f
a"!
f 5

I." I P.7'.J‘

I'm“; ‘;
t‘ffﬁfa’f—v
-’ ff? {53 1X .
vs ‘0’ ' x‘ r

.52

V 'R‘ .
”1 V-
' i‘ "34.’;;,-_u..
f-‘lﬁmv J, f:
L .
x

v": K5?) ‘5
. 2’.“

,9

H

u

'2
‘va

- . ,
‘5”
3%...

"N’ ‘n- ~
”i cgﬁ'ﬁ-vam
«F "ma-35a”:
‘ .V‘.
'~”‘”‘*‘ ‘ “In;
A.“ "f V ,4 ‘ ‘ E2331:
‘. WU ‘ ‘ _ ‘ a?
k “3% CL: — 3; : - exé”:
w .

V dd):- ‘
6‘13? ~ l
”*E‘Tx’f-‘FLxﬁ ‘. .

1‘- ‘7‘? ‘1 ~

g
m; V“ \
th‘;‘...~..¢
' ‘-

W‘g- .

’zﬁizﬁ‘s:

;
in

‘ ékﬁiilg‘jn
r fared-r .
‘n~ $4- ». a
.‘tﬂr'hﬁf‘ j ‘

‘mr
~ «.1 ‘ R}

,

1:. , . N; . f.

rt. - 1 '.a 1-7. ‘ ‘Wf'o‘ : '.. “3?" ' a . $1,:

31%v1 :1) r . . ”13'?“ . z. ‘ ‘
"A" "J "‘7‘ ~ . Ev"; .
' 333% . . ‘ ,4
N

” “2w . ‘
&}€?%§m%§ggkhz :9," . 1
"; ‘: ﬁg“ 44%
W 3, ﬁ%* :3. :2? {a
3; " 7%. '33:“
" wiéw‘ﬁz"

‘; f
1.1" 2

I t‘ k
A " .- 4, h :V'rb,
N , ‘2‘“!
3-42; 1?: “$5555"
gii'év .

1

 

UNIVERSITY LIBRARIE

lllllllllllillllllllMilli ll'lllllll

3 1293 01029 884

Helllill

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the
dissertation entitled

ECOLOGICAL IMPLICATIONS
OF DIET-INDUCED THERMOGENESIS

IN THE PRAIRIE VOLE, MIQBOTUS OQERQQASTER

presented by

Terry Martin Trier

has been accepted towards fulﬁllment '
of the requirements for

Ph.D. degreein Zoology

 

 

6

ajor ' fessor

Date 1/12/94

 

MSU is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY
MIohIgan State
Untverslty

 

 

 

PLACE IN RETURN BOX to roman this ohockout Iron your rooord.
To AVOID FINES rotum on or baton m duo.

DATE DUE DATE DUE DATE DUE

itg‘f) ’ﬁf‘ﬁ I

w.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU In An Afﬂrmottvo AotIoNEmol Opportunity Imtttmlon
' W

”3-9.!

ECOLOGICAL IMPLICATIONS
OF DIET-INDUCED TI-IERMOGENESIS
IN THE PRAIRIE VOLE, MICROTUS OCHROGASIER
By

Terry Martin Trier

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Zoology

1994

ABSTRACT
ECOLOGICAL IMPLICATIONS
OF DIET-INDUCED THERMOGENESIS
IN THE PRAIRIE VOLE, MICROTUS OCHROGASTER
By

Terry Martin Trier

Diet-induced thermogenesis (DIT) occurs in laboratory rats consuming either a
cafeteria or a low-protein diet. Researchers suggest that DIT regulates weight by
oxidizing excess food intake. However, DIT may enhance nutrient uptake from a diet
that is nutritionally unbalanced. By selective oxidation of excess nutrients, the diet
can become more balanced and uptake of one or more limiting nutrients could
increase. Also, a hyperphagic response to an unbalanced diet can increase nutrient
intake, and excess nonessential nutrients can be oxidized via DIT.

In the wild, nitrogen can be in limited supply for many animals, especially
herbivores. Ecological studies of herbivores indicate that many herbivores may
periodically consume low-nitrogen diets which can limit growth and reproduction.
Therefore, natural selection should favor the evolution of DIT in herbivores to
enhance nitrogen uptake from diets that are typically low in nitrogen. To test the
hypothesis that DIT occurs in small, mammalian herbivores and may augment
nitrogen uptake, wild herbivorous prairie voles were studied to see if 1) DIT occurs
in voles consuming a low-protein diet, 2) the response is mediated by brown adipose

tissue (BAT), the putative effector organ of DIT, 3) hyperphagia occurs in response to

low-protein feeding, and 4) body protein is conserved and lipid deposition is low,
suggesting regulation of body protein and lipid levels during low—protein fwding.
Energy balance studies of voles fed semipuriﬁed diets showed that low-protein-
fed voles (2.5 % casein diet) had lower energy efﬁciencies and higher metabolic rates
per unit of metabolizable energy intake than high-protein-fed voles (15 % casien diet),
which is indicative of DIT. Daily food consumption increased for voles on low-
protein diets and decreased for voles on high-protein diets over a three-week period.
Both high- and low-protein groups deposited dry tissue (lipid, protein, and other) in
the same relative proportions, although total dry tissue depositon was greater for the
high-protein group for each constituent. GDP binding activity did not increase in the
BAT of low-protein-fed voles, suggesting that BAT may not be involved in low-

protein DIT observed in voles.

To my mother,
Betty Rebecca Rhoades

iv

ACKNOWLEDGEMENTS

My Sincere thanks to the members of my committee for their support and
understanding. I especially want to thank Dale Romsos for his invaluable advice and
technical support and for allowing me to use his laboratory, equipment, and supplies.
Also, special thanks go to Nancy Soloman, who provided the prairie voles used to
start my colony, Hsiao-Ling Chen, who helped with the GDP binding assay, and
Virginia Vega-Vargas, who helped with the Kjeldahl analysis.

TABLE OF CONTENTS

LIST OF TABLES .....................................

LIST OF FIGURES .....................................

INTRODUCTION ......................................

CHAPTER I. DIET-INDUCED THERMOGENESIS ................

THE CAFETERIA DIET AND DIT .........................
THE CAFETERIA DIET AND BAT-MEDIATED DIT ...............
Low-PROTEIN DIETS AND DIT ..........................
Nutrient Composition and Thermogenesis ...............
Recent DIT Research Using Low-Protein Diets ............
CONTROVERSY AND DIT ..............................
Heat Increment of Feeding and DIT ..................
The Case Against DIT ..........................
ECOLOGICAL FRAMEWORK FOR DIT RESEARCH ....................

CHAPTER II. ECOLOGICAL FRAMEWORK ....................

THE HERBIVORE TROPHIC LEVEL ........................
RELATIVE FOOD SHORTAGE ............................
Change in Plant Quality and Its Inﬂuence on Herbivores ......
TEMPORAL HETEROGENEITY ...................

SECONDARY PLANT CHEMICALS ................

STRUCTURAL DEFENSE ......................

INDUCED DEFENSE ........................

SUMMARY ......................................

Herbivore Response to 3 Variable Food Supply ...........
INFLUENCE OF SUPPLEMENTAL FOOD .............

FORAGE QUALITY .........................
COMPENSATORY MECHANISMS .................

Summary ................ * ..................
METABOLIC RATE AND THE NUTRITIONAL ECOLOGY OF HERBIVORES . .
Food Habits and Energetics in Mammals ...............

vi

Page

ix

10
12
14
15
16
16
18
19

23

24
29
30
3O
35
49
50
52
53
53
56
57
62
63

CHAPTER II (continued). Page

Food Habits and Energetic Efﬁciencies in Invertebrates ....... 71
Thermogenic Mechanisms and Dietary Protein ............ 73
CONCLUSION ..................................... 74
CHAPTER III. EXPERIMENTS AND ANALYSIS ................. 75
INTRODUCTION ................................... 75
EXPERIMENT 1 .................................... 79
Introduction ................................. 79
Materials and Methods .......................... 79
DETERMINATION OF BODY ENERGY .............. 82

Results .................................... 83
EXPERIMENT 2 .................................... 85
Introduction ................................. 85
Materials and Methods .......................... 86
Results .................................... 87
EXPERIMENT 3 .................................... 89
Introduction ................................. 89
Materials and Methods .......................... 90
Results .................................... 91
EXPERIMENT 4 .................................... 94
Introduction ................................. 94
Materials and Methods .......................... 94
Results .................................... 96
EXPERIMENT 5 .................................... 96
Introduction ................................. 96
Materials and Methods .......................... 98
Results .................................... 98
EXPERIMENT 6 .................................... 99
Introduction ................................. 99
Materials and Methods .......................... 100
Results .................................... 100
EXPERIMENT 7 .................................... 101
Introduction ................................. 101
Materials and Methods .......................... 102
Results .................................... 107

FOOD CONSUMPTION ....................... 107

BODY COMPOSITION ....................... 109

vii

CHAPTER HI (continued). Page

NITROGEN BALANCE ....................... 113

DISCUSSION ..................................... 113
Introduction ................................. 113
Thermogenesis ............................... 1 14

Brown Adipose Tissue ........................... 120

Food Consumption ............................. 125

Body Composition ............................. 129

Nitrogen Balance .............................. 130

CHAPTER IV. CONCLUSIONS AND RECOMMENDATIONS ......... 132
LIST OF REFERENCES .................................. 144

viii

LIST OF TABLES
Table Page

1.1 Inﬂuence of low-protein diets on parameters associated with an in-
crease in thermogenesis .............................. 15

2.1 Inﬂuence of supplemental food on reproduction in hares and micro-

tines .......................................... 54
2.2 Inﬂuence of supplemental food on growth in hares and microtines . . . . 56
3.1 Composition of diets, expt. 1 ........................... 80
3.2 Metabolic rate, energy intake, and retention, expt. 1 ............. 83
3.3 Composition of diets, expt. 2 ........................... 86
3.4 Metabolic rate, energy intake, and retention, expt. 2 ............. 88
3.5 Composition of diets, expt. 3 ........................... 90
3.6 Metabolic rate, energy intake, and retention, expt. 3 ............. 92
3.7 Total and weight-specific food consumption, expt. 7 ............. 105
3. 8 Post hoc polynomial contrasts, expt. 7 ..................... 107
3.9 Body composition of voles fed 2.5%, 5%, and 15% casein diets,

expt. 7 ........................................ 108
3.10 Statistical analysis of selected body composition data listed in table

3.9, expt. 7 ..................................... 109
3.11 % Composition of ﬁnal wet and dry weights, expt. 7 ............ 110
3.12 Gross nitrogen balance, expt. 7 ......................... 112

ix

LIST OF FIGURES

Figure

1.1 Lipid metabolism and the proton conductance pathway in BAT ......
1.2 Heat increments of variable protein diets ....................
2.1 Thomas’ model ...................................

2.2 Relationship between food nitrogen concentration and the efﬁciency of
conversion of ingested food (ECI) by assorted invertebrate herbivores

3.1 Distribution of Microtus ochrogaster ......................
3.2 The fate of ingested energy in an organism ..................

3.3 Regression of the ADMR of low- and high- protein groups on the co-
variate ME to test for homogeneity of Slopes in an analysis of covari-

ance (ANCOVA) design ..............................
3.4 Mean daily growth of weanling voles, expt. 2 .................
3.5 Mean daily growth of weanling voles, expt. 3 .................

3.6 General design of feeders used in expt. 7 ....................
3.7 Mean daily food consumption, expt. 7 .....................
3.8 Weight-speciﬁc mean daily food consumption, expt. 7 ............
3.9 Growth of weanling voles, expt. 7 ........................
3.10 Efﬁciency of energy retention, expt. 3 .....................
3.11 GDP binding, expts. 4 and 6 ...........................

3.12 Mean total food consumption for 21 days of expt. 7 ............

14

72

75

84

89

93

103

106

106

111

117

122

Figure Page
3.13 Mean total food consumption for the ﬁrst 10 days of expt. 7‘ ........ 126
3.14 Mean weight-speciﬁc total food consumption for 21 days of expt. 7 . . . 126

3.15 Mean weight-speciﬁc total food consumption for the ﬁrst 10 days

of expt. 7 ....................................... 126
4.1 Efﬁciency of enrgy retention, expt. 3 ...................... 141
4.2 GDP binding, expts. 4 and 6 ........................... 141
4.3 Gross nitrogen efﬁciencies, expt. 7 ....................... 141
4.4 Growth of weanling voles, expt. 7 ........................ 141
4.5 Mean daily food consumption, expt. 7 ..................... 142
4.6 Total mean food intake per vole for 10 and 21 days during expt. 7 . . . . 142

4.7 Data are for weight changes in tissue constituents over a 21-day feed-
ing period during expt. 7 ............................. 142

4.8 Relative proportions of dry tissue constituents for voles at the end of
expt. 7 ........................................ 143

xi

INTRODUCTION

The main purpose of my research is to examine ecological implications of a
physiological phenomenon known as diet-induced thermogenesis (DIT). Since 1979
when Rothwell and Stock published a paper in Nature where they proposed that diet
and brown fat metabolism are linked, DIT has been the focus of intense research by
nutritionists and biochemists whose main interest has been to determine the effects of
DIT on obesity in humans. However, the ecological implications of this phenomenon
for wild animals have yet to be investigated. DIT is a phenomenon that may have
considerable relevance to the nutritional ecology and bioenergetics of wild animals,
and therefore Should be examined within an ecological framework.

AS the amount of research on DIT continues to grow, researchers are begin-
ning to make claims that have not been explored within the context of ecology and
evolutionary biology. Rothwell and Stock (1986a) believe that DIT is an adaptive
trait and Should be labeled as such. Rothwell and Stock ( 1986a) wrote “. . . we have
opted for the terms obligatory and adaptive diet-induced thermogenesis to differentiate
between the two inﬂuences of the net availability of metabolizable energy.” Similar-
ly, according to Trayhum and Milner (1987), “Diet-induced thermogenesis (currently
the most widely used expression) and luxusconsumption are terms which are primarily

used to describe an adaptive energy-dissipative process.”

2
According to the claims of Rothwell and Stock (1986a) and Trayhum and

Milner (1987), DIT is an adaptive process. However, one of the problems associated
with labeling a term as adaptive is that the word ‘adaptive’ has several meanings in
the sciences, and therefore can lead to ambiguity. The following statement by Gould
and Lewontin (1979) summarizes this problem:
‘Adaptation’ “the good ﬁt of organisms to their environment” can

occur at three hierarchical levels with different causes. It is unfortunate that

our language has focused on the common result and called all three phenome-

na ‘adaptation’: the differences in process have been obscured and evolution-

ists have often been misled to extend the Darwinian mode to the other two

levels as well. First, we have what physiologists call ‘adaptation’: the pheno-

typic plasticity that permits organisms to mould their form to prevailing

circumstances during ontogeny. Human ‘adaptations’ to high altitude fall into

this category (while others, like resistance of sickling heterozygotes to malar-

ia, are genetic and Darwinian). Physiological adaptations are not heritable,

though the capacity to develop them presumably is. Secondly, we have a

‘heritable’ form of non-Darwinian adaptation in humans (and, in rudimentary

ways, in a few other advanced social species): cultural adaptation (with

heritability imposed by learning). Much confused drinking in human

sociobiology arises from a failure to distinguish this mode from Darwinian

adaptation based on genetic variation. Finally, we have adaptation arising

from the conventional Darwinian mechanism of selection upon genetic varia-

tion. The mere existence of a good ﬁt between organism and environment is

insufﬁcient evidence for inferring the action of natural selection.

3

Since the term adaptation can be interpreted in at least three different ways,
the question arises as to the precise meaning of the term ‘adaptive DIT.’ While most
human nutritionists are primarily interested in the study of physiological adaptations
(Waterlow 1985), it is clear that Rothwell and Stock (1981a) are labeling adaptive
DIT as a Darwinian adaptation. Rothwell and Stock (1981a) argue that DIT is
adaptive because it preserves leanness in wild animals. According to Rothwell and
Stock (1981a), obesity in wild animals would impair their locomotor activity and
reduce their ability to hunt for food or avoid capture. Also, Rothwell and Stock
( 1981a) argue that obesity could reduce reproductive capacity in wild animals.
Sterility, complications associated with pregnancy, and high perinatal mortality are
traits that are commonly associated with obesity in laboratory rodents. Therefore,
according to Rothwell and Stock ( 1981a), those animals that can preserve leanness by
DIT and thus avoid obesity, would have higher fecundity and increased survivorship
in the wild.

Although Rothwell and Stock believe DI’T is a Darwinian adaptation, it is not
clear that other researchers who study DIT use the term in similar fashion, thus
creating confusion. Moreover, by labeling DIT adaptive, researchers are introducing
a nonneutral term (‘adaptive DIT’) into the literature. Putting nonneutral labels on
terms can often lead to bias and stiﬂe alternative research. If we call DIT adaptive
without further research to understand its role in nature (or lack of one), then the
term ‘adaptive DIT’ may become ﬁxed in the literature and could hinder our ability to

think objectively about the occurrence and function of DIT. Therefore, although DIT

4

may be an adaptive trait, until we know more about it, putting an adaptive label on
DIT should probably be avoided.

If we label DIT as adaptive, then we should be able to state clearly and with
little ambiguity, how it is adaptive. However, several adaptive claims have been
made for DIT, and it is not clear which, if any, are justiﬁed. HimmS-Hagen (1986)
summarized some of the more signiﬁcant claims as follows: “The function of
diet-induced thermogenesis in the rat is suggested to be 3-fold. First, it contributes,
together with cold-induced nonshivering thermogenesis, to the maintenance of eu—
thermia in the cold. Second, because much of the excess food is used for increased
thermogenesis in brown adipose tissue, the development of obesity due to hyperphagia
is attenuated and mobility of the animal is preserved. Third, intake of nutrients can
be increased when the diet is low in protein so as to achieve an adequate protein
intake without the consequent development of obesity due to the increased energy
intake.” These proposals are interesting and deserve serious consideration. Howev-
er, as far as I know, none of the proposals above has ever been fully examined within
an ecological or evolutionary framework, where speciﬁc selective pressures known to
occur in nature which might favor the evolution of DIT are rigorously examined.
Therefore, from an ecological perspective, there is little support for labeling DIT as
adaptive.

If DIT is to be considered an adaptive trait, its adaptive function should be
readily understood within an ecological context. If we assume DIT is an adaptation
that evolved by natural selection, then we should be able to identify the selective pres-

sures that led to the evolution of DIT. However, the key factors in nature that may

5
have favored the evolution of DIT have not been rigorously deﬁned. Rothwell and

Stock (1981a) have discussed the evolution of DIT as an adaptive trait that prevents
obesity in animals, but their treatment of this topic was not extensive and, in their
own words, their arguments are based on “apocryphal sources and ex cathedra
statements” and characterized by “gross extrapolations and partisan view. ” Their
intention, however, was not to provide a rigorous ecological framework for the
evolution of DIT, but to stimulate discussion (Rothwell and Stock 1981a).

It is not the goal of my research to prove or disprove DIT as an adaptive trait.
That may require many years of research. But it is my goal to take tentative Steps in
that direction. According to Mayr (1983), “. . . when one attempts to explain the
features of something that is the product of evolution, one must attempt to reconstruct
the evolutionary history of this feature. This can be done only by inference. ” There-
fore, one of the goals of my research is to identify ecological factors that might have
favored the evolution of DIT. An ecological framework discussing the factors that
may have led to a selective advantage for the evolution of DIT is covered extensively
in Chapter II.

The ecological framework presented in Chapter 11, however, represents a
change from the focus on DIT as a weight regulatory mechanism, to one where DIT
is postulated to be an adaptation to enhance nutrient uptake in animals that may be
limited by food quality in the wild. According to this hypothesis, the main function
of DI’I‘ is not weight regulation, although perhaps that might be a proximate effect.
Instead, the hypothesized functional signiﬁcance of DIT is to permit wild animals that

normally feed on food low in quality to ingest large amounts of food and burn off

 

 

6

excess calories while distilling essential nutrients. More speciﬁcally, for many
herbivores, nitrogen may be the main limiting nutrient in their forage in the wild
(Chapter H), and DIT could function to increase nitrogen intake.

According to Mayr (1983), one of the ﬁrst Steps in demonstrating that a trait is
adaptive “ . . . consists in establishing a tentative correlation between a trait and a
feature of the environment . . .” Thus, for this research, a correlation must be
established between low-nitrogen food intake by herbivores and a DIT response,
before DIT can be seriously considered as an adaptive trait that enhances nutrient
uptake in wild herbivores. This research begins by predicting that, if DIT is adaptive
in aiding nitrogen uptake, then it should occur in wild herbivores that may be limited
by low-nitrogen forage.

To investigate the occurrence of DIT in wild mammals, prairie voles (Microtus
ochrogaster) were used in my research. Prairie voles were chosen because they are
primarily herbivorous and consume large amounts of low-nitrogen foodstuffs to satisfy
their nutritional needs in the wild (Batzli 1985). If DIT is an adaptive trait that
enhances nitrogen intake in small herbivores such as the prairie vole, we would
expect to see a DIT response when prairie voles consume diets low in protein. Thus,
much of my research involves measuring energy efﬁciencies and metabolic rates of
voles consuming diets that contain different levels of protein.

Also, food consumption rates were investigated to see if voles exhibit hyper-
phagia when fed low-protein diets. A hyperphagic response and a concomitant DIT
response to a low-protein fwding diet would suggest that voles oxidize excess caloric

intake and increase nitrogen uptake by distillation of nitrogen from the increased food

 

 

 

7

intake. Gross nitrogen efﬁciency, body tissue change (lipid, protein, water and
‘other’), and biochemical changes in brown fat, a heat-producing tissue thought to be
the primary effector organ of NST, were also studied in voles consuming diets with
variable protein content. Their relevance to this research is discussed in Chapter III.
In general, my research was undertaken to see if DIT could be observed in
wild mammals and understood within the framework of their nutritional ecology. The
rest of the dissertation is divided by chapters containing the following information:
Chapter I is a brief review of DIT research and Chapter H presents an evolutionary
and ecologically-based raison d ’étre for DIT. A detailed discussion of my research
goals and speciﬁc hypotheses tested are presented in Chapter III; research protocol,
experimental results, and a discussion of the results are also covered in Chapter III.

Concluding remarks and suggestions for further research are given in Chapter IV.

CHAPTER I

DIET-INDUCED THERMOGENESIS

Measurement of energy balance can be used to determine the efﬁciency of use
of metabolizable energy (ME) available in the food that an animal consumes. Energy
balance is the difference between the intake and expenditure of energy by an organism
over time (Hervey and Tobin 1982). In order for an animal to achieve zero energy
balance, energy intake and expenditure must be equal. If an animal increases the
amount of heat it produces for a given amount of ME, then efﬁciency and energy
balance are lowered. Conversely, if an animal reduces heat production and stores
more energy in the body for a given amount of ME, then efﬁciency is increased and
energy balance becomes more positive. Diet-induced thermogenesis (DIT), as deﬁned
by Rothwell and Stock (1986a), refers to a physiological phenomenon that regulates
energy balance for an animal by increasing its metabolic rate in response to an
increase in food intake above the level required for maintenance. The result is that an
animal can preserve its body weight over a range of food intake levels by burning
‘Surplus’ calories when food consumption is in excess.

The idea of homeostatic regulation of energy balance and body weight extends
back to the turn of the century when the German physiologists Voit and Neumann
suggested that the body could eliminate excess energy ingestion directly as heat (a

phenomenon called luxuskonsumption) (Voit 1901, and Neumann 1902, as cited by

9
Rothwell and Stock 1987a). Neumann experimented on himself and showed that he

could maintain his body weight at a constant level for an extended period of time,
even though his daily caloric intake varied. Gulick (1922) performed a Similar

experiment and achieved comparable results.

THE CAFETERIA DIET AND DIT

Current interest in the DIT phenomenon stems from research by Miller and
Payne (1962), who showed that pigs on a low-protein diet fed ad libitum consumed
ﬁve times more energy than pigs on an energy-restricted high-protein diet, with no
difference in weight gain. The experiment was repeated in 1980 by Gurr et al., who
reported that, compared to the high-protein group, there was a two-fold increase in
energy consumption and expenditure for pigs fed low-protein diets. Rothwell and
Stock (1979) began using a highly palatable ‘cafeteria’ or ‘junk food’ diet in order to
stimulate excess food consumption (hyperphagia) in rats. This allowed them to
compare two groups of animals fed ad libitum. Different levels of food intake
occurred as a result of the hyperphagic response exhibited by the group fed the
cafeteria diet, thus creating an increase in caloric intake in the cafeteria—fed (CAF)
group. Their results, published in Nature in 1979, provided the impetus for a wide
variety of research by nutritionists and biochemists into the existence and causal
mechanism of DIT.

Rothwell and Stock (1979) found that rats consuming a cafeteria diet ingested
almost twice the energy as controls fed a stock diet, but efﬁciency of energy retention

was much lower. Furthermore, CAF rats had an increased sensitivity to norepineph-

10

Tine-induced thermogenesis, which was determined by measuring oxygen consumption
(170,) after an injection of norepinephrine (NE); V0, declined when CAF rats were
treated with the B-adrenergic blocking agent propranolol. The control group exhibited
only a slight response to treatments. These results showed a striking similarity to
those observed in animals exhibiting cold-induced non-shivering thermogenesis (NST)
(see Jansky 1973), a phenomenon thought to be mediated primarily by the sympathetic
nervous system and effected by brown adipose tissue (BAT) (Foster and Frydman
1978). This led Rothwell and Stock (1979) to propose that DIT elicited by cafeteria

feeding is a result of sympathetic activation of thermogenesis in BAT.

THE CAFETERIA DIET AND BAT-MEDIATED DIT

Numerous studies have been conducted to test Rothwell and Stock’s claim.
Evidence supporting their proposal is extensive and varied. BAT hypertrophy,
hyperplasia, and an increase in protein content have been observed in CAF rats
(Himms-Hagen, Triandaﬁllou, and Gwilliam 1981; Rothwell, Saville, and Stock
1982a; Goldberg and Morgan 1983; Rothwell, Stock, and Warwick 1983b). The
speciﬁc mitochondrial protein thermogenin increases in concentration in CAF rats.
Thermogenin is a 32 kD protein thought to be responsible for the uncoupling of ATP
production from electron transport, resulting in the dissipation of ME as heat via a
proton—conductance pathway across the inner mitochondrial membrane (Figure 1.1).
Use of a cDNA probe provides evidence that there is also a concomitant increase in
thermogenin mRNA (Falcou et a1. 1985). In vitro binding of GDP to BAT

mitochondrial protein, considered a good parameter for measuring the activity of the

11
proton-conductance pathway (Nicholls 1979), is elevated in CAF rats (Nedergaard,

Raasmaja, and Cannon 1984; Brooks, Rothwell, and Stock 1980).

 

 

 

 

 

 

Noradrmaline
I
plmamhnne
CAMP ——» hn' aces
“‘0
«mum»

 

 

 

+ FFA
H amm- (-)

 

 

 

 

 

 

ACam ’+ M+
I i moron
cououcrmce
ACoA —> p-orddatlon PATHWAY

 

 

 

Figure 1.1. Lipid metabolism and the proton conductance pathway in BAT.
ACoA: acetyl coenzyme A; ACam: Acetyl-camitine; ATP, AMP, and GDP:
purine nucleotides; CAMP: cyclic adenosine-S’-monophosphate; FFA: free fatty
acids; TCA: tricarboxylic acid cycle; TG: triglyceride droplet; 32K: thermogenin
uncoupling protein. (Adapted from Bukowiecki 1986, and Nicholls and Rial 1988.)

An increase in the rate of norepinephrine turnover in BAT resulting from
cafeteria fwding reinforces the idea of sympathetic regulation of DIT (Young et a1.

1982). Mitochondrial enzyme activities (cytochrome oxidase and a-glycerophosphate

12

dehydrogenase) in BAT of CAP rats are more than twice the levels measured in
controls (Brooks et al. 1980), indicating an increase in cellular respiration in this
tissue. Additionally, many studies of CAF rats report a hyperphagic response and
simultaneous decline in energetic efﬁciency (see Rothwell and Stock 1986a, for a
review). Thus, the ﬁndings of Rothwell and Stock (1979) are supported quantitatively
by evidence gathered at the molecular, cellular, and organismal level.

Further evidence supporting the claim that BAT-mediated DIT is involved in
energy balance comes from experiments using genetically obese laboratory rodents.
GDP binding to BAT mitochondrial protein is reduced in obese (ob/ob) mice, and
binding does not increase upon cold exposure (Himms-Hagen and Desautels 1978).
GDP binding is also reduced in adult, genetically obese Zucker rats (fa/fa), and both
fa/fa rats and ob/ob mice show lower thermogenin concentrations than their lean
siblings (Ashwell et al. 1985). Also, lesions in the ventromedial region of the
hypothalamus cause obesity in rats (Hogan, Coscina, and Himms-Hagen 1982; Vander
Tuig, Kemer, and Romsos 1985). This is signiﬁcant because the hypothalamus is
implicated in the control of thermogenesis in BAT via sympathetic innervation

(Landsberg and Young 1983).

Low-PROTEIN DIETS AND DIT
The term ‘diet-induced thermogenesis’ also has been applied to an observed
decrease in the efﬁciency of energy retention in response to consumption of a low-
protein diet. In some instances, an increase in food consumption and concomitant

decrease in efﬁciency for animals on a low-protein diet have been recorded (Miller

13

and Payne 1962; Gurr et al. 1980). In those experiments, animals on low-protein
diets were fed ad libitum, with a mixture of ingredients (protein, fat, and
carbohydrates) sufﬁcient to maintain body weight; those in the high-protein group
were on an energy-restricted, weight-maintenance diet.

In other experiments, where animals were fed ad libitum for both low- and
high-protein diets, food consumption for low-protein-fed animals was the same or,
more often, lower than those fed high—protein diets, although weight-speciﬁc food
consumption was usually higher (Tulp et al. 1979; Donald, Pitts, and Pohl 1981;
Rothwell, Stock, and Tyzbir 1982b, 1983a; Kevonian, Vander Tuig, and Romsos
1984; Rothwell and Stock 1987b). One notable exception, though, was a study by
Stirling and Stock (1968) who found that rats fed low-protein diets for 20 days
consumed 69 percent more energy than their counterparts fed high-protein rations.

The results of the studies mentioned above indicate that when both low- and
high-protein diets are fed ad libitum, DIT can be elicited in animals that are not
hyperphagic, but rather hypophagic (based on gross food consumption). Since CAF
rats are given a choice of foods, the cafeteria diet has been criticized for being
uncontrolled in terms of nutrient composition (Moore 1987; see Rothwell and Stock
1988, for reply). Moore (1987) postulated that the DIT phenomenon is due primarily
to the consumption of a diet that is unbalanced in nutrient composition and has
proposed that for future energy balance studies of DIT, palatable, nutrient-controlled

diets should be used instead of cafeteria diets.

14

 

 

15 I I l I 1

 

 

 

E
I
Ero- i
N
5
55' II
S
u:
o I I I I I

0 10 20 30 40 50 60
i PROTEIN IN DIET

 

 

 

Figure 1.2. Heat increments of variable protein diets. (Data from Hamilton 1939a.)

Nutrient Composition and Thermogenesis

That unbalanced diets cause an increase in thermogenesis was noted by Rubner
(1902, as cited by Glick 1987), who called the phenomenon “speciﬁc dynamic effect”
because the speciﬁc response to a protein meal was higher than for a carbohydrate or
fat meal. In 1934, Mitchell stated that it was not protein per se that caused the
increase in metabolic rate; food consumption, in general, has a thermic effect.
Instead, Mitchell declared, it is a ‘balanced diet’ that produces the smallest amount of
heat. This claim was substantiated by Hamilton (1939a) using laboratory rats.
Hamilton demonstrated that diets with low- and high-protein content had the greatest

thermic effect, while those with intermediate protein levels had the least (Figure 1.2).

15
Kleiber (1945) reviewed this topic and concluded that a dietary deﬁciency of any

nutrient (vitamins, minerals, protein, etc.) would alter energy metabolism and stated
that “ . . . a diet is deﬁcient in any nutrient whose addition decreases the calorigenic
effect of the ration. ” Discussing low-protein diets, Kleiber Stated that “One of the

homeostatic reactions of an animal against the effect of a low protein diet may be an

increase in the combustion of carbohydrates and fats. ”

Recent DIT Research Using Low-Protein Diets
A variety of studies (summarized in Table 1.1) indicate that a thermogenic
response resulting from the consumption of a low-protein diet is similar in kind to that
invoked by cafeteria feeding. Some studies implicate BAT as an effector organ and
the sympathetic nervous system as a mediator of low-protein DIT. Thus, DIT can be

induced by cafeteria feeding and by consumption of a low-protein diet, but it is not

Table 1.1. Inﬂuence of low-protein diets on parameters associated with an increase in thermogenesis.

 

 

Response to low-protein diet Reference

Increase in thermic response' and decrease in Tulp et al. 1979; Rothwell et al. 1983a; Kevonian

efﬁciency of energy retention et al. 1984; Swick and Swick 1986; Rothwell and
Stock 1987c

 

Increase in GDP binding to BAT mitochondrial Rothwell et al. 1983a; Swick and Gribskov 1983;
protein Swick and Swick 1986; Rothwell and Stock
19870; Swick and Henningﬁeld 1989

 

Increase in V0, response to NE Stirling and Stock 1968; Kevonian et al. 1984;
Rothwell and Stock 1987c

 

Increase in NE turnover in BAT Kevonian et al. 1984; Vander Tuig and Romsos
1984; Young et al. 1985

 

Decline in V0, in response to B—adrenergic blockade Rothwell et al. 1983a

 

Increase in plasma thyroid hormones2 Tulp et al. 1979

1Refers to heat production per unit metabolizable energy.
2This hormonal change my affect the thermogenic responsiveness of BAT (Kopecky et al. 1986).

16

clear whether these are the same or different phenomena. Research into DIT has
been motivated primarily by its applicability to energy balance in humans and
relevance to weight control and obesity. However, since it can be induced under
hypophagic conditions, the physiological signiﬁcance of DIT may be multifaceted,
and not related to energy balance alone. My research centers on DIT induced by

low—protein diets and its ecological implications for animals living in the wild.

