healthmedicinet

Experimental sites and climate recordings

A traditional transhumant route used by herders of Olangchung Gola, Taplejung district,
and five pasture sites (Sites 1–5 of 0.17, 0.30, 0.23, 0.21 and 0.14 ha, respectively)
along this route, were selected for the experiment in 2010. The route is situated
within the Kanchenjunga Conservation Area in the Eastern Himalayan Mountains of Nepal.
In order to protect the areas from grazing, the pastures were fenced 2 weeks before
the respective measurement periods started. All paddocks were relatively flat and
had a North-East aspect. Sites 2 and 4 were established at the same pasture site;
however, sites were not contiguous but separated by a small ridge. Sites 1–3 (at 3,200,
4,000 and 4,500 m a.s.l., respectively, upward movement) were grazed in spring and
summer (May 19–27, July 6–14 and August 4–12 at Sites 1–3, respectively), while Sites
4 and 5 were used in late summer and autumn (4,000 and 2,600 m a.s.l., September 1–9
and October 30–November 7, respectively, downward movement). A portable weather station
was installed at each of the five pasture sites. Average temperatures measured at
01.30 p.m. and minimum temperatures, recorded at 06.30 a.m. (both in °C) for Site
1 were 16 ± 5 and 4 ± 3, for Site 2: 14 ± 1 and 2 ± 1, for Site 3: 12 ± 1 and 2 ± 2,
for Site 4: 12 ± 2 and 1 ± 1 and for Site 5: 11 ± 1 and 2 ± 1, respectively. Average
precipitation (measured at 06.30 a.m., in mm/day) recorded was 7 ± 11, 9 ± 7, 12 ± 8,
4 ± 6 and 4 ± 5 for Sites 1–5, respectively.

Experimental animals

All procedures involving the animals followed the international guiding principles
listed by the Council for International Organizations of Medical Sciences and the
International Council for Laboratory Animal Science (2012]). Nepal had no established system for ethical approval of animal experiments. In
the area, two different hybrids are common. The hybrids selected were female Bos taurus × Bos grunniens (? × ?) hybrids, locally called Dimjo chauries, which were shown to tolerate higher
altitudes better than the Urang chauries (Bos grunniens × Bos indicus) (Barsila et al. 2014]). The hybrids used originated from Yangma, Olangchung Gola (4,200 m a.s.l.). In total,
12 female hybrids which had calved between April 7 and May 2 at 2,500–3,000 m a.s.l.
were selected and allocated to two different stocking densities (SD), low and high.
Animals of similar milk yield (MY) and body weight (BW) were chosen (average initial
MY, May 14–16: 2.0 ± 0.20 and 2.1 ± 0.27 kg and initial BW, May 18: 201 ± 7.0 and
202 ± 12.2 kg in low and high SD, respectively). At Sites 2–4, high enough in altitude
for yak husbandry, four female yaks were additionally included that had calved between
May 15 and 20. The yaks had an average BW of 186 ± 3.9 kg and a MY (including the
amount suckled by calves) of 3.3 ± 0.8 kg as measured on May 18 and July 4/5, respectively.
In compliance with traditional practice, the yaks, but not the hybrids, were accompanied
by their calves during the whole experiment.

Experimental design, determination of forage biomass and paddock sizes

The hybrids were kept at two SD in fenced paddocks during measurement periods of 9 days
per site each. The paddock sizes were calculated per site based on assessments of
current forage biomass availability 2 days before starting the measurements and estimates
of animal requirements considering actual BW and MY. Therefore, samples of the total
biomass were harvested from 50 × 50 cm plots (Sites 1–5,  = 12, 3, 6, 6 and 4 plots, respectively). Only the green forage biomass was used
for calculation of paddock size. It was separated into palatable biomass (i.e., forage)
and unpalatable biomass (based on personal assessments of the herders). A dry matter
(DM) content of 200 g/kg was assumed. Feed intake was estimated based on the animals’
requirements using the equation stated in Wiener et al. (2003]) for lactating yaks:

DM intakekg/day and animal=0.008×BWkg0.52+1.369×MYkg/day and animal.

