Introduction

Humans have always used the natural materials around us to produce functional objects and works of art. Paintings and other objects that are part of our cultural heritage, including textiles, books, sculptures, archaeological objects, contain a wide variety of organic materials from natural to synthetic. Archaeological excavations often bring to light a wide variety of objects and materials that have been collected, processed, and used by humans over time. Due to their long period underground, some of these materials and objects, especially those of an organic nature, have been partially or totally altered. Organic materials are more subject to degradation than inorganic ones, so if we can understand their composition, then we can ensure that ancient artifacts will remain part of our cultural heritage (Ribechini et al. 2008).

The study of organic materials from archaeological contexts is complicated from the fact that the materials used by ancient cultures varied widely depending on their availability, which was strictly related to geography (Bonaduce et al. 2016). Natural resins, yarn, wood, plant, and insect exudates found in artifacts have been studied with Fourier transformed infrared (FTIR) spectroscopic analysis both with portable and laboratory instruments (Bisulca et al. 2017; Invernizzi et al. 2018; AMCTB No.106 2021).

A set of non-destructive analysis such X-ray fluorescence (XRF), FTIR, and imaging techniques at synchrotron facilities—where the objects are transported to the laboratory—have been applied mainly in textiles, leather and parchment, paper, papyrus, and wooden objects (Bertrand et al. 2012). Research shows a trend to develop more sensible instrumentation to perform non-destructive portable analyses, especially considering the study of large collections on museums (Adriaens 2005; Teodor et al. 2010; Matheson & McCollum 2014; La Nasa et al. 2022; Chen et al. 2023).

Moving towards invasive and destructive analysis, nuclear magnetic resonance (NMR) studies have been reported on copal, rubber, and plant exudates from conifers (Hosler et al. 1999; Lambert et al. 2000; Lambert et al. 2010; Lambert et al. 2016; Capitani et al. 2012). To characterize organic materials, the most adopted analytical approach has been with gas chromatography/mass spectrometry (GC/MS) to mention some examples: resinous exudates, copal, amber (Mathe et al. 2004; Modugno and Ribechini 2009; Scalarone and Chiantore 2009; Burger et al. 2011; Lucero-Gómez et al. 2014; Brettell et al. 2015, Kaal et al. 2020; Langejans et al. 2022).

From the literature review, information regarding chemical characterization of organic natural materials is abundant for European archaeological contexts and remains scarce for other regions. Also, characterization is focused on archaeological resins, while archaeological rubber only has yet a few publications.

When talking about rubber, it is usually associated with car tires, shoe soles, etc. (which have now been replaced with synthetic polymers), but we seldom reflect on its origin. Natural rubber is extracted from several trees, with the most studied being Hevea brasiliensis (endemic from Brazil), as it has been heavily exploited in past centuries and currently by different industries (Arreguín 1958). There is a Mexican variation called Castilla elastica that did not receive any industrial interest—save for some rare cases, such as waterproofing dresses or in the sleeves of some cowboy shirts—but was nevertheless known and used in Mesoamerica.

Rubber, the product derived from the olqáhuitl, a tropical tree botanically identified as Castilla elastica was—and still—is a precious material among the indigenous groups of Mexico. This sacred material was used for a multitude of purposes, both for religious and ritual matters and for everyday acts. The presence of rubber objects among Mesoamerican groups is recurrent and reiterated, as archeology has shown on numerous occasions. To point out some examples, it already appears in Olmec contexts, as can be seen in the offerings recovered from El Manatí, State of Veracruz, Mexico, and which will be explored further on. The same occurs in Classic Maya times (ca. 200 to 900 A.D.) with its use as an offering in the funerary complexes of Classic Maya rulers of Tikal or Calakmul.

Artifacts made of rubber are present in first-order ritual contexts such as the balls found in the tunnel under the pyramid of the Feathered Serpent in Teotihuacan, the rubber ball from the Central Sierra Nevada in Alta California, the objects found in the Sacred Cenote of Chichén Itzá, or the rubber objects documented in the offerings of the Templo Mayor of Tenochtitlan, just to name a few (Gómez and Gazzola 2015; Moratto 2009; Coggins and Ladd 1992; Filloy 2001; Carreón 2016).

This notorious and continuous presence denotes the importance of the material for the Mesoamerican people, as well as the multiple uses that the indigenous groups gave to this unique material. One of the main uses, and the one of interest for this research, is for the elaboration of balls in the Historia General de las Cosas de la Nueva España (1540–1585), the extensive work of Fray Bernandino de Sahagún whom with his indigenous collaborators, describes the use of rubber obtained from the olqáhuitl for the manufacture of different objectsFootnote 1 (de Sahagún 1906–1907).

