THE ANATOMICAL RECORD 252:355–368 (1998) Three-dimensional Direction and Interrelationship of Prisms in Cuspal and Cervical Enamel of Dog Tooth YOSHINORI HANAIZUMI, YOSHIRO KAWANO, HAYATO OHSHIMA, MASAAKI HOSHINO, KIICHI TAKEUCHI, AND TAKEYASU MAEDA* Department of Oral Anatomy II, Niigata University School of Dentistry, Niigata, Japan ABSTRACT The three-dimensional architecture of enamel prisms was examined in cuspal enamel and compared with that in cervical enamel by light and electron microscopy as well as computer-assisted reconstruction using the developing enamel of several dog teeth. Dog tooth enamel consists of two groups of alternately arranged enamel prisms oriented in opposite sideward directions basically forming thick horizontal rings, partly branching off from the stem. Along a 8–10 enamel prism-wide group, the enamel prisms emerge in parallel tilting uniformly to the same sideward direction. In cervices, groups of enamel prisms are arranged nearly in parallel displaying a regular arrangement of prisms. Approaching the cusp of tooth, the groups of enamel prisms fuse to a concentric cusp-centered arrangement and the prisms exhibit no periodic arrangement as shown in the cervical enamel. It is suggested that the three-dimensional structure of enamel becomes complicated close to the cusp, contributing to the chewing stress of tooth. Anat. Rec. 252:355–368, 1998. r 1998 Wiley-Liss, Inc. Key words: Hunter-Schreger band; enamel prism; dog tooth; three-dimensional reconstruction It is generally accepted that enamel consists of numerous enamel prisms stretching right or left in their passage from the enamel-dentin junction to the surface. The HunterSchreger bands that correspond to the zones of longitudinally and transversely cut prisms are observed on longitudinal cut surfaces of tooth enamel in the great majority of mammals. The three-dimensional course and arrangement of enamel prisms in relation to the formation of the Hunter-Schreger bands have been examined in various species using a variety of methods (Erausquin, 1949; Kawai, 1951, 1955; Applebaum, 1960; Boyde, 1964, 1969; Osborn, 1965, 1968a,b; Hirota, 1982; Koenigswald et al., 1987). However, no data sufficiently explain the exact three-dimensional orientation of enamel prisms. Accordingly, the spatio-temporal relationship between the arrangement of enamel prisms and the Hunter-Schreger bands remains unclear. Previous study of cervical enamel of the dog tooth in three dimensions has shown that groups of enamel prisms having similar courses form regular arrangements of laminated structures in the next layer, but in the opposite direction, appearing as the parazone or diazone of the Hunter-Schreger bands on the longitudinal plane (Hanair 1998 WILEY-LISS, INC. zumi et al., 1996). This study, however, failed to demonstrate the characteristic features of cuspal enamel where the Hunter-Schreger bands showed complex arrangements. As one ameloblast forms the body of one enamel prism throughout amelogenesis, study of the three-dimensional course of enamel prisms has provided accurate information about the directions in which ameloblasts have moved (Osborn, 1970, 1973; Nishikawa et al., 1988, 1990; Nishikawa, 1992; Hanaizumi et al., 1994). On the basis of the observation of enamel prisms, some investigators have speculated that ameloblasts are controlled by secretory (Osborn, 1970, 1973) or contractile forces of terminal web filaments (Nishikawa et al., 1988, 1990; Nishikawa, 1992), but conclusive evidence has not been reported. Grant sponsor: Japanese Ministry of Education, Science, Sports, and Culture; Grant number: 09771496. *Correspondence to: Dr. Takeyasu Maeda, Department of Oral Anatomy II, Niigata University School of Dentistry, Gakkochodori, Niigata, 951-8514, Japan. Received 1 December 1997; Accepted 11 May 1998 356 HANAIZUMI ET AL. Figs. 1,2. Light micrograph of longitudinal ground section of a dog canine tooth stained with hematoxylin in cuspal (Fig. 1) and cervical (Fig. 2) enamel. Dark and light zones correspond to the parazone (p) and diazone (d) of the Hunter-Schreger bands, consisting of longitudinally