*4.2. Quantification of Shape*

Geometric morphometric methods and EFA have already been used to quantify the *pars basilaris* shape changes and intrastage variability during the second and third trimesters of fetal life [43]. EFA is suitable for considering the curved morphology and small thickness of this bone, since it is difficult to digitize homologous landmarks on the surface. This difficulty, combined with the fact that the only definable landmarks are not linked to the overall object geometry, oriented us toward a mathematical description of the outline to analyze the global shape of the *pars basilaris.*

As explained by Niel et al., 2019 (pp. 40–41) [43], outline analysis (and, more specifically, Fourier descriptors) provide complex and detailed information regarding the shape. Additionally, this method has been frequently used for discriminating biological forms quantifying morphological differences [46,51,57,58,68–75], as the use of ellipses means that the shape description in EFA is global and therefore helpful for describing bones with curved edges [70,76]. This indicates that it is perfectly suited for characterizing the morphology of the *pars basilaris*.

In the development of the method, a few available landmarks were used to define the reference plane and normalize the Fourier descriptors. The normalization of the control point using GPA [46] prevents the homology problems encountered in specimen alignment on the major axis of the first ellipse, which is conventionally used for the normalization of Fourier descriptors [77]. This method was not adapted to *pars basilaris* because the ratio between the length and width changes as the child develops [11,18,37,78,79]. It has also been shown that among the various normalization methods, the one using the control point with GPA is the most appropriate to use for bones with a few homologous landmarks and circular contours [70,76], such as the *pars basilaris*.

#### *4.3. Interest in the Pars Basilaris*

Because of its early formation, between the 10 and 14 gestational weeks [11,78–85], and its robustness, the *pars basilaris* is one of the elements of the future adult occipital bone most used to establish age-at-death estimation methods for fetuses and infants. Methods using this bone generally use conventional morphometry and/or bone size ratio [11,15,18,37,78], but they do not consider the shape, which might be valuable in improving age estimation.

Thanks to geometric morphometric methods based on Cartesian landmark coordinates, some researchers have been interested in shape to document the skull base changes through development, though with no intention of age estimation. Shape is defined as the geometric properties of an object that are invariant to scale, rotation, and translation, whereas the form of an object includes both its shape and size [60,86] (Needham equation: form = shape + size) [87].

Transposed onto our biological or forensic anthropology context concerning bones, the shape corresponds to bone maturation and the size corresponds to growth. The advantage of geometric morphometric methods is their ability to precisely quantify and visualize morphological variation through powerful statistical tools [60,86]. Based on these methods, previous studies have described the fetal cranial base development as a whole [30,36,78,88,89], but the *pars basilaris* morphology has rarely been separately analyzed.

Moreover, most morphometric methods focus on a single anatomical area to estimate age. We believe that the multiplication of age estimators, in addition to increasing reliability and accuracy [90], would minimize estimation errors [32,78,91], an idea that is consistent with some previous studies [79,92,93]. For example, according to Tocheri and Molto [91], linear measurements of the *pars basilaris* make it possible to refine the estimated age according to the degree of dental eruption and the maximum length of the femoral diaphysis.

Other studies have shown that femoral length coupled with histological study and the combination of several fetal measurements (biparietal diameter, head circumference, abdominal perimeter, and femur and radius length) improve the accuracy of fetal age estimation [92,93]. Additionally, the *pars basilaris* maximum length is significantly correlated with age, crown–rump length, and humerus length [94]. These studies demonstrate that it is possible to refine age estimation through the use of conventional morphometry together with a combination of several parameters.

#### *4.4. Morphology of the Pars Basilaris*

In the literature, several authors have used traditional morphometry to demonstrate that the *pars basilaris* dimensions evolve during fetal and infant development [18,23,78,79], and the bone characteristics intensify with age [23]. The morphological characteristics of the *pars basilaris* are used not only in anatomy but also in biological anthropology, as they can give an idea about the fetal and infant age [11,15,18,37,78,79].

Using geometric morphometric methods, shape analysis confirms the increase in morphological changes from 18 to 41 gestational weeks [43]. The conclusions of our own study allow researchers to precisely quantify and visualize shape changes of the whole *pars basilaris* during prenatal development and after birth for the first time.

