Hominid cranial bone structure: A histological study of Omo 1 specimens from Ethiopia using different microscopic techniques



The microstructure of a hominid cranial vault has not previously been studied to determine its tissue histology, and differences in comparison with that of modern humans. We selected the parietals of Omo-Kibish 1, regarded as one of the oldest (about 130,000 years old) anatomically modern humans, and Omo 1 (Howell), which is a very recent human (about 2,000 years old)—both from the same area of Ethiopia. A combination of macrophotography, polarizing microscopy in the incident and transmission illumination mode, and confocal laser scanning microscopy (CLSM) was employed to examine thin sections, as well as polished and unpolished block faces of unembedded bone fragments, to minimize specimen destruction as much as possible. The methods enabled remarkably detailed information on bone microstructure and remodeling to be gleaned from tiny fragments of bone. The best method for examining fossilized human bones was shown to be that of incident light microscopy, which was the least destructive while producing the most amount of information. Unless the above methods are used, bone-filling minerals, such as calcite, can cause erroneous estimations of bone thickness, as observations with the naked eye or even a magnifying glass cannot determine the limit between the cortex and the diploe. This is particularly important for sciences such as paleoanthropology, in which, for instance, a thick cranial bone of Homo erectus may be confused with a pathological one of H. sapiens and vice versa. Cross sections of parietal bones revealed differences between Omo-Kibish 1 and Omo 1 (Howell) in diploic histology and in the relative thickness between the cortex and diploe, with the former specimen having an H. erectus ratio despite its H. sapiens gross anatomy. Omo-Kibish 1 may still retain some affinities with H. erectus despite its being classified as H. sapiens. Newly described histological structures, such as the reverse type II osteons, the multicanalled osteons, and the osteocytomata are presented here. A modern human skeletal anatomy does not necessarily imply a modern human cranial bone histology. The outer circumferential lamellae of cranial bones are in essence growth lines. Cranial histology of hominids may provide useful information concerning their taxonomy and life history, including such factors as growth rate, developmental stress, and diet. Anat Rec 267:52–59, 2002. © 2002 Wiley-Liss, Inc.

There is a paucity of knowledge concerning hominid bone histology, and what little there is is based mainly on long-bone thin sections (Thompson and Trinkaus, 1981; Walker et al., 1982; Trinkaus and Thompson, 1987; Bartsiokas and Day, 1993). There is hardly anything in the literature regarding flat bones (especially cranial bones) of hominids. This may be due to their anatomical importance, which makes their curators hesitant to provide specimens for histological studies. This work is the first attempt to study the histology of a cross section of a fossil human cranial vault microscopically. Osteocyte lacunae have been identified from Neanderthal crania (Sergi et al., 1972), but these were minute pieces from the palate that were too small to allow comparisons between compact and trabecular bone to be made. Comparisons of thickness in compact and diploic bone have been made in H. erectus skulls (Weidenreich, 1943), archaic H. sapiens (Stringer et al., 1985), and late Pleistocene crania (Webb, 1989), but they do not contain information on lamellar bone, i.e., detailed histological structure. The aim of the present study was to investigate as nondestructively as possible whether there are any histological differences between the cranial vaults of the oldest anatomically modern humans and those of living populations.


Toward that end we examined a few small pieces from the parietals of Omo-Kibish 1 and Omo 1 (Howell) skeletons. Omo-Kibish 1 is a fossil from a famous skeleton that is considered to be one of the oldest known anatomically modern H. sapiens (Leakey et al., 1969; Day and Stringer, 1991). It was found in situ in the minor discomformity separating beds 5 and 6 of Member I of the Kibish Formation at Omo Basin, Ethiopia (Butzer, 1971), by Kamoya Kimeu, a member of the Kenyan team of the International Paleontological Research Expedition to the Omo Valley, in 1967. Bed 6 has been tentatively dated by uranium/thorium determination of mollusk shells to 130,000 years B.P. (Butzer et al., 1969). Omo 1 (Howell) is a subfossil from a human skeleton, possibly 2,000 years old, from an early African population; it is still undescribed (C. Howell, personal communication). It was found by Professor Clark Howell of the above-mentioned expedition on the present-day land surface of Omo Basin. It was chosen because it was found in the same area as the former specimen and therefore is ideal for comparing with Omo-Kibish 1. Both are adult skeletons (Leakey et al., 1969). The specimens studied here were provided courtesy of Professor M.H. Day.

