Trauma analysis in paleopathology

Trauma Analysis in PaleopathologyNANCY C. LOVELLDepartment of Anthropology, University of Alberta,Edmonton, AB T6G 2H4, Canada

KEY WORDS fracture; dislocation; violence; head injury

ABSTRACT This paper reviews the mechanisms of injury and the types offractures that most commonly affect the human skeleton, presents descriptiveprotocols for cranial and postcranial fractures adapted from clinical andforensic medicine, and summarizes anatomically the injuries most likely to befound in archaeological skeletons along with their most common causes andcomplications. Mechanisms of injury are categorized as direct and indirecttrauma, stress, and fracture that occurs secondary to pathology. These areconsidered to be the proximate, or most direct, causes of injury and they areinfluenced by intrinsic biological factors such as age and sex, and extrinsicenvironmental factors, both physical and sociocultural, that may be thoughtof as the ultimate, or remote, causes of injury. Interpersonal conflict may beone of those causes but the skeletal evidence itself is rarely conclusive andmust therefore be evaluated in its individual, populational, sociocultural, andphysical context. A cautionary tale regarding parry fractures is presented asan illustration. Yrbk Phys Anthropol 40:139–170, 1997. r 1997 Wiley-Liss, Inc.

Trauma may be defined many ways butconventionally is understood to refer to aninjury to living tissue that is caused by aforce or mechanism extrinsic to the body.The anatomical importance and sociocul-tural implications of trauma in antiquitylong have been recognized and the descrip-tion of trauma in human skeletal remainsand the identification and comparison oftrauma patterns among ancient populationstherefore have a lengthy history. As thediscipline of palaeopathology has developed,the objectives of traumatic injury analysishave shifted from a focus on the identifica-tion and description of the earliest and themost unusual pathological specimens to theinterpretation of the social, cultural, or envi-ronmental causes of traumatic injury; theirrelationship to biological variables, such assex and age, that may have social or culturalrelevance; and their temporal and spatialvariation. Thus, interpretations of the causeof trauma in antiquity range from inter- andintragroup conflict (e.g., Angel, 1974; Ham-perl, 1967; Janssens, 1970; Jurmain, 1991;

Liston and Baker, 1996; Shermis, 1984;Stewart, 1974; Walker, 1989; Wood-Jones,1910; Zivanovic, 1982; and others) to environ-mentally or occupationally facilitated mis-adventure and accident (e.g., Angel, 1974;Burrell et al., 1986; Cybulski, 1992; Grauerand Roberts, 1996; Kelley and Angel, 1987;Lovejoy and Heiple, 1981; Wells, 1964; andothers). Although great advances have beenmade in paleopathological diagnosis andinterpretation in recent years, inconsisten-cies in descriptions and interpretations oftrauma in the literature, particularly asthey affect our understanding of the natureand extent of interpersonal violence in antiq-uity, have made it difficult to compare theresults of different studies and to acceptwith confidence some conclusions. The pur-pose of this paper, therefore, is to reviewtypes of fractures and the mechanisms ofinjury, critique protocols for fracture descrip-tion, and consider the problems of interpret-ing the causes of injury. Although an impor-tant source of data for the study of thehistory of medicine, a discussion of skeletal

YEARBOOK OF PHYSICAL ANTHROPOLOGY 40:139–170 (1997)

r 1997 WILEY-LISS, INC.

indicators of surgical practice and othermedical intervention is beyond the scope ofthis paper.

Scholars have categorized traumatic inju-ries in a variety of ways (Table 1), butgenerally refer to both accidental and inten-tional trauma, the former usually includingmost fractures and dislocations and the lat-ter usually including examples of surgicalintervention and weapon wounds. It may bemore prudent, however, to first sort injuriesaccording to their predominant characteris-tic, either fracture1 (any break in the conti-nuity of a bone) or dislocation (the displace-ment of one or more bones at a joint), ratherthan to classify injuries in a manner thatimplies causation or intent.

DISLOCATIONS

Traumatic injuries to joints may result inpartial or complete dislocations. A disloca-tion, or luxation, occurs when the articularsurfaces of a joint are totally displaced fromone another. A subluxation results when thearticular surfaces are partially displacedbut do retain some contact.Although disloca-tions and subluxations may be congenital orspontaneous in origin, they are most com-monly caused by trauma and in such cases itis not uncommon for the joint displacement

to be associated with a fracture. Since dis-placement cannot occur without damage tothe joint capsule and ligaments, complica-tions such as the ossification of membrane,ligament, and tendon attachments to bonemay ensue. Persistent instability of the jointalso may result, particularly in the shoulderand ankle, although this complication can-not be easily identified in archaeologicalskeletal remains. Osteoarthritis is one of themore common, and recognizable, complica-tions and results from damage to the articu-lar cartilage itself or from prolonged incon-gruence of the joint surfaces.

For joint displacements to be recognizablein dry bone the injury must have occurredsome time before the death of an individualand remained unreduced (i.e., not ‘‘set’’) longenough for bone modifications to take place.Some dislocations, such as of the digits,usually can be relatively quickly and easilyreduced (but see Dreier, 1992), while others,such as of the vertebrae, may cause immedi-ate death. In either case no evidence of theinjury will be observable in archaeologicalskeletons.

Dislocations tend to be more frequent inyoung and middle-aged adults, since in sub-adults a similar force instead causes epiphy-seal separation and in older adults causesfracture of osteoporotic bones. The glenohu-meral joint is a common site of dislocation,the shallowness of the shoulder joint mak-ing it particularly susceptible to displace-ment. Traumatic dislocation of the femoralhead from the acetabulum, in contrast, re-quires considerable force and this site is

1‘‘Infracture’’ and ‘‘infraction’’ are alternative terms for ‘‘frac-ture,’’ particularly undisplaced fractures, according to medicaldictionaries (e.g., Stedman, 1982). Although rarely used inpaleopathology, these terms may be seen in the literature with adifferent meaning: ‘‘infraction,’’ for example, has been defined asan incomplete fracture. In the interests of developing clear andstandard terminology, these alternative terms are not used inthis review.

TABLE 1. Variation in the categorization of traumatic injuries by different authors

Knowles (1983) Merbs (1989a)Ortner and

Putschar (1981) Steinbock (1976)Roberts and

Manchester (1995) This study

Fractures Fractures Fractures Fractures Fractures3 Fractures4

Dislocations Dislocations Dislocations Dislocations Dislocations DislocationsTrephination and

amputationSurgery Trephination Sharp Instruments2 Osteochondritis

dissecansWeapon wounds Weapon wounds Weapon wounds Growth arrest linesExostoses Scalping Scalping Crushing injuriesSchmorl’s nodes Dental trauma1 Deformation1

Osteochondritisdissecans

Pregnancy-related

1Includes cranial deformation, filing of teeth, and other modifications performed for aesthetic purposes.2Includes surgery and weapon wounds.3Includes piercing injuries caused by knife and sword cuts, scalping, and projectile points (i.e., surgery and weapon wounds).4Includes piercing injuries caused by knife and sword cuts, scalping, and projectile points (i.e., surgery and weapon wounds), and crushfractures caused by foot binding and by cranial binding and flattening.

140 YEARBOOK OF PHYSICAL ANTHROPOLOGY [Vol. 40, 1997

more commonly associated with congenitaldislocations.

FRACTURES

A fracture consists of an incomplete orcomplete break in the continuity of a bone.The most common types of fractures, such astransverse, spiral, oblique, and crush frac-tures, result from direct or indirect trauma.Two additional types of fractures, those re-sulting from stress and those secondary topathology, are less common and have dis-tinct etiologies. Fracture types and theirassociated mechanisms of injury are re-viewed below (and summarized in Table 2),followed by discussions of fracture healingand complications.2

Mechanisms of injury and typesof fractures

Direct trauma. When a break occurs atthe point of impact it is referred to as adirect trauma injury (Miller and Miller, 1979)and the resulting fracture may be trans-verse, penetrating, comminuted, or crush(Fig. 1). A transverse fracture results fromforce applied in, and appears as, a lineperpendicular to the longitudinal axis of thebone. Clinically, this injury often results

from a hard kick to the shin and is often seenamong soccer players. Typically, transversefractures are caused by a relatively smallforce delivered to a small area.

Partial or complete penetration of thebone cortex by cutting, piercing, drilling, orscraping, such as the excision of pieces ofcranial vault bones in the practice of trephi-nation, or the amputation of a limb segmentis classed as a direct trauma injury. Penetrat-ing fractures typically are caused by applica-tion of a large force to a small area. Inarchaeological contexts, penetration couldbe caused by a projectile point, the blade ofan axe or sword, or a musket ball (Blair,1983; Butler, 1971). Wounds from arrow orspear points often can be identified withcertainty only if the point remains embed-ded in the bone and healing would not beevident if such wounds were linked to thedeath of the individual. The human remainsin many historic cemeteries show the trau-matic results of conflicts with bullets andother projectiles (e.g., Gill, 1994; Larsen etal., 1996; Owsley et al., 1991). Early cases ofpenetrating projectile wounds can be ex-pected to show subsequent infection and/orpronounced deformity in the absence of sta-bilization or rest of the injured part. Somepenetrating fractures may also be commi-nuted, which occurs when the bone is brokenin more than two pieces. In clinical cases,

2The information provided here has been compiled from avariety of sources, including Adams (1987), Apley and Solomon(1992), Gustilo (1991), Harkess and Ramsey (1991), and Schultz(1990).

TABLE 2. Summary of mechanisms of injury and associated types of fractures

Mechanism of Injury Type of fracture Comments

Direct trauma Penetrating Partial or complete penetration of bone cortexComminuted Bone is broken in more than two pieces; most common in long

bone diaphysesTransverse Force applied in a line perpendicular to long axis of the boneCrush Most common in cancellous bone

Depression Crushing force on one side of the boneCompression Crushing force on both sidesPressure Force applied to growing bone

Indirect trauma Spiral Rotational and longitudinal stress on long axis; often confusedwith oblique fracture

Oblique Rotational and angular stress on long axis; often confused withspiral fracture

Torus/greenstick Bending of the bone due to longitudinal compression; common inchildren

Impacted Bone ends are driven into each otherBurst Found in the spine due to vertical compressionComminuted Force splits in several directions and forms a ‘‘T’’ or ‘‘Y’’ shapeAvulsion Fracture due to tension at ligament or tendon attachment

Stress Due to repetitive force, usually perpendicular to long axisMay be confused with direct trauma transverse fracture

Secondary to pathology Secondary to localized or systemic disease that has weakened thebone

141TRAUMA ANALYSISLovell]

high velocity bullets and blunt force traumato the cranium typically cause comminutedfractures.

Crush fractures most commonly occur incancellous bone and result from the applica-tion of a direct force to the bone, whichcollapses on itself. Three types of crushfractures are recognized: depression, com-pression, and pressure. The first refers tocrushing on one side of the bone (especiallycommon on the ectocranium) while the sec-ond refers to a crushing force that originateson both sides of the bone. The incompletepenetration of a bone by a low velocityprojectile may result in a crush fracture,such as depressed fractures caused by theimpact of musket balls (Liston and Baker,1996) or shotgun pellets (Swan and Swan,1989). Blunt trauma, such as that producedby a bludgeon, fist, or hammer, or when anobject is dropped on the hand or foot, resultsin crush fractures. The third type of crush-ing injury results when developing boneresponds to the application of direct force.Examples of this last type are culturallymandated bone alterations, such as the shap-ing of immature cranial and foot bones byvarious types of binding for beautification.

Rarely, direct trauma injury that bruisesa joint may fracture articular cartilage, and

sometimes the subchondral bone as well,causing separation of a fragment from themargin of the articular surface. This result-ing lesion may be confused with osteochon-dritis dissecans, which is caused by asepticnecrosis and is usually seen as the completeor incomplete separation of a portion of jointcartilage and subchondral bone, most com-monly on the femoral condyles. Osteochon-dritis dissecans is usually recognized in drybone as a pit, often 2 to 5 mm in diameter, inthe subchondral bone, although new boneformation may partially or completely fillthe defect or may produce a deposit thatexceeds the level of the normal articularsurface. The etiology of osteochondritis disse-cans is uncertain but indirect trauma isthought to play at least a contributory role.

Indirect trauma. When a fracture occursin a place other than the point of impact it issaid to result from indirect trauma (Millerand Miller, 1972). Oblique, spiral, green-stick, impacted, burst, and avulsion frac-tures are consequences of indirect trauma(Fig. 2). An oblique fracture, where the lineangles across the longitudinal axis, is indica-tive of a combined angulated/rotated force(Harkess and Ramsey, 1991). If the fractureis well healed, this break is easily confused

Fig. 1. Fractures caused by direct trauma. From left to right: transverse, penetrating, comminuted,and crush.


with a spiral line. A spiral fracture linewinds down around a long bone shaft due toa rotational and downward loading stress onthe longitudinal axis. In some cases, such aforce applied to the tibia results in a fracture-dislocation of the ankle rather than a spiralfracture (Harkess and Ramsey, 1991); atother times, the force results in an associ-ated proximal fibular fracture.

