國立臺灣大學生命科學院生態學與演化生物學研究所 碩士論文
Institute of Ecology and Evolutionary Biology College of Life Science
National Taiwan University Master Thesis
合歡山地區冷鐵杉混合林中台灣高山田鼠的覓食生態：
探討植物特性的影響
Foraging Ecology of Taiwan Field Vole (Microtus kikuchii) in a Taiwan Fir-Taiwan Hemlock Forest at the Hehuan Area:
Effects of Plant Attributes
周柏翰
Po-Han Chou
指導教授：林雨德 博士
Advisor: Yu-Teh K. Lin, Ph.D.
中華民國 103 年 6 月
June 2014
i
摘要
草食動物的覓食生態學有一重要議題在探討植物特性(如：化學特性、物理特 性以及相對豐度)如何影響草食動物的覓食選擇；同時這也協助我們預測草食動物 對植物群聚造成的影響。室內的植物可食度實驗和野外的動物食性分析，兩者結 合，提供了了解覓食生態的重要資訊。台灣高山田鼠(Microtus kikuchii)為台灣特有 種。先前研究已研究了高山草原中高山田鼠的覓食生態。本研究旨在了解在合歡 山冷鐵杉混合林中植物特性對台灣高山田鼠的覓食生態之影響。我分析三個季節 中(三月、七月以及十一月)台灣高山田鼠的食性，並同時進行餵食實驗，後者包含 了 五種 優勢植種 ：玉山 箭 竹 (Yushania niitakayamensis)，玉山鬼督 郵 (Anisliaea reflexa) ， 裂 葉 樓 梯 草 (Elatostema trilobulatum) ， 玉 山 擬 鱗 毛 蕨 (Dryopsis transmorrisonensis)以及日本曲尾苔(Dicranum japonicum)。我分別檢測五種植物的 七種化學成分、硬度和相對豐度。結果顯示，高山田鼠的食性主要由玉山箭竹組 成，並且不同植種對田鼠有不同的可食度。而在食性結果具有季節上的差異。化 學成分對可食度具有顯著的影響：粗蛋白對可食度有正向的影響。另外，硬度對 可食度有顯著的負向影響。基本上，食性分析和可食度的結果相吻合。總言之，
在高山森林內對高山田鼠來說玉山箭竹仍是最重要的食物來源，植物特性也對可 食度具有顯著的影響。
關鍵字：台灣冷杉林、食性分析、可食度、台灣高山田鼠、玉山箭竹
ii
Abstract
A key aspect of herbivore foraging ecology investigates how plant attributes,
including chemical, physical characteristics, and relative abundance affect plant
palatability and herbivore diets, which, in turn, help us predict the impact of herbivory
on plant communities. The Taiwan field vole (Microtus kikuchii) is an endemic species
in Taiwan. Previous studies have investigated its foraging ecology in alpine meadows.
In this study, I aimed to understand the foraging ecology of Taiwan field voles in a
Taiwan fir-Taiwan hemlock forest at the Hehuan area. I analyzed the diets of Taiwan
field voles and conducted palatability feeding experiments in three seasons (March, July,
and November). Five dominant plants were included in feeding experiments: Yushania
niitakayamensis, Anisliaea reflexa, Elatostema trilobulatum, Dryopsis
transmorrisonensis and Dicranum japonicum. I measured 7 chemical compounds,
toughness, and relative abundance of the 5 species. The results showed that vole diets
were mainly composed of Yushania niitakayamensis, which was also the most palatable
plant. Different species had different palatability to voles. Vole diets showed significant
seasonal effects. Chemical characteristic of plants affected palatability: crude protein
had a positive effect. Furthermore, toughness had a negative effect on palatability.
Besides, the results in diet analyses and palatability experiments were generally
iii
consistent with each other. In conclusion, Yushania niitakayamensis remains the most
important food resource for Taiwan field voles in alpine forest. Plants attributes
significantly influence palatability.
Key words: Taiwan fir forest (Abies kawakamii), diet analysis, palatability, Taiwan field
vole (Microtus kikuchii), Yushan cane (Yushania niitakayamensis)
iv
Content
摘要 ... i
Abstract ... ii
Introduction ... 1
Materials & Methods ... 7
Field survey ... 7
Vegetation survey ... 7
Vole trapping ... 8
Diet analyses ... 9
Palatability & feeding trials ... 11
Chemical analyses ... 14
Toughness analyses ... 16
Statistical analyses ... 17
Results ... 19
Field survey ... 19
Vegetation composition ... 19
Animals trapping ... 19
Diets of Taiwan filed voles ... 21
Palatability of dominant plants... 22
Effects of plant attributes on palatability ... 23
Effects of chemical characteristics of plants on palatability ... 23
Effects of physical characteristics of plants on palatability ... 24
Effects of abundance of plants on palatability ... 25
Diet analyses and palatability of dominant plants ... 25
Discussion ... 27
References... 36
Appendix ... 81
v
Content of Tables
Table 1. Percent coverage in area of dominant plant species in the study site ... 40
Table 2. The species list of vegetation in the study area ... 41
Table 3. Numbers of voles provided fecal samples in diet analyses and entered feeding trials ... 45
Table 4. Results of a two-way ANOVA that examined the effects of season and sex on body weight of voles used in feeding trials ... 46
Table 5. Vertebrates that were caught during vole trapping in the study area from June, 2011 to November, 2013 ... 47
Table 6. The relative importance by area of different food items in the vole’s diet in March ... 48
Table 7. The relative importance by area of different food items in the vole’s diet in July ... 49
Table 8. The relative importance by area of different food items in the vole’s diet in November ... 50
Table 9. Results of Chi-square tests that examined the effects of season on diet composition ... 51
Table 10. The effect of season on diet composition based on the Kruskal-Wallis tests ... 52
Table 11. Standardized palatability of five dominant plants in March, July and November, in 2012 and 2013 ... 53
Table 12. Ranking in standardized palatability of five dominant plants in March, July and November, in 2012 and 2013 ... 54
Table 13. Results of a two-way ANOVA using MCMCglmm that examined the effects of season and species on standardized palatability ... 55
Table 14. Chemical attributes of plants (aboveground parts) in March, 2012 ... 56
Table 15. Chemical attributes of plants (aboveground parts) in July, 2012 ... 57
Table 16. Chemical attributes of plants (aboveground parts) in November, 2012 ... 58
Table 17. Results of Principle Component Analysis ... 59
vi
Table 18. Eigenvectors of the Principle Components ... 60 Table 19. Correlation matrix of 7 chemical compounds and three principle components ... 61 Table 20. Simple linear regression of standardized palatability and PCs ... 62 Table 21. Toughness of five dominant plants measured in the laboratory in three
seasons ... 63 Table 22. Results of a two-way ANOVA using MCMCglmm that examined the effects of species and location (laboratory and field) on toughness ... 64 Table 23. Results of a two-way ANOVA using MCMCglmm that examined the effects of species and season on toughness ... 65 Table 24. Simple linear regression of standardized palatability and toughness ... 66 Table 25. Simple linear regression of standardized palatability and relative
abundance ... 67 Table 26. Simple linear regression of relative abundance and diets of five plant species
in three seasons ... 68 Table 27. Results of model selection ... 69
vii
Content of Figures
Figure 1. The concept map of this thesis ... 70 Figure 2. Sexual difference in body weight of voles that used in feeding trials in three seasons ... 71 Figure 3. The relationship between diets and relative abundance of five tested plants
... 72 Figure 4. Standardized palatability of five dominant plants in three seasons ... 73 Figure 5. Simple linear regression of standardized palatability and PC3 ... 74 Figure 6. Toughness of five dominant plants that measured in laboratory and field .. 75 Figure 7. Toughness of five dominant plants that measured in laboratory in three
seasons ... 76 Figure 8. Simple linear regression of standardized palatability and toughness ... 77 Figure 9. Simple linear regression of standardized palatability and abundance ... 78 Figure 10. Relationship between standardized palatability and diets after accountng for availability ... 79 Figure 11. Result of cluster anaylsis by centroids hierarchical method ... 80
1
Introduction
Foraging ecology of herbivores investigates how herbivores interact with plants
(Olff & Ritchie, 1998; Provenza et al., 2003). Herbivores can change plant communities
by selectively consuming plants, which reduce the relative abundance of plants they
prefer (Wu & Shih, 2010). Herbivores often choose from the variety of plants based on
plants’ quality, i.e., biochemical and physical attributes. The consumed plants affect
herbivores’ body growth rates, reproduction, and population density (Cole & Batzli,
1979). Plants may, in turn, respond to herbivore consumption by altering biochemical
and physical attributes. Therefore, understanding what herbivores prefer to eat in the
fields, and why, are important for predicting the effects of herbivory on plant
communities and of plants on herbivore populations. It is thus critical for inferring how
herbivores and plant communities will be impacted by the changing environment
(Litvaitis, 2000).
