Feathers addresses the dual challenges of inferring original vessel counts from sherds and inference to use life from reconstructed vessels. His solution assumes the validity of sherd assemblages as units of observation that considerable research invalidates and overlooks methods that estimate original vessels from sherds. Feathers also doubts that use life can be inferred for reconstructed vessels. Although not a focus of my article, the larger study from which it derived addresses this matter in detail that strongly warrants vessel size as use-life measure. Of course we must be pragmatic in quantifying pottery assemblages, but first we must identify valid units of observation, and only then attend to pragmatics.

Archaeologists see the value, if not the allure, of formation theory. Before inferring what happened in the past and why, we must know how the material record formed. Pottery is abundant and informative, therefore a common analytical subject. Understanding size and composition of ceramics assemblages requires formation theory, including knowledge of vessel use life. This fundamental quantity has two salient properties. The first—central tendency measured by mean or median—is widely acknowledged. Use life's second, equally important, property is the distribution of failure-age by specimen across assemblages. This article considers how and why both use-life properties affect size and composition of pottery assemblages. From a longitudinal ethnoarchaeology of household pottery in Michoacán, Mexico, it identifies vessel-size measures that correlate with use-life mean, and it demonstrates archaeologically innovative ways to characterize distributions that improve both analysis of assemblages and comparison between them.

Significant parts of the archaeological record are in private hands, including those of responsible and responsive stewards (RRS). This is not necessarily a bad thing. The Central Ohio Archaeological Digitization Survey (COADS) engages RRS in collaborative study of central Ohio prehistory. COADS leverages the mass of RRS data constructively to add depth and breadth to a regional archaeological record. We recorded over 12,000 diagnostic points and about 5,000 other stone tools from 32 RRS collections. All were scanned as two-dimensional (2D) images, and a sample as three-dimensional (3D) models for landmark-based geometric morphometric (LGM) analysis and GIS analysis of prehistoric land use. The resulting dataset is >4.7 times the number of diagnostics recorded in the Ohio SHPO database for the region, shedding new light on land use and tool use over millennia. In addition to academic research, COADS creates an accessible collection of 3D models available to RRS colleagues and to society at large. This reciprocal sharing mutually benefits professionals and the RRS community. Professionals are not the only ones who research the cultural past; the more and better we collaborate with RRS and others with legitimate interests, the better our common understanding of that past.

Projectile points are a common subject of archaeological study. In the past decade, landmark-based geometric morphometrics (LGM) has increasingly been used to analyze points as whole objects. LGM and other studies document allometric changes in points—change in shape with change in size—as a product of resharpening. Allometry registers in part because different segments or modules of points are subject to different degrees of resharpening, with blades often experiencing more reduction than stems. Different modules retain varying degrees of morphological integrity as points move through their use lives. Most previous LGM studies involved two-dimensional point models, and few tested directly for modularity. We apply LGM methods to three-dimensional models of Folsom point replicas whose degree and pattern of reduction are known, finding evidence for both allometry and modularity, with modest integration. Complementary non-LGM data reveal similar results, indicating a robust pattern and ways to approximate LGM results in other data. Moreover, our dataset's experimental control clearly identifies the results as a function of the progressive reduction in use experienced by points.

How much stone tools are reduced and their form changed from first use to discard bears upon issues such as typological integrity, curation rate, and effects of occupation span. But degree of reduction depends partly upon the measures used to gauge it. Most studies involve single indices that gauge reduction in different ways or at different scales, so results are difficult to compare between studies. In this pilot study, we compare four allometric reduction measures—one each based on length, length:thickness ratio, volume, and mass, estimated by comparing observed values in discarded tools to estimated original values—for consistency when applied to an endscraper sample from the Nobles Pond Paleoindian site in Ohio, USA. Fitted to the Weibull distribution, all measures suggest attrition compared to experimental controls, but variation among them demands reconciliation. A multifactorial method that weights individual measures by their principal-component loadings suggests attritional discard at increasing rate as reduction advances. More importantly, it addresses the growing problem of reconciling the many reduction measures in use, a major concern in this expanding research area.

Chipped-stone projectile points are used to mark the passage of time and cultures in the record. Archaeologists often recover points in surface survey, yet we do not know how many were found by private collectors before or after professional work. In a 1975–1977 Michigan probabilistic survey, professional archaeologists documented 30 private collections from 20 sample units. In those units, points found by private collectors outnumber professionally recovered ones by a factor of about 32. The survey region's point population estimated separately from the professional and private-collection samples differs by nearly an order of magnitude in favor of private collections, despite highly conservative assumptions about the latter. The number of points found in professional survey is inversely correlated with the number found in private collections, and the professional sample is more sparsely and randomly distributed. However, proportions of several common types are similar between professional and private collections. To the extent that large, reasonably complete samples of points are important for research and preservation, archaeologists must document private collections compiled in and near their survey areas.

