Archaeological Evidence Overridden by the Preconceived Significance of the Equinoxes

One of the earliest proposals about astronomical motives underlying the layout of Mesoamerican buildings was published in the 1920s, suggesting that Group E of Uaxactun, Guatemala, was an astronomical observatory, intended specifically for sighting the equinoctial and solstitial sunrises (Ricketson Reference Ricketson1928a, Reference Ricketson1928b). In the following decades, a number of other architectural compounds with a similar configuration have been found, most of them in the central Maya Lowlands. Commonly labeled E Groups, these architectural complexes deserve special attention because the early astronomical interpretation of Group E of Uaxactun was soon extended to other similar assemblages and may have been the very origin of the widespread idea about the importance of the equinoxes in the Maya area and in Mesoamerica in general. The astronomical uses commonly attributed to E Groups represent probably the most eloquent example of how a traditional prejudice is given greater weight than materially evidenced facts.

A typical E Group has a symmetrical ground plan, with its central axis running approximately east–west, from a pyramidal temple on the west side of a plaza to the center of an elongated platform that delimits the plaza on its east side and extends in a roughly north–south direction. The latter has no superstructures in the so-called La Venta type, which is the earliest variant of this assemblage. The Cenote type has a central building on the platform, slightly set backwards, whereas the Uaxactun type, the latest version of the complex, has two additional buildings on the platform's extremes (Chase and Chase Reference Chase, Chase, Freidel, Chase, Dowd and Murdock2017; Inomata et al. Reference Inomata, Triadan, Pinzón, Burham, Ranchos, Aoyama and Haraguchi2018). Most E Groups were built during the Preclassic period (around 1000 b.c. to a.d. 250).

For the eponymous Group E of Uaxactun, Ricketson (Reference Ricketson1928a, Reference Ricketson1928b) argued that, observing from the substructure of the western pyramid (Structure E-7-sub), the outer edges of Temples E-1 and E-3 on the extremes of the eastern platform recorded solstitial sunrises, whereas the equinoctial Sun rose over Temple E-2 in the middle of the platform. Soon afterwards, Ruppert (Reference Ruppert, Hay, Linton, Lothrop, Shapiro and Vaillant1940) noticed that other compounds of this type have different orientations and suggested that they had more ceremonial than observational functions. Accordingly, some scholars interpreted them as allegorical imitations of the astronomically functional template at Uaxactun (Aimers and Rice Reference Aimers and Rice2006; Aveni and Hartung Reference Aveni, Hartung and Aveni1989; Awe et al. Reference Awe, Hoggarth, Aimers, Freidel, Chase, Dowd and Murdock2017; Fialko Reference Fialko1988; Guderjan Reference Guderjan2006). On the other hand, Aveni and colleagues (Reference Aveni, Dowd and Vining2003) argued that the alignments incorporated in E Groups were astronomically functional, largely corresponding to solar positions on different agriculturally important dates. Nonetheless, and although it soon became evident that many E Groups were earlier than the supposed prototype at Uaxactun, it was still argued that their initial purpose was to record the solstices and equinoxes, but that this observational function was subsequently abandoned and replaced by a predominantly ritual use (Aimers and Rice Reference Aimers and Rice2006; Guderjan Reference Guderjan2006). It was also suggested that stone columns or wooden poles placed on the eastern platform could have served as markers of the equinoxes or solstices (Aimers and Rice Reference Aimers and Rice2006:80; Rice Reference Rice2007:155), although no specific evidence supports this hypothesis. All these conjectures share the prejudice that the only potentially significant moments of the tropical year were the equinoxes and the solstices.

E-Group assemblages have been found in a large number of archaeological sites, both in the Maya Lowlands and along the Pacific and Gulf coasts and adjacent highlands. Given the long-known diversity of their orientations, as well as the study by Aveni and colleagues (Reference Aveni, Dowd and Vining2003), it is surprising how often we can still read, even in some contributions in the recently published book on E Groups (Freidel et al. Reference Freidel, Chase, Dowd and Murdock2017), that the alignments incorporated in complexes of this type marked the equinoxes and solstices. Although some alignments approximate the equinoctial directions, such statements are inconsistent with what is known about the orientations of both E Groups and other buildings.

