When H. P. Lovecraft died in 1937, his “Instructions in Case of Decease” named the young Robert Hayward Barlow as his literary executor. Yet Barlow did not spend the rest of his life involved solely in science fiction fandom and publishing Lovecraft; while studying at the University of California he became interested in Mexican anthropology and languages, and after graduation in 1942 moved to Mexico permanently where he became a noted expert on Mesoamerican languages and anthropology, and a printer in Nahuatl and Mayan languages. He finally became head of the Department of Anthropology at Mexico City College, where he drew the attention of a young William S. Burroughs.

Despite success in his field, including a 1948 trip to Europe to study Mesoamerican codices, Barlow died on New Year’s Day 1951, the result of an overdose of Seconal tablets. The suggestion has been made that his death was the result of blackmail or threats to expose him as a homosexual. Despite being cut off in the prime of his life, Barlow’s brief but brilliant career was a substantial influence on Mesoamerican anthropology; some of his papers and the concepts he originated are still cited to this day—such as the Codex Huitzilopochtli.

Barlow’s legacies as Lovecraft’s executor, poet, printer, and anthropologist are often very distinct; readers familiar with him for his connections to Lovecraft are often ignorant of his achievements as an anthropologist, and vice versa. The matter of the Codex Huitzilopochtli is a matter purely of anthropological bibliography; chasing a telephone game of citations back to its source to see what Barlow actually wrote about something, and how that got interpreted, misinterpreted, and finally re-interpreted over the course of decades. Misunderstandings about it continue to crop up occasionally today, due to the relative availability of some older sources over newer ones in academic publishing.

To give an example of this kind of issue, consider these two passages:

The story of Indian America must be written with soft chalk, easily erased and corrected. The conclusions which yesterday seemed tenable may tomorrow be overruled by the discovery of a handful of bones in a cave or hitherto unknown utensils in volcanic ash.
—Hubert Herring, A History of Latin America, 2nd Ed. (1956) 25

“The story of Indian America,” Pablo Martínez del Río used to tell his classes, “must be written with soft chalk, easily erased and corrected.”
—Michael C. Meyer & William L. Sherman, The Course of Mexican History (1991) 4

Depending on who is speaking or writing, an anthropologist or historian might cite either Hubert Herring (1889-1967) or Pablo Martínez del Río (1892-1963) as the originator of the “soft chalk” phrase. Herring seems to have gotten it into print (in English) first, but whether he paraphrased something del Río said or vice versa—well, we don’t know, exactly. The sentiment is generally agreed upon, but the lineage of transmission is confused. This is the kind of small problem that the Codex Huitzilopochtli and R. H. Barlow presents: not a major issue of anthropology and interpretation, but how the understanding of an idea has been transmitted over time, and how that has affected how that idea has been received and understood.

In brief, this is a story about citations, the lack thereof, and what Barlow originally wrote versus how it developed in the hands of others.

My hope in tracing the history of the Codex Huitzilopochtli as a concept is to both to give readers an idea of the influence that Barlow had on Mesoamerican anthropology, and to highlight the history of a specific idea from Barlow that continues to have some currency today. For readers more familiar with Barlow as an anthropologist, this might be a more straightforward exercise in tracing citations and how ideas are transmitted over time.

General Background on Mesoamerican Codices

At the time contact between Europe and Mesoamerica was made in the late 15th century CE, several indigenous cultures had written languages and scripts, which were used to record matters including history, cosmology, and religion on both durable materials (stone monuments and buildings, ceramics, etc.) and perishable materials (bark paper, animal skin, cloth, etc.). Codices took several forms, including a long strip (tira), which might be rolled (like a scroll) or folded in and out, concertina-, accordion-, or screenfold-style. The content of these books might include both pictorial and pictographic elements.

In the 16th century (1519-1521), the Spanish Empire came into conflict with and ultimately defeated the Aztec Triple Alliance in an armed conflict. The Spanish set up new administration of their possessions, expanded against other polities in the Americas, suppressed rebellion against their rule, and attempted to suppress local religion and convert the indigenous peoples to Catholic Christianity. Toward this end, the Spanish Crown banned the production of local bark paper (forcing the use of European-made paper or vellum) and religious authorities collected and destroyed many local manuscripts. Knowledge of many local scripts and records of pre-contact indigenous culture was ultimately lost. The Mayan script, for example, fell out of use, and would not be deciphered until the 20th century.

The term codex has come to refer to both pre-contact indigenous Mesoamerican manuscripts and post-contact manuscripts that were created in the early Colonial period in whole or in part by indigenous artists and/or that contain material copied from pre-contact manuscripts. The use of the term arose in part because Europeans who collected or preserved such materials sometimes prepared and bound them as European-style books (e.g. a set of sheets called a quire is folded and bound along one edge; multiple quires are bound together to make a book. When the outer folds are cut, the individual leaves move freely).

The majority of surviving codices probably date from after European contact. Some were written and illustrated by indigenous artists on European paper, others may have been prepared by European artists copying from indigenous originals as various Europeans commissioned, copied, recorded, or otherwise preserved some of these works for their own purposes. Post-contact codices may show the influence of European contact beyond just the material used: records of post-contact events, depictions of Europeans, formatting similar to European books, etc. These scanty writings provide valuable insight into Mesoamerican history and life during this period.

The bibliographic entry for each surviving codex is unique. Since these works have survived for centuries on relatively perishable materials, they often show wear and tear, may be missing pages, repairs, include annotations or glosses in various hands, etc. Scholars who study these codices for information on Mesoamerican history do not just read the words and interpret the pictures, but analyze the style, formatting, construction, and context of the codices, noting similarities and differences with other evidence. In some cases, codices contain sufficiently similar content or style to suggest a distinct tradition or line of transmission. Such is the case with the Codex Telleriano-Remensis and the Codex Ríos.