By studying *pars basilaris* shapes, forensic anthropologists will gain a better idea of fetus or infant ages since each maturation stage is associated with an age interval. In addition, regarding the WHO definition of viability (more than 22 amenorrhea weeks) and the term of a pregnancy, maturation stages higher than 3 can indicate whether a fetus is viable, and stages 11 and 12 are helpful for marking the term of the pregnancy.

#### *4.5. Maturation and Growth Criterion*

In our method, femoral length was chosen as the growth criterion because of its strong relationship with age, and the *pars basilaris* shapes gathered in 19 consensus stages were used to characterize maturation. The grouping of shapes into stages based on consensus shapes with overlaps enable one to obtain a logical continuity of maturation for fetuses and infants while also allowing one to compensate for the low number of individuals of certain age groups.

Growth was defined according to the maturation stages, and we used percentiles, since we sometimes had few individuals per stage. As in any inferential approach based on population sampling and because we are aware that the variability in femur size is not limited to that observed in our samples, which were sometimes of limited size, we widened the range. For this, extreme percentiles were added to either side of the 0 and 100 percentiles. As with growth charts, the use of percentiles allows for growth to be precisely "quantified" with limited statistical bias. Thus, for a given stage, if the length of the femur is below or above the extreme percentiles, growth is considered to be altered.

#### *4.6. The Two Main Advantages of This Coupling Method*

The method established in this study makes it possible to analyze the link between the biometric (growth) and physiological (maturation) age of fetuses and infants by coupling the maturation process estimated by means of the *pars basilaris* outline and the growth process estimated by means of the femoral diaphyseal length.

The results obtained from the nonpathological validation sample (B) are encouraging for the fetus and infant age-at-death estimation. We reported coupling in 90.48% of samples, so not only can our method confirm the "overall normality" of this nonpathological sample

(first advantage), but we can also be confident when using a method with femoral length to assess age (second advantage).

Only 4 out of the 42 individuals of sample B showed uncoupling, and they never exceeded a shift of two stages of *pars basilaris* maturation. According to medical reports, these individuals did not have any identified pathological conditions, but in addition to the variability that we tried to include as much as possible in our learning sample (A), several factors can explain uncoupling, such as parity [95–99], parent general height and build [100,101], and the overall progress of the pregnancy, including the exchanges between the fetus and the placenta [100,102–110]. These appear to just have a slightly different variability from our learning sample and confirm that no method can be expected to be 100% reliable due to normal human variability.

## *4.7. Pathological Uncoupling*

As previously mentioned, age estimation from femoral length may be biased since the individual may have had abnormal growth [32], which is not necessarily visible at first sight. This is particularly true when there are no visible bone deformations or malformations such as those which can be seen on fetuses with thanatophoric dysplasia type I-II, osteogenesis imperfecta type IIA, hypophosphatasia, achondrogenesis type IA-II, or diastrophic dysplasia group) [111]. For example, a small stature is found in trisomy 21 fetuses, whose femoral lengths are smaller than normal [112,113] and there are no obvious bone deformations that alert about this pathological state. Additionally, various chromosomal abnormalities or chronic utero-vascular insufficiencies can bias estimations of fetal biometric age [32].

Disease-related bone conditions are not always visible on a skeleton because, for the lesions caused by these conditions to be visible, the individual must be immunologically affected enough to allow disease development yet strong enough to survive it [114]. For example, there are no visible traces on fetal or juvenile human osteological remains of individuals affected by plague, whooping cough, smallpox, measles, scarlet fever, or even osteomyelitis or congenital syphilis, since the disease causes death before any bony stigmas can develop. Thus, childhood disease is not obviously observable from a skeleton, especially when the skeleton is moderately preserved [37].

