Volume 82 1973 > Memoir No. 38: The Lapita pottery style of Fiji and its associations, by S. M. Birks et al. > Part 4: The relationships of the decorative systems of Fiji, by S. M. Mead, p 56-100
                                                                                             Previous | Next   

- 56
Part 4

The starting point of this study is the decorative system of Sigatoka. Vessel shapes upon which Lapita style decorations were imprinted or incised were described by Birks. The decorative system, consisting of an inventory of design elements and motifs, a set of zone markers and design zones, and a set of rules for constructing patterns, was then analysed. The analysis was based largely on the material from Yanuca.

Some 50 miles north of Yanuca, at Natunuku, more Lapita sherds were recovered by Shaw, who subjected the data to a parallel analysis. The two Fijian decorative systems were described as separate systems and some comparisons made one with the other.

In this final section, comparisons are extended in the first instance to include Tonga and then to include other areas in Melanesia where Lapita decorations have been found and reports published.


Nubbins occur only at Yanuca and Natunuku, but of the remaining three-dimensional design elements described for Yanuca both the vertical bar and the transverse bar occur in all three areas. The decorative bars are highly specific traits whose occurrence in both Fiji and Tonga cannot be easily explained by chance.

Counting only two-dimensional design elements and motifs now, there are 8 design elements and 33 motifs at Yanuca and 10 design elements and 32 motifs at Natunuku. Shared between the two decorative systems are 7 design elements and 20 motifs. There are 13 Yanuca motifs which do not occur at Natunuku and 12 Natunuku motifs which do not occur at Yanuca. Taking into account only the inventory of elements and motifs it would be reasonable to infer that there is some degree of divergence between the two Fijian systems. Such divergence can be explained on the one hand by regional separation of two groups of people, both inheritors of the same style of decoration, who worked independently of each other. On the other hand, the divergence might be a result of time, and in this case we postulate a single homogenous group of people inhabiting both sites but at different times.

The radiocarbon dates for Natunuku, 1290 ± 100 B.C., and Yanuca, - 57 1030 ± 90 B.C., might favour the second explanation. It is interesting to note, however, that by considering the range for each date at one standard deviation the difference is only 70 years, that is, nearly three generations of potters. At two standard deviations the dates overlap. The possibility of the two systems being contemporaneous must therefore be entertained.

Seven of the 8 two-dimensional design elements occurring at Yanuca also occur in Tonga, and, of the 33 motifs listed for Yanuca, 15 and possibly 20 occur in Tonga, but the evidence is somewhat doubtful for the latter figure. As mentioned earlier, there are in the Natunuku inventory 12 additional motifs which are not present at Yanuca, but 7 of these also occur in Tonga (Table 18).

Judgements of correspondences as shown in Table 18, and divergences, listed in Table 19, are necessarily tentative in the case of Tonga, but are more certain for Yanuca and Natunuku since similar analyses have been applied to the data. Nevertheless, it is worth while pursuing comparisons because of the kind of result which follows. Between Yanuca and Natunuku there are 20 motifs which are shared and 25 which are not. Between Natunuku and Tonga there are 21 shared motifs and 28 which occur only in Tonga. Between Tonga and Yanuca there are 15 shared motifs and 31 which occur only in Tonga.

Tonga there are 21 shared motifs and 28 which occur only at Tonga. Between Tonga and Yanuca there are 15 shared motifs and 31 which occur only at Tongs.

Before continuing the discussion it is as well to pause and point out that nearly all of the additional Tongan motifs listed in Table 19 either are early and do not appear in later levels or, except for M68, are early and occur continuously in early and late levels of the sites from which they were recovered.

What is important is that it is not so much what is shared which tends to place the sites in order of relationship, but rather the degree of divergence in their motif inventories. Yanuca and Natunuku diverge to the same degree one form the other, but both of them diverge to about the same degree from Tonga. The most reasonable explanation for such differences in the inventories is as follows. Whereas the difference between Yanuca and Natunuku inventories can be explained as a function of regional differentiation in a relatively short period of time, the difference between the two Fijian sites and Tonga requires regional differentiation and isolation plus a much longer period of time. It is assumed that elaboration of the decorative system to a point where about 30 new motifs are incorporated into the inventory requires a longer span of time than incorporation of 12 or 13 new motifs.

This may be stated in another way. In the case of Yanuca and Natunuku 20 items of the inventory are shared and 12 or 13 are not. Thus a greater number of items is shared than those not shared. On the other hand, when it happens that the number shared is far less than the number not shared, one can postulate a longer period of differentiation.

An explanation is demanded for the fact that Yanuca and Tonga share a lower number of corresponding motifs than Natunuku and Tonga. The

- 58
Comparison of Rules applied in Yanuca, Natunuku, Tonga
Element Yanuca Natunuku Tonga
DE1 2 2 2
DE2 2 Nil 2
DE3 4 2 2
DE4 2 2 Nil
DE5 2 2 2
DE6 2, 3 2 2
DE7 3 2 2
DE8.1 2 2 2
DE8.2 1 1 1
DE9 Nil 2  
DE10 Nil 2  
DE11 Nil 2  
Motif Yanuca Natunuku Tonga
M1 1 1, 1 1
M2 1, 1 1 1
M3 1 1 Nil
M4 1 Nil Nil
M5 1 1 Nil
M6 1 1 Nil
M7 1 1 Nil
M8 1, 1, 1 1, 1, 2, 2, 1, 2 1, 2
M9 1, 2, 1 Nil Nil
M10 1 1, 1, 1 Nil
M11 2, 2 2 1, 1, 1
M12 2, 2 Nil 1
M13 1, 1 Nil 1, 1
M14 1, 2 Nil 2
M15 2 2 2
M16 1, 1, 1 1, 1, 2 1, 1, 1, 1, 1
M17 3, 2, 2 2, 2, 2, 1 1
M18 1, 1, 5 1 1, 1
M19 1, 1 1, 2, 1 1, 1, 1
M20 1 Nil Nil
M21 1 Nil (?)
M22 2 Nil Nil
M23 2, 6 Nil 1 (?)
M24 1 1, 1 1
M25 1, 1 Nil (?)
M26 1 1 1(?)
M27 1 Nil Nil
M28 1, 1 1 Nil
M29 2, 2, 2 2 2
M30 Not defined 2, 2, 2 1, 1
M31 2 Nil Nil
M32 2 Nil Nil
M33 2 (?) 2 (?) or 4 (?) 2 (?)
M34 Nil 1 1
M35 Nil 2 (?) or 4 (?) 2 (?)
M36 Nil 1, 2, 2 Nil
M37 Nil 1, 1, 1, 1, 2, 2 1
M38 Nil 2 Nil
M39 Nil 1, 1 1
M40 Nil Not defined Nil
M41 Nil Not defined Nil
M42 Nil 3 4
M43 Nil 1 1
M44 Nil 1 Nil
M45 Nil 2, 2, 2 1
- 59
Additional Motifs from Tonga
Motif No. Corresponding illustration in Poulsen 1967 Relative Age
M46 K1 Fig. 50, Poulsen 1967 early
M47 K2, K4, Fig. 50 early
M48 K5, Fig. 50 early
M49 K6, Fig. 50 early
M50 K12, Fig. 50 early
M51 K13, Fig. 50 early
M52 K15, Fig. 50 early
M53 K16, Fig. 50 early
M54 L1-6, Fig. 50 early
M55 M3, Fig. 50 early
M56 N6, Fig. 50 early
M57 N7a, Fig. 50 early
M58 N7b, Fig. 50 early
M59 N8, Fig. 50 early
M60 0. 1, 2, 3, 5 , Fig. 50 early
M61 D15, Fig. 50 all through
M62 D16, Fig. 50 all through
M63 B16, Fig. 50 early
M64 D8, 12, Fig. 50 all through
M65 P7, 9, Fig. 50 early
M66 P7, 8, 9, 10, Fig. 50 early
M67 P18, 19, Fig. 50 early
M68 E1—4, Fig. 50 probably late
M69 A31, 32, Fig. 50 (possibly) all through
M70 A21, Fig. 50 (possibly) all through

There are possibly 25 motifs additional to the combined inventories of Yanuca and Natunuku.

inference that Tonga is more closely related to Natunuku than to Yanuca seems to follow, but this poses a problem. The problem increases when the number of motifs unique to each area is considered. By unique, I mean that each motif so classified does not appear in the inventories of the other two sites. It should be noted that the number of unique motifs does not coincide with the number of motifs occurring in one site and not the other when only two sites are considered. The greater the number of motif inventories compared, the greater are the possibilities that the number of unique motifs will decrease. Thus, for Yanuca there are 7 unique motifs, for Natunuku there are 5 and for Tonga there are 25. References for the Tongan motifs are given in Table 19.

The figures for unique motifs support a hypothesis of closer relationship between Natunuku and Tonga than between Yanuca and Tonga. The difficulty in accepting the hypothesis above arises when the radiocarbon dates for Natunuku and Yanuca are considered. Natunuku appears to be earlier than Yanuca, yet the evidence presented here argues for a sequence which places Yanuca first, Natunuku next and Tonga last. A radiocarbon date of 430 ± 51 B.C. for Tonga 1 offers strong evidence for placing Tonga last in the series. Thus, the problem appears to narrow down to the dates within Fiji, but it could equally be that the analysis is at fault of that the interpretation is incorrect. For example, it is reasonable to assume that a long period of continuous occupation in one locality such as in Tonga would - 60 produce a greatly elaborated inventory of elements and motifs. Probably, then, the data presented above indicate not a emporal sequence, but rather a difference in length of occupation. Lapita-making communities stayed a relatively short time at Natunuku, much longer at Yanuca and longer still in Tonga. Poulsen's carbon dates gives an incradibly long period of over 2000 years of occupation 2 in Tonga, although Groube has recently lowered this figure to about 1000 years. 3 Dates for the Sigatoka sites suggest a period of some 300 years of occupation. It is not known how long the Natunuku site might have been occupied. The interpretation suggested above could be quite a reasonable one.

Another possible interpretation is that Natunuku is the area from which the makers of Lapita pottery migrated, one gorup settling at Yanuca and another group ultimately settling in Tonga. In order for this hypothesis to be more acceptable, the date of settlement of Tonga would need to be extended to approximately 1140 B.C., the date obtained by Groube 4 from a sample of shell net sinkers. Without such an early date, the problem of the close relationship between Natunuku and Tonga remains unexplained.

Comparisons of design zones and the motifs associated with them yield little information of importance. Most areas employ a practice of dividing a pot into horizontal zones, and in all areas the standard procedure of dividing a pot into horizontal zones, and in all areas the standard procedure is to confine patterns within these zones. Watom Island provides the only exception to the rule. There a pattern may be extended beyond the zone boundaries as shown by Golson. 5

There are some motifs which must logically be placed in one type of zone and not in another. For example, patterns composed by applying Rule 1 logically fit into a Type A zone. Apart from such motifs, the choice of design zone appears to be quite arbitrary, as comparison of Mead's Table 13 with Shaw's Table 14 will show.

Comparisons of the rules used in Yanuca, Natunuku and Tonga add nothing to the discussion of temporal sequence. The question now posed is this: are the decorative systems of Yanuca, Natunuku and Tonga the same or not? Do they belong to the same style or is each one significantly different from the other? Reference to Table 18 will show fairly convincingly that the decorative systems of the three areas are indeed the same. In each area the same set of rules is used but there are difference in shape, size and combination of elements and motifs. That is to say, the differences are in the arbitrarily selected items of the inventory and not in the basic structural rules which are used in composing patterns. In all three areas, the predominant rules are Rule 1, which produces a continuous border pattern in a transverse direction, and Rule 2, which produces a discontinuous pattern in a similar direction. The direction is varied in all three areas by applying either Rule 3 or Rule 5. On present evidence, no rule additional to those isolated for Yanuca is required for either Natunuku or Tonga.

- 61

The conclusion can thus be drawn that the decorative systems of Fiji and Tonga all fall within the same Lapita style. Regional and temporal manifestations of the same style are represented at Yanuca, Natunuku and Tonga.

The analysis of rules for Yanuca yielded Rule 6, which is a rule that was applied to an extremely complex pattern. No such rule is necessary for Natunuku or Tonga. Parallels to Rule 6 are more likely to occur to the west of Fiji, for example, in New Caledonia, where very complex patterns have been found.

Similarly, parallels to the highly complex M33, which appears to be composed in a circular design zone, are closer in the west, especially at the Ile des Pins, than at either Natunuku or Tonga. These facts argue for an earlier date for Yanuca, since it shares a few specific innovations with sites to the west.

The main points which arise from the discussions of comparative material for Yanuca, Natunuku and Tonga are the following:

  • 1. The decorative systems of Yanuca, Natunuku and Tonga belong to the same Lapita style.
  • 2. Each is a regional and temporal variation of the same style.
  • 3. The decorative systems of Yanuca and Natunuku are more closely related to each other than either is to Tonga.
  • 4. The Natunuku system is more closely related to the Tongan system than Yanuca is.
  • 5. The settlement of Yanuca, Natunuku and Tonga is likely to have happened within a short period of time.

These conclusions may now be generalised into a theory. The early potters in Fiji and Tonga were heirs to a single pottery technology and to a single decorative system. If the pottery produced in Fiji and Tonga were not delivered by an outside trading group, as suggested by Groube, and the evidence in Tonga is strongly against this hypothesis, 6 then it was produced locally. The large number of motifs unique to Tonga does not favour a hypothesis of trade from Fiji. The technology and the decorative systems are so complex that the only reasonable explanation for their diffusion is that they were established by migrants of a single culture.

