SEARCH

SEARCH BY CITATION

Keywords:

  • forensics;
  • craniofacial identification;
  • odontological identification;
  • superimposition;
  • skeleton;
  • skull;
  • teeth;
  • Romanov, Tsar Nicolay;
  • Romanova, Anastasia;
  • Russia

Abstract

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

This article describes the identification of skeletal remains attributed to the family of Tsar Nicolay Romanov and other persons buried together at a site near present-day Ekaterinburg, Russia. Detailed descriptions are given regarding the objective methods of craniofacial and odontological identification that were used. Employing computer-assisted photographic superimposition techniques and statistical analysis of morphologic and other characteristics of the specimens, this study identifies with a high likelihood of certainty the remains of the Tsar, his wife, three of his four daughters, and four household assistants. Very strong evidence is presented that the Tsar's daughter Anastasia was killed in 1918. This study demonstrates the effectiveness of the methods and trustworthiness of the results, as well as the prospects of future application of the methods for the identification of skeletonized human remains. Anat Rec (New Anat) 265:15–32, 2001. © 2001 Wiley-Liss, Inc.

The circumstances of the tragic death of the family of the last Russian Tsar Nicolay Romanov were, until recently, one of the most mysterious pages of Russian history. The collapse of the Communist regime in Russia during the last decade of the 20th Century has brought many of the facts surrounding this event to light. A brief account of the historical record follows, based upon Radzinsky (1992).

The collapse of the Communist regime in Russia during the last decade of the 20th Century has brought many of the facts surrounding this event to light.

Having abdicated March 2, 1917, Nicolay Romanov and his family were kept under house-arrest from March 9th. The Provisional Government, headed by A. Kerensky, sent the family, accompanied by servants and guards, to Tobolsk, a Siberian town on the eastern foothills of the Ural Mountains. When the Bolsheviks came to power, the new leaders, fearing that adherents of the dethroned Tsar would make efforts to free him, decided to send the family farther up the mountains; to the town of Ekaterinburg (known later as Sverdlovsk), capital of the Ural Soviet.

By the end of May 1918, the family was settled in a house of Nicolay Ipatiev, a local engineer. The much reduced household consisted, at the end, of the immediate members of the Tsar's family (age in parentheses): Tsar Nicolay II (50) and his wife Alexandra (47); their children Olga (22), Tatiana (21), Maria (19), Anastasia (17) Alexey (13); and four retainers: E. Botkin, a family doctor (61); A. Demidova, a house maid (20); I. Kharitonov, a cook (48); and A. Trupp, a valet (62). During the night between the 16th and the 17th of July, 1918, the Tsar, his family, and their retainers were killed.

According to witnesses, the execution was particularly violent. The victims were shot at close range, bayoneted, and bludgeoned with rifle butts. Afterwards, the bodies were secretly moved out of town and dropped into a mine shaft. In an effort to collapse the walls of the shaft, hand grenades were detonated on the bodies. The large number of witnesses to the disposition of the bodies caused their location to be well known throughout the town. At that time the White Russian army was advancing through the Urals. Therefore, it was decided that the remains should be exhumed and more thoroughly hidden. In doing so, an attempt was made to destroy the bodies. Some witnesses who participated in the reburial said that the bodies of Alexey and one of the females were burned to ashes, but it was decided that complete cremation of the remaining nine bodies would not be practical. The rest had sulfuric acid poured over their faces, then were buried in a pit dug into a dirt road built across a bog.

UNCOVERING THE MYSTERY

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

All of this information regarding the assassination of the last Russian Tsar and his family was kept in great secrecy for a long time. In July 1991, however, it was announced that the skeletal remains hypothetically belonging to members of the Tsar's family and their entourage had been exhumed in a forest near Ekaterinburg.

The identification of the remains was complicated by the fact that the grave contained damaged and fragmented bone lying in disorder (Fig. 1). The first problem was to determine the number of individuals, and the assignment of specific bones to specific skeletons. With the help of the archaeologist's sketch and traditional osteological methods, more than 900 osseous pieces (bones, teeth, and bone fragments) were examined. It was determined that the bones were from the skeletons of nine individuals.

thumbnail image

Figure 1. Archaeological map of Ekaterinburg grave, with designation of the numbers of the skeletons and their depth.

Download figure to PowerPoint

The second problem was to identify the individuals associated with each of the skeletons. This required the comparison of unique characteristics of each skeleton with such characteristics known about a person while they were alive. Unfortunately, the only antemortem documents available for comparisons with the skeletons were photographic portraits. There were no trustworthy data concerning their stomatological (i.e., dental) records, height, build, traumas, illnesses, or congenital malformations. Nevertheless, an attempt at identification was made by matching general biological information, i.e., age, gender, race and stature, estimated for each skeleton (Table 1), with what was commonly known about the Tsar, his family and retainers (Table 2).

Table 1. General characteristics of the skeletons
SkeletonSexAgeHeight (cm)Race
1Female40–50164–168European
2Male51–60+171–177European
3Female20–25161–165European
4Maleabout 50165–169European
5Femaleabout 20166–169European
6Female20–24162–166European
7Female40–50163–168European
8Male?40–50163–167Not Determined
9Male60+173–180European
Table 2. Common characteristics of hypothetical persons (known for July 1918)
Hypothetical PersonSexAgeHeight (cm)RaceCompatible RemainsaMatch of Remains
ZviaginbMaplesc
  • a

    By skeleton number (see Table 1).

  • b

    Match of remains (by skeleton) as published in Zviagin and Zinin, 1987.