Herbivore diet and palatability of plant to herbivores are two basic pieces of
information needed to predict the effect of herbivory (Kimball & Provenza, 2003). Diet
analysis gives information on what herbivores eat in the field; palatability gives
information on how much a plant would be consumed under a controlled environment.
The two pieces of information support each other. Although some studies have found
2
that results of diet analyses were similar to those of palatability (reviewed in Batzli,
1985), either information alone is not sufficient to predict the impact of herbivory on
vegetation for two reasons: (a) preferred food is prone to be eaten first in the nature,
certain plant species tend to be underestimated in diet analyses (Batzli & Pitelka, 1983);
(b) palatability is not measured in a natural setting, thus could give artificial information
(Batzli, 1985).
Both herbivore diet and plant palatability are a consequence of the interaction
between herbivores’ ability to obtain and consume plants and external environments
(Kimball & Provenza, 2003). The former may involve intra- and inter-specific
competition, predation, and herbivores’ attributes such as physiological adaptations and
foraging strategies. The latter may include three main plant attributes: chemical
characteristics of plants, physical characteristics of plants, and the availability of plants
in the environment. For example, protein and fibers were important positive and
negative chemical factors, respectively, in determining palatability of foods
(Bucyanayandi & Bergeron, 1990; Rezsutek & Cameron, 2011). Plant secondary
metabolites such as phenolics, tannins, alkaloids, and monoterpenes could also deter the
consumption by herbivores (Barthelmess, 2001; Bergeron & Jodoin, 1987;
Bucyanayandi & Bergeron, 1990; Goldberg et al., 1980; Hartley et al., 1995; Marquis
3
& Batzli, 1989; Takahashi & Shimada, 2008). Overall, while considering chemical
characteristics of plants, both positive (e.g., proteins) and negative (e.g., plant secondary
metabolites) factors need to be considered (Bergeron & Jodoin, 1987; Torregrossa &
Dearing, 2009b). Second, the physical characteristics, such as toughness of plants can
resist the consumption of herbivores (Hanley et al., 2007; Scheidel & Bruelheide, 1999).
The physical characteristics of plants could be divided into many traits; such as tensile
and shearing strength (Laca et al., 2001). The measurements of toughness represented
the overall tissue strength of plants (Laca et al., 2001). Silica and lignin could deter the
damage to plants caused by herbivores (Kimball & Provenza, 2003; Massey at al.,
2007). It took longer time for herbivores to ingest and digest tough plant material, as a
result, reduced the total intake of food (Laca et al., 2001). Third, as the availability of a
plant species in the environment increases, the chances of the plant species being
encountered and consumed by herbivores increase. Thus, other things being equal, the
proportion of a plant in diets should increase with its relative abundance in the fields
(Boyle et al., 2012). Ideally, all three plant attributes should be considered to better
understand herbivore-plant interactions. However, most empirical studies focused on a
single attribute. Few combined all plant attributes altogether. Finally, the characteristics
of plants likely change with seasons, seasonal variation of diets and palatability should
4
be examined (López-Wilchis & Torres-Flores, 2007; Lindroth & Batzli, 1984).
Taiwan field vole (Mircotus kikuchii) is an endemic species in Taiwan, living in
alpine meadows and fir forests, where Taiwan fir (Abies kawakamii) and Taiwan
hemlock (Tsuga chinensis var. formosana) are dominant woody plants and Yushan cane
(Yushania niitakayamensis) is dominant herbaceous plants (Chen, 1998). Taiwan field
voles usually co-exist with two other rodent species, Formosan white-bellied rats
(Rattus culturatus) and Formosan field mouse (Apodemus semotus) in the forests
(Chang-Jen, 1997; Yeh, 2012). Lyu (1991) proposed that the reproductive cycles of
voles were closely linked to their food resources, especially Yushan cane. Yeh (2012),
using stable isotopes, found that the voles consumed more plants than other food
resources in both alpine meadows and fir forests. Ho (2009) examined the palatability
of thirteen plants to voles in alpine meadows, and found that Yushania niitakayamensis (玉山箭竹) and Carex spp. (薹屬) were the most palatable plants. The palatability
could be explained by the abundance of plant species, and the percentage of
hemicellulose they contained. Furthermore, Yeh et al. (2012) found that the preference of voles for different parts of Yushan cane varied with seasons, and vole’s consumption
facilitated the asexual reproduction (shooting) of Yushan canes. The above-mentioned
studies have shown, in alpine meadows, the herbivory of Taiwan filed vole strongly
5
impacted the plants, and plant attributes did influence the foraging of voles. However,
the relationship between voles and plants in fir forests remained unclear, given that the
plant communities in the fir forests are dramatically different from that in the meadow.
The alpine ecosystems are expecting to see great changes because the global
temperature has been proposed to increase in the following decades (Van Vuuren et al.,
2008). Understanding the relationship between plants and herbivores should allow us
better predict the impacts of environmental changes. The purpose of this thesis was to:
(1) understand the diets of voles in fir forests, and see if diet choice is consistent with
plant palatability over seasons, and (2) test the palatability of five dominant plants to
voles, and (3) examine the effects of plant attributes on palatability. I included all three
plant attributes: chemical characteristics, physical characteristics, and abundance. The
concept map of this thesis is shown in Fig. 1. Specifically, I aimed to test the following
hypotheses:
Diets of Taiwan field voles
(1) Taiwan field voles forage selectively, and their diets change with seasons.