In our study, uncoupling concerns: localized anomalies, constitutional bone diseases, growth disorders, and cerebral anomalies. Cerebral anomalies are related to size anomalies and malformations: there is one case of cerebral hypotrophy, one case of cerebral gliosis, one case of hydrocephalus, one case of bilateral frontal paraventricular cysts, one case of infection with necrotizing and viro-induced malformative ventriculoencephalitis cytomegalovirus, and one case of agenesis of the corpus callosum associated with microcephaly. Constitutional bone diseases form a heterogeneous group of conditions responsible for insufficient stature or abnormalities in the structure of the bone, whether or not associated with deformities [115]. Among these, uncoupling indicated one case of achondroplasia, one case of Ellis–van Creveld syndrome, one case of Jeune syndrome (or asphyxiating thoracic dysplasia), two cases of thanatophoric dysplasia, one case of femoral-facial syndrome, one VACTERL-type association case, and one case of harlequin ichthyosis.

For all the affected individuals, the femur growth did not match *pars basilaris* maturation. Some authors have further stated that the femoral length is the most suitable biometric parameter for distinguishing bone dysplasias: fetuses with a femur below 30% the mean for gestational age would have achondroplasia; fetuses with a femur between 40% and 60% the mean for gestational age would have thanatophoric dysplasia or type II osteogenesis imperfecta; and fetuses with a femur below 80% the mean for gestational age would be affected by hypochondroplasia, achondroplasia, or type III osteogenesis imperfecta [116].

For uncoupling in individuals with growth disorders, two individuals were found to have diabetes, one macrosomia, four IUGR, and the last one had a twin pregnancy. All these abnormalities or simple variations in growth (twin pregnancy is not necessarily a pathological pregnancy) could lead to either growth delays or advancements depending on the description of the symptoms, evidence for which can be retrieved with this method.

However, not all individuals in our pathological sample showed systematic uncoupling since the growth disorders associated with each disease depend on several factors such as their origin, their arrival during pregnancy, and their severity. This is the reason why only a few cases were detected. For example, the severity of macrosomia varies according to maternal, pregestational, and gestational diabetes, regardless of association with obesity [117,118]. Macrosomia is also associated with the mother's age (the more advanced, the higher the risk) and parity (the more pregnancies the mother has had, the greater the risk) [118]. Unfortunately, this information cannot be verified since it had not been entered into our database.

Regarding IUGR, a fetus will develop this condition if it cannot achieve its genetic potential for growth due to genetic or external phenomena modifying this potential, or because an abnormality during pregnancy causes growth restriction [119]. Again, the severity of IUGR depends on its cause, the timing of its occurrence during pregnancy, and the duration of the intrauterine aggression [119]. Generally, fetuses with IUGR catch up in terms of their height during the second year of life, often as early as one year [120–122]. A child over 3 years of age who has still not caught up to his height should be taken care of by a pediatrician endocrinologist for in-depth examinations on stature delay, with a view initiating growth hormone treatment from the age of four [121–124]. It should be added that in cases of IUGR, cerebral maturation is generally not affected [125,126].

Additionally, there are variations in growth for multiple pregnancies compared to single pregnancies. For twins, a difference in the mean weight for gestational age is noted from 30 weeks [119]. The differences in growth between twins can be explained by the type of pregnancy; if it is monochorial–biamniotic, the transfusion–transfused syndrome is the first explanation. In bichorium–biamniotic pregnancies, the difference can be explained by a malformation of one of the twins. Placental anomalies and poor fetoplacental exchanges (nutritional, hypoxic, or toxic) can also explain growth anomalies [119].

Finally, the uncoupling of individuals with one or more localized anomalies concern: Skull anomalies in four cases:


An anomaly of the limbs for one case:


An anomaly of the spine for one case:


Three cases of polymalformative syndrome:


Finally, the cases of uncoupling highlighted by our method suggest that when maturation and growth do not match, experts must be prepared for a possible anomaly or variation in growth that risks biasing the age as estimated from femoral length.

Thus, the proposed method should be used in forensic anthropology for age estimation to verify whether growth has been altered by possible pathological conditions. This appears to be crucial in forensic contexts, where age estimation should be as accurate as possible to assess viability, set at 22 weeks of amenorrhea or a weight of 500 g according to WHO recommendations, to determine whether an individual came to term and to provide an unbiased age-at-death for police investigations.

