Oxalate content of vegetation and termite frass in western South Africa

Abstract

Oxalate salts in organic matter are potential substrates for the oxalate‐carbonate pathway, which can sequester carbon in drylands. We compared calcium oxalate (CaOx) and water‐soluble oxalate (SOx) concentrations of samples of vegetation and termite excrement (frass) collected from termite mounds in sites across a regional rainfall gradient in western South Africa. We developed mid‐infrared (MIR) spectroscopic models to quantify oxalate components in vegetation extracts (n = 30) and frass samples (n = 39). The MIR spectroscopic method was more efficient than chemical analytical techniques of oxalate measurement. The median CaOx concentration of plants (0.311 mmol g⁻¹) was four times greater than frass (0.081 mmol g⁻¹), which may be explained by degradation of oxalates by microorganisms or selective harvesting of low‐oxalate vegetation by termites. The mean CaOx content of frass from sites in mesic regions (0.042 mmol g⁻¹) was lower relative to frass from sites in more arid regions (0.156 mmol g⁻¹), and lower in termite mounds (0.048 mmol g⁻¹) compared with off‐mound samples (0.131 mmol g⁻¹). Frass collected from sites with higher rainfall had a lower mean SOx content (0.006 mmol g⁻¹, respectively) compared with frass from sites with lower rainfall (0.013 mmol g⁻¹, respectively). This may be attributed to faster degradation of CaOx in soils with greater moisture content. Estimated annual inputs of carbon (17.6 kg mound⁻¹) and calcium (2.55 kg mound⁻¹, 20% of which occurs as CaOx) due to termite frass deposition may be instrumental in the formation of calcite via the oxalate‐carbonate pathway in soils of earthen mounds occupied by termites. This work is relevant to modeling carbon storage in drylands where termites are significant consumers of vegetation.

INTRODUCTION

Oxalate salts are one of the most abundant biominerals on Earth and are synthesized by nearly all plants (Cheng et al., 2016; Franceschi & Nakata, 2005). These salts perform various structural and biological functions in plants, including providing resistance to drought and herbivory (Molano-Flores, 2001; Tooulakou et al., 2016). Oxalates enter soils via litterfall, plant root exudates, and biosynthesis by soil-dwelling microorganisms and play an important role in carbon and calcium cycling (Cowan et al., 2024; Dauer & Perakis, 2014).

The Greater Cape Floristic Region of South Africa is home to endemic vegetation of the Fynbos and Succulent Karoo biomes, which support unusually high plant biomass and biodiversity despite the dry conditions. Plants of the Fynbos and Succulent Karoo biomes of the Greater Cape Floristic Region are known to contain calcium oxalate (CaOx) (Vermonti et al., 2025). Dominant features in these landscapes are large diameter (~30 m), low-relief (1–2 m) dome-shaped mounds (known in South Africa by the Afrikaans term, heuweltjies) occupied by the termite Microhodotermes viator, which can cover more than a quarter of the total land surface (McAuliffe, 2023). Heuweltjies represent a potentially significant store of carbon due to the accumulation of organic material within the mounds via litterfall, and the collection of plant food materials and deposition of excrement (frass) by the termites. In addition to this organic carbon, the mounds frequently contain inorganic carbon (calcium carbonate, CaCO₃) in large accumulations of pedogenic calcrete (Clarke et al., 2023; Francis et al., 2024; Francis & Poch, 2019). With long-term harvesting and transport of plant material into the mounds by termites, oxalates are continuously introduced into heuweltjie soils (Francis et al., 2013; Francis & Poch, 2019). This may create a favorable environment for oxalotrophic bacteria and fungi, which metabolize CaOx in organic matter, leading to the precipitation of calcium carbonate (CaCO₃) via the oxalate-carbonate pathway (OCP) (Cailleau et al., 2011; Cowan et al., 2024). This inorganic carbon in the form of CaCO₃ is more stable in soils relative to the organic carbon fraction and accounts for the largest carbon pool in drylands (Lal et al., 2021).

The interactions between climate, vegetation, and soil nutrients may exert complex controls on levels of oxalate salt inputs as well as rates of degradation. Heuweltjies occur principally in winter rainfall regions of the Western and Northern Cape of South Africa. Rainfall influences vegetation structure and composition, as well as distributions of nutrients of heuweltjie soils (Booi, 2011; Kunz et al., 2012; McAuliffe, 2022; McAuliffe, Timm Hoffman, McFadden, Bell, et al., 2019). The Fynbos biome hosts a high diversity of fine-leafed shrubs and receives greater amounts of precipitation than the Succulent Karoo (Cowling et al., 2015). Rainfall events in the Succulent Karoo biome, which includes arid and semiarid climate regimes, are sporadic and vegetation cover (predominantly dwarf succulent shrubs) is sparse (Luther-Mosebach et al., 2012). However, soils of the Succulent Karoo biome tend to have higher nutrient content than soils of the Fynbos biome (leached, sandy soils derived from Table Mountain sandstones) (Carr et al., 2022; Esler et al., 2015).

Studies of oxalate dynamics in agricultural soils are scarce, as quantification of oxalates by chemical techniques is limited by the requirement for expensive equipment and time-consuming sample preparation requirements (Misiewicz et al., 2023). Attenuated total reflectance (ATR) mid-infrared (MIR) spectroscopic analysis offers an efficient alternative technique to quantify analyte concentrations in samples indirectly (from spectral measurements), based on models calibrated with chemical reference data (Nel et al., 2023). Biominerals such as oxalate salts may be detected in the MIR wavelength region (Kachkoul et al., 2020).

The range of oxalate salt concentrations in vegetation of arid to semiarid zones of the Greater Cape Floristic Region is unknown. Oxalate salt input rates resulting from the transport of vegetation and the deposition of frass by M. viator in heuweltjie soils and resulting accumulation of oxalates in these soils have not been investigated. Thus, the oxalate content of vegetation and termite frass samples of various degrees of weathering needs to be assessed across different climatic regions to help quantify the cycling of Ca and carbon in these ecosystems. This study aimed to determine the CaOx and water-soluble oxalate (SOx) content in vegetation and termite frass collected from termite-affected (heuweltjie, i.e., on-mound) and surrounding (off-mound) soils across different biomes and to compare CaOx and SOx levels measured by a rapid ATR-MIR spectroscopic technique in frass samples of different degrees of weathering.