  • c

    Match of remains (by skeleton) as published in Maples and Browning, 1994.

Nikolay RomanovMale50about 170European2, 4, 8, 944
Alexandra RomanovaFemale46about 170European1, 777
Alexey RomanovMaleabout 14below 160EuropeanXX
Olga RomanovaFemaleabout 23165–170European3, 5, 633
Tatiana RomanovaFemale21165–170European3, 5, 666
Maria RomanovaFemale19165–170European3, 5, 65/X5
Anastasia RomanovaFemale17160–165European3, 5, 65/XX
E. BotkinMale61More than 170European2, 4, 8, 922
A. TruppMale62no dataEuropean2, 4, 8, 999
A. DemidovaFemale40no dataEuropean1, 711
I. KharitonovMale48no dataEuropean2, 4, 8, 988

The data given in Table 1 allow the following grouping of skeletons: a) skeletons 3, 5, and 6 are young women; b) skeletons 1 and 7 are middle-aged women; c) skeletons 2, 4, 8, and 9 are mature men. These groups all share a European ancestry. A comparison with what was commonly known of the characteristics of the Romanov family and their retainers demonstrates a strong compatibility with these skeletal groups (see Table 2).

The data given in Table 2 allow the following grouping of individuals: a) four young women (the Tsar's daughters); b) two middle-aged women (the Tsar's wife and her maid); c) four mature men (the Tsar, Doctor Botkin, cook Trupp and the Valet Kharitonov); d) one young man (the Tsar's son, Alexey). A comparison with the age- and gender-classed groups of skeletons demonstrates a close similarity. Only Alexey Romanov has no compatible set of skeletal remains. This supports the report that he was cremated and not included in the secondary burials of his family. Similarly, there are only five skeletons of females, although there were six females executed. This also supports reports that one of the females was cremated with Alexey. The missing female is likely one of the Tsar's daughters, because there were four daughters. Only three skeletons exist in the young woman grouping, whereas the mature males and females have a compatible number of remains in their respective groupings.

In the decades after the family's execution, additional intrigue entered into the story of the death of Tsar Nicolay II and his family when several women came forward and claimed to be the Grand Duchess Anastasia Romanova. One claimant, an “Anna Anderson,” persisted with a legal claim to the remaining Romanov fortune (Anonymous, 2000). In 1970, her suit as an heir was denied by West German courts. She lived her retiring years in Virginia, USA and died in 1984. In the 1990s, DNA forensic evidence definitively disproved her claim. Partly because of the lack of physical remains of one of the Tsar's daughters, however, there still exists with some people a belief that Anastasia indeed escaped execution and survived, a belief fostered in popular films and literature.

EARLY IDENTIFICATION EFFORTS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

Using osteological classification, Popov (1994, 1996), Zviagin and Zinin (1987), and Maples and Browning (1994; quoted in Massie, 1995) attempted to match hypothetical individuals with specific remains (Table 2). The expensive dental prostheses of skeleton 7 allowed for the identification of Alexandra Romanova vis-à-vis the maid Demidova, who was thereby identified by elimination as skeleton 1. The other identifications were made on the basis of subjective judgments regarding their similarities to photographs. Maples and Browning (1994) believed that among the daughters he could identify Olga, Tatiana and Marie, thereby identifying by elimination the missing female skeleton as Anastasia. He claimed that all skeletons were those of individuals older than 17 years because the annular epiphyses of the vertebrae were fused to their bodies. In addition, he believed that Anastasia was shorter than any of the skeletons exhumed, based on his estimation of heights through the measurement of the radii. Zviagin and Zinin (1987) could not decide whether the missing daughter was Anastasia or Maria. It should be noted that neither Maples and Browning (1994) nor Zviagin and Zinin (1987) used objective methods of comparison between postmortem remains and antemortem documentation to make their identifications.

It is our position that the classificatory method of osteological identification used was not specific enough to determine the individual relationships among the various remains in this case. Whereas it may have allowed researchers to make some broad groupings, within each group it was not dependable. The use of the annular epiphysis for aging has not been tested on a large sample of individuals and is not generally accepted as reliable. Estimation of stature is notoriously unreliable, because the antemortem stature of most individuals is not known with certainty, and the reconstruction of stature from the skeleton has a large standard error. Moreover, the results of Zviagin and Zinin (1987) and of Maples and Browning (1994) further suffer from the fact that the bones in the skeletons were not complete. Therefore, unavoidable errors may have taken place as the skeletons were assembled from a great number of dispersed and injured bones. Additionally, the methods used for determination of gender, age and height were not perfect. For example, the radius is not the bone of choice for making stature estimations if other bones are present (Trotter and Gleser, 1952). Therefore, the specific identifications of the Tsar's daughters and of the male retainers based on these two studies must be considered speculative. More objective methods were clearly necessary to more accurately identify the remains.

Gill et al. (1994) and Ivanov et al. (1996) used genetic methods of identification on these remains. Analysis of sex chromosomes confirmed the earlier osteological assessments of gender for each skeleton. Comparisons of both nuclear and mitochondrial DNA extracted from the femora and tibiae of each of the nine skeletons were compared with DNA collected from living relatives of the Tsar and his wife. These comparisons clearly demonstrated that skeleton 4 is Tsar Nicholay II, skeleton 7 is Alexandra, and skeletons 3, 5, and 6 are daughters of Alexandra and Nicholay. By elimination, skeleton 1 is A. Demidova, and also by elimination, skeletons 2, 8, and 9 are E. Botkin, I. Kharitonov, and A. Trupp. Because no samples of DNA were taken from relatives of any of the Tsar's male retainers, however, the genetic method could not identify these individuals specifically. In addition, because no antemortem samples of DNA were available, the method was not able to distinguish specifically among the Tsar's daughters.

