A COMPARISON OF SOUTH CHILEAN AND ECUADORIAN "FISHTAIL" PROJECTILE POINTS Junius B. Bird American Museum of Natural History In November of last year, during a brief visit to Fell's Cave near Magellan Strait, I obtained a carbon sample from the level of the first occupation. When processed by Isotopes Inc. the radio carbon age was reported as 11,000 ? 170 years (sample 1-3988). Another sample, ex- cavated by Mr. John Fell, from a different fire hearth but at the same level, had been processed in 1960 by the U.S. Geological Survey labora- tory: W-915, 10,720 ? 300 years. Both figures were computed using 5,568 + 30 years as the average C half life. When the more recently deter- 14 mined half life figure 5730, is used, the mean age figures are 11,320 and 11,032, respectively. If these figures are valid, then the associ- ated projectile points are, at this time, the oldest dated points known in South America. As such, their distribution and characteristics should be of some general interest. In 1937 I recovered fifteen whole and fragmentary examples at Fell's Cave and a single stem in Palliaike Cave (Bird 1946). In 1959 four more were found in Fell's Cave by Henry Reichlen and Annette Laming-Emperaire (Emperaire 1963). Subsequently, John Fell secured other examples, now displayed in a regional museum in Punta Arenas, Chile. All of these were found in association with bones of extinct horse (Para- hipparion saldasi), giant ground sloth (Mylodon listai) and guanaco. In 1952 Professor Osvaldo Menghin found what appear to be two broken fishtail points while excavating in a small cave on the Estancia Los Toldos (Menghin 1952a, 1952b). This is south of the Deseado River in Argentine Patagonia, 310 miles north of Fell's and Palliaike Caves. The only associated evidence of extinct fauna was a single horse tooth. The 52 points shown in Fig. 4-k, -l are of translucent chalcedony. The one remaining stem is certainly a poor example of the type and I suggest that it pertains to the group mainly because of its stratigraphic position at the beginning of occupation, its association with a discoidal stone of the same type as found with the Fell's Cave points, and the types of artifacts found in the overlying strata. Illustrated with the Los Toldos points is a comparable specimen, also of chalcedony, a surface find from the Department of Flores in Uru- guay and the property of Professor Antonio Taddei. The details of flaking and retouching are more readily visible after coating with ammonium chloride, Fig. 4-m. Had the Los Toldos items been similarly treated it would be clear that they, too, have been made from rather thin, large flakes; that form was achieved with a minimum of secondary chipping; and that one side of the stem has a basal flake scar 18 mm long by 10 mm wide. As with the Fell's Cave points, the sides of the stem on both the Los Toldos and Uruguay points have been ground or smoothed sufficiently so that the edges would not cut the spear shaft lashing. This smoothing ex- tends around the lower corners onto the basal edge. While one cannot prove that the Uruguay specimen is an example of the early fishtail type, it is sufficiently similar to warrant careful consideration. The one other locality in South America where the fishtail type point has been found is at the El Inga site, near Quito, Ecuador (Bell 1965; Mayer-Oakes 1963). The meticulous excavation by Dr. Robert Bell yielded two whole and nineteen broken fishtail points, plus several other pieces which might be grouped with them. In the great quantity of scrap, flakes and artifacts found on the surface were twenty-one more examples. These were not included in the report on the excavation and have been separately described by Dr. William Mayer-Oakes (Mayer-Oakes ms.). The site is on an open hilltop with such shallow soil deposit that no identifiable bones survived. Variable erosion, cultivation and probably other factors had mixed artifacts of different types and ages. 53 No fire hearths were encountered and only bits of charcoal scattered in the soil were available for radio carbon measurement. The oldest figure obtained, 9030 ? 