The corresponding assessments for BW and MY were made one and two consecutive days
before stocking the paddocks, respectively. Shortly before starting with the experimental
periods, the final paddocks were established within the previously fenced areas. Additional
space was allocated if part of the area was covered with unpalatable shrubs or stones.
Estimates of trampling losses were considered. Based on these considerations, the
amount of forage on offer was estimated to match the requirements of four adult hybrids
at high SD during 9 days. Low SD was defined as stocking with only half of the animals
for the same period of time in a paddock of similar size. At Sites 1–5, this resulted
in paddock sizes (m) of 41 × 42, 55 × 54, 50 × 45, 50 × 42 and 40 × 35, respectively.
The corresponding SD (m
²
/head) at Sites 1–5 was 861, 1,485, 1,125, 1,050, and 700 at low SD and 431, 743,
563, 525 and 350 at high SD, respectively. Both SD treatments were repeated by independent
paddocks per site. Two additional paddocks, established at Sites 2–4, were stocked
with two yaks per paddock (following the low SD scheme). The animals were moved into
the paddocks after morning milking at about 09.00 a.m. on day 1 and were released
at about 05.00 p.m. on day 9. Before starting the experimental periods, the animals
had been kept for 5 days in the surroundings of the respective pasture site without
entering the pre-fenced areas. In the time between the experimental periods, the animals
were kept together with a larger herd of the respective genotype and were moved along
the transhumance route according to the herder practice.

Recordings and samplings made in the herbage

The total biomass availability was measured in three random plots of 50 × 50 cm/paddock
2 days before the animals were allocated to their respective paddocks. To measure
biomass availability, samples were cut 2 cm above ground (Pande and Yamamoto 2006]), resulting in total  = 12 and  = 18 plots at Sites 1 and 5 and at Sites 2–4, respectively. The standing biomass
was separated by functional groups, first air dried and later dried at 60°C during
48 h in an oven, weighed and milled instantly in order to determine the amount of
air DM. Samples of herbages as selected by the animals were collected per paddock
from days 1–7 of the measurement period from 09.00 a.m. to 03.00 p.m. (and from 10.00
a.m. to 03.00 p.m. at Site 5). The animals were accompanied and their selection behavior
was mimicked for the whole duration of the respective observation period by randomly
switching between animals following the procedure described by Berry et al. (2002]). Every 30 min sampling was resumed in another paddock. Per paddock, two independent
samples (A and B) were collected. In the evening, the samples of 150–200 g fresh weight
were pooled across sampling days by maintaining separate A and B samples, chopped
to small pieces and spread within a tent to facilitate drying.

Measurement of milk yield and composition

Hand milking was accomplished in the morning and the evening (the latter only in the
hybrids). The MY was measured with a digital balance (Scout Pro, Ohaus, NJ, USA).
In the yaks, at first milk let-down had to be initiated by allowing the calves to
suckle for 1–2 min for two times, then hand-milking was practiced and finally the
calves were allowed to suckle again. By weighing them before and after suckling, the
milk amount consumed by the calves was determined. Keeping yaks and their calves separately
at night ensured that most of the milk was consumed in the morning. Still, the actual
total MY of the yaks was probably higher as the calves may have suckled also during
the day. Milk compositional analysis was done in duplicate after each milking in the
9-day experimental periods by a portable ultrasonic milk analyzer (Lactoscan SA-L,
Milkotronic Limited, Nova Zagora, Bulgaria). For the hybrids, aliquots (weighted means)
of morning and evening milk yields were used for calculation of the daily values.
The contents measured were adjusted using regression equations obtained from 25 milk
samples analyzed with the portable milk analyzer and the standard Milkoscan 4000 (Foss
Electric, Hillerød, Denmark) (for details see Barsila et al. 2014]). For statistical analysis, the average of the last 3 days/experimental period was
used. Energy corrected milk (ECM, kg/day and animal) was calculated as milk (kg/day
and animal) × (0.38 × fat (%) + 0.24 × protein (%) + 0.17 × lactose (%)/3.14 (Agroscope
2015]).

Measurement of body weight

Body weight (BW) was assessed with a digital balance (BH-300X, Aditsan, New Delhi,
India, accuracy of ±0.02 kg) covered by a wooden platform of a size of 200 × 50 × 5 cm.
Calves were weighed before and after each milking. Adults were weighed after morning
milking 1 day before the start of each experimental period and after evening milking
on day 9. The daily BW change was calculated as (initial BW minus final BW)/9 days.