In Mesoamerica, a ritual practice among many of its cultures was the ball game. Reports of XVI century about the ball and the practice of the game are repeated in the works of Spanish friars such as de Torquemada, Toribio Motolinía, or Diego DuránFootnote 2 (de Torquemada 1977; Motolinía 2014, Durán 1867). The ball and game descriptions are repeated in the later historical sources that record it in northwestern Mexico, as they refer to the ball with eyewitness records and reports (Carreón 2014).

Variants of the ball game are still practiced today in various parts of Mexico, such as Pok-ta-Pok in the Southeast or Ulama in the North, each with its own rules and variations. The practice of the nowadays sport is not limited to Mexico, countries such as Guatemala, Belize, and El Salvador and other Central American countries also register active teams of this ancient ritual game and the fabrication of the ball with rubber from the Castilla elastica tree is a main feature of its practice. The modern success of this sport can be highlighted by the celebration since 2015 the Mesoamerican World Cup of Ball Game (Panqueba et al. 2022).

From this concise tour of rubber, its ancient and current uses, the permanent presence of rubber among indigenous groups are observed. Having already stressed the importance of this material for Mexican culture, it is relevant to contextualize this research to focus on the nature of this material. Since it is precisely the composition of this material and its production technology in Olmec times that leads the interest of this project.

Rubber is a material that is extracted from a very diverse variety of tree families, and in the case of Mesoamerican cultures came from a tree endemic of the southeastern region of Mexico, the Castilla elastica, as noted above. Thus, the product known as rubber is obtained from processing the coagulum of dehydrated latex that sprouts from different tree species and that, due to the characteristics of elasticity and resistance, was used to make a multitude of objects. Among the properties and characteristics that surely captivated the Mesoamerican people, it is possible to highlight the chromatic transformation of latex when it dries, as well as its elasticity, its combustibility, and even its scent (Filloy 2001; Carreón 2006). Regarding its manufacture, using Helical Computer Tomography on an Aztec ball found at Templo Mayor provided evidence that the balls were made of several thin layers (strips) of rubber arranged in a spiral, with particles of powder placed on the rubber strips (Filloy 2001).

Although anthropological and iconographic studies on the ball game have provided some proposals for the manufacture of the rubber ball, still each Mesoamerican region had access to different raw materials and probably different manufacturing techniques (Filloy 2001; Stone 2002; Tarkanian and Hosler 2011). Therefore, technological and compositional studies need to be performed to the different Mesoamerican rubber balls to provide a comprehensive view of these extraordinary objects.

Sacred ancient Olmec site: El Manatí

The term Olmec is of Nahuatl origin Olman and roughly translates to “The rubber country,” as their homeland in southern Veracruz is where this plant was widely harnessed. Use of the name Olmec to refer to this ancient culture from the Gulf Coast was employed by scholars many years ago, before it had been established that rubber was discovered by these groups (Diehl 1996). The presence of rubber artifacts was recorded by chroniclers such as Sahagún and can be seen in the Mendoza codex—which relates the tributes given by the different señoríos to the Mexicas—that indicates the payment of large quantities of rubber balls and sandals made by the people from this region (Berdan and Reiff 1992).

The discovery of rubber balls dated to stage Manatí A 1700 BC (calibrated) in Mesoamerica as part of the Proyecto Arqueológico Manatí in southern Veracruz provides the oldest evidence of rubber made artifacts. These balls were associated with religious ceremonies, as an offering for the mountain, its freshwater springs, and to the specular hematite deposits. This dating is contemporary to the oldest architecture of ball court at Paso de la Amada (1400 BC uncalibrated) corresponding to the Locona phase of Soconusco region (Hill et al. 1998).

The sacred site of “El Manatí” is located approximately 15 km southeast of the archaeological zone of San Lorenzo, in the south of the state of Veracruz, near the so-called Olmec nuclear area (Fig. 1). The floodplain on which the offering site was found is annually flooded by the Coatzacoalcos River and is next to various water springs. The offerings’ site was discovered by residents of nearby communities in 1988 while constructing ponds for fish farming at the foot of the “El Manatí” hill. The archaeological work carried out yielded data that allowed dividing the occupation of the site into three construction stages, temporarily placing it between 1700 and 1040 BC. The first two stages are called phase A and B corresponding to 1700 to 1600 BC and 1500 to 1200 BC, respectively, while the third phase called Mayacal dates from 1100 to 1040 BC (Ortiz and Rodríguez 2000; Ortiz et al. 2015).

Fig. 1
figure 1

a Location of El Manatí archaeological site. b Photogrammetric image of ORD03 by A. Mitrani

Full size image

The project recovered 15 rubber balls, some of which were accompanied by jade axes and other greenstones. Interestingly, most of the balls were found in the oldest deposits and were sealed by a layer of peat or other compact vegetable material. Carbon 14 dating of this material, as well as the ceramic found at the bottom, has placed these balls as old as 1600 BC.