Furthermore, the fact that Tonga produced such a large number of unique motifs suggests that the group which settled Tonga was isolated from the Fijian communities for a considerable period of time. Since the archaeological evidence in Tonga shows no signs of later intrusion from elsewhere it is highly likely that the original Lapita pottery-makers of Tonga became the ancestors of the Polynesians. It is worth pointing out that the decorative art of Tonga, Samoa and Rotuma, at the time of first European contact, still followed the Lapita characteristics of zoning and placing border patterns in layers. These same principles were also dominant in Mangaia, especially in the adze handles produced there, and in the Austral Islands in decorated paddles. In all of these Polynesian areas, the rules for composing patterns differ little from the Lapita rules. If anything, they are less complicated. The decorative art of West and Central Polynesia supports the theory expounded - 62 by Groube and others that the ancestors of the Polynesians were founded upon the early communities in Tonga who made a distinctive kind of pottery and decorated it in the Lapita style. Parts of this style were carried into the ethnological period and applied especially to wood carving in West Polynesia and Central Polynesia.


There is now evidence of Lapita pottery in a broad expanse of Melanesia and a small corner of West Polynesia. Some specific areas are mentioned in Part 1. 7 The most recent discoveries have been in the Eastern Solomon Islands. 8 Any attempts to compare decorations from all reported sites in Oceania are fraught with difficulties. On the one hand, the quantity of material recovered from the ground varies enormously, so that in some cases few design elements and motifs are available for comparison. On the other hand, the analyses and descriptions of the material published so far vary in quality to such a degree as to make comparison plainly difficult. Often the necessary graphic evidence is not published in full, usually through no fault of the author. For an enquiry such as the present one graphic evidence is as important as the printed text. It is only in the cases of Yanuca and Natunuku that the data are analysed in the same way, thus easing the task of comparison. The Tongan material is very finely analysed by Poulsen and although his approach to the subject is different, the data can be readily fitted into the comparative scheme used in this paper. For all other published material this is not the case. Moreover, archaeological investigations of Lapita pottery have only just begun and further work will provide much of the evidence which is now lacking.

Distributions of the elements and motifs which occur in Fiji and Tonga are extended in Table 20 to include Watom Island 9 in New Britain, Lapita Site 13 10 and Ile des Pins 11 in New Caledonia, and Malo Island 12 in the New Hebrides. For reasons discussed above one hesitates to place too much reliance on the geographical distributions given in Table 20. Nevertheless, some general tendencies can be noted in terms of the present evidence.

  • 1. The most complex patterns tend to be early and thus the general drift of change is towards simplification of the motifs. Poulsen's findings in Tonga support this proposition.
  • 2. Every site where decorated Lapita pottery has been recovered yields motifs which are shared with other sites. As shown in Table 20, Yanuca shares the greater part of its inventory of elements and motifs with other sites.
  • 3. Every site produces motifs which are unique to that site and are the
- 63
Geographical Distribution of Design Elements and Motifs
Element/Motif Geographical Area No. of Areas
N — Yanuca, Natunuku 2
VB — Yanuca, Natunuku, Tonga 3
TB — Yanuca, Tonga, Site 13 1 3
GZ1 — Yanuca, Tonga, Site 13, Ile des Pins, Watom 5
GZ2 — Yanuca, Natunuku, Tonga, Site 13, Ile des Pins, Watom 6
GZ3 — Yanuca, Tonga, Ile des Pins 3
GZ4 — Yanuca, Ile des Pins(?) 2
RZ3 — Yanuca, Ile des Pins 2
DE1 — Yanuca, Natunuku, Tonga, Site 13, Ile des Pins, Watom, Malo 7
DE2 — Yanuca, Tonga 2
DE3 — Yanuca, Natunuku, Tonga, Site 13, Ile des Pins, Watom 6
DE4 — Yanuca, Natunuku, Site 13, Watom(?) 4
DE5 — Yanuca, Natunuku, Tonga, Malo 4
DE6 — Yanuca, Natunuku, Tonga, Malo(?), Ile des Pins(?) 5
DE7 — Yanuca, Natunuku, Tonga 3
DE8 — Yanuca, Natunuku, Tonga 3
M1 — Yanuca, Natunuku, Tonga, Site 13, Watom, Malo 6
M2 — Yanuca, Natunuku, Tonga, Site 13, Watom, Ile des Pins 6
M3 — Yanuca, Natunuku 2
M4 — Yanuca 1
M5 — Yanuca, Natunuku, Ile des Pins 3
M6 — Yanuca, Natunuku 2
M7 — Yanuca, Natunuku, Watom(?) 3
M8 — Yanuca, Natunuku, Tonga, Site 13, Watom 5
M9 — Yanuca, Ile des Pins 2
M10 — Yanuca, Natunuku 2
M11 — Yanuca, Natunuku, Tonga, Site 13 4
M12 — Yanuca, Tonga, Site 13(?), Watom(?) 4
M13 — Yanuca, Tonga 2
M14 — Yanuca, Tonga, Site 13(?), Watom(?) 4
M15 — Yanuca, Natunuku, Tonga 3
M16 — Yanuca, Natunuku, Tonga, Malo 4
M17 — Yanuca, Natunuku, Tonga 3
M18 — Yanuca, Natunuku, Tonga, Ile des Pins 4
M19 — Yanuca, Natunuku, Tonga, Site 13, Watom 5
M20 — Yanuca 1
M21 — Yanuca, Tonga(?), Watom(?) 3
M22 — Yanuca 1
M23 — Yanuca, Tonga(?) 2
M24 — Yanuca, Natunuku, Tonga, Site 13, Ile des Pins 5
M25 — Yanuca, Tonga(?), Watom 3
M26 — Yanuca, Natunuku, Tonga(?) 3
M27 — Yanuca, Watom 2
M28 — Yanuca, Natunuku, Site 13, Ile des Pins, Malo 5
M29 — Yanuca, Natunuku, Tonga, Site 13 4
M30 — Yanuca, Natunuku, Tonga Site 13(?), Watom, Malo 6
M31 — Yanuca 1
M32 — Yanuca 1
M33 — Yanuca, Natunuku(?), Tonga(?), Site 13(?), Watom(?), Ile des Pins(?) 6
M34 — Natunuku, Tonga, Site 13 3
M35 — Natunuku, Tonga, Watom(?) 3
M36 — Natunuku, Site 13(?) 2
M37 — Natunuku, Tonga, Site 13(?) 3
M38 — Natunuku, Site 13(?) 2
M39 — Natunuku, Tonga 2
M40 — Natunuku, Site 13(?) 2
M41 — Natunuku 1
1   indicates that correspondence is uncertain. Tongan motifs M46 to M70 appear to be unique motifs.
- 64
M42 — Natunuku, Tonga 2
M43 — Natunuku, Tonga 2
M44 — Natunuku 1
M45 — Natunuku, Tonga 2
M46 to M70 — Tonga 1

result of local development. The distributional data in Table 20 show that M4, M20, M22, M31 and M32 are unique to Yanuca, M41, M44, and possibly M36 and M38 are unique to Natunuku. The additional 25 motifs which occur in Tonga, M46 to M70, comprise a highly specific set which, on present evidence, show little relationship to inventories elsewhere. The large number of motifs unique to Tonga may be due to a long period of isolation from other Lapita pottery areas or it may be that sites intermediate between the two at Fiji and those at Tonga are yet to be found.

4. Pairs of sites may share motifs which do not occur anywhere else. Such motifs may be classed as shared innovations which provide evidence of close relationship. Examples are M3, M6 and M10, which indicate a close relationship between Yanuca and Natunuku; M39, M42, M43 and M45 shared by Natunuku and Tonga; DE2, M13 and M23 shared by Yanuca and Tonga; M36, M38 and M40 shared by Natunuku and Site 13 in New Caledonia; M27 which is one specific example of a relationship between Yanuca and Watom Island; and M9, which is shared by Yanuca and the Ile des Pins. The greater the number of shared innovations the greater the degree of relationship. Unfortunately, there are not enough data available at present to allow a full-scale study of shared innovations.

5. Groups of sites which are geographically related, for example Yanuca, Natunuku and Tonga, share elements and motifs which do not occur anywhere else. Such elements and motifs provide evidence of closer relationships within the group than with any site outside the group. For example, in Table 20 it will be seen that VB, DE7, DE8, M17, M18, M26, M39, M42, M43 and M45 are exclusive to Fiji and Tonga. These are the specific items which link the sites together and indicate a close relationship among the decorative systems.

6. On the available evidence of zoning procedures, design elements and motifs used, rules applied and techniques employed to give form to elements and motifs, the decorative systems applied to Lapita pottery discovered in Melanesia and Western Polynesia all appear to belong to a single Lapita style.


Because the decorative systems applied to Lapita pottery have not been analysed in terms of the rules used for creating patterns except for Yanuca and Natunuku, it is not possible to employ structural evidence to demonstrate that all the Lapita decorative systems belong to one ancestral system. Such evidence was used to show quite convincingly that the decorative - 65 systems of Yanuca, Natunuku and Tonga belong to a single style of decoration. If there were more descriptions of design structure available for comparison, it is more than likely that most, if not all, of the decorative systems would be found to be structurally similar and therefore all members of a single decorative style.

The distributional evidence in Table 20 can be used, however, to demonstrate the interrelationships which link all of the decorative systems which occur at Tonga, Natunuku and Yanuca in Fiji, Malo Island in the New Hebrides, Site 13 and the Ile des Pins in New Caledonia and Watom Island in New Britain. The inventory of design elements and motifs for Yanuca together with additional motifs from Natunuku forms the base from which distributions are traced. Every areal identification given in the table represents one area even though, as in the case of Yanuca and Tonga, there is actually more than one archaeological site involved. The geographical spread of an element or motif is the important thing. The wider the geographical spread of an element or motif, the more likely it is that the element or motif belongs to an older inventory which is ancestral to all decorative systems. A distribution figure of one area means that the motifs are unique and not shared. A distribution restricted to two areas means that the elements and motifs are shared innovations usually between two contiguous areas but sometimes between two non-contiguous areas.

The distributions in Table 20 (excluding the unique Tongan items M46-M70) cover 61 specific items of the inventory. The results may be summarised as follows.

  • (a) Distributed over six or more areas, seven items: GZ2, DE1, DE3, M1, M2, M30, and possibly M33. Percentage of total: 11 percent.
  • (b) Distributed over five areas, six items: GZ1, DE6, M8, M19, M24, M28:9 percent.
  • (c) Distributed over four areas, eight items: DE4, DE5, M11, M12, M14, M16, M18, M29: 13 percent.
  • (d) Distributed over 3 areas, 15 items: VB, TB, GZ3, DE7, DE8, M5, M7, M15, M17, M21, M25, M26, M34, M35, M37:25 percent.
  • (e) Distributed over 2 areas, 18 items: N, GZ4, RZ3, DE2, M3, M6, M9, M10, M13, M23, M27, M36, M38, M39, M40, M42, M43, M45: 30 percent.
  • (f) Restricted to one area, seven items: M4, M20, M22, M31, and M32 to Yanuca; M41 and M44 to Natunuku: 11 percent.

One way of interpreting these facts is as follows. Of a total inventory of design elements and motifs, 11 percent result from local development and are unique to the area; 30 percent are shared with another area because the communities concerned are either in contact with each other or spring from common antecedents; 25 percent are shared by three areas because the communities concerned all share a common ancestral group; 13 percent are shared with four areas and result from the fact that the four communities share a common ancestral group; 9 percent result from the fact that the five communities concerned share a common ancestral group; and 11 percent result from the fact that six communities share common origins. Thus there is an ever-widening sphere of interrelationships which eventually links all the Lapita sites together. Only 11 percent of items in the inventory are - 66 locally invented if the comparisons are based on the inventory from Yanuca and Natunuku. If the unique Tongan motifs are included, however, the figure increases to 37 percent of a total inventory of 86 items. Inspection of Santa Cruz elements and motifs now being analysed suggests there is every indication that this area, too, will produce as high a percentage of unique features as Tonga.

The greater part of the inventory results from diffusion. If the items distributed over five or more areas are added together the percentage of items carried over from the ancestral system is 20 percent. Any item distributed over four areas is ancestral to both Fiji and Tonga. There is thus an additional 13 percent of items which predate Fiji and Tonga. All told, 33 percent of the inventory items originate west of Fiji and Tonga. This percentage is likely to be greatly increased when more decorated sherds are found. The implication is that some sites to the west must predate Fiji and Tonga if the present line of thinking is to be maintained. Carbon dates available so far do not offer strong support for a west-to-east migration. 13

Another way of indicating relationships is to compare the percentages of elements and motifs shared between pairs of areas. There is insufficient evidence to permit a full test of the method, but in order to illustrate how it works Yanuca is paired with each of the other areas. Percentages of shared elements and motifs based on a total of 49 items considered are given in Table 21.

Percentages of Shared Elements and Motifs between Paired Areas
Areas Certain No. correspondences Percentage Possible No. correspondences Percentage
Yanuca — Natunuku 29 59 30 61
Yanuca — Tonga 27 57 32 65
Yanuca — Site 13 13 27 18 37
Yanuca — Ile des Pins 12 25 15 30
Yanuca — Watom 11 22 17 35
Yanuca — Malo 6 12 7 4

Percentages are significant only when there is a large sample of decorated sherds to compare. The first column of percentages in Table 21 was calculated on correspondences which were judged to be certain. Here the percentage figures of 59 and 57 respectively, for Yanuca paired with Natunuku and Tonga, support a claim of close genetic relationships among the three areas. According to the figures, other areas are not as close genetically to Yanuca. However, the lower percentages and especially the figure for Malo reflect a low yield in decorated material rather than anything else, so that one needs to be cautioned against placing too much faith in the figures.

In view of the limitations of the data, all that has been demonstrated at this point are, first, confirmation of close genetic relationships between Fiji and Tonga and, second, indications that genetic distance increases to the west.

- 67

Percentages in the second column indicate the upper limits of possibility when uncertain correspondences are added to the list of items shared. Thus, the possibility exists that Yanuca has closer genetic ties with Tonga than with neighbouring Natunuku, which is a surprising result. However, such a result is not supported by the evidence of shared innovations which I judge to be a more reliable index of genetic relationship than ordinary pairing. For the moment, therefore, the possibility must remain in doubt.


