Elemental Composition, Rock-Magnetic Characterization, and Archaeomagnetic Dating of Ceramic Fragments from the Paquimé Archaeological Site (Northern Mexico)

Abstract

Paquimé is a remarkable pre-Hispanic settlement that flourished between the 13th and 15th centuries in northwest Chihuahua, Mexico. This site is recognized for its distinctive fusion of Mesoamerican and Southwestern American cultural traits. Although much of the explanatory models about this settlement’s development and regional role have focused on trade, pottery from the Salado tradition, particularly Polychrome Gila and Polychrome Tonto, has generally been presumed to have originated in the American Southwest. To confirm the interaction between both cultures and contribute to the clarification of the absolute chronology of Paquimé, the geochemical characterization and rock-magnetic characterization of sherds of local and presumably foreign manufacture were carried out, including sherds with manufacture that seems to be the result of the abovementioned relationship. SiO₂ and Al₂O₃ contribute more than 75% to the observed variation. The Casas Grandes pottery shares the geochemical signatures of both local and foreign types. High-coercivity magnetic grains dominate in the foreign-type pottery samples. In contrast, relatively low-coercivity ferrimagnetic grains are the main features of local-type sherds. Essentially similar absolute intensity values were obtained for both potsherd wares. The most probable age intervals obtained for all ceramic samples studied range from 990 AD to 1310 AD, in agreement with previous surveys and local archaeological frameworks.

1. Introduction

Paquimé, also known as Casas Grandes, is located in the present-day municipality of the same name, along the western bank of the Casas Grandes River, in the state of Chihuahua, Mexico. It is recognized as the largest, most developed, and centralized pre-Hispanic community in Northwestern Mexico and the Southwestern United States [1,2]. In addition, it has been considered a link between the cultures of Mesoamerica and those of the Northwest/Southwest [3,4,5,6]. It is estimated that it was a prominent cultural development from approximately 1200 A.D. to at least 1450 A.D. [7].

The site features a public architecture zone that includes platform mounds, effigy mounds, ball game courts, plaza areas, and a water redistribution system that supplied the city from distant sources. In addition, it has approximately 2000 rooms organized into multi-story adobe rooms (housing units), which are believed to have had mainly residential functions [8]. During its peak period, the city was home to an estimated population of between 2500 and 3000 [9] and was distinguished by its complex social organization. Various authors have defined it as a society that reached an intermediate level of complexity, characterized by the existence of a marked social hierarchy and economic stratification, although without the presence of centralized political institutions or a well-defined ruling elite [1,10].

The Paquimé archaeological complex is located approximately 260 km northwest of the city of Chihuahua, Mexico, half a kilometer from the town of Casas Grandes and 5 km from the city of Nuevo Casas Grandes (Figure 1). It covers an area of ca. 70 ha, most of which has not yet been excavated [11]. It is located in an arid region characterized by high temperatures. Nonetheless, the residents managed to establish a sedentary society, obtaining vital agricultural products [12].

As mentioned by Nelson et al. [13], “During its foundation, Paquimé was part of a wide exchange network of products coming from the south of Mesoamerica, such as metals and ceramics, and from the north, such as turquoise and other minerals”.

Although the first archaeological survey in Paquimé was initiated in 1936, under the supervision of the National Institute of Anthropology and History (INAH), it was not until the period from 1958 to 1961 that detailed and systematic excavations of most houses were conducted by Charles Di Peso and Eduardo Contreras (see [11]).

Based on numerous radiometric and dendrochronology age estimations, Whalen and Minnis [1] established four occupation phases: 1. Plainware period (1 to 700 AD). 2. The Viejo period, characterized by the greatest growth of the city, is classified into three phases: Convent (700–900 AD), Pilon (900–950 AD), and Perros Bravos (950–1060 AD). 3. The Medio period (approximately 1060–1300 to 1350 AD) is subdivided into Diablo, Paquimé, and Buena Fe phases. 4. The Tardio period.

Two main hypotheses stand for the decline and abandonment of Paquimé. On the one hand, Whalen and Minnis [1] suggest a big fire and violence at the last stage of occupation. On the other hand, Posada and Reyes [11] propose environmental issues in the region as the reason for the abandonment.