144 years (R-1070/2), is not necessarily the age of the initial occupation. In comparing the Fell's Cave and El Inga points we should note first that the latter are with one exception all of obsidian while all but three of the Fell's and Palliaike Caves examples which I have seen are of a relatively fine textured basalt. Two of the three exceptions are of quartzite and one is of chert. There is enough difference in the nature of the obsidian and the basalt and quartzite, the way in which they fracture and respond to pressure flaking, that one would anticipate some distinction in the chipping or flaking patterns. Nonetheless when one compares the two lots they create an impression of close identity and cultural unity, much as the widely scattered Clovis and Clovis-like points do in North America. In this case the direct distance between the Inga and Fell's sites is 4300 miles. By present roads one must travel over 5500 miles and the route followed by wandering hunters would be appreciably longer. Where such distance is involved one can rightly question the relationship of the material. We have assumed some connection but without carefully assessing or questioning the evidence. What are the similarities and differences? Are they fortuitous and if not what do they signify? The following observations are an attempt to answer these questions. As a minor contribution to the Kroeber Anthropological Society's publication honoring John Rowe I regret that it is not more directly pertinent to the Peruvian record. If this early Ecuadorian and south Chilean material is related, then further evidence should be found within Peru in those zones where sloths, horses and probably mastodon were most abundant. Comparative comments can be listed as follows. Similarity 1, Nature of sites: Both lots come from habitation sites where the basal portions of broken points were removed from the 54 shaft slots to be replaced by new ones. The yield from such- sites can include points which have been discarded for other reasons than breakage from use,-namely, ones resharpened until they are no longer desirable, ones broken in production, and rejects that were not satisfactory. Such lots need to be evaluated against examples recovered at sites where game was killed, in 'this case not yet available. Similarity 2, Manufacture 1: At both sites the procedure of manufacture seems to have been identical. Instead of preparing an initial, bifacially shaped blank to be thinned and finally formed by secondary chipping the preference was to start with a large flake of approximately the same thickness as the finished product. This is more readily visible on the obsidian points than the basalt ones, yet is clearly common to both. In some cases secondary chipping has eliminated all but small areas of the original flake surface or surfaces. In extreme cases only a mini- mal amount of secondary chipping has been needed to achieve the form leaving most of the original flake surface unmodified. The Los Toldos and the Uruguay points fall in this category. The El Inga point, Fig. 4-g, and the Fell's Cave specimen, Fig. 5-a, are good examples. The latter, distinct in outline and a unique example, might conceivably be a simple preform, the initial stage in shaping. It was found transversely broken and has no marginal or basal smoothing nor any visible evidence of use. Others, not as extreme, are shown in Figs. 4-j, 5-k. With them the secondary chipping has produced good edges and form, yet has left intact proportionally large sections of the original flake surfaces. In this feature the South American fishtail points differ from all of the known point types associated with extinct game in North America. Similarity 3, Manufacture 2: Fluting, or relatively long basal flake scars, occurs in both lots but in totally different ratios. Among the excavated El Inga specimens four are fluted on both sides, six on one side while ten have no large basal flake scars. Among the Fell's and Palliaike Caves examples one has fluting on both sides, two on one side, 55 while fifteen have varying numbers of basal thinning flake scars up to 19 mm. in length. The breakage of some of these suggest that the maker hoped to remove longer flakes but could not achieve the desired result with the basalt and quartzite. I