To improve this method in the future, it would be of interest to include more healthy individuals to reduce the age range for some stages in order to provide greater precision in determining the consensus shape. The inclusion of samples from various origins would also allow the method to be used in different populations, and it could also be used in a clinical setting for screening for abnormal growth.

#### **5. Conclusions**

This study was focused on characterizing the link between maturation and growth by analyzing bone shape and biometry. The use of geometric morphometric methods and elliptical Fourier analysis enabled us to precisely quantify the *pars basilaris* shape changes from 16 fetal weeks to approximately one and a half years (17.7 months) in an unprecedented way.

By considering the coupling between the maturation and growth process, it is possible to detect potential anomalies or variations in growth. It is important to remember that it is difficult to macroscopically detect bone anomalies that could alert one to this possible variation and that the application of age-at-death estimation methods can be biased since they were established from reference populations with normal development but that the targeted individuals do not necessarily meet this condition.

In cases of uncoupling, experts should be warned that living conditions have altered the development of a young individual and that the age-at-death estimation based on long bone biometry may be biased. In a forensic context, the detection of uncoupling must lead an expert to be careful in their conclusions regarding the age determined for a young juvenile.

**Author Contributions:** Conceptualization, methodology, validation, and visualization: M.N. and P.A. Software, formal analysis, investigation, data curation, and writing—original draft: M.N. Writing review and editing, supervision, and project administration: P.A. Resources: K.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** This study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of the Public Hospital of Marseille (AP-HM) (code CIL/AP-HM 2017-41) and The National Commission on Informatics and Liberties (CNIL) (code 2071641 v 0 registered on the 27 June 2017).

**Informed Consent Statement:** Informed consent was obtained from the mothers of all young subjects involved in the study.

**Data Availability Statement:** As mentioned in the informed consent signed by the mothers, the data for individuals will remain strictly confidential.

**Acknowledgments:** The authors wish to thank Clémence Delteil, Louise Corron, Emmanuelle Lesieur, and Floriane Remy for the co-establishment of the immature database of the UMR 7268 ADES (AMU, CNRS, EFS).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Abbreviations**



#### **References**


**Kamryn Keys and Ann H. Ross \***

Department of Biological Sciences, North Carolina State University, Raleigh, NC 27695, USA; kkeys@ncsu.edu **\*** Correspondence: ahross@ncsu.edu

**Simple Summary:** Human remains are often burned in an effort to conceal the identity of the victim and/or obscure traumatic injuries related to the death event. Thermal exposure can produce artifacts resembling trauma and disguise preexisting trauma. However, there is a paucity of experimental studies with varied results addressing the differentiation of thermally induced artifacts from traumatic signatures. To address this gap in the literature, we conducted a small-scale study using domestic pigs as correlates to test the impact of thermal alteration on blunt force trauma to the cranium. Two tools (e.g., hammer and crowbar) were utilized to manually inflict injuries on the human analogs before controlled burning in an outdoor environment. The results of this experiment demonstrated that the most diagnostic variable to differentiate thermally induced alternations from blunt force fractures was fracture pattern.

**Abstract:** In forensic scenarios involving homicide, human remains are often exposed to fire as a means of disposal and/or obscuring identity. Burning human remains can result in the concealment of traumatic injury, the creation of artifacts resembling injury, or the destruction of preexisting trauma. Since fire exposure can greatly influence trauma preservation, methods to differentiate trauma signatures from burning artifacts are necessary to conduct forensic analyses. Specifically, in the field of forensic anthropology, criteria to distinguish trauma from fire signatures on bone is inconsistent and sparse. This study aims to supplement current forensic anthropological literature by identifying criteria found to be the most diagnostic of fire damage or blunt force trauma. Using the skulls of 11 adult pigs (*Sus scrofa*), blunt force trauma was manually produced using a crowbar and flat-faced hammer. Three specimens received no impacts and were utilized as controls. All skulls were relocated to an outdoor, open-air fire where they were burned until a calcined state was achieved across all samples. Results from this experiment found that blunt force trauma signatures remained after burning and were identifiable in all samples where reassociation of fragments was possible. This study concludes that distinct patterns attributed to thermal fractures and blunt force fractures are identifiable, allowing for diagnostic criteria to be narrowed down for future analyses.

