Society for American Archaeology

Periodic Markets in Precolumbian Mexico Author(s): Ross Hassig Source: American Antiquity, Vol. 47, No. 2 (Apr., 1982), pp. 346-355 Published by: Society for American Archaeology Stable URL: http://www.jstor.org/stable/279906 . Accessed: 01/03/2011 14:28
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=sam. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to American Antiquity.

http://www.jstor.org

RecpoRts

PERIODICMARKETS IN PRECOLUMBIANMEXICO
Ross Hassig
Considerations of markets in Precolumbian Mexico have failed to analyze the significance of periodicity for Precolumbian market systems. This essay examines periodicity and argues that the several types of recorded Precolumbian markets are qualitatively different; the 5-day market was basic, while 9-, 13-, and 20-day markets were ritual in nature, tied to specific and infrequent days, and played little or no part in market articulation.

Markets were among the first phenomena of New Spain to spark the interest of the Spaniards, and they continue to be of interest at present. However, the emphasis has shifted from marketplaces to marketing: the patterns of markets and their interactions as systems now consume much attention. The efforts now being focused on sixteenth-century Mexico have generally concentrated either on markets as systems (Smith 1979), or on the frequency of individual markets (Gibson 1964; Kurtz 1974; Bromley 1974). While both approaches are laudable, neither is without difficulties. The system approach is comprehensive and thus more vulnerable to criticism in terms of both theory and data. Theories of marketing systems, such as central place theory, are based on human interaction and seek to explain the spatial distribution of markets in terms of human behavior. Thus, one must begin with human interaction in order to establish market patterns, rather than deducing market patterns from the existence of a certain settlement pattern (Christaller 1972:608). It is both unreliable and misleading to assume market interaction from settlement patterns; this caveat highlights one of the difficulties of applying spatial marketing theories to archaeological data. Even the use of historical records is not without difficulties. Most of the data concerning Mexico that are sufficiently detailed to allow extrapolation of marketing behavior, such as the Relaciones Geograficas of 1579-1580, are from periods many decades removed from the Spanish Conquest (A.D. 1519-1521), after the indigenous economic patterns had been significantly distorted. Given the theoretical constraints upon marketing theories, and the paucity of qualitatively sound data from the relevant time period, an accurate picture of Precolumbian marketing in Mexico may not be possible. However, what we do know can be made clearer. The focus of the present analysis is on the temporal sequencing of Precolumbian markets. Periodic markets-those which meet episodically rather than daily-are a common phenomenon in much of the world today, and they were a prominent feature of the Aztec world as well. Periodic markets are viable institutions wherever demand density is too low to support a permanent market. If the threshold of a commodity (the area of demand required to support a supplier)
Ross Hassig, Department of Sociology and Anthropology, Vanderbilt University, Nashville, TN 37235

