Discrimination factors of carbon and nitrogen stable isotopes in meerkat feces

Diet and feces samples

The meerkats in this study are maintained at the Edinburgh Zoo (Royal Zoological Society Scotland). Fecal samples were taken randomly from an enclosure containing seven adult female meerkats over the course of April 2016. Subsamples of diet were collected once a week for four weeks to account for variability in diet items. Per the information of Edinburgh Zoo animal care workers, the meerkats are fed different combinations of the food items each day, but over the course of a week the amount of each item they eat is roughly equal. Each animal receives the same amount of each food item by weight. A total of 24 fecal samples were collected, along with 2–4 subsamples of each diet item: carrots and apples, horsemeat, dog biscuits, whole frozen small mice, and whole frozen chicks. The above items were used to calculate the trophic discrimination factor from diet to feces. Homogenized bulk muscle with attached skin was sectioned from the whole chicks and mice for analysis.

Stable isotope analysis

Stable isotope analysis was conducted at the Wolfson Laboratory in the School of Geosciences at the University of Edinburgh. The analysis of carbon and nitrogen isotope ratios was performed on a CE Instruments NA2500 Elemental Analyzer and the effluent gas was analyzed for its carbon and nitrogen isotopic ratios using a Thermo Electron Delta+ Advantage stable isotope ratio mass spectrometer. Sediment standard, PACS-2 (δ¹⁵N 5.215‰ (Air) and δ¹³C- 22.228‰(VPDB)) from the National Research Council Canada was used for isotopic analyses. Acetanilide standard (C 71.09% and N 10.36%) was used for elemental compositions. Isotopic data were determined relative to CO₂ and N₂ reference gases whose mean values are derived from the average value of PACS-2 samples within each daily run.

The standard deviation for five analyses of the PACS-2 standard run over the same time period as the study samples was ±0.07‰ for δ¹³C (VPDB), and ±0.14‰ for δ¹⁵N (Air). Elemental analysis, measuring the percentage of carbon and nitrogen in the samples, had an error of 1% for carbon and 4% for nitrogen. The stable isotope ratios are reported in standard notation and referenced to air for δ¹⁵N values and Vienna Pee Dee Belemnite (VPDB) for δ¹³C values. Ratios are defined as δ = (R_sample∕R_standard − 1) where R=¹³C∕¹²C or ¹⁵N/¹⁴N.

Samples were prepared first by lyophilization followed by manual crushing to form a homogenized powder for isotope analysis. Feces and diet samples were subsampled and analyzed both with and without lipid extraction treatment. The samples were lipid extracted by immersion in a 2:1 ratio of chloroform/methanol for 12 h using a Soxhlet apparatus. Following lipid extraction, samples were dried in a 50 °C oven for at least 24 h to evaporate any remaining solvent.

Statistics and data analysis

Statistical tests and analyses were performed in R (R Core Team, 2016), and plots were created in ggplot2 (Wickham, 2009). Trophic discrimination factors were calculated using the averages of all subsampled dietary material. TDFs were calculated for both δ¹³C and δ¹⁵N using the aforementioned equation ΔX_feces = δX_feces–δX_diet where ΔX_feces is calculated in per mil (‰). δX_diet is the value of the average of all non-lipid-extracted dietary samples and δX_feces is each individual fecal stable isotope value.

Parametric statistical tests were performed, as the data are normally distributed as shown through a Shapiro–Wilk test (δ¹³C: W = 0.946, p = 0.057; δ¹⁵N: W = 0.976, p = 0.544). Before hypothesis testing, an F-test was performed to see if a t-test for equal or unequal variance (Welch two sample t-test) should be done based on the results. F-tests show the variance was equal for comparisons between lipid-extracted and non-lipid-extracted feces for both δ¹³C and δ¹⁵N (δ¹³C: F_8,23 = 1.409, p = 0.491; δ¹⁵N: F_8,23 = 0.408, p = 0.191). Variance for comparisons between diet and feces for both δ¹³C and δ¹⁵N are unequal (δ¹³C: F_15,23 = 5.001, p = 0.0006; δ¹⁵N: F_15,23 = 2.633, p = 0.036). All means are reported ± standard deviation (SD), and all significance is reported for α = 0.05.