Torus or greenstick fractures result frombending or buckling of bone when stress isapplied. These often are due to indirecttrauma and are most commonly seen inchildren, whose bones are still pliable andhence less likely to break, instead producinga localized bulging on the bone. An exampleis a greenstick fracture of the clavicle thatresults during childbirth when the child’sbiacromial breadth is too large to pass easilythrough the mother’s pelvic outlet. Green-stick fractures are also characterized by anincomplete fracture involving only the con-vex side of a bone that has been subjected tobending stress. In adults the ribs are com-monly affected.

Less common fractures resulting from in-direct trauma are impacted, avulsion, andburst fractures. An impacted fracture occurswhen the bone ends at a fracture site aredriven into each other by the force of injury.

Clinically, this is often seen in the proximalhumerus as the result of a fall onto anoutstretched hand, and in the metacarpalsas a result of trauma to the fist whenpunching. An avulsion fracture is causedwhen a joint capsule, ligament, or tendon isstrained and pulls away from its attachmentto the bone, tearing a piece of bone with it. Aparticular type of avulsion fracture leaves atransverse fracture line: a transverse frac-ture may occur to the ulnar olecranon pro-cess or the patella if the extensor musclescontract forcefully while the joint is flexed,and in extreme cases the bone fragmentswill separate and may heal without uniting.

A burst fracture is located in the spine. Itresults from a vertical compression thatruptures the intervertebral disc through thevertebral end plate, forcing disc tissue intothe vertebral body (Fig. 3). A mild form ofthis injury is often seen in archaeologicalspecimens as a small, localized, typicallycircular, depression in the end plate that isusually called a ‘‘Schmorl’s node.’’

Comminuted fractures may be due to indi-rect trauma as well as to direct trauma. Theindirect comminuted fracture is patternedlike a ‘‘T’’ or ‘‘Y,’’ and is produced by a forcethat passes through the bone, splitting it inseveral directions (Perkins, 1958). Crush

Fig. 2. Fractures caused by indirect trauma. From left to right: oblique, spiral, greenstick due toangular force, greenstick due to compression, impaction, and avulsion.


fractures also can be found as a result ofindirect trauma, such as in the calcaneusafter a person has jumped from a height.

Stress fractures. Repetitive force can re-sult in a stress or fatigue fracture. The usualareas of occurrence are the metatarsal, calca-neus, and tibia (Wilson and Katz, 1969).Stress fractures in the metatarsals are some-times referred to as ‘‘marching’’ fractures,since they are often diagnosed in militarycadets. Those in the tibia have been knownfor some time to affect dancers, while theincreased interest in jogging and aerobicdancing in recent decades as well as theadoption of alternative religious practiceshas led to their higher prevalence in othersegments of western society (Burrows, 1956;Cohen et al., 1974). The fracture line isusually perpendicular to the longitudinalaxis, therefore problems may arise in tryingto distinguish between stress and direct

trauma transverse injuries. Commonly astress fracture will be visible as a nondis-placed line or crack in the bone, called ahairline fracture, which is not detectableradiologically until a bony callus has formedover the break.

Fractures secondary to pathology. Frac-tures often occur secondarily to a diseasealready present in the body. Systemic dis-eases such as metabolic disturbances andnutritional deficiencies leave bone vulner-able to spontaneous fracture or to fracturefrom minor trauma. For example, postmeno-pausal females may suffer fractures if theirbones have been weakened by osteoporosis.Other skeletal markers of specific diseasemay aid in attributing cause to the fracture:neoplastic fractures are seen when the breakis through or adjacent to a tumor that is in,or of, bone, and the collapse of vertebralbodies is not an uncommon consequence oftuberculosis in the spine (Pott’s disease).

Fracture healing

Duration of healing. Fractures begin toheal immediately after the bone is broken,but the process differs for cancellous andtubular bone. Most investigators identifyfive overlapping stages in tubular bone heal-ing (Adams, 1987; Apley and Solomon, 1992;Paton, 1984). Table 3 summarizes the activi-ties of these stages and the approximatetime after injury that each is observed.Healing normally is not visible on radio-graphs until approximately 2 to 3 weeksafter the injury, when a callus of wovenbone, the result of cell proliferation from theperiosteum, marrow cavity, and surround-ing connective tissue, appears around thesite of injury. The callus internally andexternally bridges the gap caused by thefracture and stabilizes the fractured ends.Consolidation of this woven bone into ma-ture lamellar bone occurs subsequently, butthe duration of the process depends upon thenature of the fracture and the type of boneinvolved. In a phalanx, a solidly unitedfracture may develop in less than 1 month,while the same transformation may take upto 6 months in a tibia or femur. Bones of the

Fig. 3. Burst fractures of the lumbar vertebrae in ayoung adult male from the historic Fur Trade period inAlberta.


upper limb tend to heal faster than do thoseof the lower limbs, and spiral and obliquefractures heal faster than do transversefractures.

In contrast to compact tubular bone, can-cellous bone has a meshlike structure withno medullary canal. This provides a muchlarger area of contact between fracture frag-ments, which facilitates healing. In addi-tion, this mesh can be more easily pen-etrated by bone-forming tissue than cancompact bone, so the union occurs directlybetween bone fragments instead of indi-rectly via the periosteal and endosteal cal-lus. The initial haematoma is penetrated byproliferating bone cells which grow fromopposing fracture surfaces. The developingtissues fuse when they meet and subse-quently calcify to form woven bone. Healingis thus simpler and faster in cancellous bonethan in compact bone.

Because there is a delay before healing isvisible macroscopically or radiographically,it may be difficult to distinguish some post-mortem breaks from unhealed premortemfractures. Perimortem fractures is the termgiven to such injuries, which may haveoccurred in the recent antemortem period(i.e., up to 3 weeks before death) and aretherefore unhealed, or that alternativelymay have occurred in a postmortem periodthat is of indeterminate length (perhapsweeks or months) but during which the boneis still relatively fresh and its organic compo-nents not yet deteriorated. Otherwise, distin-guishing between antemortem/perimortemtrauma and that which clearly occurred

after death is predicated upon the differentfracture properties associated with bone thatretains its viscoelastic nature and bone thatdoes not, and upon the different appear-ances of bone surfaces after various postmor-tem intervals (Buikstra and Ubelaker, 1994;Maples, 1986; Mann and Murphy, 1990;Ubelaker and Adams, 1995). Antemortem orperimortem fractures can be identified by 1)any evidence of healing or inflammation; 2)the uniform presence of stains from water,soil, or vegetation on broken and adjacentbone surfaces; 3) the presence of greenstickfractures, incomplete fractures, spiral frac-tures, and depressed or compressed frac-tures; 4) oblique angles on fracture edges;and/or 5) a pattern of concentric circular,radiating, or stellate fracture lines. Post-mortem fractures, in contrast, tend to becharacterized by 1) smaller fragments; 2)nonuniform coloration of the fracture endsand the adjacent bone surface, especiallylight-colored edges; 3) squared fractureedges; and 4) absence of fracture patterningdue to the increased tendency of dry, brittle,bone to shatter on impact.

Complications of healing. Complica-tions should be assessed when examiningfractures because they may provide informa-tion regarding mobility, morbidity, mortal-ity, and medical treatment or the lackthereof. In addition to the fracture typesdescribed above, the relationship of the frac-ture to surrounding tissue is referred to as‘‘closed’’ or ‘‘open.’’ When the fractured bonedoes not come into contact with the outer

TABLE 3. The process and duration of fracture healing in tubular bones(Adams, 1987, Apley and Solomon, 1993, and Paton, 1984)

Healing stage Healing processes Duration

Haematoma formation Blood from torn vessels seeps out and forms a haematoma 24 hoursFractured bone ends die due to lack of blood supply

Cellular proliferation Osteoid is deposited around each fragment by osteoblasts ofperiosteum and endosteum and pushes haematoma aside

3 weeks

Fracture is bridged; visible in dry boneCallus formation Callus of woven bone forms from mineralization of osteoid and

acts as a splint for periosteal and endosteal surfaces3 to 9 weeks

Visible radiologicallyConsolidation Mature lamellar bone forms from callus precursor and results

in a solidly united fracture areaVaries by skeletal ele-

ment from a few weeksto a few months

Remodelling Gradual remodelling of bone to its original form, strength-ening along lines of mechanical stress

6 to 9 years

Increased density on radiographs marks the fracture site onadult bones


surface of the skin, the fracture is termedclosed. An open fracture, also known as acompound fracture, is when the bone pro-trudes through the skin or the skin is brokento the level of the bone, as in a crushing orpenetrating wound. Open fractures are proneto infection, which hinders the union of thefracture and creates instability. A pathogen,Staphylococcus aureus in about 90% of clini-cal cases (Ortner and Putschar, 1981), maybe introduced to the body through an openfracture from surface contamination or froma penetrating instrument or contaminant.Although there is a tendency to regard local-ized infections as related to observed frac-tures, but to interpret nonlocalized infec-tions as unrelated to fracture, posttraumaticinfection may in fact be present either as alocalized condition or, due to hematogenousdissemination of the pathogen, as a systemicinfection. Whether localized or systemic, ifthe body’s immune system is unable to com-bat the infection successfully bony responseis usually visible in the form of periostitis(an inflammation of the periosteum) or osteo-myelitis (a more severe bone infection thatinvolves the medullary cavity). Periostitis isusually characterized by focal periosteal bonedeposition that may eventually form aplaquelike sheet over the cortex. Osteomyeli-tis is identified by a thickened contour in thearea of the fracture and the bone may feelheavier. Pathognomonic evidence of osteomy-elitis results from the development of sub-periosteal abscesses that deprive the bone ofits blood supply and lead to necrosis (thedead bone forms a sequestrum). The perios-teum continues to produce new, hypervascu-lar bone around the sequestrum, forming ashell of bone called involucrum. The subperi-osteal pus must escape through the involu-crum to the skin surface, however, and indoing so forms one or more sinuses (cloacae)in the involucrum for pus drainage. In drybone the sequestrum, lying under the involu-crum, may be visible through a cloacal open-ing. Posttraumatic osteomyelitis is most com-monly observed in the cranium and longbones of archaeological skeletons.

Fractures inevitably result in the ruptureof minor blood vessels but this is not usuallya serious complication. In some cases, how-ever, bone displacement can compress or

twist blood vessels and lead to ischemia.This will delay the healing process and couldlead to bone death if unrelieved. Avascularnecrosis normally occurs near the articularends of bones where the blood supply tosubchondral bone is limited. Death of thetissue begins a week after the nutrientsupply is reduced and may continue for up to4 years. During this time the bone loses itstrabecular structure, becomes granular, andbegins to disintegrate due to muscle stressor body weight. The adjacent articular carti-lage also dies as a result of deficient nourish-ment, usually resulting in osteoarthritis.

Nerve injuries also may be associatedwith fractures. Three types of nerve injuriesare generally recognized. Damage is slightin neurapraxia and results in temporaryimpairment that corrects itself within a fewweeks. In contrast, the internal nerve archi-tecture is preserved but axons are badlydamaged in axonotmesis, resulting in periph-eral degeneration that may take manymonths to heal. Such a lesion may resultfrom pinching, crushing, or prolonged pres-sure. The most serious type of nerve injury,neurotmesis, involves complete division of anerve, either through severing or severescarring, and requires surgical repair. Theconsequences of these types of nerve injuriesrange from loss of sensation to loss of func-tion. Usually the loss is temporary, butmuscle atrophy may result and if the nerveloss is prolonged or permanent the boneswill display signs of disuse atrophy as well.This sequel would be most likely in archaeo-logical cases of neurotmesis. In addition, ifthere is loss of innervation to the fracturesite, the individual will not feel pain andmay therefore continue to use the brokenbone, impairing healing. Fracture of thevertebral column may result in damage tothe spinal cord or spinal nerves, with paraly-sis below the level of the injury a possibleoutcome. Depressed skull fractures with en-docranial displacement are usually associ-ated with significant brain injury, whichmust also be considered in cases of linearfractures of the cranial vault.

Another complication is posttraumatic os-sification of a haematoma, which resultswhen absorption of the haematoma is pre-vented by excessive stress placed on the


periosteum. A smooth mass of bone is macro-scopically visible after 2 months, with calci-fication being visible radiologically a fewweeks after the injury. Although usuallybenign, movement may be restricted if thereis joint involvement.

If joint function is affected by traumaticinjury, osteoarthritis may develop as a com-plication. Stiffness caused by fibrous adhe-sions or joint swelling may lead to prolongeddisuse of the joint or limb. Shortening orangulation may result in some loss of nor-mal function in the affected limb or in thejoints directly above and below the fracture;this may be difficult to interpret since un-usual biomechanical stress at a joint inwhich a fractured bone participates maycause osteoarthritis, but it is also possiblefor a joint on an uninjured limb to be af-fected. The latter might occur, for example,if weight bearing was shifted in order tofavor the injured leg. Premature deteriora-tion of articular cartilage and subsequentdeterioration of subchondral bone are com-mon complications of breaks affecting thejoint surface itself, since cartilage repair is avery slow process. Such fractures also canresult in ankylosis of the joint.

Three final complications of fractures aredelayed union, nonunion, and malunion. Inclinical settings the union of a fracture isdefined as delayed if it has not occurred inthe time expected for that skeletal element,age, and sex of the individual, and it mayeventually be classed as a nonunion. In drybone specimens, of course, delayed but even-tually successful union cannot be distin-guished from undelayed union. Several fac-tors may impede the process of healing, butoverall poor health and/or nutrition in an-cient populations may be a largely unrecog-nized contributor (Grauer and Roberts,1996).

