Volume 70 1961 > Volume 70, No. 3 > Marianas archaeology, by Marcian Pellett, p 321-325
                                                                                             Previous | Next   

- 321
Report on an Excavation on Tinian

In this article a recent excavation by Reverend Pellett, on the island of Tinian in the Marianas, is interpreted in the light of the results of Dr. Spoehr's 1940-1950 archaeological programme in the Marianas.


IN MUCH of the world, pottery types have served as valuable indicators of prehistoric culture change and as a means of cross-dating one locality with another. Though pottery was absent from eastern Micronesia at the time of European discovery, in the islands of western Micronesia it was made in abundance. The pottery types of western Micronesia are therefore important for an understanding of prehistoric culture change in the area and eventually should be useful in ascertaining the relations of the western Micronesian islands to one another and to the regions to the west and south.

This paper summarizes an excavation conducted by Pellett on the island of Tinian in the Marianas in May, 1958. The paper refers to the following points: (1) culture change in the Marianas as reflected in pottery, (2) decorated Marianas pottery as an aid in establishing relationships with other island groups, and (3) artefacts other than pottery associated with the earliest known periods of occupation of the Marianas.


Pellett's excavation consisted of a test trench at the famous Taga site on Tinian. This site was first brought to the attention of Europeans through Lord Anson's visit to Tinian in 1742. 1 When first accurately described by Hornbostel in the 1920's, the site consisted of eighteen clusters of stone columns, or latte. 2 Today only the central group of columns, known locally as the “House of Taga”, after the legendary Chamorro giant to whom the site is attributed, is left. A more complete description of the site is given in Spoehr. 3

The test trench was located 93 feet from the House of Taga on a line 10° W. of N. (Mag.) of the only two standing columns. The trench was 3 feet by 9 feet and was excavated to a depth of 10 feet 2 inches. To our knowledge this is the deepest excavation yet made in the Marianas. Unfortunately, no material suitable for radio-carbon dating was found.

- 322

The upper 1 foot 6 inches of the excavation consisted of dark gray sandy soil, filled with numerous sherds. The next foot was a stratum of sterile yellow sand. Below this lay gray sand merging into the underlying indurated yellow sand of which the sub-structure of the site is composed. Under the sterile sand level, cultural material was again found, to a depth of 9 feet 8 inches.

The material from the excavation is deposited at the Bishop Museum in Honolulu.


Spoehr's 1949-50 excavations at two sites—Laulau Rock Shelter and Chalan Piao—on Saipan revealed a temporal relationship between the two numerically most important pottery types so far found in the Marianas, one called Marianas Plain, the other Marianas Red. Description of these two types is given in Spoehr's report. 4 Marianas Red, technically a handsomer and better made ware than Marianas Plain, dominated in the lower levels at these two sites and is the earlier. The radiocarbon date of 1527 B.C. ± 200 obtained at Chalan Piao is associated with Marianas Red.

On the other hand, Marianas Plain, though possibly always present in some degree in the Marianas, is the dominant type during the later periods and is associated with the latte sites of the Marianas, some of which were occupied up to historic times. At these sites, Marianas Red is absent. One of Spoehr's test trenches at the House of Taga was carried to a depth of 6 feet, but no Marianas Red was found.

Pellett's excavation at the House of Taga now further clarifies the picture. Going deeper than Spoehr, he found that Marianas Plain was indeed associated with the latte builders of Taga, but that underlying this occupation was another one, characterized by Marianas Red. His excavation is therefore valuable corroborative evidence that in the cultural chronology of the Marianas Islands, Marianas Plain is characteristically late, Marianas Red characteristically early.

The seriation on Marianas Plain and Marianas Red from Pellett's excvation can be simply stated. Marianas Red occurred only at levels below a surface depth of 4 feet. In these levels it was the dominant type. From 4 feet to 9 feet 8 inches below the surface, at which the lowest sherd (Marianas Red) was found, 921 Marianas Red sherds were recovered. In these same levels a total of only 90 Marianas Plain sherds were found, and 77 of these were in the 4-5 foot level. Conversely, no Marianas Red occurred above a surface depth of 4 feet. Here Marianas Plain dominated. No whole vessels of either pottery type were found and conclusions are based on sherd counts alone.

The chronological relationship between these two pottery types will be particularly important when their relationship to similar types from other island groups is eventually established. Marianas Plain is typologically the same as the unlaminated ware the Giffords found on Yap in their lower levels, associated with radiocarbon dates of A.D. 847 and A.D. 176, although the relatively unspecialized nature of Marianas - 323 Plain limits its usefulness in cross-dating. 5 As for Marianas Red, it will no doubt be shown to be part of the widespread complex of red wares found in the Malaysian region. The chronology and distribution of this red ware tradition throughout Malaysia is an important problem on which there is yet little published material. It is interesting that McCarthy and Setzler found red ware trade sherds at Macassan sites in northern Arnhem Land, and that Aga-Oglu after examining them reported that they are related in body, shape, and finish to Philippine native ware deposited at the University Museum, University of Michigan. 6 In time, it will no doubt be demonstrated that Marianas Red is the eastern fringe of a red ware tradition centering in Malaysia.


In his Marianas excavations, Spoehr found 12 sherds which he considered might have been a trade ware and for which he did not establish a formal pottery type. 7 The special nature of these sherds lies in their decoration. This consists of impressed circles and bands of opposed chevrons and parallel lines which were filled with lime, resulting in a very distinctive decorative treatment.

