INTRODUCTION
Buglossoides arvensis (L) I.M.Johnst., syn. Lithospermum arvense L., (tribe Lithospermeae, family Boraginaceae) is an annual plant 10–50 cm in height with a flowering time between April and July. B. arvensis is commonly found in Eurasian arable lands, grasslands, and forest margins. The fruits, which are often incorrectly considered as seeds of B. arvensis, are small (approximately 2 mm in diameter), ovoid, and contain CaCO3 in their epidermal cells and parts of sclerenchyma (for more information about B. arvensis see Pustovoytov and Riehl [Reference Pustovoytov and Riehl2006] and references therein) (Figure 1).
Figure 1 (Left) An approximately 1-month-old B. arvensis grown in a 250-mL plastic pot; (top right) B. arvensis flower; (bottom right) B. arvensis fruits. The arrows show the openings in the pot lid, which were used for irrigation and root labeling (see Labeling Procedure section).
Fossil fruits of B. arvensis and other members of Lithospermeae are often found in late Pleistocene and Holocene deposits as well as in cultural layers of archaeological sites (Pustovoytov and Riehl Reference Pustovoytov and Riehl2006). This calls for testing the applicability of carbon (C) isotopes in these fruits for dating purposes and paleoenvironmental reconstructions. Previously, it has been demonstrated that fruit carbonate of another taxon, the genus Celtis, can be successfully radiocarbon (14C) dated (Wang et al. Reference Wang, Jahren and Amundson1997; Quade et al. Reference Quade, Shanying, Stiner, Clark and Mentzer2014) and serve as a paleoclimate proxy (Jahren et al. Reference Jahren, Amundson, Kendall and Wigand2001). Similar results have been obtained for the tribe Lithospermeae (Pustovoytov et al. Reference Pustovoytov, Riehl and Mittmann2004; Pustovoytov and Riehl Reference Pustovoytov and Riehl2006; Pustovoytov et al. Reference Pustovoytov, Riehl, Hilger and Schumacher2010). Aside from a few under- or overestimates the achieved ages showed good consistency with independently estimated ages for the archeological layers. The underestimated ages can be explained by post-sedimentary incorporation of fruits into the deposits (i.e. via bioturbation) (Wang et al. Reference Wang, Jahren and Amundson1997; Pustovoytov et al. Reference Pustovoytov, Riehl and Mittmann2004; Pustovoytov et al. Reference Pustovoytov, Riehl, Hilger and Schumacher2010) or slight diagenetic 14C-contamination effects (Quade et al. Reference Quade, Shanying, Stiner, Clark and Mentzer2014). An approximately 400-yr overestimate for a herbarium exemplar from the early 19th century has been attributed to occasional depletion in atmospheric 14C concentration because of fossil fuel combustion (Pustovoytov and Riehl Reference Pustovoytov and Riehl2006).
However, since 1940 it has been known that plants can take up HCO3 – from soil solution via their roots (Overstreet et al. Reference Overstreet, Ruben and Broyer1940; Cramer and Richards Reference Cramer and Richards1999; Cramer et al. Reference Cramer, Gao and Lips1999; Viktor and Cramer Reference Viktor and Cramer2005). It has been shown that the amount of HCO3 – taken up can be 0.8–2% of the C assimilated through photosynthesis (Pelkonen et al. Reference Pelkonen, Vapaavuori and Vuorinen1985; Brix Reference Brix1990; Viktor and Cramer Reference Viktor and Cramer2003; Ford et al. Reference Ford, Wurzburger, Hendrick and Teskey2007). However, the HCO3 – uptake depends on its concentration in soil solution (Cramer and Lips Reference Cramer and Lips1995) and the plant species (Stolwijk and Thimann Reference Stolwijk and Thimann1957). Some species, for example oats, are tolerant of high HCO3 – concentrations in the rhizosphere (up to 6.5% CO2 concentration), but some like tomato may show toxicity symptoms at comparatively low concentrations (approximately 1% CO2) (Stolwijk and Thimann Reference Stolwijk and Thimann1957). HCO3 – uptake via roots is mostly passive and depends on transpiration rates (Stolwijk and Thimann Reference Stolwijk and Thimann1957; Brix Reference Brix1990; Amiro and Ewing Reference Amiro and Ewing1992). This may explain why soil-derived HCO3 – is found at highest concentrations in roots, and decreases with distance from the roots (Brix Reference Brix1990). However, these concentrations can increase locally in specific plant organs such as newly formed stems or fine roots, through unknown active mechanisms (Vuorinen et al. Reference Vuorinen, Vapaavuori and Lapinjoki1989; Ford et al. Reference Ford, Wurzburger, Hendrick and Teskey2007).