The practice of using point estimates produced by the King ecological inference technique as dependent variables in second-stage linear regressions leads to second-stage results that, in general, are inconsistent. This conclusion holds even when all assumptions behind King's ecological technique are satisfied. Second-stage inconsistency is a consequence of the fact that King-based point estimates of disaggregated quantities contain errors correlated with the true quantities the estimates measure. Our findings on second-stage inconsistency, as well as a fix that we propose, follow from econometric theory in conjunction with an analysis of simulated and real ecological data sets.

We address in Herron and Shotts (2003; hereinafter HS) an increasingly common statistical practice called EI-R, in which point estimates generated by the King (1997) ecological inference technique are used as dependent variables in second-stage linear regressions. Although HS argue that EI-R slope estimates are inconsistent and suffer from attenuation bias, this characterization of EI-R is excessively conservative. Indeed, we show here that EI-R estimates can suffer from sign reversals, attenuation bias, and augmentation bias, that inference based on these estimates can be misleading, and that EI-R is unfixable and should not be used.

Since Herron and Shotts (2003a; hereinafter HS), Adolph and King (2003; hereinafter AK), and Herron and Shotts (2003b; hereinafter HS2), the four of us have iterated many more times, learned a great deal, and arrived at a consensus on this issue. This paper describes our joint recommendations for how to run second-stage ecological regressions, and provides detailed analyses to back up our claims.

The field-processing model (FPM) relates degree of resource processing at procurement site to round-trip travel distance. Typically applied to food getting, its extension to stone tools is part of a larger trend to broaden the scope and strengthen the theoretical foundation of inference from lithic data. This trend guided the FPM's application at the Modena obsidian quarry in the Great Basin, which follows earlier analyses by comparing the proportion by reduction stage of biface preforms between quarry and outlying sites and the technological characteristics of debris assemblages. It also uses two ratio-scale methods, slope of the regression of preform weight upon the first principal-component of linear dimensions, which declines with reduction, and slope of cumulative-count of flakes upon flake size, which increases with reduction. Results support the FPM and previous studies that document early-stage reduction at quarries vs. later-stage reduction across the landscape. The essential next step in FPM testing requires data on pattern and extent of biface reduction as a continuous function of distance from source. As useful as are existing measures and approaches, newly defined ratio-scale measures can particularly expedite testing of the FPM in its continuous terms.

"Assemblage" is a fundamental archaeological construct. By their composition, we interpret assemblages as expressions of activity or cultural identity. Yet they are not simple products of these factors alone but also of formation processes. Assemblages accumulated over varying spans, with varying combinations of tool types and discard rates. They are contexts for the playing out of complex relationships, not static types. This is a materialist view, against the essentialist view that assemblages are exemplars of ideal types (e.g., "base camps," "Quina Mousterian"). Materialism implies that their size and composition, fundamental assemblage characteristics, are correlated variables, not fixed properties, and that composition varies as size increases. I document size dependence consistent with materialism in Paleoindian and Paleolithic assemblages. Among ways to analyze size-dependent assemblage data, I apply "SHE analysis"—the joint study of assemblage richness, heterogeneity, and evenness—to gauge data’s fit to theoretical models. Archaeologists acknowledge size dependence, but we misapprehend it as a methodological bias of assemblage measures that must be suppressed when, in materialist perspective, it reveals meaningful relationships that an essentialist view cannot.

Stone tools were reduced during use, with implications both for classification and curation rates. Ballenger's “expended utility” (EU) is a continuous reduction measure devised for Dalton bifaces, described by its mean but also its distribution among specimens. We validate EU as a reduction measure by reference to experimental and contextual controls. We compare EU between the “special context” Dalton assemblages Sloan and Hawkins in Arkansas and Ballenger's eastern Oklahoma “occupation context” ones. Then we fit EU distributions to mathematical functions to model the curation process. Results show that Oklahoma bifaces were better curated than Arkansas ones. Fitting distributions to the Gompertz-Makeham model efficiently describes distributions' shape and scale, which are as important to know as central tendency. Curation is not a categorical state but a continuous variable whose complex variation implicates complex causes.

Assemblages are composed of proportions of artifacts by category. Use life affects the formation and therefore the size and composition of assemblages. Use life is to assemblage formation as lifespan is to demography, and demographers know that a population's mean lifespan is no more important than the distribution of values around that mean. When considered at all, use life typically is expressed as a mean value. But use-life's distribution—variation around the mean—affects assemblage composition independently of the mean. Distribution is neglected because its effects are not appreciated and seem difficult to measure. To improve understanding of assemblage formation, we study use-life distribution in New Guinea Wola ethnographic artifacts, using cumulative survivorship and the two-parameter Weibull model. Then we propose estimates of use-life distribution in Paleoindian stone tools. Knowing use-life distribution as well as mean, we know better how assemblages formed and improve our understanding of the archaeological record.