Written By Aubrey Whymark 2013 - 2018



Spallation is the brittle failure of the tektite glass, primarily in response to rapid temperature changes.

Once the tektite has lost its inherited cosmic velocity the ram pressure decreases and no longer heats the tektite. The air instead cools the surface. In medium and large bodies the cooling creates an aerothermal stress shell. The stress shell is typically lost as the tensile stresses exceed the strength of the glass due to the re-entry heating having been unevenly distributed through the body. A few medium-sized distal bodies retain their flange and aerothermal stress shell. The small and very small distal bodies will only have a very weak aerothermal stress shell and it will not exceed the strength of the glass.

The tensile stresses are induced by the temperature differential between the exposed anterior surface and the interior of the tektite body. When glass is heated it expands, when cooled it contracts. Glass is a poor conductor of heat and if different parts of the tektite glass body are at significantly different temperatures then brittle failure will follow.

In order to induce brittle failure one must be dealing with a material that behaves in a brittle fashion. In the early stages of tektite formation the tektites were molten or hot bodies that behaved in a ductile fashion. Tektites cooled from the outside-in, so in the early stages the exterior became brittle, but the interior remained hot, thus limiting spallation. Proximal indochinites and large medial tektites, such as philippinite bifurcated cores, do not develop circular ‘navel’ cracks as they were likely still hot during re-entry.

A further factor to consider is that during re-entry the tektite is under significant deceleration pressure and this will also influence brittle failure of a developing line of weakness. It is thought that this deceleration pressure results in Hertzian cone formation, leading to navel production.

The temperature change that induces spallation may result from the anterior of the tektite either being rapidly heated during re-entry or due to a heated surface being rapidly cooled. It is believed that the latter is by far the most important, if not the only valid mechanism.


Spallation is the brittle failure of the tektite glass, primarily in response to rapid temperature changes. The significant temperature change appears to be when the inherited cosmic velocity is lost and the tektite, instead of being heated is now rapidly cooled in the upper atmosphere.

Discussion of spallation mechanism

Spallation occurs in all the main groups of tektites, but is more important in medial and distal forms. These forms, more so the medial tektites, also contain circular fractures. Two variants of the spallation process appear to occur, one producing circular fractures and curved spallation surfaces, the other producing featureless flat spallation surfaces.

The circular fractures are believed to be Hertzian cones (or incipient cones when seen in sectional view). Hertzian cones form by point pressure. These circular fractures are surrounded by a second separate conchoidal, sometimes apparently smooth, fracture that curves inwards to meet the circular fracture.