METHODS

Frass and plant tissues sample collection and preparation

Frass produced by the termite M. viator and plant tissues were sampled from several locations in the Succulent Karoo and Fynbos biomes of South Africa (map of locations shown in Figure 1 and mean annual rainfall of each site shown in Table 1). Sections of plant leaves and stems (30 samples from separate plants), as well as frass samples (n = 39), were collected in late spring to early summer (October to December of 2020) and late summer (February to March of 2021) from the arid Succulent Karoo and semiarid Fynbos biomes in the Western Province of South Africa. An assessment of each heuweltjie, in terms of percentage cover and a count of plant species in quadrats, was conducted previously (Vermonti et al., 2025). The two most dominant species (highest percentage cover in each of three heuweltjie zones; i.e., center and off-mound) were selected for sampling such that four plant species were representative of each mound. Observations indicated preferentially collected fresh plant material, and plant species selected by termites for consumption were identified by observing materials collected by the termites and confirmed by stable carbon isotope analyses of the plant material and termite bodies (Vermonti et al., 2025). Not all plant samples collected for this study were targeted by termites for consumption. Prior to analysis, plant tissue samples were dried at 60°C for 12 h to prevent degradation of organic carbon, then milled to powder. The dried plant tissue samples were stored at room temperature in sealed Falcon tubes or Ziplock bags until analysis in 2022.

[IMAGE OMITTED. SEE PDF]

TABLE 1 Sampling locations and mean annual precipitation (MAP) of each region.

Sampling location	Sample regions	Latitude	Longitude	Vegetation type	MAP (mm)
Stellenbosch	Frass	−33.9433	18.8765	Fynbos	787a
Piketberg	Vegetation and frass	−32.5821	18.7738	Fynbos	542b
Koringberg	Frass	−33.0075	18.6413	Fynbos	437c
Springbok
Buffelsrivier	Vegetation and frass	−29.7631	17.6392	Succulent Karoo	167d
Komaggas	Vegetation and frass	−29.9121	17.4563	Succulent Karoo	156e

At all locations, surface frass (~50 g sample⁻¹) was collected from the center of heuweltjies (on-mound) as well as off-mound (surrounding soils). Frass pellets were readily distinguished from soil particles by their distinct oval shape and large, uniform deposits. Frass samples categorized as either fresh or weathered were also collected near Koringberg, with fresh deposits identified by markedly darker appearance and closer proximity to active ejection ports on the mound surface (Figure 2). Weathered frass samples were considered older and more decomposed. All frass samples were dried at 60°C overnight. Micrograph images of fresh and weathered frass samples were collected using a ZEISS Merlin high-resolution field emission scanning electron microscope (SEM) after pretreatment of loose particles with a thin layer of gold sputter-coating. Frass samples collected from all locations were milled to powder prior to chemical analyses. The total C and N content of the frass samples was determined with an Elementar Vario Macro Cube CN, Elementar GmbH, Germany. The steps of oxalate quantification in plant tissues and frass samples by liquid chromatography-mass spectroscopy and from MIR spectroscopic models are presented in Appendix S1: Oxalate salt quantification in vegetation and frass and Appendix S1: ATR-FTIR spectroscopic technique. This procedure is summarized in flow-chart form (Appendix S1: Figure S1). Preliminary analyses revealed that the CaOx concentrations in soils were below the detection limits, and soil oxalate concentrations were not further investigated.

[IMAGE OMITTED. SEE PDF]

Statistical analyses

Statistical analyses were conducted using RStudio 1.2.5033 software (RStudio Team, 2019). Independent two-sample t tests or Wilcoxon rank-sum tests (chosen according to normality of data distributions) were done to ascertain whether CaOx or SOx content in frass differed between sampling zones (i.e., on-mound versus off-mound) or sites located in different biomes. Statistical comparative tests (either one-way ANOVA or Kruskal–Wallis, depending on the normality of data distributions) were performed to determine whether CaOx and SOx content of frass samples differed between sites with different mean annual rainfall. The CaOx content of fresh and weathered frass was compared using these tests. Post hoc pairwise tests (Dunn tests with Bonferroni adjustments or Tukey Honest Significant Difference, depending on the normality of data distributions) were performed to identify differences in mean CaOx or SOx content of frass samples between sites with different annual rainfall.

RESULTS

MIR spectroscopic models

Absorption peaks of water and 2 M HCl extracts of a plant tissue sample are characteristic of oxalic acids and other organic functional groups such as those representing aromatic and nitrogenous compounds (Clayden et al., 2012). The MIR spectrum of the dry frass sample shows clear peaks (778, and 913 cm⁻¹) corresponding to carboxyl functional groups due to the presence of oxalates (Nel, 2025) in addition to peaks (796, 1010, 1030, 2850, and 1580 cm⁻¹) corresponding to the presence of amines and alkanes (Clayden et al., 2012) (Appendix S1: Figure S2). The MIR spectrum of the dry plant tissue sample shows clear peaks (512, 647, 778, 1315, and 1600 cm⁻¹) corresponding to carboxyl functional groups due to the presence of oxalates (Nel et al., 2024), in addition to functional groups representing aromatic compounds (Clayden et al., 2012) (Appendix S1: Figure S2).

The partial least squares regression algorithm only selected wavelength regions 1064–7497 cm⁻¹ for modeling oxalate components in vegetation (Table 2). Most wavelength regions selected for the calibration of models to predict oxalate components in frass fell within the 1060–6800 cm⁻¹ region (Table 2). A moderate to strong correlation (R² ≥ 0.65) exists between values predicted by spectral components and the true values of the measured property for all models calibrated by plant-based samples (Table 2). The ratio of performance to deviation (RPD) values of models calibrated by plant-based samples was larger than 1.6 (Table 2). Most ratio of performance to inter-quartile distance values were higher than corresponding RPD values.

TABLE 2 Figures of merit for models predicting calcium oxalate and water-soluble oxalate content (CaOx, SOx) in solid frass and vegetation samples and oxalic acid content in liquid extracts based on mid-infrared (MIR) spectra of samples.