Codex Telleriano-Remensis

The codex was formerly known as the Codex Tellerianus and the Codex Le Tellier. It was once part of the Charles-Maurice Le Tellier collection. It is made of European paper with 50 leaves, and it measures 32 by 22 centimeters (approximately 12 /2 by 8 6/8 inches). It is currently housed at the Bibliothéque Nationale de Paris.
—Manuel Aguilar-Moreno, Handbook to Life in the Aztec World (2007), 273

This Aztec codex was produced in 16th century Mexico and shows the evidence of many hands, indicating that several indigenous artists worked on various sections, and those sections were later annotated with glosses in Spanish by various writers:

Codex Telleriano-Remensis, glossed as “Heroglificos de que usavan lo…” on the cover, has three major pictorial sections in several native styles. Each is annotated in Spanish, in several handwritings. One of the latter is believed to be of Fray Pedro de los Ríos. The first section is an 18-month calendar with drawings of the gods of each period and a symbol for the nemontemi (the extra five days in the solar calendar). The second is a tonalpohualli (260-day divinatory almanac). The third is a pictorial annal for the period 1198-1562, in two major styles. Two final pages contain historical notices in Spanish without drawings, for the years between 1519 and 1557. There are leaves missing from each pictorial section […]
—Manuel Aguilar-Moreno, Handbook to Life in the Aztec World (2007), 272-273

The full codex has been digitized and can be viewed at the Bibliothéque Nationale de Paris website. A facsimile edition (chromolithographic reproduction) with commentary in French was published by Joseph Florimond, duc de Loubat in 1899, which can be viewed at the Foundation for the Advancement of Mesoamerican Studies website; and Eloise Quiñones Keber edited and annotated a photographic reproduction edition: Codex Telleriano-Remensis: Ritual, Divination, and History in a Pictorial Aztec Manuscript (1995), with English commentary and translations.

In 1948-1949, R. H. Barlow traveled to Europe to view Mesoamerican codices in archives, and as part of his ongoing work to translate and research the notes of Francisco del Paso y Troncoso (1842-1916), who had written extensively on the history of the Codex Telleriano-Remensis; it is not clear if Barlow accessed the original codex at the Bibliothéque Nationale during his time in Paris. Barlow’s Anales de Tlatelolco: históricos de la nación mexicana y Códice de Tlatelolco (1948) cites the Loubat facsimile of the Codex Telleriano-Remensis.

Codex Ríos

The Codex Ríos was formerly known as Codex Vaticanus A, Codex Vaticanus 3738, and Copia Vaticana. It is a European-paper codex with 101 leaves that each measure 46 by 29 centimeters (approximately 18 by 11 3/8 inches). It is currently housed at the Vatican’s Apostalic Library.
—Manuel Aguilar-Moreno, Handbook to Life in the Aztec World (2007), 272

The Vatican Apostolic Library had two Aztec codices, which scholars labeled Codex Vaticanus A (3738) and Codex Vaticanus B (3773); Codex Vaticanus A is the only Aztec codex annotated in Italian, which in itself unusual. The name Codex Ríos comes from the friar Pedro de los Ríos, who is one of the Spanish annotators of the Codex Telleriano-Remensis and in the Italian text is indicated as the compiler of the codex that now bears his name. Ríos’ involvement adds to the mystery surrounding the book’s origins:

The Italian gloss on Codex Ríos associates Pedro de los Ríos with the compilation of its paintings, identifies him as a Dominican friar, links him with an episode in Mexico in 1566, indicates that he supplied the data for the glossed commentary, and provides incidental biographical data, such as his firsthand knowledge of Oaxaca. Aside from this gloss and similar statements in the glosses on Codex Telleriano-Remensis, only one further datum has been reported about this elusive monk. An apparent report of his death by 1565 (cited in Jiménez Moreno and Mateos Higuera, 1940) conflicts with the 1566 date in Codex Ríos and the presumption that he was present when the codex was painted and annotated in Rome ca. 1570.
—John B. Glass, “A Survey of Native Middle American Pictorial Manuscripts” (1966) in Handbook of Middle American Indians, Volumes 14 and 15 (1975), 136

The identity of Pedro de los Ríos and his history with the Codex Ríos has been addressed by Maarten E. R. G.N. Jansen in “El Codice Ríos y Fray Pedro de los Ríos” (1984) and Eloise Quiñones Keber in Codex Telleriano-Remensis: Ritual, Divination, and History in a Pictorial Aztec Manuscript (1995).

In terms of content, the Codex Ríos differs from the Codex Telleriano-Remensis in that it seems to have had a single artist with a consistent style, and the Italian-language glosses seem to be from a single writer (and, indeed, appear to be a translation and gloss of the Spanish commentary in the Codex Telleriano-Remensis). Much of the material is similar in both volumes, with the Codex Ríos containing pages missing in the Codex Telleriano-Remensis. It is organized differently than the Codex Telleriano-Remensis, however:

The manuscript has seven major sections: 1) cosmogenic and mytholoigcal traditions with some emphasis on the four previous epochs, or suns, including notices about Quetzalcoatl and the Toltec; 2) a 260-day divinitary almanac; 3) calendrical tables without drawings for the years between 1558 and 1619; 4) an 18-month festival calendar with drawings of the gods of each period; 5) sacrificial and other customs, including portratis of Indian types; 6) pictorial annals for the years between 1195 and 1549, beginning with the migration from Chicomoztoc and covering later events in the Valley of Mexico; and 7) glyphs for the years between 1566 and 1562 without written or pictorial entries. Most of the codex has a long written commentary in Italian, but only three pages of the historical section are annotated.
—Manuel Aguilar-Moreno, Handbook to Life in the Aztec World (2007), 272

This comparison is slightly complicated by the fact that the original manuscript was apparently poorly bound with some pages out of order.

The Codex Ríos was the first Mexican manuscript codex to be reproduced in print, in part, through a series of woodcuts. In 1615 and 1626, Lorenzo Pignoria added his Seconda Parte delle Imagini de gli Dei Indiani (“Second Part of the Images of Indian Gods”) to Vincenzo Cartari’s Le Imagini con la sposizione dei dei de gli antichi (1556, “The Images of the Gods of the Ancients and their Explanations”), an expansive work of comparative mythology in which Pignoria reproduced Aztec gods from the Codex Ríos to feature alongside other world mythologies.

