Volume 73 1964 > Volume 73, No. 4 > An archaeological assemblage of Maori combs, by Wilfred Shawcross, p 382 - 398
AN ARCHAEOLOGICAL ASSEMBLAGE OF MAORI COMBS
This paper is concerned with the results of the initial study of the assemblage of wooden combs which was excavated from a swamp site at Kauri Point, in the western part of the Bay of Plenty area of the North Island. 1 This assemblage is the only such one, known to the author, to have been discovered by means of controlled archaeological excavation. Furthermore, at a rough estimate, it at least doubles the entire number of examples of this Maori artefact which are preserved throughout the world. In all, some 330 large fragments or almost complete combs were recovered—52 from the first season and 278 from the second. All of the combs were made of wood and the majority were found to have been snapped into two or three pieces along the grain and across the teeth, which has led to there being several thousand broken teeth in the deposits.
During the first season 2 only the upper part of the swamp was excavated and the greater number of the combs recovered were in a poor state of preservation, owing to the action of roots and fluctuations in the water level of the swamp. Indeed, it is clear that some combs had entirely disappeared, for the inorganic artefacts in the deposits appeared at a slightly higher level than the highest combs. However, a sufficient number of well-preserved pieces were recovered to indicate that the forms of the combs were varied, for though the most common were curved or rectangular in outlines, a few had decorations which were clearly related to the well-known motif in Maori carving called the manaia. On the other hand, there was also a knobbed design, which, while recognizable from a number of specimens in museums, was a motif whose meaning was - 383 obscure. The second season of excavations 3 went much deeper into the deposits and not only produced a larger number of pieces, but also further additions to the range of motifs. It is the purpose of this paper to distinguish some of these forms and to demonstrate that they are a reflection of a significant change in style over a period of time.
NATURE OF DEPOSITS, EXCAVATION AND CONSERVATION
The combs form part of a larger archaeological assemblage, of which the obsidian flakes have been previously described by the author in this Journal. 4 The comb fragments were distributed over an area of about 160 square feet, but three-quarters of the total were, in fact, concentrated within an area of about sixteen square feet, almost coinciding with the area of maximum density of the obsidian flakes (Fig. 1). The maximum depth of the cultural deposit is some four feet below the surface of the swamp, while the highest level at which combs were preserved was about a foot below the surface.
Plan of the excavation showing the density of the distribution of the comb fragments. The outer contour contains the total number of the combs, while the innermost contour, containing 75%, shows the maximum concentration.
The method of excavation was to work downwards, exposing an area of deposits. All structural evidence was then planned and artefacts had their position within the area and their depths recorded and were then placed in polythene bags, sealed, and packed in a cool, dark place. - 384 At the end of the excavations the materials were brought back to Auckland where they were washed and photographed before repacking to await conservation. Each comb was separately photographed, with its serial number and scale and the photographs were then prepared as index cards. Several different processes of conservation were tried. The thinness of the wood and its long burial had made the combs excessively weak, so that they could not simply be allowed to dry out because they would warp and disintegrate. The final choice of method of conservation was the Polyethylene Glycol process, whose constituent is obtainable in New Zealand and which, while it requires a considerable time for completion, can be carried out with a minimum of expensive equipment, and leaves the wood in a more attractive condition than do the other processes tried.
Combs are a widespread, though not universal, feature of material culture throughout the world, and serve a variety of functions, such as the industrial one of preparation of fibres, as well as those associated with human hair. They have been made from a wide variety of substances such as metal, animal products like bone and horn, wood, and stiff leaf ribs. The teeth which are the functional part are common to all combs, but some form of framework is necessary in order to hold the teeth together rigidly, and it is this which will serve as a handle for the purpose of combing, or as a medium for ornament. These are the two parts which make up a comb and the term teeth will be retained for the one part. A suitable term for the second part presents a slight difficulty, for to pursue the metaphor of teeth any further would be ridiculous, while terms like “handle” or “base” can have objections levelled against them. Therefore, the term frame will be employed in this paper (see Fig. 2).
The parts and dimensions of a comb used in this paper. The proportions are of an average comb.
