Stable isotopes analysis
For stable isotope analysis, 1mg muskox feces and 1.0–1.5mg plant tissue samples (extracted evenly from leaves, stems, and fruits if available) were homogenized. Each sample was freeze-dried and then prepared with a stable isotope ratio mass spectrometer system (IsoPrime 100; Cheadle, UK) with an elemental analyser (vario MICRO cube; Elementar, Hanau, Germany). Purified CO2 and N2 were used as the sample analysis gas and the isotopic reference gases. GC column resolve CO2 from N2 and reduction column filled with copper wires reduce N2. All results are reported with delta notation, in parts per thousand (‰) relative with PDB standard. Each plant sample was repeated six times of this analysis.
δ13C and δ15N = (Rsample/Rstandard -1) * 1,000 (‰)
The international reference materials of sucrose (ANU C12H22O11; NIST, Gaithersburg, MD, USA) for δ13C and ammonium sulfate ([NH4]2SO4; NIST) for δ15N were analyzed for calibration of reference gases and the internal standard (acetanilide; Thermo Scientific). The analytical precision was based on 10 replicate measurements of acetanilide and was within 0.12‰ and 0.20‰ for δ13C and δ15N, respectively.
We used simmr Bayesian mixing model in the R with five plant food sources, which were known as muskox diet (Forchhammer & Boomsma., 1995; Gustine et al., 2014; Mosbacher et al., 2016), to estimate dietary fiber composition in muskox diets (stable isotopes mixing models in R; Parnell., 2019). The detailed procedure for running the simmr mixing model is described in Swan et al. (2020).