Zhivotovsky (1999) challenged the DNA identifications on the following bases: a) the commingled conditions of the remains do not allow confident matching between the DNA samples and the complete skeletons; b) the mitochondrial DNA identified matrilineages and not individuals; c) the posterior probability of identification was based upon non-Russian demographics; d) the discovery of two loose teeth not associated with the other remains undermines confidence in the integrity of the exhumation scene.

The present study redresses the various difficulties of both the osteological and genetic analyses presented above. We make specific identifications of each of the remains by utilizing, in objective ways, the only antemortem documentation available, i.e., the photographic portraits. Two approaches are reported here: a) morphological matching of the outlines of the skulls and photographs, and b) the superimposition of the skulls and photographic portraits. Furthermore, it is shown that the use of Russian demographics supports the conclusion of Gill et al. (1994) of an extremely low probability that these remains are other than the Tsar's family and retainers. Finally, consideration of the stomatological evidence allows for the association of the two loose teeth with one of the skeletons.

MORPHOLOGICAL COMPARISONS OF THE SKULLS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

We believe that the skulls by themselves are suitable objects for examination. Most of the difficulties caused by the commingling of remains do not influence the results of the cranial examinations. In this particular case especially, the number of the skulls (nine) corresponded to the number of skeletons found in the grave. Each skull corresponded to its own post-cranial skeleton by topography of the remains in the grave, by morphological characteristics, and by gender and age attributes. Nevertheless, the identification of a skull is ipso facto the identification of the skeleton as a whole, irrespective of the fact that some individual post-cranial bones may or may not have been misclassified.

The skull is a large and complex anatomical structure. Its contours vary greatly among individuals. Some of this variation is related to the class characteristics of age, gender and race. A significant residual variation is related to the individualization of each person. Theoretically, it should be possible to match the contours of a skull with an accurate antemortem portrait. This can be done either by superimposing the image of the skull into the portrait itself (see below), or by comparing the skull's features with that of the portrait's.

Balyeva and Lebedinskaya (1991) and Gerasimov (1955) have studied the dependence of the external shape of the head and its elements upon the underlying structure of the skull. Accordingly, a special algorithm was constructed, utilizing 54 craniometric and cranioscopic characteristics that define the contours, or lineaments, of the skull or head (Snetkov et al., 1984). These lineaments may then be used to demonstrate the similarity of a specific skull to another skull or to a specific head. A mathematical method of quantitative estimation of different shapes was developed, which formed the basis for a software application of this method (Abramov et al., 1994a,b). The mathematical representation of the lineaments were invariant to the scale and the position of the object, and therefore could be used with photographs of people in various postures. The use of software for performing the comparisons minimized subjective factors in the estimation of similarities or differences in shape.

The initial step in the morphological method is the scanning of the skull or photographic image into the computer. The image is then transformed into a mathematical representation that is compared with other images, utilizing a programmer complex “TADD–TVID” (Abramov et al., 1994a,b). The result of each comparison creates a quantitative measure of similarity or difference. After all the images have been compared with each other, a frequency distribution is constructed, providing the basis for analysis of the results.

The morphological comparison was divided into two parts. For the first part, we attempted to compare the Ekaterinburg skulls with a control group of 60 unrelated skulls to determine whether the method could distinguish between the closely related skulls of the Tsar's family from those skulls of unrelated individuals (i.e., the retainers). For the second part, we attempted to compare the Ekaterinburg skulls with a selection of photographs, including those known to be of the Tsar's family and of the retainers (Figs. 2–8).

thumbnail image

Figure 2. Eight photographs of Nicolay Romanov selected for comparative examinations.

Download figure to PowerPoint

thumbnail image

Figure 3. Seven photographs of Alexandra Romanova selected for comparative examinations.

Download figure to PowerPoint

thumbnail image

Figure 4. Photographs of Olga Romanova. Images 5 and 6 were made in Tsarskoe Selo 1 year before death.

Download figure to PowerPoint

thumbnail image

Figure 5. Photographs of Tatiana Romanova. Images 3 and 4 were made 1 year before death.

Download figure to PowerPoint

thumbnail image

Figure 6. Photographs of Maria Romanova. Images 3 and 4 were made 1 year before death.

Download figure to PowerPoint

thumbnail image

Figure 7. Photographs of Anastasia Romanova. Images 4 and 5 were made 1 year before death.

Download figure to PowerPoint

thumbnail image

Figure 8. Photographs of E. Botkin, a doctor (image 1); A. Trupp, a man-servant (2); J. Kharitonov, a cook (3); and A. Demidova, a housemaid (4, 5).

Download figure to PowerPoint

The main obstacle for this method was the defective condition of the materials (Fig. 9). The following elements were partly or completely absent: upper teeth (five cases), nasal bones (five cases), one or two zygomatic bones (two cases), one or two maxillary bones (two cases), and zygomatic arches (three cases). There were huge defects in the mandible (three cases). One skull (Number 9) was presented by a facial fragment in partly damaged state (it looked like a mask); one skull (Number 8) had only a part of skullcap with upper edges of the orbits, nasal and left zygomatic processes of the frontal bone. In six of eight photographs Nikolai Romanov has a beard covering the lower part of the face (Fig. 2); E. Botkin also has a beard (Fig. 8). From the whole set of 54 variables some naturally had to be omitted from particular skulls due to these factors. Furthermore, the only undamaged or practically completely restored parts common to all the skulls were the skullcaps, and therefore only their shapes could be used for comparison. Of the nine skulls, only seven turned out to be suitable for comparison.

thumbnail image

Figure 9. Preparation of 9 restored skulls from the grave near Ekaterinburg.