In the matter of classification, the following scheme will be employed. All combs form a class of artefact, of which there are two sub-classes, - 385 based upon the method of manufacture. One includes all combs made from a single piece of material, while the other includes those composed from a number of pieces. These sub-classes will be called one-piece and composite. Within each sub-class there exist a range of types, which may be thought of as resulting from a specific image or form held in the mind of the manufacturer. No attempt will be made here to enter upon the controversy over the significance and reality of types; it must be accepted that without some concept of type, little progress can be made in archaeology and the problems appear rather to be in what cases different forms of typology are most suitable. The typology employed in this paper is unashamedly intuitive, having developed over a period of nearly three years, throughout much of which time other matters had been occupying the author's mind. No claim is made that the types described here, or the conclusions based upon them, are the only possible ones; however, certain results are seen to be consistent with the proposed typology and these are its justification.
DESCRIPTION OF KAURI POINT COMBS
All of the Kauri Point combs are one-piece and made from wood. The average tooth length is 7.8 cms., though some teeth on a comb may be shorter than the rest. The average length of the frame is 6.3 cms. The average width is 4.4 cms., though this may be slightly too low, due to the fact that the wider combs would have been broken more readily. The thickness is the most apparently stable characteristic, and averages 0.4 cms. (see Fig. 3). The number of teeth comes to an average of three per centimetre, which in turn gives an average tooth rate per comb of thirteen. These figures combine to give a good idea of the dimensions of what might be a typical comb and they are shown in the form of frequency distribution curves in Fig. 3.
ESTIMATE OF NUMBER OF COMBS IN SITE
As indicated previously, the assemblage consists of complete combs and large and small pieces. How many complete combs might be represented in the site has been estimated. On the evidence of the second season there were 75 complete combs, that is, excepting the teeth, which were broken short in almost every case. There were 151 pieces showing a single break and 28 with two breaks. The pieces with two breaks would need 56 pieces with single breaks to complete them: this would leave 95 pieces with single breaks to be joined—i.e. these would make 48 complete combs. Thus, the total number of complete combs represented in the second season was about 155. On a similar reckoning the first season would have produced 32 complete combs, making an excavation total of 187, which is a conservative estimate.
An inspection of the plates will show that the combs have been very carefully made and finished. This is seen both in the surface and decoration of the frames and also in the evenness and regularity of the teeth. The methods by which these were achieved are unknown, though it seems- 386
that some form of fine abrasive was used in the finishing of the surfaces. There are also a number of pieces in the assemblage which show signs either of having been repaired or of having been altered from their original forms. Most of the repairs consist of small, conically-bored holes and in one instance there are slight indications of former binding. A feature of the more obvious alterations, or attempts at alteration, is that - 387 they show much cruder workmanship; this, and the otherwise high degree of skill represented by the original pieces and, finally, the designs themselves, which, while they are very varied, are all sufficiently close to each other, strongly suggest that the making of combs was confined to a few skilled craftsmen and was not a wide-spread activity.
THE COMB DESIGNS
One face of the frame of an average comb would have an area of about twenty-eight square centimetres. Such an area was presumably necessary for the basic strength of a comb, but it also provided a vehicle for ornamental design, which, as the combs are so thin, must be essentially two-dimensional. There are some five methods of two-dimensional design: (1) silhouette or profile; (2) engraving; (3) fretting; (4) low relief; (5) painting. Although a number of the Kauri Point combs show signs of red pigmentation, there is no evidence of painted designs; but the first four methods are all employed, either singly or in combination.
As has been mentioned earlier, many of the comb forms appeared entirely inexplicable at the time of excavation and it was not until one particular comb was reconstructed from three fragments (Plate 1) that a solution was found. It will be seen that this comb is approximately oval in outline, with most of its profile studded with low knobs. In addition, there are three areas of low relief and engraved design. One of these designs protrudes outside the general area of the profile and is in the form of a miniature profile head of characteristic Maori design, with eye, nose and protruding lip. This may be seen as a smaller version of the full design on the frame, of which it forms the nostril of the nose. Once the complete design is understood as a profile face, the two other low reliefs can be seen as an eye and a mouth respectively, the latter complete with three triangular teeth and containing a small figure whose hands are easily recognizable. Points which should be noticed are the form of the eye, the low knobs round the profile, the slight indications in profile of the upper and lower lips and the nostril and bridge of the nose formed by the small face and the notch in the profile of the face with a low relief line parallel to it. All of these points appear in many of the other combs in the assemblage.