and transversely cut prisms. Fig. 1 ⫻80, Fig. 2 ⫻150. Scale bar ⫽ 100 µm. D, dentin; E, enamel. The present study aims to demonstrate the threedimensional architecture of enamel in cusps as compared with that in cervices and simultaneously proposes a possible force controlling ameloblast movement. The study examined an erupted dog tooth and the developing enamel of several dog tooth germs in three-dimensions by light microscopy as well as by scanning and transmission electron microscopy. A computer-assisted reconstruction of the Hunter-Schreger bands was also attempted. in 5% EDTA (pH 7.4) for 3 weeks. Tooth germs were postfixed in 1% osmium tetroxide for 3 hours, dehydrated in ethanol, and embedded in Spurr’s resin. The molar tooth germ was sectioned in one direction until the longitudinal plane near the cusp was obtained. Following the longitudinal sectioning, the resin block was remounted on the ultramicrotome at a right angle to the sectioning plane. One-µm thick horizontal serial sections perpendicular to the tooth axis were made from the cusp for a distance of 340 µm (see Fig. 6). The longitudinal and serial tangential semithin sections of canine tooth previously made (Hanaizumi et al., 1996) were used for this study. All the sections were stained with methylene blue and azure II, and photographed with a light microscope (Nikon Microphot FXA, Tokyo). MATERIALS AND METHODS Light Microscopy Ground sections of enamel were cut and polished from an erupted canine of an adult dog that had been stored in 5% neutral formalin solution. The longitudinal 80–90 µm thick ground sections were then etched for 15 seconds with 0.1 N hydrochloric acid, briefly washed in tap water, and stained with hematoxylin. A 1-month-old pup was anesthetized with Nembutal (0.5 ml/kg body weight intravenouly) and perfused with a mixture of 0.1 M phosphate-buffered solution of 2% paraformaldehyde and 1.25% glutaraldehyde at pH 7.4. The jaws were removed and immersed in the same fixative overnight. Permanent tooth germs were dissected out and molar tooth germs were demineralized Scanning Electron Microscopy Some of the molar tooth germs, dissected as described above, were immersed in buffered 2% OsO4 solution for several hours at 4°C. To remove the enamel organ and expose the enamel surface without causing mechanical damage, the remaining molar tooth germs were processed for cell-maceration (Ohtani et al., 1988; Ushiki et al., 1990) to remove cellular elements and their basal laminae, DIRECTION/INTERRELATIONSHIP OF PRISMS 357 Fig. 3. Tangential semithin section 50 µm from the enamel-dentin junction. The tangential semithin section shows many belt-like zones roughly perpendicular to the meridian. A single belt-like zone consists of a group of enamel prisms oriented in the same direction. Groups of prisms oriented in the opposite direction are arranged in alternate bands. The belt-like zones of the prisms oriented to the righthand side are labeled p1–8 and those of the inversely oriented prisms, d1–8. ⫻120. Scale bar ⫽ 100 µm. Fig. 4. Micrograph of longitudinal semithin section parallel to the tooth axis in a cervical region. Parazones (p) and diazones (d) of the Hunter-Schreger bands were arranged at equal intervals, partly showing confluence and divergence (ⴱ). Ab, ameloblast; D, dentin; E, enamel. ⫻140. Scale bar ⫽ 100 µm. leaving the extracellular matrix. The fixed specimens were immersed in a 5% aqueous solution of NaOH for 2–3 days at room temperature, rinsed with distilled water for 1–2 days, with these processes repeated several times. The NaOH-treated specimens were then dehydrated, transferred to isoamyl acetate, and critical-point dried with liquid CO2. The dried specimens were then affixed on aluminum stubs with double-sided tape and coated with platinum in an ion coater (Hitachi E-1032, Hitachi, Tokyo). The specimens were then observed under the scanning electron microscope (Hitachi S-4000) at an accelerating voltage of 10kV. graphic printing paper and the boundaries between the groups of enamel prisms traced on tracing paper. Using the outline of each section as a reference, the