Four methods of histological analysis were used: macrophotography, incident light microscopy, polarizing microscopy, and confocal laser scanning microscopy (CLSM).

Macrophotography was used on thin sections as described by Bartsiokas (2000). Thin sections were also examined by using a Leitz polarizing microscope. Bulk unembedded sections were viewed under a magnifying glass or a light microscope with fiber optic incident light, as described by Bartsiokas and Day (1993). A flat transverse natural section was polished with a fine emery paper to minimize destruction. After polishing, the bulk section was examined under a light microscope using incident light.

CLSM was used as described by Bartsiokas (1992a). For CLSM, there was no need for any mechanical preparation of the Omo-Kibish 1 specimen because the technique is nondestructive, providing optical sections from within the bone without the need for physical sectioning. The specimen was observed using a Bio-Rad MRC-600 CLS microscope under 488-nm blue light at a depth of 15 μm. CLSM allows the microscope to nondestructively penetrate a fossil and extract images that with a regular microscope would be completely blurred by light from regions above or below the inspected plane.


The results are presented in Figures 1–8 and Table 1.

Figure 1.

Thin cross sections of the two parietals examined here using macrophotography. Top: Omo-Kibish 1. Bottom: Omo 1 (Howell). Left: Ectocranial surface. Right: Endocranial surface. No lamellae are visible with this method. All the bone voids (including cracks, osteocytes, Haversian canals, and marrow cavities) in Omo-Kibish 1 have acquired a natural black staining from Mg impurities, as X-ray microanalysis has shown (Bartsiokas et al., unpublished results). The presence of osteocytomata is shown by the arrows on the ectocranial surface of the parietal. Area 1 is shown in Figure 4; area 2 is shown in Figure 5. Scale bar = 1 mm.

Figure 2.

Bulk cross section of the outer table of Omo-Kibish 1 observed with incident light microscopy (left: ectocranial surface). The bone voids are filled by calcite, as shown in Figure 5 by polarizing microscopy. The lamellae are clearly visible. The tables are circumferentially lamellar in nature. Scale bar = 0.5 mm.

Figure 3.

Bulk cross section of the outer table of Omo-Kibish 1 observed with incident light microscopy. Most of the osteons are ellipsoid. Scale bar = 0.5 mm.

Figure 4.

A “colony” of osteocytes or osteocytoma in a thin cross section of the outer table of Omo-Kibish 1 parietal observed with transmission light microscopy. The osteocyte lacunae and their canaliculi are clearly visible. Top: the ectocranial surface. Outside the osteocytomata the osteocytes are few and far between, and many lack canaliculi. Note that most osteocyte lacunae are flat, or disc-like. A few of them are rather circular, or spherical (bottom right). Scale bar = 0.5 mm.

Figure 5.

Thin cross section of the parietal of Omo-Kibish 1 examined here using polarizing microscopy. Left: Ectocranial surface. Right: Endocranial surface. The minerals (mainly calcite) in the trabecular spaces are clearly visible. The concentric lamellae of the secondary osteons and the circumferential lamellae are also visible. Note the diploic sheet and the lamellar sheet of the two tables. Internal remodeling is evident in each sheet. Note a triangular multicanalled osteon with three Haversian canals in the middle of the outer table. Scale bar = 1 mm.

Figure 6.

A bulk section of Omo-Kibish 1 showing two multihaloed osteons (center bottom and center left; one of them shown in half) and one reverse type II osteon (center top) by incident light microscopy. The two hypermineralized scalloped reversal lines of the reverse type II osteon are indicated by an arrow. The Haversian spaces are filled by authigenic calcite. Scale bar = 0.5 mm.

Figure 7.