Copyright? 1982 by the Society for AmericanArchaeology

0002-7316/82/020346-10$1 .00/1

346

REPORTS

347

exceeds its range (the distance consumers are willing to travel to a supplier), a trader cannot operate from a fixed location; however, by becoming mobile and visiting several markets, a trader draws on more and different consumers and thereby temporally concentrates consumer demand by presenting the market opportunity less frequently. The consumer demand encountered in a daily market held in a town of 5,000 can be matched in a periodic market that draws on a population of 1,000 if that market is held only once every five days. Such markets, which generally meet in smaller population centers, have three interlocking advantages: they allow exploitation of low demand areas by mobile traders, they offer consumers a greater variety of goods by periodically concentrating supply, and by meeting on different days they facilitate the redistribution of commodities in the marketing system to and from higher-level centers (Smith 1974:181-186, 1976:15-18). While a general understanding of periodic markets will not necessarily explain a marketing system in its entirety, knowing how such markets function is a vital component in comprehending the whole. Although the data on indigenous market periodicity are thin (the lack of indigenous dates is apparently due to the complementary factors of rapid conversion to the European calendar and the Christian monopoly of literacy in New Spain), they can be profitably examined in light of our knowledge of the Aztec calendar. Calendrical systems have a great influence on the temporal patterning of markets. The frequency with which markets are held, and possibly even their relative locations, are determined not only by the logic of market efficiency, but by cultural factors such as calendar systems. One of the basic concerns of markethe s implicity. It is crucial that buyers and sellers alike know markets are held. Consequently, schedules that fit when the markets are held. Consequently, schedules that fit te fundamental units of time of a the given calendrical system are best and most easily remembered (Bromley et al. 1975:531). Regularity alone is insufficient. In our ownGregorian system, weekly or monthly markets (or some multiple thereof) are more easily remembered than a sequence that does not fit these temporal units. For example, with a 7-day week, a 6-day marketing schedule, though regular, would yield the confusing pattern of a market being held on a different day each succeeding week. Markets of many different frequencies are recorded or the Aztecs, from daily markets in large cities such as Tenochtitlan (Casas 1967:366; Motolinia 1971:375) to 5-, 8-, 9-, 13-, and 20-day markets (Casas 1967:366; Gibson 1964:352; Gomara 1966:151; Motolinia 1973:30; Torquemada 1975:559). Although most scholars do not use the entire 1-5-8-9-13-20 series (the most common is 1-5-13-20), it is often assumed that the largest centers held markets most frequently (daily), and the smallest centers held markets least frequently (every 20 days) (Bromley 1974:7; Gibson 1964:357; Kurtz 1974:690). Certainly, this correlation of town size and market frequency is pleasing in its simplicity and is consistent with marketing theory, but it finds only limited support in the Aztec data. In order to assess the significance of the recorded 1-, 5-, 8-, 9-, 13-, and 20-day sequences, as well as to understand why those particular intervals were used, a brief consideration of the Aztec calendrical system is in order (based on Caso [1971] for simplicity, with all Nahuatl terms corrected in accordance with Andrews [1975]). The central Mexican calendar is based on two major cycles, one of 260 days and one of 365 days, and within each there are other shorter cycles. In the 260-day cycle (the tonalpohualli-''the counting of the days"), each day is represented by a sign, or day name, and an accompanying number. There are 20 day names and 13 numbers which run concurrently. The 20 tonalpohualli day names, in conventional Aztec sequence are:
1. cipactli 2. ehecatl 3. calli 4. cuetzpalin 5. coatl 6. miquiztli 7. mazatl alligator wind house lizard snake death deer 11. 12. 13. 14. 15. 16. 17. ozomahtli malinalli acatl ocelot) cuauhtli cozcacuauhtli olin monkey grass reed jaguar eagle vulture (earth) quake

8. tochtli 9. atl 10. itzcuintli

rabbit water dog

18. tecpatl 19. quiahuitl 20. xochitl

flint rain flower

348

AMERICAN ANTIQUITY

[Vol. 47, No. 2,1982

The 13-day sequence lacks names, but each day is accompanied by one of the thirteen Lords of the Day:
1. 2. 3. 4. 5. 6. 7. Xiuhteuctli Tlalteuctli Chalchiuhtlicue Tonatiuh Tlazalteotl Mictlanteuctli Centeotl fire god earth god water goddess sun god goddess of love god of hell corn god 8. 9. 10. 11. 12. 13. Tlaloc Quetzalcoatl Tezcatlipoca Chalmecateuctli Tlahuizcalpanteuctli Citlalinicue water god wind god god of providence god of sacrifice lord of dawn goddess of the heavens, the Milky Way

Since the lowest common multiple of 13 and 20 is 260, a given combination of number plus day name will recur only after 260 days. This 260-day period is therefore the longest period within which the cycles of 13 and 20 will uniquely identify a day. In the 365-day cycle (the xiuhpohualli-"the counting of the years"), there are 18 named "months" of 20 days each, totalling 360 days, plus 5 additional days, called nemontemi or nentemi. The nemontemi days are intercalated at the end of the eighteenth month. Although which month began the year is unresolved (see Caso [1971] and Cline [1973] for opposing positions), the generally accepted sequence of the 18 xiuhpohualli months is:
1. Atl Cahualo 2. Tlacaxipehualiztli 3. Tozoztontli 4. Huei Tozoztli 5. Toxcati 6. Etzalcualiztli 7. Tecuilhuitontli 8. Huei Tecuilhuitl 9. Miccailhuitontli 10. Huei MiccailhuitI 11. Ochpaniztli 12. Pachtontli 13. Huei Pachtli 14. Quecholli 15. Panquetzaliztli 16. Atemoztli 17. Tititl 18. Izcalli water is abandoned flaying of men short vigil long vigil dry thing the eating of bean porridge small festival of the Lords great festival of the Lords small festival of the dead great festival of the dead sweeping of the road small Spanish moss big Spanish moss macaw (literally, it is rubber at the neck) raising of the flags it is the descent in the form of water shrunk or wrinkled regeneration (literally, sprout)