images of the enamel prisms and the boundaries were read serially into a personal computer (NEC PC-9801 VM) using software for three-dimensional graphic analysis (Cosmozone 2SA, Nikon, Japan). The reconstructed images were photographed from the computer display. Three-dimensional Reconstruction The light microscopic images of horizontally or tangentially cut serial semithin sections were enlarged on photo- RESULTS Longitudinal Section In etched and hematoxylin-stained longitudinal ground sections, parazone and diazone of the Hunter-Schreger bands, consisting, respectively, of longitudinally and transversely cut prisms were clearly visible (Figs. 1,2). The distribution of parazones and diazones appeared curved 358 HANAIZUMI ET AL. Fig. 5. Tangential sections 55 µm (a) and 85 µm (b) from the enameldentin junction. (a) A straight row of enamel prisms perpendicular to the boundaries of groups of enamel prisms (E1–5) has been selected on the section and numbered 1 to 27 from the top. (b) The section 30 µm from section (a). The row of enamel prisms aligned straight in Figure 6a shows a wavy sine curve. ⫻10. Scale bar ⫽ 50 µm. with the concavities in the cuspal region (Fig. 1), each zone alternating in parallel in the cervical enamel (Fig. 2). A longitudinal semithin section parallel to the tooth axis revealed parazones and diazones distributing at equal intervals in the cervical region (Fig. 4), whereas the bands were arranged irregularly in the cuspal region (see Fig. 7a). Over the tip of cusp, enamel prisms showed a complicated course of enamel prisms. Some prisms extended from the enamel-dentin junction and changed the angle of tilt abruptly near the boundaries of the bands (see Fig. 7b). secretory face of Tomes’ process is situated at right angles to the longitudinal axis of the prism (Boyde, 1967, 1969; Wakita et al., 1981; Hanaizumi et al., 1994). With this in mind, the direction of the enamel prisms here could be estimated from the inclination of the top of the prism sheaths. A single belt-like zone consisted of a group of 8–10 enamel prisms oriented in the same direction. Groups of prisms oriented in the opposite direction were arranged in alternate bands. The belt-like zones of the prisms oriented to the right are labeled p1–8, and those of the inversely oriented prisms, d1- 8 (Fig. 3). A straight row of enamel prisms aligned perpendicular to the boundaries between adjacent groups of prisms was selected on a section 55 µm from the enamel-dentin junction and the prisms numbered from the top (Fig. 5a, 6). The marked enamel prisms were traced on the next serial section so they overlapped each other. On a section 30 µm from the section where the row of enamel prisms was aligned straight (Fig. 5a), Tangential Section The tangential semithin section cut at 50 µm from the enamel junction showed many belt-like zones roughly perpendicular to the meridian. It has been known that enamel prism sheath represents the fossilized path traced by the Tomes’ processes during enamel secretion (Boyde, 1967) and that DIRECTION/INTERRELATIONSHIP OF PRISMS 359 Fig. 6. Diagram showing the spatio-temporal relationship between the longitudinal section (plane L) and horizontal section (plane H). The middle block of enamel is reconstructed. the row of enamel prisms appeared to be arranged in a wavy sine curve (Fig. 5b). Horizontal Section Horizontal semithin section 280 µm from the cusp of the tooth showed a concentric pattern of belt-like zones (Fig. 8b). Each zone was composed of a group of 8–10 enamel prisms oriented in the same direction similar to the tangential section shown in Figure 3 (Fig. 8b). The groups of prisms oriented clockwise were labeled p1–3 and those inversely oriented, d1–6 (Fig. 8a). A straight row of enamel prisms perpendicular to the lower edge of the section crossing the tooth axis has been selected on the section 300 µm from the cusp and are numbered 1–25 from the top (Fig. 9a). On a section 10 µm from the first, the number 5 prism showed the curved portion changing the angle of tilt (Fig. 9b). At 35 µm from the first section, the row of enamel prisms formed a wavy