The interpretation of the design in Plate 1 as a profile face is the key to explaining the meaning of the designs of most of the other combs. Thus, Plate 2a shows a simplified, yet recognizable, derivative of Plate 1. The nostril and bridge of the nose are clearly the same, but the eye and mouth are missing. Taking this further, Plate 2a is related to 2b, 2c, of which there are numerous examples in the assemblage, and 2d. Similarly, Plate 1 is clearly related to Plate 2e through the bridge of the nose and the low knobs round the profile, though the eye and mouth are absent and the nostril has declined into a knob. From this point a series of steps may be made through Plates 2f and 2g. Lastly, in some of these the eye reappears either as an engraving or as a low relief, and in some cases (Plate 2h) the engraving links the eye, through the bridge of the nose, to an engraved version of the nostril.- 388
It was observed that the greater part of the Kauri Point combs followed one or the other of the two outlined courses, of which the simplest forms in each case consist either of plain, rounded profiles or rectangular ones. These two major groups become types in the system of classification employed here. Type A consists of all combs with rounded profiles, while Type B includes those with rectangular profiles. These two basic types are to some extent dictated by the nature of the material, for any other form of deep, decorative design would tend to weaken the frame, though examples of such design are found elsewhere. Such an alternative approach is, in fact, seen in the case of Samoan combs with fretted designs, where the weakening of the frames has been offset by making them of great length. Within each type there are many variations, as will be seen in Figs. 4, 5, and 6. These variations may be thought of as successive simplifications of the original profile, face design, until the only hint as to the original design is a slight notch cut into the profile or a knob left standing outside the profile. Type B illustrates this phenomenon particularly well, for the design may be seen as a series of successive modifications of a plain rectangular profile by the cutting of notches. In turn, this form of decoration may or may not be supplemented by a low relief or engraved design (Fig. 6), of which the latter, engraving, appears- i
PLATE 1- ii
Comb of Type A, consisting of a profile face. The nostril is represented by a small head and the mouth contains a small creature, recognizable by the hands.
PLATE 2- 389
Nine combs, of which the upper row are of Type A and the lower, Type B. Comb (i) has a particularly delicately carved manaia motif.
FIGURE 5- 390
Combs of Type B, showing how the final form was varied by combinations of notches cut into the profiles.
to be a substitute for the former. A similar process will be observed in Type A, where the part which is retained longest is the nostril. At the same time it seems possible to interpret some of the combs with the manaia design as varieties of Type A, in which the small engraved or low relief head representing the nostril is replaced by an open fretted manaia design (Plate 2i). The three figures, while diagrammatic, are based upon original specimens of which there are, in some cases, several examples. However, these diagrams have not been prepared as a rigid classification of the varieties. Nor, though it is suggested that the trend can be seen towards simplification, should the diagrams be taken as showing a step-by-step process of abandonment of features. Rather, it is as if the artists were deliberately exploring combinations of features.
Combs of Type B, showing further combinations resulting from engraving and low relief designs.
STYLISTIC CHANGE DURING A PERIOD OF TIME
Systems of classification are the means of bringing order to an otherwise confusing array of objects. However, it is desirable that such order should be capable of illuminating a problem; in particular, it should assist in the discovery of processes, such as changes in style between different points in time or area. At first sight the variety of forms from Kauri Point is bewildering and there would appear to be no way of knowing whether they just represent a great variety of forms in one place at a particular time, or forms from different areas, but still at a single point in time. Finally, there is the pervasive feeling that the span of time in New Zealand prehistory is too short, especially as reflected in archaeological sites, to provide much evidence for evolutionary change. However, in the excavation of this site it was observed that there were floors in the deposits, which effectively sealed the artefacts in the lower levels from those above them; in other words, though there was no way of knowing how long the site had been used, it was reasonable to consider that it had not been formed at a single instant, but that artefacts had found their way to it successively. To test this possibility it was necesssary to have a system of classification, such as has already been outlined, and also to apply it to the combs excavated from different levels. Two methods of analysis by which any typological change might be tested have been employed at Kauri Point; they both had the virtue of extreme simplicity, were rapidly completed, and produced clear, unambiguous results. The first method of - 391 analysis was carried out by the author, while the second analysis, using the same system of types, was carried out by another worker in a different part of the country. 5 As the results confirm each other they indicate that the types are valid ones and that with certain qualifications, any other worker might be expected to arrive at similar conclusions.