The method of analysing decoration advocated in this paper calls for a more rigorous way of establishing analytical units such as design elements and design motifs so that controlled comparisons can be made. In addition, an attempt has been made to analyse not only the items which make up an inventory of design elements and motifs, but also the structural system by which elements and motifs are combined into patterns. Statements about whether the decorative systems of Fiji and Tonga belong to the same style rest on structural evidence rather than on a listing of inventory items. The structural system is less susceptible to change than the inventory. Moreover, as shown in this paper, each area is bound to yield inventory items which are unique to that area. Owing to a lack of comparable material from other areas, structural comparisons were limited to Fiji and Tonga. However, the usefulness of the idea has, I believe, been demonstrated.

Controlled comparisons can be applied to inventory items if such items are defined in a recognised way. Suggestions for defining design elements and design motifs are provided in this paper. While it is difficult to rule out completely the possibilities of arbitrariness in defining units, the principles given here do go some way towards applying a more rigorous approach. Such an approach leads the student to many problems which have largely been ignored by past scholars and certainly by those dealing with Lapita decorations, but it will be idle to pretend that most of the problems are solved here. However, an advance is made by pointing them out and then trying to offer various solutions. Perhaps the most powerful justification for the method of approach demonstrated here is that while it provides opportunities for controlled comparisons it does also force us to consider some new problems.

How much reliance one is able to place upon the findings of comparisons depends ultimately on the integrity of the units compared. When the claim is made that design motif M16 appears in Yanuca, Natunuka, Tonga and Malo Island, what do we mean? Does the motif appear in exactly the same form in all four areas or is it that occurrences of a thing classed as M16 fall within its presumed alloform range? In this paper, the latter course is held to be more reasonable as it is comparatively rare to find motifs taking exactly the same shape in different areas. Thus, a motif is defined as representing a class and each specific representation of it is regarded as an alloform of that motif.

Negative information about a particular motif may be due simply to the fact that a suitable site has not yet been found and excavated. Contingencies of this type make it necessary to qualify all distributions listed and interpre- - 68 tations made as tentative and subject to change when more data are unearthed. The comparisons made in this study are given as examples of what can be done given an adequate sample of decorated sherds and given a systematic method of analysing decorative art.

- 69
1. Nubbins

A nubbin is roughly a cone-shaped feature which is repeated in series, always along a transverse line, on the external surface of a pot. The diameter of the base of a nubbin may vary from 3 mm to 6 mm and the height may range from 2 mm to 5 mm. Alloforms of the nubbin element occur as the result of applying various design processes, such as squaring, compressing and rounding. Squaring is applied when the nubbins are constructed in a narrow band and when the distance between each one is very narrow. The effect of squaring is to flatten the circumference of the nubbin, producing either a semicircular base or a four-sided base. In several cases, only the underside of the nubbins have been “squared off”, probably with a special tool.

If the nubbin element is identified as N1 and its ideal cone-shape as N1.1 (Figure 2.1), its squared alloform can be identified as N1.2 (Figure 2.2), or in terms of process it can be stated as N1 → sq.

Nubbins are also subject to compression: that is, to a general flattening top and bottom in response to environmental and aesthetic requirements. The effect of compression is to produce nubbins whose bases are oval or elliptic, the long axis lying in a transverse direction. Compression may be associated with rounding or squaring, the first producing an oval base and the second producing a roughly rectangular base. The resulting alloforms may be stated in two ways, thus: N1.3, or N1 → com/sq (Figure 2.3); N1.4, or N1 → com/rd (Figure 2.4).

Finally, the nubbin element may be subjected to rounding, which results in a hemispherical shape rather than a cone shape. Thus, N1 → rd, or the nubbin element subjected to rounding, produces the alloform N1.5.

The most common way of constructing a pattern with N1 is in accordance with the following formula: P=N1 → DisR/E-W.

The result of applying the formula may be stated as:

  • # N1 # N1 # N1 #

in which the spaces marked # are equidistant and may range from 1 mm to 10 mm. By discontinuous repetition is meant the separation of elements by space as indicated above.

Usually a pot has only one such pattern placed at the upper part of the vessel, but in the corpus there is one fragment of a pot which shows evidence of having two such patterns, one on the upper part of the pot's body and the other on the lower part of the body, 5.5 cm below the other.

Element N1 also occurs in association with the next element to be described and with impressed and incised elements and motifs. Class 2 elements usually occur above and below a nubbin pattern and not in between each nubbin. Consequently a nubbin pattern is always prominent.

2. Vertical Bars

A vertical bar is probably modelled in the first place as a long coil which is then cut up into equal lengths, each of which is attached to the external surface of - 70 the pot, usually just below or as part of a nubbin pattern. Each length is attached in vertical position and a pattern is produced by repeating the vertical bars within a transverse band. In one pot, the vertical bar pattern occurs immediately below the rim.

The dimensions of a typical vertical bar are as follows: length 2.6 cm, width 0.5 cm, and highest point 0.5 cm. The cross-section is usually semicircular but, as in the case of nubbins, the vertical bar is subject to squaring, the effect of which is to produce a triangular cross-section. The alloforms of the VB element are VB2.1, that is, VB2 → rd (Figure 2.5); VB2.2, that is, VB2 → sq (Figure 2.6). A standard pattern arrangement for the VB2 element is exactly the same as for N1, that is, P=VB2 → DisR/E-W. The result is # VB2 # VB2 # except that now the space is up to 2.0 cm wide. There is also some inconclusive evidence to show that VB2 elements are sometimes paired by varying the width of the spaces, thus:

  • # VB2 # VB2 # # VB2 # VB2 #

As mentioned earlier, elements N1 and VB2 may be combined into a single pattern complex, thus:

  • # N1 # {N1/VB2} # N1 #
  • or: # N1 # N1 # VB2 # N1 #

In the case of the second pattern, elements N1 and VB2 can coalesce to form a motif. When this happens the VB2 element undergoes a change at the point where VB2 meets N1 thus producing an entirely new thing which is no longer merely VB2 + N1 (Figure 2.7). I have classed this as a motif, in this instance a three-dimensional alloform of M12.

3. Transverse Bars

Two forms of transverse bars appear on Yanuca pottery. Alloform TB3.1 can be described as a structural bar, while the other can be described as an applied bar. The structural bar is formed originally as a basic part of the pot. The everted rim of the pot is levelled off and squared so that the outer surface is flat. Then an additional rim is added to the pot and the original rim becomes a decorative transverse bar (Figure 2.8).

In the construction of the second kind, TB3.2, the bar is first formed as a coil which is then attached either on the inside of the rim at the point where the angle of the wall changes or on the outside of vessel below the rim (Figure 2.9).

Alloform TB3.1 transverse bars are squared, producing a roughly rectangular cross-section. The bars of TB3.2 tend to be more rounded and thus semicircular in cross-section, but they are subject to squaring, which produces a roughly triangular cross-section, TB3.3 (Figure 2.10). Thus, there are three alloforms of the TB3 element, TB3.1, TB3.2 and TB3.3.

The standard arrangement for producing a pattern is the following: P=TB3 → ConR/E-W. That is, the bar is subjected to continuous repetition around the perimeter of the vessel on an east-west axis. Transverse bars are always continuous whether they are placed inside a vessel or outside.

A feature of the transverse bar is that it is subject to the application of Class 2 and 3 elements. On present evidence, this applies only to transverse bars on the outside of the pot.

- 71

Zone markers constitute a special kind of design element. Their function is not only to enter into pattern arrangements, but also to set the spatial limits within which patterns are confined. But it will be seen later that marked zones often constitute the only “decoration” applied to the pot.

An examination of many decorated sherds reveals that the first step taken by a potter is to divide the design field into transverse bands. This is followed by a secondary division of the transverse bands into smaller areas by the use of vertical zone markers, which has the effect of producing smaller four-sided areas for decoration. The resulting four-sided areas may be (a) rectangular with the long axis lying either horizontally or vertically or (b) square. In all such cases, the vertical boundary marker is joined with the upper and lower zone markers of the transverse bands. This appears to be the usual practice of Yanuca potters when the transverse band is relatively narrow, say from 0.5 cm to 1.5 cm wide.

Where the band is particularly wide, one method of breaking down the zone area is to place concentrically arranged rectangles or squares within it. On present evidence it is rare.

Tertiary division of the smaller four-sided zones may also occur. For this purpose diagonal zone markers are used. The effect of tertiary division is to break down rectangular and square shapes into two or more triangles. By the time tertiary zoning occurs the areas are so small that no further decoration may be thought necessary.

There is one exception to the zoning procedures outlined above. The exception occurs in a “pot” with a diameter of about 18.7 cm, which can be interpreted as either a flat-bottomed pot or a lid. If it is the former it will be the only case in the corpus where the bottom surface of a vessel has been decorated, but it is probably a lid, on which the application of decoration is a little more practical. In this example the decorator had a large, flat, circular area to zone and no longer a curved surface. Primary zoning was accomplished by drawing in a diameter line which halved the design field, followed by another diameter line which then divided the design field into quarters. From this point, each quarter was apparently further divided, or decorated, by drawing in a series of descending chevrons more or less equidistant from each other. There are, however, too many missing pieces in the reconstructed lid or pot to be sure of the steps taken in secondary zoning.

The zoning of the design field appears, on the evidence available, to serve two main functions. Firstly, it reduces the design field into horizontal bands which present the kind of area most preferred by decorators for setting out a pattern. Secondly, zoning adds a decorative effect. It is important to note that although there are numerous sherds which show the combination of zoning plus decoration of the zones, it is not obligatory upon the decorator to follow this sequence. The choice as to whether patterns are added to the zones of not was apparently left to the individual, as no particular rule can be inferred. Some individual decorators elected not to add patterns while a large number did. In a few cases, also, the decorator decided not to zone the design field at all. There is one particular pattern for which a clearly defined zone was not felt to be necessary at all times. For the majority of patterns, however, zoning was necessary.

The zone markers used by Yanuca decorators may be classified into two main categories; namely, general zone markers and restricted zone markers.

General Zone Markers

General zone markers (Figure 2.11) are all-purpose markers which can be used - 72 to divide the design field at the primary, secondary and tertiary stages of division. They may be used horizontally, vertically or diagonally. They consist of either single lines or double lines, the latter being the more common. In order of preference, the general zone markers may be distingished as follow:

GZ1, the double dentate marker; GZ2, the single dentate marker; GZ3, the single incised marker; GZ4, the single rubbed marker.

The last two occur very rarely in the corpus.

Restricted Zone Markers

Restricted zone markers (Figure 2.12) are more ornate than general markers, are wider, and occur only as primary dividers, that is, as horizontal lines. In no case is a restricted marker used as a vertical divider of space. Three main types occur.

RZ1 is a multiple dentate marker which consists of three of more lines of dentate stamping placed closely together. The lines appear to be horizontal, but in most cases they are slightly oblique.

RZ2 is a multiple oblique dentate marker where the three or more lines of dentate stamping are placed at a sharp oblique angle, thus each line is short. Characteristically, both RZ1 and RZ2 appear near the upper part of vessel. RZ2 often occurs as the first zone marker at the uppermost area of a pot.

RZ3 is a multiple-diamond pattern produced by impressing. On one sherd the impressing is roughly done and the diamonds are indistinct but on another they are more clearly defined. One sherd with particularly clear and clean impressions is an important find, since it demonstrates that Yanuca potters were using much the same kind of object to impress the multiple-diamond pattern as the people of New Caledonia who made the Lapita pots at Site 13. On present evidence, RZ3 is the only restricted zone marker which may be used in a circular direction.

An exceptional use of RZ3, to decorate the edge of the lip, is noted on one sherd. RZ3 is placed in an oblique position and repeated with equal spacing right around the edge. This example provides strong grounds for regarding RZ3 and the rest of the zone markers as design elements since they all present a decorative visual effect.

Although three restricted markers have been distinguished in the analysis, it is possible that there are really only two, RZ1 and RZ2 being in fact the same. The only thing which really distinguishes them is that the lines in RZ1 tend to be much longer and the angle of obliqueness flatter. Moreover, although these zone markers tend to be restricted in use at Yanuca it does not follow that they are similarly restricted in other Lapita-bearing sites.

It does not follow that a design zone of any type must necessarily be decorated. As already pointed out, many zones are marked out and left plain by the decorators. There is apparently no felt compulsion to construct patterns in design zones.

Apparently any combination of zone types may appear on a pot but there are instances where only one type occurs. For example, in several Yanuca pots the only zone decorated in Type G. In the neighbouring site at the Sigatoka Dunes only Type E zones appear on a flat-bottomed dish with shallow sides. From Yanuca, zones of Types E, A1, C and A1 occur in that order from top to bottom in a flat-bottomed vessel with an everted rim. In another partially reconstructed vessel from Yanuca the order is as follows: Types F, G, C, and A2 (Figure 2.16 and Plate 1).

- 73

Design element one is a single crescent which is placed in a vertical position within a design zone. When used in a lateral position, it is no longer a discrete element but functions as part of a design motif. DE1 can be represented in two ways: either as a single line or as a double line, hence the alloforms DE1.1 (single) and DE1.2 (double).

Alloform DE1.1 is used in various ways: for example, DE1.1 → DisR/E-W in a Type A zone, where the distance between each element is 2 mm (Figure 2.17a). Alternatively, it can be arranged in widely separated pairs in which the second element is a mirror image of the first, thus each pair equals DE1.1 # MR/DE1.1, which means the element plus the mirror image of the same element. Alternatively this may be written as DE1.1 # 1.1ED. From this point a pattern is created by the familiar rule DE1.1 # MR/DE1.1 → DisR/E-W. The minimum distance between two units of a pair is 2 mm and the distance between each pair is approximately 6.0cm (Figure 2.17b).

In triangular spaces of Type D, DE1.1 appears alone, usually inside the triangle and against one of its sides (Figure 2.17c).

Alloform DE1.2 usually occurs in a Type A zone and a pattern is formed by the rule P=DE1.2 → DisR/E-W. The spaces between each element are about 1.5cm (Figure 2.17d).


Design element two may be visualised as two crescents placed one above the other. The element occurs only in vertical position. It may be represented by a single line, as in one example where a DE2 pattern occurs on a TB3.1 structural bar. In the case of alloform DE2.1 a pattern is formed as follows:

  • (DE2.1 → DisR × 5/E-2 + MR/E-W) → DisR/E-W,
  • or (DE2.1 → DisR × N/alt with MR/E-W) → DisR/E-W,

where N stands for a number varying from 5 to 10.