curve but did not display a periodic pattern as shown in the cervical region (Fig. 9c). Scanning Electron Microscopy The exposed surface of developing enamel showed numerous pits, each consisting of a combination of a flat face and its enclosing face (Fig. 10b). A group of pits inclined uniformly toward the same sideward direction was arranged in a concentric pattern around the cusp (Fig. 10a). The neighboring groups of pits inclined toward the opposite direction. Narrow bands consisting of a few small round pits with flat faces directed perpendicular to the developing enamel surface were found between the neighboring band-like groups of pits (Fig. 10c). Three-dimensional Reconstruction of Groups of Enamel Prisms Groups of enamel prisms in the cervical enamel, oriented uniformly in the same direction (as shown in Fig. 3) were reconstructed from serial tangential sections over an interval of 80 µm, longer than the 60 µm demonstrated by 360 HANAIZUMI ET AL. Fig. 7. Micrographs of longitudinal semithin section cut parallel to the tooth axis in cuspal enamel. (a) Parazone (p) and diazone (d) of the Hunter-Schreger bands are distributed irregularly. ⫻100. Scale bar ⫽ 100 µm. (b) Enlargement of (a). Some enamel prisms change angle of tilt near the boundaries of the bands (arrowheads). ⫻320. Scale bar ⫽ 50 µm. Ab, ameloblasts; D, dentin; E, enamel; p, parazone; d, iazone. DIRECTION/INTERRELATIONSHIP OF PRISMS Fig. 8. Micrographs of horizontal semithin section 280 µm from the cusp of a tooth. (a) The semithin sections show many belt-like zones approximately arranged around the tooth axis. The groups of the prisms oriented to the clockwise direction are labeled to p1–3 (yellow area) and those of the inversely oriented prisms, d1–6. The lower edge of the horizontal section is adjusted to the longitudinal section shown in Figure 7. 361 Ab, ameloblast. ⫻170. Scale bar ⫽ 100 µm. (b) High magnification of (a). The belt-like zones consist of groups of prisms oriented in the same direction. Right pointing arrowheads (䉴) show groups of prisms oriented to the right. Left pointing arrowheads (䉳) show groups of prisms oriented to the left. There are groups of round, smaller prisms between neighboring zones (ⴱ). ⫻340. Scale bar ⫽ 25 µm. 362 HANAIZUMI ET AL. cervices. The different appearance between cuspal and cervical enamel was that cuspal enamel showed a concentric belt-like arrangement tilting apically (Fig. 12). Three-dimensional Reconstruction of Enamel Prisms A row of enamel prisms and the boundaries of the groups of enamel prisms selected in the cervical (Fig. 5) or cuspal (Fig. 9) enamel were reconstructed at 50 µm intervals and viewed obliquely from the surface (Figs. 13,14). Reconstructions of the boundaries of the groups of enamel prisms and a row of enamel prisms demonstrated the three-dimensional interrelationship between the boundaries and the prisms. Each group consisted of enamel prisms tilting to the same sideward direction, whereas those in the neighboring zones were oriented in the opposite direction. In the cervical enamel, the cut ends of the enamel prisms formed a wavy sine curve approaching the enamel surface (Fig. 13a). In a single group, the horizontal tilt angles of prisms toward the enamel-dentin junction tend to be greatest at the center and smallest near the boundaries (Fig. 13b). In the cuspal enamel, groups of enamel prisms did not display a periodic tilting pattern like in the cervical region because of the irregular distribution of the prism groups (Fig. 14a). Some prisms twisted to run in the opposite sideward direction after crossing the boundaries between the neighboring groups of enamel prisms (Fig. 14b). DISCUSSION Three-dimensional Structure of Enamel Fig. 9. Horizontal semithin sections 300 µm (a), 290 µm (b) and 265 µm (c) from the cusp of the tooth. (a) A straight row of enamel prisms perpendicular to the lower edge of the section crossing the tooth axis and the groups of enamel prisms (E1–3) has been selected and is numbered 1–25 from the top. (b) A section 10 µm from section (a). Note the number 5 prism changing its angle of tilt. (c) A section 