The first analysis resembles a part of the method often known as seriation, which has been widely used in other parts of the world and has been applied in the Pacific to the study of Hawaiian fishhooks. 6 The procedure was to divide the combs into groups of twenty on the basis of their serial numbers. As artefacts other than combs belonged to the full assemblage and had been assigned serial numbers there were no instances when a group consisted of a full twenty; the range of combs was between eight and seventeen and an average of thirteen in a group. Each group was then broken down into four parts consisting of plain round-topped combs and those with a knobbed design on the side, and the plain rectangular, or flat-topped combs, and those with notches cut into the top.
Analysis of combs by proportions of types within each group of twenty serial numbers. The column on the right shows the actual number of combs in each group. The proportions are expressed as percentages in the bar diagrams and it will be seen that the two right hand columns decline in relative importance in the upper levels, whereas the left hand columns increase.- 392
FIGURE 8- 393
Decorated specimens of Types A and B plotted against the section of the excavation. Type B is lower in the deposits and therefore the earlier form. The depth has been exaggerated to twice that of the horizontal scale.
In other words, the basic distinction between Type A and Type B was made, but emphasis was put upon the more varied, decorated forms. Within each group the proportions of the four parts were worked out as percentages and the total is represented in the bar diagram in Fig. 7. It will be seen that Type A, particularly its decorated varieties, is far more strongly represented among the lower numbers, whereas it is Type B and its variations which is predominant in the higher numbers. Two points should be made here: in the first place, it is inconceivable that if the forms were indiscriminately mixed up, such a clear pattern would develop 7; in the second place, the serial numbers are a crude indication of the order in which the combs had originally been placed in the swamp. The low numbers represent the first to be excavated, which, in turn, were the last to have been placed in the swamp, whereas the high numbers represent the last to be excavated, and these were the earliest ones to have found their way there.
Both the types and results of the foregoing analysis were tested by the following, second analysis. As previously noted, the depth of each comb was recorded during the excavation and thus its position within the deposits could be reconstructed. The decorated varieties only of Types A and B were selected and their depths and positions were plotted against the section (Fig. 8). A distinctive symbol is given for each type, and it will be seen that they are quite clearly sorted out into an upper and lower group, the varieties of Type B being the lower and therefore the earlier.
DISCUSSION OF ANALYSES
The results of each analysis are the same: combs of Type B are earlier in the deposits than those of Type A. However, in the absence of radiocarbon dates and as yet of any other method of determining the age in years of deposits in New Zealand, there is no way of estimating the length of time involved. It will be observed in Figure 8 that there is no precise dividing line between Types A and B, and that examples of Type B persist with Type A. This suggests that the change was a gradual one and not the result of the substitution of a different type by a new group of people in the area. A far more likely explanation is that as the combs were gradually discarded, newer forms became mixed for a while with older forms, until all the old ones had disappeared, and only the newer forms remained. Another point is that the evidence for depth disposes of what would have been a reasonable hypothesis, that the earliest form was the most complete design (Pl. 1), out of which both Types A and B had evolved together, and that the latest examples were those with the least amount of working.
It is worth discussing here why the first analysis, using serial numbers, should produce so clear a result. The method of seriation is usually employed in connection with materials excavated in a series of standard - 394 depths, but this method of excavation was not, as such, employed in the Kauri Point swamp; however, the deposits, being waterlogged and clearly having been formed under water-logged conditions, imposed their own control over the excavation. As the deposits were formed, at least in part, by water, they inevitably assumed an almost horizontal level. Similarly, during excavation, excavators could not dig below the general level, because any such hollow would be immediately filled with water, thus the excavation in some respects resembled a speeded up reversal of the method of formation of the deposits.