The diagnosis of nonunion is applied whenthe fracture fragments fail to unite and themarrow cavity seals. Radiologically, non-union may be identified by sclerosis at thebone ends. After a prolonged period of time,the fragments take on a rounded appear-ance at their ends, which are connected byfibrous tissue. Nonunion may result frominadequate bone healing due to infection,inadequate blood supply, insufficiency of vi-

tamin D or C or of calcium, excessive move-ment between bone fragments during heal-ing, soft tissue being caught between thefragment ends, inadequate contact betweenthe fragments, presence of foreign material,or from the destruction of bone due to pathol-ogy or the injury itself (Altner et al., 1975;Karlstrom and Olerud, 1974; Sevitt, 1981;Stewart, 1974; Urist et al., 1954; Yamigashiand Yoshimura, 1955). If there is persistentmovement between the ununited ends, apseudarthrosis, or false joint, may form,although this complication is relatively rare(Stewart, 1974). Studies of modern humanpopulations indicate a frequency of pseudar-throsis of less than 5% (Heppenstall, 1980;Urist et al., 1954), while an examination ofdata from temporally, geographically, andcultural diverse archaeological populationsreveals an average frequency of 2% (Burrellet al., 1986; Jimenez, 1994; Lovejoy et al.,1981; Stewart, 1974). Among alloprimates,data from Bramblett (1967), Jurmain (1989),Lovell (1990), and Schultz (1937, 1939) alsogive an average pseudarthrosis frequency ofapproximately 2%.

A malunion consists of a fracture thatheals leaving a deformity. This may occurwhen a fracture has not been reduced orwhen reduction was not maintained, leavingthe fragments to heal grossly angulated orexcessively shortened. Shortening is causedby overlap, substantial angulation, crush-ing, or gross bone loss. Injuries to growingbone that affect the epiphyses and lead topremature fusion of the growth plate mayresult in shortening, as may bone infarctionresulting from sickle cell disease, whichmost often affects the epiphyses of the grow-ing skeleton, especially those of the proxi-mal femur. The presence of a shortened boneis most detrimental to the lower, weight-bearing limbs, although a difference of up to20 mm is considered by clinical practitionersto be tolerable. A greater loss in length canlead to backache from pelvis tilting andlateral and rotational spinal deviation. Re-cent studies have interpreted minimal defor-mity in bones that are likely to be severelyaffected when fractured as evidence for im-mobilization of the injured part and possiblemedical treatment (Grauer and Roberts,1996).


DESCRIPTIVE PROTOCOLS FORFRACTURES

Proper description of an injury is the firststep in trauma analysis (Ortner and Putschar,1981; Steinbock, 1976) and is the basis fordetermining the mechanism, or proximatecause, of the injury. In turn, an understand-ing of the proximate cause is crucial for theidentification of the ultimate cause oftrauma, usually behavior. Proper descrip-tion of observed lesions also provides otherscholars with an opportunity to agree ordisagree with the diagnosis and/or infer-ences that are made about the socioculturalor environmental context of the injury. Al-though several models proposed recentlyhave made great strides in standardizingdescriptive protocols (e.g., Buikstra andUbelaker, 1994; Dastugue and Gervais, 1992;Grauer and Roberts, 1996; Roberts, 1991),paleopathologists have not yet reached aconsensus on descriptive standards fortrauma and many are not always familiarwith the underlying mechanisms of injury.Ideally, any method of fracture descriptionwill recognize two main sources of confusionin interpretation: the variation in appear-ance expressed by fractures caused by thesame mechanism of injury, as well as thesimilarities in appearance displayed by frac-tures caused by different mechanisms ofinjury. Ultimately, proper fracture descrip-tion should seek to improve the accuracyand reliability of interpretation without ex-ceeding the limits of inference that are setby the descriptive data themselves.

Although fracture types are here sub-sumed under their proximate cause, whendescribing and interpreting injury the frac-ture type is usually recognized first. Identifi-cation of the mechanism of injury then fol-lows logically, and the third step in traumaanalysis involves interpretation of the ulti-mate cause of the injury. For example, animpacted fracture of the distal radius withposterior displacement of the distal frag-ment may be recognized as a Colles’ fracturedue to its characteristic location and defor-mity. The proximate cause of the injury maythen be identified as indirect trauma. Inter-preting the ultimate cause may be difficult,but a fall onto the outstretched hand would

be a logical conclusion. If the fracture wasobserved in an older female, the possibilityof the fracture occurring secondary to osteo-porosis could also be considered.

The principal aim of most protocols hasbeen to establish standardized descriptionsfor fractures observed in dry bone, althoughadditional objectives, such as the evaluationof evidence for treatment of traumatic inju-ries are sometimes also stated (e.g., Grauerand Roberts, 1996; Roberts, 1991). Threerecently developed protocols are outlinedhere.

The repatriation of Native American pre-historic and historic skeletal remains drovethe development of standards for data collec-tion that includes procedures for document-ing fractures (Buikstra and Ubelaker, 1994).These procedures recognize nine types offractures and eight varieties of shape charac-teristics. All types and varieties are notmutually exclusive, but may have restrictedapplication. Shape characteristics, for ex-ample, describe lesions caused by blunt orsharp force and by projectiles, as well asradiating fractures and amputations. Peri-mortem fractures are identified at a thirdlevel of description, followed by sequelaesuch as healing status and various complica-tions. Dislocations are classed separately.The recommended data collection forms anddescriptive protocol do not provide for mal-union as a component of fracture descriptionspecifically, but rather under the pathologycategory of ‘‘abnormality of shape,’’ in whichmalunion would be identified as either barelydiscernable or clearly discernable angulation.

A second method was designed specificallyto describe fractures in a way that wouldprovide the information necessary to exam-ine the technology and knowledge of treat-ments in past societies (Grauer and Roberts,1996; Roberts, 1991). The method describesthe location and type of fracture and empha-sizes evaluation of the success of long bonehealing. Macroscopic and radiographicmeans are employed to assess complicationsof shortening and deformity, and sequelaesuch as infection and osteoarthritis. Skullfractures are described as resulting fromblunt or sharp force and are evaluated interms of healing as well as evidence fortrepanation. The need for radiographic


evaluation of fractures in order to determinethe amount of healing and the particulars ofdeformity and/or displacement is stressed,as is the importance of radiography fordetecting and interpreting well-remodelledfractures (Grauer and Roberts, 1996; Rob-erts, 1991). Unfortunately, radiographicequipment is not always available, espe-cially in field settings, and the interpreta-tion of radiographs may be made difficult bypostmortem alterations common in archaeo-logical contexts, such as soil inclusions thataffect density or the differential identifica-tion of osteoporosis versus diagenetic boneloss (Roberts, 1991).

Finally, a third system concerns cranialvault injuries, categorizing them as pierc-ings, depressions, gashes, cuts, and slices(Filer, 1992). The first category is describedas consistent with a penetrating injury, thesecond with blunt force trauma, and the lastthree as resulting from edged/bladed imple-ments, including sharp projectiles. The ma-jority of these lesions were interpreted asresulting from interpersonal violence, anassessment not inconsistent with the appar-ent culture-historical context of the remains.

It is likely that no one system of fracturedescription will suit all investigators, sincesome will be more or less concerned with theaffected body part, specific complications, orpossible causative behaviors. Most proto-cols, however, share similar basic categoriesof description. The method for fracture de-scription that is presented below incorpo-rates these categories in a system adaptedfrom clinical and forensic medicine. It ispredicated on identification of the skeletalelement(s) involved and the type of injury, as

well as detailed descriptions of deformationand of any associated nontraumatic lesionsthat may indicate causality or postinjurycomplications. The information thus ob-tained then serves as a basis for inferencesabout the mechanism of injury, which can inturn provide clues as to the social, cultural,or environmental associations of the injury.The method outlines descriptive features forcranial and long bone fractures since thesepredominate in the paleopathological litera-ture.

Description of cranial fractures

The interpretation of the mechanism ofinjury of cranial fractures relies on a varietyof characteristics of the fracture, such as thebones involved, patterning of fracture lines,and presence of deformation (Gurdjian,1975; Gustilo, 1991; Hooper, 1969; for acomprehensive discussion of lesions of thecalvarium, see Kaufman et al., 1997). Stressfractures and fractures secondary to pathol-ogy are uncommon in the cranium. The mostcommon fractures of the cranium affect thevault and are caused by direct trauma.These can be described according to theirbasic type, usually linear, crush, or penetrat-ing (Fig. 4), which are not necessarily mutu-ally exclusive. Although vault fractures aremost common, the base, maxillae, nasalbones, orbits, and/or zygomae may be frac-tured alternatively or additionally, and thetemporomandibular joint may be traumati-cally dislocated.

Low velocity, blunt trauma to the headmay result in simple linear fractures ordepressed (crush) fractures. The kineticsinvolved may relate to acceleration injuries,

Fig. 4. Common fractures of the cranial vault. From left to right: simple linear fracture due to blunttrauma, comminuted depressed fracture due to blunt trauma, and comminuted penetrating fracture froma high velocity projectile.


in which the head is struck by an object andset in motion, or deceleration injuries, inwhich the moving head suddenly comes to ahalt. In either case, the curve of the skull atthe point of impact tends to flatten out, andas a result the force of the impact is distrib-uted over a relatively large area. The bonesurrounding the area of impact bends out-ward, and, if the deformity of the cranium isgreat enough, fracture lines begin, usuallyin the areas subjected to bending outward.The areas of bending are not uniformlycircular, since the degree and direction towhich the fracture lines extend dependsupon both the magnitude of the applied forceand the local bony architecture.

Penetrating injuries of the cranium arecharacterized by a small area of impact witha localized area of distortion and are usuallycaused by sharp-edged objects or projectiles.With higher velocity impact, the inbendingof the skull remains localized but the depthof penetration increased. As a general rule,when the area of impact decreases thestresses are more localized but greater inmagnitude and the stresses in surroundingareas diminish. The severity of impact indirect cranial trauma is usually determinedfrom the extent and separation of linearfractures, by the extent of comminution of alocalized fracture, or by the displacement ofbone fragments in penetrating wounds.

Indirect trauma injuries are relativelyrare, but may result from vertical loadingforces transmitted from the feet or buttockswhen a person falls from a height. A basilar‘‘ring’’ fracture around the foramen magnumis an example of such an injury; it reflectsimpact forces transmitted up through thecervical spine and occipital condyles. Basilarfractures through the petrous bones andfractures of the mandibular condyles havebeen observed to result from impact to thechin (Harvey and Jones, 1980).

Description of long bone fractures

In contrast to fractures of flat and irregu-lar bones, fractures of appendicular longbones (and, by extension, short bones) oftenrequire more comprehensive descriptionsince their positions in the skeleton andtheir functions make them more susceptibleto a variety of forces. Long bone fractures

from archaeological contexts can be de-scribed in a manner adapted from that usedin clinical orthopedics (e.g., Gustilo, 1991;Harkess and Ramsey, 1991; Schultz, 1990)and can be first classified as intraarticular(involving a joint, including the metaphys-eal region) or extraarticular. Intraarticularfractures are described as either linear, com-minuted, or impacted. Extraarticular frac-tures are described as linear, comminuted,or segmental.

Linear fractures fall into three subtypes,transverse, oblique, and spiral, all of whichhave been previously described. Commi-nuted fractures are categorized according tothe size of the fragments (multiple or ‘‘butter-fly’’) and the percentage of the shaft (,50%or .50%) that is involved. A butterfly frac-ture is formed from a combination of com-pression and tension stresses that result inthe separation of a triangular fragment ofbone. Segmental fractures are identified bythe multiple fracture lines that divide thebone into at least two segments along alongitudinal axis. The location of the frac-ture should be noted as occurring at theproximal end, distal end, or shaft (either theproximal, middle, or distal third of the shaftor one of the junctions thereof).

The final components of long bone frac-ture description are length, apposition (shift),rotation, and angulation (alignment), identi-fied by the acronym, LARA. Conventiondecrees that when describing the four compo-nents the distal fragment is measured inrelation to the proximal fragment. The prin-cipal aims here are to describe fractures sothat the mechanism of injury can be de-duced, and to distinguish fractures with noor slight deformity from those with markeddeformity.

Length of the bone is measured with anosteometric board and the maximum lengthis recorded (per Bass, 1987). Length is re-corded as normal, distracted, or shortened,and is determined by comparing the injuredbone to its counterpart, if possible. Distrac-tion is a lengthening of the bone and iscaused by the separation of bone fragments,often due to muscular forces. Bones them-selves may distract a fracture, however,such as when an intact ulna pulls apart thefragment ends of a fractured radius or when


a fractured tibia is associated with an intactfibula. Distraction also may be caused whentissue is caught between fragment ends. Incontrast, shortening results when muscularforces pull the fragments over each other.This typically occurs when broken boneshave not been set, often due to severe pain ormuscle spasm, or if a fracture reductionfailed because of instability.