Pellett found an additional series of 26 lime decorated sherds in the levels below 4 feet at Taga. An important point is that the distinctive surface treatment had been applied to Marianas Red sherds. Also, four black ware sherds exhibited a variation of this treatment. Figure 1 illustrates a sample of Pellett's lime-decorated sherds.

The significance of these sherds is that their distinctive decoration should prove of substantial aid in relating the Marianas to other pottery-making islands where this type of decoration will one day surely be found and reported. It is for this reason that these lime-decorated sherds are described here. We strongly suspect a tie with the Phiilppines.


Although a complex of stone, bone, and shell artefacts have been found associated with Marianas latte sites and with Marianas Plain pottery, this is not the case in the lower levels of excavations characterized by Marianas Red pottery, where little besides sherds has been discovered. Pellett's excavation uncovered a small series of artefacts associated with Marianas Red sherds and help fill out the cultural picture of the earlier periods of occupation of the Marianas. A selection of these artefacts is illustrated in Figure 2.

In addition to those illustrated, the following were found:

Stone flakes, possibly used as knives or scrapers. Provenience: Level 5-6 feet—1; 6-7 feet—10; 7-8 feet—6; 8-9 feet—3; 9-10 feet—5.

Small stone pebbles, probably hammerstones. Provenience: Level 7-8 feet—1; 9-10 feet—1.

Small basalt flake, ground and polished except on one broken surface. Probably chip from poll of cylindrical adze. Provenience: Level 5-6 feet below surface.

- 324

Sea urchin spines, with ground terminal ends. Provenience: Level 6-7 feet—4; 7-8 feet—3; 8-9 feet—1.

Ground section of cowrie shell, probably unfinished ornament. Provenience: Level 4-5 feet—1.

Fragment of worked bone, use unknown. Provenience: Level 8-9 feet—1.

a. Marianas Red rim sherd. Lime-filled impression along top surface of rim. Provenience: Level 4-5 feet below surface., b. Marianas Red rim sherd. Lime-filled circles along top surface of rim. Provenience: Level 5-6 feet below surface., c. Marianas Red body sherd. Lime-filled circular impressions on outer surface. Provenience: Level 7-8 feet below surface., d. Variant form of Marianas Red rim sherd. Band of lime-filled impressions below rim on outer surface. Provenience: Level 7-8 feet below surface., e. Variant form of Marianas Red rim sherd. Band of lime-filled circles and inverted triangles below rim on outer surface. Provenience: Level 6-7 feet below surface., f. Marianas Red rim sherd. Band of lime-filled circles and hatched line below rim on outer surface. Provenience: Level 5-6 feet below surface., g. Marianas Red rim sherd. Band of lime-filled circles below rim on outer surface. Provenience: Level 6-7 feet below surface., h. Black ware sherds, with curvilinear lime-filled design on outer surface. Provenience: Level 5-6 feet below surface., i. Black ware sherd. Lime-filled decoration, probably stamped, on one surface. Other surface has red slip without decoration.,
- 325
a. Shell pendant or sinker. Tips broken off at points where suspension holes were bored. This type of artefact is relatively common at latte sites, in association with Marianas Plain pottery. Provenience: Level 5-6 feet below surface., b. Flat shell beads. Provenience: Level 4-5 feet below surface., c. Pink, chalk-like stone cylindrical pendant. Holes bored transversely through sides and ends. Provenience: Level 5-6 feet below surface., d. Fragment of shell bracelet. Provenience: Level 4-5 feet below surface., e. Fragments of shell bracelets. Provenience: Level 7-8 feet below surface., f. Ground shell ornament of Conus. Provenience. Level 4-5 feet below surface., g. Small basalt flake, with ground bevel at one end. Used as adze or knife. Provenience: Level 6-7 feet below surface., h. Fragment of small basalt adze. Provenience: Level 5-6 feet below surface., i. Fragment of median section of shell adze (Tridacna?). Provenience: Level 9-10 feet below surface., All objects are drawn to scale. Shell pendant (a) is 68 mm. long.,
  • ANSON, George (Lord), 1748. A Voyage Round the World in the Years 1740-44. Compiled by Richard Walter. London, John and Paul Knapton.
  • GIFFORD, Edward W. and Delila S., 1959. Archaeological Excavations in Yap. Berkeley and Los Angeles, University of California Press, Anthropological Records, Vol. 18, No. 2.
  • MCCARTHY, Frederick D. and SETZLER, Frank, 1960. The Archaeology of Arnhem Land. Sydney, Records of the American-Australian Scientific Expedition to Arnhem Land, Vol. 2, No. 5.
  • SPOEHR, Alexander, 1957. Marianas Prehistory: Archaeological Survey and Excavations on Saipan, Tinian, and Rota. Chicago, Chicago Natural History Museum, Fieldiana: Anthropology, Vol. 48.
  • THOMPSON, Laura, 1932. Archaeology of the Marianas Islands. Honolulu, Bernice P. Bishop Museum, Bulletin 100.
1   Anson 1748.
2   Thompson 1932.
3   Spoehr 1957.
4   Spoehr 1957.
5   Gifford and Gifford 1959.
6   McCarthy and Setzler 1960.
7   Spoehr 1957.