The first part of the rule indicates that the element is repeated in varying numbers, to produce groups of five to ten elements, and the second part indicates that each alternate group is reversed to form a mirror image of the previous group (Figure 2.18a). The element appears to be shell-impressed.

Alloform DE2.2 is represented by double lines. The only sherd containing this element is unfortunately not large enough to indicate what type of zone it is used in. The element is placed about 10 degrees to the left to vertical position. A pattern is produced by the simple rule P=DE2.2 → DisR/E-W. The space between each unit varies from 1.0cm to 1.6cm (Figure 2.18b).

It is possible that there is a third alloform of DE2; namely DE2.3, in which three lines are used to represent the element. There are three sherd fragments exhibiting a part of the suggested alloform. In none of them is the element complete. Neither is confirmation possible from the Natunuku site, but the alloform definitely occurs in Tonga. It is highly likely, therefore, that the incomplete representation on the Yanuca fragments does in fact represent the same alloform as appears in Tonga. A pattern is produced as follows P=DE2.3 → DisR/E-W (Figure 2.18c).


This element is a small circle which appears to be stamped into the clay with either a hollow or a solid cylindrical stamp. A pattern composed of DE3 is - 74 usually placed in a Type A2 zone and is nearly always arranged as a discontinuous series: P=DE3 → DisR/MULTI (Figure 2.19). Unlike a nubbin pattern, which is always arranged as a horizontal series, a DE3 pattern may be arranged in any direction following either a straight or a curved line, hence the term multidirectional.

The element is often used as a part of a motif as in M20 and M33, or association with another pattern; for example, with an M9.2 pattern.


This element is in the form of an oval in which the ends tend to be sharp pointed. The element is subject to rounding, especially when used as part of a motif and when the element is placed with the long axis in a vertical position. Nothing very definite can be said about this element since there is no sherd which shows a complete pattern sequence. The element does, however, occur at Natunuku in vertical position, and at Site 13 in New Caledonia it is a dominant element in many motifs. From the available evidence, it does not appear to be of much importance at Yanuca.

As inferred from two fragments of a single vessel, the element is used in horizontal position in the following manner: P=DE4.1 → DisR/E-W (Figure 2.20).


Design element five is a straight line in vertical position. The element may be used by itself; for example, on either side of a vertical bar, as in two examples where the element appears to be evenly spaced around the vessel. Here the rule followed is P=DE5 → DisR/E-W (Figure 2.21a), in which the space between each pair of DE5 is about 1.0cm.

In another example, the element is repeated in the same way, but the vertical lines are close together (about 2 mm apart). With so simple an element, the chances of producing a pattern which is aesthetically unstimulating are high. Variety is introduced by grouping sequences of DE5 and widening the space between each group, thus: DE5 # DE5 # DE5 # & 3 DE5 # DE5 # etc. (Figure 2.21b).


This element is an oblique line. Its most common alloform is represented by a double line (Figure 2.22a). The zones most favoured for a DE6.1 pattern are Type A1 and Type E. Here, the element is usually drawn completely across the zone from the lower boundary to the upper boundary. When used in a Type A2 zone, the element does not touch the upper boundary. In most zones, the pattern is arranged as follows: P=DE6.1 → DisR/E-W (Figure 2.22b). The space between each DE6.1 element is always greater than the thickness of the element itself.

Alloform DE6.2 is represented by three lines and appears to be used as a minor decoration in the corner of a divided zone. In one case, the element occurs in a Type C zone and the pattern is arranged in a vertical series as follows:

  • P=DE6 → DisR/N-S.

This element takes the shape of a chevron. There are at least two sherds, representing different pots, which show a chevron as a discrete element (DE7.1). Here, the element is not in contact with any other element. In both cases, a pattern is formed as follows: P=DE7.1 → DisR/N-S/DECR (Figure 2.23). That is, the chevrons are repeated one below the other and decrease in size (DECR) with each - 75 repetition. In other words, a diminishing series is drawn. In the case of one flat-bottomed dish or lid, the fragments indicate that four sets of diminishing chevrons comprise the decoration. It is postulated that when decorating a flat “lid” the decorator turns the work around so that the area to be decorated is directly in front.

Alloform DE7.2 is the same as DE7.1 except for the addition of a feature to the extremities. This additional feature may be described as “fingers”. In the only example in the Yanuca sample, each of the double lines representing the chevron terminates in two fingers, thus providing four altogether. In this example alloform DE7.2 is above a DE7.1 chevron.


Also functioning as a discrete element is DE8, a triangular shape, which is best represented by the notching which occurs on vessel rims. To produce a pattern the DE8 element is arranged in three different ways.

In the first pattern the triangular notches are widely spaced. The pattern rule is

  • P=DE8.1 → DisR/E-W.

In the second pattern the notches are made to touch each other. The result of this arrangement of the element produces a positive and continuous series of triangles as well as a negative series. To simplify procedure, alloform DE8.2 is described as two triangles placed side by side. The pattern rule is

  • P=DE8.2 → ConR/E-W.

In the third pattern, in which an opposing series of widely spaced notches is shifted half a place to the right, a positive series of continuous chevrons is produced, flanked on either side by a series of notches. Again, to simplify the rule alloform DE8.3 is defined as shown in Figure 2.24. The pattern rule is

  • P=DE8.3 → ConR/E-W.

It is a characteristic of this element that pattern arrangements often produce highly complex effects and, despite their small size, are probably better classed as motifs. Such arrangements nearly always have both a positive and a negative aspect.

- 76
Motif 1

There are several motifs which appear to be derived from element DE1. Elaboration of DE1 occurs when it is placed in a horizontal position with the extremities of the crescent either uppermost or facing downwards. Either way the crescent element presents the decorator with opportunities to link other crescents to it in various combinations. The simplest arrangement is a single linking in an east-west direction. This arrangement is typical of M1. Motif 1 is formed first by reposing DE1, that is, by lying it on its side, and then by repeating the reposed element once so that the basic motif contains at least two DE1 elements (Figure 2.25a).

These processes are summarised in the formula

  • M1 pattern=DE1 (rep) → ConR/E-W (Figure 2.25b).
  • A simpler statement of the rule is P=M1 → ConR/E-W.

The occurrence of this pattern at Yanuca is restricted to one definite and one indeterminate case. It does, however, occur at Natunuku and at other Lapita-bearing sites.

Motif 2

Motif 2 is a combination of two chains of the M1 pattern. The first chain is produced as for M1, that is by the formula P=M1 → ConR/E-W. The next is produced by the same rule but it is placed against the previous chain so that the points of each crescent touch the highest point (or midway point) of the arc of the first crescent. The result is to produce a mesh of crescents. To simplify the rule the process of meshing crescents in this way will be represented by the form (½DM) which stands for a half-drop mesh arrangement. Though the most common arrangement at Yanuca is to mesh two rows of crescents, one sherd bears witness to the fact that three rows were also meshed. The number of rows will be indicated thus: ½DM × 2 for two rows, and ½DM × 3 for three rows.

Thus the most common pattern arrangement for M2 is

  • P=DE1 (rep) → ½DM × 2/ConR/E-W (Figure 2.26).

However, if the alloform M2.1 is isolated as shown in Figure 2.26a, the rule for constructing a pattern is P=M2.1 → ConR/E-W. Similarly, if the alloform M2.2 is isolated as shown in Figure 2.26b, the rule becomes P=M2.2→ ConR/E-W. Such a pattern, which occurs in at least five pots, may be distinguished from a two-row mesh by setting up two alloforms of M2. Alloform M2.1 is used only in a two-row mesh while M2.2 is used in a three-row mesh.

Motif 3

In another five pots, a complex chain is produced not by dropping one row below the others as in the previous case but by superimposing a second row over the first in such a way that the points of the crescents of the second row are placed in a position midway between the points of the first set of crescents (Figure 2.27b). This process of superimposing one set of elements over another will be indicated by the form (SUP). Thus, an M3 pattern is constructed from DE1 in the following way: P=DE1 (rep) → SUP × 2/ConR/E-W (Figure 2.27). In the formula, SUP × 2 indicates that two chains of DE1 elements are superimposed in the same horizontal zone.

In one pot, this pattern effect has been varied by changing the rule for the first - 77 set of crescents from DE1 (rep) → ConR/E-W, to DE1 (rep) → DisR/E-W. When a unit of M3 is isolated as in Figure 2.27a the rule is simplified to

  • P=M3 → ConR/E-W.

The M4 motif represents the application of one further design-making process to an M2.1 pattern; namely, the process of rounding. The sharp points formed by joining two crescents side by the side are rounded to produce a wavy line. What happens, therefore, is that the DE1 element is first reposed and then the points are tailed and rounded, hence DE1 (rep/rd), which produces the basic unit of M4. An M4 pattern is produced by the formula

  • P=DE1 (rep/rd) → MR × 2/ConR/E-W.

In the formula, MR × 2 indicates that two rows are opposed. However, if a unit of M4 is isolated as shown in Figure 2.28a the rule becomes P=M4 → ConR/E-W (Figure 2.28b). This pattern occurs on one sherd.

Motif 5

A pattern composed of this motif is formed by rounding the crescents in an M3 pattern. A pattern may thus be represented as

  • M5 pattern=DE1 (rep/rd) → (SUP × 2),

or more economically as P=M5 → ConR/E-W (Figure 2.29b), given the unit illustrated in Figure 2.29a. This pattern occurs on three sherds.

Motif 6

A different pattern effect is produced by opposing two rows of continuous crescents. An opposition which is complete or full has been referred to as a mirror image (MR). Therefore, (MR × 2) can be taken to mean “oppose two rows fully so that each is a direct opposite of the other”. Another way of opposing two rows of elements is to shift the second row half a place to the right. Instead of a full mirror image the result is a half-drop mirror image which will be represented as (½MR). As the opposing rows of crescents are not touching and are actually separated by space, this fact must be included in the formula. Thus, (½MR #) means a mirror-image effect in which the opposing row of elements is shifted half a place to the right and is separated from the other row by a regular interval of space. The pattern is formed from a DE1 element and may be represented thus:

  • P=DE1 (rep) → ½MR # X2/ConR/E-W (Figure 2.30b).

The more economical solution, however, is to regard a unit of M6 as that illustrated in Figure 2.30a. The rule now becomes P=M6 → ConR/E-W.

Motif 7

This is another variation of the rounded DE1 element. If development of the pattern from the basic DE1 elements is followed, the rule for producing a pattern is as follows: M7 pattern=(DE1(rep/rd) → ConR/E-W) → DisR/N-S (Figure 2.31). That is, a chain of rounded crescents is constructed in an east-west direction and then the chain is repeated once in a north-south direction. A simpler solution may be to regard the main process as doubling the chain, since, in fact, doubling is quite common in Yanuca decoration. Given such an interpretation, the rule can be stated as M7 pattern=DE1 (rep/rd/db) → ConR/E-W.

An even simpler rule is to take one unit of DE1 (rep/rd/db) as illustrated in Figure 31a and call it M7. The rule is then P=M7 → ConR/E-W.

Motif 8

Several comple× patterns are built around a U-shaped motif which is elaborated - 78 in various ways. The simplest form may be derived from a simple M1 pattern on to which is grafted a series of vertical lines, each one being a continuation of the point where two crescents meet, as in Figure 2.32. The motif may thus be seen as a deepened crescent which in time has lost its association with the crescent element.

Alloform M8.1 is the basic motif without elaborations and is represented by a single line (Figure 2.32a). A pattern is formed thus: P=M8.1 → ConR/E-W/CLS (Figure 2.32b). When the unit is repeated, the junction of each U is represented by a single fused line, hence the letters CLS standing for coalescence. This pattern is constructed in a Type E zone on a flat dish with shallow sides. This form of the motif occurs only in the Sigatoka Dunes site and is not represented in the Yanuca sample.

Alloform M8.2 is an elaborated form of M8.1 in which a single vertical line is drawn inside each U. This form of the motif occurs at Natunuku and also at Site 13, in New Caledonia. The Yanuca alloform contains one further elaboration. Three short straight lines project from the bottom of the U (Figure 2.33). A pattern is formed as follows: P=M8.2 → ConR/E-W/CLS. But this reconstruction is based on only one sherd which does not show a complete sequence of the pattern. As in the Sigatoka Dunes example, coalescence occurs at the junction of each repetition.

Alloform M8.3 is the basic U-shaped motif with two or more vertical lines drawn inside it and spaced equally from the sides of the U. A pattern is formed by the same rule as applies to M8.2, thus: P=M8.3 → ConR/E-W/CLS (Figure 2.34). It is not clear whether there is any significance in the number of internal vertical lines within the basic motif. On one sherd two vertical lines appear. There are three vertical lines on a second sherd in which all lines are the same height. There are also three vertical lines in a third example, but the central vertical is longer than the other two. In another example the decorator appears to have built up a pattern by first drawing in a single chain of connected crescents and then afterwards drawing in the vertical lines. The main vertical lines by which a U is formed do not always occur at the junction of crescents. This badly executed example of M8.3 seems to indicate that successful decorators did apparently visualise M8 as the unit isolated here and not as two sets of basic elements combined together.

Motif 9

The visual form of M8 changes slightly when it is no longer represented by a single line. When two lines are employed to depict the basic unit, the result is that the U shape is given width and consequently solidity. There are, however, no sherds in the sample which show the basic solid U without embellishments.

Alloform M9.1 represents the basic solid motif with one internal vertical line. It is not clear from the fragment how a pattern is built up, but it appears to be as follows: P=M9.1 → ConR/E-W (Figure 2.35a). It is also not clear whether the arms of two juxtaposed basic motifs coalesce or remain separate.