35 µm from section (a). The row of enamel prisms form a wavy curve but does not display the periodic pattern shown in the cervical region. ⫻310. Scale bar ⫽ 50 µm. our previous reports (Hanaizumi et al., 1996). The reconstructed images viewed obliquely from the tooth surface showed both the tangential and longitudinal planes. In cervical enamel, groups of unidirectionally oriented enamel prisms showed a nearly parallel belt-like arrangement, with occasional confluences and divergences (Fig. 11a). Each group of enamel prisms corresponded to the parazone and diazone of the Hunter-Schreger bands on the longitudinal plane (Fig. 11b). Groups of enamel prisms in the cuspal enamel, oriented uniformly in a clockwise direction (Fig. 12a) or in a counterclockwise direction (Fig. 12b) in Figure 8b, were reconstructed from serial horizontal sections at 60 µm (Fig. 12 a-c) or 80 µm intervals (Fig. 12d). The reconstructed images viewed obliquely from the tooth surface showed both horizontal and longitudinal planes. The basic arrangement pattern of groups of enamel prisms in cuspal enamel was the same as that of groups of prisms from This study demonstrated that dog tooth enamel consisted of two groups of enamel prisms alternately oriented in opposite sideward directions basically forming thick horizontal rings, partly branching off from the stem ring. Along a certain ring composed of 8–10 enamel prisms wide, the enamel prisms emerged in parallel tilting uniformly to the same sideward direction. Groups of enamel prisms showed a nearly parallel belt-like arrangement in cervical enamel, whereas groups of prisms came to be arranged in a concentric cusp-centered lamellae close to the cusp (Fig. 15). A reconstruction of enamel prisms and their successive Tomes’ processes have demonstrated that a group of ameloblasts with the similarly inclined secretory faces of Tomes’ processes formed a single group of enamel prisms directed in the same sideward direction (Hanaizumi et al., 1994). This suggests that each ameloblast develops a Tomes’ process projection inclining the secretory face either clockwise or counterclockwise around the tooth axis at equal intervals in the beginning of amelogenesis and migrates away from the enamel-dentin junction. Therefore, the ameloblasts groupings appear to correlate with the parazones and diazones of the Hunter-Schreger bands on the longitudinal cut surface. Considering that the distribution of the Hunter-Schreger bands are different among species (Kawai, 1951, 1955; Shobusawa, 1952; Boyde, 1964; Kozawa, 1978; Rensberger and Koenigswald, 1980; Boyde and Fortelius, 1986; Gilkeson and Lester, 1989; Stefen, 1997), the patterns of ameloblast group arrangements may also vary in mammals. DIRECTION/INTERRELATIONSHIP OF PRISMS Fig. 10. Scanning electron micrographs of the exposed surface of the developing enamel cleaned of cellular components by the cell-maceration method, viewed from the cusp. (a) Enamel surface shows many groups of pits arranged in concentric circles approximately central to the cusp (arrowhead). ⫻750. Scale bar: 20 µm. (b) Each pit consists of a flat (S) and an enclosing face (N). ⫻5,400. Scale bar ⫽ 2 µm. (c) Groups of pits 363 with flat faces inclined in the same direction are arranged in a band-like zone. The pits in the neighboring groups are inclined in the opposite direction. The boundary between the neighboring groups forms a narrow zone of small pits (ⴱ) with flat faces directed toward the developing enamel surface. ⫻1,500. Scale bar ⫽ 10 µm. 364 HANAIZUMI ET AL. Fig. 11. Three-dimensional reconstructed displays of the groups of enamel prisms (p1–8 green area, d1–8 blue area) marked in Figure 3, representing the data from 30–120 µm from the enamel-dentin junction. The reconstructed groups of enamel prisms are viewed obliquely down to the tangential (left) or longitudinal (right) planes. The groups of enamel prisms marked d1–8 are shown semitransparent. Comparative anatomical studies have paid attention to potential variability in fossil and extant marsupial enamel of early mammals and indicated that the Hunter-Schreger bands first appear