Contrary to the Central and Western Mexico regions, archaeomagnetic investigations in Northern Mexico are scarce. Madingou et al. [14] obtained the first full-vector determinations in Northern Mexico. Through not-oriented material, Alva-Valdivia et al. [15] carried out the archaeomagnetic dating of pottery from the Casas Grandes region. More recently, Goguitchaichvili et al. [16] carried out the archaeomagnetic dating of deep pit ovens from Paquimé, Casas Grandes culture.

The Salado Polychromes represent one of the most widely distributed ceramic wares in the archaeology of the Southwestern United States and Northeastern Mexico. Currently, it is recognized that the Polychrome Gila and Tonto wares were produced during the fourteenth century in a wide geographical region that includes the states of Arizona and New Mexico in the United States. However, these Polychromes have also been recovered in the northern states of Mexico, such as Sonora and Chihuahua. These ceramics are part of a larger ceramic complex known as Salado or Roosevelt Red.

Although much of the explanatory models regarding the settlement’s development and regional role have focused on trade, pottery from the Salado tradition, particularly Polychrome Gila ware and Polychrome Tonto ware, has generally been presumed to have originated in the American Southwest [17,18].

The objective of this research is to determine if the Gila and Tonto Polychromes recovered at the Paquimé site were manufactured with local raw materials or if, on the contrary, their presence in this settlement is due to the exchange networks established by Paquimé with its neighbors to the north. To determine their origin, the elemental composition of the Salado types’ pastes was compared to those of local ceramic types, such as Escondida Polychrome, Ramos Polychrome, and Casas Grandes Liso. Additionally, rock-magnetic characterization of sherds of local and presumably foreign manufacture was carried out. Likewise, the archaeomagnetic dating of the sherds was carried out.

2. Materials

2.1. Geological Background

The study area is located in the Sierras y Llanuras Tarahumaras physiographic subprovince, in the Chihuahua tectonostratigraphic terrain. The basement is composed of metamorphic Precambrian rocks, which are overlain by Cretaceous limestones, sandstone, siltstone, gypsum, and fossiliferous horizons. During the Tertiary, extrusive igneous rocks are abundant and include andesites, rhyolites, ignimbrites, and basalts. During the same period, magmatic activity formed granitic and granodioritic intrusive bodies with an age of 35 Ma. The end of the Tertiary is characterized by conglomerate deposits, which remain during the Quaternary with intercalations of alluvium [19].

The Paquimé archeological site, in the Casas Grandes region, was built on Quaternary detrital polymictic conglomerate and alluvium resting on rhyolitic ignimbrite Oligocene rocks. It is located in the eastern foothills of the Sierra Madre Occidental, on a fertile floodplain adjacent to the Río Casas Grandes. The Casas Grandes region is situated in the Basin and Range province of Northwest Mexico, though its western and southern peripheries extend into the Sierra Madre, and to the east are desert and semi-desert plains [20].

2.2. Pottery Samples

The samples analyzed for elemental composition comparison comprise Salado wares (Gila Polychrome and Tonto Polychrome) of likely foreign origin and local ceramic wares (Escondida Polychrome, Ramos Polychrome, and Casas Grandes Liso). The archaeological information of the pottery samples analyzed is summarized in

Table 1. Archaeological information of the pottery samples analyzed, including pottery ware, chronology, technology, pottery shape, and decorative style.