Confocal micrograph showing a tangential optical section to the outer table of Omo-Kibish 1 parietal bone. Circular structures are (sub)surface pitting, i.e., porosity. A straight crack is also evident. Scale bar = 100 μm.

Figure 8.

A longitudinal optical section of a Haversian canal in Omo-Kibish 1 ectocranial side of the parietal bone obtained with CLSM. A Volkmann's canal branches obliquely off the Haversian canal (right). Some osteocyte lacunae with their canaliculi are also present. Some canaliculi are connected with the Haversian canal. Scale bar = 50 μm.

Table 1. Cranial bone thickness in mm of various bone compartments
 Omo-Kibish 1 parietalOmo 1 (Howell) parietal
 Bulk section; naked eye or magnifying glass in mmThin section; macrophotography and/or microscopy in mmBulk section; naked eye or magnifying glass in mmThin section; macrophotography and/or microscopy in mm
  1. The values of the 1st and the 3rd columns have no biological meaning and they merely demonstrate the pitfalls one should avoid in measuring the thickness of compact and trabecular bone. The differences between bulk and thin sections are independent of the total cranial thickness. They only depend on how many of the trabecular spaces are filled by calcite. For instance, if in bulk sections the spaces are completely filled by calcite, the c/d will be close to infinity.

  2. c/d, ratio of compact to diploic bone.

Outer table2.
Inner table2.
Total thickness6.


Comparison of the Specimens

A comparison of the two Omo specimens can be made using macrophotography (Fig. 1), in which cross sections of the two parietals are displayed. The thicknesses of various parts are provided in Table 1. Various similarities and differences can be observed between the two sections, as follows.

The Omo-Kibish 1 section shows that the histology has been preserved, whereas in Omo 1 (Howell) recrystallization of bone apatite has destroyed the lamellar structure of the bone and the osteocyte lacunae, preserving only the Haversian canals and the trabecular spaces. This paradox (i.e., fossil bones are preserved while recent bones are histologically destroyed) is not uncommon in our experience and may be caused by a quick collagen leaching in recent bones before fossilization (i.e., crystallinity increase) takes place (Bartsiokas, 1992b). As a result, the apatite crystallites become less coherent, which leads to eventual destruction of the bone.

The bone voids in both specimens have been filled partially or completely by calcite (Figs. 2–6), as shown by polarizing microscopy. Bone void is a collective term for structures such as the osteocyte lacunae, Haversian canals, marrow cavities, and post-mortem cracks. Calcite is a common bone-filling mineral in fossils because groundwater that impregnates bone is usually rich in ions that are deposited as calcium carbonate.

The anatomically modern human parietal, Omo1 (Howell), has a thickness of about 9 mm, and the fossil human parietal, Omo-Kibish 1, has a thickness of about 6 mm (Fig. 1, Table 1). The range of parietal thickness in most modern humans or archeological populations is 3–11 mm (Smith et al., 1985; Ross et al., 1998). Therefore, both specimens are within the range of H. sapiens as regards total parietal thickness.

Measurements taken with the naked eye or under magnification (Table 1) differed within specimens as regards the thickness of the diploe and tables. These differences are due to infilling of trabecular spaces by calcite; with the naked eye or a magnifying glass, the compact bone in the bulk section appears to be much thicker because the smaller marrow cavities have been filled by calcite—only the largest marrow cavities in a block face that have not been fully filled by calcite are discernible by the naked eye. Therefore, care should be taken when measuring relative bone thickness in fossils to perform these measurements on histological sections using only microscopy or macrophotography, so that trabecular bone does not get confused with compact bone because of infilling of bone voids by minerals. This is particularly important for sciences such as paleoanthropology, in which, for instance, a thick cranial bone may be confused with a pathological one.