The HCO3 – concentration in soil solution is determined by the dissolution of root- and microbe-respired CO2, exchange of CO2 between the soil and atmosphere and dissolution of carbonate containing minerals such as CaCO3. The isotopic composition of C in these HCO3 – sources differs: while HCO3 – from carbonate minerals is often totally 14C depleted, the 14C content of respired CO2 is almost identical with the 14C concentration in modern atmospheric CO2. Therefore, even a few percent of old C from carbonate minerals can modify 14C ages of a sample. We hypothesize that 14C ages based on fruit carbonate could overestimate the true age of a sample if part of the C comes from soil HCO3 –. Therefore, the main aims of this experiment were (1) to identify the origin of C in CaCO3 of fruits, (2) to quantify the contribution of absorbed HCO3 – from soil, and (3) to calculate the potential effect of root HCO3 – uptake on 14C dates based on fruit carbonates of B. arvensis.
MATERIAL AND METHODS
Experimental Layout
We used 250-mL plastic pots with lids (Sartorius AG, Germany) for plant growth (Figure 1, left). The lids had one main hole in the middle, for the growing plant stem, and three smaller openings, which were used for soil labeling and irrigation. To make a carbonate-containing and a carbonate-free medium for plant growth, a carbonate-free loamy soil (Haplic Luvisol, originated from loess) was mixed with loess and sand particles, respectively, at a 1:1 ratio (200 g of loamy soil to 200 g of loess or sand). The loamy soil, loess, and sand particles were air-dried and passed through a 2-mm screen before mixing. Loess samples containing 30% CaCO3 were taken from an open mine at Nussloch, southwest Germany, from 10 m below the soil surface (see Kuzyakov et al. [Reference Kuzyakov, Shevtzova and Pustovoytov2006] for details). Carbonate-free sand in the size range 0.5–1.5 mm was used. Water content was adjusted to 60% of water-holding capacity by adding 96 mL of distilled water to the loamy soil+loess (hereafter called Loess) and 84 mL to the loamy soil+sand (hereafter called Sand). The water content of Loess and Sand was kept at 60% of water holding capacity during the whole experiment by weighing the pots and adding water when needed.
Fruits of B. arvensis were pre-germinated in the dark on wet filter paper. When plant height was around 1 cm they were transplanted into the growth pots, the lids closed, and placed into a growing chamber at 25–27°C with a 14-hr photoperiod and a 180 μmol m–2 s–1 light intensity.
Labeling Procedure
Labeling started one week after the first flowers developed and was repeated five times over a one-month period thereafter. Labeling was applied to either the roots or the shoots. In both cases, 200 kBq of 14C in the form of Na2 14CO3 solution was used at each labeling occasion. The applied 14C activity for labeling was several orders of magnitude higher than natural abundance of 14C in plant organs or soils. Hence, the initial 14C activity of plant organs or soils had no effect on the results of labeling. Before starting the procedure, the space between the stem and the main opening in the lid was filled with cotton and covered with petroleum jelly to provide an air-tight seal, which was maintained for the one-month period. The three small openings were only closed for the few hours of each labeling procedure, using tight-fitting plastic pins. Separation of the root and shoot atmospheres during the labeling procedure was necessary to prevent dissolution of 14CO2 in the soil solution while labeling the shoots, and to avoid photosynthetic assimilation of labeled 14CO2 that might be released from the soil solution during root labeling (Amiro and Ewing Reference Amiro and Ewing1992; Cramer and Richards Reference Cramer and Richards1999).