**Keywords:** forensic anthropology; forensic science; blunt force trauma; thermal alteration; thermal fractures

#### **1. Introduction**

Trauma interpretation is arguably one of the most valuable services a forensic anthropologist can perform to assist criminal investigative proceedings. This is evidenced by the consistent theme of trauma-focused research in the forensic anthropology literature, spanning several decades [1–12]. Special interest is, in part, due to the fact that biomechanical signatures of skeletal trauma are not fully understood [13–16]. As such, considerable research replicating traumatic force has been produced to document the resultant characteristics seen on bone [1–11,13]. Efforts to identify the source of trauma are only half of the assessment, as it is imperative for the timing of the injury to be established as well. To interpret injury timing, characteristics of the defect are noted concerning the reaction

**Citation:** Keys, K.; Ross, A.H. Identifying Blunt Force Traumatic Injury on Thermally Altered Remains: A Pilot Study Using *Sus scrofa*. *Biology* **2022**, *11*, 87. https:// doi.org/10.3390/biology11010087

Academic Editor: Andrés Moya

Received: 21 November 2021 Accepted: 22 December 2021 Published: 6 January 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

of the immediate surrounding bone [8,9,13–16]. Following visual observation of trauma, it is assigned to an ante-, peri-, or postmortem temporal context [3,4,8]. By establishing the timing of the defect, the anthropologist can provide insight into whether the injury potentially contributed to the death event [3,4,8,13–16].

Although the characteristics of blunt force impact are the focus of many studies [1–4,6–8,11,13,16], there is still much ambiguity surrounding trauma assessment. Major issues encountered when interpreting skeletal trauma include the influence of the deposition environment, endogenous and exogenous taphonomic processes, postmortem disturbance from scavengers, or relocation to secondary deposition sites [17–22]. All of these variables introduce the potential for trauma alterations that must be accounted for during skeletal analysis [17–22]. Although the variables influencing trauma interpretation differ from case to case, and even across elements of the same skeleton, the laws of bone biomechanics that guide these analyses stay constant [6,13–16]. The main consideration is that wet or living bone will respond to slow loading force (e.g., blunt force) by first absorbing the force through plastic deformation until the force overloads the bone causing it to fail (e.g., break) [6,11–14]. Plastic deformation is expressed in the bone as crushing of the cortical layer into the internal cancellous region, as the bone slowly absorbs force without exceeding its yield strength. Per contra, when a bone is exposed to rapid force, such as is seen with gunshot trauma, it will react as a more brittle material and fracture with little to no associated plastic deformation [6–8,13–16]. This brittle reaction is also characteristic of postmortem dry bone breakage [17–26]. Both plastic deformation and complete bone failure leave distinct signatures on the bone when observed both macro- and microscopically [6–8,17–26]. There is general agreement that if the biomechanics of bone's reaction to force remain as a constant variable, then interpreting the timing of traumatic injury should be possible despite postmortem taphonomic events and alterations [6–8,17–23].

Although the structural reactions of traumatized bone are well understood, postdepositional events can complicate interpretation. Taphonomic processes can introduce secondary fractures, alter fracture margins, or conceal impact sites [6,24–26]. Due to this, trauma signatures are addressed in variable depositional environments [1–11,15–28] to identify criteria that can be informative for trauma identification in specific contexts. However, few studies have addressed the influence of thermal alteration on blunt force trauma, specifically on the cranium. In forensic contexts, it is not uncommon for decedents to be disposed of by means of fire, as perpetrators of a crime often correlate the idea of a quick coverup with burning the body until only ash remains [20,21,24–28]. However, bodies exposed to fire burn slowly and are often recovered with intact skeletal or fleshed elements remaining [21,24–28]. When bodies are exposed to fire for a significant time, heat will alter the bone by degrading its organic components, leaving only the mineral structure [6,23–28]. The organic components, which are quickly dehydrated and destroyed from thermal modification, are what allow plastic deformation in living bone [6,12–15]. Therefore, thermal fractures express features similar to bone impacted by rapid force [6,17–22,25–28]. Further investigation is needed to understand the modifications caused by thermal exposure to perimortem trauma, as conclusions from the existing literature are unclear, inconsistent, and without validation [1,3–7]. Research derived for applications in forensic contexts is unique in its necessity for the method to pass the rigors required in legal proceedings [29–32]. All methods applied in forensic testing must be guided by strict sets of procedures and criteria. Thus, this research aims to identify characteristics that are indicative of fracture origin in thermally altered remains. Specifically, this paper highlights the characteristics of thermal and mechanically derived fractures of the cranium using *Sus scrofa* analogs, this being one of the most commonly traumatized regions in forensic contexts.

