The relation between ameloblasts, developing enamel surface pits and enamel prisms is well established . Each pit contains the Tomes’ process of one ameloblast. The number and pattern of pits correspond to the number and packing pattern of ameloblasts. The topography of the ameloblast–enamel interface may be an important factor for the orientation of enamel crystals, the crystals tending to be oriented perpendicular to the mineralizing front or to the secretory aspect of the ameloblasts. The resulting differential crystal orientation is the basis for a distinction between prisms and interprism in the fully formed enamel. The crystals developing in the matrix secreted at the bottom of a pit are integrated into a prism, while the crystals developing in the matrix secreted at the enamel surface proper (at the top of the ridges delimiting the pits) are integrated into the interprism. The surface of the Tomes’ process facing the pit wall is non-secretory. The line of junction between the pit floor and the pit wall corresponds to the future prism sheath.

In the enamel of primates, including man, the pits and prism profiles are typically arcade-shaped and most commonly arranged in staggered horizontal rows . The convexity of the arcade of pits and prisms generally points incisally/occlusally, i.e., in the direction of the prism inclination and the tangential component of ameloblast movement. The direction of entry into the pits, i.e., the direction along the pit wall and along the prisms, is correspondingly inclined incisally. When ameloblasts move with a transverse component, as during formation of the Hunter–Schreger bands, the prisms attain a transverse component of inclination. The orientation of arcades of pits and prism profiles and the entry direction of pits correspondingly attain a transverse component.

Thus, the orientation of pit arcades reflects the orientation of prism-profile arcades, the inclination of prisms, and the direction of tangential ameloblast movement. Pit-arcade orientation is therefore an important measure for characterizing enamel structure and development. After a mapping of the occurrence and distribution of various pit outlines in human deciduous enamel, it became clear that a large number of pits appeared to have their arcades pointing cervically, i.e., in the opposite direction to what was expected. Our aim now was to reinvestigate the same material in order to solve the apparent paradox of pit-arcade orientation.

The deciduous tooth germs of a formalin-fixed human fetus of 239 mm crown–rump length were carefully enucleated and the ameloblasts mechanically removed. Formalin fixation is known to facilitate the removal of ameloblasts from the developing enamel surface . The tooth specimens were critical point-dried, mounted on aluminium stubs and sputter-coated with gold–palladium. The specimens were observed in a Cambridge stereoscan 150 MKII operated at 20 kV and viewed at different angles of observation relative to the enamel surface.

The pits on the developing enamel surface of the deciduous tooth germs presented variable outlines, but they were often arcade-shaped . Over large areas most of the arcades appeared to point in a cervical direction . Where cracks were present in such regions, allowing observation of the subsurface enamel structure, it could be demonstrated that the prisms were inclined incisally . When the specimens were tilted so that the direction of observation approached the direction of prisms, the pits appeared compressed cervico-incisally and attained the shape of arcades pointing incisally . Stereo-pair micrographs (a) and (b), (c) and (d)) revealed that the pit floor was flat and small, and was positioned in the cervical part of the pit. The curved side wall of the pit was considerably larger and constituted the incisal and lateral parts of the pit. The transverse interpit ridges tended to be somewhat swaybacked , contributing to a convex cervical and a concave incisal pit border when the enamel surface was observed at right angles. In addition, the transition from the pit to the surface appeared more abrupt in the cervical than the incisal direction.

Although the pits on the surface of developing human deciduous teeth exhibit variable outlines when viewed perpendicular to the developing surface , i.e., perpendicular to the future Retzius lines, many of the pits are arcade-shaped. The micrographs in the two works cited above and the present study revealed that many of the pits give an impression of having their arcades pointing cervically. This is contrary to what was expected. With a cervical orientation of the pit arcades the corresponding prisms should be inclined cervically, i.e., they should pursue a cervical course from the enamel–dentine junction to the enamel surface. This is known not to be the case. Also, the presence of vertical cracks in the surface of the teeth studied allowed direct inspection of prism direction in the subsurface enamel, disclosing an incisal inclination of prisms. This apparent paradox was solved when the specimens were tilted so that the pits could be observed in the direction of the prisms, changing their outline to a compressed arcade-shape and reversing their orientation so that they now preferentially pointed in an incisal direction. That the angle of observation is important for the appearance of pits has also been noted by Boyde. Thus, when analysing pits on developing enamel surfaces in relation to ameloblast movement and prism inclination, it is not a question of finding the best angle of observation, but the correct angle of observation. This usually implies an incisal angle, as prisms tend to be inclined incisally. However, in rhinoceros enamel, where prism decussation occurs in a cervical–incisal/occlusal direction, group of prisms are inclined cervically. Their pit and prism-profile arcades point cervically and the entry direction of pits inclined cervically.

The following circumstances contributed to the visual illusion of pit arcades pointing cervically. With a perpendicular angle of observation and an incisal pit-entry direction, (a) the cervical pit border stood out more distinctly than the incisal pit border, and (b) the swaybacked tendency of the horizontal interpit ridges made them distinctly convex in cervical direction. In addition, as the transition from a prism to the interprism generally is more gradual at the open side of the arcade than at the arcade convexity , the more gradual transition from pit to surface in an incisal direction than in a cervical direction observed here reinforced the impression of arcades pointing cervically. It is uncertain to what extent shrinkage of the tissue during specimen preparation may contribute to the observed appearance.

It appears that the pit-entry direction may be a more reliable measure than the pit-arcade orientation for judging the related orientation of prism-profile arcades, prism inclination and direction of tangential ameloblast movement. In this connection, stereo-pair micrographs are valuable tools for identifying pit floor and wall.

In conclusion, proper attention should be paid to the angle of observation when interpreting the outline of pits on developing enamel surfaces. Pit arcade orientation should be judged at an angle of observation parallel with the prism direction.