(2) Taiwan field vole diets reflect the palatability of plants.
Palatability of dominant plants to Taiwan field voles
(1) The palatability of different plant species differs, and the palatability within the
6 same plant species changes with seasons.
Plant attributes on palatability
(1) High palatability is associated with high nutrients and low digestion inhibitors.
(2) High palatability is associated with low toughness.
(3) High palatability is associated with high availability.
7
Materials and Methods
Study area
This study was conducted in a Taiwan Fir and Taiwan Hemlock mixed forest at the
Hehuan Mountains (24°09’41.1”N, 121°17’10.4”E, 3005 m in altitude) of the Taroko
National Park, Taiwan. The forest is nearby the High-Altitude Station of the Endemic Species Research Institute, Taiwan. The annual mean temperature was 7.0℃ and
rainfall 366 mm (Yeh, 2012). Taiwan fir (Abies kawakamii, 台灣冷杉) and Taiwan
hemlock (Tsuga chinensis var. formosana, 台灣鐵杉) are dominant woody plants, and
Yushane cane (Yushania niitakayamensis, 玉山箭竹) is dominant herbaceous plant
(Yeh, 2012).
On a 30 degree slope, I established an 11-by-11 sampling grid composed of 11
parallel lines (A to K), each with 11 trapping stations. The distances between lines and
between stations were 10 meters. Trapping stations were marked with aluminum stakes.
Line A was soon abandoned because it was at the edge of a cliff.
Field survey
Vegetation survey
To estimate the coverage of vegetation, 3 stations were randomly sampled along
8
the slope gradient of each grid line. Thirty stations were sampled in total. I randomly
overlaid a 2-m-by-2-m frame on the ground, and estimated the percent coverage in area
of each plant species within the frame. Vegetation was surveyed 3 times a year during
vole trapping (see below).
Vole trapping
Vole trapping and vegetation survey were done in March, July, and November
from July, 2011 to November, 2013, a total of 8 trapping sessions. Voles were trapped
with a multiple-capture Ugglan special live trap (LxWxH=25-cm x 7.8-cm x 6.5-cm)
and a squirrel cage (LxWxH=27-cm x 17-cm x 27-cm) at each station. Traps baited with
sweet potato and oats mixed with peanut butter were serviced for five to six consecutive
days. Traps were opened on the first evening, and checked twice in the morning and in
the afternoon each day. Whenever a vole was captured, I collected its fresh fecal pellets immediately. Pellets were preserved in 70% alcohol and stored in a -80℃ refrigerator
before diet analyses (see below). New individuals were marked with a fingerling ear tag.
The following information was recorded: trapping station, ID, sex, body weight,
reproductive condition (testes scrotal or abdominal for males; vaginal perforated or non-perforated for females). Adult voles (Male body weight ≧ 27 g; Female body
9
weight ≧ 26 g; Lyu, 1991) were brought back to the High-Altitude Station of the
Endemic Species Research Institute for feeding trials (see below). All voles were
released where they were captured after feeding trials.
Diet analyses
I followed the procedures used by several studies (Johnson et al., 1983; Lin & Lee,
2003), except that fecal samples were not sieved, to examine food fragments in fecal
pellets to quantify vole diets. I first prepared reference images of the majority of plant
species found at the study site. Preparation procedures were as follow: I collected
aboveground parts of each plant species, separated leaves and stems, and treated them
as different reference samples. Samples were cut into 0.5 cm fragments, and soaked in
95% warm alcohol to dissolve pigment. I then soaked samples in 3 N NaOH, and
replaced NaOH daily until the samples were transparent that I could see epithelial cells
clearly. I rinsed NaOH away with water, then preserved samples in 70% alcohol. I then
took photographs of the epithelial cells through microscope.
I analyzed the fecal contents of 10 captured voles (randomly chosen) in each
season. I first crumbled and homogenized fecal pellets of a vole in 70% alcohol with a
glass rod. Three slides were made for each vole, and examined under a microscope with
10
400X magnification. For each slide, I looked for epithelial cells in 20 ocular fields, and
took photographs of each field through microscope. There was a total of 60 (3x20)
photographs per sample.
I compared the epithelial cells observed in the microscope field against the
reference images of known plants to identify plant species consumed by voles. I
recorded both the frequency and area of food items, including plant and animal
fragments observed using the ImageJ 1.47v software. The five plant species, Yushania niitakayamensis, Ainsliaea reflexa (玉山鬼督郵), Elatostema trilobulatum (裂葉樓梯
草), Dryopsis transmorrisonensis (玉山擬鳞毛蕨), and Dicranum japonicum (日本曲
尾苔), tested in the feeding trials (see below) were particularly noted. Food items
observed were identified to species if possible, then grouped into three categories:
plants, insects and unknown (unidentifiable tissues or spores). Plant fragments were also
further grouped into monocots (e.g., Yushania niitakayamensis), dicots (e.g., Ainsliaea
reflexa and Elatostema trilobulatum), ferns (e.g., Dryopsis transmorrisonensis) and
moss (e.g., Dicranum japonicum). The data from the 20 fields were pooled, and
averaged over the 3 slides. The amount of each food item consumed by a vole was
expressed as percentage based on the area of food items following the formula:
Pia＝(Ai / T)×100%
11 Pia: Percentage of i food item based on area
Ai: Area of i fragment
T: Sum of i fragment areas in the 20 microscope fields
Alternatively, the amount of each food item consumed by a vole was expressed as
percentage based on the frequency of food items following the formula:
Pic＝(Ci / T)×100%
Pic: Percentage of i food item based on frequency (count)
Ci: Numbers of i food item counted
T: Sum of i fragments counted in the 20 microscope fields
Both calculated percentages gave the relative importance of a food item in the vole’s diet (Hansson, 1970). Spores of fungi and ferns were excluded from analysis
because they were too small to be compared with other fragments (Hung, 2002).
Palatability & feeding trials
The palatability of plants was measured as the amount of plant material consumed
by voles over 12 hours. Trials were carried out over two years in three seasons: spring (3/21–4/7), summer (7/2–7/21) and autumn (11/12–12/1) in 2012; spring (3/23–
4/11), summer (6/30–7/22) and autumn (11/05–11/24) in 2013 in the laboratory of the
12
High-Altitude Station of the Endemic Species Research Institute, Taiwan. Voles
captured from the study area were immediately transported to and maintained in the
laboratory under a 12-L:12-D light regime in ambient temperature. Voles were housed
individually in standard rat cages (D47.0 x W25.5 x H21.5 cm3) with 5-cm-thick aspen
chip bedding (TAPVEI ® ) for at least 5 days prior to the feeding trials to allow them
accommodate to the housing environment. Water and food (fresh sweet potatoes and
oats) were available ad libitum during this period. Females, if found pregnant, during
the course of feeding experiment were excluded from further analyses because of their
additional nutritional needs (Provenza et al., 2003).
