Fractures in ancient bones
Breaks found in bones excavated from archaeological sites may have occurred:
- (a) prior to death;
- (b) at or around time of death (termed perimortal injuries);
- (c) after burial, and may be divided into: (i) ancient breaks occurring whilst the bone is in the soil, and (ii) fresh breaks occurring during excavation or post-excavation processing.
Fractures which occurred some time before death are distinguished from those in categories (b) and (c) by the fact that they show bone production as part of the healing process. If the fracture was in the fairly early stages of healing when death intervened, woven bone callus will be present at the fracture site. Woven bone has a distinctive appearance (see Figure 1.4) so fractures in the early stages of healing should be distinguishable from those which occurred long before death. In the latter, callus appears as a cuff of fairly smooth bone at the fracture site.
An example of a healed fracture in a tibia from an archaeological site is shown in Figure 10.7. Looking first at Figure 10.7a, the bone at the fracture site has a fairly smooth, well-remodelled appearance, suggesting that the fracture occurred long before death. The appearance of the bone also suggests that the broken ends have united in less than perfect alignment - there appears to be some lateral displacement of the distal part. The radiograph (Figure 10.7b) shows that there has been some over-riding of the broken ends resulting in shortening of the bone. The radiograph also shows that it was a spiral fracture. Spiral fractures generally result from twisting forces so it was probably sustained as a result of a severely twisted ankle.
Diagenetic alteration tends to make archaeological bone rather fragile (see Chapter 2 and below). This means that bones are vulnerable to breakage due to careless handling during or after excavation. This damage is relatively easy to identify as it results in 'fresh' breaks; that is, the broken surfaces are white and unweathered. The fragile nature
Figure 10.7 (a) Anterior view of the right tibia of an adult showing a healed fracture in its distal third. The fracture site (arrowed) shows smooth, well-remodelled callus, (b) A radiograph of the same bone
of archaeological bone also means that breakage readily occurs as they lie in the soil, due to soil movements associated with changes in moisture content or temperature (Bod- dington, 1996), or to pressure of overburden (McGowan & Prangnell, 2015). Such breaks may be of considerable antiquity. Hence they show weathering to the broken surfaces. This weathering indicates that they are not of recent origin, but not whether they were perimortal or post-depositional events. Neither fractures in category (b) nor in (c)(i) will show healing, and both types will show weathered edges. However, the differences in mechanical properties between fresh and ancient bone means that they can often be distinguished.
Dry, fresh bone contains about 10% water (Boskey & Robey, 2013), which is loosely bound within the bone's structure. In the weeks and months after death, water content normally decreases (Wieberg & Wescott, 2008; Jordana et al„ 2013). This causes alteration to the bone's mechanical properties: it loses its slight 'give' or elasticity, becoming more brittle. This change in properties is accentuated over time as the collagen degrades (Turner-Walker & Parry, 1995). As well increasing its fragility, these alterations also change the way in which it fractures. This is most evident in elements with high cortical bone content: the bones of the cranial vault and the diaphyses of tubular bones. In a long-bone shaft, perimortal fractures tend to be curvilinear, with broken surfaces that are smooth and have sharp edges; post-depositional breaks in long-buried bone tend to be jagged or stepped in outline with rough surfaces (Outram, 2002; Cattaneo & Cappella, 2017). Turning to the cranial vault, perimortal blunt injury tends to cause a 'plastic response' at the site of impact resulting in a depressed fracture in which bone is pushed inwards; on the inner side of the injury, fragments may be 'hinged' inward (Fleming-Farrell et al„ 2013). Linear fractures may radiate from the point of impact or else concentric fractures may surround it (Christensen & Passalacqua, 2018: 189-190); the broken surfaces tend to be smooth with straight edges (Fleming-Farrell et al„ 2013). By contrast, post-depositional breaks to the skull of a long-buried skeleton tend to be haphazard in orientation, have rough surfaces, do not produce plastic deformation, and may shatter the vault, sometimes into many fragments.
Despite the above differences, it is sometimes difficult to distinguish perimortal from post-depositional breaks. This is due to factors which include overlap in mechanical properties of fresh and ancient bone, and that taphonomic damage may obscure perimortal injuries. For example, in a study of 210 fractures of known origin (27 perimortal, 183 post-depositional) in human remains exhumed 15 years after burial, 76% could be ascribed to their correct origin (perimortal or post-depositional) (Cappella et al„ 2014). Most of the errors were for bones with only thin cortical bone (e.g. pelves and long-bone metaphyses). In another experiment, this time focused solely on the cranium and using CT images, Fleming-Farrell et al. (2013) found that fracture surface texture (rough - smooth) and fracture surface outline (straight - irregular) in each case correctly identified whether a break was perimortem or post-depositional in 92% of the 39 cases studied. Applying the criteria of plastic deformation and inward hinging of internal fragments correctly identified only 69% and 56% of cases respectively. However, these latter two features only occurred in perimortal injuries, so their presence would seem a good identifier of perimortal blunt force injury, even if their absence cannot be used to exclude it.
