...

This discussion is in large part based on Kuper's excellent treatment of nineteenth-century evolutionism ([Kuper 1988]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Belaboring the point, we might add that Darwin appears in neither the index nor the bibliography of Steward's Theory of Culture Change ([Steward 1955]). Neither does he appear in Service's Primitive Social Organization ([Service 1962]), and he is mentioned only once in Origins of the State and Civilization to quote a short passage on social equality "retarding" civilization ([Service 1975]: 50).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Marx appears on fewer pages even than Darwin: exactly one.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It was while explaining this scheme that Sahlins and Service coined the sublimely absurd phrase ``thermodynamic accomplishment'' ([Sahlins and Service 1960]: 33).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Kup88).

I will not be concerned here with the various distinctions that have been made between cultural evolution, social evolution and political evolution. For the purposes of this essay they are taken to be synonymous.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Carneiro provides an illuminating discussion of the history of the chiefdom concept in anthropology ([Carneiro 1981]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

White said that Morgan's primitive and civil societies are logically distinct categories, "... just as reptile and mammal are" ([White 1959]: 302). Lowie called this ``palpable nonsense.''

Bourdieu: ``Nothing is more misleading than the illusion created by hindsight in which all the traces of a life ... appear as the realization of an essence that seems to pre-exist them'' ([Bourdieu 1990]: 55).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Interestingly, the phrase `primitive communism' can be traced to Morgan, who called it ``communism in living'' (see [Lee 1990]: 232).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It is sometimes startling how effortlessly such phrases as `primitive religion,' `primitive economies' and so on seem to flow from Mann's pen.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Yof93).

White provides perhaps the most succinct example of such a `world growth story': ``The crest of the wave of cultural development has been moving westward ever since the Pyramid Age" ([White 1959]: 369).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

``A formula that explains the behavior of all mankind cannot explain culture" ([Steward 1955]: 8).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The modern lake level average is approximately 3810 m.a.s.l. Thus, the peaks of the Taraco Hills are rarely more than 200 meters above modern lake level.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Trompos are also produced from quartzites, andesite, and a wide variety of other materials. However, in the Taraco Peninsula sites tuff is the most common material. This may be a clue to their function, since these tuffs are extremely low density, and the resulting artifacts very lightweight.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

As does the upper surface of the Taraco Formation. See my description of the Taraco Formation above.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Wheat and barley are, of course, European introductions.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This discussion draws heavily on my published account of the history of investigations at the site of Chiripa ([Bandy 1999a]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I participated in all of these seasons; and as co-director since 1998.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

What Kohler and Blinman call a ``calibration data set'' ([Kohler and Blinman 1987]: 2). I am aware of two previous attempts to mathematically separate mixed archaeological assemblages. One ([Kohler and Blinman 1987]) employed multiple linear regression to separate mixtures of six phases each defined in terms of relative frequencies of 18 ceramic types. Their case is considerably more complex than my own. The other attempt ([Stahle and Dunn 1982]) used regression to estimate the contribution of various stages of biface manufacture to a mixed debitage assemblage.

My own solution to the mixture problem relies on the computational brute force made available by recent personal computers, rather than on the more elegant statistical analyses of earlier researchers.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Frequency profile analysis is superficially related to standard seriation practice ([Marquardt 1978]), but only superficially. Kohler and Blinman: ``Unlike seriation, which would normally force a mixed collection into a single position on a linear continuum usually interpreted as time, this approach identifies mixing, if present, and apportions the collection into one or more temporal components'' ([Kohler and Blinman 1987]: 2-3).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

See [Steadman 1999] for a much more detailed account of Chiripa ceramic chronology.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...bandy1999a:24).

This ``nucleated'' vs. ``dispersed'' distinction is of course not a new one. Drennan, for example, captures the same dimension of variability in his ``compact'' vs. `` dispersed'' settlement patterns ([Drennan 1986]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

``The Aymara household, although variable, typically consists of a nuclear or extended family living together in an architecturally defined compound, or at least near one another in a nucleated compound cluster. Household members ... ideally share a common patio and other outdoor spaces ...'' ([Janusek 1994]: 35).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Wis93-a:109).