The most impressive of the offerings was the one containing six balls of different sizes together with 46 jade axes, the largest of which was 30 cm in diameter. Another ball 10 cm in diameter was found with six axes made with the finest translucent light green jade and was deposited over the bedrock. Three other small balls (diameters of 12, 13, and 8 cm) were found with a north-south alignment, with no axes. The last rubber ball was found flattened (with 10 cm), also without any axes, but had an appendix, presumably used to hang the ball (Ortiz et al. 2018). The variation in size of the balls suggests that multiple ballgames that required different ball sizes were being played in this one place at the same time.

Two other balls (20 and 22 cm in diameter) interestingly correspond to the second stage of the massive wooden busts offering and were found with two wooden canes associated to them. These are dated to 1200 BC. Additionally, two more balls were handed over by villagers, and two more by Mr. Alfonso Olamendi, all of them without context.

However, their presence may suggest that the Olmecs were among the first who practiced this ritual game that has practically remained up until the present day. How and where it was played, and its rules for this early stage are unknown; however, four possible ballcourts formed by two small, elongated platforms with head were found at the nearby site of El Mayacal, contemporary to El Manatí (1200–900 BC). In later stages of the pre-Hispanic era, the architectural structure of the ballcourt evolved into its characteristic double T form, which can be found throughout Mesoamerica, highlighting El Tajín in Veracruz and Cantona in Puebla where these structures are abundant.

Since their discovery, the main concern has been the conservation of the balls, which has proven to be complex. Following a request by D. Hosler, M. Tarkanian, and Ing. J. García Bárcenas in 1998, different samples were taken on two balls to determine their chemical properties and manufacturing technique. The analysis performed on this samples aimed to explain on how the physical properties of Castilla elastica rubber are modified by varying the proportion of Ipomea alba juice in reproductions of mixtures of latex and the extract of the plant Ipomea alba (Hosler et al. 1999; Tarkanian and Hosler 2011). These studies present the role of the Ipomea alba in the mixture with 13C and 1H Magic Angle Spinning-NMR, GC/MS, and FTIR with reference samples and two Manatí rubber samples. The results indicate that to obtain properties such as greater elasticity, it was necessary to maintain a ratio between both components corresponding to 50-50; however, they do not address a detailed compositional analysis or manufacturing technique for the archaeological samples. There is no archaeological evidence for the manufacture of the Olmec balls found at El Manatí offerings so far.

This work is innovative in applying imaging techniques as well as non-destructive portable techniques for the study of the collection of rubber balls found at El Manatí. These studies were complemented by performing chemical analyses to selected samples with the objective of presenting a complete panorama of the manufacture of the balls and their composition, while also providing data on their current state of conservation.

The objects: the Olmec rubber balls and contemporary rubber balls

We studied 14 archaeological rubber balls from El Manatí. The first of these balls were extracted in 1988. Due to the stable conditions present in the site, the balls had good conservation conditions at the time of their discovery. However, their extraction initiated a series of physical and chemical alterations. Studies dealing with the conservation of archaeological rubber were scarce at their time of recovery.

Currently, these rubber balls present diverse preservation states, which have been considered for the interpretation of the analytical results presented in Table 1. Apart from their previous historical factors, the rubber balls have different conservation histories which have impacted their current physical conditions. ORB02 and ORB06 were donated by Mr. Olamendi, a local inhabitant who first discovered the offerings. Six balls with different recovery dates were sent to restoration labs at the National Institute of Anthropology and History (INAH) in Mexico City, after the archaeological excavation season of 1996. They were slowly dried and stored in plastic polyethylene bags where they remained until November 2012 when they were returned to Veracruz, without any further treatment. ORD07 was preserved in a soil block from the moment of its excavation.

Table 1 List of the studied rubber balls with the nomenclature used for this work indicating their physical characteristics, properties, and a referred preservation state
Full size table

Today, most of these dry objects are seemingly stable (Fig. 2), protected in anoxia enclosures inside bags with oxygen absorbers. The most frequent deterioration in these pieces are cracks and fine fissures, a brittle state, easy fragmentation, and a little shrinkage with weight loss.

Fig. 2
figure 2

Photogrammetric image of ORD07 by A. Mitrani. The rubber ball is preserved on a block the block of soil

Full size image

There are seven balls preserved in wet state; ORW13 and ORW14 have been preserved immersed in a solution of formaldehyde, alcohol, and acetic acid. This last ball was affected by microorganisms, as it was handled frequently to be shown to other archaeologist, conservators, geologists, and other interested academics. The other five wet objects are preserved in refrigeration at 10 °C; however, before 2007, all of these had spent many years inside black bags at room temperature.