Matrix	R2	RMSEP	RMSEPnorm	RPD	RPIQ	Rank	Wavelength regions (cm−1)	Preprocessinga	nb
CaOx (mmol g−1)
Plant tissue	0.69	0.081	0.16	1.82	2.87	5	4638.8–3922; 3209.3–2492.4; 1779.7–1062.9	1st der + SLS	30
Frass	0.93	0.017	0.08	4.85	6.85	6	4638.8–3207.2; 1779.7–1062.9	1st der + SNV	38
SOxc (mmol g−1)
Plant tissue	0.83	0.149	0.08	2.46	3.23	2	5353.6–3922; 1779.7–1062.9	1st der + SNV	30
Frass	0.71	0.004	0.13	1.85	2.31	2	6783.1–6066.3; 1779.7–1062.9	1st der + SNV	38
Oxalic acid (M)
2 M HCl Plant extract	0.87	0.004	0.07	2.77	3.39	4	4638.8–3922; 1779.7–1062.9	1st der + SNV	30
H2O Plant extract	0.95	0.002	0.04	2.07	6.09	1	4638.8–3922; 1779.7–1062.9	SLS	30
2 M HCl Frass extract	0.71	0.002	0.16	1.97	2.93	2	2494.5–1777.7; 1064.9–348.1	None	39
H2O Frass extract	0.73	<0.001	0.12	2.48	2.48	3	1779.7–1062.9; 1779.7–1062.9	Nminmax	39

The accuracy of IR predictive models was evaluated by calculating the normalized RMSP (RMSEP_norm) by expressing RMSEP of each model as a fraction of the component range (Bellon-Maurel et al., 2010; Nel et al., 2023). The performance results of the MIR spectral-based models for prediction of oxalates in organic materials are presented in Appendix S1: Figures S3 and S4. Models for predictions of SOx concentrations in plants have RMSEP_norm < 0.1, but CaOx concentration predictive models have RMSEP_norm < 0.2 for analysis of dry plant tissues (Appendix S1: Figure S3). Models predicting total oxalate anion concentrations (2 M HCl plant extracts) as well as models for predicting oxalate anions in water-based plant extracts have RMSEP_norm values <0.1 (Table 2; Appendix S1: Figure S3).

Models for predictions of CaOx concentrations in dry frass samples achieved good accuracy (RMSEP_norm < 0.1) (Table 2). Models did not predict SOx concentrations in dry frass samples with a quantitative level of accuracy (RMSEP_norm < 0.2 but not <0.1; Table 2; Appendix S1: Figure S4). Models predicting oxalic acid concentrations in 2 M HCl extracts as well as models for predicting oxalic acid concentrations in water extracts of frass samples had poor model fit (R² < 0.25) and RMSEP_norm values <0.2 but not <0.1 (semiquantitative only; Table 2; Appendix S1: Figure S4).

Characterization and oxalate salt concentrations of plant tissues and frass

The UV and total ion chromatogram results for an oxalic acid standard and a plant extract (following oxalate quantification by liquid chromatography-mass spectroscopy as described in Appendix S1: Oxalate salt quantification in vegetation and frass) are shown in Figure 3. The m/z value (163.051) of the peak representing 2,3-dihydroxyquinoxaline (derivatized oxalic acid) and its chromatographic retention time (3.55 min) closely matched existing chemical data (National Center for Biotechnology Information, 2022). The CaOx and SOx concentrations in dry plant tissues (sections of stems with leaves) varied greatly (0.0–0.9 and 0.0–1.9 mmol g⁻¹, for CaOx and Sox, respectively), with a median of 0.3 and 0.2 mmol g⁻¹ for CaOx and SOx, respectively (Table 3).

[IMAGE OMITTED. SEE PDF]

TABLE 3 Calcium oxalate (CaOx) and water-soluble oxalate content (SOx, includes NaOx and oxalic acid) of different plant samples (stems with leaves) collected from various sites in the Greater Cape Floristic Region of South Africa.

Species and mounda	CaOx (μmol g−1)	SOx (μmol g−1)	Targetedb
Succulent Karoo
Buffelsrivier
Mesembryanthemum hypertrophicum
BH4	81	544	No
BH2	170	804	No
BH2	585	360	No
Tetragonia microptera
BH3	136	621	Yes
BH1	154	533	Yes
BH2	324	545	Yes
BH4	406	290	Yes
BH1	505	421	Yes
BH1	574	577	Yes
Komaggas
Cheiridopsis denticulata
KH2	481	80	Yes
KH7	593	92	Yes
KH3	645	137	Yes
KH3	676	80	Yes
KH5	905	77	Yes
KH2	938	111	Yes
Jordaaniella cuprea
KH7	240	65	Yes
KH3	244	81	Yes
KH5	286	38	Yes
KH2	299	49	Yes
KH3	407	96	Yes
Mesembryanthemum barklyi
KH5	0.0	1898	No
KH3	65	1219	No
KH2	96	1166	No
Ruschia sp.
KH2	385	359	Yes
KH5	550	528	Yes
Tetragonia microptera
KH3	105	964	Yes
Fynbos
Piketberg
Chlorophytum triflorum
PH4	110	26	No
PH5	155	126	No
Metalasia muricata
PH4	38	49	Yes
Ruschia suaveolens
PH1	424	77	No

The micrographs of frass particles (Figure 4) show that the material was fibrous and associated with plant detritus of a woody appearance which is deposited along with the frass. The C/N ratios of fresh and weathered frass samples collected from heuweltjies at the Koringberg site were 38.0 ± 6.93 (mean ± SD) and 32.3 ± 5.73, respectively, but the difference between these values was not statistically significant (Welch two-sample t test, p = 0.3392). The CaOx and SOx concentrations in frass samples varied considerably (0.008–0.211 and 0–0.026 mmol g⁻¹, for CaOx and SOx, respectively; Table 4). The median SOx content of frass was an order of magnitude lower than the median CaOx content of frass, which was in turn an order of magnitude lower than typical oxalate concentrations in plant tissues (Tables 3 and 4). The ratio of CaOx to SOx in frass was greater than in plants (Tables 3 and 4).

[IMAGE OMITTED. SEE PDF]

TABLE 4 Summary of distribution of calcium oxalate (CaOx, assumed monohydrate form) and water-soluble oxalate (SOx) concentrations in frass samples as determined by liquid chromatography-mass spectrometry.