Circular fractures are positioned in a ring roughly one third to half way from the anterior margin of the tektite to the centre of the anterior surface (see Figures 10.1 and 10.2). The circular fractures are roughly centrally placed between the radial fractures of the ‘triangular’ shell that is lost, with the distance from the margin possibly controlled by the curvature of the body. The circular fracture, or Hertzian cone, arrangement is not random. Shell loss appears to be divided into two: An area between the margin and circular fracture and an area from the circular fracture to the centre of the anterior.
FIGURE 10.1: The anterior view (left) and side view (right) of a 67.9 gram (48 mm maximum diameter) shield core from Paracale, Bicol, Philippines. Note the circular arrangement of the navels (naturally etched). (Catalogue # PB1115730).
FIGURE 10.2: The anterior view of a 70 gram core from the Davao region, Philippines. This specimen has suffered very little natural etching. One can observe navels in the middle of lost triangular shell fragments.
The circular fractures are usually flat-topped with a broken surface, but sometimes contains an attached piece of shell (see Figures 10.3 to 10.6) or a central mushroom-like protrusion as seen in Figure 10.7. Circular fractures only occur on anterior core surfaces. They do not occur on proximal or large medial (bifurcated core) forms that would have probably been too hot during re-entry. They do not occur on small pyramidal-type distal australite cores which were heated throughout during re-entry. The straight and circular fractures are often chemically enhanced to form u-grooves termed gutters and navels, respectively.
FIGURE 10.3: The anterior view (left) and side view (right) of an 18.5 gram (31x30x18 mm) indicator shield core from Sipalay, Philippines. This core is relatively fresh. It is possible that there has been further anterior shell loss on the ground, but etching has been minimal.
FIGURE 10.4: The anterior view (left) and side view (right) of an 25.6 gram (37x29x20 mm) indicator shield core from Sipalay, Philippines. This core is relatively fresh. It is possible that there has been further anterior shell loss on the ground, but etching has been minimal.
FIGURE 10.5: The anterior view (left) and side view (right) of a 45.5 gram (38 mm diameter) indicator globular core from an unknown locality in the Philippines. This core is relatively fresh. It is possible that there has been further anterior shell loss on the ground, but etching has been minimal. (Catalogue # PX1117112).
FIGURE 10.6: Left: The anterior view (left) of a 64.5 gram (44 mm diameter) indicator shield core from an unknown locality in the Philippines. (Catalogue # PX1117109). Right: The anterior view (left) of a 67.1 gram (46 mm diameter) indicator shield core from an unknown locality in the Philippines. (Catalogue # PX1117111).
FIGURE 10.7: A 33 gram (42x29x21 mm) naturally etched tektite with ‘mushroom’-like protrusions from Davao region, Philippines. The top surface is the anterior surface. This specimen acquired the name 'Shrek' after an animated film character. (Catalogue # PD1117459).
Stresses, which lead to cracking and brittle failure of the anterior surfaces are clearly caused a rapid temperature change: In this case probably very rapid cooling of the exterior once a significant proportion of the inherited cosmic velocity is lost. Cooling results in contraction of the glass and creation of mode I opening tensile stresses. When these tensile stresses exceed the strength of the glass a crack will form, ultimately resulting in shell loss. Spallation in all tektites is taken to occur in the latter stages of re-entry. Although the thermal regime is by far the most important factor influencing fracturing, the intense deceleration pressure also appears to play an important role in the creation of Hertzian cones.
Circular fractures, or navels, are considered to represent sectional views of Hertzian cones (see Figure 10.8). To form a Hertzian cone, pressure is required from a point source. The pressure is considered to come from intense deceleration pressures during re-entry, but it is somewhat problematic as to how the pressure is focused to a point source in order to produce a Hertzian cone. Alternative explanations for the circular navel structure were investigated, but through experimentation it appears that navels are genuine Hertzian cones. The breakthrough came in recognition of effectively a two-phase formation process as oppose to a single fracturing event meeting centrally to produce a circular detachment scar.

Hertzian Cones

Hertzian cones are formed by point pressure. They are believed to form as the anterior shell begins to crack and break away due to thermal stresses. Due to body force, the last remaining point of attachment (in the center of the triangular or polygonal shell fragment breaking away) is under point pressure and a Hertzian cone may be formed if pressure is sufficient.
FIGURE 10.8: A Hertzian cone. The arrow represents the Hertzian load or direction of point force.
The issue of navel formation was discussed with Dr Hugo Anderson-Whymark, the author’s twin brother, who works as an archaeologist specialising in lithics and flint knapping. He suggested the cone shape is actually akin to a pot-lid fracture, also known as a fire-pop fracture, and more broadly termed a thermal fracture or heat fracture in flint knapping circles. This explanation had its problems as the navels are not randomly distributed over the anterior surface of tektites and the similarity appeared limited. From illustrations and samples personally examined, pot-lid fractures appeared to be simple hemispherical pits and lacked the navel-like feature. Pot-lid fracture initially appeared to be a poor analogue. These types of fractures are probably caused by the mechanical action of steam or ice formation which forces the surfaces apart through expansion.