These archaeological objects present several deterioration processes, some of these caused by limitations in the infrastructure hindering the application of preventive conservation strategies such as relative humidity, temperature, light exposure, managing, and storing. Common problems are the development of microorganisms due to the relative humidity conditions in Veracruz city, and the cracking and subsequent loss of fragments due to the drying process. Despite this, the following pieces remain stable: ORW09, ORW10, ORW11, and ORW12, while ORW08 is in very bad condition.

Additional to the rubber balls studied, three soil samples (S1, S2, and S3) were obtained from the offering to compare the results from the non-destructive techniques due to the remaining sediment on the surface of the Olmec rubber balls.

Four contemporary rubber balls named CRB 16, CRB17, CRB18, and CRB19 were also characterized as reference material for the manufacturing technique and chemical composition. They were made in Acayucan, Veracruz (for location, see Fig. 1) with latex extracted from a few remaining Castilla elastica trees in the area. In this site, the balls are manufactured by creating a small nucleus and then growing its size by adding new rubber strips, and finally applying a homogeneous layer of latex on top of the ball, as registered and illustrated in Fig. 3. Balls CRB 16 and CRB17 were manufactured in 2020 while CRB18 and CRB19 were made in 2016, 4 years earlier.

Fig. 3
figure 3

Contemporary making of the rubber balls from the Castilla elastica latex. Manufacturing sequence from left to right. Scientific illustration by M. Domínguez indexed at INDAUTOR

Full size image

Methodology of analysis

The studied objects required a set of comprehensive and non-invasive analytical techniques applied in situ, due to the vulnerability of the objects and the best managing for their preservation (Fig. 4a). After the set of imaging and spectroscopic techniques were performed and data analyzed, the next phase of the study required to take a microsample to complement compositional questions.

Fig. 4
figure 4

a Non-contact optical microscopy analysis. Photograph by N. Cano. b Hole made for a previous study in 1998; taking advantage of this feature, a microsample for this study was taken. Photogrammetric image of ORW10 by A. Mitrani

Full size image

To perform additional laboratory analyses, a microsample from the interior of the ball of approximately 50 mg was acquired from ORW10, taking advantage of the previously made hole for the 1998 analyses (Fig. 4b). The rubber mechanical properties were already lost even though the sample was taken from the inside, and it broke into little pieces which were separated in two sections: the closest to the nucleus (INF) and the closest to the external side (SUP). This sample is from the center non-exposed to the ambient part of the ball.

A summary of the applied techniques is presented in Table 2. Olmec rubber balls ORD03, ORD06, ORW08, and ORW11 were not selected for XRF and FTIR analysis since they were brittle, and the required handling could affect the stability of the balls. ORW13 and ORW14 due to their size and immersion conditions were also not analyzed by these techniques.

Table 2 List of characterization techniques employed with the analyzed objects by each technique
Full size table

The analytical techniques performed on the rubber materials were:

Color identification by Munsell color chart

Color measurement of the latex was conducted before and after coagulation on site comparing directly against the color tables included in The Munsell Color Book (Munsell Color Firm 2010).

UV-induced visible fluorescence imaging (UV)

A short-wave (254 nm) UV light was employed to induce visible fluorescence on the surface of the balls to study and register, physical characteristics, organic materials such as the natural rubber and microorganisms.

Radiography imaging (RI)

RI provides information resulting from illuminating the studied object with an X-ray beam, which typically has a photon intensity distribution as a function of photon energy. Contrast on the image is ruled by the ratio of the transmitted and absorbed photons with respect to the ones in the original incident beam, which is described by the Beer-Lambert relation. The probability of a photon being absorbed by the object depends on the elemental composition, the thickness, and the density of the object as well as on the wavelength of the X-ray photons, and the type of interaction of the photons with the object is defined by their wavelength. In this study, the main interaction is represented by the photoelectric effect, along with a significant contribution of Rayleigh elastic scattering and Compton inelastic scattering.

Radiographic images were generated by a combination of a PXM-40BT POSKOM system, based on a TOSHIBA D124 X-ray tube, along with an image acquisition system VIDISCO Flash X PRO array. Distance from the irradiator to the object was fixed at about 200 cm and from the object to the detector at 10 cm. Irradiation conditions were adjusted depending on the thickness of the studied object, ranging from 40 to 70 kv and 40 to 50 mAs.

The images were analyzed with the ImageJ software using the linear attenuation coefficient from the mass attenuation coefficient of rubber from the tables of Hubbell and Seltzer 1995. Simulations were performed considering different densities on the modelled sphere, and the obtained results were compared through the gray-scale histogram of the archaeological balls, thus obtaining data directly related to the object density.

Optical microscopy (OM)

Microscopic observation was made with a Dino-Lite Edge YUY2 USB Microscope. The micrographs were acquired from different areas selected by color changes, fractures, and surface deformations. Images were taken with ×10, ×15, and ×20 magnifications depending on the details observed. Image capture was performed using the DinoCapture 2.0 software.