The methods of feeding trials were adapted from the experimental protocols
reported in previous researches (Batzli & Lesieutre, 1991; Marquis & Batzli, 1989; Ho,
2009). The top five dominant plants recorded in the vegetation survey (Table 1) were chosen as study targets. They were Yushania niitakayamensis (玉山箭竹), Ainsliaea
reflexa ( 玉 山 鬼 督 郵 ), Elatostema trilobulatum ( 裂 葉 樓 梯 草 ), Dryopsis
transmorrisonensis (玉山擬鳞毛蕨), and Dicranum japonicum (日本曲尾苔). Large
overhead woody plants, including Abies kawakamii (台灣冷杉) and Tsuga chinensis var.
formosana (台灣鐵杉) were not chosen for two reasons: First, their heights, exceed 10
meters on average, likely exclude accessibility by voles (Chen, 1998). Second, a
13
preliminary feeding trial indicates voles do not consume pine cones or pine seeds.
Upon the start of feeding trial, voles were moved to a new cage of the same size.
Sheets of paper were used as bedding. Each vole was given one of the tested plant
species per day for five consecutive days. The order of provision was random. The
details follow. Each day, I collected fresh plant materials (aboveground parts; flowering
buds and fruits were excluded) in the afternoon and soaked them in water to prevent
dehydration. Before feeding trials, voles were weighed. Plant materials were dabbed dry
with paper towel, and 15 g fresh plant materials were provided to each vole. To prevent
plants from dehydration during trials, the plant stems or leaf petioles were wrapped in
wet paper towels placed in a shallow dish. The voles were also given 15 g sweet potato
and 8 g oats to assure that the consumption of a particular plant species was not affected
by hunger (Kimball & Provenza, 2003), and that vole would not die from hunger (Ho,
2009). Water was provided ad libitum. Feeding trial lasted 12 hours, started from P.M.
8:00 and ended at A.M. 8:00.
After 12 hours, voles were weighed and moved to a new cage. Unconsumed plant
materials, sweet potato and oats were carefully sorted and collected, dabbed dry with
paper towels and weighted immediately. In order to control for the plant weight loss due
to dehydration, a control cage with tested plant materials only were established during
14
feeding trails and weighed after 12 hours. The consumption of plant materials was
calculated as below:
C= TO × (CL / CO) － TL
Co: Weight of plant material consumed
TO: Weight of plant material offered in the beginning of the trial
TL: Weight of plant material left at the end of the trial
CO: Weight of control plant material in the beginning of the trial
CL: Weight of control plant material at the end of the trial
The values of C were negative in some cases, probably because plant materials
used in feeding trials and control cages differed in the relative amount of leaves,
petioles, and stems. Nevertheless, the values were very small, I regarded it as no
consumption (zero). The value of consumption was divided by the square root of tested vole’s body weight to correct for the different metabolic requirements of animals of
different sizes (Grodzinski & Wunder, 1975). They will be referred to as standardized
palatability, hereafter.
Chemical analyses
Since it was not feasible to analyze all chemical compounds in plants, I chose to
15
analyze those compounds that were often associated with palatability, including dry
matter, crude protein, neutral detergent fibers (NDF), acid detergent fibers (ADF), acid
detergent lignin (ADL), ash, and total phenolics (Bergeron & Jodoin, 1987; Hartley et
al., 1995; Marquis & Batzli, 1989). Fresh plants samples were collected in March, July
and November in 2012 at the final day of feeding trials. Samples were put in plastic
bags to prevent water loss and brought back to laboratory within 24 hours, freeze-dried immediately, and stored in -20℃ freezers before chemical analyses.
I followed standard methods described in related literatures to perform chemical
analyses: dry matter (AOAC, 1984), ash (AOAC, 1984), crude protein (AOAC, 1984),
neutral detergent fibers (NDF) (van Soest et al., 1991), acid detergent fibers (ADF)
(Goering & van Soest, 1970), acid detergent lignin (ADL) (AOAC, 1984), and total
phenolics (Velioglu et al., 1998). Detail procedures for analyzing each chemical
compound are given in the Appendix 1. Nutritional contents (water, crude protein,
neutral detergent fibers, acid detergent fibers, acid detergent lignin, and ash) were
conducted in the laboratories of either Dr. Jih-Tay Hsu in the Department of Animal
Science and Technology, National Taiwan University (NTU) or Dr. Han-Tsung Wang
in the Department of Animal Science, Chinese Culture University (CCU). Total
phenolics was done in the laboratory of Dr. Shaw-Yhi Hwang in the Department of
16
Entomology, National Chung Hsing University (NCHU).
Toughness analyses
Since the leaves, petioles and stems of plants were provided at the same time
during trials, the measurements of toughness of different plant species depended on the
part of plant consumed by voles during trials. Based on the consumption pattern, I
measured the toughness of leaves in all tested plant species except Dicranum japonicum,
which I measured stems. The measurement procedures follow: I
stratified-random-sampled 10 stations, and collected three plants per species at each
station. Plants were put in plastic bags individually to prevent water loss, brought back
to laboratory immediately. Toughness was measured with a digital force gauge
(Chatillon ® force measurement, DFE II series). Each plant was measured only once,
the main vein was avoided during measurements. The values from the 30 plants were
averaged. A preliminary study (in July 2013) showed that the toughness measured in the
laboratory may increase slightly in all five plants (Table 22) likely due to water loss.
Because I offered clipped plants to voles in the feeding trials, I used the toughness
measured in the laboratory in the subsequent analyses.
17 Statistical analyses
I examined the normality and variance homogeneity of all data sets using
Shapiro-Wilk and Levene’s tests, respectively. Data sets that did not meet the
assumptions of parametric statistical analyses were properly transformed to meet the
assumptions. Otherwise, appropriate non-parametric statistical analyses or Markov
chain Monte Carlo methods for Generalized Linear Mixed Models (MCMCglmm)
would be applied.
First of all, I used a two-way ANOVA to examine the effects of sex and season
(both as fix factors) on body weights of voles in feeding trials. I tested if the diet
composition of voles changed with season by using a Chi-square test. The three main
food categories: plants, invertebrates and unknown were then examined separately using
the Kruskal-Wallis tests. I used MCMCglmm to run a two-way ANOVA to examine the
effects of season and plant species on standardized palatability. Year was also put in as
a random factor. I looked for the relationship between palatability and individual plant
attributes: chemical compounds, toughness and relative abundance. I pooled together
the data of standardized palatability and plant attributes from three seasons. Since the 7
chemical attributes measured were highly correlated with one another, I used a Principle
Component Analysis to produce new independent variables (PCs). I used simple linear
18
regression to examine the relationship between palatability and each PC separately. I
used MCMCglmm to run a one-way ANOVA to compare the toughness measured in the
laboratory and in the field. A two-way ANOVA by MCMCglmm was performed to
examine if toughness differed among plant species and seasons. I used a Pearson
correlation to examine the relationship between the ranking in diet and the ranking in
palatability. Statistical tests were performed by using the SAS 9.2 software, while
MCMCglmm was performed by the R studio 3.0.2 software. Differences were
considered statistically significant when p < 0.05.