Apposition is the percentage of bony con-tact between fragment ends in fresh injuriesand is measured on radiographs. Appositionfrom an x-ray is measured using a ruler andis expressed as a percentage, the horizontaldisplacement being a function of the surfacearea of bone. Therefore, if there is no horizon-tal displacement between the fractured boneends when healed, that is, the bone ends arein perfect alignment, the bone is 100% ap-posed. In dry bone, however, shifting of thedistal fragment in relation to the proximalend can be recorded in the absence of radio-graphs. If the bone is viewed in anatomicalposition, a medial or lateral shift may beseen; if viewed in a lateral position, anterioror posterior displacement may be observed.The shift in both the anteroposterior (AP)and lateral planes should be noted, as thebone can be displaced in both directions.

Rotation occurs when the distal fragmenthas turned relative to the proximal frag-ment. There is no measurement, but thedistal portion is recorded as being internallyor externally rotated. This is usually easilyidentifiable in dry bone, especially if theaffected bone can be compared to the contra-lateral element. If rotation is observed, theadjacent joint surfaces should be examinedsince rotation may result in osteoarthritis,or in ankylosis of a joint if ligaments weretorn in the injury.

Angulation at the fracture site is mea-sured in degrees with a goniometer. Thismeasurement is easily obtained from a radio-graph but also may be obtained from thebone. One end of the goniometer is placed onthe midline of the proximal fragment’s longi-tudinal axis, the other end on the axis of thedistal fragment with the center of the goni-ometer directly over the fracture site. Thenumber of degrees the distal fragment hasdisplaced in relation to the midline of theproximal fragment is the angulation. The

direction of movement must also be noted.In the AP view, the distal portion of thedistal fragment will move medially (varus)or laterally (valgus). In the lateral view,anterior angulation refers to the distal por-tion of the distal fragment moving anteri-orly so that the fracture site appears posteri-orly bowed. Posterior angulation refers tothe distal portion of the distal fragmentmoving posteriorly; the fracture site appearsanteriorly bowed. Degree and direction ofangulation should be measured in the APand lateral positions as both planes areoften affected.

Examples of long bone fractures

The value and application of standardizedfracture descriptions is illustrated here withthe description of radiographs from fourclinical cases at the University of AlbertaHospital in Edmonton. With known mecha-nisms of injury, treatment, and follow-up,these cases unambiguously illustrate theskeletal effects of trauma and their variabil-ity of expression. For each set of radio-graphs, fractures were noted for their type,the bone(s) involved, area of involvement,degree of healing, length, apposition, rota-tion, and angulation. Apposition data arereported to 65%. The measurement of angu-lation was found to be the most problematicand consequently all angulation data arepresented to 62°.3 The sex and age of eachpatient were recorded although other per-sonal, identifying information was not re-vealed. Although the degree of deformity issometimes used to assess the existenceand/or quality of medical treatment in thepast, these examples reinforce the observa-tion that the association is not always direct(Grauer and Roberts, 1996). Fracture inju-ries often improve with time but converselythey may deteriorate and individual re-sponses to fracture vary widely.

A direct trauma transverse fracture re-sulted when a 20-year-old male was kickedin the shin while playing soccer. Figure 5a isa lateral view radiograph taken immedi-

3In order to evaluate interobserver error all clinical cases wereindependently scored by N. Lovell and C. Prins. The error inmeasured length as 61 mm; in apposition 65%; and in angula-tion 62°; all adequate for distinguishing between none, slight,and marked deformity.


ately postinjury that shows a transverse,midshaft tibial fracture with no fibular in-volvement. The bone was perfectly alignedon both the initial x-rays and those takenmore than 5 months later (Fig. 5b). Thedegree of callus formation may appear tothose who are inexperienced with clinicalcases to be excessive, given the apparentlack of angular deformity, displacement, orcomminution, but this example is fairly typi-cal of such injury. Not all transverse frac-tures heal as nicely, however, and nonunion,despite good alignment, may often occur inthe tibia and/or fibula due to their inherentinstability when supporting the body’sweight in locomotion.

In some cases fractured bones heal well inone dimension, only to deteriorate in an-other. Initial radiographs of a 24-year-oldmale injured in a motor vehicle accidentshowed a transverse fracture at the junctionof the mid and distal thirds of the left tibia,with two fibular fractures, one at the same

level as in the tibia and another at theproximal end of the shaft. Both tibia andfibula were in perfect alignment when viewedanteroposteriorly immediately after the in-jury, although 3° of posterior angulation ofthe tibia was noted in the lateral view.Figures 6a and 6b were taken more than 4months after the injury and the fracturelines are still visible. In the AP view, thetibia has now shifted laterally by about thewidth of the bone cortex, and shows 4° ofvalgus angulation. The midshaft fibular frac-ture also displays 4° of valgus in the APplane. The lateral view, however, now showsthe tibia in good alignment. Although theeffects of high velocity vehicular accidentsmay appear to have little relevance to ar-chaeological remains, multiple fractures ofthis type have been noted in historic cases ofinjury in horse-drawn cart and carriageaccidents, and this case has obvious rel-evance for modern forensic investigations ofdry bone lesions.

Fig. 5. Transverse fracture due to direct trauma. a: Radiograph taken immediately postinjury,showing a transverse midshaft fracture without fibular involvement. b: Radiograph taken more than 5months later, showing callus around the fracture site.


Figure 7a is an immediate postinjury x-ray of a 54-year-old male who fell and suf-fered oblique, midshaft fractures of the righttibia and fibula. In the AP view there is 11°of valgus angulation in the tibia and 13° ofvalgus angulation in the fibula. The lateralview shows 10 mm of shortening in both thetibia and fibula and 1° of posterior angula-tion in the tibia. Both bones have shiftedanteriorly about the width of the bone cor-tex. Figures 7b and 7c were taken 6 monthslater. The fracture lines are still very evi-dent, little callus is seen, and the fragmentends are rounded, suggesting nonunion inboth bones. Shortening has lessened to 6mm in both the tibia and fibula. On the APview the tibia retains 11° of valgus angula-tion but the fibula now displays only 4° ofvalgus. In the lateral view, posterior angula-tion in the tibia has increased to 4°, 1° ofposterior angulation is seen in the fibula,and both bones retain their anterior shifting.

A good example of a rotation injury withno fibular involvement is a spiral fracture ofthe distal third of the right tibia in a 12-year-old male who fell off his bicycle (Fig. 8).Callus is evident since the radiographs were

taken about 2 months postinjury, and on theAP view there is 5° of valgus angulation. Thetibia is not shortened, probably because theintact fibula helped it maintain normallength. On the lateral view, there is 6° ofanterior angulation in the tibia.

ANATOMICAL SUMMARY OFFRACTURES AND DISLOCATIONS

COMMONLY SEEN INARCHAEOLOGICAL BONE

To aid the diagnosis of trauma accordingto the mechanism of injury, this section isorganized as an atlas and describes thosefractures and dislocations commonly seen inarchaeological bone according to their ana-tomical location. The possible complicationsof the injury and their affects on healing alsoare described.

Cranium

Fractures of the bones of the cranium varyconsiderably, but perhaps the most com-monly described are those involving the flatbones of the vault. Typically, the patterningof fracture lines on the cranium is correlatedwith the severity of the force: whether a

Fig. 6. Single tibial fracture with double fibular fracture. These radiographs were taken more than 4months postinjury and the fracture lines are still visible. a: Anterior view. b: Lateral view.


blow lands on the frontal, occipital, or pari-etal region, a single linear fracture lineindicates less force than does a pattern ofconcentric and radiating stellate fracturelines (Gurdjian et al., 1950). The position offracture lines can sometimes be used toidentify the point of impact. Stellate, or

star-shaped, fracture lines form at the pointof impact, for example, and radiating frac-ture lines run laterally, away from the pointof impact. Concentric heaving fractures arecaused by shearing forces and have charac-teristics, such as bevel angle, that can distin-guish between high velocity and blunt

Fig. 7. Oblique, midshaft fractures of the right tibiaand fibula. a: Immediately postinjury. Anterior view (b)and lateral view (c) 6 months postinjury. Nonunion isevident in both bones.


trauma injury (Berryman and Haun, 1996).With blunt trauma the concentric fracturesare caused by force from outside the cra-nium, which leads to beveling on the innertable, whereas with high velocity projectiletrauma the fractures are caused by pressurefrom within the cranium, which producesbeveling on the outer table. Identification ofthe point of impact and the direction of theforce becomes increasingly difficult withmore severe trauma but the sequence ofmultiple impacts usually can be determinedsince a subsequently produced fracture willnot cross a preexisting one.

Direct trauma injuries to the craniumoften occur when the head is struck by amoving object. Trauma from high velocityobjects, such as bullets and motorized ve-hicles, is seen commonly in clinical cases,but that from lower velocity objects (e.g.,bricks, rocks, bludgeons, push carts, wag-ons) is also observed today and undoubtedlyoccurred in the past. Direct trauma to thecranium also occurs if the head strikes theground after a fall or jump from a height or

when balance is lost after landing on thefeet. These low velocity impacts usuallyresult in linear fractures. Linear fracturelines tend to sweep around the thick, bonybuttresses of the cranium (i.e., the petrousbones, mastoid process, etc.) unless theyapproach these areas perpendicularly. Sincethe structurally weak areas of the craniumare most prone to develop fracture lines, theunfused cranial sutures in children willreadily separate to accommodate the forcesof impact. Alternatively, in very young chil-dren the cranial bones may bend inwardwithout fracturing and the depressed defor-mity may persist.

Clinically, blunt trauma injuries to thecranium usually cause linear fractures ofthe vault and the appearance of these frac-ture lines may help identify the point ofimpact and the mechanism of injury. Blunttrauma to the frontal bone, for example,produces fracture lines that radiate throughthe frontal sinus, the cribriform plate, andthe orbital roofs, although transverse frac-ture lines affecting the temporal regionsmay also appear. Anterior temporal impactleads to fracture lines that radiate down,across either the orbital plate or the sphe-noid-temporal region. In contrast, lateral orposterior temporal impact produces fracturelines that radiate downwards either in frontof or behind the petrous portion of thetemporal bone and extend across the cranialbase. Impact to the occipital bone usuallyproduces fracture lines that radiate down tothe foramen magnum or the jugular fora-men, and that may extend anteriorly acrossthe cranial base. Trauma to the cranial basemust be severe in order to cause a fracture,since the bone here is heavily buttressed. Abase fracture is therefore considered to rep-resent a severe injury.

After vault fractures, sphenoid fracturesare the most common clinical result of blunttrauma to the cranium (Unger et al., 1990).Unfortunately, sphenoidal structures arevery fragile and thus prone to postmortemdamage as well as to fatal consequences offracture and therefore it may be difficult toidentify sphenoid fractures in archaeologi-cal skeletons. Facial fractures, either due todirect or indirect trauma, are often verycomplex but commonly heal adequately with-

Fig. 8. Spiral fracture, 2 months postinjury. Anteriorview is on the left and lateral view is on the right.


out medical treatment. Since the zygoma,maxilla, and orbital margin are mutuallysupportive, a fracture of one of these bonesusually involves a fracture of at least one ofthe others. Fractures of the nasal bones,while usually not severe, are not uncommon.These are often called depressed fractures(e.g., Filer, 1992) although this descriptionrefers to the observed deformity of the nasalbridge, not the type of injury. Clinically,interpersonal violence often produces smallfractures of the nasal and zygomatic bones.

Crush fractures of the cranial vault arecommonly seen in archaeological humanremains and are caused by low velocitydirect trauma (Fig. 9). Lesser force is indi-cated by the lack of displacement of bonefragments, while greater force is character-ized by inward displacement. A portion ofbone might be completely detached if greatforce is applied, particularly if the object hasa small striking surface, but more often seenin archaeological remains is the incompletedetachment of the bone (Fig. 10). The frac-ture line on the ectocranium is usually ir-regular and may be comminuted, producinga cobweb or mosaic pattern. The depressedarea indicates the point of impact, fromwhich linear fractures radiate. Clinically,blows from hammers, fireplace pokers, and

the butt ends of axes are commonly respon-sible for incompletely detached depressionfractures, as are falls onto the sharp edge offurniture or concrete steps (Polson et al.,1985).

Penetrating injuries of the cranium arecaused by pointed and edged objects (e.g.,knives, swords) or by bullets. Heavy cutting-edged weapons that are used in a choppingmanner will produce crush injuries in addi-tion to penetration, and further injury maybe caused if the embedded weapon is re-moved with a twisting motion. This damageis often indicated by splintering of the bonewith outward displacement near the initialimpact site. The type and size of woundproduced by a projectile depends upon thesize of the projectile, the speed at which itstrikes the bone, and the distance it travels.Historical skeletons may exhibit evidence ofgunshot trauma, although these injurieswould be less severe in terms of bone frag-mentation and destruction than those typi-cally found in a metropolitan trauma centertoday. Early musket balls, for example, hadlow velocity characteristics due to theirspherical shape and the poor quality of gunpowder (Butler, 1971). High velocity bullets(.3,000 ft/sec) were not developed untilalmost 1900, a date that usually places

Fig. 9. Well-healed, crush fracture of the right parieto-temporal region.


human remains in a forensic rather thanarchaeological context. Details of the inter-pretation of modern gunshot wounds can befound in many textbooks on forensic medi-cine.