suspect that the lower frequency of fluting in the south is due almost'entirely to the nature of the materials available. Similarity 4, Manufacture 3: Marginal grinding of the stem edges occurs in both lots. On the Fell's and Palliaike examples it starts along the lower edge of the shoulders, continues down around the corners and across the concave base. The grinding is only enough to blunt the sharpness of the edges, not enough to modify the outline. It is detect- able by touch and is clearly visible under magnification. As the stem edges tend to be concave, the base lines even more curved and the pro- jecting corners rather easily broken, smoothing must have been done with a rather narrow-edged, fine textured piece of stone. I have not detected scratch lines to establish direction of stroke. The only one'of the southern group completely devoid of any edge smoothing is the one pre- viously mentioned, the exception to the usual form, Fig. 5-a. Professor Bell reports that all of-his El Inga examples have the side edges of the stem smoothed and that this continues across most of the concave bases. While we have never had the two lots together for direct comparison, my impression of El Inga edges is that the degree of smoothing is the same, never enough to modify outline beyond the elimi- nation of the most minute projections. Similarity 5, Form: The only easily measurable aspects of form are presented in Tables 1-5. Beyond these we can note that at both sites there appears to be great variation in length, considerable variation in blade width and greatest unity in stem outline and dimensions. One might describe the fishtail point as a barbless, stemmed form with and without fluting, with rounded shoulders, the stem tapering towards a concave base, the stem sides generally but not always terminating in slightly expanded, 56 rather sharp prongs or corners. Where these were present the stem sides tend to be concave in profile and minimum stem width occurs forward of or above the base. Stem width, which is less than stem length, must have been closely correlated with the diameter of the forward end of the spear shaft. Stem length is impossible to measure with certainty as it expands and curves into the rounded shoulders. Maximum blade width nor- mally occurs somewhere forward of the shoulders but without any standard place in proportional relation to the length of the point. All of these comments under Form apply equally well to the Fell's Cave and the El Inga specimens. One detail of the average form, the slightly expanding, pro- jecting corners at the base, would have had a functional justification if the diameter of the shaft tip was less than the span or width of the corners and either equal to or less than the minimum width of the stem. The projections would contact the binding and the point would be less apt to separate from the shaft when it was withdrawn from or slipped out of a wound. This would justify the smoothing of both the outer and lower margins of the corners. While on the subject of this detail I must caution those who make tracings or drawings of these points to be especially careful to render the basal prongs accurately. A simple tracing almost inevitably tends to enlarge or distort them and in subsequent renderings, without reference to the specimen, even greater discrepancy can appear. Similarity 6, Breakage: There is such marked similarity in the nature or pattern of breakage that a listing is warranted: Southern Chile Ecuador Total specimens involved 20 26 Number of breaks scored 13 27 Position of breaks: 1. Transversely across stems forward 3 9 of the narrowest portion 57 Southern Chile Ecua or 2. Transversely across blades in 5 4 the area from near the maximum width back to the shoulders 3. Two fractures in both of preceding 1 1 areas, i.e., two breaks per specimen 4. Transversely where stem widens to 1 31 shoulder behind forward turn of shoulder outline 5. Diagonally across blade forward of 1 3 shoulder 6. Diagonally across blade and into stem - 3 7. Basal corner section of stem 1 2 8. Tip missing (Four southern examples - 1 seem to have been resharpened after such breakage) Most if not all this breakage appears to have occurred