Distribution metric	CaOx (μmol g−1)	SOx (μmol g−1)
Min	8.4	0.0
Q1a	20	4.0
Q2a	61	6.6
Mean	81	8.0
Q3a	136	12
Max	212	26

The median CaOx content of frass (0.081 mmol g⁻¹) was an order of magnitude greater than the median SOx content of frass (0.008 mmol g⁻¹; Table 4). The mean CaOx content of fresh frass (0.146 mmol g⁻¹, SD = 0.013) was significantly higher (t_4,2.5% = 6.1252, p < 0.01) than that of weathered frass, in which the median CaOx content (0.036 ± 0.029 mmol g⁻¹; Figure 5a) was less than half that of fresh samples. The mean SOx content in fresh frass (0.009 ± 0.000 mmol g⁻¹) was also significantly higher than that of weathered frass (0.005 ± 0.002 mmol g⁻¹) (t_4,2.5% = 3.5961, p < 0.05; Figure 5a).

[IMAGE OMITTED. SEE PDF]

Frass samples collected from the center of heuweltjies had lower mean CaOx contents (0.048 ± 0.048 mmol g⁻¹) than frass from off-mound soils (0.131 ± 0.060 mmol g⁻¹, W = 40, p < 0.001; Figure 5b). CaOx levels in frass from the sites studied in the Succulent Karoo biome were higher (0.156 ± 0.044 mmol g⁻¹) than the sites in the Fynbos biome (0.042 ± 0.034 mmol g⁻¹, t_37,2.5% = −8.9071, p < 0.001; Figure 5c). The mean CaOx content of frass differed significantly (F_2,35,5% = 51.83, p < 0.05) between sites located across the regional rainfall gradient, with the greatest concentrations in frass in the lowest rainfall region (Springbok, 0.156 ± 0.044 mmol g⁻¹), lower concentrations in the intermediate rainfall region (Piketberg, 0.057 ± 0.035 mmol g⁻¹) and the lowest levels in frass from the higher rainfall region (Stellenbosch, 0.015 ± 0.004 mmol g⁻¹; Figure 5d).

Figure 5d shows significantly lower concentrations of SOx in frass from sites studied in the Fynbos biome (0.006 ± 0.004 mmol g⁻¹) compared with the sites in the Succulent Karoo biome (0.013 ± 0.001 mmol g⁻¹, t_37,2.5% = −4.1229, p < 0.001). Frass collected from center mound soils had lower mean Sox concentrations (0.006 ± 0.004 mmol g⁻¹) compared with frass from off-mound soils (0.011 ± 0.001 mmol g⁻¹, W = 79, p < 0.05; Figure 5b). While the mean Sox content of frass was not significantly different between the intermediate rainfall (Piketberg, 0.008 ± 0.003 mmol g⁻¹) and low rainfall (Springbok, 0.013 ± 0.001 mmol g⁻¹) zones, Figure 5d shows that frass samples from these regions had significantly higher Sox concentrations than the higher rainfall zone (Stellenbosch, 0.002 ± 0.001 mmol g⁻¹, H = 23.13, df = 2, p < 0.001).

DISCUSSION

There is evidence that vegetation of the Greater Cape Floristic Region provides a source of oxalates in termite mound soils via litterfall, termite foraging, and frass deposition, which may ultimately contribute to C sequestration in heuweltjie soils via the OCP (Francis & Poch, 2019; Vermonti et al., 2025). Here we report the first quantitative data on soluble and sparingly soluble oxalate salt concentrations in termite frass and vegetation from the Fynbos and Succulent Karoo biomes in the Greater Cape Floristic Region. We also test a rapid MIR spectroscopic method to quantify these analytes in whole samples and extracts.

The MIR spectroscopic technique was much more efficient than the chemical extraction and quantitative analysis procedure. Validation of the MIR models revealed strengths and limitations of this method. While there were similarities between wavelength regions selected as model components for predictions of CaOx and Sox, a greater range of wavelengths was used to calibrate models for prediction of CaOx compared with Sox concentrations in dry plant tissues and for models predicting total oxalate anion concentrations in 2 M HCl extracts compared with water extracts of frass samples (Table 2). This need for a greater range of wavelengths used as predictors may be due to interference from other organic acids present in plant tissues that enter solution in acid extracts (Kyllingsbæk, 1984). Morphological differences between oxalate crystals in plant tissues resulting from their different structural and biological roles (Franceschi & Nakata, 2005) may also have challenged the models. Differences in selected components for predictions of CaOx and Sox in frass (Table 2) showed that the models could distinguish between the two analytes in complex materials based on their unique MIR absorption characteristics, confirming the specificity of this technique for analysis of Sox and insoluble oxalate salts (Kyllingsbæk, 1984).

Both total oxalates and Sox content in plant extracts and CaOx in dry frass samples were quantified accurately (RMSEP_norm < 0.1; Table 2; Appendix S1: Figures S3 and S4) by MIR spectroscopic analysis. Predictions of Sox in dry frass samples or oxalic acid concentrations in frass extracts were less accurate. Therefore, MIR spectroscopy is a suitable technique for CaOx quantification and semiquantitative Sox analysis of frass samples. This method circumvents the requirement for chemical extraction of frass samples and LC–MS analysis of derivatized oxalic acid and is thus more rapid and cost-effective compared with alternative techniques. This method is useful for future studies of geochemical transformations of oxalates via the OCP in termite-affected soils.

The large ranges of CaOx and Sox contents observed in the vegetation samples (0.0–0.9 and 0.0–1.9 mmol g⁻¹, respectively; Table 3) are expected of diverse plant species from the Greater Cape Floristic Region, as oxalate concentrations in the range of 0.2–5.8 mmol g⁻¹ have been measured in plants of other regions (Franceschi & Horner, 1980). Globally, plants that accumulate CaOx are most prevalent in arid regions (Karabourniotis et al., 2020). Median values obtained in this study were like measured oxalate concentrations in pine, bagasse (Liu et al., 2015), beech and fir samples (Dauer & Perakis, 2014). The lower CaOx content in frass compared with plant material (Tables 3 and 4) may be due to degradation of CaOx by oxalotrophic microbiota in the termite gut (Cowan et al., 2024; Suryavanshi et al., 2016) or termites may selectively consume plant tissues with lower oxalate content (Vermonti et al., 2025). Larger ratios of CaOx to Sox in frass compared with plants (Tables 3 and 4) may be due to preferential metabolism of Sox by microorganisms in the termite digestive system. Different oxalate salts may be metabolized selectively by oxalotrophs and influenced by bioavailability and soil characteristics (Bravo et al., 2015; Robertson & Meyers, 2022; Sahin, 2003).