#### **2. Materials and Methods**

This study used 11 adult pigs as proxies for human remains, due to an established similarity in tissue thickness and structure between humans and pigs [2,3,11]. The pigs were procured from a local pork center and were humanely euthanized the morning of the

experiment with a blank bullet (following NC laws for humane slaughter) and shipped via cooling container to the pick-up site, where they were then transferred to the forensic laboratory. The samples consisted of skulls containing the complete cranium and mandible, and all were disarticulated from the axial skeleton prior to experimentation. No soft tissue was removed before experimentation, as removing tissue before blunt force trauma would not be consistent with an actual forensic event [24]. Of the 11 samples, 3 of the specimens were used as non-traumatized controls but were still subjected to burning. The remaining eight samples were divided into two groups, each containing four specimens (Table 1). One group was manually struck with a rounded crowbar and the other a flat-faced hammer. Two types of tools were used due to the different surface areas. Two samples of each group were traumatized with the head lying supine and the other two with the head positioned in the horizontal plane. This positioning allowed the recreation of forensic scenarios where a decedent is struck standing up (horizontal) or fallen (supine) with a buttressed surface creating secondary fractures opposite the initial impacts. Each specimen was struck on the frontal, zygomatic, parietal, and nasal bones. Samples were traumatized several times until fractures could be manually felt and then radiographed to document perimortem fracture patterns (Figure 1). Following radiographic documentation, the samples were taken to the burn site.

The burn site was located on the North Carolina State University dairy farm and the fire was constructed within a livestock feeding trough surrounded by cinderblock walls (Figure 2). An open-air, outdoor fire was implemented for this research, as this type of deposition is commonly encountered with forensic burning scenarios [1,5,24–26]. Materials involved in the creation and maintenance of the fire included wood logs and coals from previous fires. No accelerants were used in the process. Each sample was positioned with the head in the horizontal plane and maintained this position for the duration of the burn cycle (Figure 3). Specimens were placed directly on top of the logs in two rows, and documentation of the progressive thermal destruction was noted via photographs during the experiment. Total burn time was 1 h and 40 min, and the samples were burned until the calcined bone was seen across the samples. Once the degree of burning was sufficient to produce largely calcinated bone, the logs were removed from the fire to slowly decrease the temperature until the samples only remained on ash. The samples were left within the fire pit overnight to allow the specimens to completely cool before removal. The following morning the skulls and associated fragments were collected by hand from the pit, placed within individual containers, and returned to the laboratory for analysis.


**Table 1.** Sample distribution. Tool column indicates classification of instrument used during manual trauma. Position lists skull pose during trauma. Identifier # lists the classification system assigned to samples throughout experimentation.

**Figure 1.** Radiographic image of sample CBH1. (**a**) Pre-burn radiograph of a specimen after manual trauma with a crowbar. Red dotted lines denote the area containing blunt force trauma. (**b**) Close-up image of the area within the red rectangle. Red arrows point to areas of incomplete fractures as a result of blunt force trauma, with associated fragments still attached.

**Figure 2.** Burn site.

**Figure 3.** Sample placement.

Laboratory analysis began with preprocessing photographs of each skull to document differential soft tissue destruction and note any thermal signatures before soft tissue removal. Following photography, any loosely adhering soft tissue was removed with a fine, soft-bristled brush. The skulls were reconstructed by refitting fragments using an adhesive. Control samples were analyzed first so that features of thermal fractures could be noted and established before comparison with the traumatized specimens. Location of fracture origin and termination, fracture type, skeletal color changes, and areas of soft tissue survival were recorded. Traumatized samples were reconstructed in the same manner as the controls and their pre-burning radiographs were compared post-burning (Figures 4 and 5).