Supposing that the equinoctial alignments were inaccurate and more symbolic than observationally functional, they would have been distributed evenly both south and north of the true equinoctial direction; in fact, the orientations of most E Groups, indicated by their central axes, are skewed south of east, in agreement with the long-known trend observed in buildings of different types throughout the Maya Lowlands, as well as in other parts of Mesoamerica (Aveni Reference Aveni2001; Macgowan Reference Macgowan1945).

Conversely, even though the intended orientations of particular structures, due to their present state of conservation, often cannot be accurately determined, the analyses of large samples of alignment data from different Mesoamerican regions have revealed that the orientations marked astronomical events quite precisely. The solar orientations, which prevail, allowed the use of observational calendars composed of multiples of 13 or 20 days and intended for facilitating the scheduling of agricultural and related ritual activities. Low-precision alignments would have been useless for such purposes. Since we thus know which dates were most frequently recorded and particularly significant, it is worth noting that in several buildings, where the alignments that most likely served for observations are sufficiently well-preserved and can be measured with precision (e.g., the axes of symmetry of upper structures, indicated by double pairs of jambs), these dates are recorded with surprising accuracy (Figure 2; Šprajc Reference Šprajc and Ruggles2015b:Figures 52.3–52.8; Šprajc and Sánchez Reference Šprajc and Nava2015:Figure 4.22; Šprajc et al. Reference Šprajc, Nava and Ortiz2016:Figures 4.7 and 4.14).

More specific conclusions regarding the astronomical significance of E Groups have been reached in a recent study of alignments in 71 complexes of this type in the central Maya Lowlands (Šprajc Reference Šprajc2021a, Reference Šprajc2021b). The analyses have shown that, although built primarily for ritual purposes, these assemblages were astronomically functional, but their use was not essentially different from that of other astronomically oriented buildings. While the distribution of lateral alignments (from the western building to the extremes of the eastern platform) offers no support to the idea that they were used systematically for sighting celestial events, the orientations of E Groups marked by their central axes belong to widespread alignment groups, mostly materialized in buildings of other types. With very few exceptions, the central axes lie within the angle of annual solar movement along the horizon, matching various dates commonly recorded by other buildings, often at one and the same site, but the equinoxes are not among them. In Figure 4, which shows the distribution of declinations marked by the central axes, there is no concentration centered on 0°, whereas the peak at 0.8° suggests that some near-equinoctial alignments targeted sunsets on the quarter-days. In sum, while a few E Groups were oriented to solstitial positions of the Sun, their relationship with the equinoxes can be safely rejected.

Figure 4. Relative frequency distributions of declinations marked on the eastern and western horizon by the central axes of 71 E Groups in the central Maya Lowlands; only declinations within the solar span are shown (based on data published in Šprajc Reference Šprajc2021a).

Paradoxically, even Group E of Uaxactun, which originated the most popular hypothesis about the astronomical significance of the compounds of this type, cannot be related to the equinoxes and solstices. According to archaeological excavation data, amply discussed elsewhere (Šprajc Reference Šprajc2021a, Reference Šprajc2021b), the allegedly solstitial and equinoctial alignments connect buildings from different periods and thus could have never been observationally functional (Kováč et al. Reference Kováč, Karlovský, Desprat, Navarro, Arroyo, Salinas and Paiz2015:1034; Rosal and Valdés Reference Rosal, Valdés and Valdés2005; Rosal et al. Reference Rosal, Valdés, Laporte, Laporte and Valdés1993; Valdés Reference Valdés1987).

The popular associations of E Groups with the equinoxes represent probably the most notorious example of preponderance of preconceptions over the available archaeological data. To mention but two more cases, Vadala and Milbrath (Reference Vadala and Milbrath2014) suggest that the orientation of the Late Preclassic Structure 4 of Cerros, Belize, targeted the equinoctial sunrises. As they recognize that its east–west axis, with an azimuth of 88°, corresponds to sunrises several days after/before the equinox, they assume that, due to the Sun's fairly rapid movement along the horizon around the equinoxes, the builders were able to establish only an approximate equinox alignment, perhaps by bisecting the angle delimited by solstitial directions, or by halving the time span between the solstices. However, given the difficulties mentioned above and the procedures leading to inaccurate results, it is hard to understand how the builders could have been aware of the equinox in the first place.