For shoot labeling, the pots were placed in an air-tight labeling chamber made of Plexiglas (0.5×0.5×0.6 m3), which was fitted with four connections and a fan for circulating 14CO2. To produce 14CO2, 5 mL of 2.5 M Na2 14CO3 was acidified by addition of H3PO4. The 14CO2 was pumped into the chamber using 2 inlets. After 1 hr the chamber was connected via the 2 outlets to a glass bottle with 20 mL of 1 M NaOH to trap unassimilated 14CO2. The trapping period was also 1 hr. Afterwards, the plants were returned to the normal conditions outside the labeling chamber and the plastic pins were removed.
For root labeling, 3 mL of 0.002 M Na2 14CO3 solution was injected deeply into the soil in each pot via the three small openings in the lids (1 mL each) (Figure 1, left). This fairly low concentration of sodium carbonate had no effect on plant growth or fruit production compared to the shoot-labeled plants.
14C Analyses
One week after the 5th labeling, 14C activity was measured in plant organs (shoots, roots, and fruits), bulk soil and soil solution. After collecting the fruits, the plant stems were cut at the base and soils were washed with distilled water to separate the roots and to collect soil solution. To wash the soils, 1000 mL of distilled water was used for Loess and 880 mL for Sand. The bulk soils, shoots, roots and fruits were dried overnight at 40°C to determine dry weights. Afterwards, 14C was measured in a subsample of each material.
14C in fruits was measured separately in carbonate and organic components. The fruits were acidified with H3PO4 and the released CO2 was trapped in 1 M NaOH solution. The alkali solution was mixed with scintillation cocktail (Rotiszint EcoPlus, Carl Roth, Germany) and 14C was measured after decay of chemiluminescence with an Automatic TDCR liquid scintillation counter (HIDEX 300 SL, Turku, Finland). The acidified fruits were washed again with distilled water, dried at 40°C and weighed again to determine the weight lost from carbonates. The weight loss after acidification was taken as the fruit carbonate content. The remaining fruit material (i.e. the organic part) was combusted at 900°C using a biological oxidizer (OX 400) to yield CO2. The produced CO2 was trapped in NaOH and 14C activity was measured as described above.
14C measurement in the bulk soil was similar to that for fruits. For soil acidification, 0.1 g of Loess and 2 g of Sand were used. 14C measurements of shoots and roots were performed in the same way as for the organic parts of fruits, but as finely ground powders. 14C in soil solution was measured after addition of scintillation cocktail. To differentiate between dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC), a part of the solution was acidified before addition of scintillation cocktail. This provided a 14C determination in DOC. The difference between 14C activity of total dissolved carbon and that of DOC was the 14C activity in DIC.
Calculation of Carbon Incorporation into Plant Organs and Age Overestimation
The C amounts incorporated into plant organs (mg) were calculated based on the C content of the labeling solution (mg) added to each pot, total 14C activity applied to each pot, and the 14C activity measured in plant organs (i.e. fruit carbonates, fruit organics, roots, shoots) (see Kuzyakov et al. [Reference Kuzyakov, Shevtzova and Pustovoytov2006] for more details).