A facsimile edition (chromolithographic reproduction) of the Codex Ríos with preface in Italian by Franz Ehrle was published by Joseph Florimond, duc de Loubat in 1900, correcting the order of pages; a second facsimile by Coroña Núñez with Spanish translations of the Italian text in 1964 followed the Loubat’s foliation, and a third facsimile without transcription or commentary and with photographic reproduction of the pages as they exist in the original was published by Akademische Druck- u. Verlagsanstalt (ADEVA) in 1979 as part LXV of their Codices Selecti series; the latter can be viewed at the Foundation for the Advancement of Mesoamerican Studies website. The most recent facsimile (photographic reproduction, with corrected page ordering) is Religión, costumbres e Historia de los antiguos Mexicanos: libro explicative del llamado Códice Vaticano A (1996, ADEVA) edited by Ferdinand Anders and Maarten Jansen.

It does not appear that Barlow made it to Rome during his European survey of codices in 1948-1949, so any research he did on the Codex Ríos would have involved photographs or reproductions, the Loubat facsimile, and second- or third-hand descriptions of the contents. Barlow’s Anales de Tlatelolco:históricos de la nación mexicana y Códice de Tlatelolco (1948) cites the Loubat facsimile of the Codex Ríos.

Codex Huitzilopochtli

While the Codex Ríos and Codex Telleriano-Remensis are distinct works, there are close similarities in the content of several sections. Correspondences between the two codices had been drawn up by Francisco del Paso y Troncoso in 1898; correspondences were incorporated in the commentary matter of the 1900 Loubat facsimile of the Codex Ríos by Ehrle; and John B. Glass in his “A Survey of Native Middle American Pictorial Manuscripts” (1966) provides a helpful comparative table between the two codices in English:

To give an idea of how close the relevant sections of these two codices are, compare these two pages from the beginning of the pictorial annals in each codex:

Note: The footprint-trail on this page is left out of the Loubat facsimile of the Codex Ríos, so this is a lower-resolution photo from the 1979 ADEVA facsimile.

As codices go, this is an unusual level of similarity, not just in the details of the depiction of the figures, but in the arrangement of the figures on the page and the glosses applied to them. The figure in the upper left on each page is Huītzilōpōchtli, patron god of the Mexica, the dominant ethnic group in the Aztec Triple Alliance. The codices correspond so closely in parts that the scholarly consensus in the first half of the 20th century is that the Codex Ríos was a copy of the Codex Telleriano-Remensis (see “Algunas observaciones acerca del Códice Vaticano 3738 o Códice Ríos” (1925) by Dr. B. Reina), or at least very close to the original (i.e. a copy only one or two generations removed from the original).

Però il contenuto dei due codici in tutto ciò, che loro è comune, è essenzialmente identico, in maniera che il Telleriano, benchè non sia probabilmente l’ originale ommediato del Rios, è certamente ad esso molto vicino.	But the content of the two codices in all that is common to them is essentially identical, so that the Telleriano, although probably not the immediate original of the Rios, is certainly very close to it.
“Introduzione.—Storia del Codice Rios”, Loubat facsimile 13	English translation.

However, in 1941, J. Eric S. Thompson, the leading English archaeologist and ethnologist in Mesoamerican studies, published a paper titled “The Prototype of the Mexican Codices Telleriano-Remensis and Vaticanus A” that proposed a different theory: that the two codices both derived from an unknown original indigenous manuscript, and the similarities came from different artists copying from the same work. Thompson’s justification for this belief involved a close analysis of the art in the comparable portions, with an emphasis on irregularities in the Codex Telleriano-Remensis and its deviation from depictions of certain gods featured in other codices compared to the Codex Ríos. To give one example:

The figure of Xolotl in Codex Telleriano-Remensis (19v) is scarcely recognizable as that of a dog. The equivalent deity in Codex Vaticanus A is definitely canine.
—Thompson, “The Prototype of the Mexican Codices Telleriano-Remensis and Vaticanus A” in The Carnegie Maya III: Notes on Middle American Archaeology and Ethnology (2011), 17

The two images above are from the sections of the codices called tonalamatl (“pages of days”), a kind of divinatory almanac structured around the 260-day sacred year (tonalpohualli). Tonalamatl also appear in the Codex Borbonicus, Codex Borgia, and other codices.

Other cases might be cited in which Codex Vaticanus A is more correct in presenting details than its supposed prototype, Codex Telleriano-Remensis. It is hardly credible that the very poor artist who painted the figures of the former codex, and who himself omitted important attributes (e.g. black marking around Ixcuina’s mouth, 29v), had the skill or the knowledge to correct in his version errors in the work he was copying. The conclusion is inescapable that Codex Vaticanus A was copied or recopied.
—Thompson, “The Prototype of the Mexican Codices Telleriano-Remensis and Vaticanus A” in The Carnegie Maya III: Notes on Middle American Archaeology and Ethnology (2011), 17

In essence, J.E.S. Thompson doesn’t find it credible that the Codex Ríos artist could have corrected some of the errors of the artists in the Codex Telleriano-Remensis and still made other errors of their own. Thompson also ruled out the possibility that the Codex Telleriano-Remensis was copied from the Codex Ríos: the Codex Telleriano-Remensis is dated 1563 and the watermarked paper is from about that time; the Codex Ríos is dated 1566—and was possibly created c.1570. (Full discussion of the dating of these two codices and their history is outside the scope of this article, but the general consensus is that the Codex Telleriano-Remensis came first.)