Similarly, Rice and Pugh (Reference Rice and Pugh2017:12) suppose that the orientations skewed 3–4° south of east, shared by the Middle and Late Preclassic E Groups at Nixtun-Ch'ich', Guatemala, and several other sites, suggest a calibration to an approximately eight-day period before the vernal and after the autumnal equinox. They also presume that the equinoxes, being difficult to determine with empirical observations, might have been estimated by counts of days from the solar zenith passages. But again, if the observers were unable to fix the equinox, they could not have known how many days to count. Instead, the orientations of E Groups of Nixtun-Ch'ich', apparently determinant of the main east–west axis of the site, belong to the most widespread alignment group materialized in the complexes of this type, corresponding to sunrises on March 11 and October 2, separated by 160 (= 8 × 20) days (Šprajc Reference Šprajc2021a, Reference Šprajc2021b).

Alignments Not Evidenced in the Archaeological Record

In some cases, two of which will be discussed below, the alignments claimed to be equinoctial are not indicated in material remains, or the observational scheme proposed has no support in archaeological evidence.

At Dzibilchaltún, in the northwestern extreme of the Yucatan Peninsula, Structure 1-sub, also known as the Temple of the Seven Dolls (Figure 5; Andrews and Andrews Reference Andrews and Wyllys Andrews1980), is one of the increasingly popular focuses of modern equinoctial pilgrimages, which reflect a common belief that the passage shaped by four east–west aligned doorways of the building was intentionally oriented to the rising Sun on the equinoxes (Casares Reference Casares Contreras2016; Dowd Reference Dowd, Dowd and Milbrath2015; Maldonado and Hernández Reference Maldonado Cárdenas and González2013; Méndez et al. Reference Méndez, Barnhart, Powell and Karasik2005). The widely publicized photographs showing the solar disk nicely framed by the doorways contribute to the popularity of the event. However, virtually identical photos can be taken from different points and on various consecutive days, but only after the Sun has reached a considerable altitude, because the orientation of the temple, skewed about 1° clockwise from cardinal directions, does not correspond to the equinoctial position of the Sun on the horizon. Since an exhaustive discussion of this case was presented elsewhere (Sánchez and Šprajc Reference Sánchez Nava and Šprajc2015:140–142; Šprajc and Sánchez Reference Šprajc and Nava2018), I will summarize only the most relevant facts.

Figure 5. Dzibilchaltún, Yucatan, Mexico, Temple of the Seven Dolls, looking southeast. Photograph by the author.

The causeway leading westward from the temple (Sacbé 1) is skewed 1°40' north of due west, and its central axis extended eastward passes a few meters north of the center of the temple (Andrews and Andrews Reference Andrews and Wyllys Andrews1980:14, Figure 2; Stuart et al. Reference Stuart, Scheffler, Kurjack and Cottier1979). Therefore, the convenient points for observing the phenomenon are located along a line running considerably south of—but not exactly parallel to—the central axis of the causeway, as well as several meters south of Stela 3, which stands on a platform in the middle of Sacbé 1, some 130 m west of the temple (Figure 6), and would thus appear to have been an appropriate marker of the observation spot. As a consequence, the multitudes visiting the site on the equinoxes concentrate in a space south of the stela, evidently searching for the most convenient spot for sighting the Sun through the temple's doorways. The central doorway of Structure 7, located about 40 meters west of Structure 1-sub, might represent another archaeologically preserved marker for observing the equinox Sun through the opening shaped by the four east–west aligned doorways of Structure 1-sub. However, from that vantage point, given the size of the opening, the Sun can be seen during various consecutive days at different altitudes, but there is no architectural element that would have allowed the observer to pinpoint the equinox date.

Figure 6. Dzibilchaltún, Temple of the Seven Dolls, looking west along the axis of symmetry of the western pair of doorways, with Structure 7, Stela 3, and Sacbé 1 in the background. Photograph by the author.

Instead, the orientation of the Temple of the Seven Dolls was likely functional in the western direction. Skewed about 1° north of west, it pertains to the widespread alignment group for which the sunsets on quarter-days of the year represent the only conceivably significant target. These dates could have been determined not only by direct observation of the setting Sun along the axis of symmetry of the four doorways (Figure 6), but also by observing light-and-shadow effects produced by pairs of windows and smaller openings in the eastern and western walls of the temple (Šprajc Reference Šprajc1995).