To calculate the age overestimation because of incorporated HCO3 – carbon, we used the usual 14C decay equation (Bowman 1995)
$${\rm T{\equals}}{\minus}\!{\rm 8267} \cdot {\rm ln }\left( {{{{\rm A}_{{{\rm SN}}} } \mathord{\left/ {\vphantom {{{\rm A}_{{{\rm SN}}} } {{\rm A}_{{{\rm ON }}} }}} \right. \kern-\nulldelimiterspace} {{\rm A}_{{{\rm ON }}} }}} \right)$$
where ASN is the normalized number of measured 14C atoms in a given sample and AON is the initial normalized number of 14C atoms at the beginning of decay and T is the time elapsed since the beginning of decay. Assuming a constant atmospheric 14C concentration over time,
$${\rm A}_{{{\rm SN}}} {\rm {\equals}A}_{{{\rm ON}}} \cdot e^{{{\minus}\!{\rm \lambda T}}} $$
where λ=1/8267. This law remains true as long as no new fractions of 14C or radiometrically dead C are added to a sample. If a portion P of radimetrically dead carbon is added to a sample, the 14C concentration in such a sample becomes lower by a factor 1/(1+P), which modifies Equation 2 in the following way:
$${\rm A}_{{{\rm SN}}} {\equals}{\rm A}_{{{\rm ON}}} \cdot e^{{{\minus}\!{\rm \lambda T}}} \cdot {1 \over {{\rm 1}{\plus}P}}$$
Combining Equations 1 and 3, we obtain a formula for the measured age T′ of a sample with a portion of radiometrically dead carbon P
$${\rm T}\prime{\equals}{\minus}\!8267 \cdot {\rm ln}\left[ {{{\left( {{\rm A}_{{{\rm ON}}} \cdot e^{{{\minus}\!{\rm \lambda T}}} \cdot {1 \over {1{\plus}P}}} \right)} \mathord{\left/ {\vphantom {{\left( {{\rm A}_{{{\rm ON}}} \cdot e^{{{\minus}{\rm \lambda T}}} \cdot {1 \over {1{\plus}P}}} \right)} {{\rm A}_{{{\rm ON }}} }}} \right. \kern-\nulldelimiterspace} {{\rm A}_{{{\rm ON }}} }} } \right]$$
It is further apparent that
$${\rm T}\prime{\equals}{\minus}\!8267\cdot{\rm ln}\,\left( {{{e^{{{\minus}\!{\rm \lambda T}}} } \over {1{\plus}P}}} \right){\equals}{\rm T}{\plus}8267\cdot{\rm ln}\,\left( {1{\plus}P} \right) $$
Equation 5 can provide the offset between the measured age of a sample with admixtures of dead carbon and its true age ΔT under stable 14C atmospheric concentration
$${\rm \Delta T}{\equals}{\rm T}\prime{\minus}\!{\rm T}{\equals}8267\cdot{\rm ln}\,\left( {1{\plus}P} \right)$$
As it follows from Equation 6, this offset does not depend on time and is only determined by the quantity of dead carbon admixture.
Statistics
Mean values and standard errors were calculated for 6 replicates of each treatment. The significance of differences between shoot- and root-labeled plants was assessed using the post-hoc Fisher LSD test at α=0.05 significance level. Statistical analyses were done in STATISTICA 10 (StatSoft Inc., Tulsa, USA).
RESULTS
The 14C distribution via shoot- and root labeling showed obvious and significant differences (p<5%) between various organs (Table 1). 14C specific activity after shoot labeling was the highest in shoots, followed by fruit organics and roots. In contrast, the highest 14C activity after root labeling was recovered in the roots, followed by DOC and DIC. 14C fraction recovered in shoots was around 6 times higher (43–47%) after shoot labeling than root labeling (7–8%). Recovery after shoot labeling was also about 9 times higher in fruit organics, but around 3 times lower in roots.
Table 1 Percentage of 14C label recovered in different plant organs and soils via photosynthesis (shoot-labeling) or taken up by roots (root-labeling). Standard errors are shown in parentheses.
Total incorporation of C from shoot labeling by Loess-grown plants was 90.6 mg, lower (p<5%) than for the Sand-grown plants (92.1 mg). Incorporated C from root labeling was 74.1 and 103 mg for Loess and Sand, respectively (Table 2). Fruit carbonate had greater incorporation from shoot labeling than from root labeling: 1.5 times higher in Sand and 1.9 times in Loess (Table 2).
Table 2 Amounts of incorporated labeled carbon (mg) in plant organs after shoot or root labeling of Sand- or Loess-grown plants. Standard errors are shown in parentheses.