Both the tonalpohualli and the xiuhpohualli operate simultaneously, running in independent cycles. The multiples of the 20 day names, 13 day numbers, and 18 months (plus 5 days) generate unique combinations through 52 xiuhpohualli or 73 tonalpohualli cycles, that is, 18,980 days, and constitute the Calendar Round, after which the entire cycle begins anew. The xiuhpohualli years are named for the day sign and number on which they begin. The formal xiuhpohualli year is 360 days, of 18 complete 20-day cycles. The xiuhpohualli cycle does not include the nemontemi days, omitting 5 days each 365-day year, but the tonalpohualli is oblivious to the nemontemi days. Each 365-day year runs 18 complete 20-day cycles plus 5 additional days each solar year. Consequently, the years always begin on 1 of 4 days, known as the yearbearers, which are 5 days apart in the day-name sequence. These days-calli, tochtli, acatl, and tecpatleach designate a year 13 times during the Calendar Round (Serna 1892:313). An additional cycle of days, although by no means the only other one, is that of the nine Lords of the Night, the Yohualteuctin:
1. 2. 3. 4. Xiuhteuctli Itztli or Tecpatl Piltzinteuctli Centeotl fire god obsidian or flint the lord of princes, or child lord, the sun corn god

REPORTS

349

5. Mictlanteuctli 6. Chalchiuhtlicue 7. Tlazolteotl 8. Tepeyollohtli

9. Tlaloc

god of hell or of the dead water goddess goddess of love the jaguar "heart of the mountain"
rain god

This series of nine is generally attributed to the tonalpohualli, but does not fit into it evenly (unless there was a doubling of the gods on the last day of the cycle), running in continuous cycles apparently independent of both the tonalpohualli and the xiupohualli cycles (Serna 1892:345). The central Mexican calendar can be viewed as a series of different sized cogs, each turning at its own rate, but combining as the various cycles coincide, to create larger temporal units. The issue is how the 1-, 5-, 8-, 9-, 13-, and 20-day sequences fit the calendar. The 20-day sethe quence is based on the 20-day month, 13-day sequence is based on the 13-day Lords of the Day cycle, and the 9-day sequence is based on the 9-day Lords of the Night cycle. The 8-day sequence, however, does not fit the calendar in any transparent fashion. The probable explanation of its occurrence lies in the misinterpretation of early Spanish chronicles. Seven-day markets, based on the European calendar (initially Julian, then Gregorian), appeared after the Conquest. Although English practice states "every seven days" to describe a weekly event, Spanish practice includes the day the count is begun as well as the day it ends. Thus, "every seven days" is rendered "de ocho en ocho dias," literally every 8 days, but meaning weekly (e.g., Paso y Troncoso 1906:27). This, in conjunction with the absence of any mention of an 8-day market in Molina's (1970) classic Nahuatl dictionary, which does refer to 5-, 9-, and 20-day markets, and the lack of supporting evidence from sixteenth-entury chronicles and colonial documents, indicates that the 8-day market was probably a post-Conquest chimera. The 5-day sequence is not transparent in the Aztec calendar either, but it does constitute a basic unit of time. In the 20-day month, the 4 yearbearer days have a greater significance than the a remaining 16 days, each yearbearer beginning 5-day segment of the month (Sahagpun1957:138-139). Thus, despite not being a formally recognized cycle, as were the 9-, 13-, and 20-day cycles, the 5-day cycle emerges from the way the calendar functioned. The frequency with which sixteenth-century chroniclers and colonial documents refer to the 5-day market, as well as its reference in Molina, leave no doubt as to its existence and importance. While it has been established that 1-, 5-, 9-, 13-, and 20-day markets existed, how these fit together and what they meant is still unclear. That there were articulating periodic markets is repeatedly supported by the sixteenth-century chronicles and later colonial practices, but the assumption that the 5-, 9-, 13-, and 20-day cycles represent similar market cycles, as the sixteenthcentury accounts seem to indicate and as modern scholars have assumed, is misleading. These cycles represent two distinct types of marketing sequences. If the 5-, 9-, 13-, and 20-day cycles were actual marketing cycles, as has been assumed, what would be the consequences? The use of all these cycles, even at different levels in the marketing hierarchy, would result in adjacent markets meeting on the same day with relative frequency, thus undermining the primary advantage of periodic markets (see Figure 1). Furthermore, such conflicts would appear to occur irregularly, so that deleting one of the conflicting markets would introduce an undesirable element of randomness into the system. The problem would not arise with just the 20-day cycle, as it is a regular multiple of the 5-day cycle, but the uneven number of days in the 9- and 13-day cycles would lead to numerous conflicts. Although the sequence of the resultant conflict between adjacent markets can be calculated, it is a sophisticated matter involving patterns of 9 and 13 years, respectively. More concretely, the market deletions necessary to avoid conflicts would appear to be irregular.1 Now let us consider what these cycles meant and how, in fact, they functioned. The 5-day sequence was the basic Mesoamerican marketing unit. A market might be held on any day within the 5-day unit. For example, a town might have its market on the third day of the 5-day sequence or twice within the sequence, as was the case in Zapotitlan (Motolinia 1973: 110-111), just as we may have a Saturday market, or a Tuesday and Friday market. Daily markets