**Figure 4.** Radiographic image of specimen HH1. Red arrows point to areas of inwardly crushed bone, with displaced fragments.

**Figure 5.** Post-processing reconstruction showing fragment association retaining depressed impact area.

#### **3. Results**

All three controls showed similar thermal alterations, which consisted largely of longitudinal fractures. These longitudinal fractures were inter-connected by transverse fractures or terminated transversely into an adjacent suture. In essence, thermal fractures appeared as patterns of long, rectangular fractures all over the cranium. Further, thermal fractures were found to be associated with cranial foramina and sutures. Thermal fracture propagation consistently originated from cranial sutures or foramina and terminated into longitudinal fractures or nearby sutures Figures 6 and 7. This finding was also highlighted in the study of Macoveciuc et al. (2017), who noted that due to the lack of accessory (traumatic) fractures in controls, heat accumulation caused fractures to originate from areas of the bone that could more easily vent, in this case being foramina and sutures. Thermal degradation was further characterized by cortical flaking and patina, a result of the rapid loss of organic components in the bone, and curved transverse fractures due to tissue regression [6,7,15,17–21].

Traumatized samples featured distinct characteristics observed only in the specimens that underwent mechanical force, which included the post-burning retention of plastic deformation, and impact areas that featured comminuted fracturing (Figure 8). Only green (e.g., wet, living) bone can respond to force as plastic deformation, as the impacted surface absorbs compressive force causing the opposite (internal) surface to tear from tension [6,12–14]. In all of the samples, blunt force fractures retained the pre-burning depressed areas and, in some samples, fragments were still connected to the associated fragment through an incomplete fracture. Due to the loss of the more pliable, organic components which allow for plasticity in bone, burning bone responds as a brittle material incapable of plastic deformation [6–9]. Only traumatized regions of the crania exhibited features of depressed fractures Figure 9. When considering fracture type, blunt force impacts were almost exclusively associated with comminuted fractures, a feature absent in controls or untraumatized regions. This difference in fracture type, being primarily longitudinal or comminuted, allowed for easier identification of suspect areas of trauma. The summary of fracture type and occurrence is presented in Table 2.

**Figure 6.** Red lines denote cranial sutures. White arrows point out thermal fractures originating and terminating within other sutures or foramina.

**Figure 7.** Red lines denote cranial sutures. Red arrows point out thermal fractures originating and terminating within other sutures or foramina.

**Figure 8.** Sample CBH2 (**a**) **A** longitudinal thermal fractures interconnected with transverse fractures, **B** blunt force trauma retained showing depressed region with associated fragments, **C** comminuted fracture at impact site; (**b**) closer image of impact sites **B** and **C**.

**Figure 9.** Hammer sample (**a**) region of impact exhibiting retained plastic deformation; (**b**) closer image of depression.


**Table 2.** Summary of fracture pattern observations. √ indicates presence of feature, X indicates absence of feature.

#### **4. Discussion**

The results of this pilot study demonstrated that consistent patterns of thermal alteration were noted that allowed for the differentiation of perimortem trauma after burning. Further, we did not find that thermal alterations obscured the blunt force trauma in any of the samples. Since this study incorporated the analyses of criteria noted to be of diagnostic value in similar research [1,3,6,7], the results of these analyses are discussed further.