FT-infrared spectroscopy (FTIR)

Surface analyses of the balls were made on an Agilent Technologies FTIR 4300 Handheld spectrometer with Diamond ATR in a 600–4000 cm−1 region with 8 cm−1 resolution. The infrared spectra were processed using the Origin 2021 software. The microsamples were measured in an Agilent Technologies FTIR Cary 660 spectrometer with Diamond ATR in a 400–4000 cm−1 region with 4 cm−1 resolution and 128 scans.

X-ray fluorescence spectroscopy (XRF)

XRF spectra for both the rubber balls and the samples were acquired with the SANDRA spectrometer developed at IFUNAM (Ruvalcaba et al. 2010) with a molybdenum X-ray source at 45 kV, 0.200 mA, and a 300-s acquisition time. PyMCA software was used to process the spectra and obtain the relative intensities of the detected elements (Solé et al. 2007). Data from relative intensities were processed using the Origin 2021 software.

Combustion elemental analysis (CEA)

Combustion elemental analyses of the microsamples were performed in an Elementar VarioMICRO Cube equipment. The samples were weighted in a microbalance Mettler Toledo XP. Sulfanilamide (Elementar, Batch 109K0086) was used as standard. The analyses were made by duplicate.

13C magic angle spinning-nuclear magnetic resonance (MAS-NMR)

Solid state 13C CPMAS spectra from the samples were obtained in a Bruker Avance 500 spectrometer at 125.78 MHz. The sample was deposited in a 3.2-mm Zirconia MAS rotor with a VESPEL cap. The spectra were obtained with high power decoupling (80 kHz). Adamantane was used as reference (37.7 ppm).

Gas chromatography coupled with mass spectrometry (GC-MS)

The analytical protocol employed was designed to identify chemical markers according to Peters (2005), thus complementing the spectroscopic information from the previous techniques. The following analytical process based on Knapp 1979 was used: a microsample of 5 mg of the internal part of each generated test specimen was weighed, 1 mL of a 1:1 hexane:chloroform mixture was added, the mixture was placed in an ultrasonic bath for half an hour, after which time 50 μL of N,O-Bis(trimethylsilyl)-trifluoroacetamide was added, and allowed to react with constant stirring for half an hour, heating to 60 °C at the conclusion of the reaction, 1 μL was injected into the GC-MS.

The microsamples were analyzed on an Agilent Technologies (CA, USA) equipped with a MMI injection port, linked to a single Quad (EI 70 eV, ion source temperature 230 °C, scanning m/z 50–600, interface temperature 300 °C). The MMI injector was operated in the “constant temperature splitless with purge” mode at 275 °C. GC separation was performed on an HP-5MS chemically bonded fused silica capillary column (Agilent Technologies; 5% phenyl 95% methylpolysiloxane, 30 m × 0.25 mm I.D., 0.25-m film thickness). The GC conditions were as follows: initial temperature 80 °C, 2 min isothermal, 10 °C min−1 up to 200 °C, 6 °C min−1 up to 280 °C, 35 min isothermal. Carrier gas: He (purity 99.9995%), constant flow 1.2 mL min−1 (Mathe et al. 2004; Ribechini et al. 2008).

Peak assignments were performed by interpretation of mass spectra, comparison with mass spectral libraries (NIST 14.0) and with published mass spectra and chromatograms.

Manufacturing El Manatí’s Olmec balls

The ball manufacturing technique begins with the latex extraction from the tree. Liquid latex is white color (N9/0.5 Munsell notation) and once in contact with air, it dries and gradually changes its color to dark brown (5YR/2.5m Munsell notation) as seen in Fig. 5 and illustrated in Fig. 3; this observation is as described by Spanish friar Motolinía (Motolinía 2014).

Fig. 5
figure 5

Color change from a latex from the Castilla elastica tree (N9/0.5 Munsell) to b coagulated (dried) rubber (5YR/2.5m Munsell)

Full size image

The extracted latex is a biomaterial composed by approximately 60 wt% water with the remaining 40 wt% being composed of cis-1,4-polyisoprene, proteins, and other compounds (Jacob et al. 1993). The chemical analysis by FTIR confirms the presence of polyisoprene at bands 2929, 2855, 1445, 1371, 1010, 775, and 730 cm−1 with the cis isomer-specific band at 831 cm−1 (Table 3). None of the rubber balls analyzed presented the bands at 1092, 1021, and 470cm−1 related to a vulcanization process (Socrates 2015).