The ruins of the Templo Mayor of the Aztec capital of Tenochtitlan, in modern downtown Mexico City, represent another illustrative case. Referring to this building, the sixteenth-century friar Toribio de Benavente o Motolinía (Reference Motolinía and O'Gorman1971:51) writes that the Aztec feast of Tlacaxipehualiztli fell when the sun was in the middle of Uchilobos, which was the equinox. Since the temple is skewed notably south of east, Aveni and Gibbs (Reference Aveni and Gibbs1976) and Aveni and colleagues (Reference Aveni, Calnek and Hartung1988) proposed that the equinoctial observations were made from a spot at ground level west of the temple; given the building's height, the rising Sun would have appeared in the notch between the two upper sanctuaries only after it had moved considerably south of the east–west line. However, as extensively argued before (Šprajc Reference Šprajc2000b, Reference Šprajc2001:402–410), the hypothesis is hardly acceptable because, as the temple grew in height, the observation point would have had to be moved either farther away or to a higher level above natural ground; neither of the two scenarios finds support in archaeological evidence. Precise measurements in the Templo Mayor precinct revealed that one of the two sunset dates matching the east–west axis of the temple's late construction stages is April 4, which in the sixteenth-century Julian calendar corresponded to March 25, which in 1519 was the last day of the month of Tlacaxipehualiztli, and we can recall that the main feast of every month was celebrated on its last day (Caso Reference Caso1967; Prem Reference Prem2008). On the other hand, March 25, the Feast of the Annunciation, was commonly identified in medieval Europe with the vernal equinox (the canonical date of the ecclesiastical equinox established in a.d. 325 by the Council of Nicaea was March 21, but the Roman tradition associating the equinox with March 25 also survived: McCluskey Reference McCluskey, Ruggles and Saunders1993; Newton Reference Newton1972:22–27). We can thus conclude that Motolinía (Reference Motolinía and O'Gorman1971:51) did not refer to the astronomical equinox (the date of which would hardly have been known to a non-astronomer at that time), but rather only noted the correlation between the day of the Mexica festival—which in the last years before the Conquest coincided with the sunset along the axis of the Templo Mayor—and the date of the Christian (Julian) calendar that corresponded to the traditional day of the spring equinox. The dates marked by the building could be incorporated in an observational scheme composed of calendrically significant intervals. Also worthy of note is that orientations similar to that of the late stages of the Templo Mayor of Tenochtitlan (5°36' south of east) were identified in the neighboring area of Texcoco (Šprajc Reference Šprajc2001).

Interpretations Based on Deficient or Erroneous Alignment Data

As discussed earlier, any assessment of putative astronomical alignments must be based on reliable data (azimuth, horizon altitude). Since incomplete or imprecise data are given in a number of publications, the astronomical interpretations cannot be verified. As discussed by Šprajc and Sánchez (Reference Šprajc, Nava, Šprajc and Pehani2013), the orientation of the so-called Aposento de Moctezuma on the eastern slopes of Cerro de Chapultepec in Mexico City, related to the equinoctial sunrise by Galindo (Reference Galindo Trejo2003), cannot be established with precision, whereas Hohmann's (Reference Hohmann1995:104–106, Figure 195) equinoctial interpretation of the east–west alignment composed by Structures 9N-81 and 9N-83 in the Sepulturas group at Copán, Honduras, does not take into account the fact that, given the altitudes of the eastern and western horizon (more than 6° and 2°, respectively), the alignment corresponds to neither sunrises nor sunsets at the equinoxes, but may have been intended to record sunsets on the quarter-days of the year.

Exploring possible astronomical aspects of the Caracol tower at Chichen Itza, Yucatan, Mexico, Aveni and colleagues (Reference Aveni, Gibbs and Hartung1975) suggested that the diagonal line from inner right to outer left jambs of the west-facing window 1 corresponds to equinoctial sunsets. While a photograph supports their finding (Aveni et al. Reference Aveni, Gibbs and Hartung1975:Figure 9), they were careful enough to mention both the possibility that some of the blocks composing the window have shifted from their original positions and the difficulties involved in observing the event (which could be seen only from the floor level, due to the oddly inclined and apparently misplaced block of the inner right jamb). For these reasons they admitted that the “solar equinox alignment in window 1 remains problematical” (Aveni et al. Reference Aveni, Gibbs and Hartung1975:985). The fact that no other similar devices have been found further weakens the intentionality of this alignment, and even of other astronomical sightlines that they suggest were incorporated in the Caracol. Based on these and other arguments, Schaefer (Reference Schaefer, Bostwick and Bates2006a, Reference Schaefer, Bostwick and Bates2006b) dismissed their hypotheses, but it is fair to add that his harsh criticism, even if some issues are hardly debatable, contains assertions that are clearly objectionable (see rebuttals by Aveni Reference Aveni, Bostwick and Bates2006a, Reference Aveni, Bostwick and Bates2006b).