The record from the Juli-Pomata area for this time period is somewhat better understood ([Stanish et al. 1993,Stanish et al. 1997]). However, the sites from the western basin may not be directly compared to the southern case. This is because the basic organizational unit of LIP domestic groups in the western basin is the domestic terrace, and terracing is conspicuously absent in the South.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The specific conditions of the southern Titicaca Basin - relatively high rainfall, soil erosion, intensive agriculture, high population density and the use of adobe for domestic architecture - conspire to make the identification and excavation of prehistoric residential structures an extremely difficult undertaking.

Ironically, the best-documented Tiwanaku architecture is not from the altiplano at all, but from the coastal Moquegua Valley. Here the Tiwanaku state established a series of colonial enclaves, and the architecture at these sites is relatively well-preserved (cf. [Goldstein 1989a,Goldstein 1993a]). However, I believe that these colonies were populated by a rotating labor force, and not by long-term inhabitants (see [Bandy et al. 1996]). Therefore, these households are not in any way comparable to altiplano ones, and cannot be used for my purposes here.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Bermann ([Bermann 1993]: 125) contends that the household compound came into being in the Late Formative, and that in earlier phases people lived in single-structure domestic units. The data for this are extremely tentative, however, and I consider the question open.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

There are some indications that in certain times and places in prehistory the amount of habitation area per compound was greatly compressed. Bermann, for example, excavated two patio groups at Lukurmata which averaged only 0.02 ha (160 m$^{2}$) of total occupation area each ([Bermann 1994]: 178, 194). However, he takes no account of public space, walkways or roads in these figures. The fact remains, though, that using a constant figure for habitation area per household for all periods in prehistory may well have the effect of decreasing population estimates of urban or nucleated habitation contexts as opposed to rural or dispersed ones. Though the problem is at present without a solution, this fact should not be forgotten in the analysis to follow. Bermann's estimate of 376 persons in a 0.60 ha area ([Bermann 1994]: 178), however, is almost certainly greatly exaggerated.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Excepting non-habitation space occupied, for example, by public architecture, plaza space, and so on. For present purposes this complication will be ignored, though I acknowledge the difficulty of identifying habitation vs. non-habitation space in sites without preserved surface architecture.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The actual SQL command used in constructing the database (see schema in Appendix F) is given below. All table and column names are references to the database schema documented in Appendix F.

UPDATE site SET area_corrected=((sqrt(area*10000)-20)*(sqrt(area*10000)-20))/10000 WHERE area >= 0.25;

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

UPDATE site SET area_corrected=0.09 WHERE area > 0 AND area < 0.25;

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

UPDATE site SET area_corrected=area WHERE type IN ('Raised Fields','Terraces');

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Sites which have accumulated over very long time periods are also palimpsests ([Foley 1981]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Therefore:

The SQL statement used combines the steps from sections 4.1.4 and 4.1.5 for habitation sites only:

UPDATE site SET pop_index = (area_corrected / 0.09) * 6 WHERE type NOT IN ('Raised Fields','Terraces');

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 13.1).

Incidentally, 50 persons/ha seems to be a comfortable settlement density worldwide (see [Fletcher 1995]: 75, 107).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

My own calculations were originally carried out in the field using an equation I derived without benefit of reference material. Though unwieldy, it is equivalent to Hassan's.

$r=\left(\frac{P_{f}}{P_{i}}\right)^{\frac{1}{T}}-1$

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...carneiro-hilse-66,cowgill-1975).

This is contra Schlanger, who estimates that the intrinsic growth rate of Colorado Plateau Anasazi populations was on the order of 2.4% annually ([Schlanger 1985,Schlanger 1988]: 786). Her estimate, equivalent to a doubling time of less than 30 years, is almost certainly much too high. It is roughly the same as the modern world population growth rate.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Ser62,Sah58,Sah63,Fri67).