Since post-depositional breaks in ancient buried bone are so common, the best course is to regard all breaks as post-depositional unless application of the above criteria provides compelling evidence that they are perimortal injuries.
A long-bone showing some of the features typical of perimortal breakage is shown in Plate 7a. It is from a collection of disarticulated human remains recovered from a pit in the settlement part of the Mediaeval Wharram Percy site. Bones from the pit showed abundant evidence of burning, knife-marks indicating dismemberment of corpses, and perimortal breakage (Mays et al., 2017). The break at the proximal end of the tibia in Plate 7a is curvilinear in form, the face of the break is smooth and the edges are sharp. The observation that the broken surface shows similar weathering to the rest of the bone supports the idea that this is a perimortal break. By contrast, the break in the bone in Plate 7b shows features typical of a post-depositional break: a jagged, stepped outline and a rough broken surface. The white, unweathered surface indicates that the break occurred after excavation.
A cranial perimortal blunt injury is illustrated in the case of a Mediaeval child excavated from the churchyard at Wharram Percy. The child was about 5-6 years old at death. The lesion is a depression situated on top of the head, just to the right of the midline (Figure 10.8). The cranium is not perforated, but the bone on the external surface shows plastic defomity, being pushed down about 5 mm below the surrounding surface.
Figure 10.8 Superior view of a skull from Wharram Percy showing an unhealed blunt injury (arrowed)
Looking at the inner side of the injury (Figure 10.9), it can be seen that some of the inner table of the cranium has splintered away exposing the underlying trabecular bone. Some fragments remain attached and are hinged inward. There is no sign of healing, so the injury must have occurred at or around time of death.
In the case of the Wharram Percy child it is impossible to determine whether the cause of the injury was accident (for example a hard object falling on the head) or assault. If the latter, the nature of the wound suggests that the blow was probably delivered using a blunt weapon with a fairly small contact area (to produce the small diameter fracture). The posterior side of the wound slopes more steeply and is deeper than the anterior side (Figure 10.8). This suggests that if it was the result of a blow with a hammer or similar instrument, it was probably delivered from behind the victim. Another type of perimortal injury, where there is normally little doubt that intentional violence was the cause, is the blade injury.
When a sharp metal weapon, such as a sword, strikes bone in a living individual (or, indeed, in a fresh corpse), it tends to slice rather than shatter it. Perimortal blade injuries tend to be linear, without large irregularities to the line of the injury; their edges are generally well defined and clean; the surfaces of the cut edges tend to be fairly flat and smooth and may show polishing (Wenham, 1989; Lewis, 2008).
A blade injury on a Mediaeval skull from Ipswich, England is shown in Figure 10.10. The skull is of an adult male, and the injury takes the form of a linear cut extending from the
Figure 10.9 The inner surface of the Wharram Percy skull at the site of the injury
Figure 10.10 A skull from Ipswich showing an unhealed blade injury area of the left orbit (eye socket) diagonally across the left side of the skull, terminating in the area of the crown of the head. The two sides of the injury do not fit together well due to some distortion of the skull as a result of soil pressure as it lay in the ground. The cut surfaces are fairly flat and parts of them have a polished appearance (Figure 10.11). Had the individual survived the injury for any length of time, new bone would have begun to form at the fracture site and started to obliterate the honeycomb pattern of the trabecular bone on the cut surfaces. Figure 10.11 shows that this has not happened, the honeycomb structure is plainly visible with no signs of new bone formation. Indeed, the brain injury consequent upon the wound must have been so massive that death would have been instantaneous. The lack of fragmentation around the wound edges shows that the weapon used must have been very sharp.