GALLI CUETZPALIN COATL MIQUIZTLI MAZATL

TOCHTLI

(Q8

2<)(9 9 10 II 12
13

<3) 10 (3 11 (1)6 13 7

<1><4) II 5 12 13 5 6 7 8
2 9

) (5) 12 13 (j 8 (
3

6) 13 3 2 3 4 5 6 7 () [] 2 9 10 t1
12

<3)

) 8 9 () 11
12

2 3 4 5
6

3 4 5 6

7 8 9 10
II

E1
2 3 4

) 3
4

E
3 4 ()

9
10

5
6

ATL ITZCUINTLI OZOMAHTLI MALINALLI ACATL OCELOTL CUAUHTLI

COZCACUAUHTLI

7 8 9

EL
3 4

F(1)FI
8 2 3 4 5 6 7 9 10 11 12 13

()
6 7 8 9 10
II

9 10 11 () 12

4 II 12 5 6 7 8 9 10 ) 12 13 7 8 [

11 12 13

12

13

7 8

2 3 4 5

9 10

6 7

13 7 2 3 4 5

II 12 13

5 6 7 8 9 10

9 3 4 5 6 78 8 ( 10

EL
2
3

8 9
10

3)
4 5 6 7 8 9

12

L
2

0
9 10

12 13 7

)
7

7L
2 3 4 5

OLIN TECPATL QUIAHUITL XOCHITL

4 5 6 7 ()

II $ 12 ( 13

13

2 2 3 4 9 10 II 3 4 5

4 5 6

II 12 13

[7]
2 3

(Q)
12

E
2

as periodic markets. Figure 1. Pattern of markets if all four types-5-, 9-, 13-, and 20-day markets-operated ning on the initial day, and would vary for other towns in the same marketing system. Underlining indicates 5-da markets; squares indicate 13-day markets; and hexagons indicate 20-day markets.

