The Use of Herbage N-alkanes as Markers to Estimate the Diet Composition of Yaks on the Qinghai-Tibetan Plateau

The chemical components in plant cuticular wax can be used as markers to estimate the species composition of the diet of grazing animals. In this experiment, composition of the diet of yak on the Qinghai-Tibetan plateau was estimated using n-alkane analysis. During the grazing period, samples of whole plants of the species present, plus fecal samples voided by the yak, were collected, air-dried and ground prior to the extraction of cuticular wax n-alkanes. The species composition of the yak diets was estimated by relating fecal alkane contents to those of the plant species, using the ‘EATWHAT’ software package. The results showed that the n-alkane technique can detect the main dietary components selected by yak. The diet consumed by yak contained 33% Kobresia humilis, 67% Stipa aliena in summer pasture; 26% Potentilla anserine, 74% Carex qinghaiensis in autumn pasture; 52% Carex qinghaiensis, 32% Heteropappus bowerii and 16% Saussurea semifasciata in winter pasture and 5% Carex qinghaiensis, 95% Achnatherum splendens in spring pasture. The apparent selection for forbs is likely to be a reason for nutritional constraint of yak inhabiting alpine environments. (


INTRODUCTION
The species composition of the herbage available to and consumed by grazing animals is an important variable in studies of their nutritional status.Our understanding of the foraging behaviour of free-ranging herbivores is often constrained by our inability to make meaningful measurements of herbage diet composition.Herbage composition of feed ingested by grazing animals is difficult to assess.Direct observation of animals, use of oesophageally fistulated animals or fecal cuticle analysis have commonly been used, but all of these methods are very tedious, time consuming and/or do not allow differentiation to the level of plant species or plant parts (Dove and Mayes, 1991).Oesophageally fistulated animals have the disadvantage of requiring surgery and might be of restricted use only.
The use of n-alkanes as fecal markers for estimating the diet composition of grazing animals has provided a significant advance in such studies and has been used successfully to estimate the proportions of plant species in the diet of free-ranging herbivores (Dove andMayes, 1991, 1996;Bugalho et al., 1999).Because the different pasture species differ in their alkane patterns, and alkanes are largely indigestible during their passage through the herbivore's alimentary tract, the composition of a mixed herbage diet can be determined from the patterns of n-alkanes found in herbage and faeces (Dove and Mayes, 1996).The n-alkane technique has been successfully applied to estimate diet composition, intake and digestibility in sheep, red deer and steer (Dove and Mayes, 1991;Duncan and Mayes, 1999;Bugalho and Myes, 2002;Ru et al., 2002;Valiente and Delgado, 2003;Walter Kelman and Miguel Bugalho, 2003;Premaratne et al., 2005).
The experimental site used in the present study lies in an alpine area, and the species composition is complex.The yaks graze in different areas of natural grassland in the different seasons.The aim of the present study was to use the n-alkane method to estimate the diet composition selected by yak when grazing alpine pastures and to explore whether the method can provide accurate assessments of dietary selection by yak under these conditions.

Study area
The study area was an alpine grassland on the Qinghai-Tibetan plateau, and not very far from Qinghai Lake (37°18′N, 100°16′E) at an altitude of 3,275 m.The region has an alpine climate.Mean annual precipitation is around 327.1 mm and mean annual temperature is -0.5-0.1°C.The lowest temperature in January is about -31.7°C and the highest temperature in August is about 14.5°C.During autumn, the yaks' grazing pasture lies near the bank of Qinghai Lake which abuts with their spring and winter pasture.Their summer pasture is of a swamp type and lies in Qilian mountain valley.
The experiment was conducted from July to August 2004 in summer pasture, during September 2004 in autumn pasture, from October to December 2004 in winter pasture and from March to April 2005 in spring pasture.