Download figure to PowerPoint

The comparison of the results of craniometric and cranioscopic examinations has revealed the existence of close similarities in shapes and dimensions among Ekaterinburg Skulls 3, 5, 6, and 7. In addition to having close similarities among themselves, the results also established that this group of four skulls had remarkably stable similarities when compared with the 60 skulls from the control group. This group was found to differ significantly from the other skulls. It is therefore possible to assume with a great deal of confidence that the skulls of the skeletons 3, 5, 6, and 7 belonged to persons having close blood relationships. In addition, each of the skulls within the group was found to have some measurable difference from other members of the group, which could form the basis for their individualization.

The second part of the morphological study involved the comparisons of the Ekaterinburg skulls to photographs of the Tsar, his family members, and their retainers found in the State Archives of Russia. The reproductions of these photographs were all carefully examined to determine the consistency of their lineaments among the various portraits of each person (Zinin and Kirsanova, 1991). It was necessary to take into account a certain degree of change in a person's characteristics due to age and the quality and projection of representations. Of 76 relevant photographic portraits in the archives, 40 were selected as suitable for identification (Figs. 2–8). The 40 images were scanned into the computer and compared with the images of the Ekartinberg skulls. The results were essentially consistent with the earlier analysis (Table 3). The grouping of skeletons 3, 5, 6, and 7 as part of a family unit has support: the Tsarina identified with skeleton 7, Olga with skeleton 3, and both Maria and Anastasia with skeleton 6. Unfortunately, the photographs of Tatiana are not matched with any of the skeletons, and skeleton 5 is not matched clearly with one set of the photographs. The Tsar is identified as skeleton 4. Among the retainers, Demidova is matched with skeleton 1, Bodkin with skeleton 2, and Trupp with skeleton 9; Kharitonov is not matched.

Table 3. The results of comparisons of morphology between skulls and photographic portraits*
Hypothetical PersonSkeletons
123456789
  • *

    + indicates the cases where essential differences are absent.

Nicolay Romanov+
Alexandra Romanova+
Olga Romanova+
Tatiana Romanova
Maria Romanova+
Anastasia Romanova+
E. Botkin+
A. Demidova+
A. Trupp+
I. Kharitonov

PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

The most direct method for comparison of cranial and cephalic shapes is the superimposition of a skull into a photographic portrait. The skull is arranged in the same pose, foreshortening and scale as the photograph, and the two images are superimposed. Identity is then determined by the congruence of the two shapes and the proper anatomical placement of facial features with respect to the skull morphology. Austin-Smith and Maples (1994) have shown that the use of multiple photographs in varying postures may decrease the probability of a mismatch to less than 1%.

Attempts at superimposition were performed using all 7 of the Ekaterinburg skulls and the 40 photographs taken from the State Archives. Taking into account the results of the morphological comparisons of the lineaments, an attempt at superimposition was made for each skull with the photograph of its closest match. In addition to this, “cross” comparisons were made among the skulls and photographs that belonged to related gender and age groupings (Figs. 10–16).

thumbnail image

Figure 10. The result of photo superimposition of Skull 4 and a photograph of Nikolay Romanov. Projection correlations of 17 pairs of characteristics were examined. The results were positive.

Download figure to PowerPoint

thumbnail image

Figure 11. The result of photo superimposition of Skull 7 and a photograph of Alexandra Romanova.

Download figure to PowerPoint

thumbnail image

Figure 12. The result of photo superimposition of Skulls 3, 5 and 6 and a photograph of Olga Romanova. Results were positive for Skull 3.

Download figure to PowerPoint

thumbnail image

Figure 13. The result of photo superimposition of Skull 3, 5 and 6 (sagittal view) and a photograph of Tatiana Romanova. Results were positive for Skull 5.

Download figure to PowerPoint

thumbnail image

Figure 14. The result of photo superimposition of Skull 3, 5 and 6 (frontal view) and a photograph of Tatiana Romanova. Results were positive for Skull 5.

Download figure to PowerPoint

thumbnail image

Figure 15. The result of photo superimposition of Skull 3, 5 and 6 and a photograph of Maria Romanova. In all three cases the results were negative.

Download figure to PowerPoint

thumbnail image

Figure 16. The result of photo superimposition of Skull 3, 5 and 6 and a photograph of Anastasia Romanova. Results were positive for Skull 6.

Download figure to PowerPoint

In those cases where the initial results of the superimposition were ambiguous, two or three additional superimpositions were made by changing the main orientation points that served as a basis for determination of an angle of projection and scale of a skull representation. When transverse dimensions of the objects were identical, differences in vertical dimensions were taken into consideration, and the scale was changed in such a way that the vertical dimensions could serve as units of measure. When there were some differences in the correlation of some points of a skull projected into the head, the angle and the scale of projection were changed. By this the elimination of these differences could often be achieved. The results of the 93 photo superimpositions are given in Table 4.

Table 4. The summary results of “cross” photo superimposition*
Hypothetical PersonSkeleton
123456789
  • *

    Numbers refer to the different photographic portraits that were tested; − indicates failure at superimposition, whereas + indicates successful superimposition.