DISCUSSION
The soil properties (Loess vs. Sand) and the labeling approach (shoot vs. roots) had no effect on total plant growth or individual organs. Therefore, we can directly compare the label incorporation and distribution between the soils and labeling conditions.
The obvious differences in 14C activity of various plant organs after root labeling compared to shoot labeling reveal that HCO3 – carbon was taken up by B. arvensis roots (Table 1). To determine the amount of HCO3 – carbon incorporated by B. arvensis, the total incorporated C via shoot labeling in Loess and Sand were compared. If we assume no re-uptake via HCO3 –, there should be no difference between the incorporated C from 14CO2 in Sand- and Loess-grown plants following shoot-labeling. The comparison, however, reveals 1.51 mg less photo-assimilated C in Loess than in Sand (Table 2). CaCO3 solubility in distilled water is 13.1 mg L–1 at 25°C (Aylward Reference Aylward2007). Therefore, in Loess with approximately 700 mL waterFootnote 2 , 9.1 mg CaCO3 can be dissolved. According to the C mass proportion in CaCO3 (12 mg C 100 mg–1 CaCO3), this amount of dissolved CaCO3 contains 1.1 mg C (fairly equal to the solubility of CaCO3). Root-respired CO2 can dissolve in soil solution and be reabsorbed by roots (Ford et al. Reference Ford, Wurzburger, Hendrick and Teskey2007). However, root-respired CO2 is diluted in Loess solution before re-uptake (Figure 2). Hence, total incorporated C in Loess plants was lower than in Sand plants. In conclusion, the so-called reservoir effect, i.e. incorporation of 14C-depleted carbon from soil into biologically formed carbonates which has already been proven for other types of biogenic carbonates, such as land-snail shells (Pigati et al. Reference Pigati, Quade, Shahanan and Haynes2004; Pigati et al. Reference Pigati, Rech and Nekola2010 and references therein) also takes place in fruit carbonate of B. arvensis.
Figure 2 Dilution of 14C content of plant organs by dissolved inorganic C (HCO3 –) taken from 2 sources. In carbonate-free soils, the only source of HCO3 – is dissolution of root- and rhizomicrobially respired CO2 originally from the atmosphere ( A CO2). In carbonate-containing soils, the dissolution of lithogenic carbonates (Ca L CO3) is a second source. The HCO3 – from root-respired CO2 is diluted by the HCO3 – from lithogenic carbonates (Kuzyakov et al. Reference Kuzyakov, Shevtzova and Pustovoytov2006; Gocke et al. Reference Gocke, Pustovoytov and Kuzyakov2011). For shoot-labeled plants, this process leads to a reduction of the 14C activity in the re-absorbed HCO3 –.
Since the incorporated C from soil carbonate is 14C dead, this may lead to overestimations of 14C ages based on biogenic carbonates (Goodfriend Reference Goodfriend1987). Considering the total weight of C in fruit carbonates (8.08 mg C, based on 20 fruits) and the difference between HCO3 – incorporation into fruit carbonate in Sand and Loess after shoot labeling (0.012 mg C) approximately 0.15% of C in fruit carbonate—after 10-hr labeling—originated from soil solution. The total HCO3 – incorporated into the whole plant amounted to 1.6% of dry weight. However, fruit carbonate in Loess after shoot labeling showed 7.6% more HCO3 – than in Sand (Table 2). Furthermore, extrapolating the 10-hr labeling period to the full month of this study (420-hr photoperiod) indicates around 6.3% of fruit carbonate in Loess is derived from lithogenic carbonates. A 6.3% share of lithogenic HCO3 – leads to 14C ages overestimated by 505 14C yr (Equation 6), based on fruit carbonate of Loess-grown Buglossoides arvensis.