while the points were attached to the shafts. While we cannot credit the similarities with much significance for comparative purposes, the nature and position of fractures must relate to the overall form, the relative thinness of the points and the manner of hafting. There appears to have been close identity in the manner in which the points were fitted and secured to the shafts. If the position of the transverse fractures across stems was related to firm support or sustaining pressure of the sides of the shaft slots the average measurements suggest that in the north the stems were firmly seated for a distance of 15.5 mm., in the south for 16 mm. Assuming the shaft slots overlapped the stems up to the midline of the shoulder flare, the average depth of the slots in both areas was 21 mm. While .the measurements cannot be precise the same rule applied to twelve southern and eleven northern examples yielded the same mean. Similarities, Comparison of measurements: Table 1 records the summary of stem angle measurements, Fig. 3-f, taken from photographs, most of them transparencies projected for better appraisal and the drawing 58 of angle lines. In some cases the angle can be measured with considerable accuracy. With others, where the stem edges curve, the lines must be drawn tangent to these curves and will therefore vary with individual interpretation. Three stems with little overall taper, those shown in Figs. l-g, -h, 5-f, are so difficult to measure with certainty that they have been omitted. One should note'that they are present both in the north and south. Tables 2 and 3 record the minimum stem widths and maximum blade widths. As in Table 1 the differences between the means expressed in the T-form and checked for probability indicate that there is a much closer similarity than would occur by chance. Tables 4 and 5 show similarity only in the relation of stem and blade thickness. In any one specimen the two measurements are nearly identical and this is evident in the averages. Maximum blade thickness is normally found along the medi'an line of the point in or near the area of maximum blade width. The stem thickness measurements were made as near as possible to the center of the stem. Beyond this the tables re- cord a clear distinction which must be related to the materials used. Evidently in the initial breakage to produce a flake suitable for forming a fishtail point, obsidian yielded a thinner flake than basalt. It is regrettable that the samples are not larger. Chipped stone artifacts from one site and period can differ for so many reasons that the range of variation and the nature of the standard form become really clear only when larger numbers are available. Perhaps we can con- strue this as lending weight to the results obtained from what we have. -I believe that the sum of the similarities makes a reasonable case for close connection and identity between the El Inga and Fell's Cave points. The few distinctions or differences mentioned either have reasonable explanation or are unimportant. It is more difficult to compare associated cultural material. At Fell's Cave the initial occupation debris was isolated from later 59 occupation by fallen rock slabs and sterile fill. At El Inga the points and other stone artifacts were so distributed that it is impossible to determine with certainty any specific assemblage. The associated, uni- laterally chipped scrapers for working wood and bone at Fell's Cave are identical with those from later periods so their occurrence at El Inga means little. There is also identity in large, variably sized skin scraper blades at both sites which may have more significance. One major difference is the frequent occurrence of burins and burin spalls at El Inga and their complete absence at Fell's Cave. Their plotted distri- bution suggests that the burins post-date the El Inga fishtail points so they cannot be cited as a distinction between the sites. The rarer items associated at Fell's Cave, bone chipping tools, unidentified bone objects and discoidal stones are all lacking at El Inga, so they too do not yet enter the comparative picture. There remains only to mention the occurrence of an apparent fishtail point from Madden