Nikolai Romanov−2+8−3−2
Alexandra Romanova+7
Olga Romanova+6−5−4
Tatiana Romanova−3+4−2
Maria Romanova−4−4−4
Anastasia Romanova−4−4+6
A. Demidova+2−2−1
E. Botkin+2−2−1
A. Trupp−1−1+1
I. Kharitonov−1+1−1

The results of the superimpositions were essentially the same as those of the morphological and osteological analyses. The Tsar (Fig. 10), his wife (Fig. 11), his daughters, Olga (Fig. 12) and Anastasia (Fig. 16), and the retainers Demidova, Botkin and Trupp came out to be the same in both the superimposition and the morphological tests. Therefore, in all likelihood these individual's remains are confirmed.

There were, however, some differences. Superimposition identified Tatiana as skeleton 5 (Figs. 13 and 14), whereas by the morphological test she was left unidentified. Similarly, Kharitonov, who was unidentified in the morphological test, was by superimposition identified with skeleton 8. Maria, on the other hand, who had been identified with skeleton 6 previously (along with Anastasia), did not match by superimposition with any skull and remained the only daughter unidentified.

STATISTICAL ANALYSIS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

The problem of identity can be stated in terms of probability theory. The method of determining the identity of the remains involved matching specific variables assessed in the skeletons with those known about the hypothetical individuals. If we assume that some of these variables are independent, and if we can estimate their frequency within the population, we can then calculate the probability of discovering 9 skeletons that exactly match our sample. The power of this approach can be demonstrated by utilizing just three of the independent parameters assessed in the osteological analyses, i.e., gender, age and geometrical similarity of a skull (Abramov et al., 1995; Abramov, 1996).

The frequency distributions of age and gender for the population of Russia on the 1st of January 1994 were collected by the state Committee for Statistics of Russia (Table 5). Reliable data for the early part of the 20th Century do not exist. We recognize that there are likely to be demographic differences between the 1918 and the 1994 populations. These differences in frequencies become mathematically insignificant, however, given the order of magnitude of the results obtained when calculating the probabilities.

Table 5. Relative frequency of males and females in different age groups
Age Group (Years)MaleFemaleAge Group (Years)MaleFemale
0–40.0310.02955–590.0270.034
5–100.0410.04060–640.0260.030
10–140.0370.03865–690.0180.031
15–190.0360.03570–740.0060.016
20–240.0350.03275–790.0040.014
25–290.0330.03280–840.0030.011
30–340.0420.04285–800.0010.005
35–390.0430.04390–940.00020.001
40–440.0380.03995–990.000030.0002
45–490.0240.026>1000.000020.00008
50–540.0240.028

Table 6 shows the calculation of the probabilities for false positive results in identifying the nine skulls. The probability of a false match by age and gender for each skeleton is derived from Table 5. By multiplying these individual probabilities we arrive at the probability of 2.625 × 10−12 that we have erroneously identified the whole group of nine skeletons. That is to say, we should expect to find the same combination of gender and age parameters in nine Russian skeletons in only about three of one trillion cases. This probability is further lowered to 1.389 × 10−15, however, when the skull-photograph superimposition is additionally factored in (the probability of a false match from this variable is derived from Austin-Smith and Maples [1994]), i.e., such a combination of parameters for nine skeletons can be met rarer than two times in 100 trillion cases.

Table 6. Calculation of the probability of false positive results of identification
SkeletonGenderAge Range (Years)Probability of False Results
By Sex and AgePhoto SuperimpositionBy All Parameters
1Female40–450.0930.30.0279
2Male50–640.1010.30.0303
3Female20–290.0640.30.192
4Male45–540.0480.30.0144
5Female15–240.0670.30.0201
6Female15–240.0670.30.0201
7Female45–540.0540.30.0162
8Male40–540.0860.80.0688
9Male60–690.0440.30.0132
The probability of false identification by all 9 skeletons2.625 × 10−121.389 × 10−15

STOMATOLOGICAL INVESTIGATIONS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

Forensic stomatological examinations of teeth and jaws of Skulls 1, 3, 4, 5, 6 and 7 (see Box 1 details for our results) were made at the Departments of Human Anatomy and of Forensic Medicine, Moscow State Medical Stomatological University. In addition, two loose teeth discovered during the exhumation were examined. A variety of methods were applied, including odontoscopy, odontometry, radiography, teleradiography and X-ray spectroscopic analysis. Dental formulae according to the method of Cherniavskaya (1983), are given for each individual.

Box 1. Data and discussion of stomatological evidence

Stomatology as a branch of forensics involves analysis of the teeth and jaws of skulls. Presented here are data from Skulls 1, 3, 4, 5, 6 and 7 describing the disposition of the teeth. For purposes of this presentation, the following dental symbols are employed: O, lost after death; Ox, lost during lifetime; U, uncut tooth; F, filling; C, caries; Fc, caries on crown edges (margins); B, broken tooth; W, crown from white metal; Y, crown from yellow metal; P, porcelain tooth; S, dowel crown (pin tooth); M, bridge false tooth; R, removable false tooth. , maxilla teeth (1–8) of right side; , maxilla teeth (1–8) of left side; , mandible teeth (1–8) of right side; ⌈, mandible teeth (1–8) of left side.