Alloform M9.2 is the basic solid motif containing one centrally placed vertical line plus a horizontal line inside the U and across the points where the arc of the U shape changes to vertical. A pattern is formed thus: P=M9.2 → DisR/E-W (Figure 2.35b). Basic units are placed side by side and no fusion of the arms occurs. To complete the pattern, a DE3 element is placed in the V-shaped area below the point where the two arms of the U motif meet. The full pattern is therefore:

  • P=M9.2 → DisR/E-W/DE3 → DisR/E-W

Alloform M9.3 consists of the basic unit with three (or more) vertical lines inside - 79 it. It is not clear how a pattern is constructed. From what can be observed it appears to be P=M9.3 → ConR/E-W/CLS (Figure 2.35c). The juxtaposed arms of the U coalesce to form a single arm, thus effecting continuous repetitions.

Motif 10

M10 appears to be derived from a pattern composed of alloform M9.1 in which the open end of each solid U is closed to form a mirror image of the lower end. The same result could have been produced by opposing two chains of M9.1. Whatever the derivation, the result is a new motif shaped like an oval flattened at each end and containing within it one central vertical stroke depicted as a doubled dentate-stamped line. A variant of M10.1 also occurs at Natunuku but there the motif is represented by single lines. A pattern is formed thus:

  • P=M10 → ConR/E-W/CLS (Figure 2.36)

in a type A zone marked top and bottom by RZ3 zone markers. At the junction where two M10.1 units meet coalescence occurs.

Motif 11

Several motifs are elaborations of DE4 in which the element is in vertical position. In one motif, the element is placed in horizontal position and then combined with something else.

Isolation of M11 is difficult owing to the fact that it is often constructed in two distinct stages. The first part of the motif is added during the zoning stage. The vertical zone divider becomes eventually the internal segment of the motif. The second part is added into each zone one after the other. When this second part is drawn into the first zone, the motif is incomplete, but, after it is drawn into the second zone, one finished motif emerges. The motif is thus built up progressively as is also the case when a pattern is constructed. Following this particular sequence of construction, the motif may be isolated as M11.1 which represents something like an N stretched horizontally. The steps in the construction of a pattern are as follows: First stage: Construct a Type C zone; Second stage: Place M11.1 into each design zone; Result: P=M11.1 → DisR/E-W. The result obtained by following these steps is shown in Figure 2.37. Although each repetition is actually separated from the next, the result is viewed by an observer as a continuous pattern.

It is also possible that another series of progressive steps was taken to attain the same results. The motif is constructed in three stages. The first stage is to draw a Type C zone using either single or double zone markers. Next, the element DE4 in its single line form is drawn using the vertical zone marker as its centre. Lastly the diagonal line is drawn. The alloform M11.2 thus takes the shape shown in Figure 2.38a in which the two parts are not actually touching. The resulting pattern is formed by the following rule: P=M11.2 → DisR/E-W/zone by zone (Figure 2.38b).

Motif 12

This motif is composed of a DE4 element perched above an inverted V. The result of combining these two parts resembles the shape of a highly schematised human form. There are several alloforms of the motif, but these do not depend on the shape of the DE4 element. The element appears to vary freely from an elliptical to a circular shape. Alloform M12.1 is a three-dimensional variation of the motif made by combining a nubbin element with a vertical bar element (VB2 + N1/N-S). On present evidence, alloform M12.1 is used only in combination with a nubbin pattern as follows: P=M12.1 → DisR/E-W/ # (Figure 2.39a). Between each M12.1 unit there is an unspecified number of equally spaced nubbins filling the space marked as # in the formula. The upper section of the alloform is aligned - 80 with the nubbins and the alloform is placed in vertical position in this arrangement.

Alloform M12.2 is a two-dimensional variant consisting of a DE4 element picked out in a single dentate line. A series of the M12.2 unit is placed in a Type B zone with the “head” of the motif uppermost. A pattern is formed with the simple rule P=M12.2 → DisR/E-W (Figure 2.39b). In two sherds, a wide space is left between the “head” of the motif and the upper zone boundary.

Motif 13

This motif is a development of M12. The basic motif is reversed and is given solid form by depicting it with two lines instead of one. A pattern produced with this motif is consequently more solid and more clearly articulated than the upright and thin M12 pattern. A characteristic of an M13 pattern is that it produces a negative pattern which introduces a new kind of motif to the inventory; namely, one that is bound to another and cannot occur without it. In one fragment there are two representations of the alloform M13.1 and the manner in which they are joined can be inferred. The upper zone boundary is not discernible, however, so that is possible to postulate two different patterns as follows:

  • P1=M13.1a → ConR/E-W (Figure 2.40a); and
  • P2=M13.1b → ConR/E-W (Figure 2.40b).

In the second possible pattern (P2), each alternate M13.1 may be seen as a reversed form of the other, its “legs” coalescing with those of the upright M13.1. Then each pair is subjected to continuous repetition in an east-west direction. What is in doubt, however, is not the rule for producing a pattern but rather the actual shape of the alloform. It is not known definitely whether the upper portion of the motif contains a knob. Partial confirmation of the alloform is provided by a sherd which represents a different pot. Here, however, confirmation is based on the negative pattern and it is still not clear what form the positive aspect of the motif took.

A variant of M13 is suggested in another sherd but so much of the pattern is missing that what follows can only be taken as probable. M13.2 is postulated as being basically the same as alloform M13.1 except that a further feature is added to the lower knob of the motif. The pattern is produced by the standard rule P=M13.2 → ConR/E-W (Figure 2.40c), in which the “legs” of the motif link to form a continuous series of chevrons.

Motif 14

As mentioned above, a pattern arranged with M13 produces a negative pattern consisting of two repeated series. The upper series comprises a row of triangles and the lower series comprises a row of a new shape which might be described as a spear point. The second shape is isolated as M14. A feature of M14, on the limited evidence available, is that it does not enter into a pattern arrangement on its own but only in association with M13. Two alloforms are identified.

Alloform M14.1 is shaped like a spear point with the tip uppermost. A pattern with this alloform follows the standard rule P=M14.1 → ConR/E-W (Figure 2.41a).

Alloform M14.2 is the same spear point shape with its mirror image attached. A pattern is formed in the same way as for M14.1, namely: P=M14.2 → DisR/E-W (Figure 2.41b). Something similar to the shape of M14.2 occurs in Watom Island.

Motif 15

This motif consists of several vertical lines varying in number from four to seven, grouped together. Groups are spaced out more or less equally. In one sherd fragment the groups contain five, six or seven vertical lines which stand on the - 81 bottom zone marker of a Type A2 zone. Each vertical line is dentate stamped and terminates in a single indented dot. Each complete line gives the impression of a lighted candle. The vertical lines do not touch the upper boundary and appear to fall short by about one-third the width of the zone. The dentations of the individual lines are aligned on the transverse axis and tend to be of equal length. This form of the motif is alloform M15.1 and a pattern is formed thus:

  • P=M15.1 → DisR/E-W (Figure 2.42a).

In alloform M15.2, the dentations match transversely, but in each line they may differ both in width and height. The upper extremity of the lines in one example are sharp pointed and the lower parts widen gradually from the point. In another sherd, the dentations are large and coarse. In both of these sherds, the fragments are too small to permit the discovery of a pattern-making rule. The patterns visible on the sherds are shown in Figure 2.42b.

Motif 16

Motif 16 is an ambiguous pattern in which what is visualised depends upon which part of it catches the attention of the observer. The pattern may be visualised positively or negatively. Seen positively, the motif consists of a continuous chevron which is delineated with a double dentate line. The negative aspect consists of alternating triangles, each alternate triangle being upside down and each separated from the other by a narrow space. With this type of motif, it is impossible to form a pattern by using only one unit of the basic chevron. The chevron must be repeated. Thus, a pattern is formed by the following rule: P = M16 → conR/E-W (Figure 2.43). In one pot where this pattern is used the zone decorated is Type A2, but in another sherd the zone is Type H. The pattern remains the same in character in either type of zone. There are three alloforms of this motif.

Alloform M16.1 is a true chevron represented by a double dentate line. When this alloform is applied in a pattern the apices of the negative triangles are some distance away from the zone marker (Figure 2.43a).

Alloform M16.2 is shaped like a chevron, but the outer point where the two arms meet is flat. When applied in a pattern sequence, the apices of the negative triangles are against the zone markers, as in Figure 2.43b.

Alloform M16.3 is also a chevron, but it is represented by a single dentate line. It is used in a Type C zone to produce an all-over pattern of continuous chevrons. The space in which a chevron is placed is divided vertically into two so that the equilateral triangle formed between the arms of the chevrons, as in alloforms M16.1 and M16.2, now becomes two right-angled triangles. Each unit of the alloform contains one positive chevron and four negative right-angled triangles. The rule for composing a pattern remains the same for all alloforms of M16.

Motif 17

This is a development of M16. The negative triangular pattern is emphasised by including DE1 in each triangle. The element is placed inside the triangle against one of the sides, usually the base, so that a semicircular shape is produced. When used in a border arrangement, the pattern appears to be produced by the same rule which applied to M16. A DE1 element must be added to the base of each triangle. From observation of one sherd, it appears the triangle unit is no longer a minor negative aspect of the pattern but rather the dominant unit. This being so, it is advisable to establish the triangle, plus the internal DE1 crescent, as M17. A pattern is formed by one of several rules.

  • (a) P = M17.1 (rep) → DisR/N-S (Figure 2.44a). There is no sherd which exhibits this pattern, but it is highly likely that such an arrangement was used.
- 82
  • However, because of horizontal zoning there could not be more than two or three repetitions in a north-south direction.
  • (b) P=M17.1 (rep) + MR/M17.1 → DisR/E-W. Again, there is no sherd which shows the full pattern, but one sherd suggests there is a strong possibility that a pattern was formed by this rule.
  • (c) P=M17.2 # alt with M17.2 rev # → DisR/E-W (Figure 2.44b). There are sherds which show a sequence of this pattern. Motif 17 with apex uppermost is placed against another unit of M17, but the second one is upside down and is separated from the first by a narrow space. Motif 17 and its reversed unit are then repeated discontinuously; that is, each unit is separated from the next by a narrow space.

Patterns (a) and (b) occur in a Type B zone, and pattern (c) occurs in a Type A1 zone. Motif 17 is unusual since it can apparently be arranged into both vertical and horizontal patterns. In the first two patterns, the base of the triangle is in vertical position and in the third it is in horizontal position. The first rule applies when the motif is so placed that its base is on a north-south axis. Although the third rule appears to be complicated, it is basically the same rule as that used for producing a pattern with M13.

Motif 18

This motif is based on the triangle in which the internal area is decorated with parallel oblique bars as in the universal basket-weave pattern. There are two alloforms. The first, M18.1, consists of a single triangle whose sides are stamped with a single line as in Figure 2.45a. The internal section of the triangle is filled in with several parallel bars varying in number from two to about five. This alloform is arranged into patterns as follows.

  • (a) P=M18.1 → ConR/E-W (Figure 2.45a). The resulting pattern consists of triangles whose bases fuse into the lower zone marker of a Type A1 zone. The spaces between the triangles become plain triangular-shaped areas. In one sherd, where the pattern has been incised in a Type H zone, rather than dentate stamped, the motif has been arranged in the standard basket-weave pattern (Figure 2.45c). A somewhat similar arrangement is found on another sherd. One way of stating a rule to produce such a pattern is as follows:
  • (b) P=M18.1 + M18.2 (rev/cls) → ConR/E-W/CLS (Figure 2.45c). What is repeated is the basic alloform M18.1, to the side of which is attached a reverse representation of the same alloform in such a way that they coalesce. The two sides of the juxtaposed triangles become one. This complex is then repeated continuously along an east-west axis so that all repetitions coalesce.

In alloform M18.2, the sides of the triangle are marked with a double line of dentate stamping but the internal section consists of single stamped lines. With the limited evidence available, the alloform may be used singly or in pairs in a north-south axis, that is: P=M18.2 → N-S (one unit only), or: P=M18.2 → ConR/N-S (two units or more). Since the alloform is placed into a Type B or C zone which is widened to take it, there cannot be much scope for more than two or three repetitions. The bases of the triangles are aligned on a north-south axis.

Motif 19

The basic motif consists of a triangle inside which lines radiate from the apex to the base. Units of the motif are repeated so that the points of the bases touch one another. There are two alloforms of the motif. The first, M19.1, appears frequently inside the lip of pots, in a Type F zone. In the sherds examined, the inner boundary of the zone is not defined by a transverse bar but rather by a stamped zone marker. - 83 However, it may be unmarked, being instead structurally defined by the limits of the flat rim.

Alloform M19.1 is depicted in most cases with the base of the triangle outermost, as in rim patterns, or uppermost, when applied to the sides of a pot. In one rim sherd, the triangles are placed apex outermost and in several side fragments it is difficult to calculate which way up the alloform has been stamped. The alloform is distinguished by the following characteristics: (1) it tends to be irregular in width and particularly in the number of internal radial lines included (Figure 2.46a-d); (2) it is also irregular in the way radial lines are stamped in. In many examples, the artist apparently missed the point of radiation and in others included more than one point of radiation in some of the triangles in a pattern (Fig 2.46c); (3) the apices of the triangle vary in shape from a flattened point to a sharp point and the sides of the triangle may even be extended beyond the point of the apex. In all cases, however, a pattern is constructed to the following rule:

  • P=M19.1 → ConR/E-W, or:P=M19.1 (rev) → ConR/E-W.

The second alloform of M19 is characterised by a tendency towards greater regularity. The shape of alloform M19.2 is less tied to the triangular form than M19.1. The shape resembles a web-shaped foot or hand with six or seven extensions (Figure 2.46c). The dentations tend to be matched across the alloform, and the point of radiation, that is, the apex of the triangle, is absent, being replaced by a flattened platform. A pattern is formed by the same rule as for alloform M19.1: P=M19.2 → ConR/E-W. The base of the triangular shape is uppermost.

Motif 20

This motif is again based on the triangle and can be viewed as a development of alloform M19.1. The triangle contains radial lines but at the apex DE1.2 and DE3 have been added (Figure 2.47). The apex is uppermost. Motif 20 appears on only one pot, but on a rather beautifully decorated example from Yanuca. In this one case, M20 appears in association with other motifs to form a highly complex pattern. The motif itself is arranged as follows: P=M20 → ConR/E-W (Figure 2.47).