#### *4.1. Skeletal Biomechanics: Fracture Type & Morphology*

Post-burning analyses found that the structural reactions between wet and brittle/dry bone were maintained, as fracture type and morphology reflected the material state of the bone when fractured. Plastic deformation was identified as areas of inwardly crushed bone with associated fragments still partially or completely attached. Although thermal exposure altered impact areas, exhibited as patina and flaking on fractured surfaces, the areas of impact retained the general morphology of the impact (as noted on pre-burn radiographs). In cases where no plastic deformation was retained, impact areas could be identified through the reassociation of fragments using an adhesive. Reassociated fragments displayed impact areas of clustered comminuted fractures (a feature absent in the thermally altered controls). Thermally altered controls consistently exhibited longitudinal, transverse, combination longitudinal-transverse, patina, and curved transverse (i.e., from tissue regression), but did not display any areas of comminuted patterns or depressions due to plasticity. Blunt force samples also presented these thermal alterations. However, regions of trauma were easily identified as variations from these thermal characteristics. In highly fragmented specimens, where the structural integrity of the cranium is lost and plastic deformation is absent, we find that identifying fracture patterns after fragment reassociation most consistently indicated the presence of trauma. Diagnostic importance has been given to fracture type and morphology in previous studies, and this study supports that these variables are indicative of fracture cause [1,3,6,7].

#### *4.2. Fracture Origin and Termination*

By first assessing the controls, location patterns of thermal fracture origin and termination were established. Thermal fractures appeared to originate and terminate in areas of the skull where thermal venting was present. That is, thermal fractures could be traced to cranial foramina or sutures and terminated within adjacent foramina and sutures. This finding is consistent with other studies that conclude that the pressure of high temperatures

within the cranium is released through natural vents, or openings, within the skull [1,6,7]. The pressure and heat released from this venting cause associated fractures to appear from these openings. The samples subjected to blunt force trauma showed the same fracture location patterns. However, locations of trauma deviated from this pattern as a cluster of comminuted fractures with no clear association to suture or foramina origins. Although fracture origin is variable when caused by traumatic force, thermal fractures are consistently associated with natural areas of thermal venting.

#### *4.3. Skeletal Color Change*

Previous suggestions [6,7] regarding color change as an indicator to differentiate fracture origin (e.g., due to thermal venting or burn progression) were not found to be useful in diagnosis for this experiment. Color changes appeared inconsistent in pattern or progression, exhibited as sporadic calcined islands surrounded by charred rings. Although potentially helpful for charting thermal degradation for remains consisting of a more complete body, these variables were not found to be of diagnostic value in this study. However, our study supports previous studies that fractures propagating into green or wet bone are associated with blunt force and perimortem trauma [1,6,7].

## *4.4. Tissue Thickness and Soft Tissue Survival*

After evaluating body positioning and tissue thickness, it was observed that tissue regression and subsequent first areas of bone to burn reflected the general thickness of the tissue covering the bone, regardless of the position of the crania. The first areas to burn followed a pattern from the facial and snout regions, these being the least protected by muscle or tissue, with the thick tissues of the mandible burning last. The variability of tissue thickness across each skull created highly differential degrees of burning. Overall, facial regions were nearly calcinated while the mandible retained green bone under the lower facial muscles. Each skull consistently exhibited this pattern of tissue regression. After burning concluded, the only surviving tissues were those of the posterior portion of the mandible. Further, it appeared that fat acted as an accelerator of thermal destruction, while muscle acted as a protector. It was noted that regions of the skulls that were highly cartilaginous or fatty, such as the ears, burned more quickly than regions of the skull in more direct contact with the fire or with more densely concentrated muscle.

#### **5. Conclusions**

After incorporating analyses deemed to be of diagnostic value or indicative of blunt force trauma after thermal exposure in other studies, we found that the most valuable variable to identify the cause of fracture (e.g., blunt force or thermal alteration) is the fracture pattern. The result of this study found that tissue thickness is more indicative of thermal progression than body positioning and warrants further study when evaluating the progression of thermal destruction across skeletal elements.

**Author Contributions:** Conceptualization, K.K.; methodology, K.K. and A.H.R.; formal analysis, K.K.; writing—original draft preparation, K.K.; writing—review and editing, A.H.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** This study met the requirements of U.S. Code Title 7, Chapter 48, Code §1902a: Humane Methods of Livestock Slaughter. This study was compliant with the ethical standards of the State of North Carolina.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Images will be made available upon request.

**Acknowledgments:** The authors would like to thank Nikki Long, Abdullah Zaben, Ginny Roman, Jordan Dalton, and Johann Galilea for their assistance with sample preparation and data collection. We would also like to thank Wes Watson for his assistance with site preparation and fire management throughout the experiment.

**Conflicts of Interest:** The authors declare no conflict of interest.