The C/N ratios of frass samples were high (32–38) and fell within the range that results in equal rates of microbial mineralization and immobilization (Brust, 2019). The high C/N ratio of frass samples probably reflects the diet of M. viator termites, which feed on plant material such as bark, leaves, and twigs, tending to forage for wood more than other materials (Coaton & Sheasby, 1974). Partial digestion of recalcitrant components of vegetation may also increase the C/N ratio of the excrement. The woody material observed in the SEM images of frass samples (Figure 4) may also have contributed to the high C/N ratio. The lower C/N ratio of weathered frass samples (32) compared with fresh frass samples (38) may be a result of decomposition by wood-degrading microbes (Brady & Weil, 2017).

Higher mean CaOx content of fresh frass relative to weathered frass (Figure 5a) confirmed that CaOx becomes unstable once released from organic matter via physical decomposition, as the salt becomes susceptible to chemical degradation by microbial enzymatic action (Cailleau et al., 2011; Cowan et al., 2024). Given the rapid and continuous deposition of termite frass (Swanepoel, 2021), we assume that this decomposition occurs within weeks, as older surface frass would become buried quickly beneath new deposits. The higher mean Sox content in fresh frass compared with that of weathered frass (Figure 5a) could contribute to variation of CaOx and Sox concentrations of frass samples within biomes, rainfall regions, and positions relative to the center of the termite mound. Most frass samples collected for regional comparisons in this study were weathered.

Lower mean CaOx content of frass samples collected from the center of heuweltjies compared with frass from off-mound soils (Figure 5b) may be due to faster decomposition of oxalates due to greater levels of microbial activity stimulated by higher nutrient or moisture contents of heuweltjie soils relative to off-mound soils (Booi, 2011; Kunz et al., 2012). We expect the occurrence of the OCP in soils to be related to the abundance of oxalotrophic microorganisms (Khammar et al., 2009). Heuweltjies may thus be niche environments where oxalotrophic bacteria can proliferate in the presence of oxalate-releasing organic matter and in association with fungi that colonize the plant rhizosphere (Hervé et al., 2016). We therefore expect the occurrence of the OCP in soils to be positively related to the abundance of oxalotrophic microorganisms, particularly in heuweltjies.

Decreasing CaOx concentrations of frass samples with increasing rainfall (Figure 5c,d) may be related to CaOx inputs via plant tissues or rates of CaOx degradation by microbes. However, termites selectively harvest plant species according to their diet preferences (Vermonti et al., 2025), CaOx content of frass does not necessarily reflect regional trends in plant tissue CaOx content. Another factor controlling CaOx content in frass samples may be the regulation of microbial decay processes by soil moisture. We expect faster breakdown of CaOx under moister conditions, as microbial decay processes rely on moisture-dependent diffusion of extracellular enzymes via water films as well as supply of resources to soil organisms to support growth and respiration (Cruz-Paredes et al., 2021; Zanne et al., 2022).

Higher Sox content in frass from sites with low to intermediate rainfall relative to sites with higher rainfall (Figure 5c,d) as well as higher SOx content in frass from off-mound soils compared with heuweltjie soils (Figure 5b) may be related to a greater proportion of SOx-synthesizing plants in more arid climates (Cowan et al., 2024; Tooulakou et al., 2016) or to off-mound soils, which tend to hold less moisture (Kunz et al., 2012; Schmiedel et al., 2015). Further research efforts are required to quantify the contribution of different plant species to oxalate content of termite frass in different climatic zones. The MIR spectroscopic method developed in this study is recommended as a useful tool for rapid data collection to build upon the oxalate concentration database for plants of the Greater Cape Floristic Region.

Heuweltjie soils tend to have higher C content compared with off-mound soils (Clarke et al., 2023; Francis et al., 2024). This has been attributed to the cycling of organic matter into the mounds (Clarke et al., 2023; McAuliffe et al., 2018). Establishing the amount of frass produced by a colony or per mound is challenging, as not all frass produced is ejected onto the mound surface. Some frass is stored underground in frass chambers. If a mound is disturbed and underground channels are exposed through a trench or a road cutting, termites often dispose of frass through outlets in the exposed trench wall. The fossil dig site at the West Coast National Park, approximately 50 km from the Koringberg site, offered a unique opportunity to estimate frass production of an M. viator colony, for which no reported values exist in the study region. Frass that had collected in the exposed walls of the dig site was collected over a two-year period and weighed. The average frass production at the dig site (presumably by a single colony) was ca. 69 kg year⁻¹ (Joh Henschel and Amouré Robinson, unpublished data). Using the mean C of six frass samples from Koringberg (25.0%) multiplied by the rate of frass production measured at the West Coast National Park, we calculate a total input of 17.6 kg C into a termite mound per year due to frass deposition. Frass inputs may therefore be an important source of C enrichment in termite mound soils; however, it must be noted that frass input rates may differ between climatic zones.

Assuming that one mole of SIC is sequestered per mole of CO₂ released via oxalotrophy, we calculated that it would take approximately 195,000 years (Appendix S1: Table S1) to accumulate the amount of CaCO₃ that formed in a heuweltjie in Klawer (Clarke et al., 2023). The calculated timeline is longer than the reported lifespan of heuweltjies (13–34 ka) (Francis et al., 2024) and this suggests that sources of C other than oxalate biominerals contributed to the formation of CaCO₃. The annual Ca input requirement for CaCO₃ formation within the heuweltjie lifespan relies on the total Ca content of frass (2.79%, resulting in 2.554 kg annual Ca input) (Vermonti et al., 2025). The Ca content from CaOx in fresh frass samples from Koringberg (0.59%) represents an important component of total Ca inputs (20%). The OCP may therefore be instrumental in carbonate formation due to the release of free Ca, in addition to the rise in soil pH required for the precipitation of CaCO₃.

