About TCAW - Subscription Info
February 2002
Vol. 11, No. 2
pp 27–28, 31.
Instruments & Applications
Smoking out the past

Anthropology uses GC-MS to determine the botanical origins of residue in ancient clay pipes from Chile.

opening art
Anthropological excavations suggest that pipe smoking has been important to Chilean culture since the time of Christ. Although most pipes are found on the surface, making it impossible to assign them to an exact archaeological period, the pipes can be assigned to a cultural period based on style and other artifacts that are found in the same area.

An exciting collection of pipes at La Granja on the northern end of the Cachapoal River, about 75 miles south of Santiago, is the subject of a current excavation led by anthropologist Maria Teresa Planella (1). Whereas most sites yield only a few fragments, the La Granja site yielded 643 pipe fragments along with vases, beads, nose rings, grinding tools, a figure of a human, remains of cultivated and wild plants, and an ellipsoidal structure made of river stones. Based on the presence of nose rings, this site has been assigned to the Complejo Llolleo cultural tradition, which dates from approximately 200 B.C.E. to 900 C.E. (2).

Almost all of the pipes are made of clay, shaped as an inverted “T” with two mouthpieces, and predominantly without ornamentation. Planella and her colleagues proposed a ritual context for the pipes due to their number and the piles of stones, which indicate that ritual was important in the La Granja culture. Also, the two mouthpieces suggest the pipes were smoked simultaneously by a shaman and an initiate.

Historical Smoking Habits
When the Spanish colonists arrived in South America, they documented many aspects of the native culture, including tobacco smoking. The availability of hallucinogens, however, has been documented only recently. In 1991, Costino Torres and others found the hallucinogenic compounds dimethyltryptamine, 5-methoxydimethyltryptamine, and 5-hydroxy-N,N-dimethyltryptamine in pre-Hispanic snuff found in a dig along the San Pedro River in the Atacama desert of northern Chile (3). These compounds are indicators of hallucinogenic plants from the genus Anadenanthera, which has been used in recent times by several South American Indian tribes, including the Mapauche Indians of Chile. The San Pedro discovery proves that these plants were indigenous to that area of pre-Hispanic Chile. If hallucinogens were available, it is possible that the Indians smoked them. Consequently, the anthropologists desired an analysis for hallucinogenic residues in the La Granja pipes.

Initial Analysis
In 1994, Planella brought fragments of four ceramic pipes from the La Granja site to Orlando Munoz, a chemist at the University of Chile who enlisted our help. The pipe material is tan and coated on the inner surfaces with a black residue. The residue, however, was integrally bonded to the pipe body, so we were unable to remove a pure sample.

Tools of the Trade
We used a Hewlett-Packard 5890 GC with a Hewlett-Packard 5971A mass selective detector. A slow temperature ramp from 50–250 °C over 2 h was used. The column was a 95% methyl–5% phenyl Alltech econo-cap column, 0.25 mm wide and 30 m long. The injector was set at 275 °C, and the transfer line to the mass selective detector was at 280 °C. The Wiley 130K Mass Spectral Database Rev. A.00.00 was used to match mass spectra.
A small piece of pipe was ground into powder and soaked in ethanol. The supernatant was then decanted, concentrated, and injected into the GC-MS (see box, “Tools of the Trade”). The extracts obtained from outer and inner surfaces of the four pipes were found to contain hydrocarbons, phthalates, several essential oils, and benzoquinone. For example, the GC-MS analysis of the soft black interior of pipe #1 indicated the presence of myrcene, carene, geranyl acetate, linalyl acetate, and caryophyllene. We were disappointed, however, by the lack of nitrogenous material that would have indicated alkaloids.

Several spot tests for nitrogenous materials were then performed, using Dragendorff, dimethylaminobenzaldehyde-FeCl3, and potassium iodoplatinate reagents to visualize TLC spots, and Marquis and Mecke’s reagents on a spot plate. All gave positive tests for the controls (a morphine tablet and/or tobacco) but negative tests with the pipe extracts (except for the Dragendorff, which gave positive tests occasionally but inconsistently).

Frustrated by the failure to find compounds that are clearly hallucinogenic, we extracted larger samples and repeated the Marquis test, which showed a faint but definitely positive signal of alkaloids. Another sample was tested for nitrogen (as cyanide) by sodium fusion and the 4-nitrobenzaldehyde-1,2-nitrobenzene reagent. The solution turned purple, positively indicating nitrogen.

Encouraged, we repeated the GC-MS analysis, but first soaked the pipe samples in ammonia before they were extracted with chloroform or methanol. These samples gave similar results to the previous samples that were extracted with only ethanol.

Burning Questions
It occurred to us that we did not know what effect smoking and lengthy burial might have had on the plant material. A study of embalming materials used in ancient Egypt that investigated the effects of smoldering suggested that the products resulting from smoking the plants might be quite different from the compounds in the plants originally (4). Consequently, we burned the alkaloid-containing substances Ma Huang (ephedra), morphine, and a commercial cigarette. These and an unburned sample of each were treated with ammonia, extracted with methanol, analyzed by GC-MS, and the results compared with those of a similarly treated pool of all eight pipe samples.