in Vombatus ursinus of marsupials (Lester et al., 1987; Gilkeson and Lester, 1989). From an evolutionary point of view, the early appearance of the Hunter-Schreger bands in tooth enamel may coincide with adaptation to changes on occlusion and in diet diversity as well as in scale (Koenigswald et al., 1987). An increase in chewing force would produce the chances of a putative fracture and crack propagating right through the enamel surface. The Hunter-Schreger band formations resulting from the grouping of ameloblasts strengthen the enamel in respect to the functional load owing to the enamel prisms decussation in overlying and underlying bands (Osborn, 1965; Sundstrom, 1966; Rensberger and Koenigswald, 1980; Boyde, 1983; Boyde and Fortelius, 1986; Koenigsald, 1987). In comparing cuspal enamel where groups of enamel prisms show a concentric pattern arrangement tilting apically with cervical enamel where groups of prisms are arranged horizontally in parallel bands, it appears advisable to consider that the structure of the former enamel restrains horizontal bending stress besides compressive force more strongly than that of the latter, because more a complicated three-dimensional course and arrangement of prisms appear to increase the strength of enamel. We were not able to explain conclusively why ameloblasts are grouped to form belt-like arrangements synchronously inclining the secretory faces of Tomes’ processes toward the same direction. A further examination of the developmental changes of Tomes’ processes in amelogenesis, including a comparative analysis of various species, will be necessary for this. Possible Power Inducing Sideward Displacement of Ameloblasts There have been two hypothesis on the origins of the power inducing the sideward displacement of ameloblasts. Nishikawa et al.(1988, 1990) observed that contractile proteins of actin, myosin, tropomyosin, and alpha-actinin, as well as adherent junction protein of vinculin and radixin, were located at the terminal webs in the secretory ameloblasts of the rat incisor. They considered that the terminal web filaments contribute to the power-generating machinery of ameloblast transverse movement to produce enamel prism decussation (Nishikawa et al., 1988, 1990; Nishikawa, 1992). Although the contractile power may explain decussation of single layers of enamel prisms as in the inner enamel of rodents, it remains unclear how to demonstrate more complex decussation of enamel prisms in another species. Further, sideward movement of ameloblasts resulting from shrinking seems to need an anchor of fixation, but such an anchor has not been shown. DIRECTION/INTERRELATIONSHIP OF PRISMS groups in both the tangential and longitudinal planes. (b) Groups of enamel prisms (d1–6, blue area) oriented in the opposite direction of those in (a). (c) Image combining (a) and (b). One group of enamel prisms (d1–6, blue area) made semitransparent. (d) A reconstructed image of two groups of enamel prisms (green area; p1–8, blue area; d1–6) over an interval of 70 µm. One group of enamel prisms (d1–6, blue area) made semitransparent. 365 Fig. 12. Three-dimensional reconstructed displays of the groups of enamel prisms (p1–3, d1–6) in Figure 8b, representing the data from 290–340 µm (a-c) and 270–340 µm (d) from the cusp of the tooth. In these images, the reconstructed belt-like zones are viewed obliquely down to the tooth axis to display both the horizontal and longitudinal planes. (a) Extracted images of groups of enamel prisms (p1–3, green area) oriented in the same direction over an interval of 50 µm. Groups of enamel prisms appear almost concentrically arranged around the tooth axis. Note the confluence and divergence of the 366 HANAIZUMI ET AL. Fig. 13. Three-dimensional reconstructed display of the numbered enamel prisms (white area) and the groups of enamel prisms (E1,3,5 green area, E2,4 blue area) in Figure 5 over an interval of 50 µm, viewed obliquely from the surface. (a) The cut ends of enamel prisms (pink area) initially perpendicular to the boundaries of the groups of prisms come to show a periodic tilting pattern approaching