The increased size of the cotyledons at 300 °C is evident in Fig. 3, Fig. 4. There is little change in the L/W ratio as noted by Kaplan, 1956, Wagner, 1987; for pinto there is a moderate positive monotonic correlation for temperature and L/W ratio (ρ = 0.57881, p = 0.00000) and for kidney there is a negligible, nonsignificant monotonic correlation (ρ = 0.0005, p = 0.99582). The seeds lost mass throughout the sequence, with median decreases from ~ 20% at 220 °C to ~ 80% at 400 °C; the largest percentage decrease occurred between seeds heated at 220 °C and 240 °C. Temperature and mass loss percent have a very strong negative monotonic correlation for both varieties (pinto: ρ = −0.99287, p = 0.00000; kidney: ρ = −0.99147, p = 0.00000).
Percent change in mass and the force needed to break whole seeds have a moderate, positive monotonic correlation for pinto (ρ = 0.64279, p = 0.00000) and a weak, positive monotonic correlation for kidney (ρ = 0.31845, p = 0.00124). The cotyledons became brittle with heating, breaking with audible cracks. Cotyledon compression strength decreased throughout the sequence. Percent change in mass and the force needed to break cotyledons have strong positive monotonic correlations (pinto: ρ = 0.85621, p = 0.00000; kidney: ρ = 0.76004, p = 0.00000). Given that cotyledon lengths extended beyond the area of the indenter, it is possible that length influenced breakage force. However, partial correlation of length and breakage force controlling for temperature resulted in negligible, non-significant correlations (pinto: r = −0.07197, p = 0.45707; kidney: r = 0.06309, p = 0.51458) indicating that the length of the cotyledon had little if any effect on breakage force.
3.2. Rehydrated seeds
The rehydrated seeds were visually different from the dried seeds after heating at all temperatures (Fig. 7). Of particular note is that the testae survived largely intact at all temperatures up to 320 °C with glossy exterior surfaces. On seeds heated at 380 °C fragments of the testa remained attached but were mostly absent on the seeds heated at 400 °C. Warping of the cotyledons occurred at all temperatures. No surficial fissures occurred at higher temperatures as in the dried seeds. Partial exterior surface consumption of one or both cotyledons occurred on seeds heated at temperatures ≥ 280 °C. The plumule, epicotyl, and hypocotyl preserved in whole or part in some cotyledons at all temperatures. Partial splitting of the seed into cotyledons was common at all temperatures, with warping of the cotyledons often occurring at the anterior end of the seed, where the cotyledons separated and spread outward. Seeds lacking such splitting and warping often exhibited burned-out hila areas where gasses vented. Contrasting with the dried seed cotyledons, most of the rehydrated seed cotyledon interiors were burned away at higher temperatures. This often resulted in shell-like appearances with rippled or wavy interior surfaces. Large voids occurred within the cotyledons at all temperatures as evident in cross-section (Fig. 8). Cotyledon cross-section textures appear similar throughout the temperature range except for those of seeds heated at 400 °C, which were coarser. The cotyledons were at least partially fused to one another in all kidney seeds heated at 380 °C and half of those at 400 °C. Cotyledon fusing was less apparent in pinto bean seeds.
Fig. 7. Example rehydrated pinto and kidney bean seed cotyledon exteriors (left) and interiors (right) after charring at indicated temperatures (°C). Each pair is the same cotyledon, the interior surface image has been rotated and flipped so that the anterior end is at the top and hilum region is to the right. Brightness and contrast have been adjusted to make features visible. Arrows point to embryos.
As would be expected because of higher water content, the loss of mass was greater in the rehydrated than in the dried seeds (Fig. 9). There are very strong monotonic negative correlations between temperature and loss of mass percentages (pinto ρ = −0.98264, p = 0.00000; kidney: ρ = −0.98615, p = 0.00000). Mann-Whitney U tests indicate no significant difference between loss of mass percentage for dried seeds heated at 400 °C and rehydrated seeds heated at 280 °C (pinto: exact permutation p = 0.1333; kidney: exact permutation p = 0.20996). There are very strong negative, monotonic correlations between temperature and change in length (pinto ρ = −0.91915, p = 0.00000; kidney: ρ = −0.90611, p = 0.00000) and strong, negative monotonic correlation for change in width for pinto (ρ = −0.82648, p = 0.00000) and moderate, negative monotonic correlation for kidney (ρ = −0.53683, p = 0.00000). There is a negligible, negative, non-significant monotonic correlations between temperature and L/W ratio for pinto (ρ = −0.17742, p = 0.17885) and moderate, negative, significant monotonic correlation for kidney (ρ = −0. 61504, p = 0.00000).
Percent change in mass and the force needed to break whole seeds have a strong, positive monotonic correlation for pinto (ρ = 0.76059, p = 0.00000) and a moderate, positive monotonic correlation for kidney (ρ = 0.64442, p = 0.00124). Like the dried seed cotyledons, the rehydrated seed cotyledons became brittle after heating, breaking with audible cracks. However, unlike the dried seeds, breakage was irregular resulting in part from the large voids in the cotyledons. It was more common for cotyledons to break into multiple pieces at all temperatures than dried seed cotyledons which typically broke into two pieces. There is a strong positive monotonic correlation between loss of mass and the force needed to break cotyledons for pinto (ρ = 0.76059, p = 0.00000) and moderate positive monotonic correlation for kidney (ρ = 0.61063, p = 0.00000). The force needed to break rehydrated seed cotyledons was substantially less than that for the dried seed cotyledons. A Mann-Whitney U test of the force needed to break cotyledons indicates no significant difference between rehydrated pinto cotyledons heated at 240 °C and dried seed cotyledons heated at 400 °C (exact permutation p = 0.1333) and no significant difference between rehydrated kidney cotyledons heated at 280 °C and dried seed cotyledons heated at 400 °C (exact permutation p = 0.18057). Partial correlation of length and force needed to break cotyledons controlling for temperature was negligible and non-significant for pinto: (r = −0.01729, p = 0.89656) and weak for kidney (r = 0.29696, p = 0.02237) indicating cotyledon length has no to little effect on the force needed to break cotyledons.