Salado Polychrome Ware	Period	Technology	Forms	Decorative Styles
Gila Polychrome	1300–1450 d. C.	Distinct slips A thin layer of clay called a slip was applied over the brown body. Organic paint A sticky syrup made from boiled-down plants was used to paint the motifs. Oxidizing fire The pottery was fired once in an open fire that reached temperatures of around 750 °C.	Bowl and jar	Use of white slip as a background. Motifs in black, characterized by bold lines and abstract shapes, both rectilinear and curvilinear. Application of red slip only on the exterior surface of the bowls and on the base of the pots.
Tonto Polychrome	1300–1450 d. C.		Jar, bowl, and bird effigy jar	Use of white slip as a background. Motifs in black, characterized by bold lines and abstract shapes, both rectilinear and curvilinear. The segments of black-on-buff motifs are surrounded by large red slipped areas.
Casas Grandes Polychrome Ware	Period	Technology	Forms	Decorative Styles
Escondida Polychrome Gila Variant	1300–1450 d. C.	Soft thin red slip or wash A thin layer of slip was applied over the white body. Mineral Paint Presumably made of Mn, Pb, Cu, and Fe. Oxidizing Fire The pottery was fired once in an open fire that reached temperatures of around 750 °C.	Jar, bowl, bird effigy jar, and human effigy jar	The decoration of both variants (Gila and Tonto) is similar to that observed in their counterparts, the Salados types.
Ramos Polychrome	1300–1450 d. C.	Self-slip A finely textured surface that appears to be slip-coated with the same material that constitutes the light brown clay body. Mineral Paint The black paint is presumably made of Mn, Pb, Cu, and Fe. Oxidizing Fire The pottery was fired once in an open fire that reached temperatures of around 750 °C.	Jar, bowl, bird effigy jar, human effigy jar, and Tecomate	Its decoration consists of finely painted black and red lines on the exterior surface of the vessels. The red elements are typically outlined in black. The motifs were characterized by figurative, organic, abstract, and geometric patterns.
Casas Grandes Utility Ware	Period	Technology	Forms	Decorative Styles
Casas Grandes Plainware	1200–1450 d. C.	Polished surfaces Oxidizing fire The pottery was fired once in an open fire that reached temperatures of around 750 °C.	Jar, bowl, bird effigy jar, human effigy jar, and Tecomate	None

.

3. Methodologies

3.1. X-Ray Fluorescence

X-ray fluorescence (XRF) is an analytical method used to determine the chemical composition of all kinds of material. The method is fast and accurate and normally requires minimal sample preparation [21]. In this methodology, X-rays are produced by a source and irradiate the sample, which will emit fluorescence X-ray radiation with discrete energies characteristic of the elements present in the sample. It can be either a qualitative or a quantitative analysis. The detection limit of modern spectrometers goes from 10 ppm−wt./100%, and their accuracy (relative error) is <10% by using calibration with standards.

XRF is currently applied in archaeometry to analyze the chemical composition of pottery and its provenance [22]. Elemental determination (major oxides) by XRF was performed on pressed pellets (Figure 2) of 48 ceramic sherds of the Escondida (9 samples), Casas Grandes (10 samples), Ramos (7 samples), Tonto Polychrome (7 samples), and Gila Polychrome (15 samples) wares. The individual samples were washed with distilled water in an ultrasonic bath after their cutting to avoid contamination due to the manipulation process.

Pressed-pellet sample preparation for geochemical analysis was carried out as follows: A fragment from each potsherd was manually crushed using an agate mortar. Then, 3 g of the milled material was mixed and homogenized within the same mortar with 0.5 g of wax-C micro powder (Hoechst). The mixed compound was deposited into a 30 mm diameter stainless-steel die and pressed up to 20 tons for two minutes using an Atlas (Specac) hydraulic press. The analyses were performed with a Xenemetrix X-Calibur energy-dispersive X-ray fluorescence (EDXRF) spectrometer at the Laboratorio Universitario de Geofísica Ambiental (LUGA) facilities. An integration time of 60 s was used for the individual runs.

3.2. Rock-Magnetic Characterization and Archaeointensity Experiments

The magnetic characterization (isothermal remanent magnetization (IRM), backfield, hysteresis, and magnetization vs. temperature curves) of ceramic pots F14 (Tonto Polychrome), F16 (Tonto Polychrome), F23 (Gila Polychrome), L3 (Escondida), L43 (Casas Grandes), and L49 (Ramos Polychrome) was carried out using a variable field translation balance (VFTB). For these experiments, a chip fragment from each potsherd was pulverized with the agate mortar, and 300 mg were set into the balance’s sample holder. The magnetic characterization of the sherds was carried out at the Laboratorio Universitario de Geofísica Ambiental (LUGA) facilities.