the surface. (b) A substruction of the blue colored groups of prisms (E2,4). In a single group, the horizontal tilt angles of prisms toward the enamel-dentin junction tend to be greatest at the center and smallest near the boundaries. Fig. 14. Three-dimensional reconstructed display of the numbered enamel prisms (white area) and groups of enamel prisms (E1,3 blue area, E2 green area) shown in Figure 9 over an interval of 50 µm, viewed from the cusp of tooth. (a) The cut ends of enamel prisms (pink area) tilt sideward when approaching the cusp, but prisms do not display the periodic tilting pattern shown in the cervical region. (b) A substruction of the green colored groups of prisms (E2). Note the number 1–4 prisms (arrowheads) twisting to run in opposite sideward directions after crossing the boundaries between the neighboring groups of enamel prisms. DIRECTION/INTERRELATIONSHIP OF PRISMS 367 hindered by restraints between ameloblasts shifting in opposite directions, with the restrained sideward displacement being largest at the boundaries of the groups and smaller near the center of the groups. However, a precise knowledge of the origin of the power of the sideward displacement of ameloblasts remains to be more fully elucidated. LITERATURE CITED Fig. 15. Diagrammatic three-dimensional model of dog enamel structure. Groups of enamel prisms are oriented in pairs in opposite sideward directions and are drawn as light or dark belts. These groups of prisms appear to be nearly concentrically arranged around the tooth axis and show confluence and displacement of the prisms in the belt-like zones. Osborn (1970, 1973) postulated that enamel matrix secretion from the proximal portion of ameloblasts produces a potentially high hydrostatic pressure, resulting in a force pushing the ameloblasts radially outward. It has been necessary to demonstrate the precise three-dimensional structure of the distal end of ameloblasts, the Tomes’ processes, to substantiate Osborn’s interpretation. Scanning (Skobe, 1977; Wakita et al., 1981) and transmission (Kallenbach, 1977; Wakita et al., 1981) electron microscopic observations demonstrated that Tomes’ processes were composed of a flat secretory face slanted cervically and that its enclosing nonsecretory face was directed cuspally and that the crystallite arrangement within prisms and the prism arrangement within enamel were determined by the three-dimensional structure of Tomes’ processes. It is also well known that the long axis of enamel prisms is oriented at right angles to the secretory faces of Tomes’ processes (Boyde, 1967, 1969; Wakita et al., 1981; Hanaizumi et al., 1994). Moreover, the present results showed that ameloblasts were able to move their positions within the sheet plane of cells and that one ameloblast could change from one to another group by abruptly changing the sideward angle of tilt toward the enamel-dentin junction. These findings would allow the hypothesis that ameloblast movement is from the secretion. If there is a reaction force pushing the ameloblasts sideward due to secretion, this power may be almost equal because the secretory faces of Tomes’ processes in a single group of ameloblasts are nearly equivalent in size. Our previous studies (Hanaizumi et al., 1994, 1996) and the present results have demonstrated that a straight row of enamel prisms initially perpendicular to the boundaries of groups of prisms came to arrange itself in a half cycle of wavy curve as it approaches the enamel surface. With respect to the wavy curve arrangement of enamel prisms, our previous study (Hanaizumi et al., 1994) speculated that the crossing of ameloblasts as they moved outward might be Applebaum E. The arrangement of the enamel rods. New York State Dent. J. 1980;26:185–188. Boyde A. The structure and development of mammalian enamel, Ph. D. thesis, University of London, 1964. Boyde A. The development of enamel structure. Proc. Royal Soc. Med. 1967;60:923–928. Boyde A. Correction of ameloblast size with enamel prism pattern: Use of scanning electron microscopy to make surface area measurements. Z. Zellforsch 1969;93:583–593. Boyde A. Airpolishing effects on enamel, dentine and cement. Br. Dent. J. 1983;156:287–291. Boyde A, Fortelius M. Development, structure and function of rhinoceros enamel. Zoo. J. Linn. Soc. 1986;87:181–214. Erausquin J. The aspect of bands of Schreger in horizontal sections of the enamel. J. Dent. Res. 1949:28:195–200. Gilkeson CF, Lester KS. Ultrastructural variation in enamel of Australian marsupials. Scan. Electron Microsc. 