19
Results
Field survey
Vegetation composition
The relative abundance of plant species was consistent among three seasons as shown by the percent coverage (Table 1). In general, Yushania niitakayamensis (玉山箭 竹) was the most abundant plants in three seasons, followed by Dicranum japonicum
(日本曲尾苔), Elatostema trilobulatum (裂葉樓梯草), Dryopsis transmorrisonensis
(玉山擬鱗毛蕨) and Anisliaea reflexa (玉山鬼督郵). Although the latter two species
were not recorded at sampling stations in March, 2012 (Table 1), field observation
indicated that they were still present. The two species were patchily distributed. Overall,
I recorded 46 species of plants in total, including 8 species of Bryophyta, 6 species of
Pteridophyta, 2 species of Gymnosperms and 30 species of Angiosperms (including 13
monocots and 17 dicots) (Table 2). I collected the tissues of 38 species for making
reference slides for the diet analyses.
Animals trapping
The numbers of voles that provided fecal samples in diet analysis or entered
feeding trials are presented in Table 3. Fecal samples from ten adult voles were
20
randomly selected for diet analyses in each season. Some adult voles were trapped
multiple times in different season/year. Although their fecal samples were collected
multiple times, I only selected one of the samples for diet analysis. Forty-seven adult
voles, 27 males and 20 females, were selected for feeding trials. Because of low
population sizes in some seasons, a few voles were reused in different seasons. The body weights of voles entered feeding trails was 34.8±2.9 in male and 30.83±2.64 in
female in March; 35.83±3.3 in male and 33.00±2.83 in female in July and 33.05±2.63 in
male and 30.55±2.77 in female in November (Fig. 2). There was a significant difference
between sexes in body weight (two-way ANOVA: F = 12.25; p = 0.001; Table 4), while
there was no difference among seasons (p = 0.07). Other than Taiwan field voles, I
captured many other vertebrates (Table 5) including 5 species of the order Rodentia:
Apodemus semotus (台灣森鼠), Niviventer culturatus (高山白腹鼠), Dremomys pernyi
owstoni (長吻松鼠) and Tamiops maritimus formosanus (條紋松鼠), 2 species of the
order Soricomorpha: Episoriculus fumidus (台灣煙尖鼠) and Anourosorex squamipes
yamashinai (山階氏鼩鼱), 2 species of the order Carnivora: Mustela sibirica taivana
(華南鼬鼠) and Mustela formosanus (台灣小黃鼠狼), and 2 bird species Garrulax
morrisonianus (金翼白眉) and Fulvetta formosana (褐頭花翼畫眉).
21 Diets of Taiwan field voles
Examining the relative importance of different food items in the vole’s diet, either
by area or count, over the three seasons, March, July, and November (Table 6-8), I
found that plant was the most important food (95.3%, 91.5% and 92.2%, respectively),
followed by unknown (3.1%, 5.9% and 6.8%, respectively) and insects (1.6%, 2.6% and
1.0%, respectively). The plant food in the diet was composed of 55.4% monocots
(53.3% was Yushania niitakayamensis), 33.7% moss (0.2% was Dicranum japonicum),
5.2% dicots (1.1 % and 4.1% were Ainsliaea reflexa and Elatostema trilobulatum,
respectively), and 0.4% of fern Dryopsis transmorrisonensis in March (Table 6). In July,
the plant diet was composed of 70.7% monocots (all Yushania niitakayamensis), 16.2%
moss (0.09% was Dicranum japonicum), 3.6% dicots (0.4% and 0.8% were Ainsliaea
reflexa and Elatostema trilobulatum, respectively), and 1.0% of fern Dryopsis
transmorrisonensis (Table 7). In November, the relative importance of monocots and
moss were very similar to those in July. However, Dicranum japonicum increased to
2.8%, dicots decreased to 0.7%. Both dicots, Ainsliaea reflexa and Elatostema
trilobulatum, and fern Dryopsis transmorrisonensi were not recorded in November
(Table 8).
Basically voles did forage selectively since in many plant species the proportion of
22
a species in diets deviated from its relative abundance, although the two was positively
correlated with each other (Fig. 3). Moreover, the diets of voles changed significantly
with seasons (Chi-square test: χ2 = 65.99, p < 0.001, Table 9). By examining plant,
invertebrate, and unknown separately (Table 10A), I found the seasonal difference was
mainly contributed by the ‘unknown’ group. There was no significant differences
among seasons in plant (Kruskal-Wallis test: U = 3.86, p = 0.14) and invertebrate (U =
0.66, p = 0.72, Table 10A) categories. The percentage of unknown food consumed was
significantly different among seasons (U = 6.04; p = 0.05). Among the plant food items
(Table 10B), I found there was significant differences among seasons in monocots (U =
5.86, p = 0.05) and moss (U = 7.51, p = 0.02). While there was no significant difference
in dicots (U = 2.33, p = 0.31).
Palatability of dominant plants
A two-way ANOVA performed with MCMCglmm examining the effects of season
and plant species on standardized palatability (Table 11 and 12) showed that there was a
significant interaction between plant species and season (Table 13). The interaction occurred because the palatability of Yushania niitakayamensis (玉山箭竹) varied
greatly among seasons, much higher in July than March and November (p = 0.05 in
23
March; p = 0.03 in November; Table 13, Fig. 4). Moreover, the palatability of different
plant species was significantly different (p < 0.001; Table 13, Fig. 4). In general,
Yushania niitakayamensis (玉山箭竹) was the most palatable plant to voles in three
seasons, followed by Elatostema trilobulatum (裂葉樓梯草), Anisliaea reflexa (玉山鬼 督郵), Dryopsis transmorrisonensis (玉山擬鱗毛蕨), and Dicranum japonicum (日本
曲尾苔) (Table 13 and Fig. 4).
Effects of plant attributes on palatability
Effects of chemical characteristics of plants on palatability
The values of the 7 chemical attributes, including dry matter, crude protein, NDF
(neutral detergent fiber), ADF (acid detergent fiber), ADL (acid detergent lignin), ash,
and total phenolics for each plant species in March, July, and November are presented in
Table 14-16, respectively. Because these attributes are highly correlated with one
another (Table 19), I performed a Principle Component Analysis (PCA) to describe the
overall chemical attribute of a plant (Table 17). The first three principle components
(PC1–PC3) were selected. Cumulatively, they explained over 93% of variation (Table
17). PC1 was significantly correlated with dry matter (＋), crude protein (－), NDF (＋)
and ADF (＋), ADL (＋), ash (－) and total phenolics (＋); PC2 was significantly
24
correlated with NDF (＋) and total phenolics (－); PC3 was significantly correlated
with crude protein (＋) only (Table 19). Among the 3 PCs, only PC3 was significantly
correlated with palatability (Simple linear regression: p = 0.04; R2 = 0.30; Table 20C
and Fig. 5) while PC1 and PC2 had no significant relationship with palatability (Simple
linear regression: p = 0.07 for PC1; p = 0.14 for PC2; Table 20A and Table 20B,
respectively). Therefore, high palatability was positively associated with crude protein
(Table 19).