REPORTS

351

were a subset of the 5-day market, merely meeting every day to ensure the provisioning of the largest centers, an example of periodic markets filling up the calendar until there was a continuous market. Note, however, that Tenochtitlan's daily market had a "bulge" in attendance on a 5-day schedule (Torquemada 1975:559), as expected (Bromley et al. 1975:536-537). However, the 13, and 20 days-differed fundamentally from the 5-day seremaining market sequences-9, quence. The 9-, 13-, and 20-day markets were not cycles indicating the temporal spacing of markets. Rather, they were ritual temporal units with markets held on the first day of their respective cycles, and only on the first day. The 20-day cycle fits the calendar system neatly, but if it actually represented a temporal marketing unit, so that 20-day markets were held each day somewhere, they would be held on the nemontemi days, an unpropitious time when all nonessential activity was curtailed. Further, the infrequency of a 20-day market, as well as its spacing between more frequent 5-day markets, argue against its existence as an actual 20-day sequence. Rather than an interval indicating temporal market spacing, but not the specific day of the market, the 20-day market means once per 20-day month, on the first day. Ixtlilxochitl (1975:283) states that among the Toltecs the 20-day market was held on the first day of each month. Given the striking calendrical (Caso 1967) and ritual (Caso 1968; Kirchhoff 1968) similarity, if not unity, throughout Mesoamerica, this practice appears likely to have persisted for that reason as well as for others to be discussed later. A festival is associated with the 20-day month, but there is a certain amount of debate over which day the 20-day festival was celebrated, i.e., whether it was the first or the last day of the cycle. Despite Broda's (1969:33) statement that festivals began on the last day of the 20-day month, Ixtlilxochitl clearly states the market to have been held on the first day of the month. The discrepancy is whether it was day 1 or day 20 of the month, and I believe it was probably the former, for it would then correspond not only to the yearbearer days, but more specifically to the yearbearer the days which denoted the current xiuhpohualli year, making the event all the more important (Bromley et al. 1975:534). Unfortunately, sixteenth-century sources differ on the subject, some favoring the last day of the month (Motolinia 1973:25) and some the first (Duran 1967:225; Sahagun 1974:21). (See also the interpretation of the Tovar calendar by Kubler and Gibson [1951:44].) In addition to the support provided by most sixteenth-century sources for the first-day interpretation, and the enhanced status of the market falling on a yearbearer day, two additional considerations favor a first-day interpretation. First, there is a certain indeterminancy as to when the Aztec day began. Caso (1971:345) gives the four logical possibilities as midnight, sunrise, sunset, and noon, settling on the last as the most likely, a conclusion that makes the selection of a specific the Gregorian day not quite accurate. Second, in the Maya area, festivals "seat" by beginning on the day before the formal festival day (Goubaud Carrera 1935:46). Thus, the conflicting positions may have arisen by confusing the festival, variously given as the first and last day of the 20-day cycle, with the market, unambiguously stated as being held on the first day. Data concerning the 9- and 13-day markets are scarcer than those for the 20-day markets, but the shorter cycles also mark fundamental cyclical units and appear similarly ritual in nature, since they are based on ritually related calendrical units-the 9 Lords of the Night cycle and the 13 Lords of the Day cycle-as is the 20-day cycle, in marked contrast with the 5-day cycle. In all probability, the 9- and 13-day markets also met only on the first day of their respective cycles. Considering the differences in function of these sequences, it is misleading to lump together 5-, 9-, 13-, and 20-day markets. Although the 5-day market indicates a temporal unit within which markets were regularly held on any day(s), the same was not true for the 9-, 13-, and 20-day markets. Held on the first day of their respective cycles, the latter markets were relatively infrequent and were uniform throughout the area, corresponding to the dictates of the calendar. Thus, their significance for articulating markets was minimal. What is the significance of all this for Precolumbian marketing? Knowing the calendrical units on which a market system is based, if indeed it is a system, does not permit a fortiori deduction of

352

AMERICAN ANTIQUITY

[Vol. 47, No. 2,1982]