Field measurements and preparation of sample
Herbage mass (DM g m -2 ) was measured by cutting all grass and forb species to ground level within fifteen randomly placed 0.50 m×0.50 m quadrat frames.Thirty herbage or forb species were chosen to measure height with a ruler in different pastures.The cover of each species was measured using 10 equal grid portions with a 1 m×1 m quadrat.Samples of the most common individual herbage species were also collected at the same time.All the samples were ground through a 1 mm screen.One sample of each species was dried in a forced draught oven at 60°C for 24 h to calculate DM, then combusted in a muffle furnace to calculate organic matter (OM).Sub-samples were air-dried before determination of acid detergent fiber (ADF), neutral detergent fiber (NDF), crude protein (CP), fat and n-alkane concentrations.Fresh faeces were collected by following eight different yaks on different days in the same period.Visible nonfaeces objects were avoided when taking fecal samples from the ground.Fecal samples were air-dried before analysis of n-alkane concentrations.One part of the fecal sample was dried in the oven at 60°C for 24 to determine DM.

N-alkane analysis
The alkane extraction procedure was carried out using a modification of the method of Mayes et al. (1986) and Dove et al. (2006).Dried and milled herbage samples (2.0 g) and fecal samples (1.0 g) were weighed along with an internal standard (tetracosane, C24 alkane), into screw-capped flasks and were heated with 1 mol L -1 ethanolic KOH solution (13 ml) for 48 h at 90°C.After cooling to 55°C, 3 ml water and 8 ml heptane were added to each flask and the flasks were shaken vigorously.The top, non-aqueous layer was removed and evaporated to dryness, re-dissolved in 2 ml heptane and applied to a silica-gel column (Silica gel 60, Qingdao haiyang Chemical co., Ltd., 5 ml syringe).The hydrocarbons were eluted from the columns with heptane, concentrated to about 20 ml, evaporated and re-dissolved in 1 ml heptane.The solution was transferred to a centrifuge tube for injection into a gas chromatograph.
The hydrocarbons in the purified extracts were analysed by gas chromatography on a ThermoFinnigan gas chromatograph (Trace-GC2000) fitted with a flame ionization detector.The column was a 30 m×0.53 mm inside-diameter capillary column type SPB1 with 1 μm film thickness; Lanzhou Institute of Chemical Physics, China), temperature-programmed from 210 to 295°C at 6°C min -1 .The carrier gas was nitrogen (30 ml min -1 ).The injector and detector were maintained at 300°C throughout the whole process.The chromatographic data was processed using Trace GC software (Chrom-chard).Detector response factors for individual n-alkanes were determined by injecting onto the chromatograph a standard n-alkane mixture (C 29 , C 31 , C 32 ) after every 10 sample extracts.

Estimation of diet composition
The concentrations of the n-alkanes, from both fecal and vegetation samples, were compared iteratively within a least-squares optimization routine.The non-negative leastsquares procedure was implemented using the software EATWHAT (Dove and Moore 1995).

Height, cover, DM and chemical composition of the herbage
The chemical composition of herbage species is presented in Table 1.Generally, the content of NDF in the species in winter and spring pasture was higher than in summer and autumn pasture.High CP contents were found in some leguminous forages, such as Oxytropis deflexa and Taraxacum brevirostre, in the summer pasture.The height, coverage and DM weight of the main species in the four pastures are presented in the four graphs (Table 2).

N-alkane concentration
Concentrations of C29 and C31 alkanes were highest in the herbage group, and the concentration of C32 alkane in most species was very low and almost zero.Particular alkanes differentiated individual species (Table 3).In summer pasture, Kobresia huimlis and Saussurea superba had the highest concentrations of C29 alkane; whilst Gentiana squarrosa and Leontopodium leontopodioides had the highest concentrations of C31 alkane, compared with other species (Table 3).In autumn pasture, the highest concentrations of C29 alkane were found in Batrachium bungei, Setaria viridis and Triglochin palustre.In the autumn, Potentilla anserina had a higher concentration of C31 alkane than any other species (Table 3).The mean concentration of n-alkanes was lower in winter pasture than in summer and autumn pasture.In winter, most species had the highest C31 concentration except for Stellera chamaejasme, Artemisia frigida and Achnatherum inebrians.Saussurea semifasciata had a higher concentration of C29 alkane than any other species (Table 3).Two species were measured in spring pasture.The higher concentration of C29 was found in Carex qinghaiensis.However, no C31 or C32 were found in this species (Table 3).