Possible complications of cranial fracturesinclude displacement of bone fragments (mal-union), indirect trauma injuries elsewhereon the cranium due to the transmission ofimpact force, and soft tissue damage. Thelocation of the impact determines the subse-quent consequences of the injury due to thedifferent anatomical structures in the cra-nium. Linear fractures usually involve boththe inner and outer tables of the craniumbut do not involve displacement or depres-sion of the bone and thus are often notconsidered as serious as those injuries result-ing from greater force. Complications canarise, however, due to transmission of the

force of impact, such as when direct traumato the back of the cranium produces indirecteffects on the orbital plates. The conse-quences of cranial fractures can be fatal ifthe blood vessels running along the innertables of the cranium (e.g., middle menin-geal arteries) are torn, although this compli-cation is unlikely to be detected with cer-tainty in archaeological remains.

Mandible

The mandible forms what is essentiallypart of a ‘‘ring’’ structure, and therefore afracture on one side is commonly accompa-nied by a balancing fracture on the otherside. Usually the fracture affects the horizon-tal ramus or angle on one side and thecondyle on the opposite side. Fractures atthe angle very often communicate with the

Fig. 10. Depression fracture of the cranial vault, showing radiating and concentric fracture lines.Probably due to low velocity blunt trauma.


roots of the distal molars. Very few mandibu-lar fractures in ancient skulls have beendescribed but fractures of the ascendingramus, mandibular angle, and condylar pro-cesses have been reported (Alexandersen,1967). Asymmetrical tooth wear and osteoar-thritis at the temporomandibular joint arepossible complications of jaw fractures.

Hyoid

Although the hyoid bone is not alwaysrecovered during archaeological excavations,a perimortem hyoid fracture is consideredstrongly suggestive of interpersonal vio-lence through strangulation (Maples, 1986).

Vertebrae

The most common fractures of the verte-brae are due to indirect trauma, preexistingdisease, or stress. A very distinctive verte-bral fracture is the traumatic separation ofthe neural arch from the vertebral body atthe pars interarticularis (known as spondy-lolysis; Fig. 11), which appears to be acommon consequence of habitual physicalstress (Jimenez, 1994; Merbs, 1989a, 1989b,1995, 1996). Although the spondylolysis seenmost often in clinical settings is completeseparation, comprehensive surveys of ar-chaeological skeletons indicate that the con-dition begins as incomplete stress fracturesin adolescents that may heal or, conversely,may progress to complete lysis by youngadulthood (Merbs, 1995). The condition maybe initiated by an acute overload event thatcauses microfractures, but it is generallyagreed that the determining factor is chronictrauma, with repeated stressing promotingnonunion of the microfractures. These fa-tigue fractures appear to have the greatestpopulational frequency (approaching 50%)among arctic-adapted peoples following tra-ditional lifeways, but clinically they aremost often observed among athletes andlaborers whose activities involve frequentand large stress reversals between lumbarhyperextension and lumbar flexion (Merbs,1989b, 1996). Reported sex differences in theprevalence of the condition may be activity-related (Merbs, 1989b). Spondylolysis maybe unilateral or bilateral in expression, butpredominates in the lumbosacral region,particularly L5 and, to a lesser degree, L4

(Merbs, 1996). Complete separation is fre-quently accompanied by anterior slippage ofthe vertebral body (spondylolisthesis), butfunctional complications of this are rare(Merbs, 1989b). A fracture with possibly asimilar origin is the traumatic separation ofthe tip of the spinous process of the seventhcervical or first thoracic vertebra. Referredto as ‘‘clay-shoveller’s fracture’’ (Roberts andManchester, 1995), it may result from thestrenuous muscle action associated withshovelling clay, cement, or rocks.

More common in archaeological vertebraethan stress fractures are indirect traumainjuries, such as Schmorl’s nodes. Theseresult from bulging of the disc’s nucleuspulposus, which puts pressure on the verte-bral end plate and leads to bone resorptionin the affected area. Herniation of the disctends to occur gradually in adults becausethe nucleus has lost resiliency, whereas itmay occur suddenly in younger individuals

Fig. 11. Lumbar spondylolysis. This separation ofthe neural arch and the body at the pars interarticularisis usually attributed to a fatigue fracture.


in whom the nucleus still quite gelatinous(Bullough and Boachie-Adjei, 1988). Indi-rect vertebral damage also can occur in a fallor jump onto the feet, since the force ofimpact is carried up from the lower limbsthrough the spine.

Fractures secondary to pathology are alsocommon in the vertebral column. The bestknown clinical example is that of biconcavevertebrae, which result when intervertebraldisks expand into the superior or inferiorsurfaces of vertebral bodies that have beenweakened by osteoporosis.4 Similarly, com-pression flattening of vertebral end platesdue to sparse, coarse trabeculation in thevertebral body is a classic feature of sicklecell disease. Due to the greater strength ofthe rims of the vertebral end plates, theymay be spared even when the vertebral bodyis compressed.

Although direct trauma injuries to thevertebrae are rare, hairline transverse frac-tures on, or posterior to, the superior articu-lar processes of the second cervical vertebraare worth noting since they can result fromstrangulation (Maples, 1986).

Ribs and sternum

Ribs are known to incur stress fractures,usually as a result of occupational or simi-larly habitual labor but sometimes as aconsequence of persistent coughing or vomit-ing. Most often, however, rib fractures resultfrom direct trauma, such as a blow or a fallagainst a hard object (Adams, 1987). Clini-cally, rib fractures are the most commontype of thoracic injury and are observed in60 to 70% of individuals admitted to hospitalwith blunt chest trauma (Carrero andWayne, 1989). The direction of the impactusually can be determined from the locationof the fracture, i.e., ribs are usually frac-tured near the angle if the force is appliedfrom the front; beside the spine if the force isapplied from the back; and beside both thespine and the sternum if the force is appliedfrom the sides.

The fifth to ninth ribs are most oftenfractured (Fig. 12). Fracture of the first tothird ribs and/or the sternum indicates that

the mechanism of injury was a high kineticforce. Due to the flexibility of the rib cage,particularly in the anteroposterior dimen-sion, the degree of inward displacement atimpact may have been much greater thanthat discernable postinjury, and thus softtissue damage may have been more severethan might be inferred from the damage tothe bones themselves. Such soft tissue dam-age includes laceration of the pleura, lungs,or intercostal vessels, which would havebeen largely untreatable in earlier timesand thus may point to a possible cause ofdeath. Pneumothorax (the presence of air inthe pleural cavity) and hemothorax (thepresence of blood in the pleural cavity) maybe caused by rib fractures at any level in thethoracic cage and may be similarly life-threatening. Another serious complicationof rib fracture occurs when there are at leasttwo breaks in one rib. This produces afree-floating fragment and thus a risk of

4A biconcave vertebra should not be confused with a ‘‘butterflyvertebra,’’ which is a congenital malformation.

Fig. 12. Multiple healed fractures of the left ribs.The location of the fractures, near the angle, suggeststhat the force was applied from the front.


internal damage, referred to as ‘‘flail’’ injury.Unfortunately, if the soft tissue damagecaused by rib fractures is serious enough tocause death, the bony injuries observed inthe archaeological skeleton would be identi-fiable only as perimortem fractures. Healedrib fractures in the archaeological recordthus probably represent non-life-threaten-ing trauma.

Clavicle

Clavicular fractures are most often causedby a fall onto the shoulder but occasionallyresult from a fall onto an outstretched hand.The break tends to occur at the junction ofthe middle and lateral thirds, with down-ward and medial displacement of the lateralfragment being a common complication.Since clinical treatment of clavicular frac-tures is usually limited to the use of a slingfor 1 or 2 weeks for pain relief, healedfractures often exhibit some deformity. Mal-aligned clavicular fractures in archaeologi-cal skeletons therefore do not necessarilypoint to an absence of medical treatment.

Scapula

Scapular fractures are uncommon but areusually the result of direct trauma. Both theflat and irregular portions of the scapulamay be involved. Four types are commonlydescribed in the clinical literature: 1) frac-ture of the scapular body, which may becomminuted but rarely displaced because ofthe large muscles holding the bone in place;2) fracture of the neck, which may lead todownward displacement of the glenoid; 3)and 4) fractures of the acromion and cora-coid processes, respectively, which rangefrom simple cracks to comminution andwhich may be associated with downwarddisplacement. Although none of these inju-ries is usually considered serious, a possiblecomplication of scapular fracture, especiallyif the injury occurs on the left side, ispneumothorax (Carrero and Wayne, 1989).

Humerus

The most common humeral fractures inadults affect the neck, greater tuberosity,and shaft. Neck fractures are most commonin older women in whom osteoporosis hasweakened the bone. Indirect trauma from a

fall onto an outstretched hand is the usualcause and in more than 50% of these casesthe fracture is self-stabilized through impac-tion. Direct trauma, in the form of a fall ontothe shoulder or a blow, may cause fracture ofthe greater tuberosity. Shaft fractures aremost common in the middle third of the boneand may be due to direct or indirect trauma.The proximal half of the shaft is a commonsite of fracture secondary to pathology, suchas when the bone has been invaded bymetastatic disease. Complications of hu-meral shaft fractures include displacement,nonunion, and injury to the radial nerve.

In contrast to the pattern observed inadults and described above, humeral frac-tures in children tend to occur at the distalend, affecting the supracondylar, epicondy-lar, and condylar regions. Supracondylarfractures are indirect trauma injuries causedby a fall onto an outstretched arm, withdisplacement occurring posteriorly. Compli-cations include malunion, damage to thebrachial artery, and tearing of the jointcapsule with hemorrhage into the joint andsurrounding tissues. Epicondylar fracturesare usually medial and may result fromdirect trauma but more often from avulsionby flexor muscles in a fall. If the avulsedfragment enters the joint, which occurs fre-quently in children, complications in termsof reduced function and osteoarthritis usu-ally ensue. This injury may also cause dam-age to the ulnar nerve. Condylar fracturesare uncommon but also tend to result from afall. In contrast to the medial involvement inepicondylar fractures, the lateral portion, orcapitulum, is usually involved in condylarfractures. Displacement is not uncommonand this injury is often complicated by defor-mity, nonunion, and osteoarthritis.

Ulna

Ulnar fractures are not especially com-mon clinically, but when they occur theyusually affect the olecranon or the shaft.Olecranon fractures are more common inadults and result from the direct trauma of afall onto the point of the elbow. Severityranges from a simple crack to comminutionand the injury may be complicated by non-union and osteoarthritis. Diaphyseal frac-tures can result from either direct or indi-


rect trauma (Fig. 13). They are prone tosevere displacement, malunion, nonunion,and to infection because of the bone’s proxim-ity to the surface of the skin. Fracture of theproximal shaft of the ulna is often associatedwith the dislocation of the radial head. Thisinjury is referred to as a Monteggia fracture-dislocation. It is usually caused by a fall ontoan outstretched hand with forced pronation,but it may also be caused by a blow to theback of the upper forearm (i.e., a ‘‘parry’’fracture). Deformity commonly character-izes this injury in archaeological skeletonssince the fracture cannot be properly re-duced without surgery.

Radius

Another forearm injury, the Galeazzi frac-ture-dislocation of the radius, is more com-mon clinically than is the Monteggia frac-ture-dislocation. It is also usually caused bya fall onto the hand, and similarly it isdifficult to realign without surgical interven-tion. The fracture occurs near the junction ofthe middle and distal thirds of the radialshaft and is accompanied by dislocation ofthe inferior radio-ulnar joint. The most com-mon radius fracture occurs at the distalshaft and is called the Colles’ fracture. Clini-cally, it is the most common of all fracturesin adults over the age of 40, especiallyfemales, and is nearly always caused by theindirect trauma of a fall onto the hand. Thebreak usually occurs about 2 cm above thedistal articular surface of the radius, and

the distal fragment is posteriorly displacedand usually impacted. This injury may beassociated with fracture of the styloid pro-cess of the ulna. Although fractures of thedistal radius, such as this, are one hundred-fold more frequent than are fractures of theproximal radius (Knowelden et al., 1964),fracture of the radial head also can resultfrom a fall on an outstretched hand. Ob-served mainly in young adults, it usuallyappears as a crack without displacement.Comminution is possible, however, and insuch cases the injury may be complicated byosteoarthritis. In contrast to these indirecttrauma injuries, a simple fracture of theradial shaft, somewhat less common thanthe fracture-dislocation injuries, usually re-sults from direct trauma.

Malunion is the most common complica-tion of radius fractures, and absence ofmedical treatment cannot be presumed ifdeformity is observed since redisplacementis very common clinically within a week offracture reduction. In archaeological speci-mens, at least one study has reported thatradius fractures rarely healed without defor-mity (Grauer and Roberts, 1996).