To obtain carbon and nitrogen fecal trophic discrimination factors from other studies for comparison to results from this study, a literature search was conducted using Google Scholar and through mining references in other feces trophic discrimination factor papers. These were all of the papers with diet-feces trophic discrimination factors found as of August 2016 searching keywords such as “feces stable isotopes” and “scat stable isotopes”. These values appear in Tables 1 and 2.

Trophic discrimination factors of meerkat feces

Feces can be an extremely useful substrate for stable isotope analysis for estimating food input, and these data show in meerkats that in relation to carbon isotopes, feces are a fair representation of diet, but nitrogen isotopes of feces undergo isotopic enrichment. Investigations into stable isotope discrimination factors show that herbivore feces are representative of diet in large bodied animals (Sponheimer et al., 2003a) but that in small-bodied herbivores (Hwang, Millar & Longstaffe, 2007) and non-herbivores (Montanari & Amato, 2015) feces undergo enrichment of nitrogen isotopes within the digestive tract. In this study, the trophic discrimination factor of carbon isotopes from diet-to-feces is small and not statistically significant. Similar results for Δ¹³C_feces are seen in studies of insectivores (bats, Salvarina et al., 2013) and carnivores (tigers and snow leopards, Montanari & Amato, 2015). To this point, experimental research on diet-to-feces discrimination factors shows most calculated carbon TDFs are non-significant as referenced by the review of published studies in Table 1. More data are needed on mammalian omnivores and carnivores to further assess this pattern.

The means of δ¹⁵N_diet and δ¹⁵N_feces are different (Table 4); therefore the Δ¹⁵N_feces value (1.5 ± 1.1‰) is significantly different than zero. In this study and the other two non-herbivore fecal TDF studies of Salvarina et al. (2013) and Montanari & Amato (2015), Δ¹⁵N_feces is of similar magnitude and also the only TDF that is different than zero; although Δ¹⁵N_feces is only significant in one out of two species examined in each of these studies. Nevertheless, significance of Δ¹⁵N_feces could indicate ¹⁵N enrichment occurs during digestion, likely during transit through the gut as is seen in Hwang, Millar & Longstaffe (2007). This study investigated δ¹⁵N values at different parts in the digestive tract of voles and other small rodents and found digested material is enriched in ¹⁵N in the stomach, intestine, cecum, and colon relative to the diet. It has also been shown the δ¹⁵N of mucosal epithelium was higher than the diet input in some parts of the digestive tract of a sheep, which suggests the enrichment in ¹⁵N in feces may be due to the presence of endogenous proteins (Sutoh, Obara & Yoneyama, 1993). In general, significant Δ¹⁵N values point to some relationship between digestive processes and Δ¹⁵N, but the cause is unknown and more experiments comparing ¹⁵N enrichment during digestion in herbivores and non-herbivores are needed.

A number of other factors may be also affecting nitrogen flux during digestion and excretion. Different biochemical pathways related to the changes in proteins occurring during digestion (deamination/amination) could be the cause of ¹⁵N enrichment (Hwang, Millar & Longstaffe, 2007). Additionally, the presence of microorganisms in the digestive tract could also enrich fecal matter compared to diet (Hwang, Millar & Longstaffe, 2007; Macko et al., 1987). The ¹⁵N enrichment could be due to preferential absorption of ¹⁴N by the animal during digestion or removal during urine excretion, but experimental data from Sponheimer et al. (2003b) seem to indicate that at least in herbivores, ¹⁴N is not preferentially excreted. In Montanari & Amato (2015), we cautioned that more data should be collected to better understand TDFs of carnivores because the physiological mechanisms are still unknown; this study continues in the same vein and reinforces that there is a preferential loss of ¹⁴N and/or enrichment in ¹⁵N occurring during movement through the gastrointestinal tract in relation to solid waste. Further mammalian physiological experiments may explain the mechanism.

