in P. Freeman and D. Kennedy, eds.: The Defence of the Romanand Byzantine East. BAR International Series #297(i). 1986:Oxford. pp.59-75• "The Value of Trajan's Column as a Source for MilitaryEquipment." in C. van Driel-Murray, ed.: Roman MilitaryEquipment: the Sources of Evidence. BAR International Series#476. 1989: Oxford. pp.31-44Cumont, F.^Fouilles de Dura-Euf.opos, 1922-3. 1926:^Librairieorientaliste Paul Geuthner, Paris. plates XCVIII, LIVDarko, E.^"Influences touraniennes sur l'evolution de l'artmilitaire des grecs, des romains, et des byzantins." inByzantion 10. 1935: Secretariat de la RevueInternationale des Etudes Byzantines, Brussels. pp.443-469Davies, J.L. "Roman Arrowheads from Dinorben and the Sagittarii of129the Roman Army." in Britannia 8. 1977: Society for thePromotion of Roman Studies, London. pp.257-270Davies, R.W.^Service in the Roman Army.^1989:^ColumbiaUniversity Press, New York. pp.71-174Denison, G.T. A History of Cavalry from the Earliest Times. 1913(second edition 1977): Greenwood Press, Westport, Conn.pp.1-9, 80-96Eadie, J.W. "The Development of Roman Mailed Cavalry." in Journal of Roman Studies LVII. 1967: Society for the Promotion ofRoman Studies, London. pp.161-173Fink, R.O. Roman Military Records on Papyrus. 1971: AmericanPhilological Association, Case Western Reserve University, AnnArbor, Michigan. pp.340-344Fitz, J. Les Syriens a Intercisa. Collection Latomus volume 122.1972: Revue des Etudes Latines, Bruxelles. pp.45-151Florescu, F.B. Die Trajanssaule. 1969: Rudolf Habelt VerlagGMBH, Bonn. passimGilliam, J.F. Roman Army Papers. Mayors Series. 1986: J.C.Gieben, Amsterdam.Griffiths, W.B.^"The Sling and its Place in the Roman ImperialArmy." in C. van Driel-Murray, ed.: Roman MilitaryEauipment: the Sources of Evidence. BAR International Series#476. 1989: Oxford. pp.255-280Haldon, J.F. "Some Aspects of Byzantine Military Technology fromthe Sixth to the Tenth Centuries." in Byzantine and ModernGreek Studies V, 1. 1975: pp.11-47Hanson, V.D. The Western Way of War. 1989: Oxford UniversityPress, New York.Holder, P.A.^Studies in the Auxilia of the Roman Army from Augustus to Trajan. BAR International Series 70. 1980: BAR,Oxford. passimHyland, E.^Equus:^the Horse in the Roman World. 1990: B.T.Batsford Ltd., London. pp. 63-200James, S.^"Dura-Europos and the Introduction of the "MongolianRelease"." in M. Dawson, ed.: Roman Military Equipment: the Accoutrements of War. BAR International Series #336. 1987:Oxford. pp.77-84"The Fabricae: State Arms Factories of the Later Roman130Empire." in J.C. Coulston, ed.: Military Equipment and the Identity of Roman Soldiers. BAR International Series #394.1988: Oxford. pp.257-331Judson, H.P. Caesar's Army. 1888: Boston. Reprint (1961):Biblo and Tannen Ltd., New York. pp.18-19, 53-86Kaegi, W.E. "The Contribution of Archery to the Turkish Conquestof Anatolia." in Speculum 39. 1964: Mediaeval Association ofAmerica, Cambridge, Mass. pp.96-108Kennedy, D.L. "Parthian Regiments in the Roman Army." in J. Fitz,ed.: Limes: Akten des XI Internationalen Limeskongresses. 1977: Akademiai Kiado, Budapest. pp.521-532Keppie, L. The Making of the Roman Army. 1984: B.T. Batsford,London. pp.145-198Khazanov, A.M. "Cataphracts in the History of the Art of War." inVestnik Drevnei Istorii I, 103. 1968: Izdatelstvo "Nauka",Moscow. pp.180-191 (Eng. summary p.191)McEwen, E. et al "Early Bow Design and Construction." inScientific American. June 1991: Scientific American Inc.,New York. pp.76-82McLeod, W.E. "Egyptian Composite Bows in New York." in AmericanJournal of Archaeology 66. 