Table 3 FTIR band assignation and attribution indicating its presence in the studied objects. b, broad; s, sharp; sh, shoulder
Full size table

Difference in FTIR spectra of the ORD and ORW is explained by how these objects were treated and stored. The loss of moisture in the dry balls accelerated the chemical degradation of the polymer, while ORW objects were exposed to relative constant humidity and temperature which resulted in a better conservation of the material (Blank 1990). During the in situ analysis, ORD spectrum shows the highly degradation polymer chain (a strong band at 1736 cm−1 related to the presence of C=O and the absence of the 1660 cm−1 band associated to the C=C bond) at the surface of the objects; meanwhile, ORW spectrum mostly shows the water’s FTIR spectrum (Fig. 6).

Fig. 6
figure 6

Representative FTIR spectra of the different preservation state of the Olmec rubber balls (ORD red line and ORW blue line) in comparison to CRB16 black line and the sample from the center of the ORW10 green line

Full size image

The spectrum of the sample ORW10 analyzed in the laboratory shows a less degraded material in comparison to the surface analysis of the same object (Table 3). The presence of the 1660 cm−1 band associated with the C=C bond and a weak band at 1736 cm−1 related to the presence of C=O) and more like the CRB16 spectrum; this was expected since the sample is taken from the center of the rubber ball without ambient exposure. The spectrum of the sample analyzed with both the handheld and the desktop spectrometer is presented in Fig. S4 since the spectrum obtained with the laboratory equipment revealed other minor signals, but the main chemical structure is present.

The contemporary rubber balls have very similar spectra (Fig. S3). The four spectra show the characteristic bands of polyisoprene. Only CRB17 stands out for a more intense band at 1736 cm−1, which is related to the presence of C=O bonds associated with degradation of polyisoprene. This variation in the spectrum of CRB17 could be attributed to bad storage or previous use of this ball.

The sample from the archaeological object code ORW10 (Table 1) was analyzed by 13C NMR-MAS. The ORW10 (SUP) is nearer to the surface, and in the 13C NMR spectrum of this sample, the signals corresponding to poly-cis-isoprene were observed at 135.6, 125.3, 32.9, 27.9, and 24.1 ppm being this isomer the main constituent of the sample as identified in FTIR (Fig. S5). The width of the signals between 20 and 35 ppm suggested some degree of isomerization to the trans-isomer. The expected signals for the presence of disulfide of polysulfide cross-links at 12.5, 17.9, 44.9, 56.35, and 57.45 ppm (Buzaré et al. 2001) were not present, and these results correspond with the FTIR spectra. No other signals indicating some degradation processes were observed as expected for a sample taken from the center of the ball. In agreement with the absence of sulfide links, the elemental analysis of this sample indicated a low sulfur content of 0.28% (Table 4), corresponding the major percentage to carbon (68.01%) as expected for a carbon-based polymer.

Table 4 Results from the combustion elemental analysis for the ORW10 samples and archaeological soil from the archaeological site
Full size table

The second sample ORW10 (inf) spectrum is almost identical to the previous described (Fig. S6), although in this sample the signals for the olefinic carbon atoms appear wider than in the upper zone due to the presence of the trans-isomer. In agree with this fact, the signals between 20 and 35 ppm also appear more broadened at the base (Hosler et al. 1999). The signals for disulfide or polysulfide cross-links are also absent. The elemental analysis of this sample also indicated a low sulfur content of 0.31% (Table 4). The elemental analysis of one sample of the soil where the rubber ball was buried also indicated a low sulfur content of 1.45% (Table 4).

These results agree with the previous results obtained by Hosler et al. 1999 except we found no evidence to support the presence of sulfur- or carbon-containing cross-linking groups. These data together with the FTIR spectra indicate that no such as “vulcanization” process was applied in the preparation of the analyzed rubber ball.

From the XRF analysis of the CRB16 to CRB19, the main inorganic elements of the Castilla elastica rubber are Ca, K, S, Si, and Sr (Fig. 7a). This was confirmed by the CEA results of the samples presented in Table 3. These results show that the C. elastica rubber itself has sulfur in its composition with a similar relative abundance among the analyzed balls.

Fig. 7
figure 7

Identified chemical elements and its relative abundance by X-ray fluorescence in archaeological (ORD01 to ORW14) and modern objects (CRB16 to CRB19) and soil (S1 to S3). Relative abundance is presented as normalized X-ray intensities by Fe X-ray intensity for a major elements and b trace elements. The data are presented according to its X-ray intensity similarity for the archaeological rubber balls

Full size image

The relative abundances presented as X-ray intensities normalized to Fe X-ray intensity since one of the main elements present in the archaeological sediment is Fe. Other elements, such as Cr, Co, Zn, Cu, As, and V, were detected as trace elements. Some of the elements traced to the plants by its relation to the soil they grew are mainly Zn and Cu (Bañuelos and Ajwa 1999). In Fig. 7b, the graph represents the results for these trace elements in the archaeological rubber balls (ORD01, ORD02, ORD04, ORD07, ORW09, ORW10, ORW12, and ORW15), the contemporary balls (CRB16 to CRB19), and the archaeological soil at El Manatí (S1, S2, S3).