Variables that affect trophic discrimination factors

It has been shown that TDFs vary due to a number of factors, one of them being the initial isotopic ratio of the diet. Significant relationships between both δ¹³C and δ¹⁵N of diets and TDFs have been shown in bears (Hilderbrand et al., 1996; Felicetti et al., 2003), rats (Caut, Angulo & Courchamp, 2008), and birds (Pearson et al., 2003). A meta-analysis of TDFs and dietary isotope ratios values by Caut, Angulo & Courchamp (2009) finds significant negative relationships between these variables in both carbon and nitrogen. A similar study of this relationship with fecal TDFs cannot be completed due to the fact most of the calculated fecal TDFs are not significantly different from 0 (data from Tables 1 and 2). This could be due to the fact that feces represents immediately ingested diet (within hours or days) as opposed to assimilated diet and is not subjected to as many physiological processes of fractionation. This is a promising result for the use of feces in isotope studies because the variability of TDFs due to dietary input is a major issue for stable isotope food web studies. Variability that needs to be accounted for with other tissues (Caut, Angulo & Courchamp, 2009) appears to be a non-issue in feces.

Mammalian body size might influence calculated TDF (Hwang, Millar & Longstaffe, 2007). A mechanism that could explain this is a general trend of higher mass-specific metabolism in animals with smaller body mass (Kleiber’s law), which could in turn impact the fractionation of isotopes that occurs after food ingestion (Pecquerie et al., 2010). Hwang, Millar & Longstaffe (2007) did a meta-analysis of fecal TDFs in the literature combined with their rodent TDFs and found significant differences in Δ¹³C between different body sizes, but only in herbivores. A lack of published fecal TDFs for non-herbivorous mammals of different sizes means statistical test cannot be performed for herbivores vs. non-herbivores.

Isotopic variability over time

Feces can be used for point estimates of diet, but long term collection is especially useful for finding seasonal patterns in dietary variability (e.g., Blumenthal et al., 2012). Mammalian carnivores and omnivores tend to change their diets seasonally, so it is important to realize the magnitude of fecal isotope variation day-to-day for wild studies (e.g., Kincaid & Cameron, 1982; Melero et al., 2008). Other than this study, Blumenthal et al. (2012) and Salvarina et al. (2013) track stable isotope ratios of animal feces at monthly or daily intervals respectively.

I have tracked meerkat fecal stable isotopes over the course of one month, and it is clear there is variation on any given day a sample is taken (Fig. 1), as the combination of diet items they eat changes daily. There is a range of 7.3‰  in the δ¹³C and 4.5‰  δ¹⁵N_feces of meerkat feces over a month. This suggests feces can be directly reflective of a highly variable diet, and also shows δ¹³C_feces may not be buffered in small mammals as much against day-to-day variability as in larger mammals (Blumenthal et al., 2012). It is not known exactly how long it takes isotopes of feces to reach dietary equilibrium in meerkats, but in Salvarina et al. (2013) it was shown that bat feces acquired a new dietary signal in 2–3 h so it stands to reason it is also within hours for a small mammal. Examining the daily fluctuations in δ¹³C and δ¹⁵N in feces indicates timing of major dietary changes can be pinpointed quite precisely using this method, at least within days. Due to the fact the TDFs in this study were calculated using an average diet value, the TDF is also averaged and is meant to act as a general guide for a TDF in the wild. This variability is important to realize for wild studies, as it emphasizes the need for larger sample sizes of feces, such as multiple samples per day or week, to lessen the impact of day-to-day variability if researchers are seeking a long-term ecological or environmental trend (e.g., Blumenthal et al., 2012).