This was a development of Polanyi's concept of redistribution ([Polanyi 1944]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I used Surfer and Scigraphica. Many GIS and mapping programs can do this, however.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

As always, see [Steadman 1999] for a detailed description of the Chiripa ceramic phases.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... rare

Steadman reports only a single red on unslipped brown sherd from two seasons of excavation at Chiripa ([Steadman 1999]: 64). Several more sherds were encountered in the 1998/1999 seasons (Steadman pers. comm.) and in my surface collections of the site of Alto Pukara (T-430).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Nor has any been associated with any Early Chiripa site in the southern Titicaca Basin.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Amanda Cohen and Bill Whitehead.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

However, a brief reference in Padre Pedro Marabini's 1920 article in the Boletín de la Sociedad Geográfica de La Paz may also refer to Chiaramaya. Writing of Chiripa, he noted in passing another ``cerrito'' about 2 kilometers away, in the direction of the lake. Since the lake seems to have been very low at this time, he may have been referring to Chiaramaya, which is located 2 km to the east of Chiripa ([Marabini 1920]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Eri75).

He was a student at the time, working on Browman's project in Chiripa.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Many of the details of early construction at Chiripa were revealed in the 1998/1998 field season. These have yet to be published, though we did submit a fairly comprehensive report to the National Geographic Foundation ([Hastorf et al. 1998]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This almost certainly follows the course of the prehistoric road since it intersects virtually every major prehistoric site on the peninsula and in the Katari Basin.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Figure 5.4 is a variation on the conventional site size histogram. The difference is that the vertical axis does not represent the count of all sites in that size range, but rather the sum of the population index of these sites. This chart displays the distribution of population across the various size ranges.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...]).

Yanapata (T-130) is the exception, with the much faster growth rate of 0.65%. It may have experienced a modest influx of immigrants from Cerro Choncaya/Sunaj Pata.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Early Chiripa: 186; Middle Chiripa: 481.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Early Chiripa: 230; Middle Chiripa: 464.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Interestingly, this is very close to Fletcher's first C-limit of 1-2 ha for the transition from mobile hunter-gatherers to sedentary villages ([Fletcher 1995]: 89-90). The villages were 3.0 ha (T-2, 2.39 ha corrected) and 3.5 ha (T-3, 2.75 ha corrected) respectively before they fissioned. The Early Formative is the time that this C-limit was overcome, so some village fissioning would be predicted by Fletcher's model.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Browman's most explicit formulation of this thesis has been in unpublished essays and conference papers. However, there are clear references to a Chiripa polity in published sources (c.f. [Browman 1981]:414).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... mf1l).

The figurine illustrated in Figure 6.1l is very interesting. It can be considered a pun or joke using ceramic technology as its medium and referent. Its eyebrows are formed by use of more dense than usual fiber temper. Its eyes and ``turban'' are formed by careful placement of the angular quartz chunks which are characteristic of most Late Chiripa ceramics. Thus, the normal tempering agents used during this phase become design elements, and are foregrounded by their careful arrangement. To one familiar with Late Chiripa ceramics, this figurine is really quite witty and surprising.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Bro78b).

See my section in [Hastorf et al. 1998] for some discussion of post-Upper House Level mound stratigraphy.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... erroneous

Karen Mohr-Chávez recognizes the outlier status of these dates explicitly ([Mohr 1966]: 162).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The contemporary Early Sillumocco phase in the Juli-Pomata area has been dated to 900-200 B.C., though solely on the basis of stylistic cross-dating and a single radiocarbon date from Tumatumani ([Stanish and Steadman 1994]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This discussion is largely a reprise of a presentation made by myself and colleagues at the 1998 meetings of the Institute for Andean Studies in Berkeley, CA ([Bandy et al. 1998]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Kid56).

For a detailed discussion of pre-TAP excavations at Chiripa, see [Bandy 1999a].