thumbnail image

Skull 1. the maxilla is absent. In the mandible there is a canine, a second premolar and a second molar in the right side; in the left side there is a bridge made of unoxidized yellow metal with a support on the ⌈3 and 8 (the 3rd and the 8th teeth). the bridge prosthesis was stabilized with soldering. The dental formula indicates a degree of tooth wear. The following defects were revealed: paradontosis, moderately expressed; chronic periodontitis; a crown and repeated caries for tooth 7⌉; teeth 8, 6⌉, and ⌈4–7 were lost during the lifetime; X-ray analysis showed a compaction of bone tissue clearly outlined with a thin clear band like a rim (odontoma) at the place of the absent tooth ⌈4. As the maxillary teeth are absent it is not possible to determine the type of bite.

thumbnail image

Skull 3. the right maxilla and mandible are absent and the skull is defective. The enamel on the teeth of the left maxilla is gray with a yellowish tone, matte, and has numerous surface cracks of various lengths; the roots are brown with surface defects to the dentine and the cement, also with small cracks. The bite is orthognathous. The skull showed the presence of initial caries in ⌊ 3, 5, 6, 7 and ⌈4, repeated caries in 7, 5 ⌉, fillings in the crowns ⌊5, 6, 7 and 7, 6, 5 ⌉ ⌈5, 7 and initial signs of paradontosis.

thumbnail image

Skull 4. the maxilla is mainly represented in the anterior areas by an atrophied alveolar process. The anterior walls of the sockets ⌊l– 4 are absent along the whole length, and there are some postmortem defects. The tooth enamel is dark gray, matte, with numerous surface cracks of various depths and lengths. The surface of the roots and neck is not plane, and there is not even a tubercle due to the defects to the cement and dentine. These changes are mainly observed in the anterior group of teeth. The changes found in the mandible demonstrate the presence of chronic odontogenic osteomyelitis in the roots ⌈5–7, periodontitis of tooth ⌈4, initial caries in the tooth ⌈5, repeated caries of the crown, neck and root of the tooth ⌈7, clearly seen paradontosis; the teeth 6⌉ and ⌈8 were lost not less than 2–3 months before death, but teeth 8⌉ and ⌈5–8 were lost long before death.

thumbnail image

Skull 5. The right maxilla is absent and the bones of the facial skeleton are defective. The enamel of the teeth is dark gray, bright in some places and matte in others, with numerous surface cracks of various depths and lengths. The roots are dark brown, with surface defects of cement and dentine. There are primary caries of the crowns in the teeth ⌊6, 6, 5⌉, and repeated caries in 6, 7⌉, ⌈6, 7, 5 and ⌊6, 7 at the places of fillings.

thumbnail image

Skull 6. The enamel of the teeth is gray or in some places brown, matte, with numerous cracks of various depths and lengths. The roots are dark gray, smooth with small surface cracks and defects of cement. The initial caries 4, 6⌋, ⌊4 and repeated caries 3, 6⌋, and 2, 3, 5, 6, 7⌉, ⌈6, 7 are observed as well as the initial signs of paradontosis. There are also amalgam and cement fillings in l, 2, 3, 5, 6, 8⌉, ⌊1, 8.

thumbnail image

Skull 7. The anterior walls of alveoli of both the maxilla and the mandible were severely damaged postmortem, leaving the tooth roots exposed along the whole length of the dentition. Considerable dental work is evident. Teeth 1⌋ and ⌊1 are presented by dowel crowns of the Richmond type, their crowns are fixed to the remaining roots of natural teeth. The vestibular surface of the crowns are faced with porcelain. X-rays demonstrate that the pins were put halfway down the root canals and were banded by fillings that are as hard as cement. Tooth 5⌋ is represented by a dowel crown of the Davis type holding on the remaining tooth root. The tooth roots are defective on the vestibular side, causing the buccal margin of the crown and the pin to be exposed. The alveolar margin is atrophied. The crown of the tooth is made of white metal, like porcelain, and is kept in place by means of a thin safe-plate that is soldered above the root to a deoxidized white metal pin by an alloy of yellow metal. The whole occlusive surface of the crown, except for its medial facet, is plane due to wear. Only in the area of the facet are the remaining signs of modeled fissures seen. The X-ray showed that the tooth pin penetrates into the widened canal by only one third of its length. The walls of the canal are coated with cement. At the apex of the root a periodontal fissure is widened as a granuloma. Tooth 7⌋ has a crown of deoxidized white metal. Perfectly modeled cusps and fissures without signs of wear can be seen on the occlusive surface. The X-rays show that the crown is held with a short pin put into the palatal root canal without any filling materials. Teeth ⌊4 and 5 have a dowel crown of the Davis type that replaced the absent crown. The crown of this pin tooth is made of material reminiscent to porcelain. Clear outlines, with places of wear, are seen on the occlusive surface. The X-rays show that the tooth is held by a pin put into an artificially widened canal. Tooth ⌊6 is a dowel crown of the Richmond type at the place of the absent crown. It has a white crown made of metal reminiscent of porcelain. Caries are seen on the vestibular surface of the tooth neck up to the middle third part of the distal neck root, then passing to the posterior surface of the tooth. The margin of the alveolar process has eroded to the level of the middle third part of the tooth root. Tooth ⌈7 has a crown made of white metal. This artificial crown covers the tooth down to the anatomical neck to which it tightly joins. Perfectly modeled cusps and fissures with signs of mechanical punching are seen on the occlusive surface of the crown. There is wear in the central crown fovea. Caries on the vestibular surface of the tooth neck occupy the proximal part of the medial root. X-rays show a cyst widening a periodontal crack with elements of resorption of the bone tissue and almost complete obliteration of the canals. The alveolar margin is atrophied and has eroded to the level of the upper third part of the root. The enamel of the mandibular teeth is gray with a yellowish tone at some places, matte, flat, bright with single longitudinal small cracks and small defects along the enamel-cement border. The dental arch is saddle-shaped, as the left teeth 6 and 7 are inclined to the tongue side. The bite is characterized by deep (incisor) overlapping of the maxilla. The tooth roots are of grayish-brown color, matte with small surface defects of cement and dentine, and also with surface cracks like a net.