In other cases, of which only a few will be mentioned here, the hypotheses derive from evidently erroneous data. Córdova and Martínez (Reference Córdova Tello, Mora, Tello, Mora and Espinoza2012:28) relate the east–west axis of the archaeological site of Tamtoc, San Luis Potosí, Mexico, with the rising Sun on the equinox. However, not a single, even approximate equinoctial alignment can be identified in architectural orientations and overall layout of this site (Šprajc and Sánchez Reference Šprajc and Nava2015:79).

Coggins and Drucker (Reference Coggins, Drucker and Aveni1988) defined the group of the Temple of the Seven Dolls of Dzibilchaltún as an equinoctial observatory. However, as argued elsewhere (Sánchez and Šprajc Reference Sánchez Nava and Šprajc2015:141–143), their proposal about the importance of equinoxes, aside from depending on their interpretation of iconographic elements, is flawed by erroneous dates associated with the alignments.

Affirming that the early Middle Preclassic stages (ca. 900–600 b.c.) of Structure 19 on Mound 1 San José Mogote, Oaxaca, Mexico, are skewed 8° north of east, while the subsequent stages adopted a true east–west orientation, Flannery and Marcus (Reference Flannery and Marcus2015:7, 15, 17–19, 86–87, 119, 141; Marcus and Flannery Reference Marcus and Flannery2004:18258) associate these orientations with the equinoxes and the solstices. Field measurements at the site revealed that Structure 19 maintained the same orientation throughout its construction stages; with an azimuth of about 93.5°, it must have been intended to record sunsets on March 31 and September 12, separated by 200 (= 10 × 20) days and marked by other Middle Preclassic structures (Šprajc Reference Šprajc2018:232–233; Šprajc and Sánchez Reference Šprajc and Nava2015:44–49). Since the orientations marking calendrically significant intervals would have been useful only in association with the formal calendrical system, it is noteworthy that the existence of the 260-day count in mid-first millennium b.c. is attested in Monument 3 of San José Mogote (Flannery and Marcus Reference Flannery and Marcus2015:180–192).

For the Late Classic (ca. a.d. 600–800) Structure A-6 of Acanceh, Yucatan, Mexico, Quintal and Casares (Reference Quintal Suaste, Contreras and Salas2014) suggest that the central axis of the substructure records sunsets on the equinoxes, its azimuth and the corresponding horizon altitude being 268° and 5°, respectively. Such a horizon altitude, which they also assign to other putative astronomical alignments, is inconceivable in the flat landscape of the northern Yucatan Peninsula, invalidating their proposals.

Interpretations Based on Subjectively Selected Alignments

In a number of cases, the claimed equinoctial alignments were selected subjectively, apparently only because they fit the hypothesis. For example, Méndez and colleagues (Reference Méndez, Barnhart, Powell and Karasik2005) found an equinoctial alignment in the Temple of the Sun at Palenque, Chiapas, Mexico. In fact, in the complex architectural configuration of this building and in its relationship with others and with the surrounding landscape, they identified a number of alignments, relating them to the Sun's positions on the solstices, the equinoxes, the days of its zenith and nadir passages, and the extreme (standstill) positions of the Moon. The fact that, aside from the solstices, none of the dates most commonly recorded by axial orientations in the Maya area appears among the astronomical referents of these alignments raises the suspicion that their selection was not objective, encompassing all possible visual lines, but guided by prejudices about the importance of certain phenomena. Besides, as shown elsewhere (Sánchez and Šprajc Reference Sánchez Nava and Šprajc2015:173–174), the data provided for some of the alignments discussed do not correspond to reality.