At the time when New Zealand was first explored by Europeans, the Maori comb was a recognized head ornament. It is impossible to estimate how common it was, but the well-known engravings of the paintings of Sydney Parkinson, the draughtsman on Captain Cook's first voyage, show that feathers and combs were worn by men. Furthermore, two of Parkinson's pictures show combs which are recognizably of Type A—indeed, one shows an example which might almost have come from the Kauri Point assemblage. There are also scattered later references to the use of the comb during the earlier years of European contact; for example, J. L. Nicholas refers to a single comb in the Bay of Islands and says that it was treated with great reverence. 8 But it is difficult to locate exactly when combs went entirely out of use. Thus, J. S. Polack, who left New Zealand in 1837, after a stay of some six years, stated that “In addition to feathers, combs of wood, fancifully carved, are also inserted. Formerly, a comb . . . from six to fifteen inches long, was placed perpendicularly at the back of the head, but have become long since discontinued”. 9 Similarly, many of the early drawings of Maoris, culminating in those made by G. F. Angas, which were published in 1846, fail to show men wearing combs. On the other hand, Ernest Dieffenbach, writing at about the same time, implies that combs were still being worn. 10
Probably, the explanation lies in the fact that many of the early observers were in contact mainly with Maoris connected with missionary settlements, such as was the case of the famous Hone Heke, whose portrait was drawn by Angas. Those observers who went outside the area of missionary influence would have found the comb still in use. The reason why the influence of the missionaries led to the abandonment of the comb was that they introduced the practice of cutting the hair short. This was certainly a missionary influence in other parts of the Pacific, as is noted by Sir Peter Buck for the Cook Islands, where he says that the “prevailing style of short hair for men started with the Missionaries”, and quotes the missionary, Williams: “the heathen wear very long hair; and as the Christians cut theirs short, to cut the hair had become a kind of first step in renouncing heathenism;” 11 These would seem good grounds for associating the discontinuance of the wearing of combs with the influences - 395 of the missionaries: however, it seems possible that they were still worn in some isolated areas, though too much should not be made of the idea of isolation in New Zealand. Lastly, they may still have been treasured for their sacred qualities some time after they ceased to be worn.
There is little evidence for any changes in the designs of combs during the years of early European contact; the illustrated examples appear to belong to Type A, but there are varieties not found in the Kauri Point assemblage, while a number of carvings of human figures illustrated by Augustus Hamilton, 12 but of unknown age, possess composite combs. Dr. H. D. Skinner has suggested that the whale bone comb is a later development than the wooden one, and he has given an instance of such a whale bone form, which had been cut out by means of a steel saw. 13 This may well be so, but there were certainly bone forms in use as early as 1773 when William Bayly, one of the astronomers on Captain Cook's second voyage, visited Queen Charlotte Sound. 14
Combs have been found in several archaeological assemblages in New Zealand, of which two, Oruarangi 15 and Murdering Beach, 16 in spite of their unreliable associations, may be reasonably considered to be late. In both cases the combs are of Type A, made of bone. Those from the southern site, dated to the first years of the 19th century, resemble crudely made versions of some of the Kauri Point specimens. The northern site may be a little more recent and includes several large examples with the more elaborate manaia decorations.
The only evidence for potentially earlier forms at present comes from two sites excavated in the 19th century, whose assemblages were subsequently studied by Dr. Skinner. These were the sites of Moa Bone Point Cave and Monck's cave, near Christchurch. In each case the assemblages are mixed, part of them being in association with Moa remains. But interestingly enough, both sites produced coarse-toothed combs of wood with large open design, fretted frames, and other uncharacteristic decorations. 17
THE DISTRIBUTION OF COMBS OUTSIDE NEW ZEALAND
The comb appears to have an uneven distribution among the Pacific Islands, though throughout much of the area the evidence is based upon ethnographic observations, in many cases made long after European contact. Percy Smith illustrated composite and round-topped, one-piece forms from Niue, 18 while E. G. Burrows recorded the use of composite combs, made from coconut leaflet midribs, from Uvea. 19 From Samoa come both composite forms and elaborate, perforated intaglio carved combs, which Buck suggested were post-European introductions. 20 Else- - 396 where in Polynesia they appear to be virtually absent or unrecorded and for obvious reasons in areas such as Hawaii, with its own distinctive style of wearing hair. But the Polynesian outlier Rennell Island formerly possessed both one-piece and composite examples which are noted by Kaj Birket Smith to have been expressly stated as imitations of Melanesian types. 21 Combs are also well-represented in recent times in the Ellice Islands, where there are one-piece forms whose profiles were cut into geometric shapes. 22 Elsewhere, to the West, combs are still found among the New Guinea Highlanders 23 and at least until recently among the Negritos of Perak on the Malayan Mainland. 24
In the Northern Pacific the distribution of the comb is far more thoroughly documented, and assemblages have been preserved in a number of dated archaeological sites. Professor A. Leroi-Gourhan's distribution map 25 shows one-piece forms ranged from Hokkaido, the Kuriles and Kamchatka, through the Bering Straits, Alaska and Northern Canada to as far afield as Greenland and British Columbia.