CONCLUSIONS

Termites contribute significant quantities of organic matter (vegetation and frass) to heuweltjie soils in the Greater Cape Floristic Region. This may be an important source of organic C substrates for carbonate formation via the oxalate-carbonate pathway (OCP). The extent of the OCP relies on the abundance of oxalate inputs in organic matter, but oxalate concentrations in vegetation and frass in the Greater Cape Floristic Region are unknown. This study provided insights into patterns of environmental oxalate salt inputs due to termite frass deposition in heuweltjie soils in sites with different rainfall amounts within the Greater Cape Floristic Region. The higher oxalate content in plants compared with frass may be due to metabolism of oxalate salts (SOx to a greater degree than CaOx) by microbiota in the termite gut, or selective consumption of plant tissues with lower oxalate content by termites. The higher CaOx concentrations in frass collected from sites in more arid regions and off-mound soils may be a result of slower microbial degradation due to lower moisture content of these soils relative to those in mesic regions but may also point to greater colonization of heuweltjie soils by oxalate-degrading microorganisms compared with off-mound soils. A greater predominance of oxalate-accumulating plants in more arid regions (Karabourniotis et al., 2020) may also contribute to the regional differences; however, more extensive vegetation data are needed to confirm this hypothesis. The mean SOx concentration of frass samples collected from sites with lower rainfall and dominated by Succulent Karoo plant tissues was significantly higher than that of frass from sites in mesic, Fynbos-dominated regions. This trend, as well as higher mean SOx content of frass samples collected from heuweltjie soils relative to off-mound soils, may be linked to a greater proportion of SOx-synthesizing plants in the Succulent Karoo biome compared with Fynbos. Termite frass deposition was calculated to add significant quantities of Ca and C to heuweltjie soils on an annual basis, which can contribute to calcite formation via the OCP within timeframes that correspond to reported heuweltjie ages.

Our studies suggest that mid-IR spectroscopy is a more efficient and cost-effective alternative to chemical methods for total and SOx salt quantification in termite frass samples and plant tissues. This technique is applicable to quantify the contribution of OCP to carbon sequestration in oxalate-rich ecosystems globally.

AUTHOR CONTRIBUTIONS

T. Nel: Data curation; formal analysis; investigation; methodology; project administration; visualization; writing—original draft. C. E. Clarke: Conceptualization; funding acquisition; project administration; supervision; writing—review and editing. M. L. Francis: Conceptualization; investigation; supervision; writing—review and editing. D. Babenko: Data curation; writing—review and editing. D. Breecker: Writing—review and editing. D. A. Cowan: Investigation; writing—review and editing. T. Gallagher: Funding acquisition; project administration; writing—review and editing. J. R. McAuliffe: Writing—review and editing. M. Trindade: Investigation; writing—review and editing. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS

This work is based on the research supported in part by the National Research Foundation of South Africa (Ref Number PMDS22051410685) as well as the NRF-NSF Biodiversity on a Changing Planet program (NSF Ref Number 2224993, NRF Grant 150452). We thank Nicola Vermonti (Stellenbosch University) for the provision of plant tissue samples and species identification. We thank Amoure Robinson and Joh Henschel for assistance in collecting the frass depositional rate data, and Pippa Haarhoff (West Coast Fossil Park) for permission. We thank the Stellenbosch University Postgraduate Office and The Skye Foundation for bursary provision. We also acknowledge the DSI-BIOGRIP soil biogeochemistry node for IC method development efforts.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

DATA AVAILABILITY STATEMENT

Data (Nel, 2025) are available from Mendeley Data: .

References

Bellon‐Maurel, V., E. Fernandez‐Ahumada, B. Palagos, J. M. Roger, and A. McBratney. 2010. “Critical Review of Chemometric Indicators Commonly Used for Assessing the Quality of the Prediction of Soil Attributes by NIR Spectroscopy.” TrAC ‐ Trends in Analytical Chemistry 29(9): 1073–1081. https://doi.org/10.1016/j.trac.2010.05.006.

Booi, N. 2011. “Structure and Function of Heuweltjies across a Rainfall Gradient in the South‐Western Cape.” Master thesis, University of Stellenbosch.

Brady, N. C., and R. Weil. 2017. The Nature and Properties of Soils, 15th ed. Essex: Pearson Education Ltd.

Bravo, D., O. Braissant, G. Cailleau, E. Verrecchia, and P. Junier. 2015. “Isolation and Characterization of Oxalotrophic Bacteria from Tropical Soils.” Archives of Microbiology 197(1): 65–77. https://doi.org/10.1007/s00203-014-1055-2.

Brust, G. E. 2019. “Management Strategies for Organic Vegetable Fertility.” In Safety and Practice for Organic Food, edited by D. Biswas and S. A. Micallef, 196–197. London: Academic Press.

Cailleau, G., O. Braissant, and E. P. Verrecchia. 2011. “Turning Sunlight into Stone: The Oxalate‐Carbonate Pathway in a Tropical Tree Ecosystem.” Biogeosciences 8(7): 1755–1767. https://doi.org/10.5194/bg-8-1755-2011.

Carr, A. S., B. M. Chase, A. Boom, M. E. Meadows, and J. M. Sanchez. 2022. “Variability in Soil and Foliar Stable Carbon and Nitrogen Isotope Compositions in the Winter Rainfall Biomes of South Africa.” Journal of Arid Environments 200(May): 104726. https://doi.org/10.1016/j.jaridenv.2022.104726.

Cheng, Z. Y., D. C. Fernández‐Remolar, M. R. M. Izawa, D. M. Applin, M. Chong Díaz, M. Fernandez‐Sampedro, M. García‐Villadangos, T. Huang, L. Xiao, and V. Parro. 2016. “Oxalate Formation Under the Hyperarid Conditions of the Atacama Desert as a Mineral Marker to Provide Cluesto the Source of Organic Carbon on Mars.” Journal of Geophysical Research: Biogeosciences 121(6): 1593–1604. https://doi.org/10.1002/2016JG003439.

Clarke, C. E., M. L. Francis, B. J. Sakala, M. Hattingh, and J. A. Miller. 2023. “Enhanced Carbon Storage in Semi‐Arid Soils through Termite Activity.” Catena 232(November): 107373. https://doi.org/10.1016/j.catena.2023.107373.