The two specimens also differ in the relative thickness of compact to trabecular bone (Table 1). In Omo-Kibish 1, the thickness of the outer and inner tables combined (c) is slightly higher than that of the diploe (d) (c:d = 1:0.90), whereas the opposite is true for the Omo 1 (Howell) specimen (c:d = 1:2.78). This feature of Omo-Kibish 1 is also found in H. erectus, wherein all three parts of the cranial vault take equal part in the thickening, with the two tables combined slightly thicker than the diploe (Weidenreich, 1943; Webb, 1989). It should be noted that since the parietal of Omo-Kibish 1 had suffered a minor post-mortem exfoliation on its ectocranial and endocranial surfaces, the thickness of the compact bone presented here is somewhat underestimated. The Omo 1 (Howell) feature is similar to that of Late Pleistocene crania (Stringer et al., 1985; Webb, 1989) and living humans (Reynolds, 1962). Also, in both specimens the outer table is slightly thicker than the inner table, a feature that is also found in modern humans (Webb, 1989; Boyde et al., 1990; Schultz, 1993). This is due to the presence of more lamellae in the outer table. However, as we shall see below, this is not a characteristic restricted to humans.

In Omo 1 (Howell) there are normal cancellous trabecullae, i.e., long, thin, and more or less equal in width and size length (Fig. 1). Their edges are outwardly quite smooth. The marrow cavities lying between the trabeculae are more or less equal in size, and similar in shape. The ratio of trabecular width to the width of the marrow cavities is about 0.15. This ratio is termed here the “spongiosa index.”

In Omo-Kibish 1, the cancellous trabecullae are wide, shapeless, and irregular, with rougher edges compared to the other specimen. Many of the marrow cavities have irregular and peculiar shapes. The spongiosa index is about 0.9, although it is more difficult to measure in this specimen than in Omo 1 (Howell). The spongiosa index is more objectively determined using image-processing techniques to measure the ratio of trabecular surface area to the area of the marrow cavities. In any case, there is a substantial difference between the two specimens in the microstructure and morphology of the diploe.

It can be deduced that the relative thickness of compact to trabecular bone of the Omo-Kibish 1 parietal is within the range of that of H. erectus, even though the specimen is classified as H. sapiens (Day and Stringer, 1991; Day et al., 1991). Indeed, the diploe of Omo-Kibish 1 is not only different in relative thickness; its structure is altogether different from that of modern humans.

Further Histological Observations in Omo-Kibish 1

In Omo-Kibish 1 there is a zone in the inner and outer tables next to the diploe that contains numerous Haversian canals not previously described in histological sections of adult cranial bones (Figs. 1 and 5). It has only been described in a 4½-year-old boy (Boyde et al., 1990). Thus, each of the two tables can be divided into two sheets of bone: an outer one in which the outer circumferential lamellae are numerous and the osteons are scarce, and an inner one in which the secondary osteons are numerous and the circumferential lamellae are less common. The inner sheet is termed here the “diploic sheet of the table,” and the outer one is termed the “lamellar sheet of the table.” This histology apparently results in a higher vascularization in the sheets of bone adjacent to the diploe, where secondary osteons are abundant. Remodeling has been most active in the diploic sheet of the table, where bone is resorbed, and less active in the lamellar sheet of the table, where bone is deposited (Boyde et al., 1990; Enlow and Hans, 1996). Most of the layers of lamellae in the diploe are interrupted. Thus, in Omo-Kibish 1, the parietal received a bone deposition on the surfaces of both the ectocranial and endocranial sides. The diploic sheets of the inner and outer tables were of a resorptive nature due to intense remodeling. This increased the overall thickness and expanded the diploe.

It is suggested here that the lamellar sheet of the table may be indicative of the growth rate since the ectocranial surface of the calvaria is a depositional one, i.e., the lamellae (which are in essence growth lines) are deposited centrifugally and the tissue does not suffer substantial disruption by the secondary osteon remodeling (Enlow and Hans, 1996). The same applies to the inner table, but the outer table is preferable for growth studies because it is usually thicker. In this respect, calvaria bones resemble teeth, which also have growth lines in both the dentine and the enamel. In the outer table of Omo-Kibish 1, there are distinct bundles of lamellae with similar spacing, separated by wider spaces (Fig. 5). Whether this corresponds to a seasonal banding or to periods of stress is not yet clear. In rabbits, the increase in thickness of the cranial vault is the result of bone deposition on the ectocranial and endocranial surfaces of parietal bones (Hong et al., 1968). However, the deposition of bone in the outer table is more extensive than that in the inner table (Hong et al., 1968). The same may apply to humans. Therefore, the relative thickness of the two tables appears to be a mammalian characteristic of growth that is not limited to modern humans. If we use the average rate of bone deposition of lamellar bone in ribs of modern humans (0.9 μm/day (Yen and Shaw, 1977)) as a model, the 2 mm thickness of the outer table of Omo-Kibish 1 might correspond to a period of about 6 years. Therefore, this histology may represent the last (6) years of this individual's life.