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... another

It should be noted that the wide error ranges of the dates in question make it impossible to have total confidence in this scenario. It remains, nevertheless, and intriguing possibility. See [Bandy 1999b] for more on the problems of dating these structures.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Browman recognizes but does not emphasize the fact that the Upper House level platform is located on a terrace and not on a true mound:

The temple was placed in the center of a larger mound... The larger mound measures 50 m on a side; three sides on a downhill slope were faced of and revetted with a fieldstone wall up to 3 m high. ([Browman 1981]: 414)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

And with this I hope to lay to rest Conklin's erroneous reconstruction drawing of the Upper House complex ([Conklin 1991]: 287-8) which has been popularized in Moseley's generally excellent textbook ([Moseley 1992]: Figure 59) and elsewhere ([Conklin and Moseley 1988]: Figure 5.6). I have the utmost respect for Conklin's substantial contributions to Andean studies, but this particular reconstruction is clearly incorrect.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Several of the Llusco structure's walls are too badly destroyed to allow a precise determination of its plan. I believe it to have been rectangular, however.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

See my brief discussion in section 6.1 (also [Steadman 1999]: 66).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Upper House burials were excavated by Bennett ([Bennett 1936]), by Portugal Zamora ([Portugal Ortíz 1992]), and possibly by Kidder and Cordero Miranda in House 5 ([Layman and Mohr 1965], see Table 1).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The exception here, of course, is Janko Kala (T-394). This is one of the four principal Late Chiripa sites on the peninsula, and yet it has no clearly discernible corporate constructions. The surface of the site is heavily disturbed, however. It may be that more careful topographic work will reveal features too subtle for detection by the cursory investigation which I was able to undertake.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I would like to thank Sergio Chávez (pers. comm. 1999) for helping to show me the importance and wide distribution of this rock type.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... phases

As identified by Lee Steadman (pers. comm. 2000). Note that the table only tallies artifacts from ceramically unmixed proveniences; the large majority of RMT 61 and 65 artifacts were recovered from plow zone and from mixed fill.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Seddon reports that a ``gray andesite'', quite possibly the same material I am discussing here, makes up upwards of 90% of the lithic sample at the Early and Late Sillumocco and Tiwanaku site of Tumatumani, near Juli ([Stanish and Steadman 1994]: 70). The mixed nature of the excavated fill deposits precludes any determination of when the material appeared and/or disappeared at the site.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Recall that this rock is found in the form of hoes, mostly broken. Since the ``use locus'' of hoes is agricultural fields, we can expect that much of the breakage and discard of these tools took place off-site, outside of the village. The inhabitants of Sunaj Pata in the MF were almost certainly farming the area that was later to be the village of Kumi Kipa. It is therefore not surprising to find broken hoes in that area. Indeed, they are found over the entire peninsula, on- and off-site.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The total weight of RMT 61 and 65 excavated from Chiripa by TAP is 5361.8 g. This compares to 87.1 g of obsidian, 108.6 g of sodalite and 5.3 g of shell, as detailed earlier.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... pelts

The skirt on the Ponce Monolith in Tiwanaku ([Ponce Sangines 1995]: Figures 146-148) and the Bennett Monolith currently in La Paz ([Bennett 1934]) almost certainly represent spotted feline pelts. Many Pukara textiles also depict pelage markings using crosses, dots, or rhomboid shapes ([Conklin 1983]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I refer here to the abandonment of Sonaji (T-271) at the beginning of the Late Chiripa phase, and the simultaneous reoccupation of Sunaj Pata (T-268) and Cerro Choncaya (T-2). This event is clearly evident in Figure 6.11.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Which is to say that Roland Fletcher's first C-Limit had been successfully overcome ([Fletcher 1995]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I touched upon this topic in Chapter 4.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The exception in Janko Kala (T-394). It is exceedingly interesting that it is also the only one of these four large sites which lacks clear evidence of Middle Formative public architecture.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

These first three points basically define what Chávez and Chávez have called 'the Yaya-Mama Religious Tradition'' ([Chávez 1988]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

If I have not stated it explicitly enough, I will do so now: Chiripa (T-1) was not the seat of a complex chiefdom or multi-community polity in the Late Chiripa phase, or at any point in its history. There is positive evidence to the contrary. This was an early interpretation which should now be laid to rest.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Ben34).