Taking into consideration the status of dental science and technology from the end of the 19th Century to the beginning of the 20th Century, the prostheses found in Skulls 1 and 7 had been made by a highly qualified dentist. That is demonstrated by the following:

  • use of precious metals of very high quality (e.g., gold, platinum, silver) and their alloys;

  • dowel crowns (pin teeth) and porcelain were used in the preparation of the prostheses;

  • on the bridge prosthesis of Skull 1, soldering was performed very carefully on the small places between the crowns and artificial teeth; this soldering had undergone no deterioration even after long use;

  • metal crowns closely embraced tooth crowns down to the enamel—cement border and their anatomic shape was modeled exactly;

  • the occlusal surfaces of the prostheses were modeled very close to natural anatomic shapes of the tooth crowns to provide a normal bite.

According to the results of a microradiospectral (zondal) analysis made at the Ural Research Laboratory of Forensic Examinations of the Ministry of Justice of Russia, the materials from which all artificial fillings were made contained silver, iron, silicon; also copper and magnesium (in minimum quantities). They are all practically the same by their elemental compositions, but the elements are not distributed in equal portions in all the fillings. Thus, for example, in the places where silver content is the maximum, iron content is decreased and vice versa. The tooth prosthesis (Skull 1) is made of a gold-copper alloy, and the gold content reaches at some points as high as 96%. The pin teeth of the mandible in the Skull 5 are made practically of pure platinum, and the crown represents a gold-copper-silver alloy, where the silver content reaches 10%.

The composition and the state of the crowns of Skulls 3, 5, and 6 indicate that these persons had systematic, highly qualified odontological assistance for most of their lives. The appearance of caries on the occlusive and precervical surface of the tooth crowns, however, suggest that they had no odontological treatment during their last one or two years before death. The similarity of body shapes, mental protuberance of the mandibles as well as tooth arches, tooth inclinations, and the morphological structure of the Skulls 3, 5, 6, and 7 provide strong evidence of possible blood relationships of persons to whom the given skulls belonged. Morphological changes in the teeth and jaws of the Skulls 4 and 7 revealed long chronic inflammatory-dystrophic processes in periodontal and parodontal tissues. This suggests that the teeth of these persons were not treated for some years.

IDENTIFICATION OF LOOSE TEETH

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

The two separate teeth found at Ekaterinburg are permanent teeth (see Table 7). Their overall size, the shapes of their crowns and roots, and the degree of formation and maturity of their crown enamel identify them as maxillary molars. They both could be either from the right or the left sides; they both could be either the second (the 7th teeth) or the third (the 8th teeth). More definite conclusions of their exact position is not possible because of the great variability in the shapes and sizes of the third molars. If we assume that they are both second maxillary molars, they must belong to different persons with biological ages of 12–13 and 14–15 years respectively (Boume, 1995; Scheff Jun et al., 1989; Vorobiev et al., 1936). Accordingly, as second maxillary molars, these teeth could not belong to the tooth rows of the skulls selected for the examination because of their ages, as well as their specific morphological characteristics.

Table 7. The results of the odontometry of the teeth (in mm)
ParametersTooth 1Tooth 2
Height of a tooth15.3–15.913.8–14
Width of a crown on a level of the equator9.510
Width of a buccal-palatal crown7.38
Height of a crown5.35
Height of roots:
 posterior buccal root129
 anterior buccal root11.18.5
 palatal root8.78.4

On the other hand, if we assume that they are third molars, then they could belong to one person with a biological age of not less than 17–18 years (Boune, 1995; Scheff Jun et al., 1989; Vorobjev et al., 1936). Morphological comparison of these teeth with the remaining maxillary dentitions of Skulls 1, 3, 4, 5 and 7 allow us to exclude the possibility of their match by differences in color, shape, dimensions and development. A close comparison between the loose teeth, assumed to be third maxillary molars, with the maxillary teeth of the Skull 6 demonstrate a similarity in their relative proportions, however, as well as a close similarity of structure, notably in the area of anterior foveae of the paracones, and also with the first and second molars in the form of feathers. From the superimposition analysis, it is most likely that Skull 6 representsthat of Anastasia. Therefore, it is probable that the two loose teeth are the third maxillary molars of Anastasia.

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED

There is broad agreement among the osteological studies by Popov (1994, 1996), Zviagin and Zinin (1987), Maples and Browning (1994), and those reported here. We all agree on the identifications of the Tsar, his wife, his retainers and his daughters Olga and Tatiana. The only significant difference is whether the missing daughter is Maria or Anastasia. Genetic studies confirm the osteological identifications as far as they are able to determine. These studies confirm the gender assessments of the skeletons, identify the Tsar and his wife specifically, and identify his three daughters generally. Unfortunately, the genetic studies are unable to specifically identify the Tsar's daughters individually.