Similarly, Lelgemann (Reference Lelgemann1996) relates several alignments in the Ciudadela compound at La Quemada, Zacatecas, Mexico, to the equinoxes and the solstices, but the architectural elements they are composed of are of different types (corners of the patio, of the central altar and of the annexed pyramid, a column, a jamb, the diagonal between two stelae found in the central altar) and seem to have been selected arbitrarily. Furthermore, the phenomenon of light and shadow observed on the equinox occurs half an hour after sunrise (Lelgemann Reference Lelgemann1996:106–107, Figure 4). There is no doubt that comparably impressive phenomena, produced by other architectural elements, could be observed on other dates and times of the day.

According to Tomasic (Reference Tomasic2009), some archaeological sites in the Holmul region, Guatemala, are located along the solstitial and equinoctial lines. Since only three alignments are discussed, and given their estimated errors of up to 2°, their possible astronomical correlates cannot be determined with any confidence. Moreover, two architectural groups claimed to mark the equinoctial and summer solstice sunrises, observing from Holmul, were found as a result of survey transects made in these directions precisely because of their preconceived astronomical significance, a procedure raising legitimate methodological objections: considering the high settlement density in the Maya Lowlands, one can only wonder how many other sites or architectural compounds would have been found had the surveyors followed other directions. Clearly, only objectively collected data on site distribution and typology in an area may serve as a reliable basis for assessing possible astronomical potential of intervening alignments.

Aveni and Hartung (Reference Aveni, Hartung and Tichy1982) mention that the alignment involving the Pyramid of the Sun at Teotihuacan and two petroglyph complexes located to the west (labeled TEO 11 and TEO 16) marks the equinoctial sunsets with considerable accuracy. As already argued by Ruggles and Saunders (Reference Ruggles and Saunders1984), the problem concerning the intentionality of this alignment is that the two groups of rock carvings are of different types and that there is a large number of petroglyphs in the area of Teotihuacan.

Alternative Interpretations Are More Likely

The proposals discussed below are based on reliable alignment data, but the cases are rather exceptional and, in the light of contextual and comparative data, the purpose of recording the equinoxes is less likely than other explanations.

As pointed out by Galindo (Reference Galindo Trejo1996) and confirmed by my measurements, the Postclassic Templo de las Caritas at Cempoala, Veracruz, Mexico, faces quite accurately the sunrises at the equinoxes. However, considering that this orientation also matches sunsets on the quarter-days and that the nearby Templo Mayor has a very similar orientation, but less compatible with the equinoctial sunrises, it is much more likely that both structures were intended to record sunsets on the quarter-days. Possibly the structure called El Pimiento marked the same dates, although on the eastern horizon (Šprajc and Sánchez Reference Šprajc and Nava2015:57).

Aveni and colleagues (Reference Aveni, Hartung and Charles Kelley1982) proposed that one of the factors that conditioned the location of Alta Vista, Zacatecas, Mexico, was the equinoctial alignment marked by Mount Picacho on the eastern horizon and materialized in a passage in the compound called El Laberinto. Furthermore, Medina and García (Reference Medina González and García Uranga2010:195) reported that the equinoctial Sun rose along the diagonal of the Southwest Plaza. Both alignments, indeed, correspond almost exactly to the equinoctial declination of the Sun. Their astronomically motivated intentionality is likely, considering that both the passage of El Laberinto and the diagonals of the two main quadrangles of the sites are aligned to prominent mountains (Šprajc et al. Reference Šprajc, Nava and Ortiz2016:19–24, Figures 4.1–4.4). Since many Mesoamerican buildings are oriented both astronomically and to conspicuous peaks on the local horizon, such correspondences cannot be attributed to chance (Šprajc Reference Šprajc2018:218). However, in light of the data from the rest of Mesoamerica, it seems unlikely that the concept of the equinox would have been known in a peripheral region. Even if such a possibility cannot be discarded a priori, an alternative explanation seems more plausible. The fact that the sunrise dates matching the passage of El Laberinto and the diagonal of the Southwest Plaza coincide with the equinoxes may well be fortuitous; rather than reflecting the builders’ awareness of the equinox, these dates must have been significant because, together with those marked by the diagonal of the Southeast Plaza on the western horizon, they could have been incorporated in a single observational calendar composed of multiples of 20 days (Šprajc et al. Reference Šprajc, Nava and Ortiz2016:19–24, Table 4.1). More data on architectural orientations in northwestern Mesoamerica would be needed to support this hypothesis. In the light of the currently available evidence from other regions, however, it seems more likely than the equinoctial one: in contrast to the absence of equinoctial orientations, the orientation patterns documented throughout Mesoamerica clearly reflect the use of observational calendars composed of calendrically significant intervals.