In discussing Late Formative ceramic chronology, I will closely follow the work of John Janusek (see especially [Janusek 2002] for a recent statement; also [Janusek 1994]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Ponce apparently never made this association himself. Rather, it seems to have been made by Luís Lumbreras ([Lumbreras 1974]: 143-144; see [Janusek 2002]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Janusek and Lémuz have apparently developed substantially similar Late Formative chronologies independently, Janusek using his materials from Tiwanaku and from the Katari Basin, and Lémuz using his data from the Santiago de Huatta Peninsula.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...janusek2001).

Indeed, Janusek has suggested precisely such a division ([Janusek 2002]). I have found his formulation to be difficult to apply to surface ceramic assemblages, however, and for this reason make no use of it here.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I should note that both Janusek and Lémuz have inspected my type collection of this paste and agree that it dates to the LF1.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Figure-Ceramics-LF-2b-e).

Comparing the illustration in Figure 7.2 to those provided by Mathews for his ``Early Formative Lateral Banded Incised'' group ([Mathews 1992]: Figures 3.11-3.13b), I am forced to conclude that his LBI ceramics date, in fact, to the LF1 and that his column at T'ijini Pata (TMV-79) was at least partially stratigraphically inverted.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Figure-Ceramics-LF-1r-s).

These sherds are distinguished from later Tiwanaku wares by the ``fugitive'' quality of the red slip and by a very fine, soft and light-colored paste.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It is my belief that Burkholder's Huchani A, B and C styles all belong to the LF2 period. In her dissertation, Burkholder places the Huchani materials from ``some time after 1000 B.C. and ... before AD 600'' ([Burkholder 1997]: 172, also 168, 170). The early end of this range is owing to a radiocarbon date ($925\pm 85$ B.C., calibrated) obtained by Albarracín-Jordan in his earlier excavations (see [Albarracín-Jordan 1992]: 148, [Albarracín-Jordan 1996a]: 133). It has since become apparent (though not yet in print, unfortunately) that this date is an anomalous outlier, and that the initial occupation of Iwawe actually dates to somewhere around 300 A.D. If this is true, then the Huchani ceramics are LF2 in date, as indeed their stylistic attributes would suggest.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... 159).

Curiously, however, no dark, micaceous paste is included in Burkholder's catalog of wares from the site of Iwawe ([Burkholder 1997]: Chapter 5). Since LF2 levels were clearly excavated (the Huchani styles, [Burkholder 1997] Figures 6.8, 6.9, 6.10, 8.1, 8.2), as witness finds of Qeya polychrome ([Burkholder 1997] Figures 6.12, 8.4) and incised sherds, this must remain for the moment an unexplained anomaly.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... data

Janusek's data agree with my own observations of the ware's low frequency.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Table-Basalt-surface-density.

The data available on exchange in the LF are extremely sparse. No firm conclusions can be reached in this regard without further research, specifically designed to study regional exchange. However, the suggestions offered here are certainly provocative, and will hopefully inspire more fieldwork.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The Santa Rosa group has collective population index values of 277 for the MF and 1216 for the LF1.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This population relocation was quite possibly associated with a lake level drop around 100 A.D.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... sites

Which include Chiripa Pata (T-4), Waka Kala (T-421), Alto Pukara (T-430) and what is now the island of Sikuya (T-319/T-320). At the time, of course, Sikuya was not an island, but a lakeshore site, the lake level being considerably lower than it is today (see Figure 7.5).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This site is discussed in considerably more detail earlier in this chapter.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

In fact, I believe I encountered the center of such a polity in 2000 while traveling with Kirk Frye south of Yunguyu. At the site of Kanamarka/Lakaya (see [Stanish et al. 1997]:92-93 for a brief description and photographs of part of the site). The LF1 component at the site seems to be at least 12 ha in area. An effigy mound, in the form of a stylized catfish (approximately 60X25 meters), was evident on the surface, and zoned-incised Kalasasaya ceramics were observed by myself and my companions. This would be the center of Stanish's ``southern Ccapia polity''.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This figure was produced by John Janusek using the soundings of Bennett, Kidder and the Wila Jawira project to define the boundaries of the LF2 ceramic distribution in the Tiwanaku urban core. It is obviously tentative.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I should emphasize that this estimate is anything but precise. However, it does underscore the tremendous importance the incorporation of the Taraco Peninsula had in the process of Tiwanaku urbanization and state formation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The site and its sculpture were described by Alfred Kidder II, and more recent work has been done by Kirk Frye ([Frye and Steadman 2001] ;[Kidder 1943]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