4. Discussion
Given that the identification of charred common bean cotyledons from archaeological sites is based primarily on overall shape, the charred cotyledons at all temperatures resulting from these experiments would be identified as common bean. Distortion is evident in changing length and width, but the oval to renal shape of the cotyledons was evident throughout, and there is little change in L/W ratios for both varieties. However, it is apparent that the cotyledons charred at high temperatures would be unlikely to have survived intact in the archaeological record if at all. The fissures and partial consumption of the interiors along with the loss of mass in dried beans resulted in substantial loss of compression strength with median N at 400 °C being ~ 11% of that needed to break a pinto or kidney cotyledon heated at 220 °C. The decrease in compression strength is particularly evident at all temperatures > 280 °C for pinto and > 300 °C for kidney. The rehydrated bean cotyledons exhibited less resistance to breakage than did dried bean cotyledons. While not exhibiting the external surficial fissures, large voids occurred within the cotyledons, and much of the cotyledon interiors were consumed, weakening their structures.
From their earliest occurrences, bean cotyledons in the Northeast are frequently recovered from the same contexts as charred maize kernels (e.g., Heckenberger et al., 1992, Ounjain, 1998, Ritchie and Funk, 1973, Wagner, 1987). Multiple experiments have demonstrated that maize kernels heated at temperatures ≥ 250–260 °C extrude their endosperms or become unrecognizable black masses and would be unlikely to survive in the archaeological record. At temperatures ≤ 200 °C maize kernels and bean seeds only partially char and so are unlikely to survive in the archaeological record. As a result, it is likely that common bean seed and maize kernels found in the same archaeological contexts were heated in reducing environments > 200 °C and < 250 °C. In this temperature range bean cotyledons do not lose substantial compression strength, do not exhibit fissures on their external surfaces, and exhibit relatively minor changes in dimensions. Mean change in length in this range is −5.7% for pinto and −7.4% for kidney and mean change in width is −9.4% for pinto and −7.8% for kidney. Most of the embryo, absent the radicle, preserves in this temperature range, although this is true for cotyledons of some seeds heated at temperatures to 400 °C. Identification of common bean in macrobotanical assemblages is most confident for whole or mostly whole seeds and complete or mostly complete cotyledons. It is certainly possible that fragments of common bean seeds charred at high temperatures survive in the archaeological record, but identifying them as such is difficult (e.g., VanDerwarker and Wilson, 2016:95).
AMS dates on bean cotyledons indicate that preservation in the Northeast archaeological record occurs primarily after ca. cal. CE 1250. By this time and somewhat after depending on region, settlement patterns generally include large, nucleated villages (Birch, 2015, Hart and Means, 2002) and subsistence systems generally focus on agricultural production (Smith 2006). These two factors may account for the appearance of bean macrobotanical remains—larger human populations occupying specific locations for extended periods of time resulted in more opportunities for crop remains to be lost, charred, and preserved in the archaeological record (Hart, 2008). It is perhaps telling that bean macrobotanical remains are not found at earlier nucleated villages in the central Mississippi River and Illinois River valleys (e.g., Simon and Parker, 2006, VanDerwarker and Wilson, 2016). If common bean was used in earlier contexts it should be archaeologically visible if dried bean seeds were charred in contexts that resulted in the preservation of maize kernels.
At temperatures ≥ 260 °C dried beans lose substantial mass, exhibit fissures on external surfaces, and burning in the center of the interior side. These features result in substantial loss of cotyledon compression strength, suggesting that it is unlikely the cotyledons would survive intact in the archaeological record. Similarly, the rapid loss of mass in rehydrated beans throughout the tested temperature range resulting from consumption of large portions and large voids within the remaining cotyledon, make it unlikely that they would survive complete the archaeological record if at all. These results substantiate claims that processing of common beans may account for its relatively scarcity in the archaeological record. If lost and exposed to high temperatures and charred in reducing environments, rehydrated common bean seeds seem unlikely to preserve in the archaeological record.
5. Conclusions
Common bean is the last of the three primary crops of late prehistoric and historical Native American polycultural agricultural systems to be adopted in the Northeast, many centuries after maize and squash. Archaeological visibility of macrobotanical remains in the form of charred seeds and their cotyledons occurs around CE 1250. While regionally visible, the occurrence of bean macrobotanical remains on any given site is generally scarce, especially when compared to maize. At the same time, the experimental results related here indicate that when charred at the same temperatures resulting in maize kernel preservation, common bean cotyledons are likely to survive in the archaeological record. As a result, the hypothesis that charring does not preserve common bean seeds well can be rejected, at least in the lower temperature range. The hypothesis that the preparation of common bean seeds for consumption is detrimental to preservation in the archaeological record cannot be rejected—high water content prior to charring appears to be detrimental to common bean seed preservation. That dried common beans were often pounded into pieces before being added to food preparations (Parker, 1910, Waugh, 1916) would also limit preservation and identification.
The experimental results related here provide an initial attempt to understand the changes that occur to common bean seeds when heated at high temperatures. Unanswered is why when recovered from the archaeological record do macrobotanical remains of common bean occur more often as individual cotyledons rather than whole seeds.