Archaeointensity (AI) experiments were carried out using the Thellier–Coe experiments [23,24] on the six pot types mentioned above using the following process:

(1)

Cutting each pot into at least eight fragments/specimens (Figure 3).

(2)

Pressing each specimen into a salt pellet with dimensions similar to standard paleomagnetic cores (Figure 4).

AI determinations were carried out on six specimens (b–g) per pot. Ten double-heating temperature steps, as well as partial thermoremanence verifications (pTRM checks [25]) and tail checks [24], were implemented in the AI experiments.

Subsequently, anisotropy of thermoremanence (ATRM) correction was performed following the procedure described by Veitch et al. [26]. Correction for differences in cooling rate (CR) was performed on specimens identified with the “a” label following the procedure described by Chauvin et al. [27], as modified by Morales et al. [28].

The AI determinations were carried out at the Servicio Arqueomagnético Nacional (SAN) facilities, using the Paleointensity.Org software, V1.1.0 [29].

3.3. Archaeomagnetic Dating

Archaeomagnetic dating comprises the comparison of direction (declination, inclination) and/or intensity, obtained from an archaeological artifact (or a lava flow) of unknown age, against a reference PSV curve produced based on a set of well-dated materials. A detailed description of the procedure can be found elsewhere (e.g., [28]). Archaeomagnetic (specifically archeointensity) dating was accomplished via the archaeo_dating software, v8, of Pavón-Carrasco et al. [30] using the regional secular variation curve of Mahgoub et al. [31].

4. Results

4.1. Elemental Composition

The individual results are presented in

Table 2. Elemental composition (major oxides) of the analyzed fragments. Concentration values in [wt./%].