Recording of the activity pattern in the hybrids

Times spent walking, standing and lying, and the number of steps were recorded on
Sites 2–4 in always the same two hybrids per paddock ( = 8) by accelerometer-based sensors to record animal behavior (IceTag 3D, IceRobotics
Ltd., Edinburgh, UK). The sensors were fixed at the right hind leg in the morning
of day 1 of the experimental periods during milking and detached when the animals
were released at the evening of day 9. Thus, complete data sets were only available
for 8 days. Data was evaluated by the corresponding software (IceTag Analyser 2008,
IceRobotics Ltd., UK). Analysis was done per minute following Aharoni et al. (2009]) by considering the activity as either standing or lying according to the highest
proportion (i.e. 50%/min). Standing was further separated into walking (3 steps/min)
and standing (?3 steps/min). To account for the times needed for gathering of the
animals and for milking both in the morning and evening, a total of 122 min was removed
from the dataset of each daily record, resulting in 539 min of time allocated to daytime
(09.00 a.m. to 05.00 p.m. = 480 min of the first day + 08.00 a.m. to 08.59 a.m. = 59 min
of the following day), and 779 min of nighttime (06.00 p.m. to 06.59 a.m.). Total
time evaluated per day and animal was 1,318 min. The averages of all 8 days per animal
were used for statistical analysis.

Laboratory analysis

Herbage samples were oven-dried at 60°C for 48 h at the Institute of Agriculture and
Animal Science, Rampur, Nepal, weighed and ground to pass a sieve of 0.75 mm mesh
size on a Thomas mill. In Switzerland samples were analyzed for proximate contents
following standard procedures (AOAC 1997]; Van Soest et al. 1991]). Dry matter and total ash were assessed on an automatic thermogravimetric determinator
(TGA-500, Leco, St.Joseph, MI, USA: AOAC index no. 942.05) and nitrogen with a C/N
analyzer (Analysator CN-2000, Leco, St.Joseph, MI, USA; AOAC 977.02). Crude protein
(CP) was calculated as 6.25 × N. Contents of neutral detergent fiber (NDF), acid detergent
fiber and acid detergent lignin were assessed on a Fibertec System M (Tecator 1020
Hot Extraction, Höganäs, Sweden). The methods were based on Van Soest et al. (1991]), data were corrected for ash content, and for NDF analysis ?-amylase and sodium
sulfite were used. Hemicellulose was calculated as neutral detergent fiber ? acid
detergent fiber, and cellulose was computed as acid detergent fiber ? acid detergent
lignin.

Statistical analysis

The mixed procedure of SAS (version 9.3) was used to perform the analysis of variance.
The botanical composition was analyzed with Model 1 treating pasture site (S) as single
fixed effect and paddock as replicate
Yij=Si+?ij

. The plots used for biomass cutting (three/paddock) were averaged per paddock, and
paddock was used as replicate resulting in  = 4 (Sites 1 and 5) and  = 6 (Sites 2–4).

For ECM yield, milk composition, BW and behavior data, a model was applied exclusively
on data of the hybrids in order to determine site (S) and SD effects. This Model 2
reads:

Yijk=Si+SDj+S×SDij+?ijk

where S and SD as well as their interaction were treated as fixed effects. Site was
treated as repeated variable, animal (replicate) nested within SD was set as subject
and paddock as random effect. For milk data, the last 3 days were averaged.

Model 3 compared the two genotypes for BW data, ECM yield and milk composition at
Sites 2–4 with low SD only. It reads as follows:

Yijk=Si+Gj+S×Gij+?ijk

with G (genotype) and S (site) and the interaction. Both S and G were treated as fixed
effects, S was treated as repeated variable, and animal (replicate) nested within
genotype was considered as subject. Paddock was set as a random effect. The data obtained
on the last 3 days were averaged for milk data. Multiple comparisons among means were
performed with Tukey’s method considering P  0.05 as significant. The data on the chemical composition of the herbage selected
by the experimental animals are arithmetic means. Means of variables subjected to
analysis of variance and given in the tables are Least Square means.