Composition of the diet
Although fifteen species were analysed from summer pasture, the EATWHAT software showed that there were two species (Kobresia huimlis, Stipa aliena) in yak diet composition (Figure 1).Kobresia humilis was 33%, and Stipa aliena was 67% of the diet.Potentilla anserine (34%) and Carex qinghaiensis (66%) were detected in the yak diet composition in autumn pasture (Figure 2).The yak diet contained three species in winter pasture (Figure 3): Heteropappus bowerii, Saussurea semifasciata, and Carex qinghaiensis.The proportion of Carex qinghainsis was the highest (52%) and that of Heteropappus bowerii and Saussurea semifasciata was 32% and 16%, respectively.In spring pasture, a high proportion (95%) of Achnatherum splendens was found in the diet composition.The other species was Carex qinghaiensis with 5% proportion (Figure 4).

DISCUSSION
The results of this study indicate that the differences between species in their cuticular wax alkane levels can be exploited successfully to estimate the main species composition of mixtures of herbage.Species other than those detected by the EATWHAT software were present in the diet, but only in small quantities.These species may be detected under the conditions of a large range of alkanes analysed.If we want to measure the major diet composition under grazing conditions, the n-alkane technique and the 'EATWHAT' software are a good choice.
The botanical composition of the diet of large herbivores can sometimes be similar to the botanical composition of the top layer of sward (Hodgson, 1986).Carex qinghaiensis was one of the most abundant browse species in autumn and winter pasture and the proportion in  2).
In this study we also found that yak consumed certain forbs in most circumstances, e.g.Potentilla anserine, Heteropappus bowerri, Saussurea semifasciata.The ingestion of forbs could be partially related to the requirements of yak for crude protein, a limiting factor in the nutrition of ruminants.Usually, concentrations of crude protein are very low in the herbage group layer, especially the senescent herbage layer (Fonseca, 1998).This is also proved by our results (Table 1).For example, in winter pasture, Heteropappus bowerii and Saussurea semifasciata contain relatively high levels of crude protein and are an important source of crude protein for yak.However, although Artemisia frigida also contains high levels of crude protein, the yaks rarely choose it.We found it has a strong smell which may have inhibited selection by the yak.The other forbs were not detected in the yak diet.The proportion or the height of these species is very low in the pasture.For example, the height of Stipa crylovii was lower than other herbage species; the herbage mass of Pedicularis alaschanica was lower than other species; and the cover of Pedicularis alaschanica, Stellera chamaejasme, Achnatherum inebrians and Agropyron cristatum was very low (Table 2).On the other hand, these species may contain some anti-nutritional compounds that have inhibitory effects on grazing animals.Further research remains to be investigated on this issue.
The alpine environment is very harsh, with low oxygen, low temperature and high ultraviolet radiation.In order to maintain a sustainable grazing ecosystem, effective and reasonable grazing management is indispensable.Although summer pasture is a growing season, the stocking rate has to be controlled in order to maintain good pasture, for example, to reduce the number of livestock and retain enough forage in the ground.There is a lack of high quality food provided by winter and spring pastures when most of the herbage layer becomes senescent.Yaks are likely to suffer nutritional constraints before re-growth of the   herbage layer towards the end of the spring season.Supplementary feeding at the end of the winter and in the early spring or, if economically viable, irrigation or fertilizer application to pastures, may improve the capacity of grassland by providing additional food to yak.

Figure 1 .Figure 3 .
Figure 1.Composition of the diet of yak during summer pasture.Vertical line on each bar represents standard error of the mean (n = 8).

Figure 2 .
Figure 2. Composition of the diet of yak during autumn pasture.Vertical line on each bar represents standard error of the mean (n = 8).

Figure 4 .
Figure 4. Composition of the diet of yak during spring pasture.Vertical line on each bar represents standard error of the mean (n = 8).

Table 2 .
The height (cm), cover (%) and herbage dry matter mass (DM, g m -2 ) of predominant species in summer, autumn, winter and spring pasture (mean±SD)

Table 3 .
The n-alkane content of species in summer, autumn and winter pasture (mg/kg DM) (mean±SD) (n = 2) Achnatherum splendens provided the greatest cover and herbage mass of the two species present, and constituted the majority (95%) of the diet selected.The dominant species, Kobresia humilis was just a small proportion of the diet compared with Stipa aliena on summer pasture.The reason is that Stipa aliena has a good texture, and the height of Kobresia humilis is too low for yaks to graze (Table