Another interesting example is the line connecting the centers of the entrances to the upper sanctuaries of Temples I and III of Tikal, Guatemala. The azimuth of this line, determined by both Aveni and Hartung (Reference Aveni, Hartung and Aveni1988:9, 12) and my own measurements, is 269°54'. Although this line reproduces almost exactly the true east–west direction, it can hardly be related to the equinoxes, as suggested by Aveni and Hartung (Reference Aveni, Hartung and Aveni1988) and Malmström (Reference Malmström1997:169–171), because the equinoctial sunrises or sunsets along this alignment could have been observed only on the natural horizon (with an altitude of about 0°). In fact, due to their heights, both temples mutually block the view to the horizon. As already pointed out by Hartung (Reference Hartung and Aveni1980:148), there was evidently no opening in the back wall of either building, hence the equinoctial observations were not possible once the construction was completed. While the existence of non-functional, merely symbolic alignments cannot be discarded, such a hypothesis in this case can hardly be substantiated, both because of the lack of equinoctial alignments in Mesoamerica and because the case at hand has a more plausible rationale. Observing in front of the upper sanctuary of Temple I westward, the angular altitude of the roof comb of Temple III is 3°36'; the declination corresponding to this altitude and the azimuth of 269°54' is 0°55'. Inversely, observing from the entrance to the sanctuary of Temple III, the roof comb of Temple I, rising to 2°36' of altitude, marks the declination of 0°47' (originally the roof combs could not have been much higher). Remarkably, the two declinations are very similar, corresponding to the positions of the Sun on the quarter-days of the year (Figure 7). It seems very likely that the relative placement of both temples and their heights were conditioned by the purpose of recording these dates, not only because they are marked by a number of orientations, but also because, observing from Temple IV, the Sun rises aligned with Temple III at the December solstice (recall that quarter-days are mid-points in time between the solstices) and with Temples I and II on two other significant pairs of dates (February 12 and October 30, and February 9 and November 1; Sánchez and Šprajc Reference Sánchez Nava and Šprajc2015:192–197).

Figure 7. Tikal, Peten, Guatemala, the rising Sun over the roof comb of Temple I, observed from the upper sanctuary of Temple III on a quarter-day (September 21, 2011). Photograph by Dieter Richter.

Light-and-Shadow Phenomenon at Chichen Itza

The most famous Mesoamerican building believed to reflect the importance of the equinoxes is undoubtedly El Castillo of Chichen Itza, also known as the Pyramid of Kukulcán. Being a rather singular case, it does not neatly fit any of the above categories of methodological inconsistencies and will be discussed separately.

Every year on the vernal equinox, thousands of visitors gather to watch the light-and-shadow effect produced in the late afternoon on the northern balustrade of the pyramid, giving the impression of a descending rattlesnake with illuminated dorsal triangles. The ophidian heads at the base of the northern stairway make this visual effect even more persuasive. The popularity of the phenomenon, first described by Rivard (Reference Rivard1969) and Arochi (Reference Arochi1976), is also mirrored in numerous publications interpreting it as the result of a carefully planned architectural design, whose objective was to record the equinoxes (see a summary in Carlson Reference Carlson1999).

However, some researchers have noted that the interplay of light and shadow can be observed also a few days before and after the equinox (Aveni et al. Reference Aveni, Milbrath and Lope2004). In fact, a series of photographs taken in 2017 and 2018, from March to May every day, from 3 p.m. to sunset, at five-minute intervals, revealed that the phenomenon can be observed over a much longer period, at least from early March to mid-April. As the Sun goes down, the stepped bodies of the pyramid are projected onto the west face of the northern stairway. The increasingly jagged shadow rises and its tips begin to touch the edge of the balustrade, starting at the top and ending at the bottom, until several illuminated triangles are formed. The photos in Figure 8 were all taken at the moment when all triangles have formed, their apices touching the upper edge of the balustrade. In mid-March this occurs about 1.5 hours before sunset, when six triangles can be observed, while on subsequent days this moment approaches the time of sunset and the number of triangles increases. It has been suggested that the seven illuminated triangles and six intermediate shadows, appearing at the equinoxes, support the intentionality of the solar “hierophany,” considering that 13, the sum of these elements, had an obvious calendrical and symbolic significance. However, the seventh triangle begins to appear at the lower extreme of the stairway a few days before the March equinox, and since the whole set of seven triangles also can be seen during several following days, it could not have served as an accurate marker of the equinox.