Data from the X-ray analysis found differences in K, Ca, Cu, and Zn; each rubber ball found at El Manatí may have different origins and they could have been offered at different moments. Another explanation is that latex as raw material must have been extracted at different times from trees of the same region, and only the ORW10 and ORD04 seem to be from the same raw material tree and the same time of latex extraction, with the rest of the rubber balls presenting variations in their trace element content.

It is important to point out that the contemporary balls made in the same year (ORB 16 and ORB 17; ORB 18 and ORB19) have comparable results in the elemental contents of S, K, Ca, Sr, Ni, Cu, and Zn (Fig. 7). These results are consistent since they were made following the same procedure and using raw materials collected from trees from the same region in a same moment without a mixture of raw materials from other regions.

As reported by the specialist in the manufacture of contemporary rubber balls from Veracruz, the latex can be used to form a small handmade ball, on to which different partially coagulated rectangular strips are applied in different directions to make the ball grow. This technique of applying different strips was observed in the radiographic images showing superposition of the strips in different directions (Fig. 8) confirming the description of Spanish friars Sahagún and Torquemada (de Sahagún 1906–1907; de Torquemada 1977).

Fig. 8
figure 8

Radiographic images (negative) a ORW09 and b ORD01 Olmec rubber balls where the superposition of rubber bands can be appreciated c CRB16 and d CRB19. Contemporary rubber balls showing a homogeneous gray scale

Full size image

From the images, it cannot be established if the Olmec rubber balls have an inner nucleus as it reported in the manufacture of the contemporary balls (Fig. 3), since at the center it may be that the rubber bands are still compressed forming a higher density area. The images from the wet Olmec balls did not provide any information since they appear homogeneous due to the liquid media absorbed. In the contemporary balls, the presence of thin strips can only be slightly appreciated in Fig. 8c, but no evidence of the nucleus can be observed since the rubber is highly compacted.

The band manufacturing feature can also be macroscopically observed in some balls (as shown in Fig. 4a), with the microscopic images providing additional evidence, especially for the dried Olmec balls (Fig. 9), it can be observed that the bands can be in one direction or superimposed in several directions.

Fig. 9
figure 9

Micrographs of the rubber balls: a ORD01, d ORD07, and g ORD05 dried Olmec rubber balls; b ORW10, e ORW10, and h ORW09 wet Olmec rubber balls; c CRB17, f CRB17, and i CRB18 contemporary rubber balls

Full size image

The wet Olmec balls micrographs also show the bands manufacture and, in some cases, a probable thin layer of latex covering (Fig. 9b) as seen in the preparation of the Acayucan contemporary balls. Comparing the Olmec and the contemporary rubber balls, as rubber appears white if the layers are thinner, while it is darker if the layers are thicker.

In the contemporary balls, the bands have different directions, while in the Olmec balls, only one direction is regularly observed in their exterior, with a few balls presenting strips’ superposition in different directions. Additionally, there may be a presence of a possible thin layer finish in two wet balls.

Preservation of El Manatí’s Olmec balls

The analyses performed were applied to understand the technique, materials, and degradation of the rubber balls. The UV-induced fluorescence is a highly used technique in conservation science to differentiate organic materials. The obtained images (Fig. 10) showed that the lighter colored rubber areas in the balls have a yellowish or orange fluorescence, while a green fluorescence indicated the presence of microorganisms in the surface, which were not observable under visible light. The C. elastica rubber balls without biodeterioration did not show fluorescence. Still, there is the need to study further the biological species of the organisms and whether they are active or not as shown in micrographs (Fig. S2).

Fig. 10
figure 10

Comparison of the appearance of the balls with visible light and with UV-induced fluorescence: a ORD02 visible light image; b ORD02 UV image where the yellowish-orange fluorescence from lighter areas of the natural rubber; c ORD07 visible light image; d ORD07 UV image with green fluorescence due to microorganism presence

Full size image

The X-ray imaging analysis showed that the inner region is denser in comparison to the borders (Fig. 11). It suggests this area still preserves the ball form and compression on the rubber bands made to obtain this form, while in the exterior area, the bands have separated, leaving spaces between them; therefore, in the analysis, it is determined as of lesser density.

Fig. 11
figure 11

Radiographic image analysis of ORD01: a Histogram data of the gray level in the images. b Intensity ratio of the gray scale for the simulation. c, d Resulting images showing in black and white the density difference simulation

Full size image

The results can be understood as the archaeological burial soil allowed the balls to maintain their form, while providing the necessary weight and pressure on them, but also the finer soil transported by the water in the offering’s site was transferred to the spaces left as the rubber bands started to separate due to aging as observed in the micrographs (Fig. 9). The separation of the rubber bands from the exterior to the interior was accelerated after the excavation as result of the water evaporation process. This physical transformation can be corroborated by comparing with the radiographic images of the contemporary balls in Fig. 8, where the manufacture technique can only be slightly appreciated in one ball, due to the still present compaction of the bands and on the immersed balls on which water compresses the ball and only shows surface fractures (Fig. S1).