R. H. Barlow read Thompson’s article no later than 1948, because in that year in a footnote on the Codex Tlatelolco he wrote:

En los códices Telleriano y Vaticano Ríos, que Thompson ha señalado como copias de un original perdido, tenemos una representación vivida de la defensa del peñol de Nochistlán, de los ataques por las fueras del Virrey, y de la muerte de Alvarado.	In the Telleriano and Vaticano Ríos codices, which Thompson has pointed out as copies of a lost original, we have a vivid representation of the defense of the Peñol de Nochistlán, the attacks by the Viceroy’s forces, and the death of Alvarado.
“El Códice de Tlatelolco, Interpretación por R. H. Barlow” in Anales de Tlatelolco:históricos de la nación mexicana y Códice de Tlatelolco (1948; 1980 reprint) 108n2	English translation

By 1949, R. H. Barlow would first reference this “lost original” in his critical work Extent of the Empire of Culhua Mexica as:

Códice Huitzilopochtli. Migration map divided up and copied as pages of the twin codices Telleiiano and Vaticano Rios (141)

This is the first reference in print to what would become known as the Codex Huitzilopochtli.

In 1966 when John B. Glass (with Donald Robertson) did his monumental “A Survey of Native Middle American Pictorial Manuscripts,” a comprehensive survey of all known Mesoamerican codices and fragments, he placed the Codex Telleriano-Remensis and Codex Ríos together as related codices in what he called the “Huitzilopochtli Group.” His justification for that grouping and that name was Barlow’s use of the term.

J. E. S. Thompson (1941b) has advanced iconographic reasons to show that Ríos cannot be a copy of Telleriano-Remensis and that the two therefore derive from a common original. Barlow (in unpublished lecture notes) has named this hypothetical lost manuscript Codex Huitzilopochtli after the god who appears at the beginning of the migration history in both manuscripts “as a traveller guiding his people.”
—John B. Glass and Donald Robertson, “A Survey of Native Middle American Pictorial Manuscripts” (1966) in Handbook of Middle American Indians, Volumes 14 and 15 (1975), 136-137

The fact that Barlow was apparently keen on the idea of a hypothetical precursor codex isn’t unusual. In 1945 Barlow published “La Crónica ‘X’: Versiones Coloniales de la Historia de los Mexica Tenocha,” a paper which postulated a hypothetical common textual or pictorial source (Crónica X/Chronicle X) for the Mexica histories in Durán’s Historia de las Indias…, Alvarado Tezozomoc’s Crónica Mexicana, and Tovar’s Relación del Origen… The idea was convincing enough that Glass even saw fit to group these works together as the Crónica X Group, just as he had formed the Huitzilopochtli Group.

While Lovecraft fans might compare how the Al Azif was the precursor to the Necronomicon, a more likely inspiration would probably be New Testament textual studies, whose scholars had hypothesized a Q document as a source for the Synoptic Gospels. As Mesoamerican scholars strove to document and analyze the mass of codex material, they would utilize all the textual analysis tools at their disposal, and they were already aware that the surviving codices represented a fraction of what had been a much larger literary corpus.

Glass is careful to specify that it is only the migration history portion—beginning with the appearance of Huītzilōpōchtli—which he claims Barlow attributed to the hypothetical Codex Huitzilopochtli in his unpublished lecture notes, not the calendar or other material. This distinction was noted and continued by others, for example:

 Another Central Mexican pictorial which deserves special comment is the historical portion of Codex Telleriano-Remensis/Vaticanus A (apparently cognate derivatives from a common original—see Thompson 1940-43; Robertson 1959: 107-15). An atypical version of the Mexica migration and founding of Mexico-Tenochtitlan is followed by a more standard historical narrative (pictorial scenes until 1548-49, colored year count to 1554-55, final pen additions to 1562) featuring Tenochca history but also devoting considerable attention to other neighboring communities. Nothing is known concerning the original sources of this history, but particularly for the migratory portion (Barlow’s [1950] “Codex Huitzilopochtli”), an ultimately pre-Hispanic source(s) seems likely for the section covering events anterior to the Conquest. As in the case of the Matrícula-Mendoza, although a mid-sixteenth-century date for the completion of the prototype of Telleriano-Remensis/Vaticanus A is undeniable, I would regard significant Hispanic influence in the formation of its place and name signs as quite unlikely.
—H. Be. Nicholson, “Phoneticism in the Central Mexican Writing System” in Mesoamerican Writing Systems: A Conference at Dumbarton Oaks, October 30 and 31st, 1971 (1973) 22-23

Nicholson cites the unpublished Barlow notes as his source:

Barlow, R. H. n.d. Anthropology 307-A: Codices and Mesoamerican Picture Writing. Unpublished lecture notes taken by Anne Garges and corrected by lecturer. Mexico, 1950. [ibid 37]

Donald Robertson cites the same source in Mexican Manuscript Painting of the Early Colonial Period (1959):

Robert H. Barlow, “Anthropology 307-A: Codices and Mesoamerican Picture Writing,” unpublished lecture notes taken by Anne Garges and corrected by Barlow, Mexico, winter 1950. (163n18)

Anne Garges de Forrest is listed among the Bachelor of Arts candidates on page 3 of the Mexico City Collegian for 1 June 1950, which puts her in exactly the right place and time to take Barlow’s class and those notes from his lecture. Together, Glass, Roberston, Nicholson, and others incorporated the idea of the Codex Huitzilopochtli into their view of the Mesoamerican codices’ historiography, much as they did with Crónica X. However, unlike Crónica X, there was no readily available paper or textbook to refer to…so the idea began to change as others adopted it. For example:

In 1941 Thompson related TR and Vaticanus A (Codex Rios) to a lost prototype that Robert Barlow called Codex Huitzilopochtli.
—Howard F. Cline, “The Chronology of the Conquest: Synchrologies in Codex Telleriano-Remensis and Sahugun” (1973) in Journal de la Société des Américanistes

Cline was doing a close analysis of correspondences between Codex Telleriano-Remensis, Codex Ríos, and similar Mexica calendar and historical material in other codices; he built tables to showcase what parts of these codices agreed and where they differed. Yet, he makes a slip: where Glass and Nicholson had been careful to say that Barlow’s hypothetical Codex Huitzilopochtli had been a precursor only to the historical portion, Cline’s gloss suggests the Codex Huitzilopochtli was the prototype for both codices en toto. Whether or not Cline understood and intended this meaning isn’t clear, but it is common misunderstanding, one possibly made easier by the name “Huitzilopochtli Group,” which perhaps implies more than it should.