Pelvis

Isolated fractures of the pelvic bones mostcommonly appear on the superior and/orinferior ischio-pubic ramus and the wing ofthe ilium. These fractures are quite benignunless displacement occurs. More seriouspelvic fractures are those that disrupt the

Fig. 13. Healed oblique fracture of the ulnar shaft. This type of fracture suggests that the mechanismof injury was indirect trauma (e.g., falling onto an outstretched hand) rather than direct trauma (e.g.,parrying a blow).


pelvic ring through the rami or at the pubicsymphysis, with associated dislocation ofthe sacroiliac joint. The mechanism of injuryin these cases tends to be anterior-posteriorcrushing, lateral compression, or verticalshearing force. Complications are usuallyserious and would likely be life-threateningin the absence of modern medical treatment.A fracture-dislocation of the hip occurs whenthe head of the femur is driven through thefloor of the acetabulum. This is usually theresult of a heavy blow upon the lateralfemur, due to a serious fall or a similarimpact (e.g., vehicular trauma in clinicalcases). The injury tends to comminution andserious complications. A poorly understoodlesion that may result from the incompleteand temporary dislocation of the hip is theacetabular flange lesion, which appears as aflattening of the superoposterior rim of theacetabulum (Knowles, 1983). Osteoarthritisis a common sequel to any injury that in-volves the acetabulum.

Femur

Clinical data indicate that fractures of thefemoral neck and trochanteric region arevery common in the elderly and are seri-ously disabling. Femoral neck fractures areoften a consequence of osteoporosis andtherefore appear most often among olderfemales, although physical activity duringthe reproductive years (i.e., age 15 to 45)diminishes significantly the long-term riskof femoral neck fractures (Åstrom et al.,1987). When secondary to osteoporosis, femo-ral neck fractures may result from very mildtrauma, such as a stumble. They usually aredue to rotational force and cause lateralrotation and upward displacement of theshaft. Avascular necrosis is a serious compli-cation of femoral neck fracture and is causedby damage to vessels in the neck that supplythe femoral head. The result is that theblood supply to the head may rely overly onvessels in the ligamentum teres, which isonly one of three routes of blood supply andwhich is usually inadequate on its own.Necrosis is usually sufficiently advancedwithin 2 to 6 months postinjury that thehead collapses. Unless rigidly immobilized,femoral neck fractures are prone to non-union, with avascular necrosis being the

most important contributor. Osteoarthritisis another common sequel and may be due tomechanical damage at the time of injury,impairment of blood supply to the basallayer of articular cartilage, and/or to mal-alignment of a united fracture. An impactedabduction fracture of the femoral neck oc-curs less commonly, but it usually uniteswithout surgical intervention and may becomplicated only by slight shortening of thelimb and possibly by arthritis.

Fractures of the trochanteric region (i.e.,roughly between the greater and lesser tro-chanters) are common clinically but are notlikely to be observed in archaeological popu-lations because they are almost always seenin adults over 75 years of age. Decreasedphysical activity among older individuals islikely a contributing factor (Nilsson et al.,1991), further limiting their occurrence inpast populations who followed traditional,nonmechanized, lifeways.

Femoral shaft fractures are often due tosevere direct or indirect trauma. They mayoccur at any location on the shaft and maybe of any fracture type. They are compli-cated by simultaneous hip dislocation; arte-rial damage that can compromise the viabil-ity of the limb; nerve damage, especially tothe sciatic nerve; and delayed, mal-, or non-union. Four months is considered the aver-age healing time in clinical cases of shaftfractures. Associated deformity of shaft frac-tures tends to be shortening or angulation,the latter having a tendency to facilitatedevelopment of osteoarthritis at the knee.

Supracondylar fractures of the femur aremore or less transverse and located justabove the epicondylar region. Malunion isnot a common sequel in clinical cases wherethere has been proper medical treatmentbut is a possible complication in earliertimes if the knee could not be kept immobi-lized for 2 to 3 weeks. Condylar fractures areuncommon, but when these do occur theyare almost always due to direct trauma. Theseverity of the fracture can range from anundisplaced crack to complete separation ofa condyle with marked upward displace-ment. In the absence of treatment, a dis-placed fracture is likely to heal out of align-ment and osteoarthritis of the knee joint is a


probable consequence. Occasionally a trans-verse supracondylar fracture combines witha condylar fracture and forms a T-shapedfracture line that splits apart the two con-dyles.

Patella

Fractures of the patella may be caused bydirect or indirect trauma. Indirect traumacauses an avulsion fracture, usually a clean,transverse separation of the bone, due to thesudden and violent contraction of the quadri-ceps muscle. In contrast, direct trauma froma fall or a blow onto the patella tends tocause a crack fracture or a comminutedfracture. Undisplaced crack fractures tendto heal without complication because thefragments are held in position by the aponeu-rosis of the quadriceps muscle, but fracturesthat involve separation of the fragments andthose that are comminuted will produce anirregular articular surface unless surgicallyrepaired, and osteoarthritis is then an obvi-ous sequel.

Tibia and fibula

Injuries to the knee joint most commonlyinvolve the menisci and ligaments, not bones,and therefore there may be no evidence oftrauma other than soft tissue ossificationafter ligament strain or tear, or avulsion ofthe tibial spine from injury to the cruciateligaments. When fractures do occur aroundlower limb joints they tend to affect theankle, not the knee. Clinically, the bonesforming the ankle are injured more oftenthan any other bone except the distal radius.Isolated fractures of the malleolus of thetibia or fibula are especially common, andoccur with or without dislocation of thetalus. The usual mechanism of injury forthese is either abduction and/or lateral rota-tion for fractures of the lateral malleolus ofthe fibula; and adduction for fractures of themedial malleolus of the tibia. Simultaneousfractures of both malleoli are less common.Vertical compression forces can lead to frac-ture of the anterior margin of the distaltibia, or, if severe, fragmentation of thedistal tibial articular surface, the latter in-jury being prone to osteoarthritis. Fracturesand dislocations at the ankle are often com-plicated by ligament damage, which could

be identifiable in archaeological skeletonsdue to soft tissue ossification.

Most diaphyseal fractures of the leg in-volve both the tibia and fibula. If the mecha-nism of injury is an angular force, it will leadto transverse or short oblique fractures ofthe shafts at roughly the same level. If theinjury is due to a rotational force, spiralfractures will result and will occur at differ-ent levels in the two bones. Distal tibialshaft fractures, i.e., above the medial malleo-lus, are commonly accompanied by proximalfibular shaft fractures. Conservative treat-ment of tibial shaft fractures is to manuallyreduce the fracture and to minimize weight-bearing on the limb. Immobilization for 2 to3 weeks is recommended in cases of stablefractures, but for up to 6 weeks if the injuryis likely to be displaced; full union may takeas long as 4 months. Malunion is rare inclinical cases but common archaeologicallybecause of the greater likelihood of fracturesremaining unreduced and the difficulty ofimmobilizing the leg. Because it is so close tothe surface of the skin, the tibial shaft is themost common site of an open (compound)fracture and hence infection from contamina-tion. Infection can also lead to nonunion. Afibular shaft fracture is not considered seri-ous by most clinical practitioners because itunites readily and is of such little functionalimportance that surgical removal of a por-tion of the shaft is not only tolerable but isadvocated when the bone acts as a strut,distracting the fragment ends of a tibialfracture and promoting nonunion. Thus, fibu-lar shaft fractures are expected to be rare inthe archaeological record but when observedthey may be healed with deformity.

Fractures of the tibia at the knee joint areuncommon. Clinically, the most common isfracture of the lateral tibial condyle, causedby a lateral force against the knee such asexperienced by football players or by pedes-trians struck by a vehicle bumper. If thearticular surface is fragmented in the injurythen osteoarthritis is a predictable sequel.

Hand, wrist, foot, and ankle

According to clinical evidence, the irregu-lar bones most commonly fractured are thescaphoid and triquetral in the hand, and thecalcaneus in the foot (Adams, 1987). Frac-


ture of the scaphoid tends to occur in youngadults, usually due to indirect trauma froma fall onto an outstretched hand. The typicalinjury is a transverse break through the‘‘waist’’ of the bone, and should not be con-fused with a congenitally bipartite scaphoid.Scaphoid fractures are accompanied by ahigh frequency of complications, includingdelayed union, nonunion, avascular necro-sis, and osteoarthritis. Injury to the trique-tral also usually results from a fall, butcauses a chip fracture on the dorsal surfaceof the bone. Almost all calcaneal fracturesare caused by a fall from a height onto theheels, such as occurs as an occupationalhazard among builders and window clean-ers (Wells, 1976) and a result of misadven-ture among parachutists, hikers, and rockclimbers, for example. A fall from a tree, cliff,or roof of a dwelling is a possible cause ofsuch injury in past populations, although ithas been suggested that these injuries wouldbe uncommon in falls of less than 4 m (Wells,1976). The fracture may be observed as asplit or crack in the subtalar tuberosity, butmore often the articular surface of the calca-neus fails to withstand the stress and re-sults in a crush injury. Fracture lines mayradiate to the front and appear also on thecalcaneo-cuboid joint. Associated crush frac-ture of lower thoracic or upper lumbar verte-brae may be noted. The usual complicationof calcaneal fracture is osteoarthritis. Frac-tures of the talus are relatively uncommon.

Metacarpals, metatarsals, and phalangesare common sites of traumatic injury. Meta-carpals are often fractured due to longitudi-nal compression impact such as from boxing.If the fracture line enters the joint thenosteoarthritis is a likely complication. Theneck and distal shaft of metacarpals also areprone to transverse or oblique fractures,often complicated by displacement. Manualphalanges tend to exhibit spiral or trans-verse fractures of the shaft or oblique frac-tures of the base. Comminution is mostlikely in the distal phalanges. Dislocationsof phalanges are mainly due to forced hyper-extension, but at least one archaeologicalexample appears to result from forced adduc-tion with associated tearing of the medialcollateral ligament of the interphalangealjoint that made reduction impossible and led

to malalignment of the joint and subsequentankylosis (Drier, 1992). Metatarsal shaftsmay exhibit transverse or oblique fractures(Fig. 14). The base of the fifth metatarsal is acommon site of avulsion fracture caused by atwisting injury. The pedal phalanges, par-ticularly that of the great toe, often suffercomminuted crushing injuries from directtrauma.

INTERPRETING THE ULTIMATE CAUSEOF INJURY

Once the injury has been described, itsproximate cause determined, and any com-plications identified, the ultimate cause ofthe trauma can be evaluated. This evalua-tion must consider three types of informa-tion: 1) the characteristics of the fractureitself; 2) the skeletal pattern of trauma inthe individual and the population; and 3) the

Fig. 14. Healed fracture of the left fifth metatarsal,with the unaffected contralateral element for compari-son. Although the fracture line is not clear, the rotationof the head indicates a spiral fracture, probably due to acombined angular/rotational force.


social, culture historical, and/or environmen-tal context of the human remains, includingthe presence of artifacts. Clinical researchhas considered the role of many variables intrauma causation (e.g., Agarwal, 1980;Åstrom et al., 1987; Barber, 1973; Bjornstiget al., 1991; Buhr and Cooke, 1959; Busch etal., 1986; Cogbill et al., 1991; Donaldsonet al., 1990; Fife and Barancik, 1985; Fife etal., 1984; Garraway et al., 1979; Grimm,1980; Johansson et al., 1991; Jonsson et al.,1992; Jones, 1990; Knowelden et al., 1964;Madhock et al., 1993; Nilsson et al., 1991;Prince et al., 1993; Ralis, 1986; Sahlin, 1990;Shaheen et al., 1990; Zylke, 1990) and pro-vides valuable aids for the interpretation offractures in antiquity, particularly with re-gard to skeletal patterning and the contextsof injury.

The use of the characteristics of the frac-ture itself to identify the mechanism ofinjury and point to logical causes of fracturecan be illustrated by reference to a particu-larly problematic interpretation of trauma,that of the ‘‘parry’’ fracture. Although thisinjury has been described variously as asimple midshaft fracture (Lovejoy andHeiple, 1981), the result of a blow to anupraised arm during a fight (Janssens, 1970),or the fracture of both radius and ulna(Wells, 1964), the term is interpreted bymany scholars to identify the involved bone(the ulna), indicate the location of the injury(on the shaft), and imply specific social andcultural circumstances, i.e., interpersonalconflict. A true parry fracture would indeedbe caused by a direct blow to the forearm,but a proper diagnosis must rely on descrip-tions of the fracture location, fracture type,bones involved, apparent direction of force,and evidence of any complications. Althoughsimple transverse fractures of the ulna areconventionally viewed as parry fractures, aconsideration of fracture types and mecha-nisms of injury indicates that a parry frac-ture could be indicated by either a trans-verse line or a comminution; the fracturecould be closed or open; and the radius couldbe affected in addition to the ulna. The term‘‘parry fracture’’ does not appear in the clini-cal literature on fracture description, etiol-ogy, and treatment (e.g., Adams, 1987), andother terms commonly used in orthopedic

medicine should therefore be used to de-scribe relevant injuries. An example is theMonteggia fracture of the ulna with associ-ated dislocation of the proximal radius, de-scribed previously in this paper. Althoughthis term describes the type of fracture, itdoes not specify a cause. In fact, this injurycan be caused by a direct blow to the poste-rior ulnar shaft or by the indirect trauma ofa forced pronation injury. A fatigue fracturemight also manifest as a simple transversebreak, as illustrated by the case of a mid-shaft ulnar fracture caused by the repeatedphysical stress of forking manure (Kitchin,1948). Thus, although the prevalence of‘‘parry fractures’’ is often reported in paleo-pathological studies it is not always clearwhether all or any of the injuries actuallyresulted from parrying a blow.