Motif 21

In the complex pattern referred to in the description of M20, several motifs are brought into combination. A row of M20 occupies the lower portion of the pattern and just above are two rows of M21. Motif 21 is basically an expanded chevron which has internal decoration (Figure 2.48a). That portion of the complex pattern occupied by M21 may be seen as two series of P=M21 → ConR/E-W (Figure 2.48b), one placed directly above the other, the two coalescing at the point of contact. The rule may be restated as follows:

  • P={M21/M21} → ConR/E-W.
Motif 22

Motif 22 appears just above the M21 pattern referred to above (Figure 2.49). The basic unit is as illustrated in Figure 2.49a. The pattern formed by repetitions of this motif is discontinuous and interrupted, each repetition being separated by another motif which will be described as M23. A pattern is constructed by the following rule: P=M22 → DisR/E-W.

- 84
Motif 23

Also forming a part of the complex pattern described above are repetitions based on M23. The basic motif is a diamond which is decorated internally in various ways. The shape itself results negatively from combinations of M20, M21 and M22. Although these motifs are all “bound forms” in the sense that each is bound to some other motif in a complex pattern, a richer sample of decorated sherds may show that these motifs can appear independently. On present evidence, however, they must be regarded as bound motifs. Motif 23 is even more a bound form since it is in some ways a by-product of other motifs. It is essentially a filling-in device of a secondary nature rather than a primary motif which is used to form a dominant and positive pattern.

Alloform M23.1 contains an internal cross (Figure 2.50a) and alloform M23.2 is decorated internally with three crescents (Figure 2.50b). As they are used in the complex pattern, the alloforms appear to be in free variation and may be paired vertically so as to have two similar alloforms together, as in Figure 2.50c. In one instance, however, alloform M23.2 is paired with alloform M23.1. The pairs are repeated discontinuously in an east-west direction, thus:

  • P={M23/M23} → DisR/E-W.

The complex pattern consisting of motifs 20, 21, 22 and 23 is composed as follows:

Complex P={M22 M21 M21} alternates with {M23 M23 M20}

Motif 24

This motif consists of two sets of parallel diagonal lines which cross to form four-sided shapes, mostly the diamond shape. The effect is generally described as cross-hatching. On one sherd the cross-hatching is done by rubbing and in two others it is done by the more characteristic dentate stamping. A pattern is repeated in an all-over arrangement which fills the entire space, thus:

  • P=M24 → ConR/E-W (Figure 2.51).
Motif 25

Motif 25 is a distinctive form which occurs on a few Yanuca sherds. The basic shape of the motif is a Y which the arms of the upper portion are extended vertically. Repetitions of the basic shape fuse into each other, producing a continuous pattern. Three alloforms are isolated, each appearing on a different vessel. Alloform M25.1, which occurs in a Type A2 zone, is the basic shape without embellishments. A pattern is produced thus: P=M25.1 → ConR/E-W/CLS (Figure 2.52a).

Alloform M25.2 is a development from M25.1 in which the spaces are filled with two vertical lines. A single unit can be visualised as an extended Y-shape plus two internal vertical lines. The pattern making rule is

  • P=M25.2 → ConR/E-W/CLS (Figure 2.52b).

Alloform M25.3 is the basic unit with a doubled outline. It occurs on two sherds. However, in the one example there is less evidence of mere doubling of the outline. The shape between the units is emphasised by repeating a smaller version of it (Figure 2.52c). All alloforms of M25 follow the same pattern-making rule.

Motif 26

This represents a simplification of M25 and the development of a new shape. The - 85 basic unit now resembles a schematically drawn house with a centrally-placed door. It is not known, however, whether Yanuca potters thought of this shape as representing a house. Variations of the motif occur at Natunuku but it is a rare form at Yanuca. In the present sample of sherds the motif occurs twice only. On one sherd the motif appears upside down and on the other it appears to be the right way up, if the motif is visualised as representing a house. A pattern is formed as follows: P = M26 → ConR/E-W/CLS (Figure 2.53).

Motif 27

Motif 27 forms a simple meander pattern. The basic unit may be visualised as a squared U-shape or as a squared S-shape which is reposed (Figure 2.54a). A simple rule can be stated for the latter, P = M27 → ConR/E-W/CLS (Figure 2.54a, b), in which the juxtaposed arms of the S-shape coalesce into a single arm when repeated side by side.

Motif 28

Before describing this motif, it is well to point out that patterns consisting of continuous chevrons packed closely together and resembling a half-closed concertina occur very frequently in Yanuca decorated sherds. In some sherds, the chevrons are placed with their points along a north-south direction and in others the points are placed in an east-west direction. Although in either case the basic motif appears to be similar, the patterns made from them are different. The difference is great enough to warrant isolation of two separate motifs.

Motif 28 is based on chevrons which can be repeated horizontally. There are two main variants of the motif. Alloform M28.1 is visualised as a mirror image of an N-shaped form in which the angles of the arms are very acute and the arms themselves tend to be straight. When the basic shape is repeated, the juxtaposed arms coalesce to become one. A pattern is formed as follows:

  • P = M28.1 → ConR/E-W/CLS (Figure 2.55a, b).

When this alloform is applied to a design zone, the apices tend to be stamped much deeper than the arms. In one sherd, the upper apices appear sharper and deeper than the lower points (Figure 2.55a). In other sherds, both the upper and lower points are sharp and deep. Patterns based on M28.1 appear in zones of Type A1, A2 or H.

Alloform M28.2 results from the application of rounding to M28.1. Very slight rounding or curving of the arms of the unit occurs in one example but it is sufficient to change the appearance of the pattern. Rounding of the arms produces the effect illustrated in Figure 2.55d and thus the basic shape of alloform M28.2 is that of an ornate N reversed. A pattern is formed by the rule:

  • P = M28.2 → ConR/E-W/CLS.
Motif 29

In shape, this motif is similar to M28 except that it is repeated only in a north-south direction and it appears usually at the upper part of a vessel in a Type H zone or in an unmarked zone. The points of the chevrons tend to be greatly elongated. While repetitions of the basic N-shaped form are reasonably regular on the north-south axis, repetitions of the vertical group around the perimeter of the vessel tend to be irregular. The over-all pattern is built up in two stages. Stage 1 is as follows: P = M29 → ConR/N-S/CLS (Figure 2.56a). In the second stage, the pattern which results from state 1 is regarded as a unit and is then repeated horizontally in various ways. If the stage 1 pattern is characterised as “M29 pattern one” the horizontal patterns may be stated as follows.

- 86
  • (a) P = M29 pattern one → DisR/E-W, in which each vertical pattern is separated by considerable space, as in Figure 2.56b.
  • (b) P = M29 pattern one + ½MR of M29 pattern → DisR/E-W. This produces a negative “lightning” shape as in Figure 2.56c.
  • (c) P = M29 pattern one → DisR/E-W/INT in which the points of one vertical group of chevrons overlap those of the next group so that all lots appear to interlock (INT). In this case, however, there appears to be no attempt made to produce a formal and regularised interlocking arrangement. The alloforms of M29 appear on at least ten sherds and the evidence indicates that this motif was used more often than M28 in Yanuca.
Motif 30

This motif is comparatively rare. One alloform occurs at Yanuca and a second alloform occurs at the Sigatoka Dunes. The basic unit comprises a space bounded by four straight sides of which two or three are marked by boundary markers and the others are defined structurally. The four-sided space is shaped like a square. Within the square, two diagonals are stamped in so that the space is divided into four smaller triangular shapes. In alloform M30.1, nothing further is added internally but in alloform M30.2 each triangle is further decorated with a crescent which is placed on the base (Figure 2.57a, b). In both sherd fragments it is not clear how a pattern is built up. In the Yanuca example, alloform M30.1 appears to alternate with some other motif or alloform of M30 and this appears also to be the case in the sherd from the Sigatoka Dunes.

Motif 31

Motif 31 resembles a paddle which has a blade at each end. There is no sherd which shows the complete motif, but in one fragment the “paddle” blade is to the left and in another sherd the blade is to the right. There is a high possibility, therefore, that the motif is shaped as illustrated in Figure 2.58. In the absence of other evidence, however, M31 must remain an unconfirmed motif. The fragment indicates that the motif is applied to a very wide Type A zone and that the composition rule is P=M31 → DisR/E-W. Each unit is separated by an interval equal in length to half that of the motif.

Motif 32

As in the case of M31, there is no sherd which shows the complete shape of M32. On present evidence the basic motif consists of concentrically placed oblongs outlined in double dentate lines (Figure 2.59). The pattern-making rule appears to be P=M32 → DisR/E-W and it is placed within an especially wide Type A zone.

Motif 33

There are two sherd fragments which indicate the presence of M33. This motif resembles the Watom Island “series of concentric circles and arcs” described by Specht. There is no sherd which shows a complete unit of the motif and thus M33 must remain unconfirmed. The shape indicated by the fragments in the corpus is as illustrated in Figure 2.60 and it appears that the pattern-making rule followed was probably P=M33 → DisR/E-W.

- 87

N1.1. The best example of N1.1, the cone-shaped piece, occurs set horizontally on a small shoulder angle above an elaborate design displayed in Zone J. It is associated with similar nubbins merged with vertical bars below. Both sets of nubbins are 5 mm wide at the base, just over 5 mm high, and are of a definite conical shape with the apex rounded a little. Two other nubbins in the category N1.1 are both wider at the base more rounded.

The squared alloforms of the nubbin are not found at Natunuku.

N1.5. The remainder of the nubbins are of the category N1.5 and appear more bun-shaped than conical; that is, N1 → flattened. All are small, 4 mm in diameter, and rounded, and most are set between two horizontal lines of GZ1 or GZ2 or between two lines of GZ5. They are spaced up to 3 mm apart, but wider spacings occur, as many sherds show one nubbin only. One nubbin of this category has a flattened top but is otherwise circular.

Vertical bars

Four sherds from Natunuku have vertical bars which occur above a marked shoulder angle on a body sherd.

VB2.1. These vertical bars are of the rounded variety in cross-section, of similar shape and dimensions to those from Yanuca. A minor difference occurs in the shape of those from Natunuku, which have parallel sides and do not splay out at the lower end.

Transverse bars

TB3.1. The structural bar TB3.1 is found on several Natunuku sherds and is constructed in the same fashion as those from Yanuca. The lip edge is sometimes finished by squaring, but in other cases the edge is rounded and convex, or distorted by impressing. The square edges are narrower than the base of the bar.

TB3.2. Other transverse bars constitute the alloform TB3.2. They are all rounded in cross-section, two being narrow, 3 mm high and set clearly against the clay surface, while two others are twice as wide and high and have edges which are carefully smoothed into the clay. All forms of TB3.2 are found on the outer surfaces of sherds. There are no examples of this form used on the inner rim to define a decorative Zone F as at Yanuca.

TB3.3. One large sherd has a triangular cross-sectioned transverse bar, otherwise referred to as a flange. It is a structural part of the pot, a solid ridge set on a small shoulder angle. It stands 1.5 cm out from the side of the pot and is 2.25 cm in width at the base. There are no examples of the smaller and more decorative bar which comes from Yanuca.

TB3.4. One example of a bar with square cross-section is found on a sherd with dentate-stamped decoration of DE5. This transverse bar is approximately 1 cm square with a row of vertical impressions on the outer face.

TB3.5. Two sherds have slightly raised ridges in horizontal positions and these are TB3.5 for Natunuku. Whether these were originally appliqué pieces is doubtful and they could result from tooling of the clay surface into a low ridge of triangular cross-section, as the clay rises steadily to form the bar with no marked outline.

- 88

The basic elements at Yanuca, with the exception of DE2, are all found at Natunuku. In addition there are three more design elements, DE9, DE10, DE11, and a modification of DE4 (Figure 3.1).


The examples illustrated in Figure 2.17a, c and d are found at Natunuku. The rules are P = DE1.1 → DisR/E-W; P = DE1.1 (rep) → DisR/E-W.


The small stamped circle is the same as that from Yanuca and shares the same rule.


The alloform DE4.2 is found at Natunuku as a small oval with pointed ends, set vertically 2 cm apart and usually in a Type A zone. Each element is 0.5 cm high. Double versions do not occur (Figure 3.1). P = DE4.2 → DisR/E-W.


The Natunuku element is similar to the Yanuca one, although the lines are not grouped and may be up to 4 cm high.


At Natunuku DE6 is found as a number of long parallel oblique lines which are uniformly spaced and cover a large area. Some examples show a border of GZ1 or the Natunuku GZ5. P = DE6 → DisR/E-W.

Alloform DE6.1 consists of one to four lines in a series, usually between a double border or with a single line of GZ1 if it occurs in zones G or H.

  • P = DE6 X N → DisR/E-W,

where N is a variable number of repetitions from one to four. DE6. 2 is not found.


The chevron at Natunuku is a series of single V-shaped lines of the same size, equally spaced and usually set horizontally. It is described as alloform DE7.3. P = DE7.3 → DisR/E-W.


This element is similar to the Yanuca one although DE8.3 is not present. P = DE8.1 → DisR/E-W; P = DE8.2 → ConR/E-W.


This element is a small elongated diamond (Figure 3.1).

  • P = DE9 → DisR/E-W.

This element is an upright rectangle with the shorter sides half the length of the longer two (Figure 3.1). P = DE10 → DisR/E-W.


This is also rectangular in shape but the two longer sides curve inwards towards each other, presenting a rather negative appearance (Figure 3.1). An unfinished example has no base line. P = DE11 → DisR/E-W.

Elements DE9, DE10 and DE11 are all 0.5 cm high.

- 89
Motif 1

The single horizontal line of crescents is found on four sherds at Natunuku. This forms M1.1. P = M1.1 → ConR/E-W.

The double version is M1.2. This alloform occurs at Natunuku where a second line of crescents is placed 0.25 cm above the original line.

P = M1.2 → ConR/E-W.