Subsequently, Dr. Hugo Anderson-Whymark sent me an article by Rose (1860) exhibiting a fine example of this type of fracture. In Rose's (1860) example of pot-lid fracture, see Figure 10.9, one can understand the similarity. Rose referred to the projections in the cupped surface as papillæ (nipple-like) and the surface as mastoid (breast-like). These pot-lid fractures were found in flints used to face walls. They were not found on new walls; only on old walls exposed to weathering and the action of frost. Rose examined even older Roman walls and did not find the mastoid character as well developed as anticipated, leading to the conclusion that more than one process was acting on the surface of the flint to produce the papillæ. On exposing the weathered flint to repeated forcible concussions with a hammer, a cup-like cavity, having a nipple within it, was produced. So, one imagines that this flint, already weakened by weathering processes, when struck, produces Hertzian cones. The flint is then struck further/repeatedly and material is broken off from the exterior of the cone.
FIGURE 10.9: Fragment of flint from Norfolk, England, showing the cup-like cavities and papillæ. a) papillæ fully formed; b) apex of papilla just visible through the flakes. From Rose (1860).
It is apparent that the so called true 'pot-lid' fracture is a two-stage fracture caused by point pressure impact creating a Hertzian cone internal flaw, followed by tensile failure related to changes in temperature, perhaps aided by the mechanics of ice or steam. This differs from the more typical single-stage fracture acting on flawless amorphous silica by tensile failure related to changes in temperature, which creates a simple hemispherical pit or cup-like cavity.

The next step was to take this information and determine whether it could be applied to tektites. A 59 mm clear glass sphere was moderately lightly hammered 'head-on' a number of times. Very evident Hertzian cones, sometimes direct hits and sometimes slightly glancing and asymmetrical, were produced within the glass, but no material was lost. This sphere was then heated up over a flame, resulting in instantaneous spallation of the shell fragment around the Hertzian cones (see Figures 10.10 to 10.13). The body was then rapidly cooled in water resulting in further internal fracturing which appeared to join the navels, as is commonly observed in tektites (see Figure 10.14).
FIGURE 10.10: A 6 cm diameter glass sphere hammered to produce internal Hertzian cones fractures (still visible) and then heated until thermal spallation occurs. Note how the Hertzian cone remains intact. Note the fissures on the thermal spallation surface.
FIGURE 10.12: Two oblique views of the same surface of a 59 mm diameter glass sphere hammered to produce internal Hertzian cone fractures and then heated until thermal spallation occurs. Note how the thermal spall curves inwards as it approaches the Hertzian cone.
FIGURE 10.11: A 71.8 gram (41x39x35 mm) philippinite from Sipalay, Philippines showing a well preserved navel compared with a 59 mm diameter glass sphere hammered to produce internal Hertzian cone fractures and then heated until thermal spallation occurs.
FIGURE 10.13: If a Hertzian cone is formed by a previous impact, subsequent impacts or stresses may result in spallation of the area around the cone.
FIGURE 10.14: A 59 mm diameter glass sphere hammered to produce internal Hertzian cone fractures, heated and then rapidly cooled in water. Fractures occur which join up Hertzian cones. A similar pattern is seen in the 100.5 gram (53x51x31 mm) philippinite to the left from Davao region, Philippines.
As one can see, the similarity between the glass sphere with a Hertzian cone that has undergone thermal spallation is remarkably close to that seen in well preserved, minimally etched, tektites (see Figure 10.11). This leads the author to conclude that this is the likely process in tektite spallation. Problems and differences between the experimental and actual fractures do, however, exist.

Firstly, if one concludes that the circular fractures, or navels, are Hertzian cones (or incipient cones as they are cut Hertzian cones) then one must determine how they formed. Hertzian cones require a point-pressure to form, commonly by impact but also by slowly increasing pressure to a critical value. We know that navels have a non-random configuration, so random collision events are ruled out. There appears to be a relationship between the positioning of the Hertzian cone being in the central part of the triangular shell loss area. Shell loss and Hertzian cones are likely to be closely related.