CONTROVERSY AND DIT

The idea that there is an increase in heat production that accompanies an
increase in food intake is well established in the nutrition literature. The monogas-
tric, or Rubner school of nutritionists called this phenomenon speciﬁc dynamic effect,
while the Kellner, or ruminant school called it heat increment of feeding (HIF)
(Webster 1981). Both terms refer approximately to the same event, differing mainly
in the length of time the measurements of heat production are taken. The measure-
ments are usually taken right after the consumption of a meal and, therefore, the
thermogenic response has also been called the “thermic effect of a meal” by Gar-

row (1978).

Heat Increment of Feeding and DIT
The mechanism of HIF (HIF is used because it is a more neutral term than
speciﬁc dynamic effect) continues to be a fertile topic for debate in the nutrition
literature, and the discussion is now further complicated by the phenomenon of DIT,
which may or may not be related to HIF (Glick, Teague, and Bray 1981; Glick

1987). Historically, HIF has been considered an acute response to normal food

l7

consumption; DIT is regarded as a chronic response that dissipates excess food
consumption resulting from long-term overeating (Glick 1987). Mechanisms for HIF
have been ascribed to a variety of processes (see Kleiber 1975a, for a historical
review). Webster (1983) delineates the essential (but not necessarily only)
contributors to HIF in the following statement: “. . . they must include the energy
costs of ingestion and digestion of food, the work of active absorption and secretion
of substances across the gut wall, increments in H1 in the food processing tissues such
as the gut wall and the liver associated with increasing inﬂow of substrates and the
costs of synthesis (turnover and net accretion) in body tissues, muscle, fat, mammary
gland, etc. All of these things must be considered as inevitable contributors to HIF. ”
The difference between HIF and DIT is that the amount of heat that is produced by
DIT is considered to be much greater than what can be accounted for by the processes
described by Webster.

Because the multitude of terms designating dietary heat production in animals
has led to some confusion, Rothwell and Stock ( 1986a) proposed that the terms
obligatory and adaptive DIT be used to distinguish between the unavoidable (obligato-
ry) costs of metabolizing foodstuffs, e. g., those described by Webster (1983) above,
and nonessential (adaptive) thermogenesis, which presumably dissipates excess
energy. Adaptive DIT represents heat production that cannot be accounted for by
metabolic work associated with processing of food. This terminology is convenient
because, in some ways, adaptive DIT is a more neutral term than speciﬁc dynamic

effect (referring to the speciﬁc effect of protein) or luxuskonsumption (referring to

 

1H refers to metabolic heat production (Webster 1983).

18
hyperphagic consumption). However, as I have suggested in the Introduction, there

are good reasons for avoiding this type of characterization, because what appears to
be an unbiased term in some respects, is far from neutral to ecologists and evolution-
ary biologists who expect adaptive claims to be cast within an ecological and evolu-
tionary framework, and supported by research involving organisms representative of
wild populations. Further controversy exists because there are many nutritionists who
dispute the existence of adaptive DIT. Therefore, as long as there is a major and
continuing dispute among scientists about DIT, it is important that terminology be

neutral.

The Case Against DIT

There are a substantial number of studies that have failed to conﬁrm the
occurrence of DIT, and this has sparked a major debate among nutritionists.
Research by Fuller (1983) and McCracken and McAllister (1984) failed to conﬁrm
the ﬁndings of Miller and Payne (1962) and Gurr et al. (1980) of DIT in the pig.
Similarly, a variety of studies involving cafeteria feeding of rats have also reported
negative results for DIT (McCracken 1975a; Tobin, Armitage, and Hervey 1981;
Armitage et al. 1981; Barr and McCracken 1982; Bestley et al. 1982; McCracken and
Barr 1982). DIT effected by a low-protein diet has also been questioned (McCracken
1975b), although in another study by McCracken (1976), an increase in energy
expenditure and lower efﬁciency was shown for rats fed a low-protein diet.

The controversy surrounding DIT has generated a series of highly polemical

discussions in the literature (Hervey and Tobin 1981, and Rothwell and Stock 1981b;

19
Hervey and Tobin 1983, and Rothwell and Stock 1983; McCracken and Barr 1985,

and Rothwell and Stock 1985a; Moore 1987, and Rothwell and Stock 1988). Clearly,
the issue is far from resolution, and adopting a dogmatic, siege mentality has done
little to enhance an open-minded discussion. However, this debate has stimulated
energy balance research and it provides some of the impetus for my own research.
DIT is a concept that can, and Should, be tested within an ecological framework,

before it can be conclusively dismissed or wholly embraced.

ECOLOGICAL FRAMEWORK FOR DIT RESEARCH

DIT has been investigated in only a small number of species, most of which
are bred speciﬁcally for laboratory research and can no longer be considered wild.
The laboratory rat and mouse, both of which have been selectively bred within the
stenothermal environment of a laboratory colony, are not necessarily fair representa-
tives of natural populations of small mammals. Both rodents are from the same
family, Muridae, and therefore are closely related. If we are to increase our
knowledge about DIT, we Should make an attempt to broaden our understanding of
the occurrence of this phenomenon in other species and families of organisms.

If DIT is an adaptive process, we should be able to make predictions about its
relevance to wild animals and then test them within an ecological framework. For
instance, Rothwell and Stock (1985b) hypothesized that, since DIT is an adaptation to
prevent obesity caused by excess food consumption, and NST is an adaptation to
preserve homeothermy in the cold, then organisms living in environments where high

temperatures and persistent food shortages are normal, may Show a complete absence

20

of thermogenic capacity. Rothwell and Stock suggested that some desert rodents
(e.g. , spiny mouse and tuco tuco) may represent examples of this phenomenon, Since
they quickly develop obesity when ample food is available (Rothwell and Stock
1985b). More research, however, is needed to substantiate these claims.

Rothwell and Stock (1985c) used an ecologically-based argument to link DIT
to the presence of BAT and the high capacity for thermogenesis observed in the
common marmoset (Callithrix jacchus). Rothwell and Stock observed a 63 percent
increase in oxygen consumption in marmosets injected with noradrenaline. They also
found large BAT depots that showed a high degree of binding to GDP. Rothwell and
Stock argued that, since marmosets evolved and normally live in a warm climate, it is
unlikely that the high thermogenic capacity shown by marmosets is due to
thermoregulatory NST. They suggested that, since marmosets primarily consume
low-protein fruits, and since low-protein diets are known to activate DIT in rats, it is
possible that the large BAT depots and high thermogenic capacity in marmosets are
more related to their diet than to thermoregulatory demands. Rothwell and Stock
proposed that DIT in marmosets could permit ingestion of large quantities of low-
protein food while dissipating excess energy intake, thus resulting in ‘concentrating’
dietary protein (Rothwell and Stock 1985c).

Because the use of wild animals in DIT research has been limited, more
research of this type is needed. The full range of experimental techniques used to
investigate DIT in rats and mice has yet to be implemented in the study of wild
species. More studies of DIT involvement in different species are needed, and their

results Should be examined within the context of each organism’s nutritional ecology

21

and thermal environment. An increase in the use of wild species will allow
researchers to compare DIT responses (or parameters associated with DIT) between
different species. This information can then be analyzed within the context of the
ecology of each species, and researchers will be better able to observe any trends in
responses that may be ecologically related. In this way, DIT responses viewed within
the context of trophic levels, habitat quality, thermal conditions, or other ecological
factors, can be compared for Similarities and differences. Without this type of
research, our understanding of DIT, and its existence or relation to the ecology and
evolution of different species, will remain minimal.

My research focuses on DIT induced by low-protein intake and its relationship
to the nutritional ecology of herbivores. If DIT evolved as an adaptive trait by
natural selection, then an examination of the ecology of an organism exhibiting DIT
could provide evidence for the selective pressures that led to the evolution of this
trait. In the next chapter, an ecological framework is presented Showing that for
many herbivores, high-quality food is a limiting resource in the environment. A lack
of high-quality food can exert selective pressures that favor traits that enhance the
ability to increase nutrient intake.

For many herbivores, the quality of their food supply is determined primarily
by the amount of metabolizable nitrogen it contains. The evidence presented in
Chapter 11 suggests that a food supply low in nitrogen can have a negative inﬂuence
on herbivore growth and fecundity, and therefore could be selective for traits such as
DIT that may enhance the use of available nitrogen. Since the question of whether

animal populations can be limited by resources is a perennial topic of controversy in

22

population and community ecology, it is discussed in depth. If animal populations
cannot be limited by food resources, then it is unlikely that there would be sufﬁcient
selection pressure for the evolution of a trait or traits that enhance the use of food

resources .

CHAPTER II

ECOLOGICAL FRAMEWORK

Of the three mechanisms mentioned in the Introduction proposed by
Himms-Hagen (1986) as possible functions for DIT, I chose the third proposal to
investigate as a possible adaptive trait for a wild mammal. Himms-Hagen suggested
that a small mammal could consume more energy than is required for daily
maintenance, and DIT may function to optimize protein intake by providing a
pathway through which excess energy or nonessential non-protein energy can be
oxidized. An interesting feature of this hypothesis is that it implies that mammals
have evolved a metabolic pathway to ‘waste’ energy. This notion is contrary to
assumptions made for many optimality models of foraging strategy, in which it is
presumed that animals function to minimize energy costs and maximize energy
efﬁciency (Krebs and Davies 1987). However, some animals (e. g. , herbivores) may
be limited more by the quality (nutrient content) of their diet than the quantity (energy
content) (White 1978). Herbivores facing Shortages of high-quality nitrogenous foods
in the environment may be forced to consume large quantities of low-protein food to
maintain nitrogen balance. During growth and gestation, when protein synthesis is at
its maximum, the ability to get enough nitrogen in the diet could be critical (Mattson
1980). The main hypothesis of my research, then, is that DIT is an intrinsic physio-

logical response occurring in small, herbivorous mammals, which allows them to

23

24

increase nitrogen intake when dietary protein is low by oxidation of excess energy
intake. The result is that essential dietary nitrogen is distilled from a bulky, low-
nitrogen food supply. The following is a review of the research and discussion of the

ecological ideas that are relevant to this hypothesis.

THE HERBIVORE TROPHIC LEVEL

If DIT evolved as an adaptive trait in herbivores to facilitate the maintenance
of nitrogen balance, then resources consumed by herbivores must be limited in
metabolizable nitrogen to exert selective pressure for the trait to evolve and be
maintained. Evidence supporting the claim that nitrogen is a limiting resource for
herbivores has been reviewed several times (McNeil and Southwood 197 8; White
1978; Mattson 1980). The genesis of much of this research can be traced back to
ideas in population ecology concerning the regulation of the size of natural
populations. One school of ecologists proposed that density-dependent factors (factors
that are inﬂuenced by population size) such as predation and competition tend to
stabilize populations at some equilibrium level (Nicholson 1933). Others argued that
density-independent factors (factors that are independent of population size), such as
those created by changing weather conditions, exert the greatest inﬂuence on
population size by limiting population growth to periods of favorable environmental
and climatic conditions (Andrewartha and Birch 1954).

Many researchers who supported the idea of density-independence saw
resources as limiting. They viewed the natural world as unpredictable and, at times,

inhospitable, where highly variable extrinsic inﬂuences such as weather controlled

25
primary productivity and inﬂuenced herbivore food availability and population size.

As a consequence, food abundance and population size for subsequent trophic levels
in the food chain were also affected. This type of population regulation was termed
“bottom up” regulation, since population size at all trophic levels was seen to be
controlled by productivity at the lowest trophic level (plants) (White 1978).

A different view presented by Hairston, Smith, and Slobodkin in a controver-
sial paper written in 1960, outlined a set of ideas that became known as the HSS
hypotheses. In brief, HSS proposed the following:

1. “Populations of producers, carnivores, and decomposers are limited by
their respective resources in the classical density-dependent fashion.

2. lnterspeciﬁc competition must necessarily exist among the members of
each of these three trophic levels.

3. Herbivores are seldom food-limited, appear most often to be predator-
limited, and therefore are not likely to compete for common resourc-
es. ”

Their third point, that herbivores are seldom food-limited, is a conclusion based on
the deductive premise that, since plants generally appear abundant, herbivores must be
incapable of depleting them. Therefore, depletion of plants by herbivory is prevented
by predators that limit the population size of herbivores to a level that has no appre-
ciable impact on plant populations.

The “top down” regulation hypothesis, where populations levels are assumed

to be controlled by the inﬂuence of higher trophic levels on lower ones, was heavily
criticized by Murdoch (1966), and Ehrlich and Birch (1967), who claimed that HSS

failed to consider alternative hypotheses. They claimed that herbivore populations

may be regulated by plants through a relative shortage of food brought on by a spatial

26
and temporal variability in plant distribution and quality. They also suggested that the

evolution of spines and secondary plant compounds can render plants inedible and
thus reduce resource availability. For the next twenty years, resource-based ideas of
population ecology dominated research into plant-herbivore interactions, especially
insect herbivore research. The HSS view continued to inﬂuence competition and
predation theory. However, HSS failed as a whole to achieve dominance (Oksanen
1988).

Recently, there has been a renewed interest in HSS. Oksanen et al. (1981)
proposed that trophic structure is linked to physical factors and primary productivity
(referred to as the CF model after Fretwell 1977 and Oksanen et al. 1981). Accord-
ing to OF, as primary productivity increases, the length of the food chain increases.
Food chains can vary in length continuously rather than discretely, depending on the
number of organisms present in each trophic level. For instance, in a barren habitat,
the food chain length is zero; as plants begin to grow, the food chain length increases
incrementally. A food chain of length one is established with the appearance of the
ﬁrst herbivore. Food chain length two occurs when herbivores begin to impact the
population size of the plant community. The food chain lengthens continuously to
three when predators begin to inﬂuence the herbivore population; at food chain length
four, a top—level predator regulates the population of the herbivore predator. In this
model, HSS is viewed as depicting an ecosystem with a food chain length of three.
However, in the OF model the herbivore population is now interacting dynamically
with plant and predator populations, and the entire trophic structure is ultimately

dependent on primary productivity.

27
The OF model predicts that in plant communities with low productivity,

herbivore population size will be low and plant populations will be limited by the
physical environment. As plant productivity increases, herbivore numbers will
increase and eventually play a regulatory role in primary productivity. As the
herbivore population increases, so does the carnivore population until a dynamic food
chain with three trophic levels is established (Fretwell 1987; Oksanen 1988).

Arditi, Ginzburg, and Akcakaya (1991) have criticized the OP model on the
basis that it is too simplifying in its approach to the dynamics of predator-prey
interactions. The Arditi et al. (1991) model of predator-prey interaction predicts
equilibrium states where the densities of both predator and prey rise proportionately
as primary productivity increases, which is counter to the OF model. While the OF
model predicts that no two contiguous trophic levels will Show a Simultaneous
increase, Arditi et al. cite numerous studies showing that herbivore and predator
biomass covary with an increase in plant productivity. This view is supported by
McNaughton et al. (1989), who reviewed 51 studies of a wide variety of ecosystems,
including those with three or more trophic levels, and found that herbivore biomass
increases with primary productivity. Interestingly, Moen and Oksanen (1991)
reanalyzed the same data and came to a similar conclusion.

The preceding discussion illustrates that although current ecological thought on
trophic dynamics has not reached a consensus, it is still compatible with resource
limitation theory. Furthermore, resource limitation theory is the only ecological
theory that can account for the evolution of traits that are unique to herbivores. If

plant resources are unlimited, as proposed by HSS, it is difﬁcult to imagine how

28

selective pressure by predators can account for such obviously herbivorous traits as
rumination, caecal digestion, or lophodont dentition. However, it is more likely that
both plants and predators have shaped the evolutionary history of herbivores. For
instance, the evolution of large body size in grazing ungulates may be a deterrent to
predators as well as reduce the energetic cost of transport in migrational foraging.
Similarly, the long neck of the giraffe may aid in visually locating predators and assist
the giraffe in foraging on resources inaccessible to smaller herbivores.

One of the basic assumptions of my research is that resource limitation is a
major factor in the evolution of many herbivore traits. Charles Elton (1936) once
wrote “Animals are not always struggling for existence, but when they do begin, they
Spend the greater part of their lives eating. Feeding is such a universal and common-
place business that we are inclined to forget its importance. The primary driving
force of all animals is the necessity of ﬁnding the right type of food and enough of it.
Food is the bunting question in animal society, and the whole structure and activities
of the community are dependent upon questions of food-supply. ” Elton’s emphasis on
“the right type of food” has been echoed by others (McNeil and Southwood 1978;
White 1978; Mattson 1980) who claim that the right type of food means having
enough metabolizable nitrogen in the diet to nourish the young and sustain growth.
The following is a review and analysis of ideas and research related to relative food

shortages in nature.

29

RELATNE FOOD SHORTAGE

The concept of relative food shortage is important in understanding the
selective pressures that can lead to the evolution of low-protein DIT. In this view,
gross food energy is not a limiting resource for herbivores. As HSS pointed out,
energy may be plentiful for herbivores. However, even if there is an abundant supply
of energy available for herbivores, the nutrient quality of that food supply can change.
Actual food quality2 of plants is highly variable and is dependent on a variety of
factors such as seasonal change, ontogenetic stages, secondary plant chemicals, and
plant-herbivore interactions. These factors, as well as others, can affect the spatial
and temporal availability of a nutritionally-adequate food supply for a population of
herbivores. Herbivores are often faced with what appears to be unlimited food;
however, the quality of the food—especially the amount of metabolizable nitro-
gen—may be so low that herbivores are unable to sustain normal growth and develop-
ment. The effects can be especially devastating for young, growing animals that are
highly dependent on dietary nitrogen to sustain development. Therefore, the interac-
tion of herbivores with a food source that may vary temporally in nitrogen content,
could involve major selective forces for traits that maximize nitrogen intake in what
may be a nutritionally inadequate environment (White 1978). In an environment
where energy is plentiful but the supply of nitrogen is variable, low-protein DIT could
have evolved to aid in nitrogen intake during relative food shortages. Thus, this topic

is covered in depth, since it is important to establish that a relative Shortage of high-

2High—quality food is deﬁned here as plants high in usable protein and low in
deleterious chemicals, after White (1978), and Bergeron and Jodoin (1987).

30

nitrogen food exists in the wild, if we are to believe that DIT evolved as a response

to such a shortage.

Change in Plant Quality and Its Inﬂuence on Herbivores
TEMPORAL HETEROGENEITY

The nitrogen content of different plant tissues can vary from 0.03 to 7.0
percent (0.2 to 44 percent protein)3 of dry weight, with seeds and new tissue growth
having the highest levels. For most plants, the highest total nitrogen content occurs
during seasonal periods of rapid growth. As the rate of plant growth declines over
time, nitrogen concentrations are reduced and reach their lowest levels during
senescence and abscission. Seasonal decline in nitrogen content can be as high as 80
percent of maximum (Mattson 1980).

A good example illustrating seasonal changes in crude protein content in plants
occurs in blue grama grass (Bouteloua gracilis), an important forage species in the
shortgrass ecosystem of the Great Plains (Uresk and Sims 1975). Seasonal lows of 5
percent crude protein were recorded during March, when blue grama is dormant;
crude protein increased to a high of 10 percent in June, during ﬂowering, and
progressively declined to 6 percent in December, when the grass again entered
dormancy. Similar changes have been observed for broadleaf tree leaves (Feeny

1970; Ricklefs and Matthew 1982; Schultz, Nothnagle, and Baldwin 1982), mixed

3To convert nitrogen content to protein content, multiply nitrogen content by
the gravimetric conversion factor 6.25, which is the one most commonly used. This
Value can vary, depending on the type of protein or protein mixture being studied.
For instance, the factor 6.38 is used to convert the nitrogen content of milk to protein
Content (Helrich 1990). The chief protein in milk is the phosphoprotein casein.

31
forbs and grasses (Karasov 1985), the needles of conifers (Watt 1989) and phloem

and xylem sap (Mittler 1953; Horseﬁeld 1977; Barlow and Randolph 1978).

Seasonal variation in nitrogen content in plants is partially a result of greater
accumulation of cell wall material (lignin, cellulose, hemicellulose, and silica)
compared to cell contents during growth, which dilutes the concentration of plant
nitrogen; later, the translocation of nutrients out of the leaves during the onset of
senescence causes a further reduction in nitrogen. The reduction of nitrogen content
in the leaves of evergreen species, though, is primarily inﬂuenced by cell wall
dilution, since evergreens minimize translocation of their nutrients, retaining them in
their leaves during the winter (Chapin 1980). These processes are often driven by
climatic changes in rainfall and the length of temperate growing seasons (Bell 1970;
Phillipson 1975; Coe, Cumming, and Phillipson 1976). Freeze-thaw or wetting-
drying cycles inﬂuence leaching and the breakdown of litter and organic compounds,
which determine nutrient concentrations in soil solution. Soil nutrient availability
affects nutrient absorption by plants, which eventually inﬂuences the nitrogen
concentration in primary production (Chapin 1980).

The impact of dry seasons on plant nitrogen and herbivores was examined by
Sinclair (1974; 1975; 1979), who documented the relationship between African
ungulates and the quality and quantity of available food in the East African Serengeti
grasslands. Sinclair reported that during the dry season from July to October, the

crude protein level for grasses dropped from eight percent to between one and three

32

percent4 (Sinclair 1975). Dry season crude protein was below the four to ﬁve percent
level required to maintain adult body weight in African ungulates. Such low levels
caused a loss in fat reserves, low efﬁciency of conversion of plant protein into animal
protein (0.4 percent), and an increase in mortality rate during the dry season.

Grazing time was similar during both the wet and dry seasons, indicating that
ungulates were consuming the same amount of food throughout the year, but were
unable to maintain nitrogen balance due to the low protein content of the food supply
during the dry season (Sinclair 1974).

As earlier mentioned, plant nitrogen content is strongly inﬂuenced by growth
stage. New growth is high in protein but protein content declines rapidly with age
(Uresk and Sims 1975; Mattson 1980; Jones and Wilson 1987; Watt 1987). In
tropical rainforest trees, the young leaves and shoots highly favored by folivorous
primates can contain more than 20 percent crude protein, while mature leaves have
only between 7 and 15 percent crude protein (Hladik 1978). Newly ﬂushed foliage of
the Eucalyptus has a much higher nitrogen content than older growth; as nitrogen
levels decrease with age, growth rates of leaf—eating larval beetles such as Paropsis
atomaria and P. charybdis decline below maximum rates. P. atomaria larvae feed
preferentially on new ﬂush; however, when feeding on old growth, an inverse

relationship is observed between foliage consumption by the larvae and leaf nitrogen

 

‘A 1 to 3 percent crude protein content is well below the 8 to 10 percent
protein level required to invoke DIT in laboratory rats (Rothwell and Stock 1985c).
Thus, it is likely that many herbivores typically consume diets in the wild that
potentially could cause DIT.

33

content. Presumably, this feeding pattern optimizes nitrogen intake (Ohmart and
Edwards 1991).

The seasonal decline in the nitrogen content of plants is also correlated with a
decrease in moisture content (Raupp and Denno 1983). In general, leaf water of
terrestrial plants declines with maturity (Scriber and Slansky 1981). Reduction in
plant moisture content may reduce the suitability of plants as a food source. Low
water content in forage may cause water stress in herbivores and reduce their ability
to use plant nutrients. Scriber (1977) found that, when compared to controls, the
efﬁciency of nitrogen assimilation and both gross and net energy efﬁciencies declines
in lepidopteran larvae fed wild cherry leaves with low moisture content. Reese and
Beck (1978) noted a similar phenomenon occurred in the black cutworm (Agrotis
ipsilon), when larvae were fed artiﬁcial diets of varying moisture content. The
physiological efﬁciency model proposed by Scriber (1984a; 1984b) links host-plant
suitability for insect herbivores with leaf water-nitrogen composition, and predicts
optimal host-plant utilization when leaf water and nitrogen content are at their highest
concentrations. The model is supported by evidence showing that geographical
variation in larval growth rates of the silkmoth Callosamia promethea is largely
attributable to variations in leaf water and nitrogen content of host plants (Scriber

1984b).

 

’This response is similar to the DIT response observed in vertebrates in the lab.
As the amount of nitrogen assimilated from the diet declines, net energy efﬁciency
also declines.

34
Herbivore Feeding Habits

Seasonal shifts in food quality are often reﬂected in the feeding habits of wild
animals. Lindroth and Batzli (1984a) observed that meadow voles (Microtus pennsyl-
vanicus) fed on the green shoots of monocots and dicots when available, then
switched to seeds, roots, and grasses during autumn and winter. Plants that were
preferred forage during the summer (Tnfolium pratense and Taraxacum oﬁ‘icinale)
had higher levels of crude protein (20. 8 and 11.0 percent, respectively) than less-
preferred species such as Poa pratensis and Andropogon gerardii (6.3 and 3. 8
percent, respectively). Karasov (1982) found a similar pattern for omnivorous
antelope ground squirrels (Ammospermophilus leucurus). In the spring, the dominant
component of the natural diet of Mojave Desert antelope ground squirrels is moist,
high-nitrogen vegetation that is easily digested. During midsummer, when arthropod
populations are peaking, arthropods make up a considerable portion of the squirrels’
diet. By late summer and fall, nitrogen levels and dry matter digestibility of
vegetation decline, and seeds and arthropods dominate the diet. After the winter
rains, ground squirrels readily feed on new plant growth, and md consumption
declines. During feeding trials, Karasov found that ground squirrels maintained
constant body weights while consuming vegetation collected during the spring, but
lost weight on a diet of late-summer vegetation. When the latter diet was supplement-
ed with seeds and arthropods, weight constancy was regained (Karasov 1982).

Preferential fwding by herbivores resulting from phenological or ontogenetic
variability in food quality has been observed in many different animal groups. These

include the impala (Rodgers 1976), voles (Goldberg, Tabroff, and Tamarin 1980;

35
Bergeron and Jodoin 1987), deer (Klein 1970), Icelandic ptarmigan (Gardarsson and

Moss 1970), pocket gophers (Gettinger 1984) and many Species of insects (see Denno
and McClure 1983, and references therein). Thus, the supply of high quality food for
herbivores, inﬂuenced by climate and growth stage, varies over time and can affect
feeding patterns and survivorship of herbivores.

According to Sinclair (1975), the quality of the food supply for a herbivore,
determined by its crude protein content, can become so low due to seasonal variability
that the herbivore can no longer maintain its body weight, even through sustained
grazing. In African ungulates, selective feeding may be beneﬁcial at times, but many
species are forced into opportunistic nomadism, moving to regions where rainfall
permits an adequate food supply (McNaughton 1987). During periods of sustained
drought, survivorship declines (Sinclair 1975).

Thus, the effects of climate can create conditions whereby herbivores are faced
with a food supply that is too low in usable protein to sustain normal body weight or
permit growth. Under these conditions, traits that enhance protein uptake should be
favored by natural selection. If DIT enables a herbivore to increase nitrogen uptake
under harsh conditions such as this, then it is likely that it could have evolved by
natural selection as a physiological response to low ambient protein levels in food
resources.

SECONDARY PLANT CHEMICALS

Plants produce a variety of allelochemicals that strongly inﬂuence the nutri-

tional ecology of herbivores. Plant secondary chemicals can alter the nutritional

Status of herbivores by functioning as toxins, feeding deterrents, or digestibility

36

reducers. Furthermore, they can have indirect effects on herbivore nutrition by
altering intestinal microﬂora or adversely affecting the herbivore’s biochemical
detoxication systems (Lindroth 1988). Ultimately, plant allelochenricals can affect the
feeding habits, digestion and conversion efﬁciencies, health status, and survivorship
of herbivores (Scriber and Slansky 1981; Lindroth and Batzli 1984b; Lindroth, Batzli,
and Avildsen 1986; Bergeron and Jodoin 1987; Bryant et a1. 1991).
Plant Defense Theory

Plant secondary chemicals were not always recognized by scientists as having
antiherbivore properties. Before Gottfried Fraenkel developed the theory in the 1950s
that plant secondary chemicals defended plants from herbivore attack, they were
viewed primarily as irrelevant, biologically meaningless, waste by-products of plant
metabolism, occasionally useful only to humans for medicinal purposes (Grant 1984).
Since then, considerable research has been done in an attempt to ﬁt allelochenricals
into a general theory of plant antiherbivore chemistry.

In the mid 1970s, the theory of plant apparency was developed by Feeny
(1976) and Rhoades and Cates (1976) to explain observed herbivore feeding patterns
(e. g. , Feeny 1970, and Tahvanainen and Root 1972). The plant apparency theory
provided an evolutionary framework for studies on the coevolutionary interactions
between plants and herbivores, and is based on the idea that production of secondary
compounds imposes a metabolic cost to the plant.

Apparency theory holds that the allelochemical proﬁle of a plant is primarily
determined by how easily it is found (how “apparent” it is) by a herbivore. Large

and/or long-lived plants easily found by herbivores are predicted to defend themselves

37

by investing a large amount of primary production into “quantitative” chemicals, such
as polyphenols (tannins) or resins, which function in a dose-dependent fashion to deter
herbivory. In contrast, small plants and/or ephemerals (“unapparent” plants) should
rely on “qualitative” chemicals such as toxins, which sufﬁce in small amounts to
deter most herbivores and, therefore, require less energy expenditure to produce
(Howe and Westley 1988).

Because apparent plants are easily found, quantitative chenrical defenses are
predicted to be effective against all herbivores. In contrast, Specialist herbivores can
easily neutralize qualitative defense toxins by employing enzyme defenses that
presumably have evolved through long-term interactions with plants. Production of
qualitative low-cost toxins is considered an effective defense for short-lived and scarce
plants because herbivores have minimal contact time and are therefore less likely to
evolve the means to detoxify them (Howe and Westley 1988).

To explain the heterogeneous distribution of secondary compounds within a
plant, and preferential feeding by herbivores on different plant parts, McKey (1979)
extended the principle of apparency to plant components. According to McKey, older
woody tissue and mature leaves should rely heavily on quantitative chemical defense,
while new growth and seeds are more likely to contain qualitative defensive chemi-
cals, such as toxins.

The theory of plant apparency is supported by numerous studies of plant-
herbivore interactions (Feeny 1970; Root 1973; Fee'ny 1976; Futuyma 1976; Cates
and Rhoades 1977; Ikeda, Matsumura, and Benjamin 1977; Slansky and Feeny 1977;

Cates 1980; Berenbaum 1981; Henderson 1990). However, in recent years it has

38

become clear that plant apparency cannot fully explain all patterns of allelochemic
defense, and there has been considerable revision in defense theory to accommodate
new insights. Most notable are the ideas of Bryant, Chapin, and Klein (1983) and
Coley, Bryant, and Chapin (1985), who propose that resource availability is the
evolutionary determinant that controls the carbon/ nutrient balance of plants.
According to these researchers, plant resources ultimately control the amount of
energy than can be allocated to plant defense and the nature of the defense.

In this view, when resources are limited, slow-growing plants are favored by
natural selection. These plants invest heavily in constitutive chemical defenses
because they are unable to acquire new resources to compensate for tissue lost to
heavy browsing (e.g., boreal forest trees). In contrast, plants that are not limited by
available resources have high growth rates, allocate less energy to chemical defense,
and respond to excess grazing through compensatory growth. Grarninoids that retain
large nutrient reserves underground are good examples.

The theoretical framework interpreting the ecological role of allelochenrics
continues to grow and change as our knowledge of plant-herbivore interactions and
plant secondary chemistry increases. Current theories on mammal-plant chemical
ecology and evolutionary relationships can be found in Palo and Robbins 1991 (and
references therein); for insects, see Barbosa and Letoumeau 1988 (and references
therein). Defense theory has provided and continues to provide an ecological and
evolutionary framework for research on plant-herbivore interactions. One of its major
accomplishments is that it has cast a new light on the green world; plants are no

longer viewed as passive participants in the ecology of herbivores. According to

39

Janzen ( 1978): “The plant world is not colored green; it is colored morphine,
caffeine, tannin, phenol, terpene, canavarrine, latex, phytohaemagglutinin, oxalic acid,
saponin, L-dopa, etc.” Thus, our increasing knowledge about plant secondary
compounds adds a dynamic facet to the study of herbivore-plant interactions that was
previously unknown or ignored.

Defense theory also serves to illustrate that herbivores can appear to have a
plentiful food supply, but the quality of their food resources can be limited. For
herbivores, food Shortages are not always readily apparent.

Impact of Chemical Defense on Herbivores

The quality and quantity of available food for herbivores can be altered by
plant secondary compounds. This, in turn, can lead to decreased herbivore ﬁtness
through increased mortality, reduced growth rate, or lowered fecundity (Rhoades and
Cates 1976). Also noteworthy is that the impact of plant chenricals on the nutritional
status of a herbivore is inﬂuenced by a variety of intrinsic herbivore characteristics,
including the animal’s age, sex, genetic makeup, and prior feeding experience with
various allelochemics (Lindroth 1988). Ultimately, actual food quality depends on the
interaction of both plant and herbivore traits. For instance, the arginine analog L-
canavanine, found in the seeds of the neotropical legume Dioclea megacarpa, can be
highly toxic to many herbivores (e.g., the silkworm, boll weevil, and fruit ﬂy).
However, some specialist herbivores, such as the bruchid beetle (Caryedes
brasiliensis), not only have circumvented the toxic effects of L-canavanine, but use it

to generate dietary nitrogen (Rosenthal and Bell 1979).

40
Thus, broad generalizations about the effects of speciﬁc allelochemicals (or

even a particular class of chemicals, such as tannins) are difﬁcult to make because of
the dynamic interactions between plants and herbivores. Moreover, our knowledge
about allelochemic-herbivore interaction is still growing; as the mass of information
increases, so does our awareness of the increased complexity of interaction. There-
fore, while generalizations are inevitable, they are not intended herein to be dogmatic.