1962: The ArchaeologicalInstitute of America, Boston, Mass. pp.13-19 . "The Bow in Ancient Greece, with Particular Referenceto the Homeric Poems." doc. diss. abstract in Harvard Studiesin Classical Philology 71. 1967: Harvard University Press,Cambridge, Mass. pp.329-331. "The Range of the Ancient Bow." in Phoenix 19. 1965:University of Toronto Press, Toronto. pp.1-14.^"The Range of the Ancient Bow: Addenda." in Phoenix26. 1972: University of Toronto Press, Toronto. pp.78-82Mann, J.C.^"A Note on the Numeri." in Hermes 82. 1954: FranzSteiner Verlag, Wiesbaden. pp.501-6Medinger, P. "L'arc turquois et les archers parthes a la bataillede Carrhes." in Revue Archeologique. ser. 6, tome 2. 1933:Librairie Ernest Leroux, Paris. pp.227-234Negev, A. "The Nabateans and the Provincia Arabia." in Aufstiegund Niedergang der ROmischen Welt II, 8. 1978: Walter deGruyter, Berlin and New York. pp.520-686131Paterson, W.F. "The Archers of Islam." in Journal of the Social and Economic History of the Orient 9. 1966: E.J. Brill,Leiden, Holland. pp.69-87^ . "Shooting Under a Shield." in Journal of the Society ofArcher-Antiquaries I, 12. 1969: (Publisher not given) pp.27-28. "The Sassanids." in Journal of the Society of Archer-Antiquaries I, 12. 1969: pp.29-32Richmond, I. Trajan's Army on Trajan's Column. 1982: The BritishSchool at Rome, London. pp.16-20Rossi, L. Trajan's Column and the Dacian Wars. J.M.C. Toynbee,trans. 1971: Thames and Hudson, London. pp.67-120Rostovtzeff, M.I.^"Les inscriptions caravanieres de Palmyre." inMelanges Gustave Glotz 2. 1932: Les presses universitairesde France, Paris. pp.793-811^ . "The Caravan-Gods of Palmyra." in The Journal of RomanStudies 22.^1932:^Society for the Promotion of RomanStudies, London. pp. 107-116Roxan, M. Roman Military Diplomas 1978-1984. Institute ofArchaeology Occasional Publication No. 9. 1985: Institute ofArchaeology, London. passimSaddington, D.B. "The Roman Auxiliaries in Tacitus, Josephus, andOther Early Imperial Writers." in Acta Classica XIII, 1970:A.A. Balbema, Cape Town. pp.89-124Scheiper, R. Bildpropaganda der ROmischen Kaiserzeit unterbesonderer Beracksichtiqunq der Tralanssaule in Rom undKorrespondierender Manzen. 1982: Dr. Rudolf Habelt GMBH,Bonn. pp.155-182Speidel, M.P.^"The Roman Army in Arabia." in Aufstieg undNiedergang der ROmischen Welt II, 8.^1978:^Walter deGruyter, Berlin and New York. pp.687-730. Roman Army Studies. Mayors Series 1984: J.C. Gieben,Amsterdam.^ . "Catafractarii, Clibanarii, and the Rise of the LaterRoman Mailed Cavalry." in Epigraphica Anatolica 4. 1984:Dr.R. Habelt, Bonn. pp.151-6. "The Later Roman Field Army and the Guard of the HighEmpire." in Latomus 46, 2, ay.-juin 1987: Revue des Etudeslatines, Bruxelles. pp.375-9132. "Nubia's Roman Garrison." in Aufstieg und Niederclang derROmischen Welt II, 10.1. 1988: Walter de Gruyter, Berlin andNew York. pp.767-798Sulimirski, T. "Les archers a cheval: cavalrie legere desanciens." in Revue Internationale d'histoire militaire III,12. 1952: (Publisher not given) pp. 447-461Warner, P.^Firepower. 1988: Robert Hartnoll, Bodmin, U.K.pp. 12-34Webster, G.^The Roman Imperial Army.^1969: Adam and CharlesBlack, London. pp.107-165, 221-260van de Weerd, H. and P. Lambrechts. "Note sur les corps d'archersau Haut Empire." in F. Altheim and R. Stiehl, eds. Die Araberin der Alten Welt erster Band. 1964: Walter de Gruyter &Co., Berlin. pp.661-677Yadin, Y.^The Art of Warfare in Biblical Lands.^1963: McGraw-Hill, New York. pp.6-15, 46-48, 80-113, 293-297Department of National Defence Publication B-GL-317-019/PT-001The light Machine Gun 5.56 mm C9. 1987: Queen's Printers,Ottawa. p. 2-20133


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items