There was probably a period of intense competition between and within the Taraco Peninsula communities during the first decades of the LF1. Certainly, there was time for the community of Chiripa to construct a larger and more elaborate mound, sealing - and fortuitously preserving - the Upper House Level complex. This probably would not have happened had it been immediately subjugated to Kala Uyuni.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

These trends are discussed in detail earlier in this chapter.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

There do appear to be several discontinuous MF occupations within boundaries of the later city of Tiwanaku (Blom, Janusek pers. comm. 2001). Little is known of them, however, and they probably represent hamlets or small villages.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... floor

Virtually all EF and MF occupation in the Tiwanaku Valley was on the slopes of the flanking hills.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

If the principal occupations of Kala Uyuni and the Santa Rosa group are indeed sequential. The issue is addressed earlier in this chapter.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Other factors were no doubt of equal or greater importance. Most significant, perhaps, was the rise of raised field agriculture and its integration into strategies of statecraft and surplus extraction. The Taraco Peninsula is entirely unsuited to raised field agriculture and the Tiwanaku Valley well-suited (see [Bandy 1999c]; [Bandy 2001]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The combined phase population index value of the Lower and Middle Tiwanaku Valley for the LF1 is around 1100, as opposed to 5581 for the Taraco Peninsula. The phase population index of the Upper Tiwanaku Valley for the LF1 could not have been more than another 500-600, since there are very few Formative Period sites in that part of the valley, and the largest one apparently covers only 1.5 ha ([Albarracín-Jordan et al. 1993]: 77). These figures were produced using data collected by Albarracín-Jordan and his colleagues (Lower Tiwanaku Valley: [Albarracín-Jordan 1992], Middle Tiwanaku Valley: [Mathews 1992], Upper Tiwanaku Valley: [Albarracín-Jordan et al. 1993]), and must be considered approximate.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The only significant change in military technology in prehistory seems to have been the introduction of the bow and arrow, which Bruce Owen has shown to be both earlier and less important than had been imagined ([Owen 1998]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Some classes of ceramics, typically the more elaborate and ornate forms, continued to have a restricted distribution.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Very similar forms, though with rounded bottoms and slipped orange, yellow or a lighter red, exist in the following Early Pacajes phase. See Chapter 9.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I do not pretend to offer a comprehensive review of Tiwanaku ceramics here. Many such treatments exist (e.g. [Alconini Mujica 1995,Bennett 1934,Burkholder 1997,Goldstein 1985,Goldstein 1989a,Ponce Sangines 1981]) and it is to these that I refer the interested reader.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The Proyecto Wila Jawira has also produced the beginnings of a new Tiwanaku Period chronology, primarily through the work of Alconini and Janusek (see [Alconini Mujica 1995]). This new chronology has yet to be applied successfully to mixed surface assemblages, however, or to settlement research (see [Janusek 2002,Janusek and Kolata 2002]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Thanks to James Mathews for pointing out the importance of this passage ([Mathews 1992]: 128).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... questionable