Clayden, J., N. Greeves, and S. Warren. 2012. “Determining Organic Structures.” In Organic Chemistry, 2nd ed. 43–80. Oxford University Press.

Climate Data. 2023. “Piketberg Climate.” https://en.climate-data.org/africa/south-africa/western-cape/piketberg-27276/.

Climate Data. 2023. “Springbok Climate.” https://en.climate-data.org/africa/south-africa/northern-cape/springbok-10649/.

Climate Data. 2023. “Stellenbosch Climate.” https://en.climate‐data.org/africa/south‐africa/western‐cape/stellenbosch‐6770/.

Coaton, W. G. H., and J. L. Sheasby. 1974. “National Survey of the Isoptera of Southern Africa. 6. The Genus Microhodotermes Sjöstedt (Hodotermitidae).” In Cimbebasia, Vol. 3, 47–59. Namibia: SWA Research.

Cowan, D. A., D. Babenko, R. Bird, A. Botha, D. O. Breecker, C. E. Clarke, M. L. Francis, et al. 2024. “Oxalate and Oxalotrophy: An Environmental Perspective.” Sustainable Microbiology 1. https://doi.org/10.1093/sumbio/qvad004.

Cowling, R. M., A. J. Potts, P. L. Bradshaw, J. Colville, M. Arianoutsou, S. Ferrier, F. Forest, et al. 2015. “Variation in Plant Diversity in Mediterranean‐Climate Ecosystems: The Role of Climatic and Topographical Stability.” Journal of Biogeography 42: 552–564.

Cruz‐Paredes, C., D. Tájmel, and J. Rousk. 2021. “Can Moisture Affect Temperature Dependences of Microbial Growth and Respiration?” Soil Biology and Biochemistry 156(October 2020): 108223. https://doi.org/10.1016/j.soilbio.2021.108223.

Dauer, J. M., and S. S. Perakis. 2014. “Calcium Oxalate Contribution to Calcium Cycling in Forests of Contrasting Nutrient Status.” Forest Ecology and Management 334: 64–73. https://doi.org/10.1016/j.foreco.2014.08.029.

Esler, K. J., L. von Staden, and G. F. Midgley. 2015. “Determinants of the Fynbos/Succulent Karoo Biome Boundary: Insights from a Reciprocal Transplant Experiment.” South African Journal of Botany 101(November): 120–128. https://doi.org/10.1016/j.sajb.2015.02.006.

Franceschi, V. R., and H. T. Horner. 1980. “Calcium Oxalate Crystals in Plants.” The Botanical Review 46(4): 361–427. https://doi.org/10.1007/BF02860532.

Franceschi, V. R., and P. A. Nakata. 2005. “Calcium Oxalate in Plants: Formation and Function.” Annual Review of Plant Biology 56: 41–71. https://doi.org/10.1146/annurev.arplant.56.032604.144106.

Francis, M. L., F. Ellis, J. J. N. Lambrechts, and R. M. Poch. 2013. “A Micromorphological View through a Namaqualand Termitaria (Heuweltjie, a Mima‐Like Mound).” Catena 100(February 2017): 57–73. https://doi.org/10.1016/j.catena.2012.08.004.

Francis, M. L., L. Palcsu, M. Molnár, T. Kertész, C. E. Clarke, J. A. Miller, and J. van Gend. 2024. “Calcareous Termite Mounds in South Africa Are Ancient Carbon Reservoirs.” Science of the Total Environment 926(May): 171760. https://doi.org/10.1016/j.scitotenv.2024.171760.

Francis, M. L., and R. M. Poch. 2019. “Calcite Accumulation in a South African Heuweltjie: Role of the Termite Microhodotermes viator and Oribatid Mites.” Journal of Arid Environments 170: 103981. https://doi.org/10.1016/j.jaridenv.2019.05.009.

Hervé, V., T. Junier, S. Bindschedler, E. Verrecchia, and P. Junier. 2016. “Diversity and Ecology of Oxalotrophic Bacteria.” World Journal of Microbiology and Biotechnology 32(2): 1–7. https://doi.org/10.1007/s11274-015-1982-3.

Kachkoul, R., T. S. Houssaini, M. Mohim, R. El Habbani, and A. Lahrichi. 2020. “Chemical Compounds Identification and Antioxidant and Calcium Oxalate Anticrystallization Activities of Punica granatum L.” Evidence‐based Complementary and Alternative Medicine 2020: 9424510. https://doi.org/10.1155/2020/9424510.

Karabourniotis, G., H. T. Horner, P. Bresta, D. Nikolopoulos, and G. Liakopoulos. 2020. “New Insights into the Functions of Carbon–Calcium Inclusions in Plants.” New Phytologist 228(3): 845–854. https://doi.org/10.1111/nph.16763.

Khammar, N., G. Martin, K. Ferro, D. Job, M. Aragno, and E. Verrecchia. 2009. “Use of the Frc Gene as a Molecular Marker to Characterize Oxalate‐Oxidizing Bacterial Abundance and Diversity Structure in Soil.” Journal of Microbiological Methods 76(2): 120–127. https://doi.org/10.1016/j.mimet.2008.09.020.

Kunz, N. S., M. T. Hoffman, and B. Weber. 2012. “Effects of Heuweltjies and Utilization on Vegetation Patterns in the Succulent Karoo, South Africa.” Journal of Arid Environments 87: 198–205.

Kyllingsbæk, A. 1984. “Report No. 1708. Determination of Organic Acids in Plant Material.” Lyngby. https://dcapub.au.dk/pub/planteavl_88_169.pdf.

Lal, R., C. Monger, L. Nave, and P. Smith. 2021. “The Role of Soil in Regulation of Climate.” Philosophical Transactions of the Royal Society B: Biological Sciences 376(1834): 20210084. https://doi.org/10.1098/rstb.2021.0084.

Liu, Y., C. Zhang, B. Li, H. Li, and H. Zhan. 2015. “Extraction and Determination of Total and Soluble Oxalate in Pulping and Papermaking Raw Materials.” BioResources 10(3): 4580–4587.

Luther‐Mosebach, J., J. Dengler, U. Schmiedel, I. U. Röwer, T. Labitzky, and A. Gröngröft. 2012. “A First Formal Classification of the Hardeveld Vegetation in Namaqualand, South Africa.” Applied Vegetation Science 15: 401–431.