These problems with the diagnosis of parryfractures are compounded in paleopathologi-cal interpretations of social order, particu-larly vis a vis gender relations (for a cri-tique, see Mafart, 1991). Without othersupporting evidence, Wood-Jones (1910) as-cribed female sex to ancient Nubian skel-etons that displayed forearm fractures sincehe thought the cause of the fractures wasspousal abuse by Nubian men armed withheavy staffs. Wells referred to forearm frac-tures among the ancient Nubians as indicat-ing ‘‘short tempers and aggressive conduct’’that implied wife beating or a generally lowstatus of women (Wells 1964). His accompa-nying fracture descriptions, however, referonly to the fact that the injury appeared inthe mid- or lower shaft of the ulna, oraffected the ulna and the radius. Ulnarfractures among the Cro-Magnon were simi-larly interpreted to indicate an aggressivenature (Zivanovic, 1982); however, the ac-companying photograph and x-ray of onesuch injury shows a spiral fracture, withrotation in the distal aspect. Spiral frac-tures, in particular, are unlikely to resultfrom a direct blow, but rather from a twist-ing, forced pronation injury, such as in a fall.

While features of the fracture itself mayidentify the mechanism of injury, the skel-etal pattern of fractures in an individualmay also help to clarify the probable causesof trauma. Clinically, the ulna and radiusare fractured more commonly than are any


other skeletal elements, but the cause ofinjury varies and rarely is due to assault. Acomprehensive study of long bone fracturesin Medieval Britain also indicates that theradius and ulna are the bones most com-monly affected, likely due to falls and othermishaps rather than violence (Grauer andRoberts, 1996). Indeed, fractures of the cra-nium, ribs, or hands are more likely toindicate trauma due to interpersonal vio-lence than are fractures to the forearm.Injuries that are considered to have a highspecificity for a clinical diagnosis of assaultare fractures of the skull (especially thenasal and zygomatic bones and the man-dible); and posterior rib fractures, vertebralspinous process fractures, and fractures ofhand and foot bones, which can result fromthe direct trauma of punches or kicks. Occa-sionally the palmar surfaces of the manualphalanges will exhibit healed or unhealedcutmarks, originating as defensive woundsincurred as a victim of a knife or swordattack. Although fractures that pass alongsuture lines are common on subadult skulls,the exposure of diploe in intersutural perfo-rations is a key indicator of antemortemtrauma. In addition, metaphyseal lesions,transverse fractures of the scapular acro-mion, and sternal fractures are consideredindicative of child abuse in clinical cases andmay result from shaking forces rather thandirect blows (Appleton, 1980; Brismar andTuner, 1982; Fisher et al., 1990; Fonseka,1974; Gayford, 1979; Kleinman, 1987; Shep-herd et al., 1990). According to clinical prac-titioners, the most specific finding of physi-cal abuse in both adults and children ismultiple injuries at different stages of heal-ing (Maples, 1986; Walker et al., 1997;Wilkinson and Van Wagenen, 1993).5 Mul-tiple but simultaneous fractures, in con-trast, may represent accidental trauma, asdemonstrated by the case of a Neolithic flintminer in Belgium who suffered numerousfractures and apparently died of his injurieswhen the roof of the mine collapsed on top ofhim (Knowles, 1983).

Patterns of fractures within a populationmay also be informative. Clinically, older,postmenopausal females have more frac-tures than do any other age/sex group due tothe influence of osteoporosis as a predispos-ing factor in fracture from minor, oftenindirect, trauma, and this pattern is paral-leled in an early 20th century skeletal collec-tion (Mensforth and Latimer, 1989). Whenolder females are excluded, clinical fracturerates are greatest among individuals youn-ger than 26 years of age and tend to bedetermined by their activities.Age also influ-ences the skeletal pattern of involvement.Femoral neck fractures, for example, occurcommonly in older adults but rarely in chil-dren.

Features of the physical environment alsohave been shown to influence the frequencyand nature of trauma. For example, adverseweather conditions (e.g., snow and ice) andirregular landscapes increase fracture riskfrom falls, while reduced winter daylighthours in northern latitudes increase frac-ture risk from mishaps due to limited visibil-ity. Decreased sunlight also may impaircalcium absorption and lead to fracturessecondary to osteoporosis or rickets, anddietary inadequacies of vitamin C or calciummay increase the risk of pathological frac-tures.

The sociocultural context of injuries mustalso be considered. Clinical evidence over-whelmingly indicates, for example, that mostfractures are due to daily activity ratherthan interpersonal violence or unusualevents. Traditionally, most fractures in fe-males occur in the home while most frac-tures in males occur at work or duringsports, although this trend varies accordingto country, income level, occupation, andage. High fracture risks exist in occupationsthat have been generally restricted to men,such as agriculture, mining, forestry, andconstruction. In developing nations, house-hold work such as carrying water and loadsof firewood pose high fracture risks forwomen, as do many farming activities en-gaged in by both sexes. Technology-basedtransportation, either mechanized (e.g., au-tomobiles) or unmechanized (e.g., bicyclesand horse-drawn wagons), also carries itsown fracture risk. Given circumstances such

5For further descriptions and discussions of the skeletal evi-dence for interpersonal conflict, the reader is referred to TroubledTimes: Osteological and Archaeological Evidence of Violence,edited by D. Martin and D. Frayer and forthcoming from Gordonand Breach publishers.


as these, the trauma caused by activity oroccupation may be indistinguishable fromthat due to interpersonal violence (Smith,1996; Wakely, 1996).

Finally, historical records and an under-standing of the sociocultural context of theinjuries can benefit their interpretation, suchas the explanation of fractures due to occupa-tional accident during the 1856 constructionof the Grand Trunk Railroad in Canada(Jimenez, 1994). Several cases of interpreta-tion have successfully relied on such informa-tion, such as the identification of strangula-tion from hyoid fractures and of hanging anddecapitation from vertebral injuries (Angeland Caldwell, 1984; Waldron, 1996), and thecomprehensive analyses of trauma etiologyin an aboriginal population of the Canadiannorthwest coast (Cybulski, 1992), in prehis-toric populations of the American midwest(Milner, 1995), and among soldiers of the14th century battle of Visby (Ingelmark,1939; Knowles, 1983). Often the most com-pelling evidence for interpersonal violencelies in the presence of artifacts or otherphysical evidence of violence. Projectilepoints embedded in bone or recovered fromthe abdominal cavity are diagnostic of atleast some traumatic events, and may pointlogically to the cause of other injuries as well(e.g., Bennike, 1985; Jurmain, 1991; Smith,1996; Walker, 1989; and several papers inOwsley and Jantz, 1994).

CONCLUSIONS

This review of fracture types and of theproximate and ultimate causes of injuryindicates that standardized descriptive pro-tocols can be used to improve paleopathologi-cal analysis and interpretation of trauma.These protocols should include identifica-tion of the skeletal element(s) involved; thelocation of the injury; its appearance; andany evidence for complications of the injury.These descriptions thus serve as a basis forinferences about the mechanism of injury,from which social, cultural, or environmen-tal associations may then be examined. At-tention to skeletal patterns, biological fac-tors such as age and sex, and variables suchas physical and sociocultural environmentsalso improves the diagnostic accuracy of anyinterpretations. In particular, the traumatic

effects of violence may be difficult to distin-guish from those of high-risk activities oroccupations solely on the basis of skeletalevidence. Although the actual cause oftrauma displayed by an archaeological skel-eton may remain unknown, a quantifiabledescription based on specific terminology,supplemented by photographic and radio-graphic images of the fracture, and clearlyplaced in an individual, populational, socio-cultural, and physical context will enableothers to evaluate a researcher’s interpreta-tion of the injury.

ACKNOWLEDGMENTS

I thank Carolyn Prins, who selected andevaluated the clinical case studies and whoconducted some of the library research; theUniversity of Alberta Hospital, which pro-vided the clinical fracture data; MargaretJudd, who provided helpful discussion andnumerous references; Ben Lees, who sup-plied useful information on firearms; andScott Haddow, who assisted with printing ofphotographs. I have also benefitted greatlyfrom discussions with Robert Jurmain andLynne Kilgore and from suggestions madeby the reviewers of this paper. This researchwas supported by a grant from the SocialSciences and Humanities Research Councilof Canada.

LITERATURE CITED

Adams JC (1987) Outline of Fractures. Edinburgh:Longman.

Agarwal RP (1980) Pattern of bone injuries in a hill area(Pauri-Garhwal). J. Indian Med. Assoc. 74:65–66.

Alexandersen V (1967) The evidence for injuries to thejaws. In DE Brothwell and AT Sandison (eds.): Dis-eases in Antiquity. Springfield: CC Thomas, pp 623–629.

Altner PC, Grana L, and Gordon M (1975)An experimen-tal study on the significance of muscle tissue interposi-tion on fracture healing. Clin. Orthop. 111:269–273.

Angel JL (1974) Patterns of fracture from Neolithic tomodern times. Anthropologiai Kozlemenyek 18:9–18.

Angel JL, and Caldwell PC (1984) Death by strangula-tion: A forensic anthropological case from Wilmington,Delaware. In TA Rathbun and JE Buikstra (eds.):Human Identification: Case Studies in Forensic An-thropology. Springfield: CC Thomas, pp 168–175.

Apley AG, and Solomon L (1992) Concise System ofOrthopaedics and Fractures. Oxford: Butterworths.

Appleton W (1980) The battered woman syndrome. Ann.Emerg. Med. 9:84–91.

Åstrom J, Ahlqvist S, Beertema J, and Jonsson B (1987)Physical activity in women sustaining fracture of theneck of the femur. J. Bone Jt. Surg. [Br.] 69:381–383.

Barber HM (1973) Horse-play: Survey of accidents withhorses. Br. Med. J. 3:532–534.


Bass WM (1987) Human Osteology. 3rd ed. Columbia:Missouri Archaeological Society.

Bennike P (1985) Paleopathology of Danish Skeletons.Copenhagen: Akademisk Forlag.

Berryman HE, and Haun SJ (1996) Applying forensictechniques to interpret cranial fracture patterns in anarchaeological specimen. Int. J. Osteoarchaeol. 6:2–9.

Bjornstig U, Eriksson A, and Ornehult L (1991) Injuriescaused by animals. Injury 22:295–298.

Blair C (ed.) (1983) Pollard’s History of Firearms. Mid-dlesex: Country Life.

Bramblett CA (1967) Pathology in the Darajani baboon.Am. J. Phys. Anthropol. 26:331–340.

Brismar B, and Tuner K (1982) Battered women: Asurgical problem. Acta Chir. Scand. 148:103–105.

Buhr AJ, and Cooke AM (1959) Fracture patterns.Lancet 1:531–536.

Buikstra JE, and Ubelaker DH (eds.) (1994) Standardsfor Data Collection From Human Skeletal Remains.Fayetteville: Arkansas Archaeological Survey, Re-search Series No. 44.

Bullough PG, and Boachie-Adjei O (1988) Atlas of SpinalDiseases. Philadelphia: JB Lippincott.

Burrell LL, Mass MC, and Van Gerven DP (1986)Patterns of long-bone fracture in two Nubian cemeter-ies. Hum. Evol. 1:495–506.

Burrows JH (1956) Fatigue infraction of the middle ofthe tibia in ballet dancers. J. Bone Jt. Surg. [Br.]38:83–94.

Busch HM, Cogbill TH, Landercasper J, and Lander-casper BO (1986) Blunt bovine and equine trauma. J.Trauma 26:559–560.

Butler DF (1971) U.S. Firearms: 1776–1875. New York:Winchester.

Carrero R, and Wayne M (1989) Chest trauma. Emerg.Med. Clin. Nort Am. 7:389–418.

Cogbill TH, Steenlage ES, Landercasper J, and StruttPJ (1991) Death and disability from agriculturalinjuries in Wisconsin: A 12-year experience with 739patients. J. Trauma 31:1632–1637.

Cohen HR, Becker MH, and Genieser NB (1974) Fatiguefracture in Hare Krishna converts. N. Y. State J. Med.74:1998–1999.

Cybulski JS (1992)AGreenville Burial Ground. Archaeo-logical Survey of Canada Mercury Series No. 146.Hull: Canadian Museum of Civilization.

Dastugue J, and Gervais V (1992) Paleopathologie duSquelette Humain. Paris: Societe Nouvelle desEditions Boubee.

Donaldson LJ, Cook A, and Thomson RG (1990) Inci-dence of fractures in a geographically defined popula-tion. J. Epidemiol. Community Health 44:241–245.

Dreier FG (1992) The palaeopathology of a finger disloca-tion. Int. J. Osteoarchaeol. 2:31–35.

Fife D, and Barancik JI (1985) Northeastern Ohiotrauma study III: Incidence of fractures. Ann. Emerg.Med. 3:244–248.

Fife D, Barancik JI, and Chatterjee BF (1984) Northeast-ern Ohio trauma study II: Injury rates by age, sex,and cause. Am. J. Public Health 74:473–478.

Filer J (1992) Head injuries in Egypt and Nubia: Acomparison of skulls from Giza and Kerma. J. Egypt.Archaeol. 78:281–285.

Fisher EB, Kraus H, and Lewis VL (1990) Assaultedwomen: Maxillofacial injuries in rape and domesticviolence. Plast. Reconstr. Surg. 86:161–162.