``... the distinction between uncontextualized Tiwanaku IV and V plainwares is a difficult (if not impossible) task at present, we are unable to ascertain the relative intensity of each phase of occupation at these sites with both types of decorated wares.'' ([Mathews 1992]: 132)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Tiwanaku grew at an approximate annual rate of 0.30%, as opposed to about 1% annually during the LF2. It should be remembered, though, that this is still a quite rapid growth rate and probably still implies considerable immigration from subject regions.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Titicachi being, of course, the new location of the major Formative Period village of Janko Kala.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It also seems to have been a politico-administrative boundary in the Early Colonial period, as I discuss in Chapter 10.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The three basins to which he alludes are the Tiwanaku Valley, the Katari Basin to the north (Pampa Koani), and the Jesus de Machaca/Khonko Wankani area to the south. I would add the Taraco Peninsula as a fourth component of the Tiwanaku heartland.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Other areas of the Titicaca Basin under Tiwanaku administration were not so radically changed as the core region itself. An example is the Juli-Pomata area, where the basic Late Formative settlement configuration was unchanged in the Tiwanaku period ([Stanish 1994,Stanish et al. 1997]). The LF settlement system was already oriented to raised field production, and it continued to be so during the Tiwanaku period.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I should also note that Stanish, in his report on a recent reconnaissance in the Desaguadero area, has identified a Late Intermediate Period style he terms Kelluyo. Kelluyo ceramics, as he illustrates them, are identical to Early Pacajes ceramics as described here ([Stanish et al. 1997]: 46-47, 107-111 and Figures 22 and 82), the inclusion of a solitary Late Pacajes sherd ([Stanish et al. 1997], Figure 82, leftmost) notwithstanding. The term Kelluyo will not be employed in the present study.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I refer to their joint report ([Albarracín-Jordan and Mathews 1990]), as well as to their respective dissertations ([Albarracín-Jordan 1992,Mathews 1992]) and subsequent publications ([Albarracín-Jordan 1996a]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Rydén also made this observation ([Rydén 1947]: 160).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Undulating lines are sometimes also present below the rims of the bowls (Figure 9.1q), recalling a common feature of Tiwanaku hyperboloid bowls (for example, Figure 8.1a).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Though other scholars have suggested that it indicates instead some degree of temporal overlap in the production of Early Pacajes and Saxamar ceramics (see [Mathews 1992]: 187, 194). I believe this argument to be invalid, and will address it at more length in my discussion of the Pacajes-Inka phase in the following chapter.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

And the villages accounted for the great majority of the Tiwanaku period population. This issue is discussed at more length in my discussion of the phase population index in Chapter 4.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The phase with the next-lowest occupation continuity index is the following Pacajes-Inka phase.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

There are many types of evidence which do indicate continuity between the Tiwanaku period and the Early Pacajes phase. One common argument, however, is absolutely invalid. Some investigators argue that the high rate of occupation continuity between Tiwanaku sites and Early Pacajes sites is evidence for the continuity hypothesis. An example from the Lower Tiwanaku Valley:

If Tiwanaku's collapse was induced by foreign invaders with a distinct cultural background, one would expect to find substantial differences in settlement distribution and land use. On the contrary, the Early Pacajes settlement pattern suggests a continuous occupation of sites ... Approximately 82% of the Tiwanaku V settlements exhibit an Early Pacajes occupation. [[Albarracín-Jordan 1992]: 310]

The fallacy here is this: given that the number of sites in the Lower Tiwanaku Valley increased threefold - from 137 to 441 - in the Early Pacajes phase, and that these new sites are located in nearly every possible settlement location, a high occupation continuity index is exactly what one would expect. Put another way, the reoccupation of a village site by a small farmstead, probably only ephemerally occupied, does not constitute the kind of occupation continuity that Albarracín-Jordan perceives. It means only that `` ... Pacajes people settled everywhere the Tiwanaku had been, and just about everywhere else'' ([Janusek and Kolata 2002]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

In doing so he is, of course, reversing the traditional Mayanist association between dispersed settlement and swidden agriculture.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This is discounting the abandonment of the city of Tiwanaku proper, of course

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It is certainly the case that the Early Pacajes settlement pattern includes a much greater occupation of the pampa in adjacent regions such as the Katari Basin ([Janusek and Kolata 2002]) and the Tiwanaku Valley ([Albarracín-Jordan 1992]: 273, 277, [Mathews 1992]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Which is to say that he gives areas of occupation for each sector - or separate component - of a site.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Which is only to say that the raw hectare value has been corrected by one method or the other.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

I consider the Lower Tiwanaku Valley to have zero population growth in this phase. This in itself is remarkable, since we can imagine that the disintegration of the urban core of Tiwanaku would have produced a large group of migrants.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Though elsewhere in a figure caption he ends the phase at 1660 A.D. ([Albarracín-Jordan 1996a]: 310). No reason for this discrepancy has been given, to my knowledge. Mathews, perhaps wisely, mentions no terminal date at all, though a chronological diagram ([Mathews 1992]: 232) has the Late Pacajes phase ending at 1600 A.D.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