Secondly, in tektites we do not see a neat Hertzian cone projecting out of the specimen as seen in Figure 10.12. Instead, one observes a flat fractured surface, usually nearly flush with the thermally spalled surface (see Figure 10.15). Where there is a projection then it is not a cone, but a chunk of shell fragment or mushroom-like protrusion (see Figures 10.3 to 10.7).
FIGURE 10.15: The anterior view (left) and side view (right) of a 65.9 gram (44x43x29 mm) shield core from Sipalay, Philippines. This core is relatively fresh. It is possible that there has been further anterior shell loss on the ground, but etching has been minimal.
It is hypothesised that rather than the Hertzian cone being formed independently earlier and then later acted on by thermal spallation that instead the two processes may be acting practically simultaneously, operating hand-in-hand. If the thermal spallation is occurring under high pressure then as the tensile thermal spallation crack opens up and runs close to parallel to the anterior surface of the tektite then pressure will go from being broadly distributed over the entire surface to a point source in the remaining attached area. A Hertzian cone might therefore be produced simultaneously as the thermal spallation crack front moves forward (see Figure 10.16).

The thermal spallation crack would stop at the Hertzian cone and then the weak neck joining the Hertzian cone with the shell is broken leaving a flat detachment scar where the Hertzian cone once existed. In some cases the shell is not lost and where this shell is etched a mushroom-like protrusion may form.
FIGURE 10.16: Possible explanation for spallation surfaces. Bottom surface is the anterior. A) The shell splits into triangular to polygonal fragments with a crack opening up perpendicular to the surface (not shown - see Figures 10.5 and 10.23). A tensile crack then forms parallel to the anterior surface, moving towards the centre of the triangular to polygonal attached shell fragment. B) When only a small area of the surface remains attached the deceleration pressure of the tektite is focused onto a small area and exceeds the tensile strength of the glass. A Hertzian cone forms, practically simultaneously. C) The tensile crack parallel to the anterior surface curves in as seen in experimentation. The shell fragment may remain attached or spall from the main tektite body if the small attached area, at the top of the Hertzian cone, fails. D) If the shell remains attached, it may be etched and leached, resulting in a mushroom-like protrusion.
In examining the problem of navel formation one must explain why navels are found in medium to small medial and distal cores, but not in large medial cores, all proximal bodies and small distal pyramidal australite cores. The temperature of these bodies during re-entry immediately seems to differentiate these two groups, one having navels, the other having none. Temperature during re-entry will go on to influence the timing of brittle failure, which in turn will influence the pressure regime under which failure occurs. If failure occurs under still relatively high deceleration pressures then Hertzian cones may form. If essentially no pressure is placed on the body then the thermal spallation will produce a flat crack, effectively parallel to the surface with no Hertzian cone formation.

The spallation mechanism in tektites is an area requiring further detailed investigation. Spallation has been much discussed and rewritten in this book. The excerpts probably make interesting reading, but have been removed in order to avoid confusion. Suffice to say, many types of fractures have been investigated. Electron microscopy of the fracture surfaces of fresher tektites and further experimental reproduction are required.

Please view subsequent pages for expanation of distal, medial and proximal spallation.

References (for this and subsequent spallation pages)

Chapman D. R. 1964. On the unity and origin of the Australasian tektites. Geochimica et Cosmochimica Acta. 28 (6): 841-880.

Cleverly W. H. 1986. Australites from Hampton Hill Station, Western Australia. Journal of the Royal Society of Western Australia. 68 (4): 81-93.

Florensky P. V. 1975. Irgitzy-tektity iz meteoritnogo kratern Ahamanshin (Severnoje Priatal'je). (=Irghizites - Tektites from the meteoritic crater Zhamanshin / Northern Aral's region). Astronomicheskii Vestnik. 9: 237-244. Also in: Solar System Research. 9 (4) (Apr. 1976): 195-200. Translation.

Izokh E. P. 1994. Microtektites of the Zhamanshin impact crater: key facts to the microtektite problem. Meteoritics. 29 (4): 477. (Abstract).

Rose C. B. 1860. On the Mastoid Appearances exhibited on the Faced Flints employed for the Outer Walls of Buildings. Proceedings of the Geologists' Association. Vol I: 192-194.