Glass and Robertson’s use of the Codex Huitzilopochtli in the Handbook of Middle American Indians (1975) lent additional scholarly weight to the idea of its existence. In 1979 when ADEVA published a facsimile of the Codex Ríos, the introductory note cites their work and adds:

The manuscript itself was composed between 1570 and 1589 on the Plateau of Mexico or already in Italy, in the latter case similar to the Codex Telleriano Remensis which is preserved in Paris, as a copy of a joint model which Robert H. Barlow denominated “Codex Huitzilopochtli.”

Thanks in part to such scholarly consensus and repetition, the idea of a common source for both the Codex Telleriano-Remensis and Codex Ríos continued to enjoy popularity and currency in Mesoamerican codicological studies for decades. When Emily Good Umberger wrote her doctoral dissertation in 1981, she accepted the basic idea of the Codex Huitzilopochtli as a hypothetical source and worked it into her personal view of Mexica codicology:

The migration pictorials of note are the Codex Boturini (Fig. 140) and the historical sections of Codex Telleriano-Remensis and Codex Vaticanus A/Rios, which are cognates deriving from a hypothetical lost original called the Codex Huitzilopochtli. They start with the departure from Aztlan and show the Aztecs wandering from site to site until they reach the lake area. Boturini stops before the foundation of the city of Tenochtitlan, but Telleriano-Remensis and Vaticanus A/Rios continue past the Conquest. The original format of these histories (which Boturini still has) is a single long sheet of fiber paper on which the history proceeds from left to right, pretty much in a horizontal direction. The main organizing device is a stream of hieroglyphic dates representing the years, and every year is counted. Another important pictorial history, the Codex Mendoza (Fig. 141) begins at a later point in history, the foundation of Tenochtitlan.
—Emily Good Umberger, Aztec Sculptures, Hieroglyphs, and History (1981) 33

Umberger is doing what good anthropologists and historians do: working with the information at hand and making reasonable extrapolations based on relevant examples. The Codex Huitzilopochtli is a hypothetical source for the historical migration material in these codices; what else looks like that? Answer: the Codex Boturini, whose historical material has been noted as having strong similarities to other codices, notably the Codex Aubin and the Codex Mexicanus. So now we’re getting the image of the Codex Huitzilopochtli as a 16th-century indigenous manuscript in the style of a single long sheet (tira) of locally-made paper, which is a very logical and reasonable extrapolation.

Which presents a problem. The Codex Huitzilopochtli was still essentially hypothetical, based on a series of surmises and some scanty material—the name itself came from an unpublished set of Barlow’s notes, if we can trust the citations. What if they were all wrong?

Corrections & Re-Evaluations

The idea of a single prototype codex for both the Codex Telleriano-Remensis and Codex Ríos never gained universal acceptance among Mesoamerican scholars. Even before Thompson, there was room for speculation:

For many years after Humboldt first noted their resemblance, scholars generally assumed that the Codex Vaticanus A was a copy of the Codex Telleriano-Remensis. Such a judgment seemed inevitable since the poorly drawn illustrations of the Vatican codex could not have served as the model for those of the more skillfully executed Paris manuscript and the glosses of the latter were inscribed by several hands, whereas those of the former were collated into a single text. This ordering was espoused by J. F. Ramirez ([1855] :217, cited by Paso y Troncoso 1898:337), the first serious scholar of the Codex Telleriano-Remensis after Humboldt. But discrepancies between the two manuscripts also led to alternative proposals. In his bibliographic study of the two manuscripts Francisco del Paso y Troncoso, who otherwise accepted most of Ramirez’s opinions concerning the Codex Telleriano-Remensis, proposed that both manuscripts had been copied from a lost original (Paso y Troncoso 1898:349-351). At about the same time, in his introduction to the Loubat edition of the Codex Vaticanus A, Franz Ehrle (1900:13-16, 21-22) suggested an intermediate copy between the two known manuscripts, as did B. Reina (1925:218-219) in a later detailed study of the texts of the two documents
—Eloise Quiñones Keber, “The Codex Telleriano-Remensis and Codex Vaticanus A: Thompson’s Prototype Reconsidered” (1987) in Mexicon (Jan 1987), 9

Then came Thompson’s “The Prototype of the Mexican Codices Telleriano-Remensis and Vaticanus A” (1941). Donald Robertson cited Thompson’s essay in his critical 1959 survey of Mesoamerican manuscript painting, but felt the need to add:

Thompson discussed the prototype of Telleriano-Remensis. It would be more accurate to discuss the prototypes, for there were at least two sources for the present manuscript. Kubler and Gibson have pointed out the Texcocan affiliations of the first pages in the symbols of the months of the “vague” native year. These are followed by a Tonalamatl (Plate 10). The first part of the history follows the traditions of Texcoco and the eastern shores of the lake in relating how the Valley [of Mexico] was populated (Plate 28). Mexican influences in this essentially Texcocan account will be detailed below. The second part of the historical section derives from a separate tradition and probably a distinct manuscript and is reminiscent of the historical section of the Codex Mendoza (Plate 29). The manuscript is thus a synthetic work, rather than a simple document deriving from a single source, as Thompson implied, unless the immediate source was an earlier already synthetic manuscript.
—Donald Robertson, Mexican Manuscript Painting of the Early Colonial Period (1959), 109

Robertson doesn’t mention the Codex Huitzilopochtli specifically, but the “second part of the historical section” “probably a distinct manuscript” would be equivalent to Barlow’s hypothetical source-document for the migration section. Roberston also believed that the migration section originated from a tira that showed the entire migration, which the artist then broke up into individual pages—the same conclusion Umberger would come to decades later, good anthropologists thinking alike and all that (ibid. 109, 115).