The present study introduced objective methods of comparison between the skulls and antemortem photographic portraits. These methods successfully identified those skeletons whose identities where already determined by independent means (i.e., the Tsar, his wife and her maid Demidova, and Dr. Botkin). In addition, the results were consistent with the identifications of those whose prior matching were not controversial, although not decidedly proved (i.e., Olga, Tatiana, Trupp, and Karitonov). The consistency of the results gives added confidence to the legitimacy of the methods. Therefore, the identification of Skull 6 and loose teeth as those of Anastasia has the same strength as the other identifications, and should not be dismissed.

The tragic execution of Russia's last Tsar and his family has served to stimulate much research in forensic anthropology, research that has proven useful for forensic investigators of today and tomorrow. The objective methods of examination and data analysis employed in the present study contribute to the body of work before it and have helped to answer some lingering questions regarding this chapter in modern Russian history. With the proper evidence available for examination, these trustworthy analytical techniques could be utilized in future cases for which identification of collections of multiple skeletonized human remains proves difficult by other means.

The objective methods employed in the present study contribute to the body of work before it and have helped to answer some lingering questions regarding this chapter in modern Russian history.

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. UNCOVERING THE MYSTERY
  4. EARLY IDENTIFICATION EFFORTS
  5. MORPHOLOGICAL COMPARISONS OF THE SKULLS
  6. PHOTOGRAPHIC SUPERIMPOSITION OF THE SKULLS
  7. STATISTICAL ANALYSIS
  8. STOMATOLOGICAL INVESTIGATIONS
  9. IDENTIFICATION OF LOOSE TEETH
  10. CONCLUSIONS
  11. LITERATURE CITED
  • Abramov S. 1996. O planirovanii sydebno—meditsinskikh osteologicheskikh issledovaniy v ekspertize identifikatsii lichnosti. Foren Med Expert Examin 32: 1721.
  • Abramov S, Boldyrev W, Ziakhovez A. 1994a. Optimizatsiya metoda nalozheniya pri identifikatsii lichnosti po cherepu b prishizuennym fotosnimkam. Materials of the VII All-Russian Plenary Session of Forensic Medicine. Moscow, Astrachan city: Expertiza. p 9699.
  • Abramov S, Danilevich A, Baskakov V. 1994b. O vozmozhnom reshenii voprosa o rodstve putem sravneniy konturnykh isobrazheniy cherepov. The laboratory methods of investigations in forensic medicine and the tasks of forensic medical science and practice directed to improve these methods. Materials of the VII All-Russian Plenary Session of Forensic Medicine. Moscow, Astrachan City: Expertiza. p 9296.
  • Abramov S, Korneeva N, Moiseeva N. 1995. O dostovernosti razmetki konstantnykh tochek na portretnom izobrazhenii pri identifikatsii lichnosti. 1995. The First International Conference of forensic medical workers. Proceedings. Astrachan City: Expertiza. p 2425.
  • Anonymous. 2000. Anastasia. In: Britannica.com. New York: Encyclopedia Britannica. http://www.britannica.com/bcom/eb/article/printable/6/0,5722,7456,00.html
  • Austin-Smith D, Maples WR. 1994. Reliability of skull/photograph superimposition in individual identification. J Foren Sci 39: 446455.
  • Balyeva T, Lebedinskaya G. 1991. Anthropological reconstruction. Moscow: Nauka. p 576.
  • Cherniavskaya Z. 1983. Age diagnostics by tooth wearing away in forensic medical examination. Abstracts of Ph degree thesis. Moscow. 29 p.
  • Gerasimov M. 1955. Face reconstruction by a skull. Moscow: Nauka. 585 p.
  • Gill P, Ivanov PL, Kimpton C, et al. 1994. Identification of the remains of the Romanov family by DNA analysis. Nat Genet 6: 130135.
  • Ivanov PL, Wadhams MJ, Roby RK, Holland MM, Weedin VW, Parsons TJ. 1996. Mitochondrial DNA sequence heteroplasmy in the Grand Duke of Russia Georgij Romanov establishes the authenticity of the remains of Tsar Nicholas II. Nat Genet 12: 417420.
  • Maples WR, Browning M. 1994. Dead men do tell tales: the strange and fascinating cases of a forensic anthropologist. New York: Doubleday.
  • Massie RK. 1995. The Romanovs: the final chapter. NY: Random House.
  • Popov V. 1994. Identification of the remains of the Romanovs—the Tsar family (forensic stomatological and forensic ballistic investigations). St. Petersburg: Avalanche. 61 p.
  • Popov V. 1996. Your Majesty, where are you? St. Petersburg: Peter's Academy of Science. 303 p.
  • Radzinsky E. 1992. The last Tsar: the life and death of Nicholas II. Trans. Marian Schwartz. New York: Doubleday.
  • Snetkov V, Vinichenko I, Zhitnikov V, et al. 1984. Crime detection description of human lineament. Moscow: Allunion Crime Scientific Centre. 126 p.
  • Trotter M, Gleser GC. 1952. Estimation of stature from the long bones of American whites and Negroes. Am J Phys Anthropol 10: 463514.
  • Zhivotosvsky LA. 1999. Recognition of the remains of Tsar Nicholas II and his family: a case of premature identification? Ann Hum Biol 26: 569577.
  • Zinin A, Kirsanova L. 1991. Crime detection photo portrait expert examination. Moscow: Allunion Crime Scientific Centre. 187 p.
  • Zviagin V, Zinin A. 1987. Printsipy vydeleniya priznakov vneshnosti cheloveka v tselyakh mediko—kriminalisticheskoy identifikatsii lichnosti. Foren Med Exp Examin 30: 2528.