Figure 8. Chichen Itza, Yucatan, Mexico, light-and-shadow effect on the northern stairway of El Castillo on several dates in March 2017 and April 2018. Note that the hours in April correspond to the summer (daylight saving) time. Photographs by Miguel Ángel Cab Uicab and Mariana Hernández Blanca.

Furthermore, there is no indication of a special importance of the seven triangles. In early April the number of triangles increases to eight, and in mid-April even the ninth triangle begins to form. It can barely be seen nowadays because the trees growing west of the Great Platform supporting El Castillo cast their shadows onto the lower part of the pyramid a few minutes before sunset, but likely would have been visible during the apogee of the city. One could suggest that the appearance of nine triangles, evoking the number of layers of the Maya underworld, was even more significant. However, since the daily variations are hardly noticeable, no date could have been determined by observing this phenomenon only.

In the attempt to interpret the luminous effect in a different way, Casares (Reference Casares Contreras2021) contends that, according to his observations in 2016, 2017, and 2018, the array of nine triangles appeared for the first time on April 9. Strangely, however, in his illustration (Casares Reference Casares Contreras2021:Figure 8), supposed to support the statement, only eight illuminated triangles can be seen, and my own Figure 8 shows that the ninth triangle was not completely visible even on April 12, 2018. For the other half of the year, Casares affirms that the nine triangles are visible for the last time on September 2 (when the Sun has the same declination as on April 9), but this date is thus also incorrect. It would be a digression to comment on other questionable parts of his argument, but his selection of the two dates seems to have been based on their preconceived significance derived from the conjecture that, falling 73 days before and after the June solstice, they had a connection with the 584-day synodic period of Venus (584 = 8 × 73). Although the interval of 584 days leads from September 2 to April 9, Venus would not have been visible on both dates in the same place of the sky, as Casares (Reference Casares Contreras2021:57) implies: after 584 days the planet returns to approximately the same apparent position relative to the Sun, but not relative to the starry background or the horizon, because its coordinates (right ascension and declination) do not repeat at 584-day intervals.

Returning to the equinoctial “hierophany,” it was suggested that its intentionality is supported by a comparable light-and-shadow effect produced around the December solstice on the northern stairway of a similar pyramid at Mayapan, also known as El Castillo (Arochi Reference Arochi, Broda, Iwaniszewski and Maupomé1991; Aveni et al. Reference Aveni, Milbrath and Lope2004). But also in that case the phenomenon is observable during about a month before and after the solstice, likewise with a varying number of triangles, and thus could not have served for determining any date either. Another allegedly significant detail of El Castillo at Chichen Itza was adduced by Rivard (Reference Rivard1969:52): of its four stairways, only the northern one, on which the illumination is produced, is decorated with snake heads at its base. However, this fact may well have a simpler explanation: the northern stairway ascends to the main entrance of the upper sanctuary, which faces north, in the direction of the sacred cenote to which a broad causeway leads. Besides, according to the sixteenth-century bishop Diego de Landa, the snake heads adorned the bases of the four stairways, although their presence has not been confirmed archaeologically (Tozzer Reference Tozzer1941:178).

To summarize, the light-and-shadow effect at Chichen Itza can be observed during several weeks; the day-to-day variations are barely perceptible, making it unsuitable for determining specific dates; the number of triangles varies, depending on the date and hour of observation; and there is no clue as to what the purpose of the builders might have been. Given these facts and the absence of comparable cases, the most prudent conclusion is that the phenomenon is accidental. Even assuming that it had a symbolic rather than observational function (Aveni Reference Aveni2001:295, 298–299), there are no grounds to suppose it commemorated the equinoxes; after all, the triangles are better defined and the whole display is visually more attractive only after/before the spring/fall equinox, in early April and early September (Figure 8).