Motif 2

The “mesh of crescents” of the type M2.1 is found at Natunuku invariably bordered by a zone marker either above or below, so that the points of the crescents, or the midpoints of each, touch the line. The rule is

  • P = M1 → ½DM X 2/ConR/E-W, or P = M2.1 → ConR/E-W.
Motif 3

All examples of this motif at Natunuku appear as in Figure 2.27b with the points of the crescents touching the upper line. The rule is the same as for the Yanuca example.

Motif 5

Here, the process of rounding the elements to produce Yanuca M4 and M5 is absent. Motif 5 at Natunuku is simply an opposing series of crescents of the M1 type (Figure 3.2). A band of ovals is formed and there is no attempt to fuse the crescents where they meet.

  • P = DE1 (rep) → SUP X 2/ConR/E-W, or P = M5 → ConR/E-W.
Motif 6

This motif occurs on two sherds and the rule is the same as that for Yanuca. There are no associated zone markers shown on these sherds.

Motif 7

One sherd with this motif was found at Natunuku. This is the only example of the rounding process in connection with the use of DE1. The rule is

  • P = DE1 (rep/rd/db) → ConR/E-W, or P = M7 → ConR/E-W.
Motif 8

This is one of the most commonly found motifs at Natunuku and appears with several variations, usually on sherds from the flat-bottomed dish. The basic U-shape does not occur alone or unadorned as in M8.1 and may stand upside down or the right way up. It is constructed as part of a single line of M1 to which has been added a series of parallel vertical lines, each running from the junction of two crescents to deepen the U (Figure 3.3). Several examples of this are carelessly executed so that the vertical lines, although parallel, are not evenly spaced and do not coincide with the points of the crescents, or the line of crescents and the row of vertical lines do not touch at all and stand a few mm apart. In all the completed examples of this motif, the basic vertical lines are of the same length and join a zone marker or stop just short of it. The one exception is on a base sherd from a shallow dish where the U motif (M8.5) is immediately above the base angle the right way up, and the lines between each U project beyond the other sets of vertical lines to twice their length. However, several of the other alloforms of M8 - 90 are on sherds too small to contain the complete motif and so all the rules for the length of the vertical lines are not known.

M8.2. This is the simplest alloform of M8, although it is not produced at Natunuku with the three short lines projecting from the base of the crescent. Generally all the vertical lines are of the same length. The rule is

  • P = M8.2 → ConR/E-W/CLS.

M8.3. Here, the simple U shape has two vertical lines inside it standing parallel to the sides of the U. In the two examples found, all the vertical lines are of the same length and spaced evenly apart. P = M8.3 → ConR/E-W/CLS.

M8.4. Alloform four of this motif shows the U shapes depicted without the central coalescing line, each U standing along 0.25 cm from the next. One vertical line is drawn down the centre inside each U. P = M8.4 → DisR/E-W.

M8.5. In this alloform, which is also composed of a series of complete U forms, each U is separated from the next by a vertical line which begins at the same level as the crescents but is longer than the sides of the U. One vertical line stands inside each U, the same length as the side of the U but drawn further up into the crescent.

P = M8.5 → DisR/E-W.

M8.6 and M8.7. Both alloforms have oval or almost triangular ends to the U shapes, making them extremely deep instead of the usual semicircular crescent. Alloform M8.6 is otherwise a repetition of the Natunuku version of M8.2, with the U shapes coalescing and all the verticals parallel and of the same length. Alloform M8.7 is the same as M8.4 except for the pointed base of each U shape.

  • P = M8.6 (comp) → ConR/E-W; P = M8.7 (comp) → DisR/E-W.
Motif 10

The Natunuku variant of M10 may be regarded basically as two opposing chains of M8.1 in which the open end of the U is closed by another crescent (Figure 3.4). The rule is P = M8 (MR) → ConR/E-W/CLS, and this is henceforth presented as M10. The smooth rounded outlines of the Yanuca M10 and the use of the double lines to add substance to the motif are absent at Natunuku.

M10.1. The most simple form of M10 is a single line of U shapes with open ends closed by arcs. Coalescence occurs as the junction of each shape is a single line. The motif is made up therefore of two lines of M1 set 1.5 cm apart with the points of the crescents facing inwards. Vertical lines join each opposing point. M10.1 also has a single line in the centre of each enclosed space, standing parallel to and equidistant from the basic vertical line. The rule is P = M10.1 → ConR/E-W/CLS.

M10.2. This alloform consists of two complete chains of M8.1 which are set in opposition, the lower having moved half a space in one direction so that the projecting arms of one U fit into the space created by the opposing U. The two lines of U shapes are placed close enough together for the vertical lines to reach to the furthest extent of the opposite crescent. P = M10.2 → ConR/E-W.

M10.3. Alloform M10.3 is an elaborate version of the above with extra vertical lines drawn between the already existing ones. On one sherd there are two developments of it resulting from unequal spacing and the application of different sized crescents in the lower line. On one side the design is arranged as for M10.2 with one extra vertical line drawn in each space, thus each U contains three lines. On the other end the change in the size of the crescents results in a movement along of one third space and only one extra line is added to each U. No rule is given for this as it may be the result of poor spacing.

Motif 11

Alloform M11.2 is the only alloform which occurs at Natunuku. One example - 91 shows the outer border of the motifs as GZ1 while the next three zones are marked by GZ2. The vertical marker is also GZ1. The rule is the same as at Yanuca.

Motif 15

The use of closely positioned parallel vertical lines appears to result from a desire to cover a large band of the vessel surface as fully as possible. The lines form a continuous band without gaps and are bordered below and sometimes above with a zone marker of the type GZ1 or GZ5. At Natunuku this is M15.3 and the rule is P = M15.3 → DisR/E-W.

Motif 16

Alloform M16.2 (Figure 2.43b), the Yanuca alloform present at Natunuku, occurs in both instances as a band above the base of a flat-bottomed dish. The only complete triangles in it are the smaller negative ones. Alternatively, it may be seen as a series of opposed and interlocking triangles set in a Type A1 zone. The rule is

  • P = M16.2 → ConR/E-W.

M16.4. This occurs as a single band only. A Type A1 zone 5 subdivided into squares, each of which contains two diagonal lines forming a triangle with the apex meeting the top of the square, and the opposite side acting also as the base (Figure 3.5). Single dentate-stamped lines are used for all parts of this and the following alloform. The rule here is P = M16.4 → ConR/E-W/CLS.

M16.5. Here the zone is divided into rectangles and the triangles are formed with the vertical lines as their centre. The apex of each triangle lies at the point where the vertical line meets the lower zone marker and the base of each triangle is the upper zone marker (Figure 3.5). Thus, each one lies upside down and is separated by a short distance from the next. The rule is P = M16.5 → DisR/E-W/CLS.

Motif 17

Two examples of the postulated M17.1 (Figure 2.44a) appear at Natunuku. One is an isolated motif set horizontally on a double line, but the sides of the triangle are marked by single dentate-stamped lines, and DE1 is a deeper crescent than indicated in Figure 2.44a. The rule for this is

P = M17.1a (rep) → DisR/E-W.

M17.1. A complete motif of M17.1 shows the addition of DE1 on the upper and lower sides of the rectangle, all the horizontal and vertical lines in this case being double. An alternative pattern lacks the two extra elements but has each motif separated from the next by a distance of 2 cm, although they are still set between two bands of double lines (Figure 3.6). Both of these are found in a Type F zone, a rare example of one kind of zone superimposed on another.

  • P = M17.1b (rep) + MR M17 → DisR/E-W.

M17.3. One sherd has a variation of the Yanuca alloform M17.3 which is set between two single lines that run in a slight curve and are 2.5 cm apart. The motif is therefore elongated to fit into this wide zone and only positive triangles extend from the upper line, the lower standing on its own. The pattern is composed of a series of narrow triangles suspended from the upper line, their apices standing a short distance from the lower line. Design element one is here a deepened crescent, placed on the base of each triangle and at the other end of the triangular area between them in line with the apices (Figure 3.6). In this example, the rule for the Yanuca version of M17.3 does not apply as the lower half of the motif is not a true reversal of the upper part, there being no complete triangles, merely the spaces between the positive ones. P = M17.3 → DisR/E-W.

M17.4. This alloform is closely related to Yanuca's M17.1b, but whereas in - 92 Yanuca a pattern is discontinuous, in Natunuku it is continuous. Thus the rule is P = M17.4 → ConR/E-W. The Natunuku alloform is illustrated in Figure 3.17. Its position in Figure 3.17 reflects an earlier decision to regard it as a new motif but grounds for doing so are weak.

Motif 18

Part of M18 is found at Natunuku, namely the triangles set on a double line and filled with three oblique strokes. P = M18.1 → ConR/E-W.

Motif 19

This motif occurs in all cases in a Type F zone and several sherds have single or double zone markers as the lower border. As at Yanuca, the triangles may be set either up or down, although the base uppermost is more usual, and they are not distinguished by any particular regularity. Where the zone marker is absent the triangle stands alone or the bases overlap slightly as in Figure 2.46d. Motif 19.2 is not found at Natunuku. The rules are thus

  • P = M19.1 → ConR/E-W; P = M19.1 → DisR/E-W.

M19.3. This alloform marks a departure from the Yanuca forms. It is illustrated in Figure 3.16. The rule for forming a pattern is P = M19.3 → Con/E-W/CLS. It was decided originally to regard this alloform as a new motif but grounds for doing so were later considered inadequate.

Motif 24

Two kinds of diagonal cross-hatching are found at Natunuku, both executed with dentate-stamped tools. One is similar to that illustrated in Figure 2.51, the diamonds being rather squat in shape and the lines thick and not always straight or continuous.

M24.2. The other alloform is a fine, delicately made series of lines forming small elongated diamonds which are quite regular. The rule is the same in both cases. P = M24.1 or M24.2 → ConR/E-W.

Motif 25

This is not found at Natunuku. A motif somewhat similar to Yanuca M25 is described under M37 below.

Motif 26

Natunuku has an elaborate version of M26 and it is here considered as alloform M26.2 (Figure 3.7). It lies in a Type A zone and the house shapes do not reach the upper border of the zone marker. Each house contains two lines, but they are not joined at the top, and the triangular areas between the house roof and the upper borders are each filled with two V shapes. The rule is

  • P = M26.2 → ConR/E-W/CLS.
Motif 28

This motif is represented by two examples of alloform M28.2 where the rounding process is carried out in one direction only. The lines are applied in groups of from five to eight and are found in a Type A zone, but without the order to be seen in those from Yanuca. Some lines stand alone, others are joined at the top or at the bottom, while the spacings and alignments are uneven.

  • P = M28.2 → ConR/E-W/CLS.
- 93
Motif 29

At Natunuku this appears as a wide zigzag aligned north-south and repeated in a well-spaced horizontal band. It may be considered as an alloform M29.2 and is an exploded form of Yanuca's Figure 2.56c.

  • P = M29.2 → DisR/E-W/INT/expanded.
Motif 30

This is found at Natunuku as small and clearly marked squares and rectangles. The sides of the squares are 1 cm in length and the rectangles are 1 cm by 2 cm.

M30.3. The rectangles are subdivided by a diagonal cross and each triangle so created contains a crescent, DE1, on the base, and a small stamped circle, DE3, at the apex. The rectangles stand vertically, each 0.5 cm from the next, and they are set parallel to a line of GZ5. P = M30.3 → DisR/E-W.

M30.4. This alloform is the square, also subdivided by a diagonal cross, each triangle containing a stamped circle. The squares are set in similar fashion to the rectangles of M30.3. P = M30.4 → DisR/E-W.

M30.5. Here the square is subdivided into four smaller squares by an upright cross, each of which contains two elements, a centrally placed stamped circle, partially surrounded by a crescent of DE1. P = M30.5 → DisR/E-W.

Each motif is set in a space bounded on both sides by a vertical bar, above by a nubbin, and below by the sharply angled shoulder of the pot. On present evidence, at least, three of these alloforms are found together in one zone.

Motif 33

Two sherds show considerable portions of M33, confirming that it exists in Fiji as a basic pattern of at least four concentric circles. Both examples show the use of GZ5 for the main outlines with the small stamped circle, DE3, arranged in the intervening spaces. P = M33 → DisR/MULTI.

The following motifs are present at Natunuku but do not occur at Yanuca. Some are clearly related to Yanuca motifs, others have parallels in Tonga or in New Caledonia, while others are unconfirmed as yet.

Motif 34

This is a line of diamonds set on their sides with the extremities just touching, contained within a Type A zone (Figure 3.8). They are much larger than DE9 and because the negative effect of the motif is also important in two of the alloforms this motif is probably not related to DE9. It is also possible that the diamond shape was not considered during the construction of the motif and results from another series of processes. A pattern may be formed as follows:

  • P = DE9 (rep) → ConR/E-W, and is simplified as P = M34 → ConR/E-W.

M34.1. Here the diamonds are narrow and lie in the centre of the zone, 0.25 cm from each border. The rule is as above.

M34.2. The diamonds of this alloform are larger than those of M34.1 and the upper and lower points touch the zone marker. The decorator may have constructed this as a series of superimposed and alternating lines, the diamond shapes appearing in the series of superimposed and alternating lines, the diamond shapes appearing in the areas where the two extremities touch the zone. However it was constructed, the motif accentuates the diamonds because the triangular areas left at the sides are shaded in at the top and in the opposite direction at the bottom.

A similar motif, although placed vertically, appears on one sherd and is constructed of a series of opposing triangles set in a Type A zone with the apices meeting in the middle. Each triangle, whose base is part of the zone marker, is further decorated by lines of the impressed cross-hatch zone marker. The triangles on the - 94 left are infilled with oblique lines, those on the right with horizontal lines. The effect is to draw attention to the triangles rather than to the diamonds outlined between them. The rule for this is P=M34.2 → ConR/E-W, assuming a cut so as to produce a complete triangle each time.

M34.3. This is similar to the previous alloform although unshaded, and the lines are clearly stamped, each diamond having been made separately. The rule is as for M34 above.