So even in the 1950s, the idea of a more complex relationship between the Codex Telleriano-Remensis, Codex Ríos, and hypothetical source document(s) was still there, and even Glass and Robertson acknowledged it:

The theory of a single common prototype may oversimplify a very complex historicgraphic problem that has been inadequately studied. It fails, for instance, to explain the remarkable coincidence of a change in artists and style in Codex Telleriano-Remensis at precisely the point where its page composition changes from one format to another.
—John B. Glass and Donald Robertson, “A Survey of Native Middle American Pictorial Manuscripts” (1966) in Handbook of Middle American Indians, Volumes 14 and 15 (1975), 136

Forty-six years after Thompson suggested Codex Telleriano-Remensis and Codex Ríos shared a prototype and two decades after John B. Glass cited R. H. Barlow’s unpublished lecture notes and introduced the Codex Huitzilopochtli and the Huitzilopochtli Group to a wider audience of Mesoamerican scholars, the original concept was ripe for reexamination. Which is what happened when Eloise Quiñones Keber published “The Codex Telleriano-Remensis and Codex Vaticanus A: Thompson’s Prototype Reconsidered” (1987).

Thompson based his arguments primarily on discrepancies between pictorial details in the tonalamatl sections of the two manuscripts. Since he held that the Codex Vaticanus A was more “correct” in these instances than the Codex Telleriano-Remensis, its supposed model, he concluded that the Vatican codex was neither a direct or indirect copy of the Paris manuscript but that the two derived independently from a lost common prototype. Glass and Robertson (1975:136-138) later christened this hypothetical prototype the Codex Huitzilopochtli, although Robert Barlow (in Abrams 1976:70), who originated the term, clearly intended it for the native style tira that served as the common model of the migration section, not for the entire manuscript. (9)

Eloise Quiñones Keber made considerable use of artistic comparison with comparable figures in other codices drawing on pictorial depictions in the Codex Borgia, Codex Borbonicus, and the Codex Aubin to dispute J. Eric S. Thompson’s original arguments, and traced some of these arguments to the fact that Thompson referenced the chromolithographic reproduction of the Loubat facsimile, not the original or a photographic reproduction, whose color alterations led to some honest errors on Thompson’s part. She also appears to have done something relatively few folks had thought to do: go back and look at Barlow’s lecture notes.

Barlow’s Lecture Notes

So what did Barlow actually say about the Codex Huitzilopochtli? We know that references to it in the literature apparently came from his unpublished lecture notes. Well, as it turns out, someone actually published some of those lecture notes. While it isn’t clear if these are exactly the notes Anne Garges de Forrest saved that Robertson and Nicholson cited, they appear to be the right period and content.

Barlow’s death largely cut short his career, but the material he had produced continued to see publication and sometimes re-publication, so his influence on Mesoamerican studies was more pronounced than its duration might indicate. In 1976, H. Leon Abrams, Jr. published several of Barlow’s notes as a three-part series in Katunob: A Newsletter-Bulletin on Mesoamerican Anthropology (vol. IX, no. 1-3). The relevant section of Barlow’s notes is in part 2 (Katunob vol. IX, no. 2), which originally consisted of a mimeographed document for attendees at a February 1950 graduate seminar on Mesoamerican codices.

Over the course of the lecture, Barlow discusses the content of the various codices, including the migration narratives. He begins with the Codex Azcatitlan—Barlow had published a facsimile of this codex, with commentary, in Journal de la société des américanistes in 1949, so he would have been most intimately familiar with this work—but then the lecture turns to the Codex Huitzilopochtli. His focus is on the migration narrative, which was also a feature of the Codex Azcatitlan, and how it differs from other migration narratives:

Codice Huitzilopochtli – another deviant route, known only through copies in Codice Telleriano and Vaticano A. Both published by Loubat. Of this 3 part Ms. the 2d part is the “Huitzilopochtli” migration route, originally a long strip which has been copies [sic] in leaves and made unintelligible. By having the leaves stuck together again in a long strip, the footprints at least make sense. It has been called the Codex Huitzilopochtli, because that God appears on the 1st page as a traveller guiding his people.

Without Barlow’s actual spoken lecture to guide us, there are still gaps, but a few things are clear. When Barlow discusses the Codex Huitzilopochtli, he’s referring only to the migration segment of the Codex Telleriano-Remensis and Codex Ríos. He calls it a “deviant route” because the Telleriano-Remensis/Ríos sequence differs from other Aztec migration traditions, the pictographic elements beginning in the seven caves of Chicomoztoc rather than Aztlan, and the glosses identify ethnic groups more associated with the Puebla than the Mexica. For more detail, see Elizabeth Hill Boone’s Stories in Red and Black: Pictorial Histories of the Aztecs and Mixtecs (2000), which is a comprehensive comparison of the different migration accounts in surviving codices.

As with most of his fellow Mesoamerican anthropologists, Barlow presumes that the original manuscript they drew on was a tira, a long roll or folded document that could be extended out so that the entire migration narrative could be seen as a long strip.

A little later in the lecture, Barlow discusses the rest of the Codex Telleriano-Remensis and Codex Ríos, which he grouped together, and he notes:

The two Mss. seem to have been copied from a lost anthology. Eric Thompson wrote a comment on this Ms. reconstructing the original Ms. (J. Eric Thompson “The Prototype of the Mexican Codices Telleriano-Remensis and Vaticanus A. Notes on Middle American Archaeology and Ethnology, Carnegie Institution of Washington, Division of Historical Research, No. 6, page 24, October 1, 1941.)

The result is probably much more conservative than readers might have thought. Barlow was apparently working mostly from Loubat’s facsimiles. He basically followed Thompson in his assessment that the Codex Telleriano-Remensis and Codex Ríos were independently copied from a prototype, which in turn echoed similar thoughts Barlow had about prototype manuscripts like Crónica X. Barlow’s focus was on the migration section, which was most similar to the codices he himself had studied in detail, and he mainly notes that this migration narrative is divergent from those in other codices.