In this connection, it is important to note that too-old 14C ages on fruit carbonate were reported in literature (Pustovoytov et al. Reference Pustovoytov, Riehl and Mittmann2004, Reference Pustovoytov, Riehl, Hilger and Schumacher2010; Pustovoytov and Riehl Reference Pustovoytov and Riehl2006). One of the ways to explain the discrepancy between an age measured on the carbonate fraction of fruits and the true age of the sample could be the uptake of inorganic carbon from the soil by root systems. Regarding the suitability of fruit carbonate for dating purposes, an age overestimation of order of 500 14C yr, though persistent with increasing sample age, becomes insignificant against the measurement uncertainties in relatively old samples (such as 11,000 yr and older, i.e. after 2 14C half-lives).
Some of the other findings may also deserve particular attention. As expected, the distribution of C from soil CaCO3 decreases with the distance of plant organs from the roots (Brix Reference Brix1990) (Table 2). The HCO3 – distribution in plant organs has usually been attributed to passive uptake with transpiration flow (Stolwijk and Thimann Reference Stolwijk and Thimann1957; Amiro and Ewing Reference Amiro and Ewing1992). This means that HCO3 – moves with water from roots towards stomata (Amiro and Ewing Reference Amiro and Ewing1992). However, the different 14C activities in various organs following shoot labeling in Sand and Loess, arising from the dilution effect of lithogenic HCO3 –, suggest the selective incorporation of HCO3 – carbon in specific organs (Ford et al. Reference Ford, Wurzburger, Hendrick and Teskey2007). After shoot labeling, there was 3.20 mg more labeled C in roots and shoots and 0.012 mg more labeled C in fruit carbonate of plants grown in Sand than of those in Loess (Table 2). At the same time, Sand-grown plants had 5.11 mg less labeled C in fruit organics. The higher difference indicates a higher dilution effect by soil carbonate and higher incorporation of HCO3 –. Therefore, the highest HCO3 – amount was retained in roots and shoots, followed by fruit carbonate, while fruit organics showed the lowest HCO3 – incorporation. This may suggest some active uptake processes (Vuorinen et al. Reference Vuorinen, Vapaavuori and Lapinjoki1989; Ford et al. Reference Ford, Wurzburger, Hendrick and Teskey2007) enhancing fruit carbonate compared to the fruit organics, since these components are the same distance from the roots. The apparent lower HCO3 – incorporated in Loess- compared to Sand-grown plants after root labeling, on the other hand, is partly due to substitution of added Na2CO3–C with Loess CaCO3–C (Figure 1) (Kuzyakov et al. Reference Kuzyakov, Shevtzova and Pustovoytov2006).
CONCLUSIONS
1. Buglossoides arvensis takes up dissolved inorganic carbon (HCO3 –) from the soil via roots under laboratory conditions. The source of HCO3 – can be dissolution of carbonate minerals (radiometrically dead, e.g. loess carbonate) and dissolution of root-respired CO2 (recent C) in soil solution;
2. The HCO3 – uptake is mostly passive; however, HCO3 – can be preferentially incorporated into organs such as fruit carbonate, which are formed at specific plant development stages;
3. The incorporated HCO3 – taken up by roots may contribute more than 6.0% of fruit-carbonate C in plants growing on a carbonate-containing soil. Therefore, an age overestimation of approximately 500 yr is possible. Inflated ages based on fruit carbonate can be attributed to HCO3 – uptake by roots during fruit development. This calls for further investigation of possible effects of calcareous substrates on the outcome of 14C-dating of the fruit carbonate fraction;
4. The age overestimation because of lithogenic HCO3 – incorporation in fruit carbonate, however, is insignificant in relatively old samples, approximately after 2 14C half-lives.
ACKNOWLEDGMENTS
We appreciate the German Research Foundation (DFG) for their support (KU 1184/34-1). We would like to thank Heike Strutz and Susann Enzmann for their help during labeling. Special thanks to Bernd Kopka and the staff at Labor für Radioisotope (LARI), University of Göttingen, who facilitated running the experiment and measuring 14C in plant samples. We thank the seed collection of the botanical garden of the University of Göttingen for Buglossoides arvensis fruits. Our sincere gratitude goes to Jeff Pigati, who provided many valuable suggestions on the first version of the manuscript. The authors are also grateful to the two further anonymous reviewers for their helpful comments.