Motif 35

This is the most complex motif found at Natunuku and is composed of a number of variations on a simple design as shown in Figure 3.9. It is extended in all directions to cover a large area of the pot surface. Two sherds show the motif quite fully and smaller pieces have recognisable parts of it. However, the two larger sherds show that although the rules for building the motif are simple, they may be applied in an arbitrary fashion and any of the four alloforms used at any stage. The alloforms will be described first and the rules for their use will follow. Although the alloforms are depicted in Figure 3.9 as standing vertically, they are aligned on the diagonal, all angles being at 90 degrees.

M35.1. The basic unit here is a narrow rectangle, approximately 0.25 cm by 3 cm, with one end open and a line drawn down the centre parallel to the sides. The line may or may not touch the end of the rectangle.

M35.2. The narrow open rectangle is also the alloform here but it is wider than M35.1 because two parallel lines are drawn down the centre.

M35.3. Alloform M35.3. is a double version of the basic unit with a slightly larger rectangle outside the first narrow one.

M35.4. This is also doubled but the single line is marked down the centre of the inner rectangle.

M35.5. A slanted version of the latter is formed by putting the two rectangles on the diagonal but marking the end as a vertical line. This is rare, found only at the edge of the motif or on small sherds.

The rule for M35 is P=M35 → DisR/MULTI.

A complex pattern is drawn up in the following way: one of the alloforms of the dimensions indicated for M35.1 is drawn in an oblique position and the open end is closed by being made part of the side of a second alloform placed at right angles to it. The motif is then built up by the addition of a series of parallel lines set at right angles to any of these basic lines, or by placing more of the basic alloforms in parallel positions at either end. The parallel lines and open ends of the rectangle are contiguous to an existing line but the closed ends of the alloforms are not usually joined to any other part of the design. Length of line and method of arrangement may be left to the decorator, although the outstanding quality of the motif appears in the perfect spacing and exact alignment of the various components. All examples are executed with the finest of dentate stamp tools. One sherd shows the motif bounded above by a double zone marker but unfinished 5 cm below.

Motif 36

The three alloforms of M36 are composed of triangles arranged in various ways (Figure 3.10).

M36.1. This alloform is presumably contained within a Type A zone, although the only example of the motif is incomplete. The triangles are spaced along the zone alternately, one from the bottom line, one from the top, with the apices stopping short of the opposite zone marker by 0.25 cm. The spaces between the triangles are filled with a number of lines parallel to the triangle sides. One suspended triangle has a smaller one drawn inside it and the lines drawn to the left - 95 of it appear to extend from border to border. It is not clear how the next set of lines would fit in as the sherd is broken off at this point. A complete but less complicated version of this occurs at Tonga and is illustrated in Poulsen's Category 14. The rule is P = M36.1 → ConR/E-W.

M36.2. A series of four triangles is built up with parallel sides standing on a single zone marker which is also their common base. P = M36.2 → DisR/E-W/DECR.

M36.3. This is also a series of triangles on a common base, but each triangle does not stand as a clear unit as one of the sides is extended past the apex to form also the side of a larger triangle. The other side of this triangle is extended past the apex and so on. The sherd which has the largest piece of this motif shows the triangles with the base uppermost. P = M36.3 → DisR/E-W/INT.

Motif 37

The following patterns are based on the house shapes of the Yanuca motifs M25 and M26, but at Natunuku each pair of houses is outlined and divided with a large Y and the intervening spaces are filled with numbers of parallel lines. It is therefore considered as another motif and has six alloforms. Two of these are based on the Y motif with the arms extended vertically (Figure 3.11). Some of these alloforms are no sherds large enough to show that a zone marker is used in association with them.

M37.1. This alloform stands on a single zone marker. The house shape has one vertical line in the centre, and between the house and the Y is an angled line parallel to both. P = M37.1 → ConR/E-W/CLS/DECR.

M37.2. In a more elaborate form of the above alloform there are two lines drawn between the house and the Y and further parallel lines are drawn within the arms of the Y.P = M37.2 → ConR/E-W/CLS.

M37.3. This is similar to M37.1 but here the house contains two lines and these are capped by a small crescent of DE1 with the two points downwards.

  • P = M37.3 → ConR/E-W/CLS.

M37.4. This is a simplified version of the following two. It is a series of Y shapes with the points of the arms touching. The narrow house shape so formed contains a single vertical line. P = M37.4 → ConR/E-W/CLS.

M37.5. Here the arms of the Y are extended vertically and they are placed in a band alternately, one up, one down. Each is separated from the next by three lines which parallel the stem and two arms of the Y. A short vertical line is marked between the arms. P = M37.5 → DisR/E-W/CLS/MR.

M37.6. The same long-armed Y is used here, but all are set the same way up in a band. A single line is drawn between the stems. P = M37.6 → DisR/E-W.

Motif 38

This appears to have a zigzag line as its main unit, each V being filled with a series of diminishing V shapes (Figure 3.12). P = M38 → DisR/E-W/DECR.

Motif 39

This is composed of two or more parallel zigzag lines which run horizontally (Figure 3.13). A unit of M39 consists of one V of the zigzag, doubled, with the two lower points joined by a single vertical line.

M39.1. This is a band of M39 subdivided by a series of vertical lines which join the lower points of each zigzag. Each enclosed space contains two short vertical lines. P = M39.1 → ConR/E-W/CLS.

M39.2. Here three or more lines of M39 make up the pattern which appears to cover the entire upper half of at least two pots. Every line of zigzags has a vertical - 96 extending down from every low point and up from every high point, thus producing a fine network pattern. P = M39.2 → ConR/E-W/CLS. This is confirmed from Tonga where it is Category N5. It is part of this motif which is adapted to fit into a zone whose top part becomes more restricted in area.

Motif 40

This is one of those motifs used by Natunuku potters when faced with a wide band which is not defined very closely, if at all, by zone markers. It consists of a deep double arc facing downwards to enclose a triple chevron (Figure 3.14). The outside of the arc is accentuated by closely spaced vertical lines of the same length which follow the curve of the arc and continue to repeat the shape once more to the right. Unfortunately, the sherd ends at this point and we do not know what further original designs there may have been beyond this. However, a similar motif occurs at Site 13 in New Caledonia. This is turned up the other way and shows two arcs with their additional straight lines, but the chevrons are lacking. Although this confirms the motif, it is considered too incomplete for a rule to be stated.

Motif 41

This appears to be a combination of parts of other motifs. Although it is presented as a separate motif, it is unconfirmed and is given no rule. Half of it is made up of a double zone marker and opposing crescents of the central part of M11.2, but it is otherwise unzoned and placed horizontally above a series of triangles which are an enlarged version of M16.2 The suspended triangles use the same zone marker as their base. Two of the crescents are therefore on the base of one of the triangles (Figure 3.15).

Motif 42

M42 is a vertical line of chevrons enclosed within a narrow vertical zone (Figure 3.18). It occurs in Tonga as Category P31. P = M42 → DisR/N-S.

Motif 43

A combination of DE1 and a series of triangles forms a simple pattern against a single zone marker. The triangles are set side by side below a zone marker which is also their base, and the unit of DE1 is placed below the point where each triangle meets the next (Figure 3.19). P = M43 → ConR/E-W.

Motif 44

This is an elaboration of M37, inhabiting a well-defined zone, and is sufficiently distinct for it to be considered as a separate motif. It is essentially a line of house shapes separated by the Y motif (Figure 3.20). A V is drawn inside the arms of the Y and the house is completed by a door as in M26. The lines of the door and the sides of the house are then extended obliquely outwards for a short distance and this has the effect of outlining a negative house shape upside down between the houses. The motif may thus be viewed from either direction. It is the only example so far found at Natunuku of the use of the Type I zone. P = M44 → ConR/E-W/CLS.

Motif 45

This motif is based on a band of squares, with sides of 3 cm, each separated from the next in various ways and completed with diagonal lines drawn inside the square (Figure 3.21). Although at first sight this motif may seem to be related to M30 it is considered as a motif in its own right because it is drawn twice the size of the units in M30 and forms a continuous band.

- 97

M45.1. Each square stands a distance of 0.5 cm from the next and has lines drawn from corner to corner in a diagonal cross. P = M45.1 → DisR/E-W.

M45.2. Alloform M45.2 shows the squares with vertical side lines only and another two lines drawn parallel between them. The diagonal cross is similar to that in alloform M45.1. This motif is therefore composed of an east-west line of a set of four vertical lines alternating with a diagonal cross. P = M45.2 → DisR/E-W.

M45.3. Here the squares are found as part of a Type F zone with the horizontal line as a single zone marker. Each square is separated from the next by 0.25 cm. A short diagonal line is drawn across the upper left corner of each square.

  • P = M45.3 → DisR/E-W.
  • AVIAS, Jacques, 1950. “Poteries Canaques et Poteries Préhistoriques en Nouvelle-Caledonie.” Journal de la Société des Océanistes, VI (6):111-40.
  • BIRKS, Lawrence and Helen, 1966. Archaeological Excavations at Sigatoka. Sigatoka Research Project, Fiji, Preliminary Report No.2, July 1966. Fiji Museum, mimeographed.
  • —— 1967a. “A Brief Report on Excavations at Sigatoka, Fiji.” New Zealand Archaeological Association Newsletter, 10(2):16-25.
  • —— 1967b. “Radiocarbon Dates for Sigatoka, Fiji.” New Zealand Archaeological Association Newsletter, 10(2):69.
  • CASEY, D. A., 1936. “Ethnological Notes.” Memoirs of the National Museum of Victoria, 9:90-7.
  • EMORY, K. P., 1962. “Three-Year Polynesian Research Programme.” New Zealand Archaeological Association Newsletter, 5 (1):96-7.
  • GARANGER, José, 1971. “Incised and Applied-relief Pottery, its Chronology and Development in Southeastern Melanesia, and Extra Areal Comparisons,” in R. C. Green, and M. Kelly (eds.), Studies in Oceanic Culture History Vol. 2. Honolulu, B. P. Bishop Museum (Pacific Anthropological Records, 12), pp. 55-65.
  • GIFFORD, E. W. and Richard SHUTLER Jr., 1956. “Archaeological Excavations in New Caledonia.” Anthropological Records, 18(1). Berkeley and Los Angeles, University of California Press.
  • GOLSON, Jack, 1961. “Report on New Zealand, Western Polynesia, New Caledonia and Fiji.” Asian Perspectives, 5 (2):166-80.
  • —— 1968. “Archaeological Prospects for Melanesia,” in I. Yawata and Y. H. Sinoto (eds.), Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 3-14.
  • —— 1971. “Lapita Ware and its Transformations,” in R. C. Green and M. Kelly (eds.), Studies in Oceanic Culture History vol. 2, Honolulu, B. P. Bishop Museum (Pacific Anthropological Records, 12), pp. 67-76.
  • GROUBE, L. M., 1971. “Tonga, Lapita Pottery, and Polynesian Origins.” Journal of the Polynesian Society, 80(3):278-316.
  • HEDRICK, John D., 1969. Lapita-style Pottery from Malo Island, New Hebrides. Unpublished M.A. thesis, San Diego State College.
  • —— 1971. “Lapita Style Pottery from Malo Island.” Journal of the Polynesian Society, 80 (1):5-19.
- 98
  • —— and Mary Elizabeth SHUTLER, 1969. “Preliminary Report on “Lapita Style” Pottery from Malo Island, Northern New Hebrides.” Journal of the Polynesian Society, 78 (2):262-5.
  • LENORMAND, M. H., 1948. “Découverte d'une gisement de poteries indigènes à l'lle des Pins.” Etudes Mélanésiennes, 5:54-8.
  • MEYER, O., 1909. “Funde prähistorische Topferei and Steinmesser auf Vuatom, Bismark-Archipel, Südsee.” Anthropos, 4:215-52; 1093-5.
  • PALMER, J. B., 1966. “Lapita Style Potsherds from Fiji.” Journal of the Polynesian Society, 75 (3):373-7.
  • —— 1967. “Sigatoka Research Project: Preliminary Report.” New Zealand Archaeological Association Newsletter, 10 (1):2-15.
  • —— 1968. “Recent Results from the Sigotaka Archaeological Program.” in I. Yawata and Y. H. Sinoto (eds.) Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 19-27.
  • POULSEN, J., 1964. “Preliminary Report on Pottery Finds in Tonga.” Asian Perspectives, VIII (1):184-95.
  • —— 1967. A Contribution to the Prehistory of the Tongan Islands. Unpublished PhD thesis, Australian National University. 2 vols.
  • SOLHEIM, Wilhelm G., 1964. “Pottery and the Malayo-Polynesians.” Current Anthropology, 5:360; 376-84.
  • SPECHT, Jim, 1968. “Preliminary Report of Excavations on Watom Island.” Journal of the Polynesian Society, 77 (2):117-34.
  • SHUTLER, Richard Jr., 1971. “Pacific Island Radiocarbon Dates, an Over-view,” in R. C. Green and M. Kelly (eds.), Studies in Oceanic Culture History vol. 2, Honolulu, B.P. Bishop Museum (Pacific Anthropological Records 12) pp. 13-27.
  • WATT, William C., 1966. Morphology of the Nevada Cattlebrands and their Blazons, Part One. N.B.S. Report 9050, U.S. Department of Commerce.
  • —— 1967. Morphology of the Nevada Cattlebrands and their Blazons, Part Two. Pittsburgh, Carnegie-Melton University.

- 99 Page is blank

- 100 Page is blank

1   Poulsen 1967:148
2   Poulsen 1967:156.
3   Groube 1971.
4   Groube 1971:303.
5   Golson 1968:6, and Figure 2, second illusration from left.
6   An argument against this hypothesis is given in the next section.
7   See pp. 6-7.
8   I have seen some of the decorated sherds from the Santa Cruz Group and from observation M8, M15 and M24 are similar to Yanuca forms, but there are also variants of M10, M26, M27 and M33. When analysis of the material from the Santa Cruz Group is completed a whole new range of design motifs will be added to the present list.
9   Evidence for Watom Island is based on Specht 1968.
10   Based on the records of Gifford and Shutler 1956.
11   Based on Avias 1950 and Lenormand 1948.
12   Hedrick 1969; Hedrick and Shutler 1969; Hedrick 1971.
13   See Shutler 1971:15-6.