When unhealed blade injuries are examined under a microscope, their cut surfaces often show fine, parallel scratch marks. An experiment was undertaken at Leicester University, England to try and investigate the causes of these (Wenham & Wakely, 1989; Wenham, 1989). Wakely and Wenham purchased fresh pigs' heads from their local
Figure 10.11 The anterior part of the cut surface of the blade injury in the Ipswich skull. The orbit (eye-socket) is on the left in the photograph. The fairly flat cut surface is smooth and has a polished appearance; this is most visible in the anterior part of the lesion, just above the orbit butcher. They used replicas of Mediaeval swords to strike the pigs' heads with blows whose trajectories were carefully recorded. The cut surfaces on the skulls were then examined microscopically and this showed that the fine, parallel scratches aligned with the passage of the blade through the bone. They appeared to be caused by minute irregularities on the blade's surface. Thus, the orientation of these parallel scratches in archaeological specimens may serve as an indication of the trajectory of the blow which caused the injury.
Examination of the surfaces of the cut on the Ipswich skull under a low-powered microscope revealed fine, parallel scratch marks which were vertically orientated. This suggests that the injury was caused by a downward blow.
A further type of perimortal injury occasionally seen on more recent archaeological material is gunshot wounds. Entrance wounds are often approximately circular; exit wounds are often larger and more irregular in shape. As with blunt injuries, fracture lines may sometimes radiate from the point of impact (Berryman & Haun, 1996; Christensen & Passalacgua, 2018: 191-193). Unlike blunt injuries, holes due to gunshot wounds do not generally show adherent, inwardly directed fragments at their margins: the speed of the bullet means that bone fragments break away rather than bend (Berryman & Haun, 1996). Unless the bullet is retained embedded in the bone, it can be difficult to distinguish gunshot wounds from post-depositional damage on simple visual examination of the remains. This point is illustrated by a 19th century skull from the London Spitalfields site (Cox et al„ 1990).
In order not to bias the scientific results, all the data on the Spitalfields skeletons were recorded 'blind' - i.e. without reference to the coffin plates and other sources which provide biographical information about many of the interred individuals. When the biographical data was tied in with the skeletons, it was noted that one of the burials was of William Leschallas, who committed suicide in 1852 by shooting himself in the head with a pistol. The skull in guestion showed an approximately circular hole in the right temple (Figure 10.12) and a larger more ragged hole in the back of the cranium. These holes looked typical of the sort of taphonomic damage customarily seen on archaeological remains, and indeed were accepted as such by several researchers who had previously examined the skull. When it was realised that the skull belonged to the suicide victim, a radiograph of it was examined and tiny opaque particles were found to be visible. These were lead fragments which had become embedded in the bone. This showed that the two holes were in fact entry and exit wounds from the gunshot. Because gunshot injuries are so difficult to distinguish from post-depositional defects on simple visual examination of the bones, radiography plays an important role in helping to make this distinction in archaeological material where gunshot wounds are suspected.
Strictly speaking, in the perimortal injuries just discussed we cannot determine whether they were delivered to a living person or to a fresh corpse. With the exception of the tibia from Wharram Percy illustrated in Plate 7a, where the specific context of the find suggests it was a mutilation visited upon a corpse (Mays et al„ 2017), common sense indicates that they must have been inflicted on living individuals and hence were cause of death. However, with some perimortal injuries, the situation is less clear.
In cemeteries in Britain dating from about the 3rd to the 7th centuries AD it is not uncommon to find a few decapitated burials (Figure 10.13). Decapitated burials can be
Figure 10.12 The right temple area of the skull of William Leschallas, excavated from the Spitalfields crypt, London. The orbit is visible on the right of the photograph, the ear canal towards the bottom left. The hole in the temple is the entry wound from the bullet which caused his death Source: Reproduced from Cox et al. (1990: Figure 1).
Figure 10.13 A decapitated burial dating to the 4th century AD from a cemetery excavated at Stanwick, England distinguished from burials in which the skull has been moved by inadvertent disturbance to a long-forgotten grave by the presence of cut or chop marks on the neck vertebrae (Figure. 10.14) and by the observation that the top few vertebrae and the mandible are found still articulated with the cranium (Figure. 10.15). Various explanations for these decapitated burials have been suggested, including executions of criminals or enemies, human sacrifices, mutilations to corpses to prevent ghosts walking and mutilations to dishonour the corpses of enemies (discussion in Philpott, 1991:77-89; Reynolds, 2009: 91-92; Tucker, 2013; 2015). The first two possibilities assume that the decapitation was cause of death, whereas the last two assume that it was inflicted upon a lifeless corpse. In the minority of cases where the head was severed by multiple fine knife cuts, it seems reasonable to suppose the it represented a post-mortem mutilation. But in most cases the method was by chopping with a sword or axe, and here it is impossible to say whether they were executions or occurred after death (although the clenched fingers in the skeleton in Figure 10.13 may suggest a traumatic death - see caption to Figure. 3.5).