Fonseka S (1974) A study of wife-beating in the Camber-well area. Br. J. Clin. Prac. 28:400–402.

Garraway WM, Stauffer RN, Kurland LT, and O’FallonWM (1979) Limb fractures in a defined population 1:Frequency and distribution. Mayo Clinic Proc. 54:701–707.

Gayford JJ (1979) Battered wives. Br. J. Hosp. Med.22:496–503.

Gill GW (1994) Skeletal injuries of pioneers. In DWOwsley and RL Jantz (eds.): Skeletal Biology in theGreat Plains. Washington, DC: Smithsonian Institu-tion Press, pp 159–172.

Grauer AL, and Roberts CA (1996) Paleoepidemiology,healing, and possible treatment of trauma in theMedieval cemetery population of St. Helen-on-the-Walls, York, England.Am. J. Phys.Anthropol. 100:531–544.

Grimm H (1980) Sex differences in the frequency ofbone fracture in historic and prehistoric times. In ISchwidezkef, B Chiarelli, and O Necrason (eds.):Physical Anthropology of European Populations. TheHague: Mouton, pp 347–349.

Gurdjian ES (1975) Impact Head Injury. Sprinfield: CCThomas.

Gurdjian ES, Webster JE, and Lissner HR (1950) Themechanism of skull fracture. Radiology 54:314–339.

Gustilo RB (1991) The Fracture Classification Manual.St. Louis: Mosby.

Hamperl H (1967) The osteological consequences ofscalping. In DR Brothwell and AT Sandison (eds.):Diseases in Antiquity. Springfield: CC Thomas, pp630–634.

Harkess JW, and Ramsey WC (1991) Principles offractures and dislocations. In CARockwood, DP Green,and RW Bucholz (eds.): Fractures in Adults. 3rd ed.Philadelphia: JB Lippincott, pp 1–180.

Harvey FH, and Jones AM (1980) ‘‘Typical’’ basal skullfracture of both petrous bones: An unreliable indicatorof head impact site. J. Forensic Sci. 25:280–286.

Heppenstall RB (1980) Delayed union, nonunion, andpseudoarthrosis. In RB Heppenstall (ed.): FractureTreatment and Healing. Philadelphia: WB Saunders,pp 80–96.

Hooper R (1969) Patterns of Acute Head Injury. London:Edward Arnold.

Ingelmark BE (1939) The Skeletons. In B Thordeman(ed.): Arms and Armour from the Battle of Wisby,1361. Stockholm: Vitterhets Historie Och AntikvitetsAkademien, pp 149–209.

Janssens PA (1970) Paleopathology. London: Curwen.Jimenez SB (1994) Occupational hazards in 19th-

century Upper Canada. In A Herring and L Chan(eds.): Strength in Diversity. Toronto: Canadian Schol-ars, pp 345–364.

Johansson C, Mellstrom D, Milsom I, Rundgren Å, andEkelund P (1991) Prevalence of fractures among10,000 women from the 1900 to 1940 birth cohortsresident in Gothenberg. Maturitas 14:65–74.

Jones MW (1990) A study of trauma in an Amishcommunity. J. Trauma 30:899–902.

Jonsson B, Gardsell P, Johnell O, Redlund-Johnell I, andSernbo I (1992) Differences in fracture pattern be-tween an urban and rural population: A comparativepopulation-based study in southern Sweden. Osteopo-rosis Int. 2:269–273.

Jurmain RD (1989) Trauma, degenerative disease, andother pathologies among the Gombe chimpanzees.Am. J. Phys. Anthropol. 80:229–237.

Jurmain RD (1991) Paleoepidemiology of trauma in aprehistoric central California population. In DJ Ort-ner and AC Aufderheide (eds.): Human Paleopathol-ogy: Current syntheses and future options. Washing-ton, DC: Smithsonian Institution Press, pp 241–248.

Karlstrom G, and Olerud S (1974) Fractures of the tibialshaft: A critical evaluation of treatment alternatives.Clin. Orthop. 105:82–115.

Kaufman MH, Whitaker D, and McTavish J (1997)Differential diagnosis of holes in the calvarium: Appli-


cation of modern clinical data to palaeopathology. J.Archaeol. Sci. 24:193–218.

Kelley JO, and Angel JL (1987) Life stresses of slavery.Am. J. Phys. Anthropol. 74:199–211.

Kitchin ID (1948) Fatigue fractures of the ulna. J. BoneJt. Surg. [Br.]. 30:622–623.

Kleinman PK (ed.) (1987) Diagnostic Imaging of ChildAbuse. Baltimore: Williams and Wilkins.

Knowelden J, Buhr AJ, and Dunbar O (1964) Incidenceof fractures in persons over 35 years of age. Br. J. Prev.Soc. Med. 18:130–141.

Knowles AK (1983) Acute traumatic lesions. In GD Hart(ed.): Disease in Ancient Man. Toronto: Clarke Irwin,pp 61–83.

Larsen CS, Huyn HP, and McEwan BG (1996) Death bygunshot: biocultural implications of trauma at Mis-sion San Luis. Int. J. Osteoarchaeol. 6:42–50.

Liston MA, and Baker BJ (1996) Reconstructing themassacre at Fort William Henry, New York. Int. J.Osteoarcheaol. 6:28–41.

Lovejoy CO, and Heiple KG (1981) The analysis offractures in skeletal populations with an examplefrom the Libben Site, Ottowa County Ohio. Am. J.Phys. Anthropol. 55:529–541.

Lovell NC (1990) Skeletal and dental pathology offree-ranging mountain gorillas. Am. J. Phys. Anthro-pol. 81:399–412.

Madhock R, Melton LJ III, Atkinson EJ, O’Fallon WM,and Lewallen DG (1993) Urban vs. rural increase inhip fracture incidence. Acta Orthop. Scand. 58:38–42.

Mafart B-Y (1991) Apport de l’etude des fractures os-seuses pour la connaissance des populations anci-ennes. Actes des 5e Journees Anthropologiques, Dos-sier de Documentation Archeologique no. 14. Paris:Editions du CNRS.

Mann RW, and Murphy SP (1990) Regional Atlas of BoneDisease. Springfield: CC Thomas.

Maples WR (1986) Trauma analysis by the forensicanthropologist. In KJ Reichs (ed.): Forensic Osteology.Springfield: CC Thomas, pp 218–228.

Mensforth RP, and Latimer BM (1989) Hamann-Toddcollection aging studies: Osteoporosis fracture syn-drome. Am. J. Phys. Anthropol. 80:461–479.

Merbs CF (1989a) Trauma. In MY Iscan and KARKennedy (eds): Reconstruction of Life From the Skel-eton. New York: Alan R. Liss, pp 161–189.

Merbs CF (1989b) Spondylolysis: Its nature and anthro-pological significance. Int. J. Anthropol. 4:163–169.

Merbs CF (1995) Incomplete spondylolysis and healing:A study of ancient Canadian Eskimo skeletons. Spine20:2328–2334.

Merbs CF (1996) Spondylolysis and Spondylolisthesis: Acost of being an erect biped or a clever adaptation?Yrbk. Phys. Anthropol. 39:201–228.

Miller M, and Miller JH (1979) Orthopaedics and Acci-dents. London: The English Universities Press.

Milner GR (1995) An osteological perspective on prehis-toric warfare. In LA Beck (ed.): Regional Approachesto Mortuary Analysis. New York: Plenum, pp 221–244.

Nilsson R, Lofman O, Berglund K, Larsson L, and TossG (1991) Increased hip-fracture incidence in the countyof Ostergotland, Sweden, 1940–1986, with forecastsup to the year 2000: An epidemiological study. Int. J.Epidemiol. 20:1018–1024.

Ortner DJ, and Putschar WGJ (1981) Identification ofPathological Conditions in Human Skeletal Remains.Smithsonian Contributions to Anthropology No. 28.Washington, DC: Smithsonian Institution Press.

Owsley DW, and Jantz RL (eds.) (1994) Skeletal Biologyin the Great Plains: Migration, Warfare, Health, andSubsistence. Washington, DC: Smithsonian Institu-tion Press.

Owsley DW, Mann RW, and Murphy SP (1991) Injuries,surgical care, and disease. In S Pfeiffer and RFWilliamson (eds.): Snake Hill: An Investigation of aMilitary Cemetery from the War of 1812. Toronto:Dundurn, pp 198–226.

Paton D (1984) Notes on Fractures. London: ChurchillLivingstone.

Perkins G (1958) Fractures and Dislocations. London:Athalone Press of the University of London.

Polson C, Gee DJ, and Knight B (1985) Essentials ofForensic Medicine. 4th ed. Toronto: Pergamon.

Prince RL, Knuiman MW, and Gulland L (1993) Frac-ture prevalence in an Australian population. Aust. J.Public Health 17:124–128.

Ralis ZA (1986) Epidemics of fractures during periods ofsnow and ice. Br. Med. J. 293:484.

Roberts C (1991) Trauma and treatment in the BritishIsles in the Historic Period: A design for multidisci-plinary research. In DJ Ortner and AC Aufderheide(eds.): Human Paleopathology: Current Synthesesand Future Options. Washington, DC: SmithsonianInstitution Press, pp 225–240.

Roberts C, and Manchester K (1995) The Archaeology ofDisease. Ithaca: Cornell University Press.

Sahlin Y (1990) Occurrence of fractures in a definedpopulation: A 1-year study. Injury 21:158–160.

Schultz AH (1937) Proportions, variability and asymme-tries of the long bones of the limbs and the clavicles inman and apes. Hum. Biol. 9:281–328.

Schultz AH (1939) Notes on diseases and healed frac-tures of wild apes. Bull. Hist. Med. 7:571–582.

Schultz RJ (1990) The Language of Fractures. 2nd ed.Baltimore: Williams and Wilkins.

Sevitt S (1981) Bone Repair and Fracture Healing inMan. Edinburgh: Longman.

Shaheen MAE, Bard AA, Al-Khudairy N, Khan FA,Mosalem A, and Sabet N (1990) Patterns of accidentalfractures and dislocations in Saudi Arabia. Injury21:347–350.

Shepherd JP, Shapland M, Pearce NX, and Scully C(1990) Pattern, severity, and aetiology of injuries invictims of assault. J. R. Soc. Med. 83:161–162.

Shermis S (1984) Domestic violence in two skeletalpopulations. Ossa 9–11:143–151.

Smith MO (1996) ‘Parry’ fractures and female-directedinterpersonal violence: Implications from the lateArchaic Period of west Tennessee. Int. J. Osteoar-chaeol. 6:84–91.

Stedman TL (1982) Stedman’s Medical Dictionary, 24thEdn. Baltimore: Williams and Wilkins.

Steinbock RT (1976) Paleopathological Diagnosis andInterpretation. Springfield: CC Thomas.

Stewart TD (1974) Nonunion of fractures in antiquity,with descriptions of five cases from the New Worldinvolving the forearm. Bull. N. Y. Acad. Sci. 50:876–891.

Swan KG, and Swan RC (1989) Gunshot Wounds:Pathophysiology and Management. Chicago: Year BookMedical Publishers.

Ubelaker DH, and Adams BJ (1995) Differentiation ofperimortem and postmortem trauma using tapho-nomic indicators. J. Forensic Sci. 40:509–512.

Unger JM, Gentry LR, and Grossman JE (1990) Sphe-noid fractures: Prevalence, sites, and significance.Radiology 175:175–180.

Urist MR, Mazet R, and McLean FC (1954) The patho-genesis and treatment of delayed union and non-union. J. Bone Jt. Surg. 36:931–967.

Wakely J (1996) Limits to interpretation of skeletaltrauma: Two case studies from Medieval Abingdon,England. Int. J. Osteoarchaeol. 6:76–83.

Waldron T (1996) Legalized trauma. Int. J. Osteoar-chaeol. 6:114–118.


Walker PL (1989) Cranial injuries as evidence of vio-lence in prehistoric southern California. Am. J. Phys.Anthropol. 80:313–323.

Walker PL, Cook DC, and Lambert PM (1997) Skeletalevidence for child abuse: A physical anthropologicalperspective. J. Forensic Sci. 42:196–207.

Wells C (1964) Bones, Bodies and Disease. London:Thames and Hudson.

Wells C (1976) Fractures of the heel bones in early andprehistoric times. Practitioner 217:214–298.

Wilkinson RG, and Van Wagenen KM (1993) Violenceagainst women: Prehistoric skeletal evidence fromMichigan. Midcont. J. Archaeol. 18:190–216.

Wilson ES, and Katz FN (1969) Stress fracture

analysis of 250 consecutive cases. Radiology 92:481–486.

Wood-Jones F (1910) Fractured bones and dislocations.In GE Smith and F Wood-Jones (eds.): The Archaeologi-cal Survey of Nubia Report for 1907–1908. Vol II:Report on the Human Remains. Cairo: National Print-ing Dept., pp 293–342.

Yamigashi M, and Yoshimura Y (1955) The biomechan-ics of fracture healing. J. Bone Jt. Surg. [Br.]37:1035–1068.

Zivanovic S (1982) Ancient Diseases: The Elements ofPaleopathology. New York: Pica.

Zylke JW (1990) As nation grows older, falls becomegreater source of fear, injury, death. JAMA 263:2021.


Documents

Trauma analysis in paleopathology