	Na2O	MgO	Al2O3	SiO2	P2O5	K2O	CaO	TiO2	MnO	Fe2O3	Σ
Casas Gdes.
39	2.36	1.83	19.91	68.63	0.07	3.79	1.57	0.72	0.04	3.69	102.6
40	1.24	2.98	18.52	68.11	0.07	3.50	1.82	0.51	0.05	3.88	100.7
41	1.86	1.28	25.23	61.37	0.44	3.01	1.19	0.73	0.05	4.96	100.1
42	1.60	1.56	17.81	69.16	0.28	4.08	1.45	0.30	0.04	4.72	101.0
43	1.68	2.26	17.31	61.90	0.50	4.30	1.48	0.79	0.08	5.73	96.0
44	1.55	2.21	19.83	61.89	0.64	2.58	3.25	0.74	0.06	5.07	97.8
45	1.49	2.23	17.04	69.69	0.16	3.43	1.84	0.52	0.05	4.16	100.6
46	1.50	1.24	19.94	70.49	0.05	4.64	1.08	0.42	0.03	3.41	102.8
47	1.81	1.46	18.22	65.51	0.24	3.87	1.18	0.44	0.05	5.71	98.5
48	1.83	1.19	17.68	65.34	0.25	3.89	1.02	0.43	0.04	5.34	97.0
Mean =	1.7	1.8	19.1	66.2	0.3	3.7	1.6	0.6	0.05	4.7
1 σ =	0.3	0.6	2.4	3.5	0.2	0.6	0.7	0.2	0.01	0.8
Ramos
51	1.58	1.42	19.19	73.83	0.03	4.65	0.76	0.40	0.05	2.54	104.4
52	1.52	2.08	19.12	75.49	0.12	4.39	0.89	0.28	0.04	1.69	105.6
53	1.56	2.25	17.19	76.61	0.08	3.99	1.01	0.29	0.05	2.68	105.7
54	1.77	1.58	18.17	77.07	0.04	4.46	1.00	0.18	0.05	1.29	105.6
55	1.57	1.74	15.48	73.51	0.05	3.98	0.83	0.33	0.05	3.03	100.6
56	2.02	2.01	18.67	75.66	0.05	4.64	0.90	0.27	0.05	1.61	105.9
57	2.33	1.27	17.73	76.97	0.18	4.46	0.83	0.18	0.05	1.71	105.7
Mean =	1.8	1.8	17.9	75.6	0.1	4.4	0.9	0.3	0.05	2.1
1 σ =	0.3	0.4	1.3	1.4	0.1	0.3	0.1	0.1	0.00	0.7
Escondida
4	1.45	1.69	17.83	75.30	0.03	4.35	0.81	0.32	0.04	2.51	104.3
5	1.89	1.34	17.95	78.43	0.02	4.34	0.81	0.28	0.04	1.97	107.1
6	1.80	1.29	16.82	78.24	0.04	4.16	1.11	0.26	0.05	2.02	105.8
7	1.82	1.10	17.57	76.74	0.37	4.54	0.92	0.27	0.05	1.95	105.3
8	1.82	1.56	20.46	69.15	0.07	4.58	0.92	0.52	0.06	2.78	101.9
9	1.47	1.57	17.27	73.11	0.03	4.99	1.09	0.29	0.05	1.87	101.7
10	0.91	1.72	9.78	42.47	0.30	3.04	17.41	0.44	0.02	2.95	79.0
11	1.89	1.94	18.61	72.32	0.21	4.47	0.92	0.45	0.05	2.56	103.4
12	1.77	1.19	17.37	77.98	0.01	4.53	0.88	0.17	0.05	1.29	105.2
Mean =	1.6	1.5	17.1	71.5	0.1	4.3	2.8	0.3	0.05	2.2
1 σ =	0.3	0.3	2.9	11.3	0.1	0.5	5.5	0.1	0.01	0.5
Tonto P.
16	0.95	1.60	11.29	39.29	0.12	2.57	18.42	0.45	0.04	3.61	78.3
17	1.57	2.20	15.18	65.00	0.10	3.53	2.06	0.76	0.09	5.20	95.7
18	1.76	4.43	14.12	61.17	0.24	3.19	4.59	0.70	0.07	5.20	95.5
19	1.77	4.58	14.40	59.53	0.31	3.04	4.30	0.85	0.10	6.03	94.9
20	1.45	4.87	13.25	54.40	0.48	3.63	7.81	0.64	0.07	4.99	91.6
21	1.61	4.58	16.27	63.92	0.31	3.26	3.66	0.75	0.09	5.62	100.1
22	1.70	2.26	16.09	57.09	0.19	2.77	2.68	1.22	0.09	7.14	91.2
Mean =	1.5	3.5	14.4	57.2	0.3	3.1	6.2	0.8	0.08	5.4
1 σ =	0.3	1.4	1.7	8.7	0.1	0.4	5.7	0.2	0.02	1.1
Gila P.
24	1.23	4.19	14.34	61.11	0.23	3.65	5.15	0.64	0.08	4.66	95.3
25	1.13	4.09	15.76	59.58	0.22	3.22	4.70	0.67	0.09	5.47	94.9
26	1.40	2.14	19.58	57.93	0.10	3.20	2.19	0.63	0.06	6.77	94.0
27	1.83	4.54	14.83	62.71	0.18	3.23	3.55	0.74	0.07	5.53	97.2
28	1.31	4.03	13.73	59.17	0.24	3.20	4.99	0.59	0.07	4.57	91.9
29	1.36	5.80	14.88	64.45	0.33	3.50	3.47	0.73	0.06	5.53	100.1
30	1.29	4.15	15.97	61.15	0.27	3.06	4.58	0.70	0.08	5.36	96.6
31	1.91	1.92	18.44	71.19	0.08	4.33	0.70	0.42	0.05	3.96	103.0
32	1.29	3.98	15.37	59.88	0.27	3.42	4.49	0.73	0.08	5.17	94.7
33	1.42	4.43	14.23	61.51	0.25	3.86	4.65	0.65	0.08	4.63	95.7
34	1.68	4.15	13.59	59.30	0.29	3.52	5.82	0.64	0.06	4.69	93.7
35	1.38	4.34	15.63	60.44	0.23	3.18	4.49	0.70	0.07	5.11	95.6
36	1.87	4.14	13.60	59.47	0.20	3.05	5.30	0.65	0.06	4.91	93.3
37	1.39	2.39	17.03	56.27	0.16	2.34	2.05	0.90	0.05	10.02	92.6
38	1.32	4.78	17.01	65.26	0.25	3.36	4.86	0.65	0.09	5.23	102.8
Mean =	1.5	3.9	15.6	61.3	0.2	3.3	4.1	0.7	0.1	5.4
1 σ =	0.2	1.0	1.8	3.6	0.1	0.4	1.4	0.1	0.0	1.4

.

