been classed as late protohistoric sites and are not considered within the scope of this paper. Several sites, e.g., LAn-17, were of dubious mixed origin and would require more extensive examination to determine their cultural affinities. Where omissions in the numbering occur, e.g., 7, it is because sites were reported for an area and later failed to materialize as much.

FIELD TECHNIQUES

Our major purpose in continuing field work at the Tank Site was to establish with greater certainty the relationships between the diverse artifact types and classes, and other manifestations, already recognized. In the hope that the general region in which burials had been located in 1947 would continue to be productive in this respect, we expanded from there in all directions, especially toward the center of the mound. Digging in the deeper northwestern part of the site was furthered with the intent of verifying the suggested stratigraphy and acquiring, possibly, a deep undisturbed burial in better condition than those from the upper soil horizons.

The procedure of excavation and notation was essentially unaltered from that previously employed. A grid of coördinates had already been established with reference to permanent data. Burials and features were again entered on standard University of California archaeological forms. A slight change, however, was made in the method of recording and cataloguing field data.

Originally a data sheet had been completed for each 6-inch interval of a 5-foot grid section, on which artifacts were plotted in exact horizontal location. In working up the material it became clear that the specific spatial distribution of isolated implements lacked patterning. It was therefore considered adequate, when returning to the field, to designate provenience by excavation unit and level only. In addition, the method of cataloguing was simplified and so organized that 90 per cent of the tabulation of data could be completed in the field. This was possible because the specimens derived from the 1947 field work had already been classified and constituted a sample on which expectations could reasonably be based.

The procedure followed was to strip each 5-foot section in 6-inch levels, and to sack together all the artifacts from one such test unit. At the end of the day the level bags were taken to camp where the artifacts were washed, labeled, and tabulated. All items were marked in India ink according to section number and level interval, e.g., 15R10-1, a specimen from the 0- to 6-inch level of the pit; 15R10-2 would indicate the 6- to 12-inch level, etc. A tabulation sheet was kept for each excavation unit. This sheet listed the most frequently occuring types or categories, allowing for the notation of rare forms, and was ruled vertically to indicate depth intervals. Artifacts were entered according to type, or category, and level, and then packed for transport. Atypical specimens or those to be used for illustration were set aside for separate shipment and more intensive examination.

This system had many advantages. Records were readily kept up to date, problems that suggested themselves as excavation progressed could be more closely defined and investigated, and artifacts could be expeditiously and finally cleared from the work area. The data sheets served as a field catalogue and covered the groundwork of the final statistical compilation. The number assigned each specimen referred not only to its catalogue entry but also its provenience.

In 1947 we could not anticipate what might be found, nor could we establish immediately the significance of what we did encounter. Thus it has been our policy to save all worked stone and ship it back to the Museum of Anthropology at Berkeley for study. During the second season, however, we felt a little more discrimination was warranted in order to save the museum valuable storage space. Therefore, the bulk of the hammerstones and a number of complete, and all fragmentary, manos, metates, and scrapers were tabulated and piled into pit 21R4 before backfilling.

Map 2. Tank Site LAn-1

FEATURES

Since the Tank Site showed promise of being an unusual and important deposit, considerable care was observed during both seasons of excavation to isolate and expose any concentration of lithic remains which appeared to be in any way atypical of the average mound matrix. As a result, numerous associations of stones, such as mano caches, highly weathered inverted metates, and massive piles of rejected cores, broken manos and metates, and plain cobbles, were set apart from the rest of the site and given the term “feature.” In some instances these features possess obvious meaning, as was true of the mano caches and the inverted metates, but in other instances the purpose remains unknown. If nothing else, this technique of isolating features as excavation progressed provided an adequate view of the internal structure of a village, a type of information largely lacking in southern California archaeology.

The features described below represent a continuation of the series reported for 1947.

Feature 14 (pl. 19, b).—Cache of 4 manos. The placement of these specimens precludes a chance affinity; they were closely grouped and each was standing more or less on end. No other artifacts were found in association.

Feature 15 (pl. 19, d).—Owing to its areal extent, feature 15 is somewhat difficult to define. The complex of stone by which it is characterized has been arbitrarily broken down for descriptive convenience. There is no way of knowing whether the entire complex exemplifies a single unit or if in the course of time it merely developed from a single point of departure.

Feature 15a.—This was 10 by 10 feet with an average depth of 4 to 8 inches. Four inverted metates, additional metate fragments, manos, core tools, and a single fragment of a slate pendant. In the southeast portion were 12 symmetrical stream cobbles of different sizes. This latter aggregation is of interest. The almost perfect symmetry of the stones suggests selection, and the physiographic location of the Tank Site implies such stones must have been transported to it. None shows any evidence of utilitarian use, and in the light of present knowledge the existence and function of such objects cannot be explained.

Feature 15b.—Badly weathered, fragmentary metates; altered “lumps” of sandstone; manos; and core tools. Three small pestles were found near association.

Feature 15c.—Inverted, killed, sandstone metate; sandstone slab; and core tools. Burial 11 was in close proximity, but owing to its badly disturbed condition, no positive association could be made with the feature.

Feature 15d.—Metates; altered sandstone blocks; core tools; manos; and fragments of human leg bones.

Feature 16.—Disintegrated, pitted metate and fragments of 2 other metates; mano fragments; core tools; unworked stone; and unidentifiable fragments of human bone.

Feature 17.—Characterized by a number of symmetrically water-worn cobbles. Contrasts with other features wherein metates, manos, cores, and irregular, unutilized stones predominate. Fragmentary metates, manos, and core tools were also present, as were 2 segments of human femora.

Feature 18.—Badly weathered, inverted, deep-basin metate in near association to symmetrical sandstone cobble containing a ground depression. The latter may represent the initial stage of mortar manufacture, though the smoothness and regularity of the depression surface somewhat invalidates the idea.