Effects of physical characteristics of plants on palatability
In most plant species, toughness measured in the laboratory and field were similar
(two-way ANOVA by MCMCglmm: p = 0.72; Table 22), although the former was
slightly higher than the latter (Fig. 6). The difference approached significance only in Dicranum japonicum (日本曲尾苔) (two-way ANOVA by MCMCglmm: p = 0.05;
Table 22). I consistently used the toughness measured in the laboratory in all statistical
analyses. Toughness of different plant species that measured in three seasons were
shown in Table 21. There was a significant interaction between season and species (Table 23). The interaction occurred because the toughness of Dicranum japonicum (日 本曲尾苔) varied greatly among seasons, low in March and high in November
25
(two-way ANOVA by MCMCglmm: p < 0.001 in March; p < 0.001 in November; Table 23 and Fig. 7). Elatostema trilobulatum (裂葉樓梯草) had the lowest toughness among
all plant species, followed by Yushania niitakayamensis (玉山箭竹), Anisliaea reflexa
(玉山鬼督郵), Dryopsis transmorrisonensis (玉山擬鱗毛蕨), and Dicranum japonicum
(日本曲尾苔) (Table 23 and Fig. 7). There was a significant negative correlation
between palatability and toughness (Simple linear regression: p = 0.02; R2 = 0.34; Table
24 and Fig. 8), and suggested that high palatability was associated with low toughness.
Effects of abundance of plants on palatability
Relative abundance of plants was not correlated with standardized palatability,
although the result approached significance (Simple linear regression: p = 0.08; R2 =
0.22; Table 25 and Fig. 9).
Diet analyses and palatability of dominant plants
Vole diets come from plants with different relative abundance, while plant
palatability is measured under controlled amount of plant. The more abundant a plant in
the field, the more likely it will be encountered and consumed by voles. Thus, to
examine the relationship between diets and palatability, one needs to control for the
26
relative abundance. To do so, I performed a simple linear regression between the
percentage of different plants in diets and their relative abundance in the field (Simple
linear regression: p = 0.0001; R2 = 0.69; Table 26), and obtained the residuals of diets
after accounting for abundance. Next, I correlated standardized palatability with the
residuals. The results showed that there was a significant correlation between
standardized palatability and diets (Pearson correlation: p = 0.003; Fig. 10).
27
Discussion
The aim of this thesis is to understand the relationship between Taiwan field voles
and dominant herbaceous plants in the Taiwan fir-Taiwan hemlock forest by combining
diet analyses and palatability experiments. I first examined the hypothesis that voles
foraged selectively, and their diets changed with seasons. The results showed that voles
did not forage selectively based on the relationship between the proportion in diets and
abundance of five plant species (Fig. 3). Furthermore, there was a significant overall
seasonal effect in voles’ diets (Table 9). Particularly, the unknown food items, but not
plants or invertebrates (Table 10A), tended to be lower in March than July or November
(Table 6–8). Unknown food items were mainly composed of spore-like tissues. It
indicated that some food items associated with these spore-like tissues were quite
important for voles in July and November. Dividing plants into specific categories, I
found that the plant diets of voles were mainly monocots and moss. The consumption of
these two food items together made up 88–90% of vole diets. The percentages changed
with seasons as well (Table 10B). Voles consumed more moss in March (34% of diet),
compared to July (16%) and November (17%). The genus, Microtus, has been found to
prefer monocots over dicots in the field (López-Wilchis & Torres-Flores, 2007;
Lindroth & Batzli, 1984), although monocots were sometimes overestimated and dicots
28
were underestimated (Alipayo et al., 1992). Indeed, the abilities that herbivores digest
different plant species are different; resulting in quantifying the food fragments with
bias. Nonetheless, from the vegetation survey I found that the coverage of monocots
outmatched the coverage of dicots (Table 1); therefore, voles preferred to eat those that
were abundant in the environment. Moss seemed to be another important food resource
for voles in the forest. Previous studies have found that moss helped small mammals to
persist in winter and adapted to a wider range of environments since they were available
all year (Varner & Dearing, 2014). In this study, at least four species of moss were
recorded in voles’ diets (Table 6-8), the percentage of the unknown moss (M1), which was most likely Pleurozium schreberi (赤莖苔), remained the highest among all moss
species. Dicranum japonicum (日本曲尾苔) was included in feeding trial since it was
the most abundant moss from field observation and vegetation survey (Table 1).
However, the results of diet analyses showed that voles consumed more M1 than
Dicranum japonicum in the field. This indicates that voles showed selective foraging on
moss. Fir, being too tall, was not considered a food item available to voles. Nevertheless,
the fragments of fir epithelial cells were recorded once in diet analyses in March. Thus,
voles may consume fir seedlings. Since there was only recorded once; fir is not likely a
constant food resource to voles. On the other hand, fungi have been proposed to be an
29
important food resource for voles in the forest. Yeh (2012), at the same study site as
mine, found that fungi made up 25% of vole diets in growing and non-growing seasons
in forests. In my study, I also found that there were mycelia and spores of fungi in vole
feces (especially in July). However, fungi were not included in diet analyses since they
were too small to be properly quantified. Therefore, voles did consume fungi in the
forest. Further investigation is required to know what species and how much were
consumed by voles.
Second, I aimed to examine the hypothesis that different plants had different
palatability, and the palatability within the same species changed with seasons. The
results showed that different plants had significantly different palatability (Table 13).
Across all seasons, Yushania niitakayamensis (玉山箭竹) was the most palatable plants
to voles (Table 13 and Fig. 4). If herbivory pressure imposed by voles on a plant species depended on palatability, then Yushania niitakayamensis (玉山箭竹) would face the
highest herbivory pressure. Since there was significant seasonal variation in palatability
of Yushania niitakayamensis (Table 13 and Fig. 4), Yushania niitakayamensis likely
faced different herbivory pressure in different seasons. Despite the seasonal variation in
palatability; generally speaking, the palatability ranking of Yushania niitakayamensis
remained stable across seasons (Table 11 and Table 12). It was consistent with a study
30
investigating the palatability of 8 freshwater macrophyte species to a snail, Lymnaea
stagnalis (Elger & Barrat-Segretain, 2004).
Third, I examined how plant attributes affected the palatability of different plant
species by testing the hypothesis that high palatability was associated with high
nutrients and low digestion inhibitors. The results showed that only one plant chemical
stood out—crude protein had a positive effect on palatability. Protein has been proposed
as a nutrient which encouraged the consumption of food for herbivores (Cole & Batzli,
1979; Marquis & Batzli, 1989) since it is an important compound for animals to
synthenize vital substances. I did not find the effects of negative chemical constituents,
including fibers and phenolics. Fibers have been regarded as a negative compound since
it reduced digestibility of plants to herbivores (Marquis & Batzli, 1989). Total phenolics
are digestion inhibitors that herbivores avoided and deterred further consumption of
plants (Marquis & Batzli, 1989). However, there was no significant negative correlation
between either fibers or total phenolics and palatability in this study. I think it is likely
Taiwan field voles have adapted physiologically to tackle them. Several previous study
also suggested that herbivores were capable of adjusting to changing environments,
particularly the changes in food quality because morphological, physiological and
behavioral adaptations will take place (del Valle et al., 2006; Sassi et al., 2010;
31
Torregrossa & Dearing, 2009a). Therefore, it was possible that voles had adapted to
specific chemical compounds that persisted in plants for a long time. Lovegrove (2010)
compared the length of rodent intestines and found that herbivorous voles had large
caecums and colons which indicated that voles were able to process plants with high
fiber contents. Total phenolics could be divided into various secondary metabolites,
such as tannins, lignins and flavonoids (Kimball and Provenza, 2003), different
chemicals may have different effects on palatability (either levels or directions of
effects). More detailed analyses of voles’ abilities to cope with total phenolics are
suggested.