After performing a principal component analysis (PCA), it was observed that two components (SiO₂ and Al₂O₃) contributed more than 75% to the observed variation. In this way, the values of the components mentioned above were plotted in a scatter plot (Figure 5). The green envelope shows that local-type potsherds (Ramos and Escondida) possess a very similar geochemical signature. Similarly, “apparently” foreign-type potsherds (Tonto Polychrome and Gila Polychrome) possess a similar geochemical signature, as shown in the yellow envelope. The Casas Grandes potsherds share the geochemical signatures of both local and foreign types, as shown in the purple envelope.

These similar geochemical signatures also stand out in the Fe content. Local-type potsherds have Fe₂O₃ values between 1.3 and 3.0 wt./%, while foreign-type potsherds present higher Fe₂O₃ values between 3.6 and 10.0 wt./%.

4.2. Rock-Magnetic Characterization

IRM plots evidence the presence of a high-coercivity magnetic mineralogy—likely hematite—as the principal remanence carrier within the foreign-type sherd, since it does not saturate at fields as high as 750 mT (Figure 6a). On the contrary, IRM plots of the local-type sherds saturate at relatively low magnetic fields of 250–300 mT, which suggests the presence of a low-coercivity magnetic mineralogy—likely magnetite—as the principal remanence carrier (Figure 7a). Hc values for both foreign- and local-type sherds seem similar (Figure 6b and Figure 7b). In contrast, while most foreign-type sherds present the constricted (so-called wasp-waisted) shape, local-type ones mainly show the so-called pot-bellied shape c and Figure 7c).

M-T plots mainly show a single ferrimagnetic phase with a Curie temperature of ~540–560 °C for the foreign-type sherds (Figure 6d). In contrast, corresponding M-T plots for local-type sherds show between two and three ferrimagnetic phases, the highest one with a Curie temperature above 550 °C (Figure 7d). Both reversible and irreversible curves are observed in the potsherds analyzed, which have implications for AI determinations, as will be discussed further.

4.3. Archaeointensity Determinations

Representative examples of Arai diagrams are presented in Figure 8 and Figure 9.

Individual AI results are presented in

Table 3. Archaeointensity results. Braw: uncorrected ancient geomagnetic field strength; Tmin/Tmax: min/max temperature used for the least-square linear fitting; n: number of points used for the least-square linear fitting; f: NRM fraction; β: ratio of the standard error of the slope of the selected segment in the Arai plot to the absolute value of the slope; q: quality factor; MAD_ANC: anchored maximum angular deviation; n_pTRM: number of tail checks carried out; DRAT; CDRAT; B_ATRM: ancient geomagnetic field strength corrected for ATRM effects; B_{ATRM + CR}: ancient geomagnetic field strength after ATRM and CR corrections; NR: no result.