the sequence and articulation of markets. However, it should be obvious that examination of such units may reveal something of the temporal structure involved. The frequency of a market generally indicates the population it serves; the larger the population, the more frequent the market. Thus, daily markets in central Mexico were held in the largest centers, such as Tenochtitlan, Texcoco, and Tlaxcala (Torquemada 1975:559; Motolinia 1971:375). Five-day markets were the norm for most towns (Torquemada 1975:559; Motolinia 1971:375; G6mara 1966:151). However, in towns holding daily markets, the 5-day sequence was cumulative, and the 5-day market was larger and more important than the daily one (Casas 1967:366). Which day of the 5-day sequence was the market day in specific towns in unknown. Basic market principles would indicate, however, that lower-level markets did not conflict with adjacent markets of the next higher level, but rather, dovetailed with their schedules, a pattern supported by the extant data. Within the 5-day sequence, the most important day was the yearbearer day. Thus, it is plausible to hold that the main town, or cabecera, of a region held its market day then, just as, in 1547, the city of Tlaxcala claimed exclusive use of Saturday for its market in the state of Tlaxcala (Anderson et al. 1976:123). Is there, however, any support for such a pattern, beyond the logic of it? Clavijero (1974:235) recorded the market days for Mexico City (since it had several major markets, this presumably referred to the most important of them-probably Tlatelolco). He lists Mexico City's market days as the third, eighth, thirteenth, and eighteenth days of the month, not the first, sixth, eleventh, and sixteenth, as one would assume if they were yearbearer days. However, he lists the market days as calli, tochtli, acatl, and tecpatl. Since these are the yearbearers, it is clear that Clavijero's statement about the third, eighth, thirteenth, and eighteenth days refers not to the yearbearers as they function in the system but to their order in the indigenous conventional sequence. Mexico City's market days did, then, meet on the first, sixth, eleventh, and sixteenth days of the month, the yearbearer days, leaving the remaining days for the subordinate markets. This supports the theory that major centers held the yearbearer days. Unfortunately, the paucity of marketing data recorded in the indigenous calendrical system prevents a fuller examination of this issue. Given articulating markets based on a 5-day sequence and focusing on major markets on yearbearer days, what is the place of the 9-, 13-, and 20-day markets? These markets were most likely not independent, but cumulative, superimposed on the 5-day sequence much as the 5-day market is superimposed on the daily markets. Cumulative celebrations were common in the Aztec system (Duran 1967:83, 126). Furthermore, extant documents record the occurrence of 20-day markets in larger cities rather than in small towns, just as one would expect if they were truly periodic markets, their infrequency indicating small population and small consumer demand. The neatest fit with the 5-day markets was by the 20-day market. The 20-day market, held on the first day of the month, would have coincided with the first yearbearer of the month. Although the name of the xiuhpohualli year rotated among the four yearbearers, the first yearbearer of each month was also the yearbearer after which that 365-day year was named. Presumably, this market was larger and of more importance than the other markets, including the other yearbearer markets, just as the 5-day market was more important than the daily markets. The 9- and 13-day markets do not fit as neatly into the annual cycle. The 9-day market, if it was held on the first day of the Lords of the Night, as appears to have been the case (Duran 1967: 233-234), would fall on a yearbearer market once every 9 cycles (every 45 days-every 5 cycles of 9 and every 9 cycles of 5) and on a 20-day/yearbearer market once every 36 cycles, but since the 9-day cycle does not fit the annual 365-day cycle evenly, its occurrence varied each year throughout a 9-year cycle. The 13-day cycle, likewise, occurred on yearbearer market days once every fifth 13-day cycle and on a 20-day/yearbearer market every 260 days, or 20 cycles; and since it was not evenly divisible by the 365-day year, its schedule varied each year through a 13-year cycle. The 13- and 9-day cycles coincide every ninth 13-day cycle or every thirteenth 9-day cycle. The lack of better articulation introduces an element of randomness in the system if the 9- and 13-day markets are considered fully functional markets like those of the 5-day sequence. If the 9- and 13-day markets coincided with an existing cycle, as with the 5- and 20-day

REPORTS

353

markets, there would be little difficulty. However, in a sequence of "1 market day-4 nonmarket marketmarkets would not only conflict with neighboring days-1y 5-day markets but create an additional market too close to the established markets to generate either demand or produce. For example, in a market system containing five towns, the introduction of an additional market (either 9-day or 13-day) would result in considerable confusion. If all towns celebrated it, none would attract sufficient numbers to its market. If a single center celebrated it (most likely the politically dominant center), it would conflic wwith existing market an elsewhere four out of five times, and it would be temporally near the dominant center's market four out of five times, thus reducing either its attraction or that of the conflicting 5-day market. The 9- and 13-day markets were probably celebrated only in those towns whose regular 5-day markets coincided with the 9- and 13-day markets. This practice of selective celebration would have been compatible with existing patterns, as it is clear that festivals varied in importance throughout central Mexico (DurGan 1967:32, 72, 125, 189, 270). Only one town in a marketing system would celebrate that 9- or 13-day market, making its regular market slightly more important than it would ordinarily be, and avoiding the economic chaos of every town holding a market on the ninth and thirteenth days-similar to the way in which the 20-day markets operate. Thus, the system as reflected in the documents, the calendar system, and the logic of marketing, appears to have been one of daily markets in major cities and articulating 5-day markets elsewhere, with the yearbearer 5-day markets being held in the dominant center of the area. The 9-, 13-, and 20-day markets, held on the initial day of their respective cycles, were not celebrated in every town, but only in those towns with whose market days they coincided. The 20-day markets may have been held only in large centers, as evidence of their occurrence in Tulanzingo and Huitzilopochco (Gibson 1964:357) suggests (and also in Toltec Tula, Teotihuacan, Tulancingo, Cuauhnahuac, Cholula, and Tultitlan, according to Vaillant 11966:841),whereas the 9- and 13-day markets may have rotated among the various markets of the region. This cumulative market celebration permitted special markets to be held without generating chaos in the system. What this does indicate, however, is a marketing region of basically two levels-important centers all holding their markets on the same days, the yearbearers, and their subordinate towns with markets on the nonyearbearer days. While it has been shown that use of the 5-, 9-, 13-, and 20-day cycles as periodic marketing cycles would result in apparently irregular conflicts between adjacent markets, logically, the use of a 5-day cycle alone, with the ritual 9-, 13-, and 20-day cycles being cumulative rather than independent, permits a marketing pattern which would not conflict with adjacent markets, even if the yearbearer day is reserved for the cabecera. While this consideration of Aztec marketing in light of their calendar system does not present a clear picture of what the entire system was, it does clarify some misleading data. Distinctions must be drawn between functional periodic markets (5-day) and ritual markets (9-, 13-, and these 20-day) and he lockstep nature of both ritual markets and of cabecera yearbearer markets must be taken into account. Market articulation above the cabecera level was not a matter of periodicity. Further consideration of the Aztec system cannot rely solely on town size as the indicator of market importance and articulation, but must take into account market scheduling and seek additional data detailing actual consumer interaction. NOTE This is similarto our situationwith respect to Easter.Easteris celebrated on the first Sundayafter the full moon following the vernal equinox. In other words it is loosely tied to Passover and is set according to the Jewish lunar calendar. Thus, despite its regularity, Easter appears irregular from the perspective of the Acknowledgments. I would like to express my appreciationto G. William Skinnerand JamesA. Fox for