Interestingly, Barlow himself does not claim in his notes that he came up with the idea of the Codex Huitzilopochtli—but his use of it has led, as we have seen, to the convention. The relative obscurity and limited availability of these notes probably had much to do with the misconceptions about the Codex Huitzilopochtli over the years.

In addition to these notes, there are some scattered references to the Codex Huitzilopochtli in Obras de Robert H. Barlow volumes IV and V. This is a Spanish-language reprint of Barlow’s materials, published and previously unpublished, including materials in the Barlow Archive at the University of the Americas in Mexico. Volume IV is Extent of the Empire of Culhua Mexica, with the first instance of “Codice Huitzilopochtli,” already mentioned above.

Volume V opens with an incomplete draft for a project Barlow had begun between 1945 and 1948 on sources of pre-Hispanic history (“[Fuentes para la historia prehispanica]”); it is little more than a detailed outline (in English in the original, in Spanish in Obras de Robert H. Barlow V), but it includes:

Prototipo de los anales contenido en el Vaticano A y el Telleriano. 1. Estos códices copian otro documento aparte del Códice Huitzilopochtli, uno anales de los siglos XIV a XVI. 2. Es importante ver las pinturas y no el comentario anexo que a veces es muy equivacdo.	Prototype of the annals contained in Vaticano A and the Telleriano. 1. These codices copy another document apart from the Codex Huitzilopochtli, an annal from the 14th to 16th centuries. 2. It is important to see the paintings and not the attached commentary which is sometimes very misleading.
Obras de Robert H. Barlow V.10	English translation

The interesting thing about this brief reference is that it reiterates Barlow sees the Codex Huitztilopochtli as the prototype for the migration narrative, not the other material in the Codex Telleriano-Remensis and Codex Ríos. A little later in the outline, Barlow expands:

Códice Huitzilopochtli Podemos dar este nombre al códice que está copiado en el Vaticano A (Ríos) y el Telleriano. 1. Thompson, Carnegie [Institution of Washington] Notes, [f.21] (No. 6, 1941). Señala que estos dos códices derivan de un original perdido. 2. El asunto es diferente de la Tira y el Sigüenza. Son varias migraciones, inclusive una tlaxcalteca y quizá no debe considerarse mexicana.

Codex Huitzilopochtli We can give this name to the codex that is copied in Vatican A (Ríos) and the Telleriano. 1. Thompson, Carnegie [Institution of Washington] Notes, [f.21] (No. 6, 1941). He points out that these two codices derive from a lost original. 2. The subject is different from the Tira and the Sigüenza. There are several migrations, including one from Tlaxcala and perhaps it should not be considered Mexican.

Obras de Robert H. Barlow V.11

English translation

“Sigüenza” is a reference to the “El Mapa de Sigüenza” which is an Aztec cartographic migration narrative; “Tira” in this context is a reference to the Codex Boturini, which was also known as Tira de la Peregrinación de los Mexica since it too contains a migration narrative (and had been discussed almost immediately before this part of the outline). Again, Barlow is focused narrowly on the migration narrative of the two codices, and how they differ from others—the reference to Tlaxcala is in relation to the different ethnic names, which is a major point of divergence from other migration narratives. Near the end of the draft, Barlow wrote:

Figura una migración tlaxcalteca en el Códice Huitzilopochtli. (Véase: Valle de México. Pinturas).	A Tlaxcalan migration is depicted in the Codex Huitzilopochtli. (See: Valley of Mexico. Paintings).
Obras de Robert H. Barlow V.25	English translation

A Spanish translation of Barlow’s lecture notes from Katanub are also printed in volume V.

There may be some additional materials from Barlow on the Codex Huitzilopochtli which have not yet seen print. The volume Robert Hayward Barlow: Obra histórico-antropológica (2005) contains an index of the Barlow Archive, including a folder listing for the Codex Huitzilopochtli:

Carpeta 66 5 FF. 66.1 Foto sobre cartón de un detalle de algún códice del cual no se anota el nombre. 66.2-4 Notas referentes al Códice Huitzilopochtli. 66.5 Dibujo grande a lápiz y tinta del Códice Huitzilopochtli con anotaciones.	Folder 66 5 pages 66.1 Photo on cardboard of a detail of a codex whose name is not noted. 66.2-4 Notes regarding the Codex Huitzilopochtli. 66.5 Large pencil and ink drawing of the Codex Huitzilopochtli with annotations.
Robert Hayward Barlow: Obra histórico-antropológica 159	English translation.

Scanty detail, but intriguing. The pencil and ink drawing is probably a sketch trying to reproduce the tira form of the original Codex Huitzilopochtli from the pages of the Codex Ríos. Other notes from Barlow on the Codex Huitzilopochtli might be buried elsewhere, such as among his correspondence with Thompson, also in the archive. Perhaps someday an intrepid scholar will make the trip to the University of Americas, do a little digging, and report back.

The Codex Huitzilopochtli Cannot Die!

Eloise Quiñones Keber’s study of and publications about the Codex Telleriano-Remensis and Codex Ríos and their historiography have substantially shifted the scholarly consensus. In part, this is because in 1995 the University of Texas published Codex Telleriano-Remensis: Ritual, Divination, and History in a Pictorial Aztec Manuscript, with transcription, translation, and commentary by Eloise Quiñones Keber. Here, she once again reiterates her reference to Barlow’s notes:

Because the figure of Huitzilopochtli dominates the initial folio of the migration, Barlow (in Abrams 1976:70) called this section the “Huitzilopochtli” migration route and “Codex Huitzilopochtli.” Glass and Robertson (1975:136-139) applied this name more broadly to the prototype(s) of the entire Codex Telleriano-Remensis and the Codex Vaticanus A, but it is doubtful that this was Barlow’s intent. As noted, the rule of the migration account differs from the rest of the historical section and undoubtedly derived from a different pictorial source. Because of its anomalous character, judgments based on the migration section should be applied very cautiously to other sections of the manuscript or to the manuscript as a whole. (203-204)

A reflection of this shift is a post-millennial decline in references to the Codex Huitzilopochtli in the context of the Codices Telleriano-Remensis and Ríos; sources that cite Quiñones Keber don’t tend to mention the Codex Huitzilopochtli or Barlow in this context. Elizabeth Hill Boone doesn’t mention the Codex Huitzilopochtli at all in Stories in Red and Black: Pictorial Histories of the Aztecs and Mixtecs (2000) or Descendants of Aztec Pictography: The Cultural Encyclopedias of 16th-Century Mexico (2020), which are comprehensive works in comparing Aztec Codices and identifying cognates.