McAuliffe, J. R. 2022. “Heuweltjies – The “Little Hills” of Western South Africa.” In Fairy Circles of the Namib Desert – Ecosystem Engineering by Subterranean Social Insects. Biodiversity & Ecology, Vol. 7, edited by U. Schmiedel and M. Finckh, 302–339. Goettingen: Klaus Hess Verlag. http://www.biodiversity-plants.de/biodivers_ecol/vol7.php.

McAuliffe, J. R. 2023. “Earthen Mounds (Heuweltjies) of South Africa and Their Termite Occupants: Applicability of Concepts of the Extended Phenotype, Ecosystem Engineering and Niche Construction.” Philosophical Transactions of the Royal Society B: Biological Sciences 378(1884): 20220150. https://doi.org/10.1098/rstb.2022.0150.

McAuliffe, J. R., L. D. McFadden, and M. T. Hoffman. 2018. “Role of Aeolian Dust in Shaping Landscapes and Soils of Arid and Semi‐Arid South Africa.” Geosciences (Switzerland) 8(5): 1–34. https://doi.org/10.3390/geosciences8050171.

McAuliffe, J. R., M. Timm Hoffman, L. D. McFadden, W. Bell, S. Jack, M. P. King, and V. Nixon. 2019. “Landscape Patterning Created by the Southern Harvester Termite, Microhodotermes viator: Spatial Dispersion of Colonies and Alteration of Soils.” Journal of Arid Environments 162(July 2018): 26–34. https://doi.org/10.1016/j.jaridenv.2018.11.010.

Misiewicz, B., D. Mencer, W. Terzaghi, and A. L. VanWert. 2023. “Analytical Methods for Oxalate Quantification: The Ubiquitous Organic Anion.” Molecules 28(7): 3206. https://doi.org/10.3390/molecules28073206.

Molano‐Flores, B. 2001. “Herbivory and Calcium Concentrations Affect Calcium Oxalate Crystal Formation in Leaves OfSida (Malvaceae).” Annals of Botany 88(3): 387–391. https://doi.org/10.1006/anbo.2001.1492.

National Center for Biotechnology Information. 2022. “PubChem Compound Summary for CID 27491, 2,3‐Dihydroxyquinoxaline.” https://pubchem.ncbi.nlm.nih.gov/compound/2_3‐Dihydroxyquinoxaline#section=13C‐NMR‐Spectra.

Nel, T. 2025. “Ecosphere Nel 2025.” Version 1. Mendeley Data. https://doi.org/10.17632/kwf7r6bf4s.1.

Nel, T., C. E. Clarke, M. L. Francis, W. Stone, D. A. Cowan, A. Botha, and T. Gallagher. 2024. “Efficient Quantification of Soluble and Insoluble Oxalates in Clay Mineral Mixtures.” Communications in Soil Science and Plant Analysis 1–15: 1985–1999. https://doi.org/10.1080/00103624.2024.2336574.

Nel, T., C. E. Clarke, and A. G. Hardie. 2023. “Comparison of Soil PH and Exchangeable Cation Quantification by Various Wet Methods with Near‐ and Mid‐Infrared Spectroscopy Prediction.” Communications in Soil Science and Plant Analysis 54(17): 2425–2438. https://doi.org/10.1080/00103624.2023.2223657.

Robertson, C. F. M., and P. R. Meyers. 2022. “Oxalate Utilisation Is Widespread in the Actinobacterial Genus Kribbella.” Systematic and Applied Microbiology 45(6): 126373. https://doi.org/10.1016/j.syapm.2022.126373.

RStudio Team. 2019. RStudio: Integrated Development for R. RStudio. Boston, MA: PBC. http://www.rstudio.com/.

Sahin, N. 2003. “Oxalotrophic Bacteria.” Research in Microbiology 154(6): 399–407. https://doi.org/10.1016/S0923-2508(03)00112-8.

Schmiedel, U., N. Kühne, A. Twerski, and J. Oldeland. 2015. “Small‐Scale Soil Patterns Drive Sharp Boundaries between Succulent “Dwarf” Biomes (or Habitats) in the Arid Succulent Karoo, South Africa.” South African Journal of Botany 101: 129–138. https://doi.org/10.1016/j.sajb.2015.05.001.

Suryavanshi, M. V., S. S. Bhute, N. Bharti, K. Pawar, and Y. S. Shouche. 2016. “Eubacterial Diversity and Oxalate Metabolizing Bacterial Species (OMBS) Reflect Oxalate Metabolism Potential in Odontotermes Gut.” Journal of Pure and Applied Microbiology 10(3): 2035–2044.

Swanepoel, J. 2021. The Influence of Microhodotermes Viator Nest Density and Dispersion on Ecological Processes [PhD]. University of Cape Town.

TheBestTimeToVisit. 2023. “Climate of Komaggas.” https://www.besttimetovisit.com.pk/south-africa/komaggas-3497182/.

Tooulakou, G., A. Giannopoulos, D. Nikolopoulos, P. Bresta, E. Dotsika, M. G. Orkoula, C. G. Kontoyannis, et al. 2016. “Reevaluation of the Plant “Gemstones”: Calcium Oxalate Crystals Sustain Photosynthesis under Drought Conditions.” Plant Signaling & Behavior 11(9): e1215793. https://doi.org/10.1080/15592324.2016.1215793.

Vermonti, N., C. E. Clarke, C. J. Crous, J. A. Miller, T. Nel, and M. L. Francis. 2025. “Termites as Vectors for the Lateral Movement of Elements in the Landscape: Their Influence on Nutrient Cycling and Landscape Heterogeneity.” Applied Soil Ecology 206: 105898. https://doi.org/10.1016/j.apsoil.2025.105898.

Weather Atlas. 2023. “Climate and Monthly Weather Forecast Koringberg, South Africa.” https://www.weather‐atlas.com/en/south‐africa/koringberg‐climate#:~:text=InKoringberg%2CSouthAfrica%2Cduring,(17.2%22)ofprecipitation.

Zanne, A. A. E., H. Flores‐moreno, J. R. Powell, and W. K. Cornwell. 2022. “Termite Sensitivity to Temperature Affects Global Wood Decay Rates.” Science 377: 1440–1444.