In fact, the entire question of the archaeology of the period of Spanish rule in the Titicaca Basin requires much more study. Very little has been accomplished to date in this regard, though some work has been done in neighboring regions of the Andes (e.g. [Van Buren et al. 1993,Van Buren 1999]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Early Pacajes bowls are described in Chapter 9. Their rims are often flared and almost always tapered. Their slip color tends to the orange or yellow, as opposed to the deep Saxamar reds.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Taraco

I like to think of this as ``the other Taraco''. Charles Stanish has pointed out to me a general pattern of duplicate toponyms in the northern and southern Titicaca Basin, of which this is an example.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This is similar to a common decorative technique of the Sillustani Black and White on Red style from the northern Titicaca Basin ([Tschopik 1946]: 27), though this is apparently the only point of similarity between the two styles.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Albarracín-Jordan, for example, characterizes Late Pacajes ceramics as ``essentially a debilitated extension of the previous local Pacajes-Inka style'' ([Albarracín-Jordan 1992]: 327). Also, ``[el] decorado se limita a puntos negros y blancos, y en reducidos casos a líneas paralelas'' ([Albarracín-Jordan and Mathews 1990]: 175).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

That is to say that it is equal to 100 minus the Pacajes-Inka phase site founding index value.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Mathews does not include the site of Tiwanaku itself in his site registry. Since Tiwanaku seems to have been a substantial Late Horizon town, this has the effect of underestimating the population of the Middle Tiwanaku Valley during the Pacajes-Inka phase. However, the Late Horizon population decline is a undeniably a very real phenomenon, though we may quibble about its precise magnitude.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This interest in maize is characteristic of Inka rule throughout the Andean highlands ([Hastorf 1990a,Hastorf and Earle 1985]). In the case of the Titicaca Basin it may have been related to the provisioning of the important shrine on the Isla del Sol ([Bauer and Stanish 2001]).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

These include Wankarani (T-379).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Especially, of course, the Tiwanaku Valley (see [Albarracín-Jordan 1992]: 326-333, [Mathews 1992]: 194-195).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... agriculture

If it has not in fact been abandoned earlier, at the beginning of the Early Pacajes phase, as I believe it was.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

In fact, the first encomendero of both Tiwanaku and Guaqui (1538-1541 A.D.) was none other than Francisco Pizarro, the conqueror of Peru ([Choque Canqui 1993]: 60).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Albarrac\'in-Jordan

The data are in his dissertation ([Albarracín-Jordan 1992]) in Appendix 2.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

This boundary is very dramatically visible on Albarracín-Jordan's own settlement maps (e.g. [Albarracín-Jordan 1992] Figure 15.2, [Albarracín-Jordan 1996a] Figure 12.2).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Though Stern has suggested that these population declines, while real, might not have actually been so drastic as presented in official records. He documents a pervasive pattern of local groups hiding births, exaggerating deaths rates, and otherwise skewing population figures in order to reduce the tribute demanded by the colonial government ([Stern 1982]: 121-132).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Combined EF1 population index of four villages with population index values > 100 = 571

Phase population index for EF1 = 693

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Hills

See Chapter 2 for a discussion of the geology of the Taraco Peninsula.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Circumscription theory put succinctly: "...the major change we see in political structure, the change from village to chiefdom to state to empire, is the direct consequence of competition between societies" ([Carneiro 1978]: 209).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

From a MF population index value of 361.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

The boundary is located at approximately 519500E.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

It is interesting to note that the Santiago de Huatta Peninsula was beyond the radius of this LF2 demographic effect ([Lémuz Aguirre 2001]: 202), population there increasing by more than 100% during this phase. So, apparently, was the Juli-Pomata area ([Stanish et al. ]), though the LF chronology employed there makes no clear distinction between LF1 and LF2.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

There are rumors of inundated raised field groups in the vicinity of the Taraco Peninsula. I have been unable to confirm these, however.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.