The separation of the rubber bands in the dried Olmec balls permits us to define this manufacture technique, but this physical feature allows for the burial soil to enter between the bands (Fig. 9d), and it shows also the material becomes friable by separating each band into small fibers, giving the image of a fibrous material (Fig. 9g). The case of the wet rubber balls is different with small to no separation, and the surface still looks like a polymer material (Fig. 9b and e). This physical condition is attributed to the fact that C. elastica rubber while being in water exudates polyisoprene; therefore, the available spaces are filled with these exudates that then coagulates (Guzmán 2022).

As presented before, difference in FTIR spectra of the ORD and ORW is explained by how these objects have been preserved. The loss of moisture in the dry balls accelerated the chemical degradation of the polymer, while ORW objects were exposed to relative constant humidity and temperature which resulted in a better conservation of the material (Blank 1990). In the dried balls, the degradation bands detected are related to a very slow process of photooxidation which probably started when the archaeological objects were produced and can be explained as the polymer chain scission occurs forming carbonyl groups.

The effect of light on polymers has been highly studied; however, these artifacts have not been exposed to light and the chemical degrading reaction is very slow. The bands at 775 and 730 cm−1 associated with crystallinity features of the polymer show both in the contemporary and archaeological balls to be dependent on the analysis site and not indicative of the formation of crystalline sites due to degradation but in accordance with the NMR results is suggested a transition to the trans-isomer as the material transforms.

The FTIR spectra of the contemporary balls (Table 3 and Fig. S3) show in 4 years of light exposure the diminishing of the 1657 and 1580 cm−1 polyisoprene bands and the appearance of 1742 and 1310 cm−1 bands due to photodegradation.

In general, the relative main inorganic element composition decreased in comparison to the contemporary rubber balls indicating the possible solubilization of the compounds containing these elements in the water from the burial context (Fig. 7). This solubilization process of some of the C. elastica rubber biomarker compounds was also observed in the GC-MS analysis of the samples where only seven compounds (Table 5) were separated in comparison to the 14 compounds reported for C. elastica rubber (Campos 2022). The identified compounds are reported in the composition of C. elastica rubber but are also present in other natural materials therefore cannot be considered as chemical markers for the specie.

Table 5 Identification of compounds by CG-MS in comparison to the library. All compounds have a match greater than or equal to 80%
Full size table

This shows that even the burial conditions preserved the poly-isoprene rubber organic material, there are other soluble minor concentration compounds present which are soluble and which could be used as chemical markers but were washed away. The presence of estrone is attributed to the solubilization of this hormone in the burial soil and water before the offering’s discovery due to human-related activities such as cattle-raising. No other plant markers were identified by this method, due to the running of natural water conditions at the offering’s site if plant extracts were added to the El Manatí balls as the literature reports is probable that they were solubilized and lost, but there is no chemical evidence to confirm or neglect this.

According to these results, there is a minor chemical degradation of the balls while the most important changes occurring towards its physical properties.

Conclusions

In this study, we were able to present a comprehensive analysis of 3000-year-old Olmec rubber balls discovered at El Manatí to unravel its manufacturing technique and transformation processes. These results provide new information to such balls presented as offerings with the perspective of understanding other rubber balls from Mesoamerican cultures.

The data obtained shows that the Olmec people had a well-established knowledge of how to extract the latex from the Castilla elastica tree and the coagulation times of the rubber in order to transform it into a ball. The manufacturing technique of strips as described by the Spanish friars is confirmed through the radiographic and microscopic imaging. Both the manufacture and the composition of the material suggest an established technology in the region considering that the balls were probably offered at this site in different moments as suggested by the XRF analysis. Regarding the composition, no plant additives were identified but it cannot be discarded yet, more research must be done in this sense. On the other hand, no evidence of a vulcanization process was determined on any of the balls through the comprehensive set of analytical techniques.

The information obtained indicates that the Olmec rubber balls preserved in wet state have not undergone chemical degradation, while the dried ones have initiated a very slow chemical degradation process. The radiographic images showed the main physical degradation is the rubber band separation due to drying which also causes the brittleness of the materials. Studies performed in the contemporary balls show that the Olmec rubber balls must stay away from light to prevent them from suffering photooxidation. The role of the preservative solution and biodegrading processes is the next step in research to conserve this rare and special collection.

The combination of imaging and non-destructive techniques applied to organic materials such as the Olmec balls made from Castilla elastica rubber provided new information of a material with characteristics different from the highly studied Hevea brasiliensis. This method can also be employed as diagnostic for their conservation assessment.