Lake, Panama. The water in this reservoir is used to maintain the Gatan Lake level for the Panama Canal and fluctuates about thirty-two feet annually. The washed and exposed shoreline has yielded many chipped stone artifacts. Among them is a fluted point some- what Clovis-like in form and the piece mentioned which is described also as fluted. Both were found in years of unusually low water when lake level may reach within 210 feet of sea level (Sander 1959). While I have not seen the specimen, Professor Bell reports that a cast sent him bears strong resemblance to the fishtail form. We should keep it in mind when seeking further examples to compare with the Fell's Cave and El Inga series. 60 TABLE 1. STEM PROFILE ANGLES OF POINTS FROM EL INGA AND FELL'S CAVE (in minutes) El Inga 1 600 - 699 700 - 799 800 - 899 900 - 999 1000 - 1099 1100 - 1199 1200 - 1299 1300 - 1399 1400 - 1499 1500 - 1599 Fell's Cave 2 1 1 2 3 4 2 2 3 1 1 Total Samples 10 13 Means (Calculated from ungrouped data) El Inga 958' = 150 58' t = 0.013 Fell's Cave 970' = 16- 10' Probability > .25 61 TABLE 2. MINIMllM STEM WIDTH OF EL INGA AND FELL'S CAVE POINTS (in mm.) Fell's Cave 1 1 2 2 4 2 1 Total Sample 18 Means (Calculated from ungrouped data) El Inga Sample 15. 5 mm.. t = .041 Probability > .90 Fell's Cave Sample 62 11.0 12.0 13.0 14.0 15.0 16.0 17.0 18.0 19.0 20.0 21.0 11.9 12.9 13.9 14.9 15.9 16.9 17.9 18.9 19.9 20.9 21.9 El Inga 2 4 2 3 2 2 2 1 13 15.5 mm. TABLE 3. MAXIMUM BLADE WIDTH OF EL INGA AND FELL'S CAVE POINTS (in mm.) Fell's Cave El Inga 1 1 1 1 1 1 1 1 Total Sample 8 Means (Calculated from ungrouped data) El Inga Sample t = 0.239 28.4 mm. Fell's Cave Sample Probability .80 < p < .90 63 1 19.0 20.0 21.0 22.0 23.0 24.0 25.0 26.0 27.0 28.0 29.0 30.0 31.0 32.0 33.0 34.0 35.0 36.0 - 19.9 - 20.9 - 21.9 - 22.9 - 23.9 - 24.9 - 25.9 - 26.9 - 27.9 - 28.9 - 29.9 - 30.9 - 31.9 - 32.9 - 33.9 - 34.9 - 35.9 - 36.9 2 2 1 2 3 1 1 2 15 29.0 mm. TABLE 4. MAXIMUM BLADE THICKNESS OF EL INGA AND FELL'S CAVE POINTS (in mm.) Fell's Cave 2 1 5 2 6 6 Total Sample 8 14 Means (Calculated from ungrouped data) El Inga Sample 4.6 mm. Fell's Cave Sample t = 3.122 Probability < .005 TABLE 5. MAXIMUM STEM THICKNESS OF EL INGA AND FELL'S CAVE POINTS (in mm.) El Inga 2 3.0 - 3.9 4.0 - 4.9 5.0 - 5.9 Fell's Cave 8 6 6.0 - 6.9 7.0 - 7.9 1 Total Sample 17 Means (Computed from ungrouped data) El Inga Sample t = 3.611 Probability < .005 4.59 mm. Fell's Cave Sample 64 3.0 - 3.9 4.0 - 4.9 5.0 - 5.9 6.0 - 6.9 5.6 mm. 3 7 5 2 17 5.66 mm. NOTES 1 One in each lot has two intersecting lines of fracture forming an angle. 2 An article on the discoidal stones from southern South America will appear shortly in American Antiquity. REFERENCES Bell, Robert E. 1965 Archaeological Investigations at the Site of El Inga, Ecuador. Casa de la Cultura Ecuatoriana, Quito. Bird, Junius B. 1946 The archaeology of Patagonia. In J. H. Steward, ed., Handbook of South American Indians 1:17-24. Bureau of American Ethnology, Bulletin 143, Washington. Emperaire, Jose, Annette Laming-Emperaire, and Henry Reichlen 1963 La Grotte Fell et autres sites de la region volcanique della Patagonie Chilienne. Journal de la Societe des Americanistes, nouvelle serie, 52:167-254. Paris. Mayer-Oakes, William J. 1963 Early man in the Andes. Scientific American 208:117-128. mis. El Inga projectile points. Mimeographed, April 10, 1964. Winnipeg, Manitoba, Canada. Menghin, Oswaldo F. A. 1952a Las pinturas rupestres de la Patagonia. Runa 1:5-22. Buenos Aires. 1952b Fundamentos cronologicos de la prehistoria de Patagonia. Runa 1:23-43. Buenos Aires. Sander, Dan 1959 Fluted points from Madden Lake. Panama Archaeologist 2:39-51. Panama. 65 Fig. 1. Fishtail type points, excavated at the El Inga Site, Ecuador. a. Chalcedony. b-h. Obsidian. Sections approximate. R. E. Bell, 1965. a d b C e 7ZMLT 9 b f c d 4zZLZTI: g, h i 0 k I m 0 5cm FIG. 2 68 I -.1 Fig. 3. Fishtail type points. a-e. El Inga, Ecuador, surface f inds. a-b. Broken and resharpened. W. Mayer-Oakes ms. f. Fell's Cave, Period 1, showing angle of taper. Length 55 mm. (See Fig. 5-o.) g. Madden Lake, Panama. D. Sander, 1959. I I~~ Eut d x4zy b c 9 (I)~~~~~~~~~~~~~~ 1-I .1 ~ ~~~~~~~~~~~~~~~ I e 0 5cm FIG. 3 69 A