1989;3:177–191. Hanaizumi Y, Shimokobe H, Wakita M. The three-dimensional structure of Tomes’ processes and their relationship to arrangement of enamel prism in dog teeth. Arch. Histol. Cytol. 1994;57:129–138. Hanaizumi Y, Maeda T, Takano Y. Three-dimensional arrangement of enamel prisms and their relation to the formation of HunterSchreger bands in dog tooth. Cell Tissue Res. 1996;286:103–114. Hirota F. Prism arrangement in human cusp enamel deduced by X-ray diffraction. Arch. Oral Biol. 1982;27:931–937. Kallenbach E. Fine structure of secretory ameloblasts in the kitten. Am. J. Anat. 1977;148:479–512. Kawai N. The bands of Schreger observed through the enamel surface. Okajimas Fol. Anat. Jap. 1951;26:25–28. Kawai N. Comparative anatomy of the bands of Schreger. Okajimas Fol. Anat. Jap. 1955;27:115–131. Koenigswald WV, Rensberger JM, Pretzschner HU. Changes in the tooth enamel of early Paleocene mammals allowing increased diet diversity. Nature 1987;328:150–152. Kozawa Y. Comparative Histology of Proboscidean molar Enamel. J. Stomatol Soc. Jap. 1978;45:585–606. Lester KS, Boyde A, Gilkeson C, Archer M. Marsupial and monotreme enamel structure. Scan. Electron Microsc. 1987; 1:401–420. Nishikawa S, Fujiwara J, Kitamura H. Formation of the tooth enamel rod pattern and the cytoskeletal organization in secretory ameloblasts of the rat incisor. Eur. J. Cell Biol. 1988;47:222–232. Nishikawa S, Tsukita S,Tsukita S, Sasa S. Localization of adherent junction protein along the possible sliding interface between secretory ameloblasts of the rat incisor. Cell Str. Funct. 1990;15:245–249. Nishikawa S. Correlation of the arrangement pattern of enamel rods and secretory ameloblasts in pig and monkey teeth: A possible role of the terminal webs in ameloblasts movement during secretion. Anat. Rec. 1992;232:466–478. Ohtani O, Ushiki T, Teshima A. Collagen fibrillar networks as skeletal frameworks: A demonstration by cell- maceration/scanning electron microscope method. Arch. Histol. Cytol. 1998;51:249–261. Osborn JW. The nature of the Hunter-Schreger bands in enamel. Arch. Oral Biol. 1965;10:929–993. Osborn JW. Directions and interrelationships of enamel prisms from the sides of human teeth. J. Dent. Res. 1968a; 47:223–232. Osborn JW. Directions and interrelationships of prisms in cuspal and cervical enamel of human teeth. J. Dent. Res. 1968b; 47:395–402. Osborn JW. The mechanism of ameloblasts movement: A hypothesis. Calc. Tiss. Res. 1970;5:344–359. 368 HANAIZUMI ET AL. Osborn JW. Variations in the structure and development of enamel. In: Melcher AH, Zard GA, eds. Dental Enamel, Vol. 3. Munksgard, Copenhagen: Oral Science Reviews, 1973. Rensberger JM, Koenigswald MV. Functional and phylogenetic interpretation of enamel microstructure in rhinoceroses. Paleobiology 1980;6:477–495. Shobusawa M. Vergleichende Untersuchungen über die Form der Schmelzprismen der Saugetiere. Okajimas Folia. Anat. Jpn., 1952; 24:371–392. Skobe Z. Enamel formation in the monkey observed by scanning electron microscopy. Anat. Res., 1977;187:329–334. Stefen C. Differentiations in Hunter-Schreger bands of carnivores. In Koenigswald MV, Sander, PM, eds. Tooth Enamel Microstructure. pp. 123–136, 1997. Sundstrom B. Schreger bands and their appearance in microradiographs of human dental enamel. Acta. Odont. Scand. 1966;24:179–194. Ushiki T, Ide C. Three-dimensional organization of the collagen fibrils in the rat sciatic nerve as revealed by transmission- and scanning electron microscopy. Cell Tissue Res. 1990;260:175–184. Wakita M, Tsuchiya H, Gunji T, Kobayashi S. Three- dimensional structure of Tomes’ processes and enamel prism formation in the kitten. Arch. Histol. Jap. 1981;44:285–297.