Another hypothesis I examined is that high palatability is associated with low
toughness. The results showed that not only different plants species had significantly
different toughness, there was a significant interaction between plant species and
seasons in toughness (Table 23 and Fig. 7). Low toughness of a plant was associated
with higher palatability (Table 24), which was consistent with previous findings (e.g.,
Pennings & Paul, 1992). Laca et al. (2001) proposed that the measurement of toughness
could be divided into tensile and shearing strength; accordingly, more detailed
measurement on toughness were recommended. I also examined the hypothesis that
high palatability is associated with high availability. The result showed that the relative
32
abundance was not significantly, though marginally (p=0.08), correlated with
palatability (Table 25), which was inconsistent with the findings of Ho (2009). I only
measured the palatability of 5 plant species. The small number of species, in
comparison with 13 species in Ho (2009) may render the results insignificant.
Particularly, there were many species of mosses in the study area (Table 2). Results of diet analyses suggested that at least 4 other moss species were part of voles’ diets.
Dicranum japonicum (日本曲尾苔) was the only one included simply because its
relative abundance was the highest. If more moss species were included in the
palatability experiments, probably the relationship between palatability and abundance
would be much clearer.
In general, I found plant attributes (especially chemical and physical characteristics)
affected palatability when they were examined separately. In order to know how all
these plant attributes affected palatability altogether, I used cluster analysis and AIC
model selection as tools. The cluster analysis included all plant characteristics (7
chemical compounds and toughness) except abundance to examine the similarity among
plant species in each season. The results showed that Yushania niitakayamensis (玉山箭 竹), Anisliaea reflexa (玉山鬼督郵) and Elatostema trilobulatum (裂葉樓梯草) were
close to one another (Fig. 11), which were palatable plants (Fig. 4). Dryopsis
33
transmorrisonensis (玉山擬鱗毛蕨) and Dicranum japonicum (日本曲尾苔) were
close each other (Fig. 11) (except for Dicranum japonicum in November) and they were
unpalatable plant species (Fig. 4). The result of the cluster analysis indicated that
particular combinations of plant characteristics would have higher palatability. I used
AIC model selection to understand what combinations of plant characteristics
contributed more to palatability. The results also showed that almost every plant
characteristic was included (Table 27). Results of cluster analysis and model selection
suggested that to completely understand how plant attributes affect palatability, every
characteristic of plant should be considered.
Lastly, I examined the hypothesis that the diets of voles reflected the palatability of
plants, and found that standardized diets and palatability matched perfectly (Fig. 10).
Many other studies also found that diets were quite similar to palatability (reviewed in
Batzli, 1985), although some studies found otherwise (e.g., Lantová & Lanta, 2008).
The inconsistency is probably because diets were not standardized by relative
abundance. Palatability is examined by feeding trials. They are non-choice experiments
examining the intrinsic characteristics of a plant species. Whereas, diet analysis
examines the consequence of a decision-making process in which herbivores encounter
various plant species with drastically different relative abundance in the field. Therefore,
34
by standardizing the diets with relative abundance and then correlated with palatability
was a better way to link these two parameters.
Both the results of diet analyses and palatability showed that Yushania niitakayamensis (玉山箭竹) was the most important food resource to Taiwan field voles
across seasons in the fir forest. In alpine meadows, Yushania niitakayamensis (玉山箭 竹) was also a critical food resource for voles (Ho, 2009). In addition, voles have been
found that they prefer to consume different parts of Yushania niitakayamensis in
different seasons (Yeh, 2010). Although voles seemed to be able to dwell in these two
habitats all year, I found that forests were likely to be refuges in fall and winter for
Taiwan field voles for two reasons: (1) Ho (2009) showed that in alpine meadows, the
values of standardized palatability of Yushania niitakayamensis were 0.65, 0.75 and
0.44 in March, July and November, respectively. Note that there was a sharp decrease in
November. In contrast, in this study I found the values of standardized palatability of Yushania niitakayamensis (玉山箭竹) were 0.74, 0.97 and 0.78 in March, July and
November, respectively. There was no apparent decrease in November (Fig. 4). It
indicated that the major food resource, Yushania niitakayamensis became unpalatable in
November in meadows, but not in forests. It probably encouraged voles to move into
forests in fall and winter to consume Yushania niitakayamensis. (2) In the study area,
35
Yeh (2012) found that vole population sizes were generally higher in meadows than in
forests, except in fall and winter when population sizes were higher in forests than in
meadows. Besides, number of vole captured in this study was also highest in November,
lowest in July (Table 3). Based on these two reasons, I propose that forests are liable to
be important habitats for Taiwan field voles in fall and winter. Nevertheless, it is
possible that the population of voles in these two habitats are independent to each other,
which means the increase and decrease of vole population are not due to the movement
of voles between habitats. Accordingly, further research on whether voles will move to
forests for wintering are required.
In conclusion, the results of diet analyses and palatability both showed that Yushania niitakayamensis (玉山箭竹) remained the most important food resource for
Taiwan field voles in forests. Diets of voles and palatability of particular plants changed
with seasons. The plant attributes did affect the palatability of plants to voles, especially
crude protein and toughness. Results of diets and palatability were consistent with each
other. Moreover, it seemed that forests were important habitats for voles in fall and
winter since Yushania niitakayamensis was palatable all year round in forests. These
results had shown that voles could impose strong herbivory pressure on vegetation in
the forest, especially Yushania niitakayamensis.
36
References
吳海音、施金德。2010。代表性生態系經營管理之合歡山高海拔生態系長期生態 研究網計畫（二）-生態系物種間交互作用調查及保育措施評估。太魯閣國 家公園管理處。
呂孟栖。1991。合歡山地區台灣高山田鼠(Microtus kikuchii)的生態與生殖研究。東 海大學生物學研究所碩士論文.
林宗以、李玲玲。2003。台灣水鹿(Cervus unicolor swinhoei)的食性研究。行政院 農業委員會林務局保育研究系列第 91-3 號。
洪麗慧。2002。花蓮地區兩共域鼠種-赤背條鼠與月鼠之食性研究。國立東華大學 自然資源管理研究所碩士論文。
張簡琳玟。1997。也是國寶-台灣高山田鼠。自然保育季刊, 17, 46-49。
陳玉峰。1998。台灣植被誌-第三卷：亞高山冷杉林帶與高地草原(上)。台北市: 前 衛出版社。
賀菡芝。2009。合歡山地區台灣高山田鼠與高山草原植物的交互作用：從優勢草 本植物的營養與豐富度來探討。國立臺灣大學生態學與演化生物學研究所 碩士論文。
葉威廷。2012。運用穩定同位素探討合歡山地區兩個小鼠群聚的食物資源區隔。
國立臺灣大學生態學與演化生物學研究所碩士論文。
葉素含。2010。合歡山地區台灣高山田鼠啃食作用對玉山箭竹生長之影響。國立 臺灣大學生態學與演化生物學研究所碩士論文。
Alipayo, D., Valdez, R., Holechek, J. L., and Cardenas, M. 1992. Evaluation of microhistological analysis for determining ruminant diet botanical composition.
Journal of Range Management, 45, 148-152.
AOAC. 1984. Official method of analysis. 14th ed. Washington D. C., USA:
Associations of Official Analytical Chemist.
Barthelmess, E. L. 2001. The effects of tannin and protein on food preference in eastern grey squirrels. Ethology Ecology & Evolution, 13(2), 115-132.
Batzli, G. O. 1985. Nutrition. In Biology of the New World Miroctus. (R. Tamarin Ed.
Vol. 8): American Socitety of Mammalogists Publication.
Batzli, G. O., & Lesieutre, C. 1991. The influence of high quality food on habitat use by arctic microtine rodents. Oikos, 60, 299-306.
Batzli, G. O., and Pitelka, F. A. . 1983. Nutritional ecology of microtine rodents: food habits of Lemmings near Barrowm Alaska. Journal of Mammalogy, 64(4), 648-655.
Bergeron, J. M., and Jodoin, L. 1987. Defining “high quality” food resources of
37
herbivores: the case for meadow voles (Microtus pennsylvanicus). Oecologia, 71, 510-517.
Boyle, Sarah A., Zartman, Charles E., Spironello, Wilson R., and Smith, Andrew T.
2012. Implications of habitat fragmentation on the diet of bearded saki monkeys in central Amazonian forest. Journal of Mammalogy, 93(4), 959-976.
Bucyanayandi, J. D., and Bergeron, J. M. 1990. Effects of food quality on feeding patterns of meadow voles (Microtus pennsylvanicus) along a community gradient. Journal of Mammalogy, 71(3), 390-396.
Cole, F. R., and Batzli, G. O. 1979. Nutritional and population-dynamics of the prairie vole, Microtus ochrogaster, in central Illinois. Journal of Animal Ecology 48, 455-470.
del Valle, J. C., Busch, C., and Lopez Mananes, A. A. 2006. Phenotypic plasticity in response to low quality diet in the South American omnivorous rodent Akodon azarae (Rodentia: Sigmodontinae). Comp Biochem Physiol A Mol Integr Physiol, 145(3), 397-405.
Elger, A., and Barrat-Segretain, M. H. 2004. Plant palatability can be inferred from a single-date feeding trial. Functional Ecology, 18, 483-488.
Goering, H. K., and van Soest, P. J. 1970. Forage fiber analyses (Apparatus reagents, procedures, and some applications) Washington DC, USA: ARS USDA.
Goldberg, M., Tabroff, N. R., and Tamarin, R. H. 1980. Nutrient variation in beach grass in relation to beach vole feeding. Ecology, 61, 1029-1033.
Grodzinski, W., and Wunder, B. A. 1975. Ecological energetics of small mammals.
Cambridge, UK: Cambridge University Press.
Hanley, Mick E., Lamont, Byron B., Fairbanks, Meredith M., and Rafferty, Christine M.
2007. Plant structural traits and their role in anti-herbivore defence. Perspectives in Plant Ecology, Evolution and Systematics, 8(4), 157-178.
Hansson, L. 1970. Methods of morphological diet micro-analysis in rodent. Oikos, 21, 255-266.
Hartley, S. E., Nelson, K. , and Gorman, M. 1995. The effect of fertilizer and shading on plant chemical composition and palatability to Orkney voles, Microtus arvalis orcadensis. Oikos, 72, 79-87.
Johnson, M. K., Wofford, H., and Pearson, H. A. 1983. Microhistological technique for food habits analyses. U.S. Department of Agriculture Research Paper, SO-199.
Kimball, B. A., and Provenza, F. D. 2003. Chemical defense and mammalian herbivores.
USDA. National Wildlife Research Center-Staff Publications Paper 236.
38
Laca, E. A., Shipley, L. A., and Reid, E. D. 2001. Structural anti-quality characteristics of range and pasture. Journal of Range Management, 54, 413-419.
Lantová, Petra, and Lanta, Vojtěch. 2008. Food selection in Microtus arvalis: the role of plant functional traits. Ecological Research, 24(4), 831-838.
Lindroth, R. L., and Batzli, G. O. 1984. Food habits of the meadow vole (Microtus Pennsylvanicus) in bluegrass and prairie habitats. Journal of Mammalogy, 65(4), 600-606.
Litvaitis, J. A. 2000. Investigating food habits of terrestrial vertebrates (L. Boitani and T.
K. Fuller Eds.). New York.
López-Wilchis, Ricardo, and Torres-Flores, Jose W. 2007. Diet of the Jalapan Pine Vole (Microtus Quasiater) in Mature Mountain Cloud Forest. Journal of Mammalogy, 88(2), 515-518.
Lovegrove, B. G. 2010. The allometry of rodent intestines. J Comp Physiol B, 180(5), 741-755.
Marquis, R. J., and Batzli, G. O. 1989. Influence of chemical factors on palatability of forage to voles. Journal of Mammalogy, 70, 503-511.
Massey, F. P., Ennos, A. R., and Hartley, S. E. 2007. Herbivore specific induction of silica-based plant defenses. Oecologia, 152.
Olff, H., and Ritchie, Mark. E. 1998. Effects of herbivores on plants diversity. Trends in Ecology and Evolution, 13, 261-265.
Pennings, S. C., and Paul, V. J. 1992. Effects of plant toughness, calcification, and chemistry on herbivory by Dolabella auricularia. Ecology, 73, 1606-1619.
Provenza, F. D., Villalba, J. J., Dziba, L. E., Atwood, S. B., and Banner, R. E. 2003.
Linking herbivore experience, varied diets, and plant biochemical diversity.
Small Ruminant Research, 49(3), 257-274.
Rezsutek, Michael J., and Cameron, Guy N. 2011. Diet selection and plant nutritional quality in Attwater's pocket gopher (Geomys attwateri). Mammalian Biology 76(4), 428-435.
Sassi, Paola L., Caviedes-Vidal, Enrique, Anton, Rosa, and Bozinovic, Francisco. 2010.
Plasticity in food assimilation, retention time and coprophagy allow herbivorous cavies (Microcavia australis) to cope with low food quality in the Monte desert.
Comparative Biochemistry and Physiology Part A: Molecular & Integrative Physiology, 155(3), 378-382.
Scheidel, U, and Bruelheide, H. 1999. Selective slug grazing on montane meadow plants. Journal of Ecology, 87, 828-838.
Takahashi, A., and Shimada, T. 2008. Selective consumption of acorns by the Japanese