The majority of the osteons have an ellipsoid shape, more so in the inner table. Thus, they look as if they are bilaterally compressed (Figs. 3–5). Perhaps this osteonal morphology is an adaptive means of strengthening the skull against head injuries.

Using incident light microscopy, some multihaloed osteons can be seen with as many as 13 hypermineralized lamellae, indicating arrested growth (Fig. 6). The term “halo” refers to the appearance of a lamella. Hypermineralization refers to the nature of the halo. Multihaloed osteons were first described in fossil hominids by Bartsiokas and Day (1993), who reported a maximum of 10 haloes in their specimens.

Type II osteons were also observed in the present study, with the characteristic scalloping of reversal lines indicative of arrested resorption and redeposition of new layers of lamellar bone (Fig. 6). They have previously been described and defined in archeological populations (Richman et al., 1979) and fossil hominids (Bartsiokas and Day, 1993) by the presence of a non-hypermineralized scalloped reversal line that divides the osteon into two parts: an “outer osteon” and a less mineralized “inner osteon” (Bartsiokas and Day, 1993). However, in the specimen described here (Fig. 6) there are two hypermineralized scalloped reversal lines, and the inner osteon is more mineralized than the outer osteon. This is a paradox, because inner osteons are younger than outer osteons. This may be explained by a very slow formation of the inner osteon. Because of these differences we term the osteon described here a “reverse type II” osteon. Type II osteons are evidence of a meat diet (Bartsiokas and Day, 1993; Richman et al., 1979). Thus, the reverse type II osteon described here may represent evidence of a meat-eating period (or two) followed by a period of arrested growth, possibly caused by disease or dietary stress in the Omo-Kibish 1 individual.

There is also an osteon in the middle of the outer table shown in Figure 5. An unusual feature of this osteon is that it contains three Haversian canals, which give the osteon a triangular shape. These type of osteons are termed here “multicanalled” osteons; they are defined by the presence of at least two canals in the same osteon, the lamellae of which encircle all the canals and are consequently polycentric. Multicanalled osteons have not previously been described in vertebrates. These osteons may represent the point where a small diploic vein branches into two or more smaller ones.

Another intriguing feature is the presence of osteocyte “colonies” spaced every 2–3 mm in the outer table of Omo-Kibish 1 (Figs. 1 and 4). They are characterized by the close packing of numerous osteocytes with numerous canaliculi. A close view of such a colony is shown in Figure 4. To date, this feature has not been observed in modern humans (but this does not mean it does not exist in them). These sorts of osteocyte colonies are termed here “osteocytomata” (singular osteocytoma). The areas outside osteocytomata have very few osteocytes, with very few and small canaliculi. The osteocytes have a flat shape similar to that of long bones, and a few of them are rather spherical (Fig. 4). The function of osteocytomata may be to facilitate the communication between the diploe and ectocranial surface during development or in cases of head injuries. In other words, they may play a sort of role as osteocyte anastomosis. It should be noted that what we call here osteocytes are in fact osteocyte lacunae filled with impure authigenic calcite.