There is a particularly impressive assemblage of bone one-piece combs from the site of Ust Poluyishkoe, close to Salekard in the north of the West Siberian Plain, 26 which is dated back to between the fourth and second centuries B.C. 27 However, there are considerably earlier combs, such as those from the Laurentian Cultural Stage of the North Eastern United States, dated to about 2,000 B.C. and which are thought of by J. A. Ford as possible ancestors of the later North American examples. 28 On the other hand, the one-piece comb has an even earlier ancestrv in the Old World, for it is associated with the Danish Ertebolle Culture, dated between the fifth and third millennium B.C., 29 and there are even more highly finished forms, roughly contemporary, from the Badarian of Egypt, dated to about the fourth millennium. 30
The foregoing comparative study, though a cursory one, is analogous to the frequency distribution curves employed earlier, for it indicates the dimensions to which the combs under discussion belong. It is thus most unlikely that the comb was an independent invention in New Zealand, though it seems that some time after its introduction it became highly developed along its own individual lines. It would also be absurd to trace its diffusion as through the migration of peoples from such centres as Egypt or Denmark. Instead, Leroi-Gourhan's reasoning that such an object as the comb diffused as an idea, quite independent of population migrations, may be followed. 31 The available evidence suggests that the area from which the idea dispersed throughout the Pacific, was the North, and this corresponds to a similar, though far more - 397 thoroughly documented, suggestion put forward by Bengt Anell, for the diffusion of Polynesian fishing gear. 32
Within Polynesia itself the distribution is too uneven and poorly documented to provide a satisfactory basis for discussion. The evidence need not conflict with Roger Duff's hypothesis that New Zealand was first settled from Tahiti, 33 for in this case the comb could have been preserved in New Zealand long after its use had died out in Tahiti, or it could have been subsequently introduced either with a local movement of people or as an idea.
The Kauri Point assemblage has provided evidence by which the design of Maori combs may be interpreted as the profile of a head. The assemblage is divided into two types, which are called here A and B, of which Type A is demonstrated to be the later. Varieties of Type A are shown in illustrations made at the time of Captain Cook's first voyage; they also form the majority of specimens in museum collections and appear to have continued in use until they became virtually extinct about 1840. On these grounds it is reasonable to suggest that Type B is truly Prehistoric, though there is as yet no reliable evidence for its ancestry nor for its exact age. It may also be noted in passing that this is the first instance in which a part of the evolution of a portable artefact has been identified and demonstrated by means of stratigraphic archaeology in New Zealand. Finally, it appears that New Zealand was at the end of one route by which the comb was diffused, and it is likely that the idea entered the Pacific through the North.
I wish to thank the Auckland University Archaeological Society for making the work at Kauri Point possible, and particularly those of its members who excavated in the swamp. I also wish to thank Mr. W. Ambrose for his encouragement and assistance and Mr. J. Hjarno for his assistance and work of conservation. I thank all of those who have discussed the combs with me, among whom I must mention Dr. T. Barrow. And lastly I wish to acknowledge the invaluable work of the Anthropology Department technicians, Messrs. Peters and Schollum, at Auckland University.
1 Shawcross 1962:51.
2 Shawcross 1962:53.
3 Shawcross 1963:54.
4 Shawcross 1964:7.
5 I wish to express my gratitude here to Mr. Jan Hjarno for carrying out this analysis.
6 Emory, Bonk and Sinoto 1959.
7 This statement is supported by the Chi Squared test which was subsequently applied to the data. The result was 59.5, which is considerably below the P = 0.1% level and is thus statistically “highly significant”. I wish to thank my colleague Dr. R. C. Green for advice on this point.
8 Nicholas 1817:98.
9 Polack 1840:180.
10 Dieffenbach 1843:54.
11 Buck 1944:104.
12 Hamilton 1896:Pl. XVII.
13 Skinner 1930:284.
14 Bayly 1773:12th April.
15 Fisher 1934:275.
16 Skinner 1959:158.
17 Skinner 1923:93; 1924:158.
18 Smith 1903:66.
19 Burrows 1937:137.
20 Buck 1930:627.
21 Birket Smith 1956:21.
22 Koch 1961:92.
23 Dr. Graham Jackson, personal communication.
24 Evans 1927:29.
25 Leroi-Gourhan 1946:465.
26 Moshinskaya 1953:98.
27 Talitskaya 1953:246.
28 Ford 1958:208.
29 Bronsted 1957:121.
30 Baumgartel 1960:47.
31 Leroi-Gourhan 1946:247.
32 Anell 1955:247.
33 Duff 1956:4.