Gregorian calendar.

having read and helpfully commented on an earlier draft of this paper, and to I. Richard Andrews for assisting me with Nahuatl. The research on which this paper was based was funded by Grant # PRA 57781 from the

Organizationof American States and by a grant from the Departmentof Anthropology, StanfordUniversity.

354

AMERICAN ANTIQUITY

[Vol. 47, No. 2,1982]

REFERENCES CITED Anderson, Arthur O., Francis Berdan, and James Lockhart J. 1976 Beyond the codices. University of California Press, Berkeley and Los Angeles. Andrews, J. Richard 1975 Introduction to classical Nahuatl. University of Texas Press, Austin. Broda, Johanna 1969 The Mexican calendar. Acta Ethnologica et Linguistica No. 15. J. Bromley, O., Richard Symanski, and Charles M. Good 1975 The rationale of periodic markets. Annals of the Association of American Geographers 65:530537. Bromley, R. J. 1974 Marketplace trade in Latin America. Latin American Research Review 9:3-38. Casas, Bartolom6 de las 1967 Apolog6tica historia sumeria (Vol. I). Universidad Nacional Aut6noma de M6xico, Mexico, D.F. Caso, Alfonso 1967 Loscalendarios prehisp6nicos. Universidad Nacional Aut6noma de M6xico, Mexico, D.F. 1968 ,Religi6n o religiones mesoamericanas? Internationalen Amerikanistenkongresses 38, 3:189-200. Stuttgart-Miinchen. 1971 Calendrical systems of central Mexico. In Handbook of Middle American Indians (Vol. 10), edited by G. F. Ekholm and Ignacio Bernal, pp. 333-348. University of Texas Press, Austin. Christaller, Walter 1972 How I discovered the theory of central places: a report about the origin of central places. In Man, space, and environment, edited by Paul Ward English and Robert C. Mayfield, pp. 601-610. Oxford University Press, New York. Clavijero, Francisco Javier 1974 Historia antigua de M6xico. Porrfia, Mexico, D.F. Cline, Howard F. 1973 The chronology of the conquest: synchronologies in Codex Telleriano-Remensis and Sahagiun. Journal de la Soci6te des Am6ricanistes 62:9-34. Duran, Diego 1967 Historia de las Indias de Nueva Espana e Islas de la Tierra Firma (Vol. I). Porrua, Mexico, D.F. Gibson, Charles 1964 The Aztecs under Spanish rule. Stanford University Press, Stanford, Calif. Gomara, L6pez de 1966 Historia general de las Indias (Vol. II). Iberia, Barcelona. Goubaud Carrera, A. calendarica indigena. Anales de la Sociedad de Geografia e 1935 El "Guajxaquip Bats"-ceremonia Historia de Guatemala 12:39-52. Ixtlilx6chitl, Fernando de Alva 1975 Obras hist6ricas (Vol. I). Universidad Nacional Aut6noma de M6xico, Mexico, D.F. Kirchhoff, Paul 1968 Las 18 fiestas anuales en Mesoamerica: 6 fiestas sencillas y 6 fiestas dobles. Internationalen Amerikanistenkongresses 38, 3:207-221. Stuttgart-Miinchen. Kubler, George, and Charles Gibson 1951 The Tovar calendar. Memoirs of the Connecticut Academy of Arts and Sciences (Vol. II). Kurtz, Donald V. 1974 Peripheral and transitional markets: the Aztec case. American Ethnologist 1:685-705. Molina, Alonso de 1970 Vocabulario en lengua castellana y mexicana y mexicana y castellana. Porrfua, Mexico, D.F. Motolinia (Toribio de Benavente) 1971 Memoriales o libro de las cosas de la Nueva Espana y de los naturales de ella. Universidad Nacional Aut6noma de Mexico, Mexico, D.F. 1973 Historia de los Indios de la Nueva Espana. Porrfia, Mexico, D.F. Paso y Troncoso, Francisco del (editor) 1960 Papeles de Nueva Espana (Vol. 7). Madrid and Mexico, D.F. Sahagfin, Bernardino de 1957 General history of the things of New Spain: Florentine Codex, Books 4 & 5, translated by Arthur J. 0. Anderson and Charles E. Dibble. University of Utah Press, Salt Lake City. 1974 'Primeros Memoriales' de Fray Bernardino de Sahagun, edited by Wigberto Jimenez Moreno. Instituto Nacional de Antropologia e Historia, Mexico, D.F. Serna, Jacinto de la 1892 Manual de ministros de indios para el conocimiento de sus idolatrias y extirpaci6n de ellas. Anales del Museo Nacional de M6xico VI:261-475. Mexico, D.F.