The exception is when a scholar is reiterating in a historical or historiographical context, in which case they might follow Quiñones Keber. For example, Henry B. Nicholson in Topiltzin Quetzalcoatl The Once and Future Lord of the Toltecs (2001) cites Eloise Quiñones Keber and recaps her interpretation of the Codex Huitzilopochtli. Notably, Nicholson does not cite Barlow’s notes, but only repeats Quiñones Keber. By 2001, these notes had been out of print for almost 25 years, and could be difficult to source.

However, there are still works that cite the Codex Huitzilopochtli, Barlow’s coining of the term, and even Glass and Robertston’s Huitzilopochtli Group without reflecting Quiñones Keber’s corrections to the common misconceptions that had crept in over the decades.

Danna Alexandra Levin-Rojo in A Way Back to Aztlan: Sixteenth Century Hispanic-Nahuatl Transculturation and the Construction of the New Mexico (2001), later published as Return to Aztlan: Indians, Spaniards, and the Invention of Nuevo México (2014), cited older material by Glass and Robertson, and subsequently reiterates their claims about Barlow and the Codex Huitzilopochtli. Manuel Aguilar-Moreno in Handbook to Life in the Aztec World (2007) is also reliant on older sources, and continues to present the Codex Huitzilopochtli as the source manuscript for both books (or for the Codex Telleriano-Remensis, which the Codex Ríos copies).

In 2006, Juan José Batalla Rosado published “Estudio codicológico de la sección del xiuhpohualli del Códice Telleriano-Remensis” in Revista Española de Antropologia American vol. 36 no. 2; in this codicological study Batalla Rosado concluded that the Codex Ríos was not a copy of the Codex Telleriano-Remensis, but was independently copied from the original manuscript that gave rise to the Codex Telleriano-Remensis. (This conclusion was later challenged by Gláucia Cristiani Montoro in 2010 with the more extensive study “Estudio codicológico del Códice Telleriano-Remensis” in Revista Española de Antropologia American vol. 40 no. 2.) This might be why when Batalla Rosado wrote Chapter 2 of the Oxford History of the Aztecs (2016), he resurrected the Codex Huitzilopochtli and the Huitzilopochtli Group.

The reliance of these sourcebooks in particular on older scholarship has the effect of perpetuating old ideas. For example, Batalla Rosado is cited as the source for this passage:

Likely written and drawn in Italy after 1566, Codex Ríos is a copy of the Codex Telleriano-Remensis, itself a copy of the now-lost Codex Huitzilopochtli.
—Mackenzie Cooley, “The Giant Remains: Mesoamerican Natural History, Medicine, and Cycles of Empire” in Isis vol. 112, no. 1(Mar 2021), 57

So instead of the Codex Huitzilopochtli being Barlow’s stand-in name for the tira of the migration route, it’s now being used as shorthand for the hypothetical anthology of manuscripts that were copied to make the Codex Telleriano-Remensis. Cynthia L. Stone references the Codex Huitzilopochtli in the same way in In Place of Gods and Kinds: Authorship and Identity in the Relación de Michoacán (2017), 237n28, which shows that this isn’t just a one-off.

It’s important to note that neither of these books is primarily concerned about the Codex Huitzilopochtli for its own sake; they just missed the correction that Quiñones Keber noted, depending on some standard source that had likewise missed her book and papers.

Conclusion

It’s important to note that the basic idea that the Codex Telleriano-Remensis and Codex Ríos represent an anthology of different indigenous texts, copied and then glossed or annotated by European writers, has seen more or less widespread agreement for a little over two centuries. J. Eric S. Thompson’s brief but influential article only reflected his difference of opinion on the line of transmission from those sources to the European-style codices that have come down to us. Yet neither Thompson or anyone else appears to have given this prototype a name until it crops up in Barlow’s lecture notes.

This is why R. H. Barlow’s Codex Huitzilopochtli has had sticking power. Like Lovecraft’s Necronomicon, it is an evocative name; more than that, it fills an ideological niche. While Barlow’s notes indicate he clearly intended this title to only apply to cognate parts of the two codices—the migration route—the way it was expanded by others to include the entirety of the hypothetical prototype manuscript, and even adopted by Glass and Robertson for the “Huitzilopochtli Group,” seems to indicate that Mesoamerican scholarship hadn’t really had a good terminology for blocking and discussing this material together before.

The continued use suggests maybe it still doesn’t have a better term. Although I’ve seen the reference “Ríos Pair” in at least one context, it doesn’t seem to have caught on broadly as a reference to the two codices, and there doesn’t seem to be a single more accepted term for the hypothetical prototype manuscript. Codex Huitzilopochtli is, if nothing else, pithy, catchy, and memorable.

I suspect, barring the re-publication of Barlow’s notes in a more available format or a concerted effort by Mesoamerican scholars and editors of books and journals to correct for it, the Codex Huitzilopochtli will continue to crop up here and there. That’s an ongoing issue in all sorts of scholarship: it takes time and effort to stay current, and in the gaps, some zombie ideas can continue to lurch through bibliographies and footnotes for years or decades after they’ve been identified and corrected.

†††

Thanks to Martin Andersson and Dave Goudsward for their help sourcing some materials.

---

Bobby Derie is the author of Weird Talers: Essays on Robert E. Howard and Others and Sex and the Cthulhu Mythos.

Deep Cuts in a Lovecraftian Vein uses Amazon Associate links. As an Amazon Associate I earn from qualifying purchases.