The GC-MS of the pipe samples was consistent with previous findings. However, although morphine was the major ingredient in the morphine tablet and nicotine was the major ingredient in the cigarette, no morphine was present in the burned residue of the tablet and only a trace of nicotine was present in the cigarette ashes. These results confirmed our suspicion that burning destroyed most of the active ingredients. Furthermore, these results gave us the first hint of what might have been smoked in the pipes. The chromatogram of burned Ma Huang indicated three compounds that matched the pipe extracts: benzyl alcohol, a phthalate eluting around 78 min, and another phthalate eluting just before 89 min (Figure 1). The phthalates also attracted our attention to a finding that had not seemed important to us earlier—the presence of a series of phthalates in the pipe extracts.

Phthalates are commonly encountered as contaminants, however, and may have entered the pipe material from the plastic bags in which they were shipped. But, because the compounds were also present in a second shipment that was not packed in plastic, and because of the presence of compounds in the burned Ma Huang with the same mass spectra as the corresponding phthalates from the pipe, we were convinced that at least these two phthalates were true components of the pipe residues. An attempt to match them with known phthalates, however, was unsuccessful.

The 78-min GC peak from the pipe and Ma Huang extracts gave mass spectral peaks at 149, 163, and 221, and eluted roughly midway between dibutyl and dihexyl phthalates. Similarly, the 89-min GC peak gave mass spectral peaks at 149, 207, and 263. Dihexyl phthalate, by contrast, eluted at 88.6 min and gave the expected masses of 149, 233, and 251 (5).

The Origin of Mass Species
To solve the puzzle, we decided to compare the pipes with ashes of Chilean plants that contained hallucinogens and requested samples of any plants that Munoz could obtain, with particular interest in Chilean tobacco and the hallucinogenic plant family Anadenanthera. A portion of nine plant samples, including one tobacco and two varieties of Anadenanthera, was burned and extracts of both the plants and the ashes were prepared for analysis.

All burned samples, including the pipes, have very large peaks at 5–8 min that appear to be acetamide. Comparison with earlier results showed no other consistent similarity of the pipes with the plants. Burned samples of Criftocaria alba and Anadenanthera peregina both contain a phthalate, but its retention time of 98 min does not match that of the pipes.

The similarity between the plants and their ashes was also minor. For example, Peumus boldus gave a large peak that the Wiley library identified as ascaridole (a nicotine-like compound), but after burning, this peak disappeared. Latua publiflora gave a peak identified as hyoscyamine but, again, this peak was absent from the sample that had been burned. As with the cigarette, only a trace of nicotine remained in the ashes from the Chilean tobacco. These observations confirm our previously reported finding that burning destroys active ingredients.

There is also little similarity between the plants. Even extracts of Anadenanthera peregina and Anadenanthera peregina lordidad, which are varieties of the same species, are dissimilar. Since the plants have so little in common, it seems very improbable that we will chance upon the one responsible for the black residue, assuming it still grows in Chile and has not undergone significant mutation in the past 1000–1500 years.

Previous workers have attributed the failure to detect psychoactive compounds in archaeological specimens to the small amounts of material that were available for analysis and to the lack of any identifiable plant parts. Our findings, however, suggest that alkaloids do not survive being burned or smoked. Furthermore, because only traces of alkaloids, if any, survive, determining the actual identity of the alkaloids is unlikely.

The presence of acetamide and the positive spot tests demonstrate that the Chileans did smoke alkaloid-containing material. Since many alkaloids are hallucinogenic, the pipes were possibly used to smoke plants for a hallucinogenic effect. Further, because Ma Huang contains benzyl alcohol and two of the same phthalates as the pipes and because Ma Huang contains the alkaloid ephedra, the alkaloid-containing material in the pipes was probably an ephedra-like compound.


The authors would like to acknowledge Orlando Munoz, Michael Hanna, Jr., James Patrick, Harold McNair, Dorothy Mulberry, Ivy Arbulu, Al Sneeden, and David Koontz for their technical and editorial assistance.


  1. Planella, M. T.; Falabella, F.; Tagle, B. Actas XIV Congreso Nacional de Arqueologia Chilena (in press).
  2. Westfall C. Rev Chilena de Antropología (Santiago). 1994, 12, 123–161.
  3. Torres, C. M.; Repke, D. B.; Chan, K.; McKenna, D.; Llagostera, A.; Schultes, R. E. Curr. Anthropol. 1991, 32, 640–649.
  4. Weser, U.; Kaup, Y.; Etspler, H.; Koler, J.; Baumer, U. Anal. Chem. 1998, 70, 511A–516A.
  5. Kamer, R. Amer. Lab. 1999, 31, 32–35.

Hampton H. Hairfield, Jr. is a chemistry laboratory instructor and Elizabeth M. Hairfield is a professor of chemistry at Mary Baldwin College (Staunton, VA) Send your comments or questions regarding this article to tcaw@acs.org or the Editorial Office 1155 16th St N.W., Washington, DC 20036.

Return to Top
|| Table of Contents

 CASChemPortChemCenterPubs Page