		Braw	Tmin	Tmax	n	f	β	q	MADanc	npTRM	DRAT	CDRAT	BATRM	BATRM+CR
			[°C]	[°C]					[°]				[µT]	[µT]
PQM-1	F14-B	71.0	350	560	6	0.777	0.027	20.2	0.75	3	2.6	2.0	65.3	56.3
PQM-2	F14-C	77.7	275	560	7	0.791	0.031	19.5	1.60	3	1.8	0.2	66.0	56.9
PQM-3	F14-D	70.2	350	560	6	0.749	0.031	18.3	1.19	3	2.6	3.6	66.7	57.4
PQM-4	F14-E	72.7	200	560	8	0.803	0.038	17.1	2.50	3	2.3	4.0	65.4	56.4
PQM-5	F14-F	56.4	275	560	7	0.737	0.018	32.7	0.85	3	4.8	8.5	57.0	49.1
PQM-6	F14-G	56.7	275	560	7	0.726	0.021	28.7	0.84	3	4.0	9.5	65.8	56.7
	mean =	67.5											64.4	55.4
	1 σ =	8.8											3.7	3.2
PQM-7	F16-B	NR
PQM-8	F16-C	NR
PQM-9	F16-D	NR
PQM-10	F16-E	NR
PQM-11	F16-F	NR
PQM-12	F16-G	NR
	mean =
	1 σ =
PQM-13	F23-B	70.7	415	560	5	0.629	0.096	4.6	1.00	3	6.1	5.3
PQM-14	F23-C	70.0	415	560	5	0.589	0.087	4.7	1.20	3	4.8	9.1
PQM-15	F23-D	78.7	415	560	5	0.579	0.038	10.6	1.40	3	7.6	8.9
PQM-16	F23-E	70.6	415	560	5	0.534	0.025	14.9	0.90	3	5.5	8.4
PQM-17	F23-F	76.6	415	560	5	0.630	0.083	5.3	1.10	3	6.7	4.0
PQM-18	F23-G	80.6	415	560	5	0.615	0.062	6.8	1.60	3	3.4	0.2
	mean =	74.5
	1 σ =	4.7
PQM-19	L3-B	60.0	200	560	8	0.537	0.038	11.5	1.45	3	5.0	6.4	58.2	55.7
PQM-20	L3-C	51.7	200	560	8	0.425	0.030	11.6	1.28	3	2.1	2.1	53.8	51.5
PQM-21	L3-D	60.0	200	560	8	0.707	0.020	29.6	0.65	3	2.6	1.4	55.8	53.4
PQM-22	L3-E	57.9	200	560	8	0.720	0.022	27.6	0.78	3	1.6	0.4	57.3	54.9
PQM-23	L3-F	58.8	200	560	8	0.684	0.030	19.1	1.34	3	2.6	1.3	54.1	51.8
PQM-24	L3-G	53.0	200	560	8	0.483	0.016	26.2	0.87	3	3.3	7.8	64.1	61.4
	mean =	56.9											57.2	54.8
	1 σ =	3.6											3.8	3.6
PQM-25	L43-B	NR
PQM-26	L43-C	NR
PQM-27	L43-D	NR
PQM-28	L43-E	NR
PQM-29	L43-F	NR
PQM-30	L43-G	NR
	mean =
	1 σ =
PQM-31	L49-B	68.8	350	560	6	0.658	0.076	6.5	3.36	3	2.9	1.4	59.2	56.6
PQM-32	L49-C	68.4	200	560	8	0.840	0.028	24.7	1.01	3	2.8	2.9	54.0	51.7
PQM-33	L49-D	76.4	200	560	8	0.820	0.012	55.8	1.17	3	1.0	0.4	59.6	57.0
PQM-34	L49-E	60.2	415	560	5	0.547	0.014	26.9	1.64	3	4.2	5.8	50.6	48.3
PQM-35	L49-F	71.4	200	560	8	0.883	0.025	28.7	1.23	3	0.2	0.2	57.8	55.3
PQM-36	L49-G	64.0	200	560	8	0.883	0.039	18.7	1.99	3	1.7	1.9	52.5	50.2
	mean =	68.2											55.6	53.2
	1 σ =	5.6											3.8	3.6

. Potsherd F14 yielded an average ancient field strength of (55.4 ± 3.2) µT. Potsherds F16 and F23 failed to yield reliable results; vectorial plots of all F16 specimens, except one (PQM-11), failed to go to the origin. This means that an artificial chemical remanent magnetization was produced during heating, which has no geomagnetic significance. A higher average raw ancient field strength of (74.5 ± 4.7) µT was retrieved from the F23 specimens; the corresponding ATRM-corrected ancient field strength was not possible to obtain due to significant abnormal ATRM correction factors obtained.

In the case of the local-type fragments, very similar AI values of (54.8 ± 3.6) µT and (53.2 ± 3.6) µT were obtained for L3 and L49 potsherds, respectively. No results were obtained from F43 specimens; concave-up plots, characteristic of multi-domain (MD) grain sizes, were obtained from these specimens.