Confocal laser scanning microscopy (CLSM) was also used to examine Omo-Kibish 1. Osteons and osteocyte lacunae with their canaliculi were previously observed in Omo-Kibish 1 with this method (Bartsiokas, 1992a), which was the first application of CLSM to paleoanthropology. In the present study, an optical section tangential to the outer table was obtained (Fig. 7) by placing a piece of parietal flat on the microscope table. The dark round structures thus obtained were initially taken for osteocytes. However, their size and distribution subsequently identified them as structures caused by (sub)surface pitting (i.e., porosity) of the ectocranial surface. Figure 8 shows a longitudinal section of a Haversian canal in a cross section of the Omo-Kibish 1 parietal. A Volkmann's canal obliquely branches off the Haversian canal. Figure 8 could not have been obtained without the CLSM. The fluorescence observed inside the canal is due to ethidium bromide staining. Authigenic minerals have filled the Haversian canal.

Evaluation of the Methods

Of the methods used in this study, transmission light microscopy provided the most detailed images of bone voids, such as osteons and osteocyte lacunae. However, these images lack color, and since the photons are emitted from a thin layer of bone, they resemble images obtained by a secondary electron detector of a scanning electron microscope (SEM). As a result, no images of lamellae and their degree of mineralization can be produced with this method. Polarizing microscopy can provide stunning full-color images that render the authigenic minerals identifiable and the lamellar structures clearly visible. However, these methods require sections to be cut. In contrast to the destructiveness of the above methods, incident light microscopy of bulk sections can provide impressive images that are full of natural color and show the degree of mineralization in various bone components. This can be done with minimal destruction (from a slight polishing of the surfaces examined), and since the emitted photons derive from deep inside the specimen, the images produced are comparable to those produced by the backscatter electron detector of an SEM (Boyde et al., 1990). Apart from its complete nondestructiveness, CLSM is particularly effective for fossilized bones because they are more transparent to light than are modern bones. This is apparently due to the lack of collagen and the increased crystallinity in the former, which otherwise would blur the bone images despite the bone staining caused by diagenetic effects (our observations on modern bone are not reported here). Unfortunately, CLSM does not produce images at low magnification and does not reveal clear lamellar structure. Macrophotography, with its portability and use of low magnification, can simply and quickly produce overall pictures of bone sections that are difficult to obtain using microscopy.

After comparing all the above methods, we consider that the best method for examining fossilized human bone is incident light microscopy of bulk sections, since it provides maximum information with minimum specimen destruction. It is also simple and inexpensive. To obtain an overall picture of a bone section, macrophotography is the ideal choice.


The first histological observations of a hominid cranial vault are presented. Of the methods used in this study, incident light microscopy produced the best results by being essentially nondestructive, informative, simple, and inexpensive. Without the use of microscopy or macrophotography, the borders between compact and trabecular bone are difficult to discern due to void-filling minerals that have been deposited post-mortem. Thus, estimations of compact and diploic thicknesses should be based on measurements taken using microscopy, or at least macrophotography.

Omo-Kibish 1 displays an H. erectus cranial histology as regards relative thickness of cortical to trabecular bone that is consistent with the diploic microstructure of the bone. This is unexpected since the skeleton of Omo-Kibish 1 has been identified as an anatomically modern H. sapiens. It shows that Omo-Kibish 1 may still retain some affinities with H. erectus despite being classified as H. sapiens. Thus, similar gross anatomy does not necessarily imply similar cranial bone histology. The outer circumferential lamellae of the cranial bone outer tables are much thicker than those of long bones. They can be regarded as growth lines because they are deposited centrifugally for a long time, and as such they may provide useful information concerning the development of hominids.

Newly described histological structures in humans, including the reverse type II osteons, multicanalled osteons, and osteocytomata, are presented here. The reverse type II osteons may indicate periods of meat-eating followed by arrested growth. Multihaloed osteons are indicative of developmental stress (in the Omo-Kibish 1 individual). Osteocytomata may be useful for communication between the spongiosa and the outer table, or reactions in cases of cranial trauma. Multicanalled osteons show close vascularization. Overall, cranial bone histology may elucidate the taxonomy and the life histories of hominids and past populations.


I thank Professor M.H. Day for providing the Omo-Kibish 1 specimen and, along with Professor C. Howell, the Omo 1 (Howell) specimen; Dr. Vivian Howard for allowing me to use the CLSM at the Department of Anatomy and Cell Biology, University of Liverpool; and Dr. A. Trichas for the macrophotography of the specimens.