REPORTS

355

Smith, Carol A. 1974 Economics of marketing systems: models from economic geography. In Annual review of anthropology (Vol. 3), edited by Bernard J. Siegel, Alan R. Beals, and Stephen A. Tyler, pp. 167-201. Annual Reviews Inc., Palo Alto, Calif. 1976 Regional economic systems: linking geographical models and socioeconomic problems. In Regional analysis (Vol. I), edited by Carol A. Smith, pp. 3-63. Academic Press, New York. Smith, Michael E. 1979 The Aztec marketing system and settlement pattern in the Valley of Mexico: a central place analysis. American Antiquity 44:110-125. Torquemada, Juan de 1975 Monarquia indiana (Vol. II). Porrfua, Mexico, D.F. Vaillant, George C. 1966 Aztecs of Mexico. Penguin Books, Baltimore.

MANUFACTURE OF MESOAMERICAN PRISMATIC BLADES: AN ALTERNATIVE TECHNIQUE John E. Clark

Recent analyses of primary documentary sources have demonstrated that Don Crabtree's technique of prismatic blade manufacture differs significantly from that of the ancient Mexicans. The most widely known of these ethnohistoric descriptions of blademaking, as well as several previously not considered, are reevaluated and compared to the Crabtree technique. Major discrepancies between the Aztec technique and that described by Crabtree then became the focus of replication experiments. Finally, prismatic blades were successfully produced in the manner and with the tool described by the early Spanish friars.

The method used to produce the obsidian blades common at Mesoamerican sites has long intrigued anthropologists (cf. Courtis 1865; Joly 1883; Stoll 1886; Tylor 1861). Several Spanish friars described the technique, but contemporary experimentation has shown the descriptions to be unworkable (Barnes 1947; Cabrol and Coutier 1932; Crabtree 1968; Ellis 1940). However, Crabtree made a breakthrough and was able to produce exact replicas of the blades using a chest crutch and a vise. He suggested that portions of Juan de Torquemada's account may have been miscopied or mistranslated and, theorfore, he postulated several minor changes which were more in accord with his own experiments (Crabtree 1968). Crabtree has since been criticized for his alleged misuse of the ethnohistoric data (Feldman 1971; Fletcher 1970). Additional ethnohistoric material has been brought forward in these critiques and, consequently, more descriptions of pressureblade production are now available. Although Crabtree's technique works, it does not accord with the evidence now at hand (Sheets 1977:143-144). In this paper an alternative method is described for making prismatic blades, based upon a reevaluation of the ethnohistoric sources. Previous experimental work will be summarized and compared to the accounts of the Spanish chroniclers. The final portion of the paper will describe a technique that resolves most of the discrepancies between previous work and the written descriptions.
John E. Clark, New World Archaeological Foundation-Brigham Young University, San Crist6bal de las Casas, Chiapas, Mexico Copyright ? 1982 by the Society for American Archaeology 0002-7316/82/020355-22$2.70/1