4.4. Archaeointensity Dating

The most probable age ranges obtained for the potsherds from which reliable AI values were obtained are between 990 AD and 1310 AD (Figure 10). As expected, due to the similar AI obtained for them, both local- and (likely) foreign-type potsherds possess indistinguishable age ranges.

5. Discussion

The geochemical results of this study differ from those obtained by Di Peso [32]. While he observed similarities in the pastes of the Casas Grandes Liso and Gila Polychrome wares, interpreting them as evidence that both series were made with local raw material, our analyses indicate that Polychrome Gila is clearly distinguishable from what we consider to be local pottery. In particular, Ramos and Escondida Polychromes are more similar to each other than to the Casas Grandes Liso type. In addition, greater variability is observed in the composition of Tonto Polychrome compared to Gila Polychrome.

A well-differentiated grouping of local- and probably foreign-type potsherds was evidenced by their geochemical signatures. In the case of the Casas Grandes material (local-type potsherds), its geochemical signature seems to have an affinity with both local and foreign types.

One must bear in mind that the elemental composition of the ceramics mainly depends on the raw material (clay) used for their elaboration. As mentioned in Section 2.1 (Geological Background), the Paquimé archeological site was built on Quaternary detrital polymictic conglomerate and alluvium. With such an assortment of clay sources, it is unsurprising to observe different geochemical signatures.

Moreover, elemental composition could also depend on the preparation process: mixing of more than one clay type to give the ceramic the color and/or consistency desired, a process closely related to a specific tradition (likely, the Salado tradition).

The well-differentiated grouping based on the geochemical signatures mentioned above seems to be supported by the rock-magnetic characteristics of the potsherds: (i) high-coercivity mineralogy observed in the foreign-type potsherd compared to median-coercivity mineralogy in the local-type material; (ii) wasp-waisted vs. pot-bellied shape for foreign and local types, respectively; and (iii) single vs. multiple mineral ferrimagnetic phases in foreign and local types, respectively.

The similar AI values of (54.8 ± 3.6) µT, (53.2 ± 3.6) µT, and (55.4 ± 3.2) µT obtained for L3, L49, and F14 potsherds, respectively, agree with those previously reported by Alva-Valdivia et al. [16], at least at the upper part of Table 3. Those reported at the lower part of the table, however, show significantly lower AI values (45.7 ± 3.0) µT, on average.

Indistinguishable age ranges of local- and (likely) foreign-type potsherds suggest similar temporalities for both types of potsherds. The most probable age intervals obtained for all ceramic samples studied range from 990 AD to 1310 AD.

The results of this study allow us to confirm that the age intervals obtained for the ceramics analyzed coincide with the reevaluation of the chronology of Dean and Ravesloot [7]. This places the Salado pottery recovered in Paquimé within contexts corresponding to the fourteenth century.

6. Conclusions

Aimed to determine whether the Gila and Tonto Polychromes recovered at the Paquimé archaeological site were manufactured with local raw materials or if, on the contrary, their presence in this settlement is due to the exchange networks established by Paquimé with its neighbors to the north, the elemental composition of the pastes of the Salado types were compared with those of local ceramic types (Escondida Polychrome, Ramos Polychrome, and Casas Grandes Liso). Forty-eight pottery fragments were subjected to X-ray fluorescence analysis.

The principal component analysis revealed that SiO₂ and Al₂O₃ contribute more than 75% to the observed variation. A well-differentiated grouping of local- and probably foreign-type potsherds was evidenced by their geochemical signatures.

The analyses of local clayey raw materials seem to be necessary for the provenance of the ceramics.

Additionally, the rock-magnetic characterization of sherds of local and presumably foreign manufacture carried out supports the findings of the geochemical analysis.

Essentially similar absolute intensity values were obtained for both local and presumably foreign-type potsherds, and indistinguishable age ranges of local- and (likely) foreign-type potsherds suggest similar temporalities for both types of potsherds. The most probable age intervals obtained for all ceramic samples studied range from 990 AD to 1310 AD, in agreement with previous surveys and local archaeological frameworks.