Texins mg fﬂing deterrents. Toxic allelochemics (TA) are plant chemicals
that, in small dosages, can cause physical damage, sickness, or death in a herbivore
(Grant 1984). They are widely distributed among plants; for instance, the toxin
hydrogen cyanide has been found in over one thousand species (Bell 1978).
Alkaloids, terpenoids, toxic amino acids, phytohemaglutinins, toxic lipids, glucosinol-
ates, proteinase inhibitors, saponins, and cyanogenic glycosides are some of the
known types of TAs (Bell 1978; Scriber 1984a; Bemays, Driver, and Bilgener 1989).
Their effects on herbivores include severe convulsions, organ damage, hair loss,
tumor formation, weight loss, abortion, fetal damage, growth retardation, inhibition
of sexual maturation, neurological damage, reduction in life span, and death (Freeland
and Janzen 1974). For example, many ﬁeld and laboratory studies on voles suggest
that plant secondary compounds can affect their health and survivorship (J ung and
Batzli 1981; Lindroth and Batzli 1984b; Bergeron, Jodoin, and Jean 1987; Lindroth,
Batzli, and Avildsen 1986; Bergeron and Jodoin 1987; 1989).

Because of the extreme detrimental effects and wide dissenrination of TAs,
Freeland and Janzen (1974) suggested that some herbivores use behavioral mecha-

nisms associated with sensory capabilities and memory to avoid exceptionally toxic

41

plants or plant parts. Learned aversions and preferences are well known for laborato-
ry rats that develop meal-induced food preferences from single-trial feeding (Lindroth
1988). Vaughan and Czaplewski (1985) proposed that young Stephens’ woodrats
(Neotoma stephensz) learn from their mother to avoid juniper forage high in tannins
and terpenoids, through dietary training during a long preweaning period.

Many studies have shown that food preferences are often negatively correlated
with plant toxin levels. Feeding on bracken fern (Pteridium aquilinum) by sheep,
deer, and locusts, is strongly inhibited during seasonal peaks in cyanogenesis (Coo-
per-Driver, Finch, and Swain 1977). Evidence from experimental feeding trials
indicates that some animals will voluntarily reduce food intake to well below mainte-
nance level when fed either browse containing high levels of allelochemics or
laboratory food treated with browse extracts (Bryant et al. 1991). Similarly, feeding
deterrence caused by plant allelocherrrics has been observed in meadow voles (Kendall
and Sherwood 1975; Kendall and Leath 1976), mule deer (Schwartz, Regelin, and
Nagy 1980), tundra voles and lemmings (Jung and Batzli 1981), Abert squirrels
(Zhang and States 1991), snowshoe hares (Bryant 1981), and many other wild
animals.

AS previously mentioned, young, growing plant tissue is usually higher in
nitrogen than older tissue, and senescence routinely results in a seasonal decline in
nitrogen. Although new plant growth may be high in nitrogen content, it often
contains elevated levels of TAs that can function as feeding deterrents (McKey 1979).
Bryant’s study (1981) of snowshoe hares (Lepus americanus) browsing on boreal

forest trees showed that hares were deterred from feeding on adventitious shoots high

42

in terpenes and phenolic resins. Hares preferred, instead, to feed on more mature
twigs, even though the adventitious shoots were higher in nitrogen content.
Comparable results were found for snowshoe hares feeding on Alaska feltleaf willow
(Salix alaxensis) (Bryant et al. 1985).

Freeland and Janzen (1974) suggested that feeding strategies of herbivores
depend on both speciﬁc plant defenses and the herbivore’s nutritient requirements.
According to Freeland and Janzen ( 1974), generalist herbivores sample a variety of
foods and depend on learning mechanisms, in response to adverse internal physiologi-
cal effects induced by phytochemicals, to select a diet that will maximize nutrient
intake and minimize the effects of TAs. Thus, diet diversiﬁcation results in the intake
of a variety of toxins, but limits the dosage of any one toxin. These ideas on food
sampling and diet diversiﬁcation are now being incorporated into optimality models of
herbivore foraging strategies. For instance, Belovsky and Schnritz (1991) developed a
linear programming model of optimal foraging for herbivores that incorporates
phytochemical constraints to diet selection.

Research by Lindroth and Batzli (1984a) and Lindroth et al. (1986) suggests
that both the allelochenricals and the protein content in plants can inﬂuence foraging
preferences in M. penmylvanicus in bluegrass and prairie habitats. Preferred foods
such as Trzfolium pratense and Taraxacum oﬁicinale not only had high levels of
protein (previously discussed), they also had high levels of phenolics. Poa pratensis
had low levels of both crude protein and phenolics and was least preferred by voles.
Andropogon gerardii was also avoided; A. gerardii had low protein content and

intermediate levels of phenolics. Voles showed only a moderate preference for

43

feeding on Lespedeza cuneata, even though crude protein content was relatively high.
L. cuneata had the highest concentration of phenolics of all the plants tested, suggest-
ing that, although protein content was high, it was not high enough to overcome the
deterrent effects of phenolics and stimulate feeding.

Lindroth and Batzli (1984b) also tested the effects of the interaction of allelo-
chemics and protein on the growth rate of prairie voles (M. ochrogaster). Their
results showed that the plant chemicals quercetin and tannic acid reduced growth rates
in voles when protein levels were low, but had no effect at higher protein levels.
Thus, at least in some animals, if the nitrogen level of foodstuffs is high enough, the
detrimental effects of some TAs can be reduced or eliminated. However, there may
be a point at which TA levels or potency are so high that feeding is prevented,
regardless of the plant protein content. Another plant chemical, quebracho, was
highly unpalatable and, as a result, was lethal at all protein levels (Lindroth and Batzli
1984b).

Feeding studies by Bergeron and Jodoin (1987) and Bucyanayandi and
Bergeron (1990) found that, in an old-ﬁeld community, M. pennsylvanicus preferred
food that was high in protein and low in phenolics. Contrary to many optimality
models that use energy as the optimal nutritional factor, voles did not appear to be
energy limited or choose food based on energy content. Food selection was based on
proteinzphenolic ratios. Therefore, Bergeron and Jodoin ( 1987) deﬁned “high
quality” food for voles as “plant species that contain the highest level of protein with

the lowest levels of total phenolics. ”

 

44

Plant toxins, then, can linrit a herbivore’s intake of high-nitrogen food by
reducing palatability of foodstuffs that are high in nitrogen content. Many herbivores
may be forced to feed on a variety of plants, including those with low nitrogen
content, to limit the intake of any particular toxin, and thus avoid injurious toxic
effects. Moreover, intake of food with high protein content may negate some
deleterious effects of toxins.

Herbivores that choose to fwd on foods that are low in protein because they
are also low in allelochenric content, could increase consumption of low-protein
forage to increase bulk protein intake. Under these conditions, low-protein DIT could
function to eliminate excess caloric intake and enhance protein intake.

W. The distinction between plant toxins and allelochenrics
that reduce digestive efﬁciency is not always a clear one (Meyer and Karasov 1989).
For instance, tannins can function to limit the absorption of nutrients across the gut
wall (Palo 1985). Also, they can be toxic, causing erosion of the intestinal mucosa or
lesions in the liver or kidneys (Bemays et al. 1989). Meyer and Karasov (1989)
examined the impact of tannins on ﬁve species of rodents and two species of lago-
morphs and observed that tannins reduced plant digestibility. They also noted that
there was some evidence of postdigestive toxic effects (animals were deterred from
feeding and lost weight, even when there was no signiﬁcant effect on digestive
efﬁciency), although this was not measured directly. In general, though, digestibility-
reducing allelochenrics (DRA) are functionally different from toxins. Whereas toxin

effects tend to be qualitative and act directly on herbivore tissue, the effects of DRAS

45

are usually quantitative, dosage dependent, and more indirect (Cates and Rhoades
1977).

Tannins. Tannins are polymeric phenolic compounds, synthesized primarily
from products of the shikimic acid pathway in plants that can act as complexing
agents capable of binding with nutrients, especially proteins. In addition to the
aforementioned toxic effects, tannins can also hinder plant digestion in herbivores by
fornring precipitates with dietary proteins, amino acids, fats, nucleic acids, polysac-
charides, or digestive enzymes (Bemays et al. 1989). Also, tannins may inhibit
protein absorption in the intestinal mucosa (Mole et al. 1990b). Tannins have been
shown to reduce both protein and dry-matter digestibility in runrinants (Barry,
Manley, and Duncan 1986; Robbins et al. 1987; Robbins et al. 1991).

The relative role tannins play as DRAS is a topic of current debate (Bemays et
al. 1989). Nevertheless, fecal nitrogen data show that ingestion of tannins by herbi-
vores results in signiﬁcant nitrogen losses (Glick and Joslyn 1970; Lindroth et al.
1986). A study by Hanley et al. (1992) showed that, for cervids in the genus
Odocoileus feeding on ﬁreweed (Epilobium angustzfoliwn), there was a 44 percent
reduction in digestible protein due to the effects of tannins.

Mole, Butler, and Iason (1990a) suggested that the primary source for tannin-
induced excretory nitrogen losses is endogenous, rather than dietary, proteins.
Secretion by herbivores of proline-rich salivary proteins, which have a high afﬁnity
for tannins, has been proposed as a defensive mechanism against the negative effects
of dietary tannins (Mole et al. 1990a). High-afﬁnity, tannin-binding salivary proteins

conceivably could lower tannin absorption and concomitant toxicity. Binding of

46
tannins by salivary proteins could reduce the amount of dietary proteins bound by

tannins, and also minimize the loss of digestive enzymes due to tannin binding.
Because proline-rich salivary proteins have a higher afﬁnity for tannins than dietary
proteins (more tannin is bound per unit of salivary protein than dietary protein), the
herbivore experiences less protein loss than if all the tannins were bound by dietary
proteins alone (Robbins et al. 1991).

Other sources of non—dietary fecal nitrogen besides salivary proteins include
mucus and cells sloughed off from the gut as a result of tannin abrasion (Freeland,
Calcott, and Anderson 1985). Regardless of the source of increases in fecal nitrogen,
the net effect of dietary tannins on herbivores is a decrease in the efﬁciency of
nitrogen retention, and a negative impact on nitrogen balance.

Lignins. Lignins are another group of plant phenolic polymers produced
biosynthetically from products of the shikimic acid pathway (Swain 1979). Within
plants, lignins bind to cellulose or hemicellulose in the cell walls, imparting a
structural rigidity or “woodiness” to the plant. According to Swain (1979), “Wood,
as Gertrude Stein nright have said, is wood and, one might add, the quintessence of
woodiness lies in the presence of lignins.” Lignins also stiffen and toughen the stems
and leaves of nonwoody plant tissues (Howe and Westley 1988).

Lignins are the primary dietary component of forage that limits digestibility
(Van Soest 1982). Few organisms are capable of digesting lignins; therefore, lignins
binding to plant cell walls reduce the digestibility of cell wall carbohydrates (Howe
and Westley 1988). It has been suggested that lignins reduce both dietary

carbohydrate and protein digestion by binding in vivo to these nutrients (Swain 1979).

47

Another hypothesis is that they can limit digestible protein by entrapment of nutrients
within cell walls (Bell 1971). However, there is little evidence available to support
either of these claims (Parra 1978; Van Soest 1982; Buchsbaum, Wilson, and Valiela
1986). There is some evidence, though, that lignins bind amino acids within grasses
and legumes, thus reducing nitrogen availability (Van Soest 1982).

Lignins do play a role in reducing nitrogen availability in forage by dilution
effects. The concentration of lignins in plant cells can reach 40 percent or more
(Swain 1979). A decline in protein content and a concomitant increase in lignins is
often associated with increasing plant maturity. Daylength and temperature are key
factors determining structural changes in plants as they mature. For example, as
temperature increases, so does plant metabolism, resulting in an increase in the rate of
production of plant structural components (lignins, cellulose, and hemicellulose) and a
decline in cytoplasmic metabolites. Thickening of the cell wall by structural growth
and ligniﬁcation leads to a decline in cytoplasm volume by increasing the limitation
set by the physical dimensions of the cell wall on intracellular volume. Because
cytoplasmic volume is limited, the amount of intracellular nutrients that can
accumulate is also limited (Van Soest 1982; Schwartz and Hobbs 1985).

Cellulose and hemicellulose. Along with lignin, cellulose and hemicellulose
comprise what is commonly called “dietary ﬁber” (Milton 1979). Cellulose and
henricellulose are complex polysaccharides that form the carbohydrate portion of plant
cell walls. These structural carbohydrates cannot be broken down by the digestive
enzymes of most herbivores, especially vertebrates. Digestion of the carbohydrate

portion of ﬁber in many herbivores requires symbiotic gut 'microbes that have

 

48

cellulolytic enzymes capable of catalyzing ﬁber breakdown. In vertebrates, this
process is enhanced by fermentation, which usually occurs in a special fermentation
chamber formed by modiﬁcation of the digestive tract. The chamber usually occurs
as an enlarged foregut compartment typical of ruminants and also found in macropods
and colobid monkeys. It also may exist as a modiﬁed hindgut sac (enlarged colon or
caecum) found in perissodactyls and microtines, as well as other animals (Milton
1979).

Because structural carbohydrates are difﬁcult to digest, the presence of
cellulose and henricellulose in plants reduces their digestibility. However, like
lignins, cellulose and hemicellulose have only a moderate affect on the digestibility of
cell contents (Van Soest 1982; Buchsbaum et a1. 1986). Structural carbohydrates do
affect food nutrients and nitrogen availability, though, by dilution effects similar to
those previously described for lignins.

Silica. Silica is an oxide of Silicon, and is commonly metabolized by plants
and incorporated into cell walls as a structural component. Silica is completely
indigestible by herbivores (Howe and Westley 1988), and like ﬁber, it reduces food
quality through adulteration of nutrients. Digestibility of foodstuffs can be reduced by
silica through the abrasive effects of its crystals that accelerate herbivore tooth wear
(McNaughton et al. 1985). Worn teeth limit a herbivore’s ability to masticate food
before it is passes into the gut. A reduction in mastication reduces the surface area of
chewed food that would be available for enzymatic digestion in the gut, thus reducing

forage digestibility.

49

Silica can also be toxic to herbivores. In grazing ungulates, it is known to
cause silica urolithiasis (a fatal urinary tract disease) and augment the formation of
esophageal cancer (McNaughton et al. 1985). Silica is also a well-known feeding
deterrent due to the sharp siliceous protrusions on leaf fringes. Although its occur-
rence in grazing plants is dependent on soil type and availability of silica in the soil
(Van Soest 1968), feeding pressure can also be a factor. In heavily grazed areas of
the Serengeti, the concentration of silica in plants is much greater than in adjacent
areas with less grazing pressure. This pattern persists when transplanted plants from
different grazing areas are reared in the laboratory, showing genetic differences
presumably resulting from natural selection due to grazing pressure (McNaughton et
al. 1985).

STRUCTURAL DEFENSE

The idea that plants actively discourage herbivore feeding and therefore limit
food availability is most evident when the array of external mechanical defenses is
examined. The so-called “barbed wire syndrome” is exempliﬁed by the sharp spines
of cacti, which are well-known feeding deterrents against large, mammalian herbi-
vores (Cooper and Owen-Smith 1986). Spinescence in holly trees is known to dis-
courage leaf-eating caterpillars, and specialized hairs called trichomes create a dense,
tangled barrier to many feeding insects and are even known to pierce their antagonists
(Grant 1984). Some hairs are glandular in nature; they secrete repellant oils or sticky
exudates that can entrap invertebrate herbivores. Stinging nettles have trichomes that
secrete an irritating chemical that can function as a feeding deterrent (Grant 1984).

Thistles are well-armored against herbivores due to the extensive use of trichomes,

50
and are often highly conspicuous compared to other plants in heavily grazed pastures

(Grant 1984). Agriculturalists, well aware of the effectiveness of structural defenses
in plants, select genetic strains of cotton or beans that are highly pubescent to
discourage the attack of small insects and mites (Howe and Westley 1988).

Fiber is also part of a plant’s structural defense, in that it increases leaf
toughness, making a plant more difﬁcult for herbivores to chew or penetrate
(Tabashnik and Slansky 1987; Ohmart and Edwards 1991). Leaf toughness can cause
decreased food consumption, reduced growth, and low survival rates in foliage-
chewing insects (Tabashnik and Slansky 1987).

INDUCED DEFENSE

Plant defense in response to herbivory is commonly thought to occur within a
chronic time frame by coevolutionary changes (or by what Fox (1981) calls “diffuse
coevolution”) (see Spencer 1988, and references therein). However, some plants are
capable of mobilizing defense mechanisms within an acute time frame. This type of
response by plants is called induced defense (Rhoades 1985), and it entails a
phenotypic change in plants in response to herbivory that is protective for the plants
and usually deleterious to herbivores (Myers 1988).

For example, Benz (1974; 1977, as cited by Rhoades 1983) found that
defoliation of larch (Larix decidua) by the larch budmoth (Zeirapherra diniana)
resulted in phenotypic changes in new tree growth. Fresh growth by Larix occurring
the year after defoliation consisted of smaller needles higher in ﬁber and lower in
nitrogen content than needles from the previous year. This new growth may

sometimes be covered with an exudate of oleoresin that was not present the year

51
before. For the budmoth, these changes caused reduced palatability of larch needles,

low assimilation efﬁciencies, high mortality, and reduced fecundity.

Bryant (1981) found that some species of boreal forest trees respond to severe
browsing by snowshoe hares (L. americanus) by production of adventitious shoots that
have signiﬁcantly higher concentrations of terpene and phenolic resins than normally
found in mature growth. Bryant (1981) also found that hares avoided mature-growth- l
form twigs when they were coated with adventitious root resins. Tahvanainen et al.
(1985) noted that an acute production of defensive allelochemicals occurred when

mountain hares (L. timidus) fed on northern willow (Salix spp.) during the winter in

 

the northern boreal zone. Snowshoe hares feeding on Alaska paper birch and feltleaf
willow also induced an acute, chemically-mediated defensive response (Reichardt et
al. 1984; Bryant et al. 1985).

Lindroth and Batzli ( 1986) found that phenolic concentrations increased in old
ﬁeld alfalfa under heavy grazing pressure during vole population peaks. Oksanen and
Oksanen (1981, as cited by Lindroth 1989) observed a nearly twofold rise in
phenol/nitrogen ratios in bilberry (Vaccinium mym’llus) following intense grazing by
cycling populations of lemmings. Other plants known to exhibit induced chemical
defense include brown algae (Fucus distichus) (Van Alstyne 1988), birch (Betula
pubescens) (Hanhimaeki 1989), lodgepole pine (Pinus contarta) (Leather, Watt, and
Forrest 1987), cotton (Karban 1987), soybeans (Chiang et al. 1987; Lin, Kogan, and
Fischer 1990), wild parsnip (Pastinaca sativa) (Zangerl and Berenbaum 1990),
pondwwd (Potamogeton colorants) (J effries 1990) , and quaking aspen (Popular

tremuloides) (Clausen et al. 1989).

52
Apparently, mechanical defenses of plants can also be induced by herbivory.

When browsed by goats, Acacia depranolobium responded by producing longer thorns
(Young 1987). Young (1987) noted that thorns on browsed branches were longer
than those on unbrowsed branches, both within individual trees and between browsed
and unbrowsed trees. Similar results were noted for trees browsed by giraffes
(Milewski, Young, and Madden 1991).
SUMMARY

It is clear from the proceeding discussion that herbivore food resources can be
highly variable in terms of their nutrient content. This may be especially true of the
nitrogen content of forage, which can vary temporally, becoming so low that the food
supply may no longer be adequate to sustain growth or survival. Secondary plant
chemicals and structural defenses can also reduce a herbivore’s ability to obtain high-
quality, high-nitrogen food, by deterrent effects that force herbivores to feed on less
nutritious plants and plant parts (plants and plant parts lower in nitrogen). Plant
chemical compounds can also reduce nitrogen assimilation by binding with dietary
proteins that form complexes that are excreted.

Since nitrogen can be a limiting nutrient for herbivores, it is likely that natural
selection will favor traits that enhance nitrogen uptake. If herbivores are forced to
consume more food when nitrogen levels are low, a physiological response such as

low-protein DIT could be adaptive by augmenting nitrogen assimilation.

 

53
Herbivore Response to a Variable Food Supply

INFLUENCE OF SUPPLEMENTAL FOOD

To assess the impact of food availability on vertebrate herbivore population
dynamics, a number of food supplementation studies have been conducted. The
results of these studies have been far from unequivocal, and controversy continues
over the role that food plays in demographic patterns (Boutin 1990). Most of the
debate centers around the question of whether predation or food shortage regulates
population cycles in small mammals such as hares and microtine rodents. The issue
of population cycles will not be addressed here, but some patterns that result from
supplemental feeding of herbivores are worth mentioning and point to the importance
of the availability of high-quality food in herbivore ecology.
Reproduction

An escalation in breeding activity often occurs in response to supplemental
feeding (Boutin 1990). (See Table 2.1 for a compilation of studies focusing on
changes in reproductive activity in herbivores due to supplemental feeding.) Andrze-
jewski (1975) noted a signiﬁcant increase in winter breeding by the bank vole,
Clethrianomys glareolus, after supplemental feeding with oats. According to Andrze-
jewski (1975), in food-supplemented populations, 50 percent of the individuals found
in the spring were born during the winter, indicating extensive winter breeding. This
contributed to a two-year post-winter population increase that resulted in a population
Size that was two to four times bigger than any population of the previous six years;
control-plot animals did not increase. An early onset of the breeding season was also

observed (Table 2.1). Massive reproductive activity began in January instead of at

 

54

Table 2.1. Inﬂuence of supplemental food on reproduction in hares and microtines
(+, increase; 0, no change; blank, not examined). Adapted from Boutin 1990.

 

 

 

 

 

 

Brwding
Season
Species length Intensity Reference
Hares
Lepus americanus + Windberg and Keith (1976)
+ + Vaughan and Keith (1981)
+ Boutin (1984)
0 Krebs et al. (1986)
0 Krebs, Boutin, and Gilbert
(1986)
Microtines
Clethrionomys glareolus + + Andrzejewski (1975)
Clethrionomys rutilus + Gilbert and Krebs (1981)
Microtus califamicus + + Ford and Pitelka (1984)
Microtus ochrogaster + Cole and Batzli (1978)
+ Desy and Batzli (1989)
Microtus pennsylvanicus + Desy and Thompson (1983)
Microtus townsendii + 0 Taitt and Krebs (1981)
+ + Taitt and Krebs (1983)

 

the end of March or the beginning of April, which is usually regarded as the begin-

ning of the reproductive season (Andrzejewski 1975).

Food addition can cause an increase in breeding intensity. Cole and Batzli

(1978) found that the proportion of reproductive female prairie voles in a population

fed supplementally with rabbit pellets was Signiﬁcantly greater than in a control

population. An increase in the proportion of scrotal males in the experimental group

 

55

also was noted. Table 2.1 lists Studies that have shown increases in breeding intensity
(i.e. , the proportion of females breeding) that occurred as a result of added food.

Also of interest, Negus and Berger (1977) found that feeding limited supple-
ments of fresh green wheatgrass can trigger the onset of reproductive activity in a
nonbreeding winter population of M. montanus. All of the females in an experimental
group became pregnant while controls remained in nonbreeding condition. According
to Negus and Berger (1977), reproductive activity is cued by chemical signals in the
plant food resources. This suggests that some herbivores may depend on temporal
changes in food quality to signal reproductive behavior.
Growth

Many studies have shown that supplemental fwding of small herbivores in the
wild can have a positive inﬂuence on their growth. Increases in peak body weight
and the rate of growth have been recorded for hares and voles receiving food
supplementation (Table 2.2). Both adult and juvenile animals Show higher peak
weights when fed supplemental food, compared to control animals. Desy and Batzli
(1989) observed an increase in the rate of growth of young prairie voles (weight <
25 g when ﬁrst captured) on supplemental feeding regimes. Mean adult body weight
of voles was also greater in food-supplemented populations. The importance of
growth rate and body size is underscored by the fact that they are frequently used
qualitatively to assess ﬁtness. Reproductive maturity in many mammals is often

determined by an animal’s size instead of its age (White 1983; Rayor 1985).

 

56

Table 2.2. Inﬂuence of supplemental food on growth in hares and microtines (+ ,
increase; 0, no change; blank, not examined). Adapted from Boutin 1990.

 

Body Growth
Species weight rate Reference

 

 

 

 

Hares
Windberg and Keith ( 1976)
+ Vaughan and Keith (1981)
Boutin ( 1984)
+ Krebs et al. (1986)
Krebs, Boutin, and Gilbert (1986)
+ Smith et al. (1988)
Microtines E

L. americanus

+o++++

 

C. glareolus Andrzejewski (1975)
Banach (1986)
C. rutilus 0 Gilbert and Krebs (1981)

+ Krebs and DeLong (1965)
Cole and Batzli (1978)

+ Desy and Batzli (1989)

+ Desy and Thompson (1983)

+ Taitt and Krebs (1981)
Taitt and Krebs (1983)

M. californicus

M. ochrogaster

M. pennsylvanicus

M. townsendii

+++++oo++

 

FORAGE QUALITY

Herbivore performance (survival, growth, and reproduction) is usually
enhanced in habitats with high-quality forage (Lindroth 1989). Cole and Batzli (1979)
found that prairie voles in an alfalfa habitat had higher fecundity, survival, and body
weights than those in an adjacent bluegrass habitat. Levels of digestible energy,

crude protein, calcium, phosphorous, and sodium were higher for forage in the alfalfa

5 7
habitat. White (1983) compared the growth rates of reindeer (Rangifer tarandus) that

grazed on low-quality forage ranges (dwarf birch/sedge) with those that fed on ranges
with high-quality forage (willow/ sedge). Fawn growth rates were higher on the
willow-sedge ranges; milk production was reduced for reindeer on the low-quality
ranges, which presumably led to low fawn growth rates.
COMPENSATORY MECHANISMS

As noted previously, wild herbivores must contend with a nutrient supply that

can be highly variable in both quality and quantity. Plant tissues can ﬂuctuate greatly

 

in ﬁber and protein content, and plants have developed an array of chemical and 3
physical defenses that can limit nutrient accessibility. To compensate, herbivores
have evolved a multiplicity of responses to cope with a mercurial and often nutri-
tionally inadequate food supply.
Response to Food Privation

King and Murphy ( 1985) documented a number of compensatory mechanisms
used by many animals to mitigate the effects of food privation resulting from either
relative food shortage (food is plentiful but nutritionally inadequate) or general decline
in resource abundance (food is unavailable or there is only a linrited amount). They
include 1) accumulation and use of surplus nutrients, 2) economical use of available
nutrients, and 3) a mixture of both (King and Murphy 1985). Surplus nutrients can
be in the form of food caches that are often used by many animals during the winter.
Hamilton (1941) reported ﬁnding nearly a peck (as 9 liters) of beechnuts in a
Pemmyscus food cache, and Martin ( 1956) described underground chambers used by

M. ochrogaster to store grasses and fruits that they feed on during the winter.

58

Hibernation and torpor are common methods used by many animals to con-
serve endogenous energy stores and minimize dependence on exogenous food
resources. Hibernation can reduce energy requirements to levels well below those
required for normal activity. Herbivorous mammals such as marmots can reduce
their basal metabolic rate by 92 percent during hibernation (Heldmaier 1989). Daily
torpor can decrease daily energy requirements by up to 26 percent in Djungarian
hamsters (Phodapus sungorus) (Heldmaier 1989) and up to 30 percent in white-footed
nrice (Peromyscus leucopus) (Hill 1975).

Note that the animals used in my research (prairie voles, M. ochrogaster) are
not known to hibernate or go into torpor. Therefore, they are active throughout the
day as well as year-round, and most likely must deal with food Shortages by methods
other than metabolic arrest. Since they are herbivores, they may be faced with
relative food shortages such as those caused by a seasonal decline in the nitrogen
content of forage. It is possible, then, that herbivores such as prairie voles may have
evolved adaptive mechanisms that allow them to subsist on a food supply that is
highly variable in protein content and thus remain active throughout the year.
Response to Dietary Fiber

Plant ﬁber can limit digestibility and reduce a herbivore’s access to forage
nutrients by dilution effects. In ruminants, the digestibility of cellulose and hemicel-
lulose can range from 36 to 79 percent and lignin, cutin, and silica are essentially
indigestible (Van Soest 1982). Some structural and functional alterations of the
digestive system in herbivores are conspicuous adaptive traits that evolved to process

bulky plant ﬁber. Most mammalian herbivores have sacculated digestive tracts that

 

59

facilitate the fermentation of large amounts of food by microbial action (McBee
1971). Large fermentation chambers permit an increase in retention time, which
extends the time available for digestion. They also enable herbivores to host a large
microbial population of bacteria, ﬂagellates, and ciliates, which is essential for the
digestion of cellulose and hemicellulose. These structural polysaccharides have B-
glycosidic linkages that cannot be digested by most herbivores without the aid of
microbial-mediated fermentation (Van Hoven and Boomker 1985). Other digestive
tract adaptations include a highly papillated mucosal relief in the digestive
compartments of ruminants for the absorption of large quantities of fermentive end-
products. Forestomach fermenters such as ruminants often have comiﬁed epithelium
covering much of their papillae in the rumen. Presumably, this increases the
absorptive surface area (Van Hoven and Boomker 1985).

Flattened molars used for grinding plant tissue and teeth that grow
continuously to replace the loss due to wear from chewing abrasive plant material are
common among mammalian herbivores. Insect herbivores have highly modiﬁed
appendages for feeding on plants. Chewing mouthparts are used for fragmenting
plant tissue, and many insects can penetrate the tough cell wall protecting the nutrient-
rich cytoplasm by piercing it with a Sharp beak and sucking out the cell contents
(Howe and Westley 1988).

To augment nutrient uptake from low-quality forage, many herbivores increase
food consumption and decrease throughput time, which increases the mass of food
taken in for digestion. This is a common strategy employed by hindgut fermenters

(Van Hoven and Boomker 1985). In foregut fermenters such as ruminants, the rate of

l-'¢.' ;. I
m

 

60

passage of food is limited by the rumination process. However, hindgut fermenters
are less constrained because fermentation is not as extensive as in ruminants, and
there are less sacculated compartments for food to pass through. Nonruminants such
as the horse can process food twice as fast, for instance, as the ruminant cow (Bell
1971; Sibly 1981). According to Bell (1971):

The importance of this difference emerges in a simple calculation: If the

horse is two-thirds as efﬁcient as the ruminant in extracting protein from a

food but processes twice as much food in a given period of time, its rate of

 

assimilation of protein per unit of time is four-thirds of the ruminant’s rate. E

Therefore the horse (or the zebra) can support itself on a diet that is too low

in protein to support a ruminant.

A decline in digestive efﬁciency is usually associated with an increase in
passage rate (Demment and Van Soest 1985). For instance, food in the horse is
processed twice as fast as in the cow, but with 30 percent less digestive efﬁciency
(Janis 1976). This has led some researchers to postulate that herbivores challenged by
a low-quality, high—ﬁber diet, could beneﬁt from increased gut size, thus enlarging
the surface area available for assimilation of nutrients. An increase in gut surface
area could increase digestive efﬁciency and counterbalance the reduction caused by
low retention time. In a study by Gross, Wang, and Wunder (1985), prairie voles
consuming high-ﬁber diets increased food intake and showed a signiﬁcant increase in
small intestine and caecal mass and length. Similar results for M. ochrogaster were
obtained by Hammond and Wunder (1991), who postulate a homeostatic maintenance

of digestive efﬁciency in small herbivores challenged by increased throughput.

61
Response to Allelochenricals

Herbivores have developed many adaptations to counteract plant chemical
defenses. These include behavioral and biochemical mechanisms. Behavioral
mechanisms reduce exposure to allelochenrics, typically through selective feeding
(e. g., Bergeron and Jodoin 1987). Learning plays an important role in foraging
behavior for many herbivores, and ‘learned aversions’ to plant xenobiotics have been
documented (Vickery 1984; Vaughan and Czaplewski 1985; Lindroth 1988).
Interactions between forage constituents in herbivore diets can also be important.
Selective consumption of plants that are high in saponins may reduce the binding of
dietary protein by tannins due to chelation of tannin and saponin within the intestinal
tract. In addition, the surfactant effect of saponins within the gut may prevent tannins
from binding to the intestinal lumen and thus reduce tannin absorption and the
resultant toxicity (Freeland et al. 1985). Feeding on a broad range of plants to
nrininrize the toxic effects of any particular plant allelochemical may also be an
effective behavioral strategy (McArthur, Hagerman, and Robbins 1991).

Many enzyme detoxication systems that can overcome plant toxins are present
in both insect and mammalian herbivores (Lindroth 1988; 1991). Detoxication can
occur via gut microﬂora or by speciﬁc enzyme systems located in body tissues.
Membrane-bound mixed-function oxidases are generalized enzyme systems capable of
detoxifying many types of allelochenricals. They are present in several vertebrate
organs such as the kidneys, intestines, and lungs, but are especially common in
nricrosomes of the endoplasmic reticulum of liver cells (Lindroth 1988). They can

also be found in fat bodies or the midgut of many insects (Howe and Westley 1988).

62

Their usual mode of action is to render foreign substances more hydrophilic via
oxidation, reduction, hydrolysis, or conjugation reactions. This reduces the ability of

a toxin to enter cell membranes, thus reducing toxicity and facilitating excretion

 

(Lindroth 1988; 1991; McArthur et al. 1991).
Inactivation of allelochemics can be accomplished via noncovalent complex

formation with endogenous compounds. For instance, medicagenic acid from plant P.

saponins is bound by bile cholesterol and excreted in the feces (Applebaum and Birk

1979). Salivary proteins excreted by ungulates and insects may minimize the

 

ﬁr
.1" '.

deleterious effects of tannins by binding with dietary tannin and forming inactive
precipitates or soluble complexes (Bemays et a1. 1989; Robbins et al. 1991).

Changes in gut pH can affect chemical reactivity of gut contents. The low pH
found in mammalian stomachs can reduce tannin-protein complex formation (McAr-
thur et al. 1991), and the highly alkaline conditions found in the midgut of some

insects can cause tannins to dissociate from bound protein (Berenbaum 1980).

Summary
It is apparent that herbivore ecology is intimately linked to the quality of their
food supply. The premise that herbivores cannot be limited by their food supply
because “the world is green” seems, with hindsight, without widespread foundation.
There is a wealth of evidence that suggests otherwise; food shortages—especially
relative food shortages—are part of the fabric that molds the nutritional ecology of

many herbivores. Also, herbivores and plants interact dynamically, exerting mutual

63

selection pressures that have shaped much of their ecology. Thus we see traits of
both plants and animals that evolved because of reciprocal interactions.

In the next section, I will discuss the relationship between the nutritional
ecology of herbivores and metabolism. The determinants of mammalian basal
metabolic rates are unknown. The effects of climate or animal size in relation to its
surface area are often suggested to inﬂuence basal metabolism. Rarely are nutrient
factors considered to play a role in determining basal metabolic rates (BMR). Since
DIT in laboratory rodents is inﬂuenced by the balance of nutrients in the diet (Chapter
I), it is reasonable to postulate that their relative abundance in the natural diets of
herbivores can have some inﬂuence on the evolution of herbivore metabolic rates
(either BMRS or gross metabolic rates) in the wild. To compensate for relative food
shortages (low ambient dietary nitrogen), herbivores may have evolved physiological
mechanisms such as DIT that elevate metabolic rates in response to low-protein food
intake, which ostensibly could enable the distillation of speciﬁc nutrients from

‘unbalanced’ diets.

METABOLIC RATE AND THE NUTRITIONAL ECOLOGY OF HERBIVORES
The history of the scientiﬁc study of animal energetics extends back to
the 17008, during the time of Priestley and lavoisier. Lavoisier found that oxygen
was an elementary gas removed from ordinary air by both mice and a burning ﬂame.
This discovery could not easily be reconciled with the phlogiston theory currently
favored by Priestly and others. The phlogiston theory held that phlogiston was added

to air by a ﬂame or a mouse and, in a conﬁned space, eventually made the air unﬁt to

 

 

E II." -.
A

 

64
sustain either. This theory was eventually replaced by Lavoisier’s theory6 of

combustion and metabolism, which stated that both a ﬂame and a mouse use oxygen
from air and produce heat in similar fashion. Each consumes oxygen by combining it
with an organic substance, resulting in the production of C02, water, and heat. In
1780, Lavoisier and LaPlace concluded that most of the heat produced by an animal
(in this instance, the guinea pig) is due to the in viva mixing of organic substances
with oxygen. By the nineteenth century, the idea that combustion and metabolism
were similar thermodynamic processes was ﬁrmly established (Kleiber 1975a).

During the nineteenth century, physiologists discovered that the rate of heat
production in endotherrns is highly correlated with body size. In 1839, Sarrus and
Rarneaux proposed that mammalian heat production, as measured by oxygen con-
sumption, was proportional to the free surface area and could be scaled to the ”A
power of body mass. This became known as the “surface law” and was subsequently
veriﬁed by many biologists. However, later research—especially in the twentieth
century—showed that surface area and body mass alone could not account for the
correlation of metabolism with body mass (Kleiber 1975a).

Kleiber (1932) examined the BMRS of a number of animals, varying in size
from rats to large ungulates, and described the relationship mathematically using an
allometric function with an exponent of 0.74. The equation was subsequently revised

as follows:

 

“The English biologist Adair Crawford is considered the cofounder of this
theory (McNab 1992).

65
M = 7ow°-75

where M = metabolic rate in kilocalories per day and W = body mass in kg (Kleiber
1947). General acceptance of the Kleiber equation as a standard for the study of
animal energetics was one of the main factors that caused the decline in popularity of
the surface law, since the surface law implied an exponent of 0.67 to describe the
correlation between mass and metabolism in animals.

The study of metabolism in the twentieth century branched off into many
directions, permeating human physiological and nutrition research (e. g. , Kleiber 1945;
Garrow 1978; Chapter I), as well as ecological research. Intensive studies of how
temperature and climate affected thermoregulation and metabolism were undertaken
(Scholander et al. 1950a; 1950b; 1950c). Animal energetics became a major area of
study for environmental physiologists. Researchers used laboratory measurements of
energy expenditures for various activities of wild animals (such as foraging,
locomotion, and reproduction) to estimate ﬁeld energy budgets (Pearson 1954; McNab
1963). This type of information was thought to typify the bioenergetics of animals
under natural conditions and often formed the basis of initial calculations of trophic
transfer and transformation of energy through populations and communities (e.g.,
Petrusewicz and Macfadyen 1970).

Many scientists continued to pursue a physical explanation for the 0.75
exponent in the Kleiber equation (as of yet, without much success). Although the
0.75 exponent could not be satisfactorily explained, ecological explanations for
variation above and below the Kleiber curve were pursued. Some focused on climatic

correlates (Scholander et al. 1950a) while others examined phyletic relationships

66
(I-Iulbert and Dawson 1974; Crompton, Taylor, and Jagger 1978). Scholander’s

group (1950a) concluded that “the basal metabolic rate of terrestrial mammals is
ﬁmdamentally determined by a size relation according to the formula Cal. /day = 70
kg. ", and phylogenetically nonadaptive to external temperature conditions. Equally
nonadaptive is the body temperature, and the phylogenetic adaptation to cold there-
fore rests entirely upon the plasticity of the factors which determine the heat loss,
mainly the ﬁtr insulation. ” Thus, there was little evidence to support the notion that
climate controlled the basal rate of metabolism in animals (but see McNab and
Morrison 1963, and Shkolnik and Schmidt-Nielsen 1976). Whether basal metabolic
rates are set by taxonomic afﬁliation continues to be debated (Felsenstein 1985; Hays-

sen and Lacy 1985) and will not be discussed here.

Food Habits and Energetics in Mammals

The search for the causal factors of variance in the Kleiber curve resulted in
the ﬁnding that BMRS are correlated with mammalian food habits (McNab 1986).
For instance, folivorous mammals have low BMRS, and the degree of depression is a
function of the proportion of leaves in their diet (McNab 1978). McNab (1978)
postulated three possible explanations: 1) leaves have high ﬁber and cellulose content
and therefore are a poor source of energy, 2) arboreal mammals (a) reduce their
intake of tree leaves because they are loaded with a variety of allelochenricals and (b)
use glucose to detoxify allelochemicals which reduces net energy intake, and 3)
arboreal folivores tend to be sedentary and have a low muscle mass, which could

account for a reduction in basal rate.

67
McNab (1986) predicted that animals with a readily available, high-energy

food source would have high BMRS. Those food sources with low digestibility, high
allelochemic content, and a low energy content would permit only low BMRS.
Grazing and meat-eating mammals were observed to have high BMRS (food may be
readily available, high in energy, or low in toxins), while folivores, frugivores, and
ant/termite-eating mammals had low basal rates (diet may have low availability,
reduced energy content, and/or high toxin content). Those animals that consumed a
mix of foods in the diet had intermediate rates (McNab 1986).

McNab’s research was among the ﬁrst comprehensive ecological studies to
link metabolic heat production with food habits in mammalian energetics. However,
our knowledge of this relationship is still limited. Recently, there has been some
interest in the inﬂuence of allelochemicals on mammalian metabolic rates. Thomas,
Samson, and Beregeron (1988) measured the metabolic rate of M. pennsylvanicus fed
a diet of Purina rabbit chow containing 6 percent gallic acid (a phenol). Their results
Showed an increase in metabolism by 13.6-22.6 percent above basal levels. Two
factors were discussed that may have led to the increase in metabolism: 1) metabolic
costs associated with detoxication and/or 2) the cost of tissue repair resulting from
damage due to ingestion of a toxin. The metabolic cost of either of these processes
is, however, unknown. Lindroth and Batzli (1983) have Shown that dietary tannin can
increase uronic acid levels in the urine by 21-53 times over control levels, thus
suggesting that tannins undergo detoxication via the glucuronic acid pathway.
Detoxication by this method requires ATP and involves oxidation, reduction, or

hydrolysis and conjugation with glucuronic acids or sulphates. Therefore, it is likely

ﬁ'. _

 

68

that at least some of the elevation in metabolism resulting from ingestion of gallic
acid is due to detoxication costs. Unlike McNab’S research (1986), which is based on
correlative data, Thomas et al. (1988) measured a direct link between naturally
occurring dietary constituents and metabolism.

Because much of the focus in mammalian energetics has been on the
measurement of BMRS, the speciﬁc effects of food intake on metabolism are rarely
assessed, since basal metabolism is determined for a resting, fasting animal at
tlrerrnoneutrality (Hill and Wyse 1989). Also, ﬁeld energy budget measurements are
usually calculated for the costs of speciﬁc activities, such as sleeping, walking,
running, or ﬂying. Thermogenic effects of food are usually not taken into account in
these measurements. Energy budgets determined in the ﬁeld by use of the doubly
labeled water method (D20‘8) presumably include energy costs associated with food
intake, but the speciﬁc energy costs associated with food intake are not determined by
this method.

Few ecologically oriented studies have examined the inﬂuence of food intake
(amount of food consumed) and feeding habits (type of food consumed) on the gross
metabolism of mammals. One such study, conducted by Thomas (1984) on
paleotropical fruit bats (Pteropodidae), found that the amount of food ingested by
pteropodids is determined by the protein content of the fruits in their diet. Typically,
these bats ingest up to 2.5 times their body mass every day in fruit. By manipulating
the energy and protein content in their diet, Thomas Showed that pteropodids’ food
consumption is inversely related to the amount of dietary protein available per unit

weight of food.

69
He also compared total food consumption of pteropodids to that of neotropical

fruit bats (Phyllostomatidae) and found that, on a mass-speciﬁc basis, phyllostomatids
ingest half as much energy, apparently being able to supplement their diet with insects
to obtain the necessary protein. Thus, phyllostomatids can escape the nutrient
constraints of a frugivorous diet that is naturally low in protein by supplementally
feeding on a high-protein food source.

In contrast, pteropodids are obligate frugivores and must increase food
consumption to obtain an adequate amount of dietary protein. Although their mean
daily energy intake is 194% higher than that recorded for a comparable phyllosto-
matid bat (either captive or free-ranging), body mass does not increase nor is there a
reduction in the uptake of carbohydrate from the gut. Thus, excess energy intake is
disposed of metabolically, either by diet-induced thermogenesis (proposed by Thomas
1984) or by increased activity. Interestingly, both phyllostomatids and pteropodids
have comparable BMRS (McNab 1982); they differ, however, in their daily energy
budgets, presumably because of their feeding habits (Thomas 1984).

Thomas developed a graphical model (Figure 2.1) that illustrates the hypotheti-
cal relationship between energy and protein intake for neotropical frugivorous bats
that facultatively consume different diets (insect, fruit, or mixed). This model is
based on the ratio between energy and protein intake and predicts that when the
dietary energy/protein ratio is above 199.7 kJ/g protein (61 .3/0.307), protein becomes
a limiting nutrient and is the major determinant of food intake. Below the 199.7
value, bats regulate their intake based on energy needs. At 199.7, energy and protein

needs are satisﬁed by one and the same food intake (mixed diet; Figure 2.1).

 

70

 

Asa. ass: a8. 3%.... egg? Essa Ea .888 a. as. 83...

a 3.8.. 2. as 388. .o u 8... :8 .8... SEE . ”£588 .3 a E... .88.. £98.. :3. a. £3 a :25.
a. a... o... 26.3 388 28. a. 4.8 so»... as. .2. 2.. .8... ea. 2.. .6 .808 388 .3 83538 a 23...
.59 E. ass. 2a.... .8. 3o. a as ea 88: 538.. .8... as... 5 8 2E8. .2 a a". .38. Eats: a2.
3.3 we 8A a .8883? £22.. :3. 2.. 0.23 .3 as a .523? .388 :3. 2.. 52883 a. was:
8.. 8". a... .885 8 ea. 88... a 8 88. 39.3.53. 3.. .3 an. BEEBEE . 52833 88.5.
.2953. 8.858.. a a. 23... a3 .. an... ESE .5 an. a use 532.. 862.. .88. .28.: .. .N 95...".

o... 299... was 8. 2.39...
a. Gamzm 2.... a.» >omwzm
and: H55 Hanna.
5

 

gE-Dgﬂl
. n .

 

 

 

 

3. .6203 T89

(er)
mm Aeuana

 

 

(Eur)

MINI N13103:!

71

Food Habits and Energetic Efﬁciencies in Invertebrates
The relationship between efﬁciency of food utilization and speciﬁc dietary
nutrients is a common tOpic of research in the nutritional ecology of invertebrates.
Researchers in this area tend to focus on measures of gross food consumption, rate of
consumption, and the efﬁciency of conversion of ingested food to body substance, or

ECI (Waldbauer 1968). ECI is calculated as follows:

Body Growth or Production
Ingested Food or Consumption

ECI = x 100

 

Although ECI is not a measure of metabolic rate, energy loss via metabolic heat
production signiﬁcantly contributes to its value. All things being equal (e.g.,
digestive efﬁciency and activity level), as the ECI value decreases, the metabolic rate
of an animal increases.

Mattson (1980) plotted ECI values derived from a number of invertebrate
nutritional studies against the nitrogen content of the consumer’s food supply and
concluded that ECI and the nitrogen content of the food source are signiﬁcantly
related (Figure 2.2). According to Mattson, the data plotted in Figure 2.2 also imply
that, to obtain an adequate supply of dietary nitrogen, food consumption must be
higher for those animals subsisting on diets with low nitrogen availability. To
illustrate the point, he mentions the energy consumption of xylem-sap-sucking insects,
which consume about 100-1000 times their body weight per day of low-nitrogen sap.

Also discussed was the hothouse nrillipede, a mandibulate arthropod with an ECI

 

72

value between 0.3 and 0.8 percent. This arthropod consumes ﬁve to six times its
own body weight in food per day (Mattson 1980).

Mattson examined four major studies of leaf-feeding insects using a covariance
analysis to ﬁnd the source of variation in the relationship described in Figure 2.2.
ECI may be further broken down into two factors: approximate digestibility (AD)

and efﬁciency of conversion of digested food to body substance (ECD). Therefore,

 

60F .

50-

40- . . .
E . '
ESO- o . .o.

. ' o of

20- 9 .

0.: .." .
10- ‘ o...
oo .
.o
'1‘] l l l l l I

 

1 1 3 4 5 6 1 I
NITROGEN CONTENT (s) or PLANT TISSUES

Figure 2.2. Relationship between food nitrogen concentration and the efﬁciency of
conversion of ingested food (ECI) by assorted invertebrate herbivores. (Graph from
Mattson 1980.)

another way to calculate ECI is as follows: ECI = AD x ECD where AD =

(Consumption - Feces)/Consumption and ECD = Growth/(Consumption - Feces).

This formula yields the same result as the formula for calculation of ECI previously

73

mentioned. Mattson’s analysis showed that most of the variation in ECI is due to
ECD, not AD. This means that changes in ECIS associated with changes in plant
nitrogen content are due primarily to post—digestive energy losses, i.e. , respiratory
losses (Mattson 1980). In other words, the fraction of assimilated food that is
oxidized increases as the protein concentration in the diet decreases. As in the fruit-
eating bat study previously discussed, the source of increased metabolism could be
either an intrinsic response analogous to DIT observed in mammals, or an increase in

activity.

Thermogenic Mechanisms and Dietary Protein

In both mammalian and invertebrate research, there is no evidence to support a
claim that activity and dietary protein levels are negatively correlated (although there
is little, if any, research in this area). Thus, by appeal to parsimony, the increase in
thermogenesis evoked by a decrease in dietary protein most likely occurs as a result
of intrinsic physiological mechanisms. Also noteworthy is that it occurs in both
vertebrates (in the forrrr of DIT as seen in laboratory mammals) and invertebrates
(inferred from ECI changes previously described), and thus may be an example of
convergent evolution, where different phyletic groups have adapted to a widespread
shortage or imbalance of nitrogen or another nutrient in their natural food supply.
However, the extent to which DIT occurs in wild mammals is unknown. Except for
the study by Thomas (1984) of bats, the relationship between metabolic rates, feeding
efﬁciencies, and dietary protein levels, and their inﬂuence on the nutritional ecology

of wild mammals, is unexplored.

74

CONCLUSION

The ecological framework above provides a foundation and impetus for DIT
research on wild mammals. Currently, the most popular view in the nutrition
literature is that DIT is a weight regulatory mechanism, not an adaptation to nutrient
stress. Therefore, the focus of my research on DIT represents a perspective that has
received little serious attention. My research places more emphasis on DIT as a
physiological phenomenon that aids in protein uptake rather than weight regulation.
While this perspective has not been widely stressed, there is evidence suggesting that
it deserves more attention. Food generally is such a limiting factor in the wild that it
is difﬁcult to cast DIT evolutionarily in the role of a weight regulatory mechanism.
Obesity is common in humans but less so in wild populations. Those animals that do
store large amounts of fat (usually large animals) normally beneﬁt through increased
insulation and energy storage. A periodic scarcity of resources is common in nature,
and wild animals often rely on stored fat depots for energy when exogenous resources
are limited (King and Murphy 1985). Thus, it is counterintuitive from an ecological
point of view that wild animals evolved a pathway to waste energy for weight
regulation alone. However, if a low-quality, energy-rich food supply is readily avail-
able, nutrient optimization may take precedent and provide the selective pressure for

the evolution of DIT.

CHAPTER III

EXPERIMENTS AND ANALYSIS

INTRODUCTION

The animal chosen for this study is the prairie vole (Microtus ochrogaster), a

small rodent in the family Muridae, subfamily Microtinae. The prairie vole is a

native of North America and is distributed from central Alberta to western West

Virginia and northeastern New Mexico; it can also be found in southwestern Louisi-

ana and southeastern Texas (Nowak 1991) (Figure 3.1). The prairie vole was chosen

 

 

 

 

 

Figure 3.1. Distribution of Microtus
ochrogaster. (Adapted from Hoffmann and
Koeppl 1985.)

75

because it is primarily herbivorous and
consumes large amounts of low-protein
grasses and forbs to satisfy its nutrition-
al needs (Batzli 1985). Also, prairie
vole foodstuffs are not easily digested
due to high ﬁber content, which adds to
the challenge of obtaining adequate pro-
tein uptake from a diet already low in
protein content (Batzli and Cole 1979).
Thus, prairie voles may be highly adapt-
ed to feeding on low—quality foodstuffs

because their natural diets are high in

 

 

76

ﬁber and low in nutrients. This makes them an excellent choice for studying metabol-
ic adaptations to low-protein diets.
Voles, in general, possess traits that appear to be adaptive to a herbivorous

diet, especially a high—ﬁber, low-nutrient diet. High—crowned molars allow macera-

 

tion of tough, ﬁbrous vegetation, and continuously-growing molars compensate for

wear caused by abrasive plant material such as silica. Voles also have an enlarged

 

r4
caecum that functions as a fermentation chamber for digestion of ﬁber by symbiotic
microorganisms (Batzli 1985).
Voles exhibit a compensatory increase in intake when consuming foodstuffs L

that are low in digestibility or nutrient content. This behavioral response is thought to
allow voles to meet energy and nutrient requirements when feeding on low-quality
food (Shenk, Elliot, and Thomas 1970; Batzli and Cole 1979; Batzli 1985). Compen-
satory intake, along with the morphological characteristics mentioned above, may
explain why voles do better on high-ﬁber, low-protein diets than most laboratory
rodents (Shenk et al. 1970; Batzli 1985). However, even though voles are apparently
well-adapted to herbivory, spatial and temporal variation in microtine densities in the
wild may be linked to the quality of available forage (Batzli 1985). Cyclic declines in
vole population densities occur every three to four years and may be related to low
food quality encountered by voles during population peaks (Bergeron and J odoin
1987).

If DIT occurs in small mammalian herbivores such as prairie voles, the ability
to oxidize excess food intake to obtain an adequate supply of nitrogen would be ad-

vantageous. This could allow them to increase their consumption of low-nitrogen

77

foodstuffs and lose the excess caloric intake as heat, thus increasing nitrogen intake.
An intrinsic physiological response to low-protein intake such as DIT would be
advantageous during times when the nutritional quality of resources is declining, such
as during late summer, fall, and winter (Chapter H). In this view, energy is not a
limiting resource for the prairie vole. Instead, the amount and quality of nitrogen in
the diet places a limit on survivorship, reproductive success, and ﬁtness. DIT would
have considerable adaptive value by allowing an animal to consume an adequate level
of nitrogen from foodstuffs low in nitrogen. This would be especially advantageous
during gestation and periods of rapid growth, when nitrogen is essential for protein
production.

The aim of my research is to ﬁnd out if DIT occurs in voles and, if so, how it
might be adaptive under natural conditions. To accomplish this, voles at different
stages of development were subjected to a variety of low- and high-protein diets, and
their physiological and behavioral responses were recorded. Average daily metabolic
rate (ADMR) and energy efﬁciency were measured using material balance and bomb
calorimetry methodology; daily food intake was also recorded in one experiment.
Biochemical change in brown fat was determined by GDP binding assay and used as
an indicator of thermogenic activity of the tissue. Changes in the content of lipid and
protein body tissue were studied to see if low metabolic efﬁciencies and/or increased
food intake induced by a low-protein diet affected tissue deposition. Lipid content
was measured by ether extraction, and protein levels were determined by Kj eldahl

analysis.

 

 

78
Weanling voles (21 days old) were used in all experiments except for the ﬁrst

experiment. Since the growth rate of weanling voles is high, their protein require-
ments are also high. Compared to fully grown voles, it is probable that weanling
voles experience a greater amount of stress when dietary protein is limiting, since
they require more protein to sustain growth. Therefore, if DIT is an adaptation that
aids in protein uptake from low-protein rations, weanling voles are more likely than k ,_
adult voles to exhibit DIT in response to consuming a low-protein diet. This is the

main reason most of the voles used in this research were weanlings.

 

If DIT is ecologically important to small herbivores, then voles on low-protein
diets, when compared to those receiving high-protein rations, should Show 1) low
metabolic efﬁciencies, suggesting an ability to oxidize foodstuffs in excess of basic
energy requirements (Kevonian et al. 1984), 2) heightened metabolic rates and brown
adipose tissue (BAT) activity resulting from an increase in thermogenesis via the
proton conductance pathway in BAT (Swick and Gribskov 1983), 3) a hyperphagic
response, suggesting that voles increase food intake based on nitrogen requirements
(Tulp et al. 1979; Donald et al. 1981), and 4) conservation of body protein and low
deposition of body lipid, suggesting an intrinsic regulation of protein and lipid levels
in the body during consumption of low-protein foodstuffs (Donald et al. 1981).

Seven experiments were conducted to investigate the occurrence of DIT in
prairie voles and the impact of a low-protein diet on their metabolism and behavior.
The experiments consisted of three material balance studies, three experiments

involving GDP binding assays, and an experiment that measured daily food consump-

 

79

tion and changes in body tissue (lipid and protein content). The following is a

summary of those experiments and an analysis of their results.

EXPERIMENT 1

Introduction

Experiment 1 was exploratory. Since research on DIT had not been done
using M. ochrogaster, many basic parameters were not known. Two primary factors
that had to be determined were the total dietary ingredients that would sustain voles
and be palatable, and the level of protein in the diet that might elicit a DIT response.
Also, from the results of experiment 1, the inﬂuence of age and growth rate on
metabolic rate in response to dietary protein could be determined; this information
was critical for designing subsequent experiments. The main goal of experiment 1
was to see if the efﬁciency of energy retention and metabolic rate differ between low-

and high-protein groups.

Materials and Methods
In experiment 1, 18 voles,’ ages 53-55 days, were divided into two groups and
individually housed in metal, wire-bottomed metabolic cages, in a room maintained at
22°C with a 16-hour light/8-hour dark cycle. They were fed diets with casein
contents of 7% and 17% for low- and high-protein groups, respectively. Diets were

formulated for approximately equal metabolizable energy (MB) per gram of nutrients

 

SThe prairie voles used in this research came from a Michigan State University
colony. This colony was established from animals obtained from another colony at
the University of Illinois at Urbana-Champaign, Urbana, IL.

 

80

(excluding casein and glucose, which were variables in the diet). The estimated ME
for casein is 16.7 kJ/g, while that for glucose is 15.2 kJ/g (Kevonian et al. 1984).
Therefore, 100 g of 7% casein has approximately the same estimated ME as 99 g of
the 17% casein diet. Diet ingredients were similar to those used by Kevonian et al.
(Table 3.1).

Table 3.1. Composition of diets, expt. 1. Mean ME intake Per vole

 

 

 

 

 

Ingredients 7% casein 17% casein was different between groups (Table
Casein’ 7-0 17-0 3.2) (P < 0.001, ANCOVA). This
Glucose 74.3 63.3 . .
difference IS probably due to a
Cellulose2 4.0 4.0
Corn 011 10,0 10,0 difference in gross food consump-
. . . ,
Vltamm m 1'0 1'0 tion. Because semisynthetic diets
Choline bitartrate‘ 0.2 0.2
Mineral mix‘ 35 3.5 have a high digestibility (e.g.,
100.0 g 99.0 g Hamilton 1939b), ME intake for

 

animals fed these diets can often be

lHigh protein from U.S. Biochemical Corp., Cleve-
land, OH. 2Microcrystalline cellulose from Dyets,
Inc., Bethlehem, PA. 3AlN-76 from U.S. Bio- .
chemical Corp. ‘From Bio Serv, Frenchtown, NJ. f°°d consumption 03- Romsos,
5AlN-76 from Dyets, Inc. “Determined by bomb

calorimetry. personal communication). Meta-

used to estimate differences in gross

bolic rates are affected by food consumption because when one animal consumes
more food than another, the animal that consumes the most food will also incur the
highest energy costs associated with food intake, such as the energy costs associated
with mastication, digestion, transport of nutrients in the body, and biosynthesis of
storage compounds. Therefore, ME intake was used as a covariate in the analysis of

metabolic rate data derived from this experiment and experiment 3, which also had

81
unequal ME intakes. ANCOVA means for metabolic rates were adjusted so that each

animal was compared at the same ME intake level, thus reducing the inﬂuence of ME

intake on metabolic rates. This permitted a more accurate comparison of differences

in metabolic rates that may be caused by differences in dietary protein levels.
High-protein casein was used for the nitrogen source; variation in energy

content of the diets resulting

 

 

 

 

 

 

 

from different casein levels was Flow of may in the Body

adjusted by changing the amount hm” ,

of glucose in the diets. All other 1 '7'. I N.” J

ingredients—lipid, cellulose, g

vitamins and minerals—were held l 4 I "Me-v I

constant (Table 3.1). The ani-

mals were fed ad libitum for 21 l 4 E:
on

days. Metabolic heat production

 

 

 

was measured by material bal-
_ Figure 3.2. The fate of ingested energy in an or-
ance methods, where Ingested ganism.
energy - waste energy :1; energy stored/ lost = thernric energy or energy of metabo-
lism. However, ingested energy was not measured directly, since there was some
spillage of food. Based on the chart of energy ﬂow in the body (Figure 3.2), it
follows that ME intake :1; energy stored/ lost = thermic energy or metabolic heat
production. Therefore, ME intake was used to calculate energy of metabolism.

ME intake was calculated by subtracting the energy losses in the feces, urine,

and spilled food, from gross energy intake (total food removed from feeder). Feces,

82

urine, and spilled food (collectively termed as ‘waste’) were collected every three
days. They were dried to a constant weight at 65°C, homogenized in a blender, and
stored in the freezer in airtight containers until their energy content could be mea-
sured. Using bomb calorimetry (Parr Instruments, Moline, IL), the energy content of
urine, feces, and spilled food was determined for a 0.5-1.1 g aliquot‘5 of the homoge-
nized waste mixture of each animal. In some instances, when urine production was
large, urine was pipetted on cotton and bombed to determined its energy content (the
energy content of cotton, determined separately by bomb calorimetry, was subtracted
from the total energy recorded for the cotton/ urine combination).

Body energy gain was calculated by subtracting ﬁnal body energy from initial
body energy. Initial body energy was estimated by linear regression of body energy
on body weight of voles of similar age and weight as the experimental groups (see
footnote to Table 3.2). This method was used in all the material balance experiments.
DETERMINATION OF BODY ENERGY

Voles were killed by C02 asphyxiation and their carcasses were homogenized
in water, dried to a constant weight at 65°C, and stored in airtight containers. Body
energy was determined by bomb calorimetry of a 0.2-0.9 g aliquot7 of the dried tissue

homogenate of each animal.

 

°Coefﬁcients of variation (CV) were determined for waste sampling. Three
waste aliquots per sample (mean = 0.67 g) from three different waste samples were
used. CVS ranged from 0.011 to 0.019.

7Coefﬁcients of variation were determined for tissue sampling. Three tissue
aliquots per sample (mean = 0.48 g) from three different tissue samples were used.
CVS ranged from 0.016 to 0.028.

83

Table 3.2. Metabolic rate, energy intake, and retention, expt. l.

 

% casein in diet

 

 

Low High

7% 17%
Initial body wt, g vole'l 29.3 :1: 0.9 29.4 i 1.0
ME intake,l kJ vole'l 1394.1 :1: 86.6 1612.5 :t 83.5.
Body wt gain, g vole" 1.7 i 1.0 3.0 i 1.0
Body energy gain,2 kJ vole" 86.2 :1; 33.2 78.6 i 42.1
Efﬁciency of energy retention,3 % 6.2 :1; 1.9 4.9 :1; 2.5
ADMR,"5 kJ d“vole'1 66.0 :1; 1.6 68.8 :1; 1.7
Wt speciﬁc ADMR} kJ d"g"vole’l 2.1 :1; 0.1 2.3 i 0.1

 

 

Values are means i SEM for 8-9 voles fed for 21 days. 1ME, metabolizable energy.
2Final body energy minus that estimated at the beginning of the experiment. Initial
body energy was estimated by linear regression of body energy (y) on body weight (x)
of 11 voles weighing 24-38 g (y = 16.0x - 151.2, 72 = 0.6). 3Body energy gain
divided by ME intake times 100. ‘ADMR, average daily metabolic rate. 5ANCOVA
adjusted least squares means, using ME as a covariate. ‘P < 0.001, ANCOVA.

Results

The data for experiment 1 are compiled in Table 3.2. All means were tested
for statistical differences and only ME intake was found to be signiﬁcantly different
(P < 0.001, ANCOVA). Differences between means for initial body weight, body
weight gain, and body energy gain were compared statistically using Student’s t test.

ME intake was not the same for both groups (Table 3.2); therefore, it was
used as a covariate in the calculation and statistical testing of ADMRS (kJ d"vole")
and weight-speciﬁc ADMRS (kJ d“g“vole“), for reasons previously discussed. Meta-
bolic rates were regressed on ME intake to test for homogeneity of slopes (Figure

3.3). The probability for signiﬁcant interaction between ADMRS and the covariate

 

 

 

90 r r r T

80

70

ADMR (k1 / day)

60

 

 

 

50
900 1120 1340 1560 1780 2000

ME (kJ)

 

 

 

 

Figure 3.3. Regression of the ADMR of low— and high-
protein groups on the covariate ME intake to test for ho-
mogeneity of slopes in an analysis of covariance

(AN COVA) design. Slopes were not statistically differ-
ent. P = 0.765.

(ME intake) is 0.765 (P
= 0.510 for weight-spe-
ciﬁc ADMRS), so the as-
sumption of homogeneity
of slopes is reasonable.
Covariate analysis of
ADMRS showed no statis-

tical difference (Table

 

3.2). Efﬁciency of ener- é
gy retention (body energy

gain divided by ME in-

take times 100), a param-

eter that is commonly

used by nutritionists in

energy balance studies, also did not differ. The Mann-Whitrrey U test was used for

the statistical analysis of efﬁciencies (percentage data). Statistical testing was done

with SYSTAT statistical software (Wilkinson 1989).

The lack of difference in either ADMRS or efﬁciency of energy retention

between treatment groups may have been due to the age of the animals used. Growth

rates in the age span employed for voles may have been too low and variable to be

affected by low-protein intake. These results were not unexpected since it was

unlikely that older animals with slow growth rates would be under protein stress at

the levels of protein provided in the experimental diets.

85

Since it was desired to experiment on voles at different stages of development,
adult voles were used in experiment 1. One of the problems of using older animals in
energy balance experiments where efﬁciency of energy retention is being measured, is
that it is difﬁcult to detect differences between treatment groups if growth is limited.
If there was a difference in metabolic rates or efﬁciencies, it may have been too small
to detect.

Another possibile reason a DIT response was not observed is that older voles
with low growth rates have less protein requirements for growth and therefore are less
likely to exhibit a DIT response (assuming DIT is adaptive for increasing protein
intake). Mean total growth was low in both groups for this experiment. It may be
that the level of protein in the 7 % casein diet was too high to have a negative impact
on growth. As a result, the protein level in the 7% casein diet may have been too

high to stimulate DIT (again, assuming DIT is adaptive for increasing protein intake).

EXPERIMENT 2

Introduction
Experiment 2 was designed to test the effects of low- and high-protein diets on
the metabolic rates and efﬁciencies of weanling voles. The weanling voles used in
this experiment have a much higher growth rate than the adult voles used in experi-
ment 1. Since protein is required for growth, it is possible that growing voles
consuming low-protein diets may be protein limited. Therefore, they would be more
likely than high-protein-fed voles to exhibit DIT (an increase in metabolic rate and

decrease in energy efﬁciency per unit ME intake), if DIT augments protein uptake.

86
Materials and Methods

In this experiment, 18 weanling voles, 21 days old, were used. Voles were
divided into two groups and individually housed in metal, wire-bottomed metabolic
cages, in a room maintained at 22°C with a 16-hour light/8-hour dark cycle. The
ﬁber content in the diet was increased to 20 percent to simulate the high ﬁber content
found in the natural diet of free-ranging voles (Fable 3.3). Voles exhibit improved
performance (increased weight gain and more efﬁcient nutrient utilization) when ﬁber

is added to artiﬁcial diets, indicating an apparent need for bulk (Shenk et al. 1970).

 

Based on the American Institute of Nutrition guidelines (1977) for laboratory [E
animal diets, cornstarch was added to the diet to provide a complex carbohydrate
source. The ratio of glucose to cornstarch was held constant for both groups. The
amount of dietary casein for low- and high-protein groups was reduced to 5 % and
15 % respectively. Compared to

experiment 1 diets, lower casein Table 3.3. Composmon of diets, expt. 2.

 

 

. . I ed' ts 5% ° 15% '
levels made experiment 2 diets "gr lea casein case")
Casein 5.0 15.0
more Similar 1n casein content to Glucose 42.2 35.2
diets used in previous DIT exper- ComSWCh 13-1 15-1
. . Cellulose 20.0 20.0
1ments (e. g., Kevonian et al.
Corn oil 10.0 10.0
1984). Also, it was thought that Vitamin mix 1,0 1,0
a reduction in casein for the low- 010"” bitamate 0'2 0‘2
Mineral mix 3.5 3.5

 

protein group might be required 1000 g 100.0 g

to induce a DIT response. Diets Gross energy, kJ/g 18.4 19.2

87

were formulated on a g ingredients /100 g mixture basis (Table 3.3). Voles were fed
ad libitum for 21 days.

ME intake, metabolic rates, and energy efﬁciencies were determined using the
same material balance methodology described for experiment 1. Linear regression
was used to estimate initial body energy (see footnote, Table 3.4). Preparation of
body tissue homogenates and collection of waste material were done by the same
methods described in experiment 1. Sampling procedure using single aliquots of
tissue (0.2-0.6 g) and waste (0.8-1.1 g) homogenates was also the same. Bomb

calorimetry was used to measure the energy content of aliquots of body tissue and

 

waste.

Differences between means for initial body weight, ME intake, body weight
gain, body energy gain, ADMR, and weight-speciﬁc ADMR were compared statisti-
cally using Student’s t test. The Mann-Whitney U test was used to compare efﬁcien-
cies. ME intake was not signiﬁcantly different between groups (Table 3.4). There-

fore, it was not used as a covariate when metabolic rates were compared statistically.

Results
Experiment 2 failed to show a signiﬁcant difference in metabolic rates or
energy efﬁciencies (Table 3.4). However, an interesting growth pattern was ob-
served. The 5% casein group achieved growth similar to the 15 % casein group by
the end of the experiment (Figure 3.4). Growth was initially slow for the low-protein
group, but after 21 days, body weights were not statistically different. Change in

body energy did not differ, suggesting that the relative composition of the body (fats,

88

Table 3.4. Metabolic rate, energy intake, and retention, expt. 2.

 

% casein in diet

 

 

 

 

Low High
5% 15%
Initial body wt, g vole’1 19.3 :t 0.4 20.2 :1: 0.5
MB intake, kJ vole‘l 1644.3 :1: 93.7 1584.7 2%: 75.1
Body wt gain, g vole'1 7.9 j; 1.0 8.5 :t 1.1
Body energy gain,1 1:] vole’1 229.1 1; 23.8 212.2 :1: 30.1 L.
Efﬁciency of energy retention, % 13.9 d: 1.3 13.4 i 1.6
ADMR, kJ d"vole‘l 67.5 :1: 4.0 65.4 :1; 3.0
Wt speciﬁc ADMR, kJ d“g"vole" 2.9 i 0.1 =..= 2.7 :1: 0.11

 

Values are means :1: SEM for 9 voles fed for 21 days. ‘Initial body energy was esti-
mated by linear regression of body energy (y) on body weight (x) of voles weighing
17-21 g (y = 13.61: - 137.3, 1‘2 = 0.79). None of the values listed are signiﬁcantly
different.

proteins, and water) was similar for both groups. The low-protein group could have
increased mass by increasing lipid deposition, but since lipid has twice the energy/ g
of tissue as protein, this probably would have been revealed by an increase in total
body energy.

Because of the initial slow growth rate in the 5% group, and the subsequent
marked increase in growth rate occurring sometime after seven days into the experi-
ment, it was hypothesized that there may have been an acute DIT response—possibly
during the ﬁrst 10 days of the experiment when growth was slow—which was
obscured by an increased efﬁciency during the latter part of the experiment and
therefore not detected. It is also possible that the level of protein fed to the low-

protein group was still too high to elicit a measurable DIT response. Therefore, a

89

 

 

 

 

 

 

30 u T u
15 I- d
3
10 I- -1
A 15! and:
15 1 l 1' l U 3. (hub
11 21 33 41
A03 (an)

 

 

 

Figure 3.4. Mean daily growth of weanling voles,
expt. 2. Error bars are :1: SEM.

third experiment was designed
to test the inﬂuence of a lower
level of dietary protein on

metabolism and growth.

EXPERIMENT 3

Introduction
Since it was possible
that a 5 % casein level was too
high to cause DIT in voles

(experiment 2), a 2.5% casein

group was used in experiment 3 to see if a lower level of dietary protein could

stimulate DIT. The 5% and 15% casein diet treatments used in experiment 2 were

retained, making a total of three diet treatment groups used in experiment 3 (Table

3.5). Also, since an increase in the efﬁciency of energy retention after age 27 days

(suggested by an increase in the growth rate after seven days, Figure 3.4) could have

prevented detection of a DIT response occurring early in the experiment when energy

retention may have been less efﬁcient (suggested by the low growth rate during the

ﬁrst seven days, Figure 3.4), the duration of experiment 3 was limited to 10 days.

Thus, experiment 3 differs from experiment 2 by being shorter in duration and

having an additional diet group lower in protein than the 5 % casein treatment group

of experiment 2. Experiment 3 was designed to test the effects of low-, medium-,

 

1‘...

90

Table 3.5. Composition of diets, expt. 3.

 

 

 

 

Ingredients 2.5% casein 5% casein 15% casein
Casein 2.5 5.0 15.0
Glucose 44.0 42.2 35.2
Cornstarch 18.8 18.1 15.1
Cellulose 20.0 20.0 20.0
Corn oil 10.0 10.0 10.0
Vitamin mix 1.0 1.0 1.0 L.
Choline bitartrate 0.2 0.2 0.2
Mineral mix 3.5 3.5 3.5
100.0 100.0 100.0
Gross energy, kJ/g 18.4 18.5 19.1 it

 

and high-protein diets on the metabolic rates and efﬁciencies of weanling voles over a

feeding period of 10 days.

Materials and Methods

In experiment 3, 27 voles were divided into three groups and fed diets with
casein contents set at 2.5%, 5%, and 15% (Table 3.5). Preparation of body tissue
homogenates and collection of waste material were done by the same methods
described in experiments 1 and 2. Sampling procedure using single aliquots of tissue
(0.2—0.6 g) and waste (0.9-1.4 g) homogenates was also the same. Initial body energy
was estimated by linear regression of body energy on body weight (see footnote to
Table 3.6). Voles were fed ad libitum and housed under the same conditions as in

previous experiments.

91
Statistical testing of means for initial body weight, body weight gain, and body

energy gain were done by ANOVA. Post hoc comparisons of means were analyzed
using Tukey’s HSD test. Means for efﬁciency of energy retention were compared
using Kruskal-Wallis analysis of ranks (percentage data). Post hoc comparisons were
done using the Mann-Whitney U test; differences were considered signiﬁcant for P <
0.05/3 (Bonferroni correction). ME intake between groups was signiﬁcantly different
(P < 0.001, ANCOVA) and therefore was used as a covariate when metabolic rates
were tested statistically by ANCOVA. The assumption of homogeneity of slopes was
tested for metabolic rates analyzed by ANCOVA; slopes for ADMRS and weight-
speciﬁc ADMRS regressed on ME intake were not signiﬁcantly different (P = 0.10
for both). Post hoc comparisons of ANCOVA adjusted least squares means for

ADMRS and weight-speciﬁc ADMRS were analyzed using Tukey’s HSD test.

Results

The metabolic and growth data are presented in Table 3.6. They show major
differences in metabolism and weight change between the 2.5 % casein group and the
other two groups, indicating a DIT response by the 2.5% casein group. High-protein-
fed voles are more than four time as efﬁcient at retaining energy as low-protein-fed
voles. ADMRs (adjusted for ME intake by ANCOVA) for the low-protein group
were almost 9 percent higher than the medium-protein group ADMRS, and over 16
percent higher than ADMRS recorded for the group fed the high-protein diet. The
trend is similar and even more dramatic for weight-speciﬁc ADMRS, with voles fed

low-protein diets having over a 40 percent higher weight-speciﬁc metabolic rate than

 

92

 

.28anme 8: u mz

5585 55.. u m ”5385 5585 u 2 ”5225 35 u .— 586 v m... .55 v 5.. .85 v .5. 955.60 a 3 m2 55.5
.2888 8556a 38— 8833 <>OUZ<_ and .I. «a £62 - 6.2 H .o m 3-5 wﬁnwmoa 86> .8 3 Emma? .33 :o S
.388 32... .«o .8585on .825 .3 33:58 83 .388 .33 RES .926 3 58 8.5 83> m 58 2mm H 8.85 8.8 85:5

 

 

 

in: 555 US do H W 3 H 2“ 3 H «H 55375-5 2 .529. 058% .3
:5 .25 3 H 3m 3 H 93 I H 9% .8335. 2 229..
:55 .25 3 H 5.: 3 H ad ad H 3 5 .8558. 3.08 8 58505.5
.5: 5:5 .25 8.2 H S: 35 H 5.3 5.2 H QR .59, a .58 3.58 555
:52 5:5 525 3 H on no H 3 to H E- .52, m .58 E .585
.52 .5 in H mo; 3: H as: 35 H 3% 752, a .355 m2
m2 to H «.2 3. H we no H 92 .52, m .E 588 555

888855 e 2 g m R 2

585585 555 5582 33

55.. 5 59.8 5

 

.m .55 .8555. 55 .0555 5.58 .55 058502 .55 55:.

93
the high-protein-fed voles. These types of metabolic responses, (low energy efﬁcien-

cy and heightened metabolic rate) strongly suggest that voles increase metabolic heat

 

 

2‘ t I l I I
24- -
22- -

WEIGHT (3)
3
k

 

 

 

1| - -
A 15’ Casein
O 55 Casein
1‘ I I I I I I I 2.5% Casein
21 23 25 21 29 31
A03 (den)

 

 

 

 

Figure 3.5. Mean daily growth of weanling voles, expt. 3. Error bars are ;t SEM.

production per unit of ME intake in response to low-protein feeding. Clearly, this is
indicative of a low-protein DIT response analogous to those recorded for laboratory
rats consuming low-protein diets (Chapter 1).

Growth patterns, shown in Figure 3.5, were different in all three groups, re-
vealing the importance of dietary protein in maintaining growth. Note that in the 5%
casein group, the rate of growth increased sharply after eight days (29 days old),

displaying a growth response similar to that observed in experiment 2. The 2.5%

94

casein group experienced weight loss during the ﬁrst six days and then almost no

growth thereafter.

EXPERIMENT 4

Introduction
Experiment 4 was undertaken to determine the physiological source of the
thermogenic response shown by voles on a 2.5 % casein diet in experiment 3. This
experiment was designed to assess the thermic activity of BAT in voles consuming
2.5% and 15% casein diets by measuring the GDP binding activity of the tissue. An
increase in GDP binding is known to occur in rats consuming a low-protein diet
(Swick and Gribskov 1983), and the assay is thought to be an indirect measure of

BAT thermogenesis (Nicholls 1976).

Materials and Methods

Weanling voles were divided into two groups with 14 animals in each group,
individually housed under the same conditions as those speciﬁed for experiments 1
through 3, and fed 2.5% and 15% casein diets for 10 days. Diet ingredients were the
same as those listed in Table 3.5 for 2.5% and 15% casein groups. Voles were killed
by cervical dislocation, and the interscapular BAT was quickly dissected and placed in
a chilled buffer containing 0.25 M sucrose and 5 mM K-TES (pH 7.2). The tissue
was then homogenized in glass homogenizing tubes with a Teﬂon pestle. Mitochon-
dria were isolated by a procedure similar to that described by Cannon and Lindberg

(1979), using successive centrifugations. Aliquots of mitochondrial suspensions were

95
assayed for protein content by a modiﬁed Lowry method using the Folin-Ciocalteu

reagent and spectrophotometric measurement of absorbance (after Markwell et al.
1981). Bovine serum albumin was used as the standard.

The association of 3H-GDP (radioactively labeled GDP) with BAT mitochon-
dria was measured by the method described by Nicholls (1976), with some modiﬁca-
tions. Mitochondria (0.5 to 1.0 mg mitochondrial protein/m1) were added to an
incubation medium at room temperature for seven minutes. Medium reagents were
100 mM sucrose, 1 mM EDTA, 20 mM K-TES, 2 uM rotenone, 100 pM potassium

atractyloside, and 2.5 x 10" dpm/ml 3H-GDP (10.2 Ci/mmol, New England Nuclear,

 

Boston, MA); 5.55 x 10’ dpm/ml MC-sucrose (673 mCi/mmol, New England Nuclear)
was added to estimate the volume of medium trapped in the ﬁnal mitochondrial pellet.
The incubation medium was then divided in half; 200 uM of unlabeled GDP was
added to one solution to estimate nonspeciﬁc binding of 3H-GDP. To the second
solution, IOuM of unlabeled GDP was added to determine total binding of 3H-GDP.
(Speciﬁc binding of 3H-GDP was calculated by subtracting the amount of 3H-GDP
bound to nonspeciﬁc sites plus that trapped in the ﬁnal pellet, from the total amount
bound.) Tubes were quickly centrifuged at the end of the incubation period. The
supernatant was aspirated and the mitochondrial pellet was dissolved using Beckman
tissue solubilizer (BTS-450). The dissolved pellet was then suspended in a scintilla-
tion cocktail8 and counted in a scintillation counter (TRI-CARB 4430 Liquid Scintilla-

tion System, Packard Instrument Co. Inc., Downers Grove, IL).

 

l"The scintillation cocktail was mixed using the following ingredients: 500.0 ml
Triton X-100, 1000.0 ml toluene, 7.5 g PPO, and 150.0 mg DHPOPOP.

96

Results

The speciﬁc binding for the low-protein group was 976 j; 62 pmol/mg mito-
chondrial protein (:1; SEM); for the high-protein group it was 1149 j: 86 pmol/ mg.
This was signiﬁcantly different at P < 0.05, based on a randomized complete block
ANOVA statistical design, with blocking on assay days. While these differences are
statistically signiﬁcant, there may not be much biological importance in the differenc-
es. Both groups had very high levels of GDP binding—much higher than previously
recorded for laboratory rats exhibiting DIT—indicating a high level of BAT activity
for both groups. These values fall well within the range of GDP binding measure-
ments taken by Klaus, Heldmaier, and Ricquier (1988) for winter-acclimated wild
species. For example, seasonal peak mean values recorded for Clethrionomys were
1337 j: 67 pmol/mg; for Apodemus they were 1164 i 101 pmol/mg (values are :1;
SEM) (Klaus et al. 1988). Presumably, the high values recorded by Klaus et al. were
due to the effects of cold on nonshivering thermogenesis (NST). While high GDP
binding values recorded for the low-protein group in experiment 4 could have
occurred because of DIT, the cause of the similarly high values recorded for the high-

protein group cannot easily be explained.

EXPERIMENT 5

Introduction
Experiment 4 showed that voles on a 2.5% casein diet had high speciﬁc GDP
binding levels, which were indicative of heightened BAT activity. However, this

experiment was inconclusive because the 15% casein group also had high speciﬁc

 

97
GDP binding. One possible explanation for these surprising results is that both

groups may have been thermoregulating by NST. If NST was occurring, it could
have affected the activity of BAT and, hence, the binding of purine nucleotides.

The ambient temperature during experiment 4 was approximately 22°C.
According to Webster (1983), at 24°C there is a 36 percent increase in metabolic heat
production by mice housed in a metal calorimeter when compared to mice housed in
normal cages. The voles in experiment 4 were housed in wire-bottomed metal cages,
which would make them highly susceptible to increased heat loss by conduction and
convection. Interestingly, when I ﬁrst began using wire-bottomed metabolic cages, I
measured the metabolic rates of C3H laboratory mice housed in these cages to test
material balance methodology. During the test, three mice became hypothermic and
had to be revived in an incubator. It may be that the voles in experiment 4 were
using NST to maintain body temperature, and the results of the assay were clouded by
elevated GDP binding levels due to NST. Because DIT and NST both cause an
increase in GDP binding, the effects of DIT on GDP binding could be masked if NST
is appreciably active.

Therefore, the goal of experiment 5 was to see if voles held at the same
temperature but under different caging conditions would show a difference in GDP
binding to brown fat mitochondria. The hypothesis was that voles housed singly in
wire-bottomed metal cages should be under greater cold stress because they are more
susceptible to conductive and convective heat loss than animals housed as a group in
plastic cages supplied with bedding, since group-housed voles could huddle and be

more thermally insulated. As a result of the greater cold stress, voles housed in metal

98
cages may exhibit NST and therefore have higher GDP binding levels than voles

housed in plastic cages.

Materials and Methods
To test the hypothesis that N ST may have caused a substantial increase in
GDP binding in voles placed in metabolic cages at T, = 22°C, housing status was
used as a variable and GDP binding was compared. Voles housed as a group in
plastic cages supplied with bedding were compared with those housed individually in

metabolic cages. There were 12 animals in each treatment (single or group). Both

 

treatment groups received rabbit chow ad libitum and were maintained at an ambient
temperature of 22°C for 10 days. GDP binding was determined by the same proce-
dure used in experiment 4. Data were tested using a randomized complete block

ANOVA, with blocking on assay days.

Results

The mean GDP binding value recorded for group—housed voles was 961 j; 107
pmol/ mg mitochondrial protein (1: SEM). Voles housed individually showed a
speciﬁc binding of 1712 j: 102 pmol/mg (i SEM). These values were signiﬁcantly
different at P < 0.001.

The results of this experiment suggest that housing status can inﬂuence the
speciﬁc binding of GDP to BAT mitochondria, presumably as a result of thermoregu-
latory costs. More important, high levels of GDP binding due to NST in voles
housed individually in metabolic cages at 22°C could obscure differences induced by

diet.

99

EXPERIMENT 6

Introduction

Homeotherms at thermoneutrality can regulate their body temperature by
nonevaporative physical processes without increasing their metabolic rate. The range
of ambient temperatures this occurs in is called the thermoneutral zone (TN Z), which
has boundaries set by upper and lower critical temperatures (Bligh and Johnson 1973).
Once the ambient temperature ('1‘.) drops below the lower critical temperature of the
TNZ, a homeotherm must elevate its rate of internal heat production to preserve a
constant body temperature. This is usually accomplished by an increase in shivering
or nonshivering thermogenesis. At temperatures just below the TNZ, NST is usually
the dominant form of heat production in wild mammals, whereas shivering is activat-
ed at much lower temperatures (Jansky 1973).

NST induced by cold acclimation in small mammals has been linked to BAT
thermogenesis (Foster and Frydman 1978), and BAT mitochondria of cold-acclimated
animals exhibit an increase in purine nucleotide binding (Desautels, Zaror-Behrens,
and Himms-I-Iagen 1978). Since GDP binding by BAT mitochondria is used as an
indicator in testing for both DIT and NST, the potential for confusion exists, especial-
ly when GDP binding is used as an assay in energy balance studies where animals are
kept at temperatures below thermoneutrality. The results of experiments 4 and 5
suggest that the cool ambient temperature in the laboratory (22°C), coupled with poor
insulative properties of metabolic cages, induced NST and caused the high speciﬁc

binding of GDP to BAT mitochondrial protein observed at both dietary protein levels

100

in experiment 4. Experiment 6 was conducted to examine the inﬂuence of dietary
protein on speciﬁc binding of BAT mitochondria when voles are held at

thermoneutrality.

Materials and Methods
Weanling voles were individually housed in plastic cages supplied with

bedding and placed in an environmental chamber. T, was set at 28°C, a temperature
close to the lower critical temperature for prairie voles’ (W under, Dobkin, and
Gettinger 1977). They were maintained at a 16-hour light/8-hour dark cycle and fed
diets with casein contents of 2.5% and 15%, with eight animals in each group. Diet
ingredients were the same as those listed in Table 3.5 for 2.5% and 15% casein
groups. After 10 days, the animals were sacriﬁced and speciﬁc GDP binding by

brown fat mitochondria was assayed as in experiments 4 and 5.

Results
Means for speciﬁc binding of GDP were 524 j; 72 and 712 i 86 pmol/mg
for low- and high-protein groups, respectively. These values were not signiﬁcantly
different (P = 0.12), based on Student’s t test of independent sample means. A com-

plete block ANOVA was not used for this analysis as in previous experiments,

 

9An exact value for the lower critical temperature was not determined for the
voles used in this experiment. The lower critical temperature for an animal can vary
depending on an animal’s age, size, pelage, and prior acclimation or acclimatization
history. In the study by Wunder, Dobkin, and Gettinger (1977), it was assumed that
27 .5°C was the lower critical temperature for wild-caught prairie voles. However,
not all the animals tested in that study were at thermoneutrality at 275°C. Since an
exact value for the lower critical temperature was not known for the animals used in
experiment 6, 28°C was arbitrarily chosen, which is slightly higher than the one
chosen by Wunder et al. (1977).

101
because data recorded on separate assay days were not signiﬁcantly different (P =

0.10).

These data suggest that the content of protein in the diet does not inﬂuence
BAT activity via activation of the proton conductance pathway when voles are tested
at thermoneutrality. It is not known if metabolic rates and efﬁciencies differed
between the two groups, since an energy balance study was not performed at this
temperature. Also, there is little information in the literature concerning the inﬂuence
of dietary protein on the metabolism of small mammals at thermoneutrality, although
some experiments involving cafeteria feeding of rats at 29°C show that DIT and GDP
binding are suppressed (Barr and McCracken 1982; Rothwell and Stock 1986b).

The GDP binding data for voles provide no evidence that the low metabolic
efﬁciencies of the 2.5 % casein group recorded in experiment 3 resulted from activa-
tion of the proton conductance pathway in vole BAT mitochondria. Furthermore, it is
likely that NST contributed signiﬁcantly to the high levels of GDP binding observed
in experiments 4 and 5, since GDP binding is much lower for voles at thermo-
neutrality. However, comparisons of GDP binding data between experiments may be

inappropriate, due to the sensitivity of the assay to local conditions.

EXPERIMENT 7

Introduction
Food consumption, gross nitrogen efﬁciency, and body tissue changes were
measured in this experiment. Food consumption was measured to determine the

relationship between food protein content and feeding behavior. Tulp et al. (1979)

102

and Donald et al. (1981) noted an increase in weight-speciﬁc food consumption for
rats fed low-protein diets when compared to those on high-protein diets, and Shenk et
al. ( 1970) observed that meadow voles increase intake when dietary protein levels are
low.

Nitrogen efﬁciency data will give a general idea of the relationship between
dietary protein content and the efﬁciency of its utilization in the body. The data will
be limited though, because net nitrogen efﬁciency was not measured.

Data on body composition will show how tissue constituents are affected by
the protein content of the diet. Donald et al. (1981) found that adult rats fed a high-
protein diet had signiﬁcantly more body fat than those on a low-protein diet, even
though gross food consumption was similar for both groups. This type of response
may indicate that animals on low-protein diets conserve dietary protein and fuel

energetic costs primarily by the use of non-protein calories.

Materials and Methods

As in experiment 3, 27 voles were divided equally into three dietary groups
(2.5%, 5%, and 15% casein, made up as in Table 3.5). Voles were individually
housed in metal, wire-bottomed metabolic cages, in a room maintained at 22°C with a
16-hour light/8-hour dark cycle. Food consumption was measured daily for three
weeks.

The feeders used in this experiment were designed to minimize food spillage
and eliminate contamination of food by feces and urine. Fwders were constructed by

gluing a length of 2 inch diameter SCH 40 PVC pipe to a plexiglass base. The length

103
of the pipe depended on the height from the ﬂoor to the roof of the cage. The pipe

was out long enough so that the ﬁnal height of the fwder was sufﬁciently high enough
to prevent an animal from entering the feeder through the gap between the roof of the

cage and the top of the feeder. Four feeder holes, 3% inch in diameter, were drilled

in the side of the pipe, so that the bottom of the holes was 1% inch above the base. A

section of ‘A inch SCH 40 PVC pipe, cut to the same length as the 2—inch pipe, was

 

glued to the plexiglass base inside the

 

feeder so that it was mounted parallel to -
the 2-inch pipe. The tube was placed

close enough to the feeder holes to

prevent an animal from crawling into .

the feeder, but far enough away to allow

the animal to stick its head into the V

 

 

 

 

 

 

 

feeder to eat (approximately 14 inch L J
from the feeder holes). The overall Figure 3.6. General design of feeders

. used in expt. 7. See text for details of
desrgn of the feeders forced voles to construction.

feed while standing on their hind legs. The food level in the feeder was kept 56 inch
below the feeder holes, so that voles could not dig the food out with their front paws,
but they could extend their heads into the fwder to eat (see Figure 3.6).

Voles were sacriﬁced by C02 asphyxiation and gut contents were removed.
Carcasses were softened by autoclave for 30 min and then homogenized and dried as
described in experiment 1. Carcass lipid levels were assayed using the Soxtec System

HT6 (Tecator AB, Hogana, Sweden). Three to ﬁve aliquots (061.4 g) of tissue

104

homogenate per animal were used for lipid determination. Protein content was mea-
sured by Kjeldahl analysis of aliquots (0.2-0.3 g) of fat-free tissue homogenates.
Three aliquots of fat-free tissue homogenate per animal were used for protein
determination.

Due to the light, powdery nature of the food, it was impossible to avoid some
spillage. Even with carefully designed feeders, two animals were able to dig out
signiﬁcant amounts of food that added to the waste. To reduce the variance associat-
ed with spillage, it was assumed that a large ‘consumption’ of food was actually the
result of an excessive waste. Therefore, based on total food consumption, all the
food consumption data collected for two outliers (one from the low-protein group and
one from the high-protein group) were eliminated from the feeding data set (see
footnote, Table 3.7). The two outliers were the same two animals that were able to
dig their food out of the feeders.

Initial weights of body tissue and water components were estimated by linear
regression lines derived from the analysis of nine voles matched by weight and age to
the experimental groups at the start of the experiment. Individual body constituents
(protein, lipid, water, and other10 (y variables» were regressed on body weight (x),
generating the following regression equations: y = 0.13x - 0.08 for protein, y =
0.11): - 0.73 for lipid, y = 0.71x + 0.76 for water, and y = 0.05): + 0.05 for other.
Calculations derived from these equations are presented in Table 3.9 (Initial protein,

lipid, water, and other weights) and were used to compute tissue and water changes.

 

10Other is assumed to be primarily skeletal in origin (Evans 1973).

105

.32 :8 9855 59505 65999 5&8 5.3

do .8523“. 33 bESSEK Legs .638 \e «eecgh DMD 5.85:5 9553 58 829 1855.5 «.3 God u 3 newneaaoo
v8.5 :58 co v9.3 59:53 Eat 355650 803 £053 25,—. .256 V m... .86 V m: .86 v m. .53 mm: H.355.
wean 0:8 So? 288 he 885888 8e 36% .<>Oz< .3 38:98 203 58 :< 38> 9w 58 2mm H 28:. 0.8 82‘;

 

 

mz o4 H 35 45 H 05 as H 5.5: _.9o>_.w 2 .935 85 be a

.55 08 H 9:8 «.9. H @585 4.2: H 5.82 _.9o> 5 .935 89 be 8

m2 3 H «.5. 2 H 3.. E H 99. .99.-» a .935 88 be 2

:52 5:5 :25 99. H 92 S: H o3 «.8 H E. .99, a .935 88 be 2
”859955 52 em .92
885:me :5: 5582 33

 

 

H5. 5 598 9

 

 

.5 .55 50:55:88 89 0589-559» Ba 309 .2“ 9%...

106

 

E

 

H
pa
0

 

 

DAILY ENERGY common on vol-"1
8 8

 

 

 

 

 

 

 

 

 

 

IO A 15‘ Canal:
I 5! Caaaia
l 2.5! Caaala
7° - - - - 15! Caaaia
— — - 5! Caaala
‘0 15’ Mn
0 5 10 15 20 25
DAYS
Figure 3.7. Mean daily food consumption, expt. 7.
a 12
I
a 11
z
10
9
y . a 15G Caada
O 5! Caul-
1 l 2.5! Caaala
’ - - - - 15. Cal-II
- - - 5! Caada
g ‘ 2.5: each

 

 

 

 

 

 

Figure 3.8. Weight-speciﬁc mean daily food consumption, expt. 7.

107

 

 

 

Results Table 3.8. Post hoc polynomial contrasts, expt. 7.
F001) CONSUMPTION GROUP CONTRASTS PROBABILITY
Total food consumption Willi—119.11%

LH P = 0.001
was calculated for the ﬁrst 10

LM P < 0.01
days and for the full length of MH NS
the experiment (21 days) (Table W
3.7). Results show that 10- and

LH P < 0.001
21-day total food consumption LM P < 0.001
was greatest for the 15% casein MH r NS

 

group, when compared to the ‘P < 0.001 (repeated measures ANOVA).
2.5% group. Weight-speciﬁc food consumption was the same for all three groups
during both periods.

An examination of mean daily food consumption reveals a steady increase in
food intake by the low-protein group and a decline in consumption over time for the
high-protein group (regression lines, Figure 3.7). By the end of the experiment,
mean daily food consumption was highest in the low—protein group. The increase in
food intake by voles fed the low-protein diet is more evident on a weight-speciﬁc
basis (regression lines, Figure 3.8). Repeated measures ANOVA revealed a signiﬁ-
cant difference in food consumption over time for both whole-animal and weight-
speciﬁc daily food consumption (P < 0.001). Polynomial contrasts indicate different
feeding patterns occurred between the low- and medium- and low- and high-protein

groups for both the whole animal data and per-unit-weight data (Table 3.8).

 

108

Table 3.9. Body composition of voles fed 2.5%, 5%, and 15 % casein diets, expt. 7.

 

 

Initial wet wt
Final wet wt
Change wet wt
Initial protein wt
Initial lipid wt
Initial water wt
Initial other wt
Final protein wt
Final lipid wt
Final water wt
Final other wt
Change in protein
Change in lipid
Change in water
Change in other

is] E 1"!

Protein
Lipid
Water
Other

DI! Weight; (g)

Initial dry wt
Final dry wt
Change in dry wt

0

Protein
Lipid
Other

Fr Ini'

 

 

 

2.5% 5% 15%
20.7 3}; 0.5 20.8 ;t 0.4 20.7 i 0.4
21.2 :1: 0.6 26.8 :1: 0.9 30.6 :1: 1.2

0.6 :1: 0.4 6.0 :1: 0.9 9.9 :t 1.2
2.6 :1; 0.1 2.6 :t 0.1 2.6 :l: 0.1
1.5 i 0.0 1.5 i 0.0 1.5 :l: 0.0
15.5 i 0.4 15.6 i 0.3 15.6 :1: 0.3
1.0 :1: 0.0 1.0 :1: 0.0 1.0 i 0.0
2.8 :t 0.2 4.2 :1; 0.2 4.8 :t 0.2
2.8 :1: 0.4 4.3 :1: 0.4 5.2 d: 0.7
14.4 i 0.4 16.5 :1: 0.4 18.6 i: 0.6
1.2 :1: 0.1 1.9 :1; 0.2 2.1 :1: 0.2
0.2 :1: 0.2 1.5 :t 0.2 2.1 :1: 0.2
1.3 j; 0.3 2.8 :l: 0.4 3.7 :1: 0.7
-1.1 i 0.4 0.8 :t 0.1 3.0 :1: 0.6
0.2 :1: 0.0 0.9 :1: 0.2 1.1 :t 0.2
8.0 i 6.4 58.6 i 7.8 80.7 :t 6.5
82.4 :1; 20.0 184.3 1 22.7 246.3 :I: 47.9
-7.0 :t 2.8 5.6 :1: 3.0 19.9 :I: 3.9
20.1 i 5.6 85.8 :1: 18.3 105.2 :1: 15.1
5.2 :t 0.1 5.2 :l: 0.1 5.2 i 0.1
6.8 :t 0.4 10.4 :1; 0.6 12.0 :1; 0.9
1.7 i 0.3 5.2 i 0.6 6.9 :t 0.9
4.1 :1; 3.3 30.0 :1: 4.0 41.4 j; 3.4
67.2 :t 20.0 53.0 :1: 5.2 49.0 :t 5.8
69.7 i 22.2 55.0 i 8.1 50.0 i: 5.6
17.6 :1: 15.3 30.8 3: 3.6 34.4 :1; 4.3
67.2 i 19.9 53.0 i 5.2 49.0 i 5.8
15.2 :1: 5.0 16.2 :1: 2.3 16.6 :1; 2.2

——=

 

109

BODY COMposITION

Signiﬁcant differences were found in weight gain between all groups (Change
wet wt, Tables 3.9 and 3.10); weanling voles fed 15% casein diets gained the most
(Figure 3.9). Data for the underlying tissue and water changes are compiled in Table

3.9; ANOVA and post hoc comparisons are listed in Table 3.10. All percentage data

Table 3.10. Statistical analysis of selected body composition data listed in table 3.9,

   

 

 

 
    
   
   
  

COMPOSITION TEST Post hoc comparisons
Wts (g)
Change wet wt ANOVA P < 0.001 I H, P = 0.001; LM, P = 0.001; MH, P < 0.05
Final protein wt ANOVA P < 0.001 I H, P < 0.001; LM, P < 0.001
Final lipid wt ANOVA P = 0.010 I H, P < 0.01
Final water wt NOVA P < 0.001 I , P < 0.001; LM, P < 0.05
Final other wt W P = 0.001 I , P = 0.001; LM, P < 0.01
Change in protein ANOVA P = 0.001 I H, P < 0.001; LM, P < 0.001
Change in lipid ANOVA P < 0.01 I , P < 0.01
Change in water ANOVA P < 0.001 I , P < 0.001; LM, P < 0.05; MH, P = 0.01
Change in other W P = 0.001 I H, P = 0.001; LM, P < 0.01
5 Change From Initial

Protein P < 0.001 I , P < 0.001; LM, P = 0.001
Lipid P < 0.05 I M, P < 0.01
Water P < 0.001 I H, P = 0.001; LM, P < 0.01; MH, P = 0.012
Other P = 0.001 I , P < 0.001; LM, P < 0.01

[2g Weights (g)
Final dry wt ANOVA P < 0.001 I H, P < 0.001; LM, P < 0.01
Change in dry wt [KW P < 0.001 I H, P < 0.001; LM. P = 0.001

ﬂ Change From Initial

Protein W P < 0.001 I H, P < 0.001
Lipid NS
Other W NS

Co sition of

T' Q .

Protein W NS
Lipid
Other

 

KW = Kruskal-Wallis analysis of ranks.

110
were analyzed by Kruskal-Wallis analysis of ranks (KW), as were data that violated

homogeneity of variance (Bartlett test). Post hoc comparisons for data analyzed by
KW were accomplished using the Mann-Whitney test; differences were considered
signiﬁcant for P < 0.05/3 (Bonferroni correction). Tukey’s HSD test was used for
ANOVA post hoc comparisons.

The data in Table 3.9 indicate that body weight and body protein content

Table 3.11. % Composition of ﬁnal wet and dry weights, expt. 7.

 

 

 

% COMOPSITION

 

 

 

GROUP
:1
Low Medium High Signiﬁcant
Final Wet Wt 2.5 95 5% 1596 Differences
Protein 13.4 1; 0.7 15.6 i 0.3 15.7 :t 0.6 LM‘
Lipid 11.8 :1: 1.5 15.9 i 1.1 16.4 :t 1.8 NS
Water 67.9 i 1.4 61.5 :t 1.2 61.1 :1: 1.6 LM" LH‘
Other 6.9 :1: 1.2 7.0 i 0.5 6.9 :1: 0.5 NS
PM Dg Wt
Protein 42.3 i 2.9 41.0 n 1.7 41.0 :1; 2.6 NS
Lipid 39.5 i 3.8 40.9 i 2.2 41.3 i 3.5 Ns
Other 18.1 i 0.9 18.1 i 1.0 17.7 :1; 1.2 NS

 

Percentage data analyzed by Kruskal-Wallis analysis of ranks. Post hoc comparisons were
done using the Mann-Whitney test; differences were considered signiﬁcant for P < 0.05/3
(Bonferroni correction). 'P < 0.05, 'P < 0.01.

increase with increasing dietary protein. Lipid, water, and other show similar trends.
These results are comparable to those found in studies of rats (Donald et al. 1981;
Edozien and Switzer 1978).

The proportion of protein, lipid, and other dry tissue gain did not differ (%
Composition of Dry Tissue Gain, Table 3.9). This resulted in no change in the
relative proportions of dry tissue constituents for the whole animal (Table 3.11).

Percent composition of ﬁnal wet weight showed that protein was proportionately

111

 

 

 

 

 

“TTTIIITIIIII
30- n
3
(D
E
20.. ‘7 .-
AlSiCaaain
IS‘Caaain
10 IIILIIIIIIII IZSICIIOIII
212325212931333531554142
AGEldaya)

 

 

 

Figure 3.9. Growth of weanling voles, expt. 7. Error bars are :1; SEM.

reduced in the 2.5 % casein group. Even though this group lost water (% Change
From Initial, Table 3.9), the ﬁnal carcass water content was proportionately greater
than the water content in either the medium- or high-protein group (Table 3.11).
Consequently, the percent protein in the ﬁnal wet weight of the low-protein group
was diminished by a relative increase in water content. Note that the relative
proportions of protein, lipid, and other in the ﬁnal dry weight were equal for all three
groups (Table 3.11). This is interesting because total food consumption and growth

rates were markedly different.

112

.506 V m: .56

v m. .386 88: a 588 .99 a: £93 55 <65. .3 5:55 555% .85 955 z 565

.3 825“. Saw Z boom. 5885.898 85 we magmas 85 8 .888qu :2: 358 Z 33 85m... .Eouootoo
58.8.5508 Qmod V m 58 5:85:me 3.8288 883 8088.553 :85 8523-582 85 mean 8:8

883 3on .8 885958 .63 3.6m 5555 Co 595528 5583-355 .3 “88:98 San—N .mmd \ 9 5888
u z .53 u m .8. - an; n e w 8-: 555563 85> e .8 8 5563 5.8 .5 8 59:8 5985 5.8
CO .85mb8 58:: .3 “88:58 33 539:” .33 35:: .5335 238.9 .3 “858.88.. 283 8253 Z =<

 

 

 

.52 .25 2 H 02! m5 H Won 8 H 3. 5. 5.88858 2 580
:5 :25 58 H 3% 4.8 H 2.8 98 H can .96., we 3.56» 2 58m
:52 :5 :5 QB H mama 4.: H 45; 35 H 39. .92, me .6555 2 5555
m2 ed H 29. 45 H 3.5. I: H 93. .99, we _.z .68 E55
8288th R. m _ a» m R m .N
585255 e55 5582 33

 

one 5 598 9

H 598 .8558 comp—5: 3.80 .S .m 838.

113

NITROGEN BALANCE

Gross nitrogen efﬁciency was highest for the 5 % casein group (Table 3.12).
Gross nitrogen efﬁciencies were statistically the same for low- and high-protein
groups, even though dietary nitrogen intake was six times greater in the high-protein
group. However, net efﬁciencies are not known, since fecal and urinary nitrogen

contents were not measured.

DISCUSSION

Introduction

The focus of the proceeding experiments has been to examine the effects of
different dietary protein levels on energy balance, brown adipose tissue activity, food
consumption, body tissue change, and nitrogen balance in herbivorous voles. Low
energy efﬁciencies were recorded in experiment 3 for voles consuming low-protein
diets (Table 3.6). Voles fed low—protein diets had efﬁciencies that were over 75
percent lower than voles fed high protein diets. The low efﬁciencies observed in
experiment 3 for low-protein-fed voles are characteristic of a physiological phenome-
non known as diet-induced thermogenesis (DIT), which has been observed in labora-
tory rats consuming low-protein diets (Chapter I).

The effector organ for DIT in laboratory rats is thought to be brown adipose
tissue (BAT). Speciﬁc-binding of GDP to BAT mitochondria is a parameter used to
estimate BAT activity. Contrary to what has been observed for laboratory rats,
speciﬁc binding of GDP to BAT mitochondria does not increase in voles fed low-

protein diets when compared to high-protein-fed voles (experiments 4 and 6).

114
Food consumption data (experiment 7) indicate that voles increased food intake

when fed low-protein diets. Gross food consumption was greater for high-protein-fed
voles (Table 3.7), but over a 21—day feeding period, the rate of food consumption
increased for voles on low-protein diets and declined in the high-protein group
(Figures 3.7 and 3.8).

Body composition data (Table 3.9) from experiment 7 indicate that body
weight, body protein content, lipid, water, and ‘other’ tissue were higher for voles
consuming high-protein diets when compared to animals fed low-protein diets.
However, the relative proportions of protein, lipid, and other in the ﬁnal tissue dry
weight showed no difference based on the level of protein in the diet (Table 3.11).

Gross nitrogen efﬁciency did not differ between low- and high-protein groups
but it was signiﬁcantly higher for the medium—protein group when compared to the
other two groups (Table 3.12). Dietary nitrogen intake was six times greater in the
high-protein group compared to the low protein group. Net nitrogen efﬁciency for

the three treatment groups was not determined.

Thermogenesis
Experiments 1 through 3 were designed to test the inﬂuence of dietary protein
on the utilization of energy in voles. Similar experiments have been performed on
rats (e. g., Stirling and Stock 1968), and revealed an increase in metabolic heat
production and reduction in efﬁciency of energy retention. Likewise, data presented
here show that the protein content of the diet is a major determinant of energy

efﬁciency in voles.

115

A 7% casein diet was not low enough to elicit a signiﬁcant thermogenic
response in voles in experiment 1, probably because the animals used in this experi—
ment were too old to be stressed by protein deﬁciency at the level offered. Once
voles reach 30 g, they are considered adults (Baker 1983), and growth thereafter is
highly variable and relatively slow when compared to the ﬁrst 30 days post-weaning
(personal observation). The voles used in experiment 1 (53-55 days old, with a mean
weight of 29 g), having almost reached adult weight, were at a stage where growth is
slow and unpredictable. Note that in the 17% casein group, the mean in growth was
only 3 g over a period of 21 days.

Thermogenic responses to low-protein rations have been recorded in adult rats
(Donald et al. 1981; Swick and Swick 1986). In the study by Donald et al., the 10-
week-old rats used in the experiment were considered ‘fully grown, ’ but mean weight
gain over a period of 9 weeks ranged from 100 g for rats fed a 5 percent protein diet
to 200 g for rats fed a 25 percent protein diet. Donald et al. measured ﬁnal fat-free
body mass but the initial fat-free body mass was not estimated. Final fat-free body
mass, though, was greater than the initial weights of the animals. Presumably, some
of the growth was due to an increase in body protein. It is likely, therefore, that
protein intake was important for the continued growth that occurred in these rats.

The researchers’ main focus, though, was the inﬂuence of dietary protein on adiposity
in older, ‘nongrowing’ rats. Consequently, little attention was given to changes in
body protein. The question remains as to whether ingestion of a low-protein diet
would induce a thermogenic response in animals that no longer require dietary protein

to sustain growth of protein tissue.

116

Since protein turnover can be substantial in fully grown animals (Swick and
Benevenga 1977), presumably there is a minimum level of dietary protein required by
these animals for maintenance. If DIT stimulated by low-protein intake is an adaptive
process for the distillation of dietary protein, then fully grown animals could exhibit
DIT when protein intake falls below maintenance.

In DIT experiments using laboratory rats, a 5 % casein or lactalbumin diet is
usually low enough in protein content to induce a thermogenic response (e. g. , Donald
et al. 1981, Swick and Swick 1986). DIT in rats has also been detected using 8%
casein diets (Tulp et al. 1979; Rothwell et al. 1983a). However, in experiment 2, a
5 % casein diet was not low enough in protein content to produce a signiﬁcant
increase in thermogenesis in voles. There are at least two possible explanations,
which are not mutually exclusive. There may have been an early DIT response and
low energy efﬁciency that was obscured by a subsequent increase in efﬁciency of
energy retention occurring after seven days into the experiment. It is also possible
that voles, known to have a high BMR when compared allometrically to other
animals, require less dietary protein to stimulate DIT. Since voles are herbivores,
they may have evolved elevated BMRs to aid in dissipation of excess energy intake
resulting from consumption of large amounts of grasses that are typically much higher
in energy content than in nutrients such as protein.

The growth pattern in experiment 2 (Figure 3.4) suggested that some type of
compensatory response occurred that enabled the 5 % casein group to catch up to the
15 % group after a comparatively slow growth rate at the start. Energy efﬁciencies

for the whole growth period were nearly identical, so there was no prima facie

117

 

evidence suggesting that

 

DIT was involved.

Since the experiment

1 ran for 21 days, and the
growth rate for the 5%
q casein group increased
after 7 days into the
experiment, it is con-
0 ‘ 2.5! 5.0. 15.0!

ceivable that a DIT

pa
0
I

I met
(Dom! ENERGY GAIN/In) x 100

 

 

 

 

 

 

response occurred ini-

Figure 3.10. Efﬁciency of energy retention, expt. 3. . .
Error bars are :1: SEM. tially dunng the ﬁrst 7

days, and then was masked by subsequent growth and an increase in efﬁciency.
Figure 3.10 shows the energy efﬁciencies for experiment 3. The 2.5 % casein
group differed statistically from the other two groups, but the 5% and 15% casein
groups showed no signiﬁcant difference (P = 0.047 for post hoc comparison using
the Bonferroni correction of P = 0.05/ 3). Due to the large variance inherent in
material balance studies, if a DIT response occurred in the 5% group, it may have
been too small to be detected statistically. However, the data in Figure 3.10 suggest
a trend, with energy efﬁciencies decreasing as dietary protein is reduced. Also of
note, the mean efﬁciency for the 5 % group in experiment 3 for a 10 day period was
lower than that recorded in experiment 2 for 21 days (12.0 versus 13.9 percent,

respectively). It is doubtful that this is demonstrably different, but the mean change

1 18
is in the direction predicted for a DIT response occurring during the ﬁrst 10 days of

feeding.

As mentioned earlier, another explanation relates to a vole’s BMR. Voles may
require a lower dietary protein level than laboratory rats to stimulate DIT, if elevated
BMRS known to occur in voles aid in dissipation of excess energy intake. Voles have
been shown to have a 20 to 40 percent greater BMR than expected from allometry
(Wunder 1985). McNab (1980) suggested that an abundant high-energy food supply
for herbivores permitted the evolution of higher BMRS. A high BMR is positively
correlated with a variety of ﬁtness traits, including rm, the maximal intrinsic constant
of population growth, suggesting that a high BMR is somehow adaptive (McNab
1980). According to McNab (1980), “. . . it behooves all mammals to have as high a
rate of metabolism as can be sustained by the quantity and quality of their food
resources in space and time, because this adjustment will permit them to maximize
their reproductive efforts. ”

It is also possible that herbivorous feeding not only permits a high metabolic
rate, it necessitates it. If we assume that dietary nitrogen—not energy—is the most
common limiting resource for a herbivore (Chapter 11), then higher BMRs could
facilitate processing of foodstuffs and extraction of essential nutrients to meet nutri-
tional requirements. A high metabolic rate that aids in the assimilation of nutrients
could be an essential trait required to evolve herbivory. If we compare herbivorous
voles to omnivorous rats, it may be that voles are more metabolically adapted to
process foodstuffs that are nutritionally unbalanced by having a higher than normal

basal metabolic rate. Herbivorous voles may not have to increase metabolism at the

119

7% (experiment 1) or 5 % (experiment 2 and 3) casein level because their metabolism
is already 20—40 percent higher than comparable omnivorous homeotherms that may
be less limited in nature by the nitrogen content of their food supply.

Thus, the low level of nitrogen in the diet required to increase thermogenesis
in voles may be the result of an already intrinsically high BMR. The data from
experiment 3 (Figure 3.10) clearly indicate a DIT response by the 2.5% casein group,
and suggest that voles require lower concentrations of dietary protein to elicit a
signiﬁcant increase in thermogenesis than omnivorous rodents such as rats. F urther-
more, a comparison of energy expenditures in rats versus voles on low-protein diets
shows that weight—speciﬁc ADMRS are several times higher in voles. Rothwell and
Stock (1987c) reported an energy expenditure of 0.9 k] d"g‘1 for rats maintained at 24
j; 1°C and fed low-protein dies for 15 days. This contrasts sharply with the 2.7 to
3.8 k] d“g’l ADMRS recorded in experiments 1 through 3 for voles on either high- or
low—protein diets, indicating a comparatively high weight-speciﬁc metabolic rate for
voles.

Although the rats used by Rothwell and Stock (1987c) were considerably
larger, it is unlikely that the difference in weight-speciﬁc metabolic rates was due to
size difference alone. The relationship between weight-speciﬁc BMRs (M) and body
weight (W) for birds and mammals can be expressed by the following allometric
equation:

M = 70W“
where M is in k] kg"d'l and W is in kg (Kleiber 1975b). The mean weight for voles

on a low-protein diet in experiment 3 was 18.3 g and the mean weight for laboratory

120
rats fed a low-protein diet in the study by Rothwell and Stock (1987c) was 151.5 g.

When these mean weights are used to calculate M for the low-protein-fed voles used
in experiment 3 and the laboratory rats on low-protein diets used by Rothwell and
Stock (1987c), the results indicate that the smaller voles should have a weight-speciﬁc
BMR that is 70 percent greater than that of the larger laboratory rats. Experimental
data (experiment 3 and Rothwell and Stock 1987c) indicate that the 18.3 g voles had
an ADMR that was 333 percent higher than the 151.5 g laboratory rats (3.8 k] d“g‘1
versus 0.9 k] d"g“, respectively), thus indicating that the empirically derived ADMRS
did not scale allometrically according to the above equation. These calculations also
suggest that body weight alone is not responsible for the differences in weight-speciﬁc
ADMRS between voles and rats.

According to Kleiber (1975b), the weight-speciﬁc metabolic rate reﬂects the
turnover rate of chemical energy in the body. Because of their high metabolic
turnover rates, voles consuming low-protein diets are capable of rapidly processing
large amounts of ingested energy. If nitrogen is a limiting resource in the natural
diets of voles, a high metabolic turnover rate evoked by foodstuffs deﬁcient in
nitrogen could facilitate nitrogen uptake by permitting high food intake and a concom-
itant distillation of protein, while oxidizing unneeded nonprotein energy. The results

from the energy balance experiments performed here support this hypothesis.

Brown Adipose Tissue
One of the most exciting and novel aspects about the DIT phenomenon is the

idea that it is connected with BAT metabolism. Not only is there a substantial rise in

121

heat production associated with cafeteria feeding or low-protein intake, the mechanism
of heat production ostensibly occurs by an unusual process circumventing ATP
production in specialized brown fat tissue. In 1979, when Rothwell and Stock pub-
lished their seminal paper on DIT in Nature, linking the phenomenon to BAT
metabolism, a new approach to the study of energy balance was engendered. Previ-
ously, BAT had been assigned the role of the principal effector of NST. Since BAT
thermogenesis had already been the subject of considerable research as a thermoregu-
latory process, many of the techniques used to elucidate BAT activity in response to
cold could now be used in energy balance studies (Chapter I).

One of the most common techniques used to estimate BAT activity is the GDP
binding assay. Binding of purine nucleotides to thermogenin, the uncoupling protein
in the mitochondrial membranes of brown fat, is thought to be part of the regulatory
process that determines proton conductance across BAT inner mitochondrial mem-
branes and, therefore, thermogenic activity. This assay has been applied to DIT
studies using both cafeteria and low-protein feeding regimens, and the results tend to
show that GDP binding and DIT are positively correlated (Chapter I).

An increase in GDP binding, however, was not observed in voles feeding on
low-protein diets. GDP binding at 22°C was greater in experiment 4 for the high-
protein group compared to the low protein group; at thermoneutrality (28°C), no
signiﬁcant difference was recorded between dietary groups (experiment 6) (Figure
3.11). As previously mentioned, though, a difference in speciﬁc binding was
observed for voles experiencing different thermal conditions created by disparate

housing conditions (experiment 5).

122

 

 

15. u v 1 u

11:. + .

I
l

 

 

 

 

 

 

 

 

 

no - i .
g 1. - + ..
g 5“ - 4
ﬂ 1 l l l
m It! LI IV
m4. 4.. s “I
«our

 

 

 

Figure 3.11. GDP binding, expts. 4 and 6. L = low protein, 2.5% casein; H =
high protein, 15% casein; C = cold (22°C); W = warm (28°C). Error bars are i
SEM.

Since voles in experiment 4 had very high GDP binding levels, and those in
experiment 5 were signiﬁcantly different—ostensibly due to different levels of NST
resulting from microclimate differences—it was concluded that testing voles at
thermoneutrality would eliminate variation in GDP binding caused by NST. The
experiment at thermoneutrality, however, did not uncover differences in GDP binding
relative to dietary protein differences. The most likely conclusion is that BAT proton
conductance was probably not responsible for the low energy efﬁciency observed in
experiment 3 for voles on the low-protein diet.

Rothwell and Stock (1986b) have argued that DIT is inhibited at high environ-

mental temperatures, probably to prevent hyperthermia at thermoneutrality. They

123
found that rats fed cafeteria diets and kept at 29°C showed no evidence of DIT. Barr

and McCracken (1982) had similar ﬁndings. Thus, it may be that voles at
thermoneutrality failed to show a signiﬁcant difference in GDP binding due to diet
because the thermogenic response was suppressed. Perhaps a better experiment
would be to test voles at temperatures only slightly below thermoneutrality, thus
minimizing NST, yet reducing the chances of hyperthermia.

More difﬁcult to explain are the GDP binding results observed in experiment
4, where both low- and high-protein groups had high GDP binding levels, and GDP
binding for the high-protein group was signiﬁcantly higher. Swick and Swick (1986)
noted similar results for rats fed 5% and 15% lactalbumin diets for 3% months.
Speciﬁc GDP binding to brown fat mitochondrial protein (pmol/mg protein) did not
differ statistically, but total GDP binding (nmol/IBAT pad“) did. According to Swick
and Swick (1986), total GDP binding is a more reliable parameter for measuring BAT
thermogenesis, because total interscapular BAT is accounted for. However, total
GDP binding was not measured in voles; therefore, the difference in total binding for
voles on low- and high-protein diets is not known. Total GDP binding for BAT in
voles consuming variable protein diets should be measured before any conclusions
can be made about the contribution of BAT metabolism to DIT in voles.

Complicating matters are recent studies suggesting that heat production by
BAT is not increased by cafeteria feeding. Foster and Ma (1989) and Ma and Foster

(1989) made direct measurements of BAT oxygen consumption using nonocclusive

 

"IBAT pad refers to the interscapular brown adipose tissue pad. Total GDP
binding is determined for the total amount of BAT dissected from the animal.

124

cannulae to sample the oxygen content of venous efﬂuent from BAT. Radioactive
microspheres were used to measure tissue blood ﬂow. Cannulas were also placed in
the right common iliac artery to sample arterial oxygen content. BAT oxygen
consumption was calculated using the Fick principle. Their results show that cafete-
ria-fed rats exhibiting a measurable DIT response did not exhibit a signiﬁcant
elevation in BAT metabolism. Heat production in BAT contributed only 2 to 3
percent of the total thermic effect of cafeteria feeding. Foster and Ma also found that
partial hepatectomy (’6) had a major effect on whole-body metabolic rate of cafeteria-
fed rats, with a reduction in V0, equal to or greater than the increase in V0, caused by
DIT under prehepatectomy conditions in cafeteria-fed rats (Foster and Ma 1989).

Since it is possible that the liver may be an effector organ for DIT, the degree
to which BAT is involved in heat production via DIT may not be fully known. In a
review article on BAT thermogenesis, Himms-Hagen (1990) stated, “In no case has it
been possible to quantitate the contribution made by BAT to the altered state of
energy balance.” According to Himms-Hagen, some rats with atrophied BAT do not
become obese, while others become obese, even though their BAT content increased
and is thermogenically active.

Perhaps the main reason the GDP binding results of experiment 4 seem
contrary is that our expectations are based primarily on studies of laboratory rats. It
is possible that, in some species, speciﬁc GDP binding may not be a suitable assay
for BAT thermogenic activity stimulated by low-protein DIT. Clearly, more
interspeciﬁc comparisons are needed. Both speciﬁc and total GDP binding should be

measured in more species exhibiting DIT, so that the two assays can be compared for

 

125

effectiveness in predicting BAT thermogenic activity among different animal groups.
Until more information is known about DIT-related BAT metabolism in species other
than laboratory rats and mice, it is premature to generalize about the contribution of

BAT to low-protein DIT based on the GDP binding data presented here.

Food Consumption

One of the main hypotheses in this research is that voles should show a
hyperphagic response while fwding on low-protein rations. A hyperphagic response
would allow voles to increase protein intake by increasing bulk intake. A concomi-
tant reduction in energy efﬁciency would permit voles to eliminate unneeded calories
while extracting maximal dietary protein.

Clearly, a hyperphagic response is illustrated in Figures 3.7 and 3.8. They
show that the rate of intake increases steadily for voles on a low—protein diet and
declines over time for the high—protein group. Also, by the end of 21 days of
feeding, voles on the low-protein diet are consuming as much or more food per day
than their high-protein counterparts, even though there is a mean difference in weights
of 9.4 g, which constitutes a 44 percent greater body weight for the high-protein
group.

However, looking at gross food consumption (Figures 3. 12-3. 15 and Table
3.7), the 15% casein group ingested the most energy during the ﬁrst 10 days and for
the 21-day total; weight-speciﬁc food consumption did not differ during either period.
The only time voles on a low-protein diet had signiﬁcantly greater energy to dispose

of by DIT than high-protein-fed voles was during the last few days of the 21-day

 

 

 

 

V\\\\\\\\\\\\\\m

   

 

 

 

 

 

 

F P b h

9.13%ng

 

 

 

sumption for the ﬁrst 10 days of expt.

7.

Figure 3.13. Mean total food con-

 

126

 

 
   

, w\\\\\\\\\\\\\\\\\\\\\\\s_

 

 

 

 

 

 

 
 

$8“

nnn

 

 

 

 

 

0.13%:ng

 

 

 

Figure 3.12. Mean total food con-

sumption for 21 days of expt. 7.

 

 

 

 

 

 

 

 

total food consumption for the ﬁrst 10

Figure 3.15. Mean weight-speciﬁc
days of expt. 7.

 

total food consumption for 21 days of

Figure 3.14. Mean weight-speciﬁc
expt. 7.

127
feeding period, when their daily consumption was highest (Figures 3.7 and 3.8).

However, the low energy efﬁciency recorded in experiment 3 for low-protein-fed
voles was for a lO—day period that corresponds to the ﬁrst 10 days of experiment 7,
thus making it difﬁcult to attribute the source of DIT to oxidation of gross energy
intake. Since DIT occurred in low-protein-fed voles during a time when their food
intake was lower than voles fed high-protein diets, it is more likely that the primary
cause of thermogenesis induced by low-protein diets in voles is the ratio of protein to
energy in the diet and not gross energy intake alone. Rothwell and Stock (1987c)
came to a similar conclusion after a study of rats fed low-protein diets showed
increased thermogenesis even though ME intake‘2 did not differ.

Other data suggesting that low-protein DIT occurs in the acute time frame can
be seen in growth patterns (Figures 3.5 and 3.9). At the outset, the growth rate is
considerably greater for high-protein groups. While this may be due to low food
intake by the low-protein group, it could also be, in part, the result of low efﬁciency
of energy retention by the low-protein group when compared to the high-protein
group. Energy balance studies shorter in duration or detailed measurements of daily
metabolic rates (oxygen consumption) would be nwded to resolve this question.

Glick et al. (1981) suggested that a DIT response recorded in an energy
balance study may be caused by daily heat production via meal-induced thermogenesis

(i.e. , speciﬁc dynamic effect or heat increment of feeding). A typical energy balance

 

l2Rothwell and Stock (1987c) corrected ME intake for body size differences
(kl/(kg0'75d)). Actual ME intake (intake before allometric adjustment for body size)
was different between treatments, with low-protein groups showing the smallest ME
intake.

128

study can last for one or more weeks, and the ﬁnal energy balance determination
represents energy efﬁciency integrated over the duration of the experiment. Accord-
ing to Glick (1981), the observed DIT response measured over time represents the
cumulative effects of heat production resulting from by speciﬁc dynamic effect. Glick
(1987) cites as evidence that most of the increases in BAT parameters resulting from
DIT (e. g. , hypertrophy, thermogenesis, GDP binding, and norepinephrine turnover)
have also been found to occur after the intake of a single meal.

Thus, Glick and colleagues see DIT as speciﬁc dynamic effect writ large.

 

From this perspective, the acute thermic response exhibited by voles on a low-protein
diet in experiment 3 is more understandable. Because of the magnitude of the
difference (energy efﬁciency was more than four times greater for the 15 % casein
group), it is more likely that the low efﬁciency recorded over a period of 10 days is a
cumulative response, rather than occurring in the just last few days of the experiment.
The ‘excess’ energy wasted by voles may be derived from a nutritionally unbalanced
series of meals, where daily energy consumption above maintenance is disposed of by
oxidation. In this view, DIT is an ongoing process that functions in the acute time
frame, and it is elicited by nutritionally imbalanced fwding bouts or meals. If this is
the case, it would mean that small herbivores such as voles can rapidly adjust their
metabolism to local conditions of food quality. This could confer an adaptive
advantage to herbivores living in environments where high quality food is patcth
distributed or food nutrients are temporally variable.

The results of the energy balance study of experiment 3 combined with food

consumption data indicate that DIT observed in voles occurs quickly in response to a

129

diet unbalanced in protein. However, this does not negate the possibility that voles
can exhibit DIT as a chronic response to overlngestion of energy. Clearly, daily food
consumption continued to rise after 10 days, but note that weight gain was only
minimal (Figure 3.9). This suggests that low energy efﬁciencies continued after 10
days and may have been even lower that those recorded initially. Therefore, it is
possible that voles can make both short- and long-term adjustments to their metabolic

rates to meet nutrient constraints imposed by changing levels of dietary nitrogen.

Body Composition

The body composition data (Tables 3.9 and 3.11) indicate the importance of
dietary protein in protein tissue gain. The 15 % casein group gained 10% times more
body protein. The high-protein group also gained almost 3 times more lipid and 5%
times more ‘other,’ which is probably a result of their high total food intake (Figure
3.12). Low-protein-fed voles gained only a small amount of weight (0.6 g), but their
dry tissue gain was 1.7 g. This was due primarily to water loss of 1.1 g. The low-
protein group gained protein, lipid, and other in the same proportions as the other two
treatment groups, and the ﬁnal relative proportions of dry tissue were almost identical
(Table 3.11).

It could be that growing voles homeostatically maintain the relative proportions
of solid tissue in their body when faced with a food supply that is low in nutrients.
Overall growth could be aided by distillation of dietary protein via DIT, which would
permit an increase in the growth of protein tissue, and, consequently, other tissue. If

an innate proportional constraint to tissue growth exists, it might also explain why

 

130

balanced diets result in the greatest growth increases. Excess energy that cannot be
used to maintain proportional growth in body tissues could be eliminated via DIT.
However, this is purely hypothetical and more research would be needed to test these

suppositions.

Nitrogen Balance
Mean nitrogen efﬁciency for the low-protein group (7.6 percent) in experi- h.
ment 7 exceeded the overall energy efﬁciency recorded for the low-protein group in

experiment 3 (4.0 percent), although it is doubtful that this difference is signiﬁcant.

 

It does suggest, though, that voles are able to differentially increase nitrogen efﬁcien- .1;
cy over total energy efﬁciency. Mean nitrogen efﬁciency for the 5% casein group in
experiment 7 was 30.3 percent, while their energy efﬁciency recorded in experiment
3 was 12.0 percent.

Two voles in the low-protein group lost body nitrogen; one vole lost almost as
much as the maximum gain recorded for the 2.5% group. Thus, considerable
variation in nitrogen efﬁciency occurred in the low-protein group, suggesting that the
nitrogen content of a 2.5% casein diet was near the lower limit for maintenance of
positive nitrogen balance. When dietary nitrogen was doubled at the low end of the
spectrum (2.5% to 5%), gross nitrogen efﬁciency quadrupled. Thus, small increases
in nitrogen in the diet are signiﬁcant when the total protein content of foodstuffs is
minimal.

Given that food consumption by the low-protein group was increasing daily, it

may be that eventually, an increase in intake could have supplied enough nitrogen to

131

cause a signiﬁcant increase in growth rate. Some evidence supporting this notion can
be observed in the growth curve for experiment 7 (Figure 3.9). After initially losing
weight, growth in the low-protein group began to rise. This may have been due to
the steady increase in food consumption, which means greater nitrogen intake. If
higher nitrogen efﬁciencies occur with only moderate increases in nitrogen intake,
then a low energy efﬁciency coupled with an increase in food consumption would
translate into an effective physiological and behavioral system for optimizing nutrient L

intake from a high-energy, nutrient-poor diet.

 

CHAPTER IV

CONCLUSIONS AND RECOMMENDATIONS

It has been suggested that diet-induced thermogenesis (DIT) is an adaptive
response that regulates energy balance and weight in mammals (Rothwell and Stock
1979). Proponents of this idea have labeled the phenomenon as ‘adaptive’ DIT
(Rothwell and Stock 1986a). However, because our understanding of this physiologi-
cal response is still limited, it is probably premature to use a nonneutral adaptive label
for DIT. By calling it adaptive DIT, we assume we know why it is adaptive.
However, several adaptive functions have been proposed for DIT, with no clear
demarcation between them (Himms-Hagen 1986). DIT is evoked by excess energy
consumption and by moderate ingestion of a low-protein diet (Rothwell and Stock
1979; Rothwell et al. 1987c). It is not clear what differences, if any, exist between
these responses, nor have the possibilities for different adaptive functions been fully
explored. DIT occurring in response to cafeteria or low-protein feeding is usually
treated as the same phenomenon, even though it can be elicited by different feeding
regimens.

To complicate matters, there are some who dispute the existence of DIT, and
believe that the thermic response observed in laboratory rodents to cafeteria feeding is
an experimental artifact (Barr and McCracken 1982). Furthermore, the term ‘adap-

tive’ has several meanings and can be interpreted differently by researchers with

132

133
different backgrounds. Rothwell and Stock (1981a) have implied that DIT is an

evolutionary adaptation, but the most common use of the term adaptation in the nutri-
tion literature is to denote a phenotypic physiological adaptation (Waterlow 1985) and
DIT could be interpreted as such. Therefore, to avoid confusion, minimize contro-
versy, and foster a sense of objectivity, a speciﬁc adaptive label for DIT should be
avoided. I
Since DIT is stimulated by low-protein intake, it has been postulated that it i

could be adaptive for increasing nutrient intake by distillation of essential nitrogen

 

from low-protein foodstuffs by oxidation of excess, nonprotein energy intake. Several %
researchers have made this suggestion, but an ecologically-based scenario for the

genetic selection of DI’T as a Darwinian adaptation to nutrient stress has not been

fully articulated. Most of the current discussion in the nutrition literature is focused

on the impact of DIT on energy balance and obesity in laboratory rodents (e.g.,

Himms-Hagen 1989). The role DIT plays in nutrient optimization has received little

serious attention.

The extent to which DIT is involved in augmenting nutrient intake may best be
examined by focusing on the physiology and nutritional ecology of herbivores.
Herbivores consume foodstuffs that are typically low in nitrogen. Food quality may
be so poor (low nitrogen content) that herbivores are unable to consume enough
nitrogen to satisfy maintenance needs or nourish young (Sinclair 1975).

Temporal changes in food quality and availability due to climatic changes such
as rainfall, can drastically reduce the amount of high-quality, high-nitrogen food

available to herbivores. Chemical defenses of plants can force herbivores either to

134

feed on less nutritious parts of plants or consume plants of lesser quality. Feeding
can also be deterred by structural defenses such as spines or trichomes. Digestible
protein can be diluted by high dietary ﬁber, and allelochemicals such as tannin can
reduce nitrogen availability by binding with dietary protein. Thus, the amount of
high-nitrogen foodstuffs available to herbivores in the wild can be limited and have a
negative impact on herbivore survival, suggesting that selection pressure could lead to
the evolution of speciﬁc traits that allow herbivores to cope with low ambient levels
of dietary nitrogen (McNeil and Southwood 1978; White 1978; Mattson 1980).

Laboratory rats have been the primary animals used in research on low-protein

 

1' '

DIT. Since these animals are omnivores and are not as likely as herbivores to be
limited by food resources low in nitrogen in the wild, it is not clear how ecologically
important an increase in thermogenesis in response to low-protein intake would be for
them. A study of wild frugivorous bats, though, suggests that they can overingest
energy to obtain sufﬁcient protein, and dispose of the excess energy by oxidation
(Thomas 1984). This is signiﬁcant because the bats studied by Thomas ( 1984) are
obligate frugivores that must rely on nutrient-poor fruits (typically 5 3.3 percent
protein by dry weight) to fulﬁll their nutritional demands.

Thomas’ study showed clearly that food intake and energy efﬁciencies of
frugivorous bats were dependent on the protein content of the diet. The model devel-
oped by Thomas (Figure 2.1) predicts that frugivores are signiﬁcantly more likely to
be limited by dietary nitrogen than omnivores. By comparing an obligate with a
facultative frugivore, Thomas was able to show the ecological implications of a

thermogenic response caused by low-protein intake. His research also emphasized the

135

importance of using frugivorous or herbivorous wild animals in this type of study.
Linking the nutritional ecology of an animal directly to an observed physiological
response facilitates the interpretation of the results and lends credibility to the
supposed function of that response. In this case, Thomas’ data strongly suggest that,
for those species constrained by the nutrient quality of the food they consume, DIT
could play a signiﬁcant role in the ability to achieve an adequate nutrient intake.
Nutritional studies of invertebrates provide further evidence that nitrogen can

be a limiting nutrient in the wild and that animals have developed speciﬁc adaptations

 

to deal with this problem. Mattson’s analysis of invertebrate herbivores showed that i
the efﬁciency with which a consumer can convert ingested food into body tissue is
positively correlated with the percent nitrogen in the diet (Mattson 1980) (Figure 2.2).
He also showed that the low energy efﬁciencies observed in insects feeding on plants
low in nitrogen are not due to a decrease in throughput time, but are caused by post-
digestive processes, thus suggesting metabolic changes similar to DIT observed in
mammals. That similar responses occur in both homeotherms and poikilotherms
indicate that DIT stimulated by low-nitrogen foodstuffs could be a ubiquitous response
to a widespread problem of nitrogen shortages in the wild and may be evolutionarily
convergent.

Low energy efﬁciencies recorded for herbivorous voles in the experiments de-
scribed in Chapter III provide further evidence that a DIT response to low-quality
food is an adaptation that aids in nutrient uptake. Complete material balance studies

of voles on variable protein diets show that animals on low-protein diets have lower

136

energy efﬁciencies (Figure 4.1‘) and signiﬁcantly higher metabolic rates per unit of
metabolizable energy (Table 3.6) than highoprotein-fed animals. Results suggest that
energy efﬁciencies in older animals are less affected by low dietary nitrogen than
those in younger animals. Differential responses due to age may indicate that the DIT
response is greater in younger, growing animals because they require more dietary
protein to sustain growth. However, the results presented here are far from conclu-
sive, since the older animals used in experiment 1 were not tested at the lower protein
level (2.5 % casein) used to test the young weanlings in experiment 3. More research
is needed to elucidate the relationship between low-protein DIT and the protein needs
of fully grown adult animals.

Voles required less dietary protein to stimulate DI’T than laboratory rats. The
most likely explanation for this is that voles are better adapted at assimilating nutri—
ents from a low-quality diet. This is probably due to selective pressures associated
with herbivory that may have led to the evolution of high weight-speciﬁc BMRs in
voles. Weight-speciﬁc metabolic rates for voles (experiment 3) were several times
higher than those for rats in a similar study (Rothwell and Stock 1987c), indicating a
high metabolic turnover rate for voles. Although the rats used by Rothwell and Stock
(1987c) were much larger than the voles used in experiment 3, the differences in
metabolic rates were higher than predicted based on weight differences alone. A high
metabolic turnover rate could facilitate oxidation of excess food intake and increase

the rate of uptake of essential nutrients such as dietary nitrogen by distillation from a

 

‘For convenience, a graphical summary of some of the quantitative data
discussed in Chapter III is provided at the end of this chapter.

 

137

high bulk intake of low-nitrogen forage. Thus, herbivorous voles may possess greater
intrinsic capabilities for oxidation of bulk, low-quality foodstuffs than omnivorous
species such as rats.

Data on the speciﬁc binding of GDP to brown fat mitochondria in voles
(Figure 4.2) provides no evidence that BAT is involved in the DIT response to low
protein intake observed in these animals. However, an elevated NST response,
known to be mediated by BAT in small mammals, may have clouded these results.
Therefore, more research is needed in this area before any conclusions can be made.
Energy balance studies involving analysis of BAT should probably be conducted at
temperatures just below thermoneutrality, to minimize NST and prevent the putative
hyperthermic suppression of DIT.

Prairie voles showed a deﬁnite hyperphagic response to low-protein diets,
demonstrating that they can behaviorally change their intake in response to the nutri-
tional quality of their food supply. For voles feeding on a 2.5 % casein diet in
experiment 7, the rate of food consumption increased over the 21-day feeding trial for
mean daily food consumption and weight-speciﬁc mean daily food consumption.
Over the same period, mean daily food consumption and weight-speciﬁc mean daily
food consumption declined in voles fed a 15% casein diet (Figures 3.8 and 4.5).
However, gross food consumption for the ﬁrst 10 days (total food consumed for 10
days) by the low-protein group showed no evidence of excess intake (Figure 4.6) and
was statistically lower than the 10—day food intake recorded for the 15 % casein group
(Table 3.7). Since voles consuming a 2.5% casein diet for 10 days showed a lower

energy efﬁciency than voles feeding on a 15 % casein diet (Figure 4.1), yet consumed

 

138
less food during that period, it is unlikely that the low energy efﬁciency of the 2.5 %

casein group was triggered by excess gross energy intake. It is more probable that
the low energy efﬁciency observed in low-protein-fed voles was an acute response,
possibly caused by an imbalance of nutrients in the diet due to the low ratio of protein
to energy in the diet.

For voles consuming low-protein diets, ‘excess’ energy consumption may be
meal-based and result from the daily intake of a high-energy, low-protein diet. In this
view, excess energy is energy that is present in the diet in surplus of what might be
considered a ‘balanced diet’ for protein and energy content. It is possible that voles
respond in the acute time frame and oxidize daily increments of excess energy
ingested from an unbalanced diet, thus displaying a rapid metabolic response to acute
variations in nutrient and energy intake.

However, it is also possible that voles can respond within a chronic time frame
to eliminate excess gross intake by DIT. This is evidenced by the steady increase in
food intake over a 21-day period associated with only a moderate gain in weight
(Table 3.9 and Figure 4.5). More lengthy and comprehensive studies of energy
balance and food consumption are needed to differentiate between immediate and
long-term thermogenic responses.

The data for tissue changes in voles showed that animals on high-protein diets
exhibited the greatest increases in protein, lipid, water, and other (Figure 4.7). Voles
fed diets low in protein showed only a moderate increase in dry tissue. However, the

relative increases in dry tissue components were the same for low-, medium-, and

 

139

high-protein groups, as were the relative proportions of tissue constituents in the ﬁnal
dry tissue mass (Tables 3.9-3.11 and Figure 4.8).

Why the relative proportions of tissue constituents in the ﬁnal dry tissue mass
did not differ between dietary groups is not clear, but one possibility is that growth
could be restricted to proportional increases in tissue deposition and may be dependent
on the relative availability of nutrients in the diet. If such a constraint exists, then an
ability to differentially extract limiting nutrients from an unbalanced diet could aid in
sustaining growth. Moreover, if we accept the premise that there is a developmental
limitation based on proportional growth, then a metabolic response like DIT would be
a necessity for growing animals consuming unbalanced diets, and therefore is likely to
have evolved by natural selection.

Nitrogen balance data imply that low-protein-fed voles use dietary nitrogen
more efﬁciently than other diet components. Gross nitrogen efﬁciencies were higher
than total energy efﬁciencies for both 2.5 % and 5 % casein groups (Figures 4.1 and
4.3), although the difference between mean energy and nitrogen efﬁciencies is
probably not signiﬁcantly different for the 2.5 % casein group. Also, when dietary
nitrogen is limited, a small increase in nitrogen content can substantially increase
gross nitrogen efﬁciency. This suggests that a small hyperphagic response by voles to
low-protein diets could result in a substantial increase in growth, since total protein
intake increases with increased feeding. The 2.5 % casein group did not show rapid
growth, but as they increased food intake, they did recover initial growth loss and
even had a moderate net gain (Figure 4.4 and Table 3.9). A more lengthy experiment

measuring nitrogen balance and food consumption is needed to see if hyperphagia is

140

persistent and can augment growth in body protein over an extended length of time.
If hyperphagia and low energy efﬁciencies persist during long-term experiments, then
excess gross energy consumption could be removed by oxidation, thus differentially
enhancing nitrogen retention.

In summary, nutritional studies of wild herbivores suggest that nitrogen is a
limiting nutrient in natural diets. Since laboratory animals respond to low-protein
diets by increasing metabolic heat production per unit of ME intake, it may be that
this thermogenenic response (termed diet-induced thermogenesis or DIT) is adaptive
for animals in the wild that normally consume foodstuffs low in protein. A possible
adaptive function for DIT is that it may aid in the acquisition of essential nitrogen
from a low—quality food supply. This could be accomplished by the oxidation of
nonessential caloric intake resulting in a distillation of nitrogen from a diet low in
essential nitrogen. The results of experiments using wild herbivorous voles conﬁrm
that energy efﬁciencies are reduced as dietary protein declines. Also, voles show
higher metabolic turnover rates than laboratory rats consuming similar low-quality
diets, suggesting a greater inherent capacity for responding to the deleterious effects
of low-nutrient intake via heightened metabolic rates.

Voles respond to low dietary protein by increasing food intake, reducing
efﬁciency of energy retention, and differentially increasing nitrogen retention efﬁcien-
cy. Thus, the metabolic distillation of protein from a nutrient-poor diet by DIT
appears to be a possible adaptive trait and could play a central role in the nutritional

ecology of wild animals that normally consume diets low in nutrients such as protein.

141

 

 

 

8
.1
i
i
4
.4

 

x

til-Illa“
£5
—+—

 

 

am
(Ion! WY GAIN/II) 81.

 

 

 

 

 

 

 

- q
i ~ P .
. b |+l .
r - a , 1 7 7 a
In. I'

 

1
o u: I
ma I—ll
m

 

 

 

 

 

 

Figure 4-1- Efﬁciency 0f energy reten- Figure 4.2. GDP binding, expts. 4 and

tion, expt. 3. Error bars are :1; SEM. 6. L = low protein, 25% casein; H =
high protein, 15% casein; C = cold
(22°C); W = warm (28°C). Error bars
are :1: SEM.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Wﬁ‘

hula-Ida

.9!“

d 1. llllllllllll .mu
............

19' Figure 4.4. Growth of weanling voles,
expt. 7. Error bars are :1: SEM.

Ls

 

 

 

Figure 4.3. Gross nitrogen efﬁciencies,
expt. 7 ((body N gain/dietary N intake) x
100). Error bars are :1; SEM.

142

 

 

 

 

  
   

 

 

 

 

 

 

 

 

 

 

,- I. i 1 a.
i
3 m I'. .. - ...... g”.
I. h _- _
u -" E n- -
i'r ‘ we: 1"" * $-
10 _ I mou- § N
V .. ”I“ E - \-:-.\1:;.
5 ca 1 a 1 A lama S %
I l I'mnu ’ I 1. § § :33:
I“ . I”
mm

 

 

Figure 4.5. Mean daily food consump-

tion, expt. 7. Figure 4.6. Total mean food intake per
vole for 10 and 21 days during expt. 7.
Error bars are i SEM.

 

 

 

 

 

 

 

 

 

 

5 l 1 l
4- -
3- -
E“ ‘
a 1- -
“'— .......
Ion-rein.
‘l” ‘ l'atarChaua
rat-woma-
_: I l 1 0mm
25. 5! I”
610!!!

 

 

 

Figure 4.7. Data are for weight changes in tissue constituents over a 21-day
feeding period during expt. 7. Error bars are :1; SEM.

143

 

 

    

ooooooooooooooooooooooooooooooooooooooooooooo

'l
0000000000000000000000000000000000000000000 ‘

.uu VVVVVVVVV.\VVVVVVVVVVVVVVVVVVN «-

 

 

. p . ,
a u m m
Boar run .3 a

 

 

 

Figure 4. 8. Relative proportions of dry tissue constituents for voles at the end of

expt. 7. Error bars are :1: SEM.

LIST OF REFERENCES
American Institute of Nutrition 1977. Report of the AIN Ad Hoc Committee on
standards for nutritional studies. J. Nutr. , 107:1340—1348.

Andrewartha, H.G. and L.C. Birch 1954. The Distribution and Abundance of Ani- f‘
mals. Chicago: University of Chicago Press.

Andrzejewski, R. 1975. Supplementary food and the winter dynamics of bank vole
populations. Acta Theriol. , 20:23-40.

 

Applebaum, S.W. and Y. Birk 1979. Saponins. In Herbivores: Their Interaction with
Secondaty Plant Metabolites, pp. 539-566, eds. G.A. Rosenthal and DH.
J anzen. New York: Academic Press.

Arditi, R., L.R. Ginzburg, and HR. Akcakaya 1991. Variation in plankton densities
among lakes: a case for ratio-dependent predation models. Am. Nat. ,
138(5): 1287-1296.

Armitage, G., G.R. Hervey, B.J. Rolls, E.A. Rowe, and G. Tobin 1981. Energy
balance in young ‘cafeteria’-fed rats. J. Physiol., 317:48P—49P.

Ashwell, M., S. Holt, G. Jennings, D.M. Stirling, P. Trayhum, and DA. York
1985 . Measurement by radioimmunoassay of the mitochondrial uncoupling
protein from brown adipose tissue of obese (ob/ ob) mice and Zucker (fa/ fa)
rats of different ages. FEBS Lett. , 179(2):233-237.

Baker, R.H. 1983. Michigan Mammals. East Lansing: Michigan State University
Press.

Banach, K. 1986. The effect of increased food supply on the body growth rate and
survival of bank voles in an island population. Acta Theriol. , 31:45-54.

Barbosa, P. and D.K. Letoumeau, eds. , 1988. Novel Aspects of Insect—Plant Interac-
tions. New York: John Wiley & Sons.

Barlow, CA. and RA. Randolph 1978. Quality and quantity of plant sap available to
the pea aphid. Ann. Entomol. Soc. Am, 71:46—48.

144

 

145

Barr, H.G. and K.J. McCracken 1982. Absence of ‘diet-induced thermogenesis’ in
growing rats kept at 29° and offered a varied diet. Proc. Nutr. Soc. , 41:63A.

Barry, T.N., T.R. Manley, and 8.]. Duncan 1986. The role of condensed tannins in
the nutritional value of Lotus pedunculatus for sheep. 4. Sites of carbohydrate
and protein digestion as influenced by dietary reactive tannin concentration.
Brit. J. Nutr., 55:123-137.

Batzli, GO. and F .R. Cole 1979. Nutritional ecology of microtine rodents: digestibil-
ity of forage. J. Mamm., 60:740-750.

Batzli, GO. 1985. Nutrition. In Biology of New World Microtus, pp. 779-811, ed.
R.H. Tamarin. Am. Soc. Mamm. Spec. Pub. No. 8.

Bell, EA. 1978. Toxins in swds. In Biochemical Aspects of Plant and Animal
Coevolution, pp. 143-161, ed. J .B. Harbome. London: Academic Press.

Bell, R.H.V. 1970. The use of the herb layer by grazing ungulates in the Serengeti.
In Animal Populations in Relation to Their Food Resources, pp. 111-124, ed.
A. Watson. Oxford: Blackwell Scientiﬁc Publications.

Bell, R.H.V. 1971. A grazing system in the Serengeti. Scientiﬁc American, 225:86-
93.

Belovsky, G.E. and 0.]. Schmitz 1991. Mammalian herbivore optimal foraging and
the role of plant defenses. In Plant Defenses Against Mammalian Herbivory,
pp. 1-28, eds. R.T. Palo and CT. Robbins. Boca Raton: CRC Press.

Benz, G. 1974. Negative Ruckkoppelung durch Raum- und Nahrungskonkurrenz
sowie Zyklische Veranderung der Nahrungsgrundlage als Regelprinzip in der
Populationsdynamic des Gruen Larchenwicklers, Zeriaphera diniana (Guenee)
(Lep. Tortricidae). Z. Angew. Entomol. , 76:196-228.

Benz, G. 1977. Insect-induced resistance as a means of self defense in plants. Eucatp-
ia/IOBC Work. Group Breed. Resistance Insects Mites, Bull, SROP,
1977/1978, pp. 155-159.

Berenbaum, M. 1980. Adaptive signiﬁcance of midgut pH in larval Lepidoptera. Am.
Nat., 115:138-146.

Berenbaum, M. 1981 . Patterns of furanocoumarin distribution and insect herbivory in
the Umbelliferae: plant chemistry and community structure. Ecology, 63: 1254-
1266.

 

 

146

Bergeron, J .M. and L. Jodoin 1987. Deﬁning “high quality” food resources of herbi-
vores: the case for meadow voles (Microtus pennsylvanicus). Oecologia,
71:510-517.

Bergeron, J.M. and L. Jodoin 1989. Patterns of resource use, food quality, and health
status of voles (Microtus pennsylvanicus) trapped from ﬂuctuating populations.
Oecologia, 79:306-314.

Bergeron, J .M., L. Jodoin, and Y. Jean 1987. Pathology of voles (Microtus pennsyl-
vanicus) fed with plant extracts. J. Mamm., 68:73-79.

Bemays, E.A., G.C. Driver, and M. Bilgener 1989. Herbivores and plant tannins. In
Advances in Ecological Research, vol. 19, pp. 263-302, eds. M. Begon, A.H.
Fitter, E.D. Ford, and A. Macfadyen. London: Academic Press.

Bestley, J.W., P.N. Bramley, P.M.S. Dobson, A. Mahanty, and G. Tobin 1982.
Energy balance in ‘cafeteria’-fed young ‘Charles River’ Sprague-Dawley rats.
J. Physiol. , 330:7OP-71P.

Bligh, J. and K.G. Johnson 1973. Glossary of terms for thermal physiology. J.
Applied Physiol. , 35(6):94l-961.

Boutin, S. 1984. Effect of late winter food addition on numbers and movements of
snowshoe hares. Oecologia, 62:393-400.

Boutin, S. 1990. Food supplementation experiments with terrestrial vertebrates:
patterns, problems, and the future. Can. J. Zool., 68:203-220.

Brooks, S.L., N.J. Rothwell, and MJ. Stock 1980. Increased proton conductance
pathway in brown adipose tissue mitochondria of rats exhibiting diet-induced
thermogenesis. Nature, 286:274-276.

Bryant, J .P. 1981. Phytochemical deterrence of snowshoe hare browsing by adventi-
tious shoots of four Alaskan trees. Science, 213:889-890.

Bryant, J .P., F.S. Chapin, and DR. Klein 1983. Carbon/nutrient balance of boreal
plants in relation to vertebrate herbivory. Oikos, 40:357-368.

Bryant, J .P., F.D. Provenza, J. Pastor, P.B. Reichardt, T.P. Clausen, and LT. du
Toit 1991. Interactions between woody plants and browsing mammals mediat-
ed by secondary metabolites. Annu. Rev. Ecol. Syst., 22:431-446.

Bryant, J .P., G.D. Wieland, T. Clausen, and P. Kuropat 1985. Interactions of
snowshoe hare and feltleaf willow in Alaska. Ecology, 66(5): 1564-1573.

 

 

 

147

Buchsbaum, R., J. Wilson, and I. Valiela 1986. Digestibility of plant constituents by
Canada geese and Atlantic brant. Ecology, 67(2):386—393.

Bucyanayandi, J .D. and J.M. Bergeron 1990. Effects of food quality on feeding
patterns of meadow voles (Microtus pennsylvanicus) along a community
gradient. J. Mamm., 71(3):390-396.

Bukowiecki, L.J . 1986. Lipid metabolism in brown adipose tissue. In Brown Adipose
Tissue, pp. 105-121, eds. P. Trayhum and D.G. Nicholls. London: Edward
Arnold.

Cannon, B. and O. Lindberg 1979. Mitochondria from brown adipose tissue: isolation
and properties. Methods Enzymol. , 55:65-78.

Cates, R.G. 1980. Feeding patterns of monophagous, oligophagous and polyphagous
insect herbivores: the effect of resource abundance and plant chemistry.
Oecologia, 46:22-31.

Cates, R.G. and DP. Rhoades 1977. Patterns in the production of antiherbivore
chemical defenses in plant communities. Biochem. Syst. Ecol. , 5: 185-193.

Chapin, F.S., III 1980. The mineral nutrition of wild plants. Ann. Rev. Ecol. Syst. ,
11:233-260.

Chiang, H., D.M. Norris, A. Ciepiela, P. Shapiro, and A. Oosterwyk 1987. Induc-
ible versus constitutive PI 227687 soybean resistance to Mexican bean beetle,
Epilachna varivestis. J. Chem. Ecol., 13(4):741-750.

Clausen, T.P., P.B. Reichardt, J.P. Bryant, R.A. Werner, K. Post, and K. Frisby
1989. Chemical model for short-term induction in quaking aspen (Populus
tremuloides) foliage against herbivores. J. Chem. Ecol. , 15(9):2335-2346.

Coe, M.J., D.H. Cumming, and J. Phillipson 1976. Biomass and production of large
African herbivores in relation to rainfall and primary production. Oecologia,
22:341-354.

Cole, RR. and G.C. Batzli 1978. Inﬂuence of supplemental fwding on a vole
p0pulation. J. Mamm., 59:809-819.

Cole, RR. and G.C. Batzli. 1979. Nutrition and population dynamics of the prairie
vole, Microtus ochrogaster, in central Illinois. J. Anim. Ecol. , 48:455-470.

Coley, P.D., J.P. Bryant, and RS. Chapin 1985. Resource availability and plant
antiherbivore defense. Science, 230:895-899.

 

148

Cooper, S.M. and N. Owen-Smith 1986. Effects of plant Spinescence on large
mammalian herbivores. Oecologia, 68:446-455.

Cooper-Driver, G., S. Finch, and T. Swain 1977. Seasonal variation in secondary
plant compounds in relation to palatability of Pteridium aquilinum. Biochem.
Syst. Ecol., 5:177-183.

Crompton, A.W., CR. Taylor, and J .A. Jagger 1978. Evolution of homeothermy in
mammals. Nature, 272:333-336.

Demment, M.W. and R]. Van Soest 1985. A nutritional explanation for body-size
patterns of ruminant and nonruminant herbivores. Am. Nat. ,125(5):641-672.

Denno, RF. and M.S. McClure, eds., 1983. Variable Plants and Herbivores in
Natural and Managed Systems. New York: Academic Press.

Desautels, M. , G. Zaror-Behrens, and J. Himms-Hagen 1978. Increased purine
nucleotide binding, altered popypeptide composition, and thermogenesis in

brown adipose tissue mitochondria of cold-acclimated rats. Can. J. Biochem. ,
56:378-383.

Desy, EA. and GO. Batzli 1989. Effects of food availability and predation on
prairie vole demography: a ﬁeld experiment. Ecology, 70(2):411-421.

Desy, EA. and C.F. Thompson 1983. Effects of supplemental food on a Microtus
pennsylvanicus population in central Illinois. J. Anim. Ecol. , 52: 127-140.

Donald, P., G.C. Pitts, and S.L. Pohl 1981. Body weight and composition in labora-
tory rats: effects of diets with high or low protein concentrations. Science,
211:185-186.

Edozien, LC. and B.R. Switzer 1978. Inﬂuence of diet on growth in the rat. J. Nutr.,
108:282-290.

Ehrlich, RR. and L.C. Birch 1967. The “balance of nature” and “population con-
trol.” Am. Nat., 101:97-107.

Elton, C. 1936. Animal Ecology. New York: MacMillan.

Evans, D.M. 1973. Seasonal variations in the body composition and nutrition of the
vole Microtus agrestis., J. Anim. Ecol. , 42:1-18.

Falcou, R., F. Bouillaud, G. Mory, M. Apfelbaum, and D. Ricquier 1985. Increase
of uncoupling protein and its mRNA in brown adipose tissue of rats fed on
‘cafeteria diet’. Biochem. J., 231:241-244.

 

149

Feeny, P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of
spring feeding by winter moth caterpillars. Ecology, 51:565-581.

Feeny, P. 1976. Plant apparency and chemical defense. In Recent Advances in
Phytochemistry, vol. 10, pp. 1-40, eds. J.W. Wallace and R.L. Mansell. New
York: Plenum Press.

Felsenstein, J. 1985. Phylogenies and the comparative method. Am. Nat. , 125:1-15.

Ford, R.G. and F.A. Pitelka 1984. Resource limitation in populations of the Califor-
nia vole. Ecology, 65:122-136.

Foster, D.G. and M.L. Frydman 1978. Nonshivering thermogenesis in the rat. 11.
Measurements of blood ﬂow with microspheres point to brown adipose tissue
as the dominant site of calorigenesis induced by noradrenaline. Can. J.
Physiol. Phannacol. , 56:1 10-122.

Foster, DD. and S.W.Y. Ma 1989. The effector of diet—induced thermogenesis:
brown adipose tissue of liver? In Hormones, Thermogenesis, and Obesity, pp.
165-171, eds. H. Lardy and F. Stratrnan. New York: Elsevier.

Fox, LR. 1981. Defense dynamics in plant-herbivore systems. Amer. Zool. , 21:853-
864.

Fraenkel, G. 1959. The raison d’etre of secondary plant substances. Science,
129: 1466-1470.

Freeland, W.J., P.H. Calcott and LR. Anderson 1985. Tannins and saponin: interac-
tion in herbivore diets. Biochem. Syst. Ecol. , 13(2): 189-193.

Freeland, W.J. and DH. Janzen 1974. Strategies in herbivory by mammals: the role
of plant secondary compounds. Am. Nat. , 108:269-289.

Fretwell, SD. 1977. The regulation of plant communities by the food chains exploit-
ing them. Perspect. Biol. Med., 20:169-185.

Fretwell, SD. 1987. Food chain dynamics: the central theory of ecology? Oikos,
50:291-301.

Fuller, M.F. 1983. Energy and nitrogen balances in young pigs maintained at con-
stant weight with diets of differing protein content. J. Nutr. , 113:15-20.

Futuyma, DJ. 1976. Food plant specialization and environmental predictability in
Lepidoptera. Am. Nat., 110:285-292.

150

Gardarsson, A. and R. Moss 1970. Selection of food by Icelandic ptarmigan in
relation to its availability and nutritive value. In Animal Populations In
Relation To Their Food Resources, pp. 47-71, ed. A. Watson, Oxford: Black-
well Scientiﬁc Publications.

Garrow, J.S. 1978. Energy Balance and Obesity in Man. Amsterdam: Elsevier.

Gettinger, RD. 1984. Seasonal patterns of nitrogen utilization by pocket gophers,
Thomomys bottae. Comp. Biochem. Physiol. 78A(4):657—659.

Gilbert, BS. and OJ. Krebs 1981. Effects of extra food on Peromyscus and Clethri-
onomys populations in the southern Yukon. Oecologia, 51:326-331.

Glick, Z. 1987. The thermic effect of a meal. J. Obesity Wt. Reg., 6:170—178.

Glick, Z. and M.A. J oslyn 1970. Food intake depression and other metabolic effects
of tannic aci in the rat. J. Nutr. , 100:509—515.

Glick, Z., R]. Teague, and G.A. Bray 1981. Brown adipose tissue: thermic response
increased by a single low protein, carbohydrate meal. Science, 213: 1 125-1127.

Goldberg, J .C. and B.L.G. Morgan 1983. The response of brown adipose tissue
(BAT) to cafeteria feeding in the Zucker rat. Fed. Proc. , 42(5): 1189.

Goldberg, M., N.R. Tabroff, and R.H. Tamarin 1980. Nutrient variation in beach
grass in relation to beach vole feeding. Ecology, 61(5): 1029-1033.

Gould, SJ. and RC. Lewontin. 1979. The spandrels of San Marco and the Pan-
glossian paradigm: a critique of the adaptationist programme. Proc. R. Soc.
Lond. B, 205:581-598.

Grant, S. 1984. Beauty and the Beast: The Coevolution of Plants and Animals. New
York: Charles Scribner’s Sons.

Gross, LE, 2. Wang, and BA. Wunder 1985. Effects of food quality and energy
nwds: changes in gut morphology and capacity of Microtus ochrogaster. J.
Mamm., 66(4):661-667.

Gulick, A. 1922. A study of weight regulation in the adult human body during over-
nutrition. Am. J. Physiol., 60:371-395.

Gurr, M.J., R. Mawson, N.J. Rothwell, and MJ. Stock 1980. Effects of manipulat-
ing dietary protein and energy intake on energy balance and thermogenesis in
the pig. J. Nutr., 110:532-542.

 

151

Hairston, G.N., F.E. Smith, and LB. Slobodkin 1960. Community structure,
population control, and competition. Am. Nat. , 94:421-425.

Hamilton, T.S. 1939a. The heat increments of diets balanced and unbalanced with
respect to protein. J. Nutr., 17:583-599.

Hamilton, T.S. 1939b. The growth, activity, and composition of rats fed diets
balanced and unbalanced with respect to protein. J. Nutr. , 17:565-582.

Hamilton, W.J., Jr. 1941. The food of small forest mammals in eastern United
States. J. Mamm., 22(3):250—263.

Hammond, K.A. and BA. Wunder 1991. The role of diet quality and energy need in
the nutritional ecology of a small herbivore, Microtus ochrogaster. Physiol.
Zool., 64(2):541-567.

Hanhimaeki, S. 1989. Induced resistance in mountain birch: defence against leaf-
chewing insect guild and herbivore competition. Oecologia, 81(2):242-248.

Hanley, T.A., C.T. Robbins, A.E. Hagerman, and C. McArthur 1992. Predicting
digestible protein and digestible dry matter in tannin-containing forages
consumed by ruminants. Ecology, 73(2):537-541.

Hayssen, V. and RC. Lacy 1985. Basal metabolic rates in mammals: taxonomic
differences in the allometry of BMR and body mass. Comp. Biochem.
Physiol., 81A:741-754.

Heldmaier, G. 1989. Seasonal acclimatization of energy requirements in mammals:
functional signiﬁcance of body weight control, hypothermia, torpor and
hibernation. In Energy Transformations in Cells and Organisms, Proceedings
of the 10th Conference of the European Society for Comparative Physiology
and Biochemistry, pp. 130-139, eds. W. Wieser and E. Gnaiger. Stuttgart:
Georg Thieme Verlag.

Helrich, K. , ed., 1990. Oﬁ‘icial Methods of Analysis of the Association of Oﬁicial
Analytical Chemists. Arlington: Association of Ofﬁcial Analytical Chemists.

Henderson, QB. 1990. The inﬂuence of seed apparency, nutrient content and chemi-
cal defenses on dietary preference in Dipodomys ordii. Oecologia, 82:333-341.

Hervey, G.R. and G. Tobin 1981. Brown adipose tissue and diet-induced therrnogene-
sis. Nature, 289:699.

Hervey, G.R. and G. Tobin 1982. The part played by variation of energy expenditure
in the regulation of energy balance. Proc. Nutr. Soc. , 41: 137-153.

 

152

Hervey, G.R. and G. Tobin 1983. Luxuskonsumption, diet-induced thermogenesis
and brown fat: a critical review. Clin. Science, 64:7-18.

Hill, R.W. 1975. Daily torpor in Peromyscus leucopus on an adequate diet. Comp.
Biochem. Physiol., 51A:413-423.

Hill, R.W. and G.A. Wyse. 1989. Animal Physiology. Harper and Row: New York.

Himms-Hagen, J. 1986. Cold- versus diet-induced thermogenesis in brown adipose
tissue: different strategies in different species. In Living in the Cold, pp. 93-
100, eds. H.C. Heller, X.J. Musacchia and L.C.H. Wang. New York:
Elsevier.

Himms—Hagen, J. 1989. Role of thermogenesis in the regulation of energy balance in
relation to obesity. Can. J. Physiol. Pharmacol. , 67:394-401.

Himms-Hagen, J. 1990. Brown adipose tissue thermogenesis: interdisciplinary studies.
FASEB J. , 4:2890-2898.

Himms-Hagen J. and M. Desautels 1978. A mitochondrial defect in brown adipose
tissue of the obese (ob/ob) mouse: reduced binding of purine nucleotides and a
failure to respond to cold by an increase in binding. Biochem. Biophys. Res.
Commun. , 83(2):628-634.

I-Iimms-Hagen, J ., J. Triandaﬁllou, and C. Gwilliam 1981. Brown adipose tissue of
cafeteria-fed rats. Am. J. Physiol., 241(4):E116-E120.

Hladik, A. 1978. Phenology of leaf production in rain forest of Gabon: distribution
and composition of food for folivores. In The Ecology of Arboreal Folivores,
pp. 51-71, ed. G.G. Montgomery. Washington: Smithsonian.

Hoffmann, RS. and J.W. Koeppl 1985. Zoogeography. In Biology of New World
Microtus, pp. 84-115, ed. R.H. Tamarin. Am. Soc. Mamm. Spec. Pub. No.
8.

Hogan, S., D.V. Coscina, and J. Himms-Hagen 1982. Brown adipose tissue of rats
with obesity-inducing ventromedial hypothalamic lesions. Am. J. Physiol. ,
243(6):E338-E344.

Horseﬁeld, D. 1977. Relationship between feeding of Philaenus spumarius (L.) and
the amino acid concentration in the xylem sap. Ecol. Entomol. , 2:259-266.

Howe, HF. and L.C. Westley 1988. Ecological Relationships of Plants and Animals.
New York: Oxford University Press.

 

153

Hulbert, A.J. and T.J. Dawson 1974. Standard metabolism and body temperature of
perameloid marsupials from different environments. Comp. Biochem. Physiol. ,
47Az583-590.

Ikeda, T., F. Matsumura, and D.M. Benjamin 1977. Mechanisms of feeding discrimi-
nation between matured and juvenile foliage by two species of pine sawﬂies. J.
Chem. Ecol. , 3:677-694.

Janis, C. 1976. The evolutionary strategy of the Equidae and the origins of rumen and
cecal digestion. Evolution, 30:757—774.

Jansky, L. 1973. Non-shivering thermogenesis and its thermoregulatory signiﬁcance.
Biol.Rev., 48:85-132.

Janzen, D.H. 1978. Complications in interpreting the chemical defenses of trees
against tropical arboreal plant-eating vertebrates. In The Ecology of Arboreal
Folivores, pp. 73-84, ed. G.G. Montgomery. Washington, D.C.: Smithsonian
Institution Press.

J effries, M. 1990. Evidence of induced plant defences in a pondweed. Freshwat.
Biol. , 23(2):265-269.

Jones, D.I.H. and AD. Wilson 1987. Nutritive quality of forage. In The Nutrition of
Herbivores, pp. 87-89, eds. J.B. Hacker and J.H. Ternouth. Orlando: Aca-
demic Press.

Jung, H.G. and GO. Batzli 1981. Nutritional ecology of microtine rodents: effects of
plant extracts on the growth of arctic microtines. J. Mamm. , 62(2):286-292.

Karasov, W.H. 1982. Energy assimilation, nitrogen requirement, and diet in free-
living antelope ground squirrels, Arnmospennophilus leucurus. Physiol. 2001. ,
55:378-392.

Karasov, W.H. 1985 . Nutrient constraints in the feeding ecology of an omnivore in a
seasonal environment. Oecologia, 66:280-290.

Karban, R. 1987. Environmental conditions affecting the strength of induced resis-
tance against mites in cotton. Oecologia, 73(3):414-419.

Kendall, W.A. and K.T. Leath 1976. Effect of saponins on palatability of alfalfa the
meadow voles. Agron. J., 68:473-476.

Kendall, W.A. and RT. Sherwood 1975. Palatability of leaves of tall fescue and reed
canarygrass and some of their alkaloids to meadow voles. Agron. J. , 67 :667-
671.

154

Kevonian, A.V., J.G. Vander Tuig and DR. Romsos 1984. Consumption of a low
protein diet increases norepinephrine turnover in brown adipose tissue of adult
rats. J. Nutr., 114:543-549.

King, LR. and M.F. Murphy 1985. Periods of nutritional stress in the annual cycles
of endotherms: fact or ﬁction? Amer. Zool. , 25:955-964.

Klaus, S., G. Heldmaier, and D. Ricquier 1988. Seasonal acclimation of bank voles
and wood mice: nonshivering thermogenic properties of brown adipose tissue
mitochondria. J. Comp. Physiol. , 158: 157-164.

Kleiber, M. 1932. Body size and metabolism. Hilgardia, 6:315-353.

Kleiber, M. 1945 . Dietary deﬁciencies and energy metabolism. Nutr. Abstr. Rev. ,
15:207-222.

Kleiber, M. 1947. Body size and metabolic rate. Physiol. Rev. , 27:511-541.

Kleiber, M. 1975a. The Fire of Life: an Introduction to Animal Energetics. Hunting-
ton: Krieger.

Kleiber, M. 1975b. Metabolic turnover rate: a physiological meaning of the metabolic
rate per unit body weight. J. theor. Biol. , 53:199-204.

Klein, DR. 1970. Food selection by North American deer and their response to over-
utilization of preferred plant species. In Animal Populations In Relation To
Their Food Resources, pp. 25-46, ed. A. Watson. Oxford: Blackwell Scientiﬁc
Publications.

Kopecky, J ., L. Sigurdson, I.R.A. Park, and J. I-Iimms-Hagen 1986. Thyroxine 5’-
deiodinase in hamster and rat brown adipose tissue: effect of cold and diet.
Am. J. Physiol., 251(14):E1-E7.

Krebs, CJ. and K.T. DeLong 1965. A Microtus population with supplemental food.
J. Mamm., 46:566-573.

Krebs, C.J., S. Boutin, and BS. Gilbert 1986. A natural feeding experiment on a
declining snowshoe hare population. Oecologia, 70: 194-197.

Krebs, C.J., B.S. Gilbert, S. Boutin, A.R.E. Sinclair, and J.N.M. Smith 1986.
Population biology of snowshoe hares. I. Demography of food-supplemented
populations in the southern Yukon, 1976-84. J. Anim. Ecol., 55:963-982.

Krebs, J .R. and NB. Davies 1987. An Introduction to Behavioral Ecology.
Sunderland: Sinauer Associates.

 

155

Landsberg, L. and LB. Young 1983. Autonomic regulation of thermogenesis. In
Mammalian Thermogenesis, pp. 99-140, eds. L. Girardier and MJ. Stock.
London: Chapman and Hall.

Leather, S.R., A.D. Watt, and 6.1. Forrest 1987. Insect-induced cherrrical changes in
young lodgepole pine (Pinus contorta): the effect of previous defoliation on
oviposition, growth and survival of the pine beauty moth, Panolis flammea.
Ecol. Entomol. , 12(3):275-281.

Lin, H., M. Kogan, and D. Fischer 1990. Induced resistance in soybean to the
Mexican bean beetle (Coleoptera: Coccinellidae): comparisons of inducing
factors. Environ. Entomol. , 19(6):1852-1857.

Lindroth, R.L. 1988. Adaptations of mammalian herbivores to plant chemical defens-
es. In Chemical Mediation of Coevolution, pp. 415-445, ed. K.C. Spencer.
San Diego: Academic Press.

Lindroth, R.L. 1989. Mammalian herbivore-plant interactions. In Plant-Animal
Interactions, pp. 123-206, ed. W.G. Abrahamson. New York: McGraw-Hill.

Lindroth, R.L. 1991. Differential toxicity of plant allelochemicals to insects: roles of
enzymatic detoxication systems. In Insect-Plant Interactions, pp. 1-33, ed. E.
Bemays. Boca Raton: CRC Press.

Lindroth, R.L. and GO. Batzli 1983. Detoxiﬁcation of some naturally occurring
phenolics by prairie voles: a rapid assay of glucuronidation metabolism.
Biochem. Syst. Ecol. , 11:405-409.

Lindroth, R.L. and GO. Batzli 1984a. Food habits of the meadow vole (Microtus
pennsylvanicus) in bluegrass and prairie habitats. J. Mamm. , 65(4):600-606.

Lindroth, R.L. and GO. Batzli 1984b. Plant phenolics as chemical defenses: effects
of natural phenolics on survival and growth of prairie voles (Microtus ochro-
gaster). J. Chem. Ecol., 10:229-244.

Lindroth, R.L. and GO. Batzli 1986. Inducible plant chemical defenses: a cause of
vole population cycles? J. Anim. Ecol. , 55:431-449.

Lindroth, R.L., G.O. Batzli, and SJ. Avildsen 1986. Lespedeza phenolics and
Penstemon alkaloids: effects on digestion and growth of voles. J. Chem. Ecol. ,
12:713-728.

Ma, S.W.Y. and DO. Foster 1989. Brown adipose tissue, liver, and diet-induced
thermogenesis in cafateria diet-fed rats. Can. J. Physiol. Phannacol. , 67 :376-
381.

156

Markwell, M.A.K., S.M. Haas, N.E. Tolbert, and LL. Bieber 1981. Protein
determination in membrane and lipoprotein samples: manual and automated
procedures. Methods Enzymol. , 72:296-303.

Martin, ER 1956. A population study of the prairie vole (Microtus ochrogaster) in
northeastern Kansas. Univ. Kansas Pub. Mus. Nat. Hist, 8(6):361-416.

Mattson, W.J., Jr. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev.
Ecol. Syst., 11:119-161.

Mayr, E. 1983. How to carry out the adaptationist program? Am. Nat., 121(3):324-
334.

McArthur, C., A.E. Hagerman, and CT. Robbins 1991. Physiological strategies of
mammalian herbivores against plant defenses. In Plant Defenses Against
Mammalian Herbivory, pp. 103-131, eds. R.T. Palo and CT. Robbins. Boca
Raton: CRC Press.

McBee, R.H. 1971. Signiﬁcance of intestinal microﬂora in herbivory. Ann. Rev.
Ecol., 2:165-176.

McCracken, K.J. 1975a. The effect of overfeeding on normal adult rats. Proc. Nutr.
Soc., 34:15A-16A.

McCracken, K.J. 1975b. Effect of feeding pattern on the energy metabolism of rats
given low-protein diets. Br. J. Nutr. , 33:277-289.

McCracken, K.J . 1976. A futile energy cycle in adult rats given a low-protein diet at
high levels of energy intake? Proc. Nutr. Soc. , 35:59A.

McCracken, K.J. and A. McAllister 1984. Energy metabolism and body composition
of young pigs given low-protein diets. Br. J. Nutr., 51:225-234.

McCracken, K.J. and HI. Barr 1982. Energy balance and body fat changes in young
‘cafeteria’-fed rats kept at 24°C. J. Physiol., 330:69P-70P.

McCracken, K.J. and H.G. Barr 1985. Reply to letter by Rothwell and Stock. Br. J.
Nutr., 53:192.

McKey, D. 1979. The distribution of secondary compounds within plants. In Herbi-
vores: Their Interaction with Secondary Plant Metabolites, pp. 56-134, eds.
G.A. Rosenthal and D.H. Janzen. New York: Academic Press.

McNab, B.K. 1963. A model of the energy budget of a wild mouse. Ecology,
44(3):521-532.

 

 

 

157

McNab, BK. 197 8. Energetics of arboreal folivores: physiological problems and
ecological consequences of fading on an ubiquitous food supply. In The
Ecology of Arboreal Folivores, pp. 153-162, ed. G.G. Montgomery. Washing-
ton, D.C.: Smithsonian Institution Press.

McNab, B.K. 1980. Food habits, energetics, and the population biology of mammals.
Am. Nat., 116(1):106-124.

McNab, B.K. 1982. Evolutionary alternatives in the physiological ecology of bats. In
Ecology of Bats, pp. 151-200, ed. T.H. Kunz. New York: Plenum Press.

McNab, B.K. 1986. The inﬂuence of food habits on the energetics of eutherian
mammals. Ecolog. Monog., 56(1):1-19.

McNab, B.K. 1992. Energy expenditure: a short history. In Mammalian Energetics.
Interdisciplinary Views of Metabolism and Reproduction, pp. 1-15. , eds. T.E.
Tomasi and T.H. Horton. Ithaca: Comstock Publishing Associates.

McNab, B.K. and P.R. Morrison 1963. Body temperature and metabolism in subspe-
cies of Peronryscus from arid and mesic environments. Ecol. Monogr. , 33:63-
82.

McNaughton, 8.]. 1987. Adaptation of herbivores to seasonal changes in nutrient
supply. In The Nutrition of Herbivores, pp. 391-408, eds. J.B. Hacker and
J.H. Ternouth. Orlando: Academic Press.

McNaughton, S.J., M. Oesterheld, D.A. Frank, and K.J. Williams 1989. Ecosystem
level patterns of primary productivity and herbivory in terrestrial habitats.
Nature, 341:142-144.

McNaughton, S.J., J.L. Tarrants, M.M. McNaughton, and R.H. Davis 1985. Silica
as a defense against herbivory and a growth promoter in African grasses.
Ecology, 66:528-535.

McNeil, S. and T.R.E. Southwood 1978. The role of nitrogen in the development of
insect/plant relationships. In Biochemical Aspects of Plant and Animal
Coevolution, pp. 77-98, ed. J.B. Harbome. London: Academic.

Meyer, M.W. and W.H. Karasov 1989. There is no simple dichotomy between
digestibility-reduction and toxicity in action of plant secondary chemicals. Bull.
Ecol. Soc. Amer. , 70:203.

Milewski, A.V., T.P. Young, and D. Madden 1991. Thorns as induced defenses:
experimental evidence. Oecologia, 86(1):?0—75.

 

158

Miller, BS, and P.R. Payne 1962. Weight maintenance and food intake. J. Nutr.,
78:255-262.

Milton, K. 1979. Factors inﬂuencing leaf choice by howler monkeys: a test of some
hypothesis of food selection by generalist herbivores. Am. Not. , 114(3):362-
378.

Mitchell, H.H. 1934. Balanced diets, net energy value and speciﬁc dynamic effects.
Science, 80:558-561.

Mittler, T.E. 1953. Amino-acids in phloem sap and their excretion by aphids. Nature,
172:207.

Moen, J. and L. Oksanen 1991. Ecosystem trends. Nature, 353:510.

Mole, S., L.G. Butler, and G. Iason 1990a. Defense against dietary tannin in herbi-

vores: a survey for proline rich salivary proteins in mammals. Biochem. Syst.
Ecol., 18:287-293.

Mole, S. J.C. Rogler, C.J. Morel], and LG. Butler 1990b. Herbivore growth
reduction by tannins: use of Waldbauer ratio techniques and manipulation of
salivary protein production to elucidate mechanisms of action. Biochem. Syst.
Ecol., 18:183-197.

Moore, B.J . 1987. The cafeteria diet—an inappropriate tool for studies of therrnogen-
esis. J. Nutr., 117:227-231.

Morrison, D.W. 1978. Foraging ecology and energetics of the frugivorous bat,
Artibeus jarnaicensis. Ecology, 59:716-723.

Murdoch, W.M. 1966. “Community structure, population control, and competi-
tion”—a critique. Am. Nat., 100:219-226.

Myers, J .H. 1988. The induced defense hypothesis: does it apply to the population
dynamics of insects? In Chemical Mediation of Coevolution, pp. 345-365, ed.
K. Spencer. San Diego: Academic Press.

Nedergaard, J., A. Raasmaja, and B. Cannon 1984. Parallel increases in amount of
(3H)GDP binding and thermogenin antigen in brown-adipose-tissue mitochon-
dria of cafeteria-fed rats. Biochem. biophys. Res. Commun. , 122(3): 1328-
1336.

Negus, N .C. and PJ. Berger 1977. Experimental triggering of reproduction in a
natural population of Microtus montanus. Science, 196: 1230-1231.

 

159

Neumann, R.O. 1902. Experimentelle Beitrage Zur Iehre von dem taglichen Nahr-
ungsbedarf des menschen unter besonderer Berucksichtigung der notwendigen
Eiweissmenge. Arch. Hyg., 45:1-87.

N icholls, D. 1976. Hamster brown-adipose-tissue mitochondria. Purine nucleotide
control of the ion conductance of the inner membrane, the nature of the
nucleotide binding site. Eur. J. Biochem. , 62:223-228.

Nicholls, DJ. 1979. Brown adipose tissue mitochondria. Biochim. biophys. Acta. ,
549: 1-29.

Nicholls, D. and E. Rial 1988. The function of the uncoupling protein in the intact
brown fat cell. In Integration of Mitochondrial Function, pp. 517-526, eds.
J .J . Lemasters, C. Hackenbrock, R. Thurman, and H. Westerhoff. New York:
Plenum Press.

Nicholson, A.J. 1933. The balance of animal populations. J. Anim. Ecol., 2:132-178.
Nowak, RM. 1991. Walker’s Mammals of the World. Baltimore: John Hopkins.

Ohmart, GP. and RB. Edwards 1991. Insect herbivory on Eucalyptus. Ann. Rev.
Entomol., 36:637-657.

Oksanen, L. 1988. Ecosystem organization: mutualism and cybemetics or plain
Darwinian struggle for existence? Am. Nat., 131:424-444.

Oksanen, L., F.D. Fretwell, J. Arruda, and P. Niemela 1981. Exploitation ecosys-
tems in gradients of primary productivity. Am. Nat. , 118:240-261.

Oksanen, L. and T. Oksanen 1981. Lemmings (Lemmas lemmas) and gray-sided voles
(Clethrionomys ruﬁts) in interaction with their resources and predators on
Finnmarks vidda, northern Norway. Rep. Kevo Subarct. Res. Stat. , 17:7-31.

Palo, RT. 1985. Chemical defense in birch: inhibition of digestibility in ruminants by
phenolic extracts. Oecologia, 68: 10-14.

Palo, RT. and Robbins, C.T., eds., 1991. Plant Defenses Against Mammalian
Herbivory. Boca Raton: CRC Press.

Parra, R. 1978. Comparison of foregut and hindgut fermentation in herbivores. In
The Ecology of Arboreal Folivores, pp. 205-229, ed. G.G. Montgomery.
Washington: Smithsonian.

Pearson, OR 1954. The daily energy requirements of a wild anna hummingbird.
Condor, 56(6):317-322.

160

Petrusewicz, K. and A. Macfadyen 1970. Productivity of Terrestrial Animals: Princi-
ples and Methods. Oxford: Blackwell Scientiﬁc Publications.

Phillipson, J. 1975. Rainfall, primary production and ‘carrying capacity’ of Tsavo
National Park (East), Kenya. E. Afr. erdl. J., 13:171-201.

Raup, MJ. and RF. Denno 1983. Leaf age as a predictor of herbivore distribution
and abundance. In Variable Plants and Herbivores in Natural and Managed
Systems, pp. 91-124, eds. R.F. Denno and M.S. McClure. New York:
Academic Press.

Rayor, LS. 1985. Effects of habitat quality on growth, age of ﬁrst reproduction, and
dispersal in Gunnison’s prairie dogs (Cynomys gunnisont). Can. J. Zool. ,
63:2835-2840.

Reese, LC. and SD. Beck 1978. Inter-relationships of nutritional indices and dietary
moisture in the black cutworm (Agrotis ipsilon) digestive efﬁciency. J. Insect
Physiol. , 24:473-479.

Reichardt, P.B., J.P. Bryant, T.P Clausen, and GD. Wieland 1984. Defense of
winter-dormant Alaska paper birch against snowshoe hares. Oecologia, 65:58-
69.

Rhoades, D.F. 1983. Herbivore population dynamics and plant chemistry. In Variable
Plants and Herbivores in Natural and Managed Systems, pp. 155-220, eds.
R.F. Demo and M.S. McClure. New York: Academic Press.

Rhoades, D.F. 1985 . Offensive-defensive interactions between herbivores and plants:
their relevance in herbivore population dynamics and ecological theory. Am.
Nat., 125:205-238.

Rhoades, D.F. and R.G. Cates 1976. Toward a general theory of plant antiherbivore
chemistry. In Recent Advances in Phytochemistry, vol. 10, pp. 168-213, eds.
J.W. Wallace and R.L. Mansell. New York: Plenum Press.

Ricklefs, R.E. and K.K. Matthew 1982. Chemical characteristics of foliage of some
deciduous trees in southeastern Ontario. Can. J. Bot. , 60:2037-2045.

Robbins, C.T., A.E. Hagerman, P.J. Austin, C. McArthur, and T.A. Hanley 1991.
Variation in mammalian physiological responses to a condensed tannin and its
ecological implications. J. Mamm. , 72(3):480-486.

Robbins, C.T., T.A. Hanley, A.E. Hagerman, O. Hjeljord, D.L. Baker, C.C.
Schwartz, and W.W. Mautz 1987. Role of tannins in defending plants against
ruminants: reduction in protein availability. Ecology, 68(1):98-107.

 

161

Rodgers, W.A. 1976. Seasonal diet preferences of impala from South East Tanzania.
E. Afr. Wildl. J., 14:331-333.

Root, RB. 1973. Organization of a plant-arthropod association in simple and diverse
habitats: the fauna of collard (Brassica oleracea). Ecol. Monogr. , 43:95-120.

Rosenthal, G.A. and EA. Bell 1979. Naturally occurring, toxic nonprotein amino
acids. In Herbivores: Their Interaction with Secondary Plant Metabolites, pp.
353-385, eds. G.A. Rosenthal and D.H. Janzen. New York: Academic Press.

Rothwell, NJ. and M.J. Stock 1979. A role for brown adipose tissue in diet-induced
thermogenesis. Nature, 281:31-35.

Rothwell, NJ. and M.J. Stock 1981a. Thermogenesis: comparative and evolutionary
considerations. In The Body Weight Regulatory System: Normal and Disturbed
Mechanisms, pp. 335-343, eds. L.A. Ciofﬁ, W.P.T. James, and TB. Van
Itallie. New York: Raven Press.

Rothwell, NJ. and M.J. Stock 1981b. Rothwell and Stock reply. Nature, 289:699-
700.

Rothwell, NJ. and M .J . Stock 1983. Luxuskonsumption, diet-induced thermogenesis
and brown fat: the case in favor. Clin. Science, 64: 19-23.

Rothwell, NJ. and M.J. Stock 1985a. Is diet-induced thermogenesis an experimental
artefact? Br. J. Nutr., 53:191.

Rothwell, NJ. and M.J. Stock 1985b. Biological distribution and signiﬁcance of
brown adipose tissue. Comp. Biochem. Physiol. , 82A(4):745-751.

Rothwell, NJ. and M.J. Stock 1985c. Thermogenic capacity and brown adipose
tissue activity in the common marmoset. Comp. Biochem. Physiol. ,
81A(3):683-686.

Rothwell, NJ. and M.J. Stock 1986a. Brown adipose tissue and diet-induced thermo-
genesis. In Brown Adipose Tissue, pp. 269-298, eds. P. Trayhum and D.G.
Nicholls. London: Edward Arnold.

Rothwell, NJ. and M.J. Stock 1986b. Inﬂuence of environmental temperature on
energy balance, diet-induced thermogenesis and brown fat activity in ‘cafete-
ria’-fed rats. Br. J. Nutr., 56:123-129.

Rothwell, NJ. and MJ . Stock 1987a. Diet-induced thermogenesis: concepts and
mechanisms. J. Obes. and Wt. Reg., 6:147-161.

 

162

Rothwell, NJ. and M.J. Stock 1987b. Effect of environmental temperature on energy
balance and thermogenesis in rats fed normal or low protein diets. J. Nutr. ,
117:833-837.

Rothwell, NJ. and M.J. Stock 1987c. Inﬂuence of carbohydrate and fat intake on
diet-induced thermogenesis and brown fat activity in rats fed low protein diets.
J. Nutr., 117:1721-1726.

Rothwell, NJ. and M.J. Stock 1988. The cafeteria diet as a tool for studies of
thermogenesis. J. Nutr. , 118:925-928.

Rothwell, NJ ., M.E. Saville, and M.J. Stock 1982. Effects of feeding a “cafeteria”
diet on energy balance and diet-induced thermogenesis in four strains of rat. J.
Nutr., 112:1515-1524.

Rothwell, NJ ., M.J. Stock, and RS. Tyzbir 1982. Energy balance and mitochondrial
function in liver and brown fat of rats fed “cafeteria” diets of varying protein
content. J. Nutr., 112:1663-1672.

Rothwell, NJ ., M.J. Stock, and RS. Tyzbir 1983a. Mechanisms of thermogenesis
induced by low protein diets. Metabolism, 32(3):257-261.

Rothwell, NJ ., M.J. Stock, and BF. Warwick 1983b. The effect of high fat and
high carbohydrate cafeteria diets on diet-induced thermogenesis in the rat. Int.
J. Obesity, 7:263-270.

Rubner, M. 1902. Die Gesetze des Energieverbrauchs bei der Emahrung Deuticke,
Leipzig.

Scholander, RR, R. Hock, V. Walters, and L. Irving. 1950a. Adaptation to cold in
arctic and tropical mammals and birds in relation to body temperature, insula-
tion, and basal metabolic rate. Biol. Bull. , 99:259-271.

Scholander, RR, R. Hock, V. Walters, F. Johnson, and L. Irving. 1950b. Heat
regulation in some arctic and tropical mammals and birds. Biol. Bull. , 99:237-
258.

Scholander, P.F., V. Walters, R. Hock, and L. Irving. 1950c. Body insulation of
some arctic and tropical mammals and birds. Biol. Bull. , 99:225-236.

Schultz, J .C., PJ. Nothnagle, and LT. Baldwin 1982. Seasonal and individual
variation in leaf quality of two northern hardwoods tree species. Am. J. Bot.,
69:753-759.

 

163

Schwartz, C.C. and N.T. Hobbs 1985. Forage and range evaluation. In Bioenergetics
of Wild Herbivores, pp. 25-51, eds. RJ. Hudson and R.G. White. Boca
Raton: CRC Press.

Schwartz, C.C., W.L. Regelin, and LG. Nagy 1980. Deer preferences for juniper
forage and volatile treated foods. J. Wildl. Manage, 44(1): 1 14-120.

Scriber, J.M. 1977. Limiting effects of low leaf-water content on the nitrogen
utilization, energy budget and larval growth of Hyalophora cecropia (Lepidop-
tera: Satumiidae). Oecologia, 28:269-287.

Scriber, J .M. 1984a. Host-plant suitability. In Chemical Ecology of Insects, pp. 160-
202, eds. W.J. Bell and RT. Carde. London: Chapman and Hall Ltd.

Scriber, J .M. 1984b. Evolution of feeding specialization, physiological efﬁciency, and
host races in selected Papilionidae and Satumiidae. In Variable Plants and
Herbivores in Natural and Managed Systems, pp. 373-412, eds. R. F. Denno
and M.S. McClure. New York: Academic Press.

Scriber, J.M. and F. Slansky, Jr. 1981. The nutritional ecology of immature insects.
Ann. Rev. Entomol., 26:183-211.

Shenk, 1.8., RC. Elliot, and J .W. Thomas 1970. Meadow vole nutrition studies with
semisynthetic diets. J. Nutr., 100:1437-1446.

Shkolnik, A. and K. Schmidt-Nielsen 1976. Temperature regulation in hedgehogs
from temperate and desert environments. Physiol. Zool. , 49:56-64.

Sibly, RM. 1981. Strategies of digestion and defecation. In Physiological Ecology:
an Evolutionary Approach to Resource Use, pp. 109-139, eds. C.R. Townsend
and P. Calow. Oxford: Blackwell Scientiﬁc Publications.

Sinclair, A.R.E. 1974. The natural regulation of buffalo populations in East Africa.
E. Afr. Wildl. J., 12:291-311.

Sinclair, A.R.E. 1975. The resource limitation of trophic levels in tropical grassland
ecosystems. J. Anim. Ecol., 44:497-520.

Sinclair, A.R.E. 1979. The eruption of ruminants. In Serengeti: Dynamics of an
Ecosystem, pp. 82-103, eds. A.R.E. Sinclair and M. Norton-Grifﬁths. Chica-
go: University of Chicago Press.

Slansky, F., Jr. and P. Feeny 1977. Stabilization of the rate of nitrogen accumulation
by larvae of the cabbage butterﬂy on wild and cultivated food plants. Ecology,
47 :209-228.

 

 

164

Smith, J.N.M., CJ. Krebs, A.R.E. Sinclair, and R. Boonstra 1988. Population
biology of snowshoe hares. II. Interactions with winter food plants. J. Anim.
Ecol., 57:269-286.

Spencer, K.C., ed., 1988. Chemical Mediation of Coevolution. San Diego: Academic
Press.

Stirling, LL. and M.J. Stock 1968. Metabolic origins of thermogenesis induced by
diet. Nature, 220:801-802.

Swain, T. 1979. Tannins and lignins. In Herbivores: Their Interaction with Second-
ary Plant Metabolites, pp. 657-682, eds. G.A. Rosenthal and D.H. Janzen.
New York: Academic Press.

Swick, R.W. and NJ. Benevenga 1977. labile protein reserves and protein turnover.
J. Dairy Science, 60(4):505-515.

 

Swick, R.W. and CL. Gribskov 1983. The effect of dietary protein levels on diet-
induced thermogenesis in the rat. J. Nutr. , 113:2289-2294.

Swick, R.W. and M.F. Henningﬁeld 1989. Changes in the number of GDP binding
sites on brown adipose tissue (BAT) mitochondria and its uncoupling protein.
In Hormones, Thermogenesis, and Obesity, pp. 117-127, eds. H. Lardy and F.
Stratman. New York: Elsevier.

Swick, A.G. and R.W. Swick 1986. Enhanced thermogenic response to a diet low but
adequate in protein persists in older rats. Am. J. Physiol. , 251:FA38-E441.

Tabashnik, BE. and ES. Slansky, Jr. 1987. Nutritional ecology of forb foliage-
chewing insects. In Nutritional Ecology of Insects, Mites, Spiders, and Related
Invertebrates, pp. 71-103, eds. F. Slansky, Jr. and J .G. Rodriquez. New
York: John Wiley & Sons.

Tahvanainen, J ., E. Helle, R. J ulkunen-Titto, and A. Lavola 1985. Phenolic com-
pounds of willow bark as deterrents against feeding by mountain hare. Oeco-
logia, 65:319-323.

Tahvanainen, J .O. and RB. Root 1972. The inﬂuence of vegetational diversity on the
populational ecology of a specialized herbivore, Phyllotreta cruciferae (Cole-
optera: Chrysomelidae). Oecologia, 10:321-346.

Taitt, M.J. and CJ. Krebs 1981. The effect of extra food on small rodent popula-
tions. 11. Voles (Microtus townsendir). J. Anim. Ecol.50:125-137.

165

Taitt, M.J. and Q]. Krebs 1983. Predation, cover, and food manipulations during a
spring decline in Microtus townsendii. J. Anim. Ecol. , 52:837-848.

Thomas, D.W. 1984. Fruit intake and energy budgets of frugivorous bats. Physiol.
Zool., 57(4):457-467.

Thomas, D.W., C. Samson, and J .M. Bergeron 1988. Metabolic costs associated with
the ingestion of plant phenolics by Microtus pennsylvanicus. J. Mamm. ,
69(3):512-515.

Tobin, G., G. Armitage, and G.R. Hervey 1981. Energy expenditure during cafeteria
feeding in rats. Int. J. Obesity, 5:379-384.

Trayhum, P. and R.E. Milner. 1987. Mechanisms of thermogenesis: brown adipose
tissue. J. Obesity and Wt. Reg., 6:147-161.

Tulp, O.L., P.P. Krupp, E. Danforth, Jr. and ES. Horton 1979. Charactristics of
thyroid function in experimental protein malnutrition. J. Nutr., 109:1321-1332.

Uresk, D.W. and P.L. Sims 1975 . Inﬂuence of grazing on crude protein content of
blue grama. J. Range Mgrnt., 28:370—37 1.

Van Alstyne, K.L. 1988. Herbivore grazing increases polyphenolic defenses in the
intertidal brown alga Fucus distichus. Ecology, 69(3):655-663.

Van Hoven, W. and E. A. Boomker 1985. Digestion. In Bioenergetics of Wild
Herbivores, pp. 103-120, eds. RJ. Hudson and R.G. White. Boca Raton:
CRC Press.

Van Soest, PJ. 1968. Structural and chemical characteristics which limit the nutritive
value of forages. In Forage Economics—Quality, pp. 63-76, eds. C.M.
Harrison, M. Stelly, and S.A. Breth. Madison: American Society of Agrono-
my.

Van Soest, PJ. 1982. Nutritional Ecology of the Ruminant. Corvallis: O & B Books.

Vander Tuig, J .G., J. Kemer, and DR. Romsos 1985. Hypothalamic obesity, brown
adipose tissue, and sympathoadrenal activity in rats. Am. J. Physiol. ,
248(11):E607-E617.

Vander Tuig, J .G. and DR. Romsos 1984. Effects of dietary carbohydrate, fat, and
protein on norepinephrine turnover in rats. Metabolism, 33(1):26-33.

166

Vaughan, M.R. and L.B. Keith 1981. Demographic response of experimental snow-
shoe hare populations to overwinter food shortage. J. Wildl. Manage. , 45 :354-
380.

Vaughan, T.A. and NJ. Czaplewski 1985. Reproduction in Stephens’ woodrat: the
wages of folivery. J. Mamm., 66:429-443.

Vickery, W.L. 1984. Optimal diet models and rodent food consumption. Anim.
Behav. , 32:340-348.

Voit, E. 1901. Uber die Grosse des Energiebedarfs der Tiere in Hungerzustande.
Ztschr. Biol. 41:113-154.

Waldbauer, GP. 1968. The consumption and utilization of food by insects. In
Advances in Insect Physiology, pp. 229-288, eds. J.W.L. Beament, J.B.
Treherne, and VB. Wigglesworth. London: Academic Press.

Waterlow, J .C. 1985. What do we mean by adaptation? In Nutritional Adaptation in
Man, pp. l-11, eds. K. Blaxter and LC. Waterlow. London: John Libbey.

Watt, AD. 1987. The effect of shoot growth stage of Pinus contorta and Pinus

sylvestris on the growth and survival of Panolis ﬂammea larvae. Oecologia,
72:429-433.

Watt, AD. 1989. The chemical composition of pine foliage in relation to the popula-
tion dynamics of the pine beauty moth, Panolis flammea, in Scotland. Oeco-
logia, 78:251-258.

Webster, AJ .F. 1981. The energetic efﬁciency of metabolism. Proc. Nutr. Soc. ,
40:121-128.

Webster, A.J.F. 1983. Energetics of maintenance and growth. In Mammalian Ther-
mogenesis, pp. 178-207, eds. L. Girardier and M.J. Stock. London: Chapman
and Hall.

White, R. G. 1983. Foraging patterns and their multiplier effects on productivity of
northern ungulates. Oikos 40:377-384.

White, T.C.R. 1978. The importance of a relative shortage of food in animal ecolo-
gy. Oecologia, 33:71—86.

Wilkinson, L. 1989. SYSTAT: The System for Statistics for the PC. Evanston:
SYSTAT, Inc.

 

167

Windberg, LA. and L.B. Keith 1976. Snowshoe hare population response to artiﬁcial
high densities. J. Mamm., 57:523-553.

Wunder, BA. 1985 . Energetics and thermoregualtion. In Biology of New World
Microtus, pp. 812-844, ed. R.H. Tamarin. Am. Soc. Mamm. Spec. Pub. No.
8.

Wunder, B.A., D.S. Dobkin, and RD. Gettinger 1977. Shifts of thermogenesis in the
prairie vole (Microtus ochrogaster). Oecologia, 29:11-26.

Young, J.B., L.N. Kaufman, M.E. Saville, and L. Landsberg 1985. Increased
sympathetic nervous system activity in rats fed a low-protein diet. Am. J.
Physiol. , 248(17):R627-R637.

Young, J.B., E. Saville, NJ. Rothwell, M.J. Stock, and L. Landsberg 1982. Effect
of diet and cold exposure on norepinephrine turnover in brown adipose tissue
of the rat. J. Clin. Invest., 69:1061-1071.

Young, T.P. 1987. Increased thorn length in Acacia depranolobium—an induced
response to browsing. Oecologia, 71(3):436-438.

Zangerl, A.R. and M.R. Berenbaum 1990. Furanocoumarin induction in wild parsnip:
genetics and populational variation. Ecology, 71(5): 1933-1940.

Zhang, X. and LS. States 1991. Selective herbivory of ponderosa pine by Abert

squirrels: a re-examination of the role of terpenes